The Lancet ~ Volume 373, Issue 9657, Pages 1-94 (3 January 2009-9 January 2009)

Editorial Oral health: prevention is key Oral health is a neglected area of global health and has traditionally regist...

Author: The Lancet

22 downloads 884 Views 4MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Editorial

Oral health: prevention is key Oral health is a neglected area of global health and has traditionally registered low on the radar of national policy makers. The reasons for this situation are complex and varied. In many countries oral health is not included in national health surveys. And, if data are collected, it is usually in isolation from the context of general health. Moreover, in some cultures, oral health is neglected because teeth are seen as expendable. Dentists have also taken little interest in advocacy to promote good oral health, preferring to treat rather than prevent oral diseases. And, because poor oral health aﬀects morbidity more than mortality, governments have viewed oral conditions as less important than other, more life-threatening diseases. Yet, globally, the burden of major oral diseases and conditions is high. Dental caries are one of the most common chronic diseases worldwide. 90% of people have had dental problems or toothache caused by caries, and in low-to-middle income countries most caries remains untreated. Severe periodontitis aﬀects 5–15% of most populations. Oral cancer is the eighth most common cancer worldwide and the most common in men in southeast Asia. And 40–50% of people who are HIV positive have oral fungal, bacterial, or viral infections. Access to oral care is a global problem, particularly in low-to-middle income countries. The workforce available to treat the most common oral health problems—dentists—are in short supply in these nations. Whereas countries such as Germany and the UK have one dentist per 1000 population, low-income and middle-income countries have one dentist per 50 000 people, and in some sub-Saharan African countries the ratio is one per 900 000 people. Dentists also cluster in cities where populations that can aﬀord treatment usually live, leaving rural areas deprived of even the most basic emergency dental care. But training more dentists and building dental clinics—the western curative model of care—is costly and unrealistic in most low-income and middle-income countries. Prevention of oral disease is therefore key, largely possible, and should be a routine part of other health professionals’ work. What can be done? The daily use of ﬂuoride is the most cost-eﬀective, evidence-based approach to reduce dental decay. Water or salt ﬂuoridation are possible populationwide approaches but their implementation depends on www.thelancet.com Vol 373 January 3, 2009

the development and infrastructure of the country as well as political will and community acceptance. Promoting the daily use of eﬀective ﬂuoride toothpaste is a more realistic strategy but its cost prohibits its widespread use in many low-income and middle-income countries. Governments can remove taxes on ﬂuoride toothpaste, which in some countries represent up to 50% of the product’s price, and they can work with manufacturers to produce lower cost toothpaste. In the Philippines, for example—where 97% of schoolchildren aged 6 years have dental caries—a programme that combines the promotion of daily handwashing with soap, tooth brushing with subsidised ﬂuoride toothpaste, and twice-yearly deworming, is proving eﬀective, aﬀordable, and sustainable at US$0·56 per child per year. Policies that address the risk factors for oral diseases, such as intake of sugars and tobacco use, can also be implemented, especially because these moves will help reduce chronic diseases. Oral diseases and chronic diseases, such as cardiovascular diseases, cancer, chronic respiratory diseases, and diabetes share many common risk factors. In 2007, a World Health Assembly resolution called for oral health to be integrated into chronic disease prevention programmes. Promoting good oral health could also help countries to achieve child-related development goals. Caries can negatively aﬀect a child’s ability to eat, sleep, and do school work. Preliminary studies have suggested that dental caries and related pain and sepsis might contribute to undernutrition and low weight and height in children in developing countries. In developed countries, studies show that when dental caries are treated, children start to put on weight and thrive. Oral pain is also one of the most common reasons for school absenteeism. Preventing oral disease is important and achievable. Evidence-based, simple, and cost-eﬀective preventive approaches exist, but they need to be rigorously promoted and implemented. Professionally, health workers, including physicians, nurses, paediatricians, and pharmacists can all deliver prevention messages about the use of ﬂuoride and the risk factors for oral disease. Politically, commitment is needed to integrate oral disease prevention into programmes to prevent chronic diseases and into public-health systems. Good oral The Lancet health should be everybody’s business.

The printed journal includes an image merely for illustration

For more on the Philippines school programme see Development & Coorperation 2008; 49: 8–12; http://www. inwent.org/ez/articles/082726/ index.en.shtml

1

Editorial

A favourable (molecular) signal for personalised medicine

Science Photo Library

With all the important advances in genomic sequencing, one can sometimes forget that molecular oncologists have been busy for decades to uncover what happens at the molecular level in oncogenesis. They have found mutation routes that are of more immediate practical and clinical value than knowing the whole genome. A mid-December meeting of the Oncologic Drugs Advisory Panel of the US Food and Drug Administration (FDA) discussed the retrospective analysis of biomarkers from completed trials to support drug-company labelling claims for anticancer agents in subgroups of patients. Provisos were: the analysis needs to be hypothesisdriven, adequately powered (presumably, in retrospect), and prespeciﬁed for tissue collection and data analysis. For example, the Panel used data for the monoclonal antibodies cetuximab and panitumumab, which inhibit the EGF receptor (EGFR) in metastatic colorectal cancer. About 40% of these patients have a KRAS gene mutation, and respond poorly to such antibodies. The KRAS protooncogene product lies downstream of EGFR, and the

mutation frees the receptor from normal regulation in cell-growth control. One important implication for this pair of monoclonals is that prospective trials in patients with metastatic colorectal cancer who harbour wildtype KRAS (ie, the non-mutated form) might no longer be ethical. EMEA, the European Medicines Agency, has already restricted use of cetuximab and panitumumab to such patients. The FDA’s decision is a major step towards the idea of personalised medicine, a concept that genomic research hails as one of its reasons for existence but one that has come from molecular oncology. Being able to identify potential responders or non-responders means that patients can be selected for therapies that could have side-eﬀects and, if they are one of the new biologicals, are expensive. This step by the FDA means that the wealth of data harboured in completed trials in cancer, and other diseases, could be plundered with more conﬁdence, for the beneﬁt of patients and of those who have to pay for expensive treatments. The Lancet

Getty Images

America’s commitment to global health

For the Institute of Medicine (IOM) report see www.iom.edu/ usandglobalhealth

2

Over the past decade, the US Government has spent record amounts on global health: in 2008 spending peaked at US$7·5 billion. Still, the share of the country’s gross national income allocated to development aid is only about one-third of the target set by the UN’s Millennium Development Goals. More than 70% of this money is aimed at AIDS programmes, even though chronic and non-communicable diseases account for more than half of all deaths in low-income and middleincome countries. Perhaps not surprisingly, then, mid-December’s US Institute of Medicine (IOM) report was devoted to the global-health commitment of the government under the incoming administration of President-elect Barack Obama. Created by an independent committee of four US government agencies and ﬁve private foundations, the report discusses several key recommendations. First, the new administration should highlight health as a pillar of US foreign policy. Second, a White House Interagency Committee on Global Health be formed and chaired by a White House global-health tsar, who

will serve as deputy assistant to the President. Third, a balanced aid portfolio should be created to cover diverse global-health issues including AIDS, malaria, tuberculosis, children’s and women’s health, nutrition, family planning and reproductive health, chronic and non-communicable disease, injury, and health systems. Fourth, the US Government should enhance partnerships with national governments, support the leadership of WHO, and fund research for health. Last, the budget for US global-health assistance should be increased annually and doubled to $15 billion by the year 2012. Before the second IOM report on the US commitment to global health is released in April, 2009, the new Obama administration must seize this opportunity to show leadership and take the initiative in expressing its determination to save lives overseas. As Bill Gates indicated, “sending this critical message at this critical time” will strengthen the country’s international relations and help to restore the global credibility of, and respect for, the US Government and its people. The Lancet www.thelancet.com Vol 373 January 3, 2009

Editorial

A favourable (molecular) signal for personalised medicine

Science Photo Library

With all the important advances in genomic sequencing, one can sometimes forget that molecular oncologists have been busy for decades to uncover what happens at the molecular level in oncogenesis. They have found mutation routes that are of more immediate practical and clinical value than knowing the whole genome. A mid-December meeting of the Oncologic Drugs Advisory Panel of the US Food and Drug Administration (FDA) discussed the retrospective analysis of biomarkers from completed trials to support drug-company labelling claims for anticancer agents in subgroups of patients. Provisos were: the analysis needs to be hypothesisdriven, adequately powered (presumably, in retrospect), and prespeciﬁed for tissue collection and data analysis. For example, the Panel used data for the monoclonal antibodies cetuximab and panitumumab, which inhibit the EGF receptor (EGFR) in metastatic colorectal cancer. About 40% of these patients have a KRAS gene mutation, and respond poorly to such antibodies. The KRAS protooncogene product lies downstream of EGFR, and the

mutation frees the receptor from normal regulation in cell-growth control. One important implication for this pair of monoclonals is that prospective trials in patients with metastatic colorectal cancer who harbour wildtype KRAS (ie, the non-mutated form) might no longer be ethical. EMEA, the European Medicines Agency, has already restricted use of cetuximab and panitumumab to such patients. The FDA’s decision is a major step towards the idea of personalised medicine, a concept that genomic research hails as one of its reasons for existence but one that has come from molecular oncology. Being able to identify potential responders or non-responders means that patients can be selected for therapies that could have side-eﬀects and, if they are one of the new biologicals, are expensive. This step by the FDA means that the wealth of data harboured in completed trials in cancer, and other diseases, could be plundered with more conﬁdence, for the beneﬁt of patients and of those who have to pay for expensive treatments. The Lancet

Getty Images

America’s commitment to global health

For the Institute of Medicine (IOM) report see www.iom.edu/ usandglobalhealth

2

Over the past decade, the US Government has spent record amounts on global health: in 2008 spending peaked at US$7·5 billion. Still, the share of the country’s gross national income allocated to development aid is only about one-third of the target set by the UN’s Millennium Development Goals. More than 70% of this money is aimed at AIDS programmes, even though chronic and non-communicable diseases account for more than half of all deaths in low-income and middleincome countries. Perhaps not surprisingly, then, mid-December’s US Institute of Medicine (IOM) report was devoted to the global-health commitment of the government under the incoming administration of President-elect Barack Obama. Created by an independent committee of four US government agencies and ﬁve private foundations, the report discusses several key recommendations. First, the new administration should highlight health as a pillar of US foreign policy. Second, a White House Interagency Committee on Global Health be formed and chaired by a White House global-health tsar, who

will serve as deputy assistant to the President. Third, a balanced aid portfolio should be created to cover diverse global-health issues including AIDS, malaria, tuberculosis, children’s and women’s health, nutrition, family planning and reproductive health, chronic and non-communicable disease, injury, and health systems. Fourth, the US Government should enhance partnerships with national governments, support the leadership of WHO, and fund research for health. Last, the budget for US global-health assistance should be increased annually and doubled to $15 billion by the year 2012. Before the second IOM report on the US commitment to global health is released in April, 2009, the new Obama administration must seize this opportunity to show leadership and take the initiative in expressing its determination to save lives overseas. As Bill Gates indicated, “sending this critical message at this critical time” will strengthen the country’s international relations and help to restore the global credibility of, and respect for, the US Government and its people. The Lancet www.thelancet.com Vol 373 January 3, 2009

Comment

When childhood dies

www.thelancet.com Vol 373 January 3, 2009

summarise the evidence base on interventions that are available to prevent child maltreatment and mitigate its eﬀects; and ﬁnally Richard Reading and colleagues6 acknowledge the complexity of social inﬂuences and policy involving children, and argue in favour of a human-rights approach to child maltreatment. Too often, the safety of children is debated in the polarising light of litigation or political division. Whilst this outcome is inevitable because of the public outrage child maltreatment engenders, a damaging consequence is that the evidence surrounding child neglect and abuse often fails to inﬂuence serious policy discussion. It is this marginalisation of the science of child maltreatment that we are seeking to reverse. We would like to extend warm thanks to Patricia Hamilton of the Royal College of Paediatrics and Child Health for supporting this Series, and to the authors and peer reviewers of the accompanying papers and Comments for their unstinting work. This Lancet Series will unfortunately not halt the blight of child abuse, because the phenomenon is too common, too surreptitious, and too deeply rooted in deprivation and other social ills—but we nonetheless hope to raise awareness of the scientiﬁc evidence that is available, and indeed essential, to guide paediatricians and other professionals in their practice with children who might have been abused; and to help bring a new logic and clarity to public debate about this contentious area.

Published Online December 3, 2008 DOI:10.1016/S01406736(08)61701-8 See Series page 68

PA Photos

Few topics are more emotive than child maltreatment. Human societies place a high value on the richness and diversity that parents contribute to their children’s upbringing, and bystanders and professionals alike will hesitate to intervene in or undermine the parent–child relationship unless there is deﬁnitive evidence that serious harm is being done. Yet the grating litany of child abuse by parents and other carers is familiar to all. The forced child labour endured by children in Victorian England more than a century ago remains common in developing countries today, and in the developed world formal child-protection systems document a depressing range of maltreatment, from deliberate neglect to the vilest sexual abuse. This cruel treatment can lead to serious physical and mental-health consequences, and risks creating a damaging intergenerational cycle of maltreatment, disadvantage, and ill health. Children, the most precious and vulnerable members of our societies, deserve closer attention to their care and education and better protection against abuse. Child maltreatment can be especially diﬃcult to ascertain—after all, the victim might be unwilling or unable to provide an accurate account of abuse, the perpetrator bent on concealment, witnesses silent, clinical signs equivocal, and professionals reluctant to act owing to uncertainty about the existence or severity of abuse, or concern that the treatment of abuse will itself lead to undue disruption or damage to a child’s upbringing. When severe child abuse culminates in death, however, as in the tragic recent cases of Victoria Climbié and Baby P in the UK,1,2 the law, news media, and public can be expected to cast a harsh and unforgiving spotlight on social workers and doctors who might be perceived or portrayed as having missed opportunities to help the victims. It is to clinicians and other professionals responsible for caring for children that The Lancet‘s Child Maltreatment Series is aimed, with the intention of providing them with a rigorous and up-to-date summary of scientiﬁc evidence and conceptual work on this complex and demanding topic. Ruth Gilbert and colleagues3 begin by reviewing the literature on the classiﬁcation and worldwide burden of child maltreatment, and in a second paper4 discuss the methodology and challenges involved in identifying abuse. Harriet MacMillan and coauthors5

3

Comment

Richard Turner, Richard Horton

4

The Lancet, London NW1 7BY, UK 1

2 3

Laming. The Victoria Climbié inquiry: report of an inquiry by Lord Laming. January, 2003. http://www.victoria-climbie-inquiry.org.uk/ﬁnreport/ ﬁnreport.htm (accessed Nov 13, 2008). BBC News Channel. Men found guilty of baby’s death. Nov 11, 2008. http:// news.bbc.co.uk/1/hi/england/london/7706598.stm (accessed Nov 17, 2008). Gilbert R, Spatz Widom C, Browne K, Fergusson D, Webb E, Janson S. Burden and consequences of child maltreatment in high-income countries. Lancet 2008; published online Dec 3. DOI:10.1016/S0140-6736(08)61706-7.

5

6

Gilbert R, Kemp A, Thoburn J, et al. Recognising and responding to child maltreatment. Lancet 2008; published online Dec 3. DOI:10.1016/S01406736(08)61707-9. MacMillan HL, Wathen CN, Barlow J, Fergusson DM, Leventhal JM, Taussig HN. Interventions to prevent child maltreatment and associated impairment. Lancet 2008; published online Dec 3. DOI:10.1016/S01406736(08)61708-0. Reading R, Bissell S, Goldhagen J, et al. Promotion of children’s rights and prevention of child maltreatment. Lancet 2008; published online Dec 3. DOI:10.1016/S0140-6736(08)61709-2.

The spurious advance of antipsychotic drug therapy Published Online December 5, 2008 DOI:10.1016/S01406736(08)61765-1

Clinicians are familiar with studies that claim to show major advances in therapy. They tend to greet early reports of such advances with a touch of scepticism and wait, usually for at least 10 years, for a raft of independent studies that show that the advance is genuine and not just another minor ripple in the treatment stream. In The Lancet today, Stefan Leucht and colleagues1 deviate from this pattern by suggesting that what was seen as an advance 20 years ago—when a new generation of antipsychotic drugs with additional beneﬁts and fewer adverse eﬀects was introduced2—is now, and only now, seen as a chimera that has passed

Science Photo Library

See Articles page 31

4

spectacularly before our eyes before disappearing and leaving puzzlement and many questions in its wake. Leucht and colleagues’ analysis of ten outcomes from 150 randomised trials, supported by some powerful studies,3–5 shows that the name “second-generation antipsychotics” is inaccurate. This group of drugs is in fact a heterogeneous mix of compounds, with some superior to others. Antipsychotic drugs diﬀer in their potencies and have a wide range of adverse-eﬀect proﬁles, with nothing that clearly distinguishes the two major groups. Importantly, the second-generation drugs have no special atypical characteristics that separate them from the typical, or ﬁrst-generation, antipsychotics. As a group they are no more eﬃcacious, do not improve speciﬁc symptoms, have no clearly diﬀerent side-eﬀect proﬁles than the ﬁrst-generation antipsychotics, and are less cost eﬀective.6–8 The spurious invention of the atypicals can now be regarded as invention only, cleverly manipulated by the drug industry for marketing purposes and only now being exposed. But how is it that for nearly two decades we have, as some have put it,9 “been beguiled” into thinking they were superior? Leucht and co-workers provide some clues. Of 150 trials in their meta-analysis, in 95 the secondgeneration antipsychotic was compared with the high-potency ﬁrst-generation antipsychotic haloperidol. The use of haloperidol as the ﬁrst-generation antipsychotic in these trials means that they were biased in favour of the second-generation drugs. This bias has been achieved through several routes—eg, by comparing the second-generation antipsychotic with a high-potency ﬁrst-generation antipsychotic likely to be associated with a high rate of extrapyramidal side-eﬀects. Another obvious way of favouring the second-generation drugs has been to avoid comparison with a medium-potency ﬁrst-generation antipsychotic, because these drugs are www.thelancet.com Vol 373 January 3, 2009

Comment

Richard Turner, Richard Horton

4

The Lancet, London NW1 7BY, UK 1

2 3

Laming. The Victoria Climbié inquiry: report of an inquiry by Lord Laming. January, 2003. http://www.victoria-climbie-inquiry.org.uk/ﬁnreport/ ﬁnreport.htm (accessed Nov 13, 2008). BBC News Channel. Men found guilty of baby’s death. Nov 11, 2008. http:// news.bbc.co.uk/1/hi/england/london/7706598.stm (accessed Nov 17, 2008). Gilbert R, Spatz Widom C, Browne K, Fergusson D, Webb E, Janson S. Burden and consequences of child maltreatment in high-income countries. Lancet 2008; published online Dec 3. DOI:10.1016/S0140-6736(08)61706-7.

5

6

Gilbert R, Kemp A, Thoburn J, et al. Recognising and responding to child maltreatment. Lancet 2008; published online Dec 3. DOI:10.1016/S01406736(08)61707-9. MacMillan HL, Wathen CN, Barlow J, Fergusson DM, Leventhal JM, Taussig HN. Interventions to prevent child maltreatment and associated impairment. Lancet 2008; published online Dec 3. DOI:10.1016/S01406736(08)61708-0. Reading R, Bissell S, Goldhagen J, et al. Promotion of children’s rights and prevention of child maltreatment. Lancet 2008; published online Dec 3. DOI:10.1016/S0140-6736(08)61709-2.

The spurious advance of antipsychotic drug therapy Published Online December 5, 2008 DOI:10.1016/S01406736(08)61765-1

Clinicians are familiar with studies that claim to show major advances in therapy. They tend to greet early reports of such advances with a touch of scepticism and wait, usually for at least 10 years, for a raft of independent studies that show that the advance is genuine and not just another minor ripple in the treatment stream. In The Lancet today, Stefan Leucht and colleagues1 deviate from this pattern by suggesting that what was seen as an advance 20 years ago—when a new generation of antipsychotic drugs with additional beneﬁts and fewer adverse eﬀects was introduced2—is now, and only now, seen as a chimera that has passed

Science Photo Library

See Articles page 31

4

spectacularly before our eyes before disappearing and leaving puzzlement and many questions in its wake. Leucht and colleagues’ analysis of ten outcomes from 150 randomised trials, supported by some powerful studies,3–5 shows that the name “second-generation antipsychotics” is inaccurate. This group of drugs is in fact a heterogeneous mix of compounds, with some superior to others. Antipsychotic drugs diﬀer in their potencies and have a wide range of adverse-eﬀect proﬁles, with nothing that clearly distinguishes the two major groups. Importantly, the second-generation drugs have no special atypical characteristics that separate them from the typical, or ﬁrst-generation, antipsychotics. As a group they are no more eﬃcacious, do not improve speciﬁc symptoms, have no clearly diﬀerent side-eﬀect proﬁles than the ﬁrst-generation antipsychotics, and are less cost eﬀective.6–8 The spurious invention of the atypicals can now be regarded as invention only, cleverly manipulated by the drug industry for marketing purposes and only now being exposed. But how is it that for nearly two decades we have, as some have put it,9 “been beguiled” into thinking they were superior? Leucht and co-workers provide some clues. Of 150 trials in their meta-analysis, in 95 the secondgeneration antipsychotic was compared with the high-potency ﬁrst-generation antipsychotic haloperidol. The use of haloperidol as the ﬁrst-generation antipsychotic in these trials means that they were biased in favour of the second-generation drugs. This bias has been achieved through several routes—eg, by comparing the second-generation antipsychotic with a high-potency ﬁrst-generation antipsychotic likely to be associated with a high rate of extrapyramidal side-eﬀects. Another obvious way of favouring the second-generation drugs has been to avoid comparison with a medium-potency ﬁrst-generation antipsychotic, because these drugs are www.thelancet.com Vol 373 January 3, 2009

Comment

likely to be just as eﬃcacious as the second-generation drug, but less likely than haloperidol to induce Parkinsonism. The picture can be complicated further with high doses of the ﬁrst-generation drug. This approach favours the second-generation antipsychotic because side-eﬀect rates are much lower than with the ﬁrst-generation antipsychotic.10 Moreover, there is often selective publication of trials11–13 that can skew the evidence base in favour of a drug favoured by the investigators. On present evidence from all sources it is diﬃcult not to conclude that the trials of the second-generation antipsychotics seem to be driven more by marketing strategy than to clarify their role for clinicians and patients. This is not to say that all antipsychotic drugs are the same, they are not. Individual responses vary, and so a range of drugs is needed for good clinical practice. So where should we go now? First, the time has come to abandon the terms ﬁrst-generation and second-generation antipsychotics, as they do not merit this distinction. The only second-generation antipsychotic that is obviously better than other drugs in resistant schizophrenia is clozapine,14 and this is a very old drug indeed. Second, clinicians must remember to keep the beneﬁt–risk ratio of each antipsychotic drug in constant perspective because all are associated in diﬀerent ways with serious adverse eﬀects, which should be important outcome measures.13 Finally, it is prudent to remember that although science rules during a drug’s development, the market usurps control once the drug is released for care of patients.

PT declares that he has no conﬂict of interest. TK is involved in updating the schizophrenia guideline for the National Institute for Health and Clinical Excellence, including a review of antipsychotic drugs in the treatment of schizophrenia. 1

2

3

4

5

6

7

8 9

10

11

12

*Peter Tyrer, Tim Kendall Department of Psychological Medicine, Imperial College London, London W6 8RP, UK (PT); and National Collaborating Centre for Mental Health, Royal College of Psychiatrists’ Research Unit, London, UK (TK) [email protected]

13 14

Leucht S, Corves C, Arbter D, Engel RR, Li C, Davis JM. Second-generation versus ﬁrst-generation antipsychotic drugs for schizophrenia: a metaanalysis. Lancet 2008; published online Dec 5. DOI:10.1016/S01406736(08)61764-X. Janssen PA, Niemegeers CJ, Awouters F, Schellekens KH, Megens AA, Meert TF. Pharmacology of risperidone (R 64 766), a new antipsychotic with serotonin-S2 and dopamine-D2 antagonistic properties. J Pharmacol Exp Ther 1988; 244: 685–93. Lieberman JA, Stroup S, McEvoy JP, et al, for the Clinical Antipsychotic Trials of Intervention Eﬀectiveness (CATIE) Investigators. Eﬀectiveness of antipsychotic drugs in patients with chronic schizophrenia. N Engl J Med 2005; 353: 1209–23. Jones PB, Barnes TR, Davies L, et al. Randomized controlled trial of eﬀect on quality of life of second- vs ﬁrst-generation antipsychotic drugs in schizophrenia: Cost Utility of the Latest Antipsychotic Drugs in Schizophrenia Study (CUtLASS 1). Arch Gen Psychiatry 2006; 63: 1079–87. Keefe RS, Bilder RM, Davis SM, et al, for the CATIE Investigators and the Neurocognitive Working Group. Neurocognitive eﬀects of antipsychotic medications in patients with chronic schizophrenia in the CATIE trial. Arch Gen Psychiatry 2007; 64: 633–47. Rosenheck RA, Leslie D, Sindelar J, et al, for the CATIE Study Investigators. Cost-eﬀectiveness of second generation antipsychotics and perphenazine in a randomized trial of treatment for chronic schizophrenia. Am J Psychiatry 2006; 163: 2080–89. Davies LM, Lewis S, Jones PB, et al, on behalf of the CUtLASS team. Cost-eﬀectiveness of ﬁrst- v. second-generation antipsychotic drugs: results from a randomised controlled trial in schizophrenia responding poorly to previous therapy. Br J Psychiatry 2007; 191: 14–22. Lewis S, Lieberman J. CATIE and CUtLASS: can we handle the truth? Br J Psychiatry 2008; 192: 161–63. Vedantam S. In antipsychotics, newer isn’t better. Washington Post (Washington) Oct 3, 2006. http://www.washingtonpost.com/wp-dyn/ content/article/2006/10/02/AR2006100201378.html (accessed Sept 8, 2008). Geddes J, Freemantle N, Harrison P, Bebbington P. Atypical antipsychotics in the treatment of schizophrenia: systematic overview and meta-regression analysis. BMJ 2000; 321: 1371–76. Turner EH, Matthews AM, Linardatos E, Tell TA, Rosenthal R. Selective publication of antidepressant trials and its inﬂuence on apparent eﬃcacy. N Engl J Med 2008; 358: 252–60. Whittington CJ, Kendall T, Fonagy P, Cottrell D, Cotgrove A, Boddington E. Selective serotonin reuptake inhibitors in childhood depression: systematic review of published versus unpublished data. Lancet 2004; 363: 1341–45. Tungeraza T, Poole R. Inﬂuence of drug company authorship and sponsorship on drug trial outcomes. Br J Psychiatry 2007; 191: 82–83. Kane J, Honigfeld G, Singer J, Meltzer H. Clozapine for the treatment-resistant schizophrenic: a double-blind comparison with chlorpromazine. Arch Gen Psychiatry 1988; 45: 789–96.

Assessing bleeds clinically: what’s the score? Acute upper-gastrointestinal haemorrhage is the most common life-threatening medical emergency faced by gastroenterologists, with an annual incidence of 50–150 per 100 000 people.1 Mortality has been stubbornly high (14% in 1995), although a reaudit in 2007 by the British Society of Gastroenterology showed a UK mortality of 10%.2 That reaudit identiﬁed several trends, including a doubling of cases due to www.thelancet.com Vol 373 January 3, 2009

variceal bleeding and a striking reduction in the use of surgery. Whether the falling mortality reﬂects improved management or an altered case-mix is not clear. Certainly, several mild cases do not undergo endoscopy or need blood transfusion. Measures need to be developed to identify patients at low risk, who can be discharged early or for whom admission can be avoided, as well as to improve

Published Online December 15, 2008 DOI:10.1016/S01406736(08)61770-5 See Articles page 42

5

Comment

likely to be just as eﬃcacious as the second-generation drug, but less likely than haloperidol to induce Parkinsonism. The picture can be complicated further with high doses of the ﬁrst-generation drug. This approach favours the second-generation antipsychotic because side-eﬀect rates are much lower than with the ﬁrst-generation antipsychotic.10 Moreover, there is often selective publication of trials11–13 that can skew the evidence base in favour of a drug favoured by the investigators. On present evidence from all sources it is diﬃcult not to conclude that the trials of the second-generation antipsychotics seem to be driven more by marketing strategy than to clarify their role for clinicians and patients. This is not to say that all antipsychotic drugs are the same, they are not. Individual responses vary, and so a range of drugs is needed for good clinical practice. So where should we go now? First, the time has come to abandon the terms ﬁrst-generation and second-generation antipsychotics, as they do not merit this distinction. The only second-generation antipsychotic that is obviously better than other drugs in resistant schizophrenia is clozapine,14 and this is a very old drug indeed. Second, clinicians must remember to keep the beneﬁt–risk ratio of each antipsychotic drug in constant perspective because all are associated in diﬀerent ways with serious adverse eﬀects, which should be important outcome measures.13 Finally, it is prudent to remember that although science rules during a drug’s development, the market usurps control once the drug is released for care of patients.

PT declares that he has no conﬂict of interest. TK is involved in updating the schizophrenia guideline for the National Institute for Health and Clinical Excellence, including a review of antipsychotic drugs in the treatment of schizophrenia. 1

2

3

4

5

6

7

8 9

10

11

12

*Peter Tyrer, Tim Kendall Department of Psychological Medicine, Imperial College London, London W6 8RP, UK (PT); and National Collaborating Centre for Mental Health, Royal College of Psychiatrists’ Research Unit, London, UK (TK) [email protected]

13 14

Leucht S, Corves C, Arbter D, Engel RR, Li C, Davis JM. Second-generation versus ﬁrst-generation antipsychotic drugs for schizophrenia: a metaanalysis. Lancet 2008; published online Dec 5. DOI:10.1016/S01406736(08)61764-X. Janssen PA, Niemegeers CJ, Awouters F, Schellekens KH, Megens AA, Meert TF. Pharmacology of risperidone (R 64 766), a new antipsychotic with serotonin-S2 and dopamine-D2 antagonistic properties. J Pharmacol Exp Ther 1988; 244: 685–93. Lieberman JA, Stroup S, McEvoy JP, et al, for the Clinical Antipsychotic Trials of Intervention Eﬀectiveness (CATIE) Investigators. Eﬀectiveness of antipsychotic drugs in patients with chronic schizophrenia. N Engl J Med 2005; 353: 1209–23. Jones PB, Barnes TR, Davies L, et al. Randomized controlled trial of eﬀect on quality of life of second- vs ﬁrst-generation antipsychotic drugs in schizophrenia: Cost Utility of the Latest Antipsychotic Drugs in Schizophrenia Study (CUtLASS 1). Arch Gen Psychiatry 2006; 63: 1079–87. Keefe RS, Bilder RM, Davis SM, et al, for the CATIE Investigators and the Neurocognitive Working Group. Neurocognitive eﬀects of antipsychotic medications in patients with chronic schizophrenia in the CATIE trial. Arch Gen Psychiatry 2007; 64: 633–47. Rosenheck RA, Leslie D, Sindelar J, et al, for the CATIE Study Investigators. Cost-eﬀectiveness of second generation antipsychotics and perphenazine in a randomized trial of treatment for chronic schizophrenia. Am J Psychiatry 2006; 163: 2080–89. Davies LM, Lewis S, Jones PB, et al, on behalf of the CUtLASS team. Cost-eﬀectiveness of ﬁrst- v. second-generation antipsychotic drugs: results from a randomised controlled trial in schizophrenia responding poorly to previous therapy. Br J Psychiatry 2007; 191: 14–22. Lewis S, Lieberman J. CATIE and CUtLASS: can we handle the truth? Br J Psychiatry 2008; 192: 161–63. Vedantam S. In antipsychotics, newer isn’t better. Washington Post (Washington) Oct 3, 2006. http://www.washingtonpost.com/wp-dyn/ content/article/2006/10/02/AR2006100201378.html (accessed Sept 8, 2008). Geddes J, Freemantle N, Harrison P, Bebbington P. Atypical antipsychotics in the treatment of schizophrenia: systematic overview and meta-regression analysis. BMJ 2000; 321: 1371–76. Turner EH, Matthews AM, Linardatos E, Tell TA, Rosenthal R. Selective publication of antidepressant trials and its inﬂuence on apparent eﬃcacy. N Engl J Med 2008; 358: 252–60. Whittington CJ, Kendall T, Fonagy P, Cottrell D, Cotgrove A, Boddington E. Selective serotonin reuptake inhibitors in childhood depression: systematic review of published versus unpublished data. Lancet 2004; 363: 1341–45. Tungeraza T, Poole R. Inﬂuence of drug company authorship and sponsorship on drug trial outcomes. Br J Psychiatry 2007; 191: 82–83. Kane J, Honigfeld G, Singer J, Meltzer H. Clozapine for the treatment-resistant schizophrenic: a double-blind comparison with chlorpromazine. Arch Gen Psychiatry 1988; 45: 789–96.

Assessing bleeds clinically: what’s the score? Acute upper-gastrointestinal haemorrhage is the most common life-threatening medical emergency faced by gastroenterologists, with an annual incidence of 50–150 per 100 000 people.1 Mortality has been stubbornly high (14% in 1995), although a reaudit in 2007 by the British Society of Gastroenterology showed a UK mortality of 10%.2 That reaudit identiﬁed several trends, including a doubling of cases due to www.thelancet.com Vol 373 January 3, 2009

variceal bleeding and a striking reduction in the use of surgery. Whether the falling mortality reﬂects improved management or an altered case-mix is not clear. Certainly, several mild cases do not undergo endoscopy or need blood transfusion. Measures need to be developed to identify patients at low risk, who can be discharged early or for whom admission can be avoided, as well as to improve

Published Online December 15, 2008 DOI:10.1016/S01406736(08)61770-5 See Articles page 42

5

Comment

Year

Country

Type of study

Low-risk patients (n [%])

Interventions (n [%])

Blatchford8

2000

UK

Retrospective

276 (15·8%)

5 (0·3%)

Gralnek10

2004

USA

Retrospective

14 (8·0%)

0

Chen9

2007

Taiwan

Retrospective

28 (8·9%)

1 (0·3%)

Masaoka11

2007

Japan

Retrospective

3 (3·4%)

0

Stanley13

2008

UK

Prospective

228 (18·7%)

0

Table: Studies of patients identiﬁed as low risk and their outcomes in studies that used GBS of 0 for identifying low-risk patients with acute upper-gastrointestinal haemorrhage

management of high-risk patients with bleeds. Most risk-determination systems3–5 include endoscopy, and protocols allowing safe early discharge have been developed.6,7 The Glasgow-Blatchford score (GBS),8 which was ﬁrst reported in 2000, needs only simple clinical and laboratory data, and can potentially be used for front-door decisions. Since then, a few retrospective studies have suggested that the GBS9–11 and the modiﬁed GBS12 predict outcomes accurately. In The Lancet today, Adrian Stanley and colleagues,13 who studied prospective cohorts in two phases in Scotland and England, conﬁrm these ﬁndings and speciﬁcally address the question of whether patients presenting with a score of 0 could be managed as outpatients. To achieve a score of 0, a patient would have to have a pulse of less than 100 beats per min, systolic blood pressure of 110 mm Hg or more, absence of melaena, syncope, cardiac failure, or liver disease, haemoglobin of 130 g/L or more for men or 120 g/L or more for women, and urea of less than 6·5 mmol/L. In phase one of Stanley and colleagues’ study (case ascertainment), 105 (16·2%) of 649 patients had a GBS of 0 (low risk), with no interventions or death in this group. In phase two (validation), 123 (21·5%) of 572 patients were considered low risk, of whom 84 (14·7% of the total population in phase two) were not admitted, with none of the low-risk patients needing interventions. The GBS outperformed the Rockall scores (both pre-endoscopic and full) in this population in phase one of the trial. The pre-endoscopic Rockall score takes only clinical parameters (age, shock, and comorbidities) into account and has been shown to predict death but not rebleed rates.4 The full Rockall score includes the pre-endocscopic score and the results of endoscopy (diagnosis and presence of stigmata of bleed) and predicts mortality more reliably than does the pre-endoscopic score alone.4 The Rockall score had been designed to assess mortality and not rebleed or need for 6

interventions, unlike GBS which was designed to assess mortality or need for in-hospital interventions.8 Has the GBS done enough to merit being rolled out in all emergency departments? The table lists the studies (including that by Stanley and colleagues) with data on the use of a score of 0 in predicting outcomes. In three small retrospective studies, a score of 0 stratiﬁed less than 10% of patients with uppergastrointestinal haemorrhage as low risk, and all had excellent outcomes.9–11 By contrast, the Scottish studies (by Stanley13 and Blatchford and colleagues8) identiﬁed 15–20% of patients as low risk. These ﬁndings would imply either a diﬀerent case-mix in Scotland or that the GBS is not transferable geographically. However, the score does have excellent predictive ability and results were similar over time in both Scottish studies. Stanley and colleagues’ study is the ﬁrst prospective one to assess the use and feasibility of the GBS system in the emergency department. Their calculation of a mean reduction of 1·2 days in bed-days per patient presenting would translate to a direct saving of £13·6 million in a population of 60 million, with the assumptions that hospital cost is £227 per day14 and an annual rate of uppergastrointestinal haemorrhage of 100 per 100 000 people. Further studies will need to compare the costeﬀectiveness of the GBS with outpatient endoscopy and strategies that involve early endoscopy and discharge.6,7 Another approach uses an artiﬁcial neural network with 27 clinical variables to predict the need for endoscopic intervention.15 In a recent study,15 this model was no better than the complete Rockall score in an external validation cohort, which would imply better use for the simpler GBS. Avoiding unnecessary admission of such patients will reduce cost and exposure of patients to hospitalassociated hazards. But to reduce mortality from uppergastrointestinal haemorrhage, the standard of care in the UK needs enhancement. Socially, a sustained www.thelancet.com Vol 373 January 3, 2009

Comment

successful campaign to reduce alcohol consumption and stem the rising incidence of variceal bleeding as well as the other tragic consequences of alcoholic liver disease is essential.16 Organisationally, the patchwork service in which patients can still be admitted to hospitals without out-of-hours endoscopy needs to be replaced with one that ensures that all necessary diagnostic and treatment facilities are available.

6

7

8

9

10

Venkataraman Subramanian, *Christopher J Hawkey Wolfson Digestive Diseases Centre, University of Nottingham, Nottingham NG7 2UH, UK [email protected] VS declares that he has no conﬂict of interest. CJH has received research funding and/or honoraria from GlaxoSmithKline, MerckSerono, and Bayer. 1 2

3

4 5

British Society of Gastroenterology Endoscopy Committee. Non-variceal upper gastrointestinal haemorrhage: guidelines. Gut 2002; 51 (suppl 4): iv1–6. Clinical Eﬀectiveness and Evaluation Unit, British Society of Gastroenterology, NHS Blood and Transplant. UK comparative audit of upper gastrointestinal bleeding and the use of blood. December, 2007. http://www.bsg.org.uk/pdf_word_docs/blood_audit_report_07.pdf (accessed Sept 4, 2008). Hay JA, Lyubashevsky E, Elashoﬀ J, Maldonado L, Weingarten SR, Ellrodt AG. Upper gastrointestinal hemorrhage clinical guideline—determining the optimal hospital length of stay. Am J Med 1996; 100: 313–22. Rockall TA, Logan RF, Devlin HB, Northﬁeld TC. Risk assessment after acute upper gastrointestinal haemorrhage. Gut 1996; 38: 316–21. Saeed ZA, Winchester CB, Michaletz PA, Woods KL, Graham DY. A scoring system to predict rebleeding after endoscopic therapy of nonvariceal upper gastrointestinal hemorrhage, with a comparison of heat probe and ethanol injection. Am J Gastroenterol 1993; 88: 1842–49.

11

12

13

14 15

16

Cipolletta L, Bianco MA, Rotondano G, Marmo R, Piscopo R. Outpatient management for low-risk nonvariceal upper GI bleeding: a randomized controlled trial. Gastrointest Endosc 2002; 55: 1–5. Longstreth GF, Feitelberg SP. Successful outpatient management of acute upper gastrointestinal hemorrhage: use of practice guidelines in a large patient series. Gastrointest Endosc 1998; 47: 219–22. Blatchford O, Murray WR, Blatchford M. A risk score to predict need for treatment for uppergastrointestinal haemorrhage. Lancet 2000; 356: 1318–21. Chen IC, Hung MS, Chiu TF, Chen JC, Hsiao CT. Risk scoring systems to predict need for clinical intervention for patients with nonvariceal upper gastrointestinal tract bleeding. Am J Emerg Med 2007; 25: 774–79. Gralnek IM, Dulai GS. Incremental value of upper endoscopy for triage of patients with acute non-variceal upper-GI hemorrhage. Gastrointest Endosc 2004; 60: 9–14. Masaoka T, Suzuki H, Hori S, Aikawa N, Hibi T. Blatchford scoring system is a useful scoring system for detecting patients with upper gastrointestinal bleeding who do not need endoscopic intervention. J Gastroenterol Hepatol 2007; 22: 1404–08. Romagnuolo J, Barkun AN, Enns R, Armstrong D, Gregor J. Simple clinical predictors may obviate urgent endoscopy in selected patients with nonvariceal upper gastrointestinal tract bleeding. Arch Intern Med 2007; 167: 265–70. Stanley AJ, Ashley D, Dalton HR, et al. Outpatient management of patients with low-risk upper-gastrointestinal haemorrhage: multicentre validation and prospective evaluation. Lancet 2008; published online Dec 15. DOI:10.1016/S0140-6736(08)61769-9. Netten A, Curtis L. Unit costs of health and social care. 2006. http://www. pssru.ac.uk/pdf/uc/uc2006/uc2006.pdf (accessed Sept 4, 2008). Das A, Ben-Menachem T, Farooq FT, et al. Artiﬁcial neural network as a predictive instrument in patients with acute nonvariceal upper gastrointestinal hemorrhage. Gastroenterology 2008; 134: 65–74. Department of Health, Home Oﬃce, Department for Education and Skills and Department for Culture, Media and Sport. Safe. Sensible. Social. The next steps in the National Alcohol Strategy. June 5, 2007. http://www. dh.gov.uk/en/Publicationsandstatistics/Publications/PublicationsPolicyAn dGuidance/DH_075218 (accessed Sept 4, 2008).

Can antiretroviral therapy eliminate HIV transmission? In The Lancet today, Reuben Granich and colleagues (including two of us, KMDC and CFG) use mathematical modelling to assess the impact of expanded HIV testing and earlier antiretroviral therapy (ART) on HIV transmission.1 These researchers evaluated a theoretical programme of annual universal HIV testing and immediate treatment on HIV diagnosis, irrespective of CD4+ cell count, in an HIV epidemic with southern African population dynamics. The exercise suggested that HIV transmission could be substantially reduced within a few years. Elimination of HIV transmission, deﬁned as an incidence below one case per 1000 population per year, could be achieved within a decade, and the overall prevalence of HIV infection reduced to below 1% before the middle of the century. Compared with current practice of starting ART at a speciﬁc CD4+ count, deaths would be halved between now and 2050. www.thelancet.com Vol 373 January 3, 2009

The uncontrolled number of new HIV infections, despite years of prevention eﬀort, represents a crisis: 2·7 million incident HIV infections occurred globally in 2007.2 By contrast, only 1 million additional individuals were placed on ART that year;3 about 6·7 million were in danger of their lives for lack of treatment, and over 20 million other HIV-infected individuals were waiting, mostly unknowingly, for immune deﬁciency to progress. Universal access to HIV treatment and care is an unrealistic aspiration unless HIV transmission can be interrupted. Behavioural interventions are essential but incompletely eﬃcacious.4 For biomedical interventions, only four of 24 randomised trials showed protection; three assessed male circumcision.5 All vaccine and microbicide trials were negative. Unsurprisingly, how best to use ART for prevention has emerged as the most pressing question that faces HIV/AIDS science.

Published Online November 26, 2008 DOI:10.1016/S01406736(08)61732-8 See Comment page 9 See Articles page 48

7

Comment

successful campaign to reduce alcohol consumption and stem the rising incidence of variceal bleeding as well as the other tragic consequences of alcoholic liver disease is essential.16 Organisationally, the patchwork service in which patients can still be admitted to hospitals without out-of-hours endoscopy needs to be replaced with one that ensures that all necessary diagnostic and treatment facilities are available.

6

7

8

9

10

Venkataraman Subramanian, *Christopher J Hawkey Wolfson Digestive Diseases Centre, University of Nottingham, Nottingham NG7 2UH, UK [email protected] VS declares that he has no conﬂict of interest. CJH has received research funding and/or honoraria from GlaxoSmithKline, MerckSerono, and Bayer. 1 2

3

4 5

British Society of Gastroenterology Endoscopy Committee. Non-variceal upper gastrointestinal haemorrhage: guidelines. Gut 2002; 51 (suppl 4): iv1–6. Clinical Eﬀectiveness and Evaluation Unit, British Society of Gastroenterology, NHS Blood and Transplant. UK comparative audit of upper gastrointestinal bleeding and the use of blood. December, 2007. http://www.bsg.org.uk/pdf_word_docs/blood_audit_report_07.pdf (accessed Sept 4, 2008). Hay JA, Lyubashevsky E, Elashoﬀ J, Maldonado L, Weingarten SR, Ellrodt AG. Upper gastrointestinal hemorrhage clinical guideline—determining the optimal hospital length of stay. Am J Med 1996; 100: 313–22. Rockall TA, Logan RF, Devlin HB, Northﬁeld TC. Risk assessment after acute upper gastrointestinal haemorrhage. Gut 1996; 38: 316–21. Saeed ZA, Winchester CB, Michaletz PA, Woods KL, Graham DY. A scoring system to predict rebleeding after endoscopic therapy of nonvariceal upper gastrointestinal hemorrhage, with a comparison of heat probe and ethanol injection. Am J Gastroenterol 1993; 88: 1842–49.

11

12

13

14 15

16

Cipolletta L, Bianco MA, Rotondano G, Marmo R, Piscopo R. Outpatient management for low-risk nonvariceal upper GI bleeding: a randomized controlled trial. Gastrointest Endosc 2002; 55: 1–5. Longstreth GF, Feitelberg SP. Successful outpatient management of acute upper gastrointestinal hemorrhage: use of practice guidelines in a large patient series. Gastrointest Endosc 1998; 47: 219–22. Blatchford O, Murray WR, Blatchford M. A risk score to predict need for treatment for uppergastrointestinal haemorrhage. Lancet 2000; 356: 1318–21. Chen IC, Hung MS, Chiu TF, Chen JC, Hsiao CT. Risk scoring systems to predict need for clinical intervention for patients with nonvariceal upper gastrointestinal tract bleeding. Am J Emerg Med 2007; 25: 774–79. Gralnek IM, Dulai GS. Incremental value of upper endoscopy for triage of patients with acute non-variceal upper-GI hemorrhage. Gastrointest Endosc 2004; 60: 9–14. Masaoka T, Suzuki H, Hori S, Aikawa N, Hibi T. Blatchford scoring system is a useful scoring system for detecting patients with upper gastrointestinal bleeding who do not need endoscopic intervention. J Gastroenterol Hepatol 2007; 22: 1404–08. Romagnuolo J, Barkun AN, Enns R, Armstrong D, Gregor J. Simple clinical predictors may obviate urgent endoscopy in selected patients with nonvariceal upper gastrointestinal tract bleeding. Arch Intern Med 2007; 167: 265–70. Stanley AJ, Ashley D, Dalton HR, et al. Outpatient management of patients with low-risk upper-gastrointestinal haemorrhage: multicentre validation and prospective evaluation. Lancet 2008; published online Dec 15. DOI:10.1016/S0140-6736(08)61769-9. Netten A, Curtis L. Unit costs of health and social care. 2006. http://www. pssru.ac.uk/pdf/uc/uc2006/uc2006.pdf (accessed Sept 4, 2008). Das A, Ben-Menachem T, Farooq FT, et al. Artiﬁcial neural network as a predictive instrument in patients with acute nonvariceal upper gastrointestinal hemorrhage. Gastroenterology 2008; 134: 65–74. Department of Health, Home Oﬃce, Department for Education and Skills and Department for Culture, Media and Sport. Safe. Sensible. Social. The next steps in the National Alcohol Strategy. June 5, 2007. http://www. dh.gov.uk/en/Publicationsandstatistics/Publications/PublicationsPolicyAn dGuidance/DH_075218 (accessed Sept 4, 2008).

Can antiretroviral therapy eliminate HIV transmission? In The Lancet today, Reuben Granich and colleagues (including two of us, KMDC and CFG) use mathematical modelling to assess the impact of expanded HIV testing and earlier antiretroviral therapy (ART) on HIV transmission.1 These researchers evaluated a theoretical programme of annual universal HIV testing and immediate treatment on HIV diagnosis, irrespective of CD4+ cell count, in an HIV epidemic with southern African population dynamics. The exercise suggested that HIV transmission could be substantially reduced within a few years. Elimination of HIV transmission, deﬁned as an incidence below one case per 1000 population per year, could be achieved within a decade, and the overall prevalence of HIV infection reduced to below 1% before the middle of the century. Compared with current practice of starting ART at a speciﬁc CD4+ count, deaths would be halved between now and 2050. www.thelancet.com Vol 373 January 3, 2009

The uncontrolled number of new HIV infections, despite years of prevention eﬀort, represents a crisis: 2·7 million incident HIV infections occurred globally in 2007.2 By contrast, only 1 million additional individuals were placed on ART that year;3 about 6·7 million were in danger of their lives for lack of treatment, and over 20 million other HIV-infected individuals were waiting, mostly unknowingly, for immune deﬁciency to progress. Universal access to HIV treatment and care is an unrealistic aspiration unless HIV transmission can be interrupted. Behavioural interventions are essential but incompletely eﬃcacious.4 For biomedical interventions, only four of 24 randomised trials showed protection; three assessed male circumcision.5 All vaccine and microbicide trials were negative. Unsurprisingly, how best to use ART for prevention has emerged as the most pressing question that faces HIV/AIDS science.

Published Online November 26, 2008 DOI:10.1016/S01406736(08)61732-8 See Comment page 9 See Articles page 48

7

Reuben Granich

Comment

Sculpture by Zimbabwean Mike Munyaradzi, 2007 “Symbolizing the fragile world we live in today, the hollowed sphere signiﬁes the destruction of our globe and the diﬃcult issue of our time including AIDS. The leaf represents the life we are struggling to sustain.”

ART may be used to prevent acquisition of HIV infection when delivered as prophylaxis before or after HIV exposure, or to prevent transmission from HIV-infected persons when provided as treatment that renders them less infectious. Prophylactic approaches are challenged by the large numbers having to be treated to prevent a single infection, and ART for preventing transmission from infected persons will have the greatest impact.6 Biological plausibility is provided by evidence that viral load is the major risk factor for all modes of HIV transmission and that ART lowers viral load in plasma and in genital secretions.7 Proof of concept is oﬀered by the virtual elimination of paediatric HIV disease in highincome countries by universal voluntary HIV testing of pregnant women and appropriate provision of ART. By lowering viral load, ART reduces transmissibility as well as duration of infectiousness, the key determinants (along with rate of partner change) of the basic reproductive rate (R₀), the number of secondary infections generated by one primary case.8 Epidemic control requires that R₀ be reduced below 1. Observational data have shown reduced HIV transmission in discordant couples after the 8

introduction of combination ART,9 and programmatic data supported reduction in HIV transmission at the population level.10 Current clinical practice, however, limits the preventive eﬃcacy of ART because substantial HIV transmission can occur around the time of seroconversion, when viral load is highest, ample opportunity for transmission exists before reaching the CD4+ count threshold for treatment, and many HIV-infected individuals are tested late or not at all. Dominant concerns of the approach modelled by Granich and colleagues relate to necessary protection of voluntariness and individual rights. Feasibility is challenged by: weak health systems and insuﬃcient health personnel; choice of appropriate drug regimens; treatment adherence; drug toxicity; drug resistance and need for durable second-line and third-line regimens; the logistics, reliability, and acceptability of regularly testing a whole population for HIV infection; and behavioural risk compensation.11 Initial costs would be high and scepticism should be expected towards large-scale topdown eﬀorts that appear to medicalise HIV prevention. By contrast, advantages of immediate treatment at diagnosis could include: simpliﬁed clinical management; reduction in the high mortality rates from late diagnosis;12 control of HIV-associated tuberculosis;13,14 and eﬀective prevention of mother-to-child transmission of HIV, including through breastfeeding. The approach could be cost-saving over the longer term, because over time fewer individuals would be living with HIV and eﬀorts would increasingly focus on treatment adherence and localised HIV transmission. Appropriately, Granich and colleagues emphasise that all current preventive interventions should be enhanced. Currently, only 20% of individuals with HIV in lowincome and middle-income countries know their serostatus,3 but most infected individuals will eventually require ART if they are not to die of AIDS. Because of newly recognised, non-AIDS-deﬁning complications of uncontrolled HIV replication,15 as well as better treatment options,16 the pendulum around when to start therapy is again swinging towards earlier starting.16,17 Expanded HIV testing and immediate treatment would oﬀer opportunity for highest quality positive prevention, a holistic approach that protects the physical, sexual, and reproductive health of individuals with HIV, and maximally reduces onward transmission.18 Provided www.thelancet.com Vol 373 January 3, 2009

Comment

coercion is avoided and conﬁdentiality and dignity maintained, individual health and societal safety should beneﬁt through reduced HIV transmission, which would enhance human rights overall. The approach could represent a conceptual redeﬁnition of universal access and combination prevention, reframe the debate around when to start ART, and counter prevention pessimism. There is currently inadequate evidence for WHO to deﬁne policy and guidance about the role of ART in HIV prevention. Research is needed to assess the feasibility, safety, acceptability, impact, and cost of innovations such as the universal voluntary testing and immediate treatment approach, and broad consultation must address community, human rights, ethical, and political concerns. Research on male circumcision oﬀers encouraging as well as cautionary experience about translation of research ﬁndings on controversial interventions into practice.19 Discussion will also be needed on ART for prevention in populations most at risk in epidemics of varying intensity. WHO is committed to promoting consultation among countries and stakeholders about the pressing biomedical issue of ART for HIV prevention.

2

3

4 5 6

7

8

9

10

11

12

13

14

*Kevin M De Cock, Charles F Gilks, Ying-Ru Lo, Teguest Guerma

15

HIV/AIDS Department, WHO, 1211 Geneva 27, Switzerland (KMDC, TG); Antiretroviral Treatment and HIV Care Unit, HIV/AIDS Department, WHO, Geneva, Switzerland (CFG); and Prevention in the Health Sector Unit, HIV/AIDS Department, WHO, Geneva, Switzerland (Y-RL) [email protected]

16

We thank Christopher Dye, Brian Williams, and Reuben Granich for helpful discussions about this Comment. We declare that we have no conﬂict of interest. 1

Granich RM, Gilks CF, Dye C, De Cock KM, Williams BG. Universal voluntary HIV testing with immediate antiretroviral therapy as a strategy for elimination of HIV transmission: a mathematical model. Lancet 2008; published online Nov 26. DOI:10.1016/S0140-6736(08)61697-9.

17 18

19

UNAIDS. 2008 Report on the global AIDS epidemic: executive summary. July, 2008. http://data.unaids.org/pub/GlobalReport/2008/JC1511_GR08_ ExecutiveSummary_en.pdf (accessed Nov 12, 2008). WHO. Towards universal access: scaling up priority HIV/AIDS interventions in the health sector. 2008. http://www.who.int/hiv/pub/towards_ universal_access_report_2008.pdf (accessed Nov 12, 2008). Coates TJ, Richter L, Caceres C. Behavioural strategies to reduce HIV transmission: how to make them work better. Lancet 2008; 372: 669–84. Cohen J. Treatment and prevention exchange vows at International Conference. Science 2008; 321: 902–03. Montaner JS, Hogg R, Wood E, et al. The case for expanding access to highly active antiretroviral therapy to curb the growth of the HIV epidemic. Lancet 2006; 368: 531–36. Janssen RS, Holtgrave DR, Valdiserri RO, Shepherd M, Gayle HD, De Cock KM. The serostatus approach to ﬁghting the HIV epidemic: prevention strategies for infected individuals. Am J Public Health 2001; 91: 1019–24. May RM, Anderson RM. The transmission dynamics of human immunodeﬁciency virus (HIV). Phil Trans R Soc Lond B Biol Sci 1988; 321: 565–607. Castilla J, Del Romero J, Hernando V, et al. Eﬀectiveness of highly active antiretroviral therapy in reducing heterosexual transmission of HIV. J Acquir Immune Deﬁc Syndr 2005; 40: 96–101. Fang CT, Hsu HM, Twu SJ, et al. Decreased HIV transmission after a policy of providing free access to highly active antiretroviral therapy in Taiwan. J Infect Dis 2004; 190: 879–85. Velasco-Hernandez JX, Gershengorn HB, Blower SM. Could widespread use of combination antiretroviral therapy eradicate HIV epidemics? Lancet Infect Dis 2002; 2: 487–93. Lawn SD, Harries AD, Anglaret X, Myer L, Wood R. Early mortality among adults accessing antiretroviral treatment programmes in sub-Saharan Africa. AIDS 2008; 22: 1897–908. Corbett EL, Marston B, Churchyard GJ, De Cock KM. Tuberculosis in subSaharan Africa: opportunities, challenges, and change in the era of antiretroviral therapy. Lancet 2006; 367: 926–37. Williams BG, Dye C. Antiretroviral drugs for tuberculosis control in the era of HIV/AIDS. Science 2003; 301: 1535–37. Phillips A. Morbidity and mortality in the HAART era. 15th Conference on Retroviruses and Opportunistic Infections, Boston, MA, USA; Feb 3–6, 2008: 8 (abstr). http://www.retroconference.org/2008/Abstracts/33423. htm (accessed Nov 12, 2008). Hammer SM, Eron JJ, Reiss P, et al. Antiretroviral treatment of adult HIV infection: 2008 recommendations of the International AIDS Society USA Panel. JAMA 2008; 300: 555–70. Hirsch MS. Initiating therapy: when to start, what to use. J Infect Dis 2008; 197 (suppl 3): S252–60. WHO. Essential prevention and care interventions for adults and adolescents living with HIV in resource-limited settings. 2008. http://www.who.int/hiv/ pub/prev_care/OMS_EPP_AFF_en.pdf (accessed Nov 12, 2008). WHO, UNAIDS. New data on male circumcision and HIV prevention: policy and programme implications. http://data.unaids.org/pub/Report/2007/ mc_recommendations_en.pdf (accessed Nov 12, 2008).

Treating our way out of the HIV pandemic: could we, would we, should we? HIV prevention is easy in theory—the practice is hard. In models, HIV can be eliminated if risk behaviours or viral transmissibility are reduced substantially. Unfortunately, in many places, achievement of these reductions has not been possible and HIV incidence has remained high. In The Lancet today, Reuben Granich and colleagues use mathematical models to show that annual screening of most adults for HIV, with immediate commencement www.thelancet.com Vol 373 January 3, 2009

of antiretroviral therapy (ART) for all those infected, would dramatically reduce HIV incidence.1 This strategy would be a bold move away from the current approach of treatment on the basis of clinical need, in which the hoped-for synergies between treatment and prevention will remain limited because testing and counselling focus on individuals who have probably already transmitted infection.2

Published Online November 26, 2008 DOI:10.1016/S01406736(08)61698-0 See Comment page 7 See Articles page 48

9

Comment

coercion is avoided and conﬁdentiality and dignity maintained, individual health and societal safety should beneﬁt through reduced HIV transmission, which would enhance human rights overall. The approach could represent a conceptual redeﬁnition of universal access and combination prevention, reframe the debate around when to start ART, and counter prevention pessimism. There is currently inadequate evidence for WHO to deﬁne policy and guidance about the role of ART in HIV prevention. Research is needed to assess the feasibility, safety, acceptability, impact, and cost of innovations such as the universal voluntary testing and immediate treatment approach, and broad consultation must address community, human rights, ethical, and political concerns. Research on male circumcision oﬀers encouraging as well as cautionary experience about translation of research ﬁndings on controversial interventions into practice.19 Discussion will also be needed on ART for prevention in populations most at risk in epidemics of varying intensity. WHO is committed to promoting consultation among countries and stakeholders about the pressing biomedical issue of ART for HIV prevention.

2

3

4 5 6

7

8

9

10

11

12

13

14

*Kevin M De Cock, Charles F Gilks, Ying-Ru Lo, Teguest Guerma

15

HIV/AIDS Department, WHO, 1211 Geneva 27, Switzerland (KMDC, TG); Antiretroviral Treatment and HIV Care Unit, HIV/AIDS Department, WHO, Geneva, Switzerland (CFG); and Prevention in the Health Sector Unit, HIV/AIDS Department, WHO, Geneva, Switzerland (Y-RL) [email protected]

16

We thank Christopher Dye, Brian Williams, and Reuben Granich for helpful discussions about this Comment. We declare that we have no conﬂict of interest. 1

Granich RM, Gilks CF, Dye C, De Cock KM, Williams BG. Universal voluntary HIV testing with immediate antiretroviral therapy as a strategy for elimination of HIV transmission: a mathematical model. Lancet 2008; published online Nov 26. DOI:10.1016/S0140-6736(08)61697-9.

17 18

19

UNAIDS. 2008 Report on the global AIDS epidemic: executive summary. July, 2008. http://data.unaids.org/pub/GlobalReport/2008/JC1511_GR08_ ExecutiveSummary_en.pdf (accessed Nov 12, 2008). WHO. Towards universal access: scaling up priority HIV/AIDS interventions in the health sector. 2008. http://www.who.int/hiv/pub/towards_ universal_access_report_2008.pdf (accessed Nov 12, 2008). Coates TJ, Richter L, Caceres C. Behavioural strategies to reduce HIV transmission: how to make them work better. Lancet 2008; 372: 669–84. Cohen J. Treatment and prevention exchange vows at International Conference. Science 2008; 321: 902–03. Montaner JS, Hogg R, Wood E, et al. The case for expanding access to highly active antiretroviral therapy to curb the growth of the HIV epidemic. Lancet 2006; 368: 531–36. Janssen RS, Holtgrave DR, Valdiserri RO, Shepherd M, Gayle HD, De Cock KM. The serostatus approach to ﬁghting the HIV epidemic: prevention strategies for infected individuals. Am J Public Health 2001; 91: 1019–24. May RM, Anderson RM. The transmission dynamics of human immunodeﬁciency virus (HIV). Phil Trans R Soc Lond B Biol Sci 1988; 321: 565–607. Castilla J, Del Romero J, Hernando V, et al. Eﬀectiveness of highly active antiretroviral therapy in reducing heterosexual transmission of HIV. J Acquir Immune Deﬁc Syndr 2005; 40: 96–101. Fang CT, Hsu HM, Twu SJ, et al. Decreased HIV transmission after a policy of providing free access to highly active antiretroviral therapy in Taiwan. J Infect Dis 2004; 190: 879–85. Velasco-Hernandez JX, Gershengorn HB, Blower SM. Could widespread use of combination antiretroviral therapy eradicate HIV epidemics? Lancet Infect Dis 2002; 2: 487–93. Lawn SD, Harries AD, Anglaret X, Myer L, Wood R. Early mortality among adults accessing antiretroviral treatment programmes in sub-Saharan Africa. AIDS 2008; 22: 1897–908. Corbett EL, Marston B, Churchyard GJ, De Cock KM. Tuberculosis in subSaharan Africa: opportunities, challenges, and change in the era of antiretroviral therapy. Lancet 2006; 367: 926–37. Williams BG, Dye C. Antiretroviral drugs for tuberculosis control in the era of HIV/AIDS. Science 2003; 301: 1535–37. Phillips A. Morbidity and mortality in the HAART era. 15th Conference on Retroviruses and Opportunistic Infections, Boston, MA, USA; Feb 3–6, 2008: 8 (abstr). http://www.retroconference.org/2008/Abstracts/33423. htm (accessed Nov 12, 2008). Hammer SM, Eron JJ, Reiss P, et al. Antiretroviral treatment of adult HIV infection: 2008 recommendations of the International AIDS Society USA Panel. JAMA 2008; 300: 555–70. Hirsch MS. Initiating therapy: when to start, what to use. J Infect Dis 2008; 197 (suppl 3): S252–60. WHO. Essential prevention and care interventions for adults and adolescents living with HIV in resource-limited settings. 2008. http://www.who.int/hiv/ pub/prev_care/OMS_EPP_AFF_en.pdf (accessed Nov 12, 2008). WHO, UNAIDS. New data on male circumcision and HIV prevention: policy and programme implications. http://data.unaids.org/pub/Report/2007/ mc_recommendations_en.pdf (accessed Nov 12, 2008).

Treating our way out of the HIV pandemic: could we, would we, should we? HIV prevention is easy in theory—the practice is hard. In models, HIV can be eliminated if risk behaviours or viral transmissibility are reduced substantially. Unfortunately, in many places, achievement of these reductions has not been possible and HIV incidence has remained high. In The Lancet today, Reuben Granich and colleagues use mathematical models to show that annual screening of most adults for HIV, with immediate commencement www.thelancet.com Vol 373 January 3, 2009

of antiretroviral therapy (ART) for all those infected, would dramatically reduce HIV incidence.1 This strategy would be a bold move away from the current approach of treatment on the basis of clinical need, in which the hoped-for synergies between treatment and prevention will remain limited because testing and counselling focus on individuals who have probably already transmitted infection.2

Published Online November 26, 2008 DOI:10.1016/S01406736(08)61698-0 See Comment page 7 See Articles page 48

9

Comment

12

Relative transmissibility Total new infections generated by a case (R0) 10 5 2

Relative transmissibility or cumulative infections

10

8

6

4

2

0 0 R0=10

1

2

3

4

5

R0=5

6

7

8

9

10

R0=2

Point needed to treat for elimination Time since infection of an individual (years)

Figure: Cumulative number of new HIV infections transmitted per infected individual As function of time since infection for scenarios in which each individual generates two, ﬁve, and ten new infections. Infectiousness is assumed to be ten-fold higher in ﬁrst 2 months than in long asymptomatic period, and four-fold higher in ﬁnal 1·5 years of average infection of 10 years.

The “treat early, treat hard” approach of the early era of triple ART fell out of favour due to drug toxicities, concerns over the evolution of drug resistance, and potential limits to the duration of treatment eﬃcacy in the patient. Therefore, when Velasco-Hernandez and colleagues ﬁrst showed in simple models that ART could eliminate HIV infection, their ﬁndings appeared extremely unrealistic.3 Since then, clinical opinions have changed with improved regimens and evidence of lower death rates if ART is started before CD4+ cell counts fall below 350 per μL, which makes earlier treatment more acceptable.4 When early treatment is considered as a prevention tool, success will require substantial resources and depend on a remarkable degree of acceptance and cooperation across populations. If we could eliminate HIV this way would we, given the will needed, and should we, given the conﬂict between utilitarianism and individualism5 inherent in this strategy? Granich and colleagues’ model ﬁndings appear robust and the critical mathematics is straightforward. At the start of an epidemic, the number of subsequent 10

infections generated by each infection (the basic reproductive number, R₀) determines how diﬃcult it will be to control the infection.6 Reductions in variables that contribute to R₀, such as the average transmission probability, will reduce the spread of infection. There is a simple rule that, for a given value of R₀, an equivalent fold reduction, reducing R₀ to its tipping point of one, will eliminate the infection. Secondary infections over the course of infection and when individuals need to be treated are shown in the ﬁgure. To eliminate infection, the individual must be treated and become noninfectious at the point when the cumulative number of infections passes through one. To allow for early high infectiousness and potentially high R₀ values, Granich’s model represents a very intensive screening programme. Such a programme and additional interventions in those with highest risk would probably be required if elimination was truly the goal. However, real elimination (ie, an incidence of zero in a large geographic area7) is unlikely. There will be some people whose risk is suﬃcient to maintain pockets of transmission within a very small fraction of the population. Probably, some individuals would not be screened and treated, and if these individuals have high risks of HIV acquisition and transmission, they will maintain infection within the population. Nonetheless, dramatic reductions in incidence are likely to be achieved. Granich and colleagues’ suggested strategy would be extremely radical, with medical intervention for public-health beneﬁts rather than individual patient’s beneﬁts; more so, if the costs of toxicity and inconvenience for the individual outweigh the clinical beneﬁts. Because screening and treatment would be for the public good, resources would have to come from the public purse. Current studies of uptake of and adherence to ART represent the case when mostly sick people are motivated to seek care. The strategy would rely on people, who feel well, regularly seeking care and adhering to ART. The prevention impact of the treatment depends on those who receive ART having a much reduced transmissibility, which seems reasonable, but there would be concern over other sexually transmitted infections, poor adherence, and treatment failure all reconstituting infectiousness, or the evolution and spread of resistant virus undermining treatment. There would also be questions about the epidemic contexts in which such a strategy should be used, the www.thelancet.com Vol 373 January 3, 2009

Comment

required frequency of testing, and how that frequency would change as prevalence falls. The suggested strategy would reﬂect public health at its best and its worst. At its best, the strategy would prevent morbidity and mortality for the population, both through better treatment of the individual and reduced spread of HIV. At its worst, the strategy will involve over-testing, over-treatment, side-eﬀects, resistance, and potentially reduced autonomy of the individual in their choices of care. The individual might gain no personal beneﬁt from testing and early treatment, but they would beneﬁt from protecting partners—and who could object to that, unless they were recklessly exposing others to infection? It is easy to see how enforced testing and treatment for the good of society would follow from such an argument. Partial success would lead to infection becoming concentrated in those with a high risk, with an increased danger of stigma and coercion. The history of the control of sexually transmitted infections documents several examples of compulsory screening and treatment of stigmatised populations,8 and there is a danger of a well-meaning paternalistic medical model following such a route. Capacity—ﬁnancial, physical, and human—as well as eﬀectiveness and cost-eﬀectiveness are important. The relative power of curative versus preventive medicine

and the priority of health enforcement would be crucial in implementation. Challenges will rightly come from those concerned about individuals’ rights and patients’ autonomy, as well as from those who moralistically fear an “easy” solution to HIV rather than behavioural change. *Geoﬀrey P Garnett, Rebecca F Baggaley Department of Infectious Disease Epidemiology, Imperial College London, London W2 1PG, UK [email protected] We declare that we have no conﬂict of interest. 1

2 3

4 5 6 7

8

Granich RM, Gilks CF, Dye C, De Cock KM, Williams BG. Universal voluntary HIV testing with immediate antiretroviral therapy as a strategy for elimination of HIV transmission: a mathematical model. Lancet 2008; published online Nov 26. DOI:10.1016/S0140-6736(08)61697-9. Baggaley RF, Garnett GP, Ferguson NM. Modelling the impact of antiretroviral use in resource-poor settings. PLoS Med 2006; 3: e124. Velasco-Hernandez JX, Gershengorn HB, Blower SM. Could widespread use of combination antiretroviral therapy eradicate HIV epidemics? Lancet Infect Dis 2002; 2: 487–93. Wilkin TJ, Gulick RM. When to start antiretroviral therapy? Clin Infect Dis 2008; published online Nov 6. DOI:10.1086/593311. Looker KJ, Hallet TB. Individual freedom versus collective responsibility: too many rights make a wrong? Emerging Themes Epidemiol 2006; 3: 14. Anderson RM, May RM: Infectious diseases of humans: dynamics and control. Oxford, Oxford University Press, 1991. Molyneux DH, Hopkins DR, Zagaria N. Disease eradication, elimination and control: the need for accurate and consistent usage. Trends Parasitol 2004; 20: 347–51. Porter D, Porter R. The enforcement of health: the British debate. In: Fee E, Fox DM, eds. AIDS the burdens of history. Berkeley and Los Angeles, CA: University of California Press, 1988: 97–120.

Malaria scale-up progress: is the glass half-empty or half-full? In The Lancet today, Abdisalan Noor and colleagues analyse progress in achieving coverage in Africa with insecticide-treated bednets.1 In 2007, 90 million children in stable endemic areas were not protected by such nets. Yet, overall, coverage has increased about six-fold since 2000. So, is the glass half-empty or half-full for malaria scale-up progress? Some would say half-empty. Noor and colleagues show that some large countries with substantial populations at risk, such as the Democratic Republic of Congo, Nigeria, and Sudan—that together account for a major part of Africa’s malaria-related morbidity and mortality—have so far achieved minimal scaleup. Coverage as documented by Noor and elsewhere2 remains unacceptably low. The scale of what yet needs to be achieved is both sobering and a reality check. The usage rates of insecticide-treated bednets reported by Noor and colwww.thelancet.com Vol 373 January 3, 2009

leagues are for children under 5 years of age. Although this criterion is the accepted indicator for bednet use as deﬁned by Roll Back Malaria, it is worth reminding ourselves that the goal of Roll Back Malaria is now universal coverage—deﬁned as one long-lasting insecticide-treated bednet for every two people in the household—with 80% usage. To achieve universal coverage by 2010, about 250–300 million new long-lasting insecticide-treated bednets will need to be distributed in Africa.3 Clearly, to achieve this ambitious goal in about 2 years, the pace of distribution of insecticide-treated bednets will need to increase dramatically, especially in places where little has been achieved so far. In countries with well-developed national distribution strategies, various approaches to catch up and keep up with coverage may be useful.4 However, rapid progress towards 2010 goals of Roll Back Malaria for bednets in countries where usage currently languishes in single

Published Online November 18, 2008 DOI:10.1016/S01406736(08)61597-4 See Articles page 58

11

Comment

required frequency of testing, and how that frequency would change as prevalence falls. The suggested strategy would reﬂect public health at its best and its worst. At its best, the strategy would prevent morbidity and mortality for the population, both through better treatment of the individual and reduced spread of HIV. At its worst, the strategy will involve over-testing, over-treatment, side-eﬀects, resistance, and potentially reduced autonomy of the individual in their choices of care. The individual might gain no personal beneﬁt from testing and early treatment, but they would beneﬁt from protecting partners—and who could object to that, unless they were recklessly exposing others to infection? It is easy to see how enforced testing and treatment for the good of society would follow from such an argument. Partial success would lead to infection becoming concentrated in those with a high risk, with an increased danger of stigma and coercion. The history of the control of sexually transmitted infections documents several examples of compulsory screening and treatment of stigmatised populations,8 and there is a danger of a well-meaning paternalistic medical model following such a route. Capacity—ﬁnancial, physical, and human—as well as eﬀectiveness and cost-eﬀectiveness are important. The relative power of curative versus preventive medicine

and the priority of health enforcement would be crucial in implementation. Challenges will rightly come from those concerned about individuals’ rights and patients’ autonomy, as well as from those who moralistically fear an “easy” solution to HIV rather than behavioural change. *Geoﬀrey P Garnett, Rebecca F Baggaley Department of Infectious Disease Epidemiology, Imperial College London, London W2 1PG, UK [email protected] We declare that we have no conﬂict of interest. 1

2 3

4 5 6 7

8

Granich RM, Gilks CF, Dye C, De Cock KM, Williams BG. Universal voluntary HIV testing with immediate antiretroviral therapy as a strategy for elimination of HIV transmission: a mathematical model. Lancet 2008; published online Nov 26. DOI:10.1016/S0140-6736(08)61697-9. Baggaley RF, Garnett GP, Ferguson NM. Modelling the impact of antiretroviral use in resource-poor settings. PLoS Med 2006; 3: e124. Velasco-Hernandez JX, Gershengorn HB, Blower SM. Could widespread use of combination antiretroviral therapy eradicate HIV epidemics? Lancet Infect Dis 2002; 2: 487–93. Wilkin TJ, Gulick RM. When to start antiretroviral therapy? Clin Infect Dis 2008; published online Nov 6. DOI:10.1086/593311. Looker KJ, Hallet TB. Individual freedom versus collective responsibility: too many rights make a wrong? Emerging Themes Epidemiol 2006; 3: 14. Anderson RM, May RM: Infectious diseases of humans: dynamics and control. Oxford, Oxford University Press, 1991. Molyneux DH, Hopkins DR, Zagaria N. Disease eradication, elimination and control: the need for accurate and consistent usage. Trends Parasitol 2004; 20: 347–51. Porter D, Porter R. The enforcement of health: the British debate. In: Fee E, Fox DM, eds. AIDS the burdens of history. Berkeley and Los Angeles, CA: University of California Press, 1988: 97–120.

Malaria scale-up progress: is the glass half-empty or half-full? In The Lancet today, Abdisalan Noor and colleagues analyse progress in achieving coverage in Africa with insecticide-treated bednets.1 In 2007, 90 million children in stable endemic areas were not protected by such nets. Yet, overall, coverage has increased about six-fold since 2000. So, is the glass half-empty or half-full for malaria scale-up progress? Some would say half-empty. Noor and colleagues show that some large countries with substantial populations at risk, such as the Democratic Republic of Congo, Nigeria, and Sudan—that together account for a major part of Africa’s malaria-related morbidity and mortality—have so far achieved minimal scaleup. Coverage as documented by Noor and elsewhere2 remains unacceptably low. The scale of what yet needs to be achieved is both sobering and a reality check. The usage rates of insecticide-treated bednets reported by Noor and colwww.thelancet.com Vol 373 January 3, 2009

leagues are for children under 5 years of age. Although this criterion is the accepted indicator for bednet use as deﬁned by Roll Back Malaria, it is worth reminding ourselves that the goal of Roll Back Malaria is now universal coverage—deﬁned as one long-lasting insecticide-treated bednet for every two people in the household—with 80% usage. To achieve universal coverage by 2010, about 250–300 million new long-lasting insecticide-treated bednets will need to be distributed in Africa.3 Clearly, to achieve this ambitious goal in about 2 years, the pace of distribution of insecticide-treated bednets will need to increase dramatically, especially in places where little has been achieved so far. In countries with well-developed national distribution strategies, various approaches to catch up and keep up with coverage may be useful.4 However, rapid progress towards 2010 goals of Roll Back Malaria for bednets in countries where usage currently languishes in single

Published Online November 18, 2008 DOI:10.1016/S01406736(08)61597-4 See Articles page 58

11

Science Photo Library

Comment

or low-double digits will need to rely initially on mass free distributions to achieve major improvements in coverage in this compressed timeline.5 Moreover, the progress in scale-up of other malaria interventions (case management with artemisinincontaining combinations, intermittent preventive treatment in pregnancy) has lagged behind that for insecticide-treated bednets. For example, a recent estimate in 14 African countries of use of artemisinin combinations in febrile children was 6% or less.2 In view of the fact that many antimalarial treatments are sought outside the public health system, private sector and community access to eﬀective antimalarials and diagnostics will need to be vastly improved if we are to make signiﬁcant inroads in improving these numbers. We have our work cut out for us. Nonetheless, there are several reasons to view the malaria scale-up glass as half-full. First, as noted by Noor and colleagues, some success has been achieved: seven African countries had projected use of insecticide-treated bednets of more than 40% by July, 2007. The challenge will be to document the lessons learned from those experiences and apply them elsewhere.6,7 Second, international funding for malaria control has increased three-fold from 2004 to 2007.8 Recent commitments include US$1·1 billion to expand the World Bank Malaria Booster Program (with targeted funding for the Democratic Republic of Congo and Nigeria),9 and $1·62 billion in new malaria grants from 12

the Global Fund to Fight AIDS, Tuberculosis and Malaria.10 Bilateral donors, such as the US President’s Malaria Initiative11 and the UK’s Department for International Development,10 have also pledged substantial amounts. About $2–3 billion is needed annually in Africa over the next several years to achieve control targets. Demonstration of impact from these investments will be critical for sustained commitment. Third, leadership has been more eﬀective at global and national levels, and has resulted in improved coordination and more eﬀective advocacy. After the Change Initiative in 2006,12 the Roll Back Malaria Partnership has become an authoritative convener for discussion of global strategies—including the recently released Global Malaria Action Plan3—and ways to improve country support. Most importantly, some African leaders have stepped forward to promote ambitious national programmes based on sound action plans.13 Without this national ownership and commitment, sustained success will not be achieved. Noor and colleagues highlight that there are mismatches between malaria burden and coverage with insecticide-treated bednets. To facilitate local decision making, it will be critical moving forward to increase support for gathering useful programme data on malaria burden and intervention coverage at subnational levels. Local data about conﬁrmed cases of malaria, and bednet ownership, use, and durability should help to better match availability of bednets with needs. Some would question the wisdom of discussing regional malaria elimination or global eradication in view of reports such as that of Noor and colleagues. We believe, however, that those working on malaria today can and must strike a delicate balance. Focus must be maintained on improving and sustaining scale-up in highburden countries. At the same time, the strategies and planning for strengthening surveillance, health systems, human-resource capacity, and regional coordination mechanisms must proceed. Lessons learned over the next 5–10 years in pursuing elimination in areas with lower transmission rates will help to guide future approaches in countries with currently intolerable malaria burdens.14 *Laurence Slutsker, Robert D Newman Malaria Branch, Centers for Disease Control and Prevention, Atlanta, GA 30307, USA [email protected] Our views are not necessarily those of the Centers for Disease Control and Prevention. We declare that we have no conﬂict of interest.

www.thelancet.com Vol 373 January 3, 2009

Comment

1

2

3 4 5 6

7 8

Noor AM, Mutheu JJ, Tatem AJ, Hay SI, Snow RW. Insecticide-treated net coverage in Africa: mapping progress in 2000–07. Lancet 2008; published online Nov 18. DOI:10.1016/S0140-6736(08)61596-2. UNICEF, Roll Back Malaria Partnership. Malaria and children: progress in intervention coverage, 2007. http://www.unicef.org/health/ﬁles/Malaria Oct6forweb_ﬁnal.pdf (accessed Oct 25, 2008). Roll Back Malaria. The global malaria action plan. http://www. rollbackmalaria.org/gmap (accessed Oct 25, 2008). Lengeler C, deSavigny D. Programme diversity is the key to success of insecticide treated bednets. Lancet 2007; 370: 1009–10. Teklehaimanot A, Sachs JD, Curtis C. Malaria control needs mass distribution of insecticidal bednets. Lancet 2007; 369: 2143–46. Steketee RW, Sipilanyambe N, Chimumbwa J, et al. National malaria control and scaling up for impact: the Zambia experience through 2006. Am J Trop Med Hyg 2008; 79: 45–52. Chambers RG, Gupta RK, Ghebreyesus TA. Responding to the challenge to end malaria deaths in Africa. Lancet 2008; 371: 1399–401. Snow RW, Guerra CA, Mutheu J, Hay SI. International funding for malaria control in relation to populations at risk of stable Plasmodium falciparum transmission. PLoS Medicine 2008; 7: e142.

9

10

11

12

13 14

World Bank. World Bank to give $1 billion boost to ﬁghting malaria in Africa. Sept 25, 2008. http://web.worldbank.org/WBSITE/EXTERNAL/ COUNTRIES/AFRICAEXT/0,,contentMDK:21914969~menuPK:258 658~page PK:2865106~piPK:286512 8~theSitePK:258644,00.html (accessed Nov 2, 2008). Malaria No More. World leaders make an unprecedented commitment to ﬁght malaria. Sept 25, 2009. http://www.malarianomore. org/news/features/mdg_092508.php (accessed Nov 4, 2008). Fast Facts. The President’s malaria initiative (PMI). April, 2008. http://www.ﬁghtingmalaria.gov/resources/reports/pmi_fastfacts.pdf (accessed Nov 4, 2008). Rollback Malaria Partnership. RBM Partnership Framework. December, 2006. http://rbmwhoint/changeinitiative/ PartnershipFramework.pdf (accessed Nov 4, 2008). Campbell CC. Halting the toll of malaria in Africa. Am J Trop Med Hyg 2008; 78: 851–53. Feachem R, Sabot O. A new global malaria eradication strategy. Lancet 2008; 371: 1633–35.

Helsinki discords: FDA, ethics, and international drug trials Since 1964, the Declaration of Helsinki has stood as one of the world’s most authoritative statements on ethical standards for human research.1 Drafted by the World Medical Association to provide medical researchers with ethical guidance, the Declaration has undergone six major revisions, most recently in October, 2008.2 For many years the US Food and Drug Administration (FDA) has required that foreign clinical studies supporting applications for drug licensure comply with the Declaration. However, on Oct 27, 2008, the FDA formally discontinued its reliance on the Declaration and substituted the International Conference on Harmonization’s Guideline for Good Clinical Practice (GCP).3 The rationale behind the FDA’s action is complex, and no doubt reﬂects an eﬀort to balance important interests and public-policy goals. Among the FDA’s reasons are the need to assure the quality of foreign data submitted to the agency, a wish to prevent confusion among researchers when the Declaration of Helsinki undergoes revision, and a worry that future modiﬁcations could “contain provisions that are inconsistent with US laws and regulations”.3 The FDA’s latest action completes a process begun in 2001 when the agency declined to recognise the 2000 revision, in part due to the Declaration’s restrictive stance on placebo-controlled trials in economically developing countries.4 The practical consequences of the FDA’s current action are unclear because the ruling applies to only a subset of clinical trials—ie, international trials. Moreover, several countries that host such www.thelancet.com Vol 373 January 3, 2009

research have regulations that endorse or emulate the Declaration of Helsinki. Nevertheless, at a time when the volume of overseas trials is increasing,5 the FDA’s new policy is troubling. First, the Declaration of Helsinki has a moral authority that GCP lacks. The Declaration has long been recognised as a leading international ethical standard for research. Whereas the World Medical Association includes 85 national medical societies from every part of the globe, the International Conference on Harmonization consists of only voting members from the USA, the European Union, and Japan. Indeed, the authors of GCP acknowledge the authority of the Declaration of Helsinki when they state that a goal of GCP is “consisten[cy] with the principles that have their origin in the Declaration of Helsinki”.6 The FDA regulates the largest drug market in the world and we worry that its replacement of the Declaration of Helsinki with a less morally authoritative document may cause others to follow suit, thereby undermining international ethical standards for research. Second, the Declaration of Helsinki has a breadth and depth that GCP lacks.1 For sure, GCP covers similar topics to the Declaration, but the focus of GCP is regulatory harmonisation, not the articulation of ethical commitments. Careful examination of the two documents reveals several important ethical issues that are addressed in the Declaration about which GCP is silent (panel). Thus reliance on GCP rather than on the Declaration of Helsinki may result in less protection for research 13

Comment

1

2

3 4 5 6

7 8

Noor AM, Mutheu JJ, Tatem AJ, Hay SI, Snow RW. Insecticide-treated net coverage in Africa: mapping progress in 2000–07. Lancet 2008; published online Nov 18. DOI:10.1016/S0140-6736(08)61596-2. UNICEF, Roll Back Malaria Partnership. Malaria and children: progress in intervention coverage, 2007. http://www.unicef.org/health/ﬁles/Malaria Oct6forweb_ﬁnal.pdf (accessed Oct 25, 2008). Roll Back Malaria. The global malaria action plan. http://www. rollbackmalaria.org/gmap (accessed Oct 25, 2008). Lengeler C, deSavigny D. Programme diversity is the key to success of insecticide treated bednets. Lancet 2007; 370: 1009–10. Teklehaimanot A, Sachs JD, Curtis C. Malaria control needs mass distribution of insecticidal bednets. Lancet 2007; 369: 2143–46. Steketee RW, Sipilanyambe N, Chimumbwa J, et al. National malaria control and scaling up for impact: the Zambia experience through 2006. Am J Trop Med Hyg 2008; 79: 45–52. Chambers RG, Gupta RK, Ghebreyesus TA. Responding to the challenge to end malaria deaths in Africa. Lancet 2008; 371: 1399–401. Snow RW, Guerra CA, Mutheu J, Hay SI. International funding for malaria control in relation to populations at risk of stable Plasmodium falciparum transmission. PLoS Medicine 2008; 7: e142.

9

10

11

12

13 14

World Bank. World Bank to give $1 billion boost to ﬁghting malaria in Africa. Sept 25, 2008. http://web.worldbank.org/WBSITE/EXTERNAL/ COUNTRIES/AFRICAEXT/0,,contentMDK:21914969~menuPK:258 658~page PK:2865106~piPK:286512 8~theSitePK:258644,00.html (accessed Nov 2, 2008). Malaria No More. World leaders make an unprecedented commitment to ﬁght malaria. Sept 25, 2009. http://www.malarianomore. org/news/features/mdg_092508.php (accessed Nov 4, 2008). Fast Facts. The President’s malaria initiative (PMI). April, 2008. http://www.ﬁghtingmalaria.gov/resources/reports/pmi_fastfacts.pdf (accessed Nov 4, 2008). Rollback Malaria Partnership. RBM Partnership Framework. December, 2006. http://rbmwhoint/changeinitiative/ PartnershipFramework.pdf (accessed Nov 4, 2008). Campbell CC. Halting the toll of malaria in Africa. Am J Trop Med Hyg 2008; 78: 851–53. Feachem R, Sabot O. A new global malaria eradication strategy. Lancet 2008; 371: 1633–35.

Helsinki discords: FDA, ethics, and international drug trials Since 1964, the Declaration of Helsinki has stood as one of the world’s most authoritative statements on ethical standards for human research.1 Drafted by the World Medical Association to provide medical researchers with ethical guidance, the Declaration has undergone six major revisions, most recently in October, 2008.2 For many years the US Food and Drug Administration (FDA) has required that foreign clinical studies supporting applications for drug licensure comply with the Declaration. However, on Oct 27, 2008, the FDA formally discontinued its reliance on the Declaration and substituted the International Conference on Harmonization’s Guideline for Good Clinical Practice (GCP).3 The rationale behind the FDA’s action is complex, and no doubt reﬂects an eﬀort to balance important interests and public-policy goals. Among the FDA’s reasons are the need to assure the quality of foreign data submitted to the agency, a wish to prevent confusion among researchers when the Declaration of Helsinki undergoes revision, and a worry that future modiﬁcations could “contain provisions that are inconsistent with US laws and regulations”.3 The FDA’s latest action completes a process begun in 2001 when the agency declined to recognise the 2000 revision, in part due to the Declaration’s restrictive stance on placebo-controlled trials in economically developing countries.4 The practical consequences of the FDA’s current action are unclear because the ruling applies to only a subset of clinical trials—ie, international trials. Moreover, several countries that host such www.thelancet.com Vol 373 January 3, 2009

research have regulations that endorse or emulate the Declaration of Helsinki. Nevertheless, at a time when the volume of overseas trials is increasing,5 the FDA’s new policy is troubling. First, the Declaration of Helsinki has a moral authority that GCP lacks. The Declaration has long been recognised as a leading international ethical standard for research. Whereas the World Medical Association includes 85 national medical societies from every part of the globe, the International Conference on Harmonization consists of only voting members from the USA, the European Union, and Japan. Indeed, the authors of GCP acknowledge the authority of the Declaration of Helsinki when they state that a goal of GCP is “consisten[cy] with the principles that have their origin in the Declaration of Helsinki”.6 The FDA regulates the largest drug market in the world and we worry that its replacement of the Declaration of Helsinki with a less morally authoritative document may cause others to follow suit, thereby undermining international ethical standards for research. Second, the Declaration of Helsinki has a breadth and depth that GCP lacks.1 For sure, GCP covers similar topics to the Declaration, but the focus of GCP is regulatory harmonisation, not the articulation of ethical commitments. Careful examination of the two documents reveals several important ethical issues that are addressed in the Declaration about which GCP is silent (panel). Thus reliance on GCP rather than on the Declaration of Helsinki may result in less protection for research 13

Comment

Panel: Requirements in latest revision of Declaration of Helsinki2 but absent in GCP6 • Investigators to disclose funding, sponsors, and other potential conﬂicts of interest to both research ethics committees and study participants • Study design to be disclosed publicly (eg, in clinical trial registries) • Research, notably that in developing countries, to beneﬁt and be responsive to health needs of populations in which it is done • Restricted use of placebo controls in approval process for new drugs and in research done in developing countries • Post-trial access to treatment • Authors to report results accurately, and publish or make public negative ﬁndings

participants. If so, the FDA’s action might lower the bar for international research under its purview—a scenario that has worried previous commentators.7,8 Third, the FDA’s departure from the Declaration of Helsinki could undermine its stated goals of clarity and regulatory harmonisation. For example, if many countries continue to use the Declaration, US researchers will encounter the same “confusion” that the FDA is attempting to prevent with its new rule. Similarly, should other countries follow the FDA’s lead and abandon the Declaration of Helsinki, the result could be the balkanisation of ethical standards in international research. In view of these concerns, we suggest the new US administration suspend this rule pending a review of the implications for US-sponsored research overseas. If such review conﬁrms our concerns, the FDA should be directed to rejoin the international community in requiring that studies be done in accordance with the Declaration of Helsinki. We also see an important role for major medical societies—though they lack regulatory authority, collectively these organisations can give voice to the commitment of medical researchers to

the Declaration’s high ethical standards. The American Society of Gene Therapy is considering policy on this issue9 and others should follow suit. *Jonathan Kimmelman, Charles Weijer, Eric M Meslin Biomedical Ethics Unit and Department of Social Studies of Medicine, McGill University, Montreal, QC, Canada H3A 1X1(JK); Departments of Philosophy and Medicine, University of Western Ontario, London, ON, Canada (CW); and Indiana University Center for Bioethics, Indianapolis, IN, USA (EMM) [email protected] We are members of the Canadian Network for the Governance of Ethical Health Research Involving Humans, supported by the Canadian Institutes of Health Research (CIHR). JK’s research is supported by a New Investigator award from CIHR. CW’s research is supported by a Canada Research Chair and CIHR operating grant. EMM’s research is supported in part by grant R25TW006070 from the Fogarty International Center at the National Institutes of Health. Our views do not necessarily represent the oﬃcial views of the CIHR, Fogarty International Center, or the NIH. We thank Daryl Pullman and Heather Sampson for contributions to this Comment. 1

2 3

4

5 6

7

8

9

Aschcroft RE. The Declaration of Helsinki. In: Emanuel EJ, Grady C, Crouch RA et al, eds. The Oxford Textbook of Clinical Research Ethics. New York: Oxford University Press, 2007. World Medical Association. Declaration of Helsinki. October, 2008. http://www.wma.net/e/policy/b3.htm (accessed Oct 25, 2008). Department of Health and Human Services, Food and Drug Administration. Human subject protection; foreign clinical studies not conducted under an investigational new drug application. Fed Reg April 28, 2008. 22800-16. http://www.fda.gov/cber/rules/forclinstud.pdf (accessed Nov 9, 2008). Department of Health and Human Services, Food and Drug Administration. Guidance for industry: acceptance of foreign studies. March, 2001. http://www.fda.gov/cber/gdlns/clinical031301.pdf (accessed Nov 9, 2008). Thiers FA, Sinskey AJ, Berndt ER. Trends in the globalization of clinical trials. Nat Rev Drug Discov 2008; 7: 13–14. International Conference on Harmonization. Guideline for good clinical practice (e6). 1996. http://www.ich.org/LOB/media/MEDIA482.pdf (accessed Oct 25, 2008). National Bioethics Advisory Commission. Ethical and policy issues in international research: clinical trials in developing countries, vol 1. 2001. http://bioethics.georgetown.edu/nbac/clinical/Vol1.pdf (accessed Oct 25, 2008). Nuﬃeld Council on Bioethics. The ethics of research related to healthcare in developing countries. 2002. http://www.nuﬃeldbioethics.org/ﬁleLibrary/ pdf/errhdc_fullreport001.pdf (accessed Nov 9, 2008). Friedmann T. The ASGT and ethical codes for research. Mol Ther 2008; 16: 1643–44.

Cardiology: a call for papers To submit a paper go to http://ees.elsevier.com/thelancet

14

The Lancet is dedicating a special issue to cardiology to coincide with the American College of Cardiology meeting to be held in Orlando, FL, USA, on March 28 to April 1, 2009. In particular, we will consider high-quality research papers that report results of randomised trials. If your work is being presented at the meeting and falls under an embargo policy, please tell us the date, time, and manner of presentation (poster or oral). If your paper is accepted here, publication on

our website can be scheduled to coincide with the presentation. The deadline for submissions is March 2, 2009, via our online submission system. Please state in your covering letter that the submission is in response to this call for papers.

Stuart Spencer The Lancet, London NW1 7BY, UK

www.thelancet.com Vol 373 January 3, 2009

Comment

Panel: Requirements in latest revision of Declaration of Helsinki2 but absent in GCP6 • Investigators to disclose funding, sponsors, and other potential conﬂicts of interest to both research ethics committees and study participants • Study design to be disclosed publicly (eg, in clinical trial registries) • Research, notably that in developing countries, to beneﬁt and be responsive to health needs of populations in which it is done • Restricted use of placebo controls in approval process for new drugs and in research done in developing countries • Post-trial access to treatment • Authors to report results accurately, and publish or make public negative ﬁndings

participants. If so, the FDA’s action might lower the bar for international research under its purview—a scenario that has worried previous commentators.7,8 Third, the FDA’s departure from the Declaration of Helsinki could undermine its stated goals of clarity and regulatory harmonisation. For example, if many countries continue to use the Declaration, US researchers will encounter the same “confusion” that the FDA is attempting to prevent with its new rule. Similarly, should other countries follow the FDA’s lead and abandon the Declaration of Helsinki, the result could be the balkanisation of ethical standards in international research. In view of these concerns, we suggest the new US administration suspend this rule pending a review of the implications for US-sponsored research overseas. If such review conﬁrms our concerns, the FDA should be directed to rejoin the international community in requiring that studies be done in accordance with the Declaration of Helsinki. We also see an important role for major medical societies—though they lack regulatory authority, collectively these organisations can give voice to the commitment of medical researchers to

the Declaration’s high ethical standards. The American Society of Gene Therapy is considering policy on this issue9 and others should follow suit. *Jonathan Kimmelman, Charles Weijer, Eric M Meslin Biomedical Ethics Unit and Department of Social Studies of Medicine, McGill University, Montreal, QC, Canada H3A 1X1(JK); Departments of Philosophy and Medicine, University of Western Ontario, London, ON, Canada (CW); and Indiana University Center for Bioethics, Indianapolis, IN, USA (EMM) [email protected] We are members of the Canadian Network for the Governance of Ethical Health Research Involving Humans, supported by the Canadian Institutes of Health Research (CIHR). JK’s research is supported by a New Investigator award from CIHR. CW’s research is supported by a Canada Research Chair and CIHR operating grant. EMM’s research is supported in part by grant R25TW006070 from the Fogarty International Center at the National Institutes of Health. Our views do not necessarily represent the oﬃcial views of the CIHR, Fogarty International Center, or the NIH. We thank Daryl Pullman and Heather Sampson for contributions to this Comment. 1

2 3

4

5 6

7

8

9

Aschcroft RE. The Declaration of Helsinki. In: Emanuel EJ, Grady C, Crouch RA et al, eds. The Oxford Textbook of Clinical Research Ethics. New York: Oxford University Press, 2007. World Medical Association. Declaration of Helsinki. October, 2008. http://www.wma.net/e/policy/b3.htm (accessed Oct 25, 2008). Department of Health and Human Services, Food and Drug Administration. Human subject protection; foreign clinical studies not conducted under an investigational new drug application. Fed Reg April 28, 2008. 22800-16. http://www.fda.gov/cber/rules/forclinstud.pdf (accessed Nov 9, 2008). Department of Health and Human Services, Food and Drug Administration. Guidance for industry: acceptance of foreign studies. March, 2001. http://www.fda.gov/cber/gdlns/clinical031301.pdf (accessed Nov 9, 2008). Thiers FA, Sinskey AJ, Berndt ER. Trends in the globalization of clinical trials. Nat Rev Drug Discov 2008; 7: 13–14. International Conference on Harmonization. Guideline for good clinical practice (e6). 1996. http://www.ich.org/LOB/media/MEDIA482.pdf (accessed Oct 25, 2008). National Bioethics Advisory Commission. Ethical and policy issues in international research: clinical trials in developing countries, vol 1. 2001. http://bioethics.georgetown.edu/nbac/clinical/Vol1.pdf (accessed Oct 25, 2008). Nuﬃeld Council on Bioethics. The ethics of research related to healthcare in developing countries. 2002. http://www.nuﬃeldbioethics.org/ﬁleLibrary/ pdf/errhdc_fullreport001.pdf (accessed Nov 9, 2008). Friedmann T. The ASGT and ethical codes for research. Mol Ther 2008; 16: 1643–44.

Cardiology: a call for papers To submit a paper go to http://ees.elsevier.com/thelancet

14

The Lancet is dedicating a special issue to cardiology to coincide with the American College of Cardiology meeting to be held in Orlando, FL, USA, on March 28 to April 1, 2009. In particular, we will consider high-quality research papers that report results of randomised trials. If your work is being presented at the meeting and falls under an embargo policy, please tell us the date, time, and manner of presentation (poster or oral). If your paper is accepted here, publication on

our website can be scheduled to coincide with the presentation. The deadline for submissions is March 2, 2009, via our online submission system. Please state in your covering letter that the submission is in response to this call for papers.

Stuart Spencer The Lancet, London NW1 7BY, UK

www.thelancet.com Vol 373 January 3, 2009

World Report

Text messages could hasten tuberculosis drug compliance Mobile phone technologies are being harnessed to help patients with tuberculosis keep up with their treatment regimens, with some promising preliminary results. Eliza Barclay reports. A handful of new technologies designed to connect tuberculosis patients with their caregivers using text messaging, or SMS (short message service), hold potential for helping to improve adherence to gruelling drug regimens. The treatment for tuberculosis is a combination of strong antibiotics that must be taken daily for at least 6 months. But side-eﬀects, such as nausea and heartburn, dissuade some patients from sticking with the treatment. Other patients endure the side-eﬀects only to drop the medication once they feel better 1 or 2 months into the regimen. The Stop TB Partnership reports that on average 5% of tuberculosis patients abandon treatment, although the ﬁgure is as high as 20% in some countries. Those patients who cannot or do not adhere to the treatment remain infectious longer and are more likely to relapse and die. They are also vulnerable to developing multidrug-resistant tuberculosis, a strain resistant to two or more ﬁrst-line drugs, or extensively drug-resistant tuberculosis, a strain resistant to three or more second-line drugs. To help people complete their treatment WHO recommends a strategy known as DOTS (directly observed treatment, short course). As part of this strategy, a health worker or a tuberculosis treatment supporter watches the patient take their antibiotics every day. Patients also receive counselling and support to ensure they do not stray from their treatment course. Although DOTS has helped to dramatically improve tuberculosis control around the world, it is insuﬃcient or inaccessible for thousands of patients. Many seek treatment in private clinics without close monitoring from their doctor. “Patients that most often fall through the cracks are usually the ones www.thelancet.com Vol 373 January 3, 2009

that ﬁrst [go] to private clinics”, said Aamir Khan, the executive director of Interactive Researchand Development in Karachi, Pakistan and a faculty member at Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA. “While the DOTS programme is not perfect, my experience is that there

“...disease control and technology specialists are now looking to SMS as a costeﬀective way to communicate with...hard-to-reach patients...“ are many more patients in the private sector that have been mismanaged and abused by unscrupulous health practitioners.” Khan noted that private doctors often overprescribe. He has seen patients taking nine expensive antibiotics when only four are required. Other experts note that DOTS is expensive and human-resource intensive. They say it is unrealistic to expect health workers to monitor patients on a daily basis. Several disease control and technology specialists are now looking to SMS as a cost-eﬀective way to communicate with and monitor hard-

to-reach patients in remote locations. “The problem is enormous, and everything has to be done in order to prevent patients from defaulting”, said Mario Raviglione, director of WHO’s Stop TB Department. “Anything that can be done technologically to help solve this issue like these cellphone technologies would be useful.” Although tuberculosis is a disease aﬀecting poor people, even those living on US$1 per day increasingly have access to mobile phones. There are more than 3·3 billion mobile-phone subscriptions worldwide. By the end of 2006, according to the International Telecommunications Union, 68% of those subscriptions were in developing countries. South Africa has proven a fertile testing ground for new drugcompliance technologies. According to WHO, South Africa had a 71% DOTS treatment success rate in 2005; most patients who were not successfully treated under DOTS defaulted on their treatment. Several companies have addressed the compliance problem by developing devices that remind patients to take their medication, and feature back-

The printed journal includes an image merely for illustration

A nurse explains the treatment for tuberculosis to a patient in the Democratic Republic of the Congo

15

Getty Images

World Report

A mobile phone owner in Khayelitsha, one of the poorest townships in South Africa

up links to health workers, friends, or family members if the patient fails to respond to the ﬁrst reminder. One reminder product is a small pill bottle made by London-based SIMpill that contains a SIM card and when opened, the SIM card delivers a SMS with a unique pill box identiﬁcation number to a central server. The central server receives the incoming SMS and stores the data, but if no SMS is received at the designated time, the server contacts the patient via phone alerting them to take their medication. If the patient does not respond, the server contacts a caregiver who can follow-up with the patient. From July, 2006, to April, 2007, SIMpill did a pilot study in 155 tuberculosis patients at three clinics in the Cape Town area with the Western Cape Department of Health. After patients used the SIMpill for 10 months, drug adherence stabilised between 86–92% with a treatment success rate of 94%, according to SIMpill. Ann-Mari Albertsen, managing director of SIMpill, says, in addition to helping patients adhere to their treatment, SIMpill also frees up health workers from daily observation of patients taking their medication. Albertsen noted that with SIMpill, a nurse could keep tabs on 50–60 patients rather than just ten patients. “We see staﬀ focusing on other parts of their job, like counselling, training, and actual follow-up with 16

patients who need more attention, instead of keeping up with tablets and glass of water”, said Albertsen. Albertsen would not reveal the cost of the technology but says the company is signing a contract with a major international public-health organisation working in South Africa, which plans to deploy the product on a larger scale. SIMmed, developed by CompuTainer, is a direct competitor of SIMpill in South Africa, and claims to be a less expensive and equally eﬀective product. Instead of monitoring patients through a pill bottle, SIMmed asks patients to press the speed dial button on their mobile phone after taking their medication. The number dials into a server, which records the medication event, and reminds the patient by SMS or contacts a caregiver if the patient fails to call. The ﬁrst SIMmed trial in Khayelitsha, South Africa, yielded a compliance rate of over 90%. CompuTainer is now working with the South African Government to reach 45 000 new patients with the programme. Other groups are using SMS messaging in combination with economic incentives to improve treatment adherence. X out TB is an SMS-based system invented by a group of scientists and entrepreneurs with the Innovations for International Health project at the Massachusetts Institute of Technology, Cambridge, MA, USA. Each day after taking their medication, tuberculosis patients urinate on a ﬁlter paper diagnostic, which detects the metabolites of the ﬁrst-line tuberculosis drug isoniazid and reveals a code. The patient sends the code over SMS where it is stored on a server. Based on accumulated right answers (correct codes) at the end of the month, patients receive rewards. In the ﬁrst trial in Nicaragua, patients received $2 worth of mobile phone minutes each month. “We are concerned people just don’t behave rationally in conditions like DOTS”, said Jose Gomez-Marquez, programme director of Innovations

for International Health. “If they can shortcut something, they’ll do it. But if there’s an economic reward involved, there’s a bigger incentive to follow the rules of the programme.” In the next X out TB trial in Pakistan, Khan will target patients attending DOTS clinics. He said X out TB would like to compare compliance between the routine DOTS programme and one with X out TB integrated into it. However, some experts question whether SMS technologies will eﬀectively replace thorough face-toface monitoring from a community health worker. Partners in Health (PIH), a Boston-based non-proﬁt organisation, working in several of the poorest developing countries has developed one of the most eﬀective programmes for tuberculosis treatment using community health workers. Every day, the workers visit patients in their homes to supervise treatment. PIH also developed DOTS Plus, a regimen for treating multidrug-resistant tuberculosis, where patients receive daily nutritional support as a supplement and incentive for treatment. According to Hamish Fraser, director of informatics and telemedicine for PIH and assistant professor at Harvard Medical School, the success of PIH’s programmes without the use of SMS communication indicate that SMSbased health technologies may be unnecessary. “I think in developing countries, having a DOTS worker visit patient in their home is extremely eﬀective”, said Fraser. “We don’t immediately feel there’s a big gap there so I’m less sold on cellphones.” Raviglione, however, believes that SMS health technologies could have a role in improving communication in tuberculosis treatment and care. “Though the human aspect of tuberculosis care and control must not be forgotten or underemphasised, there’s always great value in increasing communication between the patients and the clinicians”, he said.

Eliza Barclay www.thelancet.com Vol 373 January 3, 2009

World Report

Alcohol use on the rise in India With more than half of all alcohol drinkers in India falling into the criteria for hazardous drinking, alcohol abuse is emerging as a major public-health problem in the country. Raekha Prasad reports.

www.thelancet.com Vol 373 January 3, 2009

that the “average age of initiation” had dropped from 19 years to 13 years in the past two decades. The centre points out that a “powerful international and domestic alcohol lobby” is purposely targeting young Indians. The local industry has introduced ﬂavoured alcohol drinks to attract previously non-drinking women and young men. Multinational companies have identiﬁed India with

“The shifting composition of Indian drinkers has seen a rise in the number of Indian women drinking regularly and heavily.” its vast unexploited markets as one of the world’s most sought after places for investment. Many alcohol adverts now feature spirited groups of young people having a good time. Although alcohol advertising is banned in the electronic and print media, surrogate advertising is rife, argues Monika Arora, director of the NGO, Health Related Information Dissemination Amongst Youth Student Health Action Network. “Drinking water and apple juice is packaged by alcohol companies. It’s all about getting young people to start early and be life-long consumers. Bollywood ﬁlms now glorify alcohol where the good guys drink.”

The shifting composition of Indian drinkers has seen a rise in the number of Indian women drinking regularly and heavily. One recent study in the southern state of Karnataka found young women consumed similar amounts of alcohol to young men on any typical drinking occasion. What is of particular concern—and an important indicator of health risks—is that the signature pattern of alcohol consumption in India is frequent and heavy drinking. More than half of all drinkers fall into the criteria for hazardous drinking, which is characterised by bingeing and solitary consumption to the point of intoxication. Moreover, spirits account for 95% of the beverages drunk in India. Another problem for policy makers is the fact that two thirds of the alcohol drunk in India is unrecorded because it is either illicit local home brew or has been smuggled into the country. Employers in poor, marginalised communities sometimes pay wages in alcohol rather than cash, according to WHO. The hazards of spurious liquor can be fatal, with frequent reports of death, disability, and hospitalisation resulting from its consumption across the country. One barrier to developing a national alcohol policy for India, experts say, is the woeful lack of data and research on

Getty Images

India’s reputation as a country with a culture of abstinence especially in matters regarding alcohol is underserved, say experts. The country, which has seen a rapid proliferation of city bars and nightclubs in recent years, is fast shedding its inhibitions about alcohol as a lifestyle choice. This situation has led to fears of an undocumented rise in alcohol abuse not only among poorer classes but also in sections of society that were previously considered dry. The health minister has recognised the scale of the problem—and has called for a policy that will regulate sales and the pricing of drink. Many experts say that although this move is welcome it may not be enough to curb the harmful eﬀects of the rise in alcohol consumption in society. The increasing production, distribution, and promotion of alcohol has already seen drink-related problems emerging as a major publichealth concern in India. Sales of alcohol have seen a growth rate of 8% in the past 3 years. Oﬃcially, Indians are still among the world’s lowest consumers of alcohol— government statistics show only 21% of adult men and around 2% of women drink. But up to a ﬁfth of this group—about 14 million people—are dependent drinkers requiring “help”. The concern, say experts, is that there has been a rapid change in patterns and trends of alcohol use in India. Chief among them is people are beginning to drink at ever-younger ages. The percentage of the drinking population aged under 21 years has increased from 2% to more than 14% in the past 15 years, according to studies in the southern state of Kerala by Alcohol and Drugs Information Centre India, a non-governmental organisation (NGO). Alarmingly, the study found

17

World Report

The printed journal includes an image merely for illustration

Billboard advertising alcoholic drinks in Mumbai

its national health, social, and economic eﬀect. What is known is that alcoholrelated problems account for more than a ﬁfth of hospital admissions; 18% of psychiatric emergencies; more than 20% of all brain injuries and 60% of all injuries reporting to India’s emergency rooms. The role of alcohol in domestic violence is substantial: a third of violent husbands drink, according to a WHO study in 2004. Most of the violence took place during intoxication. There is evidence even to suggest that the poor are beginning to drink more than they earn—a deadly spiral of alcohol and debt. One recent study by the National Institute of Mental Health and Neuro Sciences (NIMHANS) in households of rural, urban, town, and slum populations of 28 500 people in and around the city of Bangalore, Karnataka, found that the average monthly expenditure on alcohol of patients with alcohol addiction is more than the average monthly salary. Although the Indian constitution includes the prohibition of alcohol among its directive principles, alcohol policy is devolved to individual states—as is the levying of taxes on it. Since most states derive around a ﬁfth of their revenue from alcohol taxation—the second largest source after sales tax—they are generally ambivalent towards stemming its ﬂow. Moreover, there is a long history 18

in India of a powerful alcohol lobby with industry ﬁgures inﬂuencing the political process, both in the form of party donations and as representatives. But experts argue that Indian society is losing considerably more than it gains. “Because of the political expediency surrounding prohibition, what is not being looked at is demand reduction strategies”, says Vivek Benegal, one of the authors of the report and assistant professor of psychiatry at NIMHANS. Using their ﬁndings in the Bangalore study, researchers from NIMHANS have calculated that the direct and indirect costs attributable to alcohol addiction is more than triple the proﬁts of alcohol taxation and several times more than the annual health budget of Karnataka. Extrapolating their ﬁndings to the whole of India they estimate the total alcohol revenue for 2003–04 of 216 billion rupees falls 28 billion rupees short of the total cost of managing the eﬀects of alcohol addiction. These included the tangible costs of health care, occupational, ﬁnancial, social, and legal factors. The oﬃcial response to India’s problem remains focused on those in acute need rather than on prevention. This situation means that oﬃcial policy concentrates on just the 4% of the alcohol-dependent adult male population—and ignores the 20% of the population who are “at risk” of serious alcohol abuse. Experts argue that government thinking on how best to mitigate the risks for alcohol are 20 years behind that of tobacco. Under its National Drug De-addiction Programme, the Government of India has funded 483 detoxiﬁcation and 90 counselling centres. Almost half of attendees are being treated for alcohol dependency. But the success of the programmes is low and states fail to adequately fund them, health professionals say. Doctors working with addicts in government hospitals report a “complete lack” of non-pharmacological care and training. “Once we’ve treated them there’s no social worker or clinical psychologist

to refer them to so we just send them to AA (Alcoholics Anonymous)”, says Smita Deshpande, a senior psychiatrist working in a Delhi state hospital. The problem is that the treatment of alcoholism is a low priority in Indian’s health sector, says Rajat Ray, professor and chief of the National Drug Dependence Treatment Centre at All India Institute of Medical Sciences. (AIIMS). Just 600 doctors have been trained to treat alcohol abuse in the past decade. “It’s seen as deviant behaviour among most doctors: a hopeless situation that is unrewarding to treat and so there’s no motivation or ﬁnancial incentive on doctors to work in this ﬁeld”, Ray says. To address this, the Indian Government has set a target to train, via AIIMS, 1000 doctors, as many paramedics, and 500 nurses to specialise in alcohol-abuse treatment in the next 4 years. Once trained, the plan is to deploy them across India’s 560 district hospitals to increase access to treatment. Ray and his team are currently piloting three district training projects in Madhya Pradesh, Assam, and Uttar Pradesh. There is, however, a growing lobby urging the health ministry to act. Indian Alcohol Policy Alliance, an NGO aiming to prevent alcohol-related harm through evidence-based policy intervention, says that the key is to break the stranglehold of state revenue departments who see increasing consumption of alcohol as a boon to treasury coﬀers. It is pressing the ministry of health, headed by a minister who has advocated prohibition in certain states, to take a lead in passing a law that privileges public health over tax receipts. The lack of a national alcohol policy creates “a very diﬃcult situation” for health professionals working to tackle alcoholism, Ray says. Discussions are taking place, but an actual policy, he says, “is still in a formative state”.

Raekha Prasad www.thelancet.com Vol 373 January 3, 2009

Perspectives

Book Time to medicalise child maltreatment Last year marked the 50th anniversary of the founding of one of the original multidisciplinary childprotection teams, in the US city of Denver, by paediatrician C Henry Kempe and psychiatrist Brandt F Steele. Their experience with nine children treated on the paediatric service of Colorado General Hospital led to the publication of their landmark paper, “The Battered Child Syndrome” (JAMA 1962; 181: 17–24). That paper estimated that there might be as many as 749 abused children in the USA annually at that time. On the basis of the public and professional horror that such maltreatment could be happening, all 50 US states passed laws in the mid-1960s that mandated all professionals to report suspected cases of abuse and neglect, and ﬁxed the responsibility for investigating and treating the children and families with public child-welfare agencies. In 1975, Kempe brought several dozen of his colleagues from around the world to a meeting in Bellagio, Italy. This gathering led to the founding of the International Society for the Prevention of Child Abuse and Neglect, the First International Congress held in Geneva, in 1976, and many other eﬀorts to address child maltreatment throughout the world, including work by WHO and UNICEF. The next three decades saw consistent increases in the number of reports of child maltreatment in the USA: from 60 000 in 1970, to 669 000 in 1980, to well over 2 million in the early 1990s. The re-recognition of sexual abuse in the late 1970s and 1980s signiﬁcantly complicated the ability of child-welfare agencies to function adequately, since that form of abuse is clearly a crime and requires law-enforcement involvement (whereas the beating of children in the earlier period was not noted for such interest by the law-enforcement www.thelancet.com Vol 373 January 3, 2009

agencies). In 1990, the US Advisory Board on Child Abuse and Neglect called the situation “a national emergency”. Their report, among many recommendations, decried the lack of data collected by federal, state, and local agencies that could guide policy on child maltreatment, called for the professionalisation of childprotection-services personnel, pled

“…the single greatest diﬃculty we faced as a specialty in the USA is that child maltreatment is viewed by the public and our politicians as a social problem, not as a health problem.” for the speciﬁc elucidation of childprotection policies at national and local levels, and suggested a focus on prevention. The Board’s recommendations were mostly ignored and the Board was eliminated by the Clinton Administration. Nearly two decades later, David Finkelhor, a long-time scholar in the ﬁeld, has published a must-read book, Childhood Victimization: Violence, Crime and Abuse in the Lives of Young People. There is a great deal to learn about the ﬁelds of child maltreatment and violence in this book, but in particular the chapter called “Good news: child victimization is declining—but why?” is enormously important from my perspective. I have worked in child maltreatment since 1981. I moved to Denver in 1968 to work with Henry Kempe after hearing his presentation on “The Battered Child” as a medical student in New York in 1967. For all those years, the assumption by the public and professionals has been that child maltreatment was increasing. So caught up were many of us in our zeal to help increase public and professional awareness that we always spoke of the epidemic of more and more cases in somewhat desperate attempts to

get attention from those in political power and from those philanthropists who have fuelled most of the research, training, and programmatic activity in the USA. We decried the lack of precise data that would tell us how many child abuse fatalities there were each year, and how many children were being maltreated. We always assumed things were getting worse. We almost needed it to be getting worse. Now, it seems, things are improving in the USA. Finkelhor has tracked the notoriously poor (but consistently poor) child maltreatment estimates of incidence in the USA and has shown that physical and sexual abuse has declined by 46% and 51%, respectively, between 1990 and 2005. I guess I could claim that the US Advisory Board reports were a success after all, but I know better than that. I will not go into all the proposed reasons why this decline could be happening—there are 11 that are enumerated by Finkelhor—but I will highlight one because in my personal view, just the possibility that it could be true may be enough to tweak the interest of medical scientists and researchers and help maintain the decline, and even accelerate it. I refer explicitly to the possibility that the dramatic increase in the use of psychopharmacological agents in our population over the past decade could explain part of the decline in rates of child maltreatment. I believe this is an important hypothesis to study further because, quite independently, there has been increasing attention paid to the genetic and biological aspects of abusive and neglectful behaviour in animal and human studies. In a 1996 editorial written as editorin-chief of Child Abuse and Neglect: The International Journal, I asked our readers “Suppose it were a genetic problem?” I asked this on the basis of a study of knockout mice missing

Childhood Victimization: Violence, Crime, and Abuse in the Lives of Young People David Finkelhor. Oxford University Press, 2008. Pp 248. US$37·95. ISBN 0-195-34285-2.

19

Perspectives

the immediate early gene fosB. I was accused by many of our readers of “medicalising” a social problem. The accusation is valid, but I do not view it as a pejorative. Rather, I have always felt that the single greatest diﬃculty we faced as a specialty in the USA is that child maltreatment is viewed by the public and our politicians as a social problem, not as a health problem. Truth be told, we do not raise private funds for social problems, and further, our politicians will not put government resources where there is not a lot of public support. In the past 5 years, the US National Institutes of Health allocated only about US$36 million of its nearly $30 billion budget to study child maltreatment. The American Board of Pediatrics has certiﬁed Child Abuse Pediatrics as a subspecialty, but there are few funds

available for training or research to fuel the growth of this discipline. Finkelhor picks up many of the recommendations of the Advisory Boards and Commissions of yesteryear in his constructive suggestions that include the professionalisation of workers in child-protection services and increasing community-based prevention programmes. He also points out that no one has ever accused the child-protection system anywhere of having an evidence base to their practice. Few, if any, systems collect and maintain data for the decades necessary to know whether their practice has been of quality. This is certainly something the medical profession can bring to bear. Could we imagine that certain drug therapies could help depressed parents treat their children better? Could we imagine that certain therapies could

even out the behaviour of individuals who ﬁnd themselves exploding with rage at some of the behaviours of their children? Could we actually have clinical trials that include the standard social interventions of “parenting classes” with and without pharmacotherapy? We have spent a half century building the child welfare and legal responses to child maltreatment. These systems are struggling under the weight of the task. Although we must acknowledge how much they have done given the capacity and resources they have had, the time is now right to spend the next decade building what has been a far less well developed medical response to its full potential. Abused children and their families deserve no less.

Richard Krugman [email protected]

In brief Seventh Art Releasing

Film A child soldier’s story

Feuerherz (Heart of Fire) Directed by Luigi Falorni. TV60 Filmproduktion, 2008. See http://www.feuerherz-info.de.

20

Modern ﬁrearms are lightweight. Automatic weapons such as the AK-47 weigh a few kilograms. They don’t have the crashing recoil of earlier models. With practice, a child aged 9 or 10 years can operate an AK-47: it’s just a case of becoming accustomed to ﬁring the thing—machine guns don’t require much aim. The superb Heart of Fire is set in 1981. Eritrea is in the midst of a lengthy war of independence with Ethiopia and Eritrean militias are ﬁghting among themselves. A previous conﬂict had been resolved in 1974, but in 1980 it reignited. The Eritrean Liberation Front—also known as the Jebha— battle the Eritrean People’s Liberation Front—the Shabia—for control of the ﬁght against the Ethiopians. Letekidan Micael plays the ﬁerytempered Awet, a 10-year-old girl delivered by her blustery father into the hands of the Jebha. The troop is

led by the charismatic Ma’aza (Seble Tilahun), a wild-haired and pretty woman in her twenties. Children learn from adults: Awet copies Ma’aza’s hairstyle, and she and the other children—known within the troop as “pioneers”—demand guns of their own. They are provided with wooden machine guns, and given rudimentary lessons in geography. But their education is soon trammelled into lessons on guerilla warfare, their ersatz weapons replaced by real ones. These are children without sanctuary: the relentless sun and dusty landscape oﬀers no protection, the adult world is endlessly and violently treacherous. The mayhem of war is powerfully conveyed. It is brutalising to dodge bullets—director Luigi Falorni reminds us—to watch those around you die, and to be forced to kill. It’s an excellent performance from Micael and an absorbing and moving ﬁlm. Awet’s burgeoning understanding and

unshakeable morality lead her to resist Ma’aza’s murderous inculcations. Senait Mehari—on whose memoir this ﬁlm is based—ﬂed to Germany in 1983, where she remains today. In Eritrea, the Shabia drove the Jebha into Sudan, thus putting an end to the second act of the civil war, itself an entr’acte in a war of independence that served as prelude to the 1998– 2000 war between the separated nations of Eritrea and Ethiopia; that war destabilised the entire region, and last year, UN Secretary-General Ban Ki-moon warned that hostilities could easily resume. Eritrea is presently involved in skirmishes with Djiboutian soldiers. War exerts a force of its own, dragging those around into the chaos; UNICEF estimates that 300 000 children serve as soldiers in armed conﬂicts around the world.

Talha Burki [email protected]

www.thelancet.com Vol 373 January 3, 2009

Perspectives

the immediate early gene fosB. I was accused by many of our readers of “medicalising” a social problem. The accusation is valid, but I do not view it as a pejorative. Rather, I have always felt that the single greatest diﬃculty we faced as a specialty in the USA is that child maltreatment is viewed by the public and our politicians as a social problem, not as a health problem. Truth be told, we do not raise private funds for social problems, and further, our politicians will not put government resources where there is not a lot of public support. In the past 5 years, the US National Institutes of Health allocated only about US$36 million of its nearly $30 billion budget to study child maltreatment. The American Board of Pediatrics has certiﬁed Child Abuse Pediatrics as a subspecialty, but there are few funds

available for training or research to fuel the growth of this discipline. Finkelhor picks up many of the recommendations of the Advisory Boards and Commissions of yesteryear in his constructive suggestions that include the professionalisation of workers in child-protection services and increasing community-based prevention programmes. He also points out that no one has ever accused the child-protection system anywhere of having an evidence base to their practice. Few, if any, systems collect and maintain data for the decades necessary to know whether their practice has been of quality. This is certainly something the medical profession can bring to bear. Could we imagine that certain drug therapies could help depressed parents treat their children better? Could we imagine that certain therapies could

even out the behaviour of individuals who ﬁnd themselves exploding with rage at some of the behaviours of their children? Could we actually have clinical trials that include the standard social interventions of “parenting classes” with and without pharmacotherapy? We have spent a half century building the child welfare and legal responses to child maltreatment. These systems are struggling under the weight of the task. Although we must acknowledge how much they have done given the capacity and resources they have had, the time is now right to spend the next decade building what has been a far less well developed medical response to its full potential. Abused children and their families deserve no less.

Richard Krugman [email protected]

In brief Seventh Art Releasing

Film A child soldier’s story

Feuerherz (Heart of Fire) Directed by Luigi Falorni. TV60 Filmproduktion, 2008. See http://www.feuerherz-info.de.

20

Modern ﬁrearms are lightweight. Automatic weapons such as the AK-47 weigh a few kilograms. They don’t have the crashing recoil of earlier models. With practice, a child aged 9 or 10 years can operate an AK-47: it’s just a case of becoming accustomed to ﬁring the thing—machine guns don’t require much aim. The superb Heart of Fire is set in 1981. Eritrea is in the midst of a lengthy war of independence with Ethiopia and Eritrean militias are ﬁghting among themselves. A previous conﬂict had been resolved in 1974, but in 1980 it reignited. The Eritrean Liberation Front—also known as the Jebha— battle the Eritrean People’s Liberation Front—the Shabia—for control of the ﬁght against the Ethiopians. Letekidan Micael plays the ﬁerytempered Awet, a 10-year-old girl delivered by her blustery father into the hands of the Jebha. The troop is

led by the charismatic Ma’aza (Seble Tilahun), a wild-haired and pretty woman in her twenties. Children learn from adults: Awet copies Ma’aza’s hairstyle, and she and the other children—known within the troop as “pioneers”—demand guns of their own. They are provided with wooden machine guns, and given rudimentary lessons in geography. But their education is soon trammelled into lessons on guerilla warfare, their ersatz weapons replaced by real ones. These are children without sanctuary: the relentless sun and dusty landscape oﬀers no protection, the adult world is endlessly and violently treacherous. The mayhem of war is powerfully conveyed. It is brutalising to dodge bullets—director Luigi Falorni reminds us—to watch those around you die, and to be forced to kill. It’s an excellent performance from Micael and an absorbing and moving ﬁlm. Awet’s burgeoning understanding and

unshakeable morality lead her to resist Ma’aza’s murderous inculcations. Senait Mehari—on whose memoir this ﬁlm is based—ﬂed to Germany in 1983, where she remains today. In Eritrea, the Shabia drove the Jebha into Sudan, thus putting an end to the second act of the civil war, itself an entr’acte in a war of independence that served as prelude to the 1998– 2000 war between the separated nations of Eritrea and Ethiopia; that war destabilised the entire region, and last year, UN Secretary-General Ban Ki-moon warned that hostilities could easily resume. Eritrea is presently involved in skirmishes with Djiboutian soldiers. War exerts a force of its own, dragging those around into the chaos; UNICEF estimates that 300 000 children serve as soldiers in armed conﬂicts around the world.

Talha Burki [email protected]

www.thelancet.com Vol 373 January 3, 2009

Perspectives

Proﬁle Paulo Sérgio Pinheiro: giving a voice to children In early 2003, the then UN Secretary-General Koﬁ Annan asked the Brazilian human rights expert Paulo Sérgio Pinheiro to lead a study into the violence perpetrated against the world’s children. During the next 3 years, he worked to expose the scope of violence against children and its impact on their lives. Pinheiro’s report made unhappy reading. Violence against children was widespread, he found: about 53 000 children died in 2002 as a result of homicide, while 150 million girls and 73 million boys under the age of 18 years experienced forced sexual intercourse or other forms of sexual violence. In 2004, 218 million children were involved in child labour, of whom 126 million worked in hazardous conditions. But these ﬁgures do not tell the full story, as Pinheiro told the UN General Assembly in 2006. Although the world’s media focuses attention on brutal crimes, such as sexual abuse and human traﬃcking, other more insidious forms of violence take place every day in homes and schools. “Much violence against children, whether in the family, schools, alternative care and justice institutions, the workplace or the community is implicitly socially condoned or legally sanctioned”, he told the international community. In an interview with The Lancet, Pinheiro describes these everyday places as the “invisible sanctuaries” of violence against children, and argues that they challenge the fundamental principles of democracy. “There is a basic contradiction in the world today”, he says. “A contradiction between the advancement of democracy all over the world and the continuation of the authoritarian treatment of children. There are millions of children suﬀering this authoritarianism. It’s a kind of power-struggle. Governments, families, and caregivers feel threatened if the child has a voice.” Bringing together the report allowed Pinheiro a direct channel to the voices of the world’s children. Their message came through loud and clear. “All over the world, from Europe to Palestine, they can’t bear parents beating them. It was a sort of chorus—children complaining about teachers, about parents, about caregivers. For me this was very dramatic.” Although the report’s scope stretched far beyond the realm of corporal punishment, Pinheiro says he found himself becoming “militant” about banning all corporal punishment. “I think this is the basic message of the report”, he says. “This the children understood very well.“ Pinheiro’s own childhood in Rio de Janeiro, in the 1940s and 1950s, was happily free from violence. “I think that I had a very protected childhood”, he says. “My grandmother was the oldest of her sisters and I remember having many great-aunts around me. Looking back, I think this helped me to have a lot of self-assurance. It was a very important www.thelancet.com Vol 373 January 3, 2009

foundation for my evolution.” Pinheiro adds that “I think most of the people of my generation have been smacked, but I on the contrary have a sensation of protection and I think that this has helped me to do more things than my social class would give conditions for me to do.” Pinheiro’s family had him earmarked for a diplomatic career, so his grandmother paid for private tuition in English and French from an early age. Later, he earned a law degree but was then unwilling to work with the military dictatorship that had been established in Brazil in 1964. Instead, he left for Paris, France, where he earned a doctorate and was profoundly changed by the student protests in May, 1968. “I think my real turning point was coming to France.” Returning to Brazil, he began what has proven to be a long and fruitful international academic career, focusing on topics including social history and police violence. In the mid-1990s, Pinheiro was rapporteur of the Brazilian National Human Rights Plan, and began working with the United Nations—among other things as a Special Rapporteur on the situation of human rights in Burundi and, later, in Burma (Myanmar). He also served as Secretary of State for Human Rights in Brazil from 2001 to 2002. “Looking back, I think that most of the phases of my career were unexpected”, he says. “I never planned to do what I have done. I think that all these phases were possible because I have a sense of self-suﬃciency. Not a single one of these involvements was very dramatic—they just came.” Taking on the role of Independent Expert for the United Nations Secretary General’s Study on Violence Against Children in 2003, Pinheiro was surprised at the positive attitude of countries. “I was able to work very closely with the Arab League and the Conference of Islamic countries in a very positive basis, and countries like Thailand or China. Even countries that are very enthusiastic about corporal punishment, they supported the report as a whole.” Now 64-year-old Pinheiro is still working to further the rights of children, as one of the seven commissioners of the Organization of American States’s Inter-American Commission on Human Rights. His academic work in political science, at Brown University, Providence, RI, USA, also continues. Meanwhile, he hopes the UN report on violence against children may have an impact. “Of course the law is not a magic wand, but even if law isn’t immediately implemented, the report is a framework for implementation by civil society. Without a framework, the struggle to combat violence against children will be much more diﬃcult.”

Stephen Pincock [email protected]

21

Perspectives

The art of medicine Landscape and health

The Bridgeman Art Library

“My soul hurt”, writes Paul Theroux at the beginning of The Happy Isles of Oceania, “my heart was damaged. I was lonely. I did not want to see another big city. I wanted to be puriﬁed by water and wilderness.” His marriage coming apart, the threat of cancer hanging over him, Theroux gives in to what must be an ancient impulse: he turns his back on the known and the civilised and strikes out into the wild. “Where there is wilderness”, he writes a little later, “there is hope”. The link between health and certain kinds of natural landscape is interesting and strong. Although the technology of medicine is increasingly metropolitan, the search for health has often led out of the city and into nature. Where the city has been seen as a source of sickness and contagion, of confusion and disorder, the rhythms of the natural world have oﬀered the hope of cure or relief. It has been said that fewer than half of all ill people have an identiﬁable underlying pathology. They may feel ill, they may in fact be ill—perhaps the two amount to the same thing—but there is no recognisable organic dysfunction. Although we have to be wary of our terminology—it is easy to slip between illness, the subjective sense of being ill, and disease, the identiﬁable biochemical disorder—it would seem that much sickness lies beyond the reach of

Bill Jacklin, Walking Down Broadway (1998)

22

orthodox medicine. Where biochemical interventions fail or fall short, where a literal cure is unavailable, we sometimes have recourse to symbolic means. Theroux’s escape into wilderness, his search for puriﬁcation, has obvious symbolic content. He sees his sickness as a case of impurity. The natural world will cleanse him and return him to a state of native health. Given the lack of an organic origin for so much of our sickness, this use of symbolic resources may not be entirely fanciful. Human beings inhabit a world of matter and of meaning, of fact and symbol. If our sickness is of symbolic origin it may be amenable to symbolic cure, or at least to symbolic amelioration. Psychoanalysis is only the most obvious, the most organised example of such a cure; homoeopathy may be another. In our search for symbolic sources of healing, we have heavily mined the idea of the natural. Ideas of wellbeing and human wholeness have helped structure our idea of landscape itself. Take this habit of opposing the natural landscape to the urban. It extends at least as far back as the Greek pastoral poets of the 3rd century BC and has set up an enduring, if mutable, set of contrasts between two ways of life: one rooted in agriculture and the natural world, the other in the complex economic practices that have their origin in trade and ﬁnance and are located in the city. Although the lived experience of the country and the city may have borne no real relation to this set of oppositions—poverty, ignorance, disease, and malnutrition were as much a part of rural life as fresh air and Edenic plenty—the symbolic link between health and natural landscapes has endured. There is clearly a moral dimension to this association. The word “sick” is often used to describe depraved or immoral behaviour and sickness has often been seen as an embodied punishment for sin. Moll, the harlot in Hogarth’s A Harlot’s Progress, falls apart physically as well as morally, the one a vivid visual counterpart to the other. This association is still with us. Despite the alleged neutrality of the state towards the life choices of its citizens, the question of personal responsibility for health is never far away. As Lord Darzi has said: “The NHS in the 21st century increasingly faces a disease burden determined by the choices people make: to smoke, drink excessively, eat poorly and not take enough exercise. Today, countless years of healthy life are lost as the result of these known behavioural or lifestyle factors.” The association between the city and immorality and therefore with illness is also an ancient one. The simple economics of pleasure concentrates the possibilities for vice in centres of population: Sodom and Gomorrah were both cities of the plain. The city also oﬀers anonymity, and with www.thelancet.com Vol 373 January 3, 2009

it the possibility of escaping moral surveillance. Take one of the greatest of rural novels, George Eliot’s Middlemarch, one of the few novels, according to Virginia Woolf, written for grown-up people. Its moral order depends upon a network of cause and eﬀect in which good and bad actions have inevitable consequences. This moral causality only works, however, in face-to-face communities, in the sort of human associations that the German sociologist Ferdinand Tönnies called Gemeinschafts, communities of shared values and relatively simple structures in which people know each other over time. It is diﬃcult to imagine how such a moral order could survive transplantation to a city. Cities are just too impersonal, too anonymous, too full of possibilities for evading the slow-moving moral consequences of our actions. The association of the city with immorality also owes something to the fact that the city, especially the capital city, has traditionally been the location of the Court. Intrigue, corruption, sophistication, and vice all dance attendance on the Crown. In the country, manners may be rude, and wits may not sparkle, but plain living and open air are more likely to keep us on the path of righteousness, both physically and morally. Jean-Jacques Rousseau, writing in the mid-18th century, bequeathed to modernity the most passionate and systematic account of the moral and physical superiority of the state of nature. “I dared to strip man’s nature naked”, he writes in Confessions, “to follow the evolution of those times and things which have disﬁgured him; I compared man-made man with natural man, and I discovered that his supposed improvement had generated all his miseries.” It would be diﬃcult to exaggerate the extent and importance of Rousseau’s inﬂuence. Sigmund Freud is indebted to him. Rousseau’s fearlessly honest selfscrutiny in his Confessions lies close to the root of our developmental understanding of human personality, and Freud’s Civilization and its Discontents at times reads like the work of a disciple. Karl Marx can also feel like an heir. “Man is born free”, wrote Rousseau, in words that could easily have come from Marx, “but everywhere lives in chains”. Rousseau was a central ﬁgure in the Romantic onslaught on the Enlightenment—on the great 18thcentury belief in reason and progress, in science and the perfectibility of mankind. For Rousseau, so-called progress was itself the problem, leading us further and further from our original nature: our civilisation is our sickness, and recovery lies in a return to our essential, our natural goodness. The Romantic mistrust of science and progress, of the relentless march of western technological civilisation, can be felt in much contemporary unease about medicine. The sense that the sophisticated, invasive, and alienating technologies of medicine are somehow generative, or www.thelancet.com Vol 373 January 3, 2009

Getty Images

Perspectives

Vincent van Gogh, Wheatﬁeld with Lark (1887)

at lease symptomatic of the very maladies they seek to cure, lies behind the appeal of much alternative and complementary therapy with its stress on “natural” remedies and non-invasive techniques. Against the forces of technology, are amassed the symbolic resources of nature, which seem to lie so close to the sources of health and, in the end, to the sources of life itself. Modern western medicine is often criticised for mission creep. The net of sickness is cast ever wider, bringing more of our life into its purview. States that were once considered ordinary parts of human character or experience—shyness, heartbreak, inattentiveness— are gradually becoming pathologised, oﬀered up for biochemical tweaking, seen less as facts to be lived with than disorders to be purged or overcome. And yet a reaction to the medicalisation of human experience is setting in, a reaction that is mining a deep stratum of symbolic resources linked to the idea of the natural. Against ideas of engineered human perfectibility is set the sheer, fertile, haphazard, and exuberant contingency of nature. Thinking about this reaction, it is beginning to look as if we are fearful that we might have paid a high price for our triumph over the natural landscape. We may have subdued the external world but the idea of nature has come back to haunt us, holding out the possibility of a wellbeing we fear we may have lost forever.

Julian Sheather Deputy Head Ethics, British Medical Association, London WC1H 9JP, UK [email protected]

23

Obituary

PA Photos

The printed journal includes an image merely for illustration

Adrian Kantrowitz Pioneering heart surgeon who invented devices to assist failing hearts. He was born on Oct 4, 1918, in New York City, NY, USA, and died from complications of heart failure on Nov 14, 2008, in Ann Arbor, MI, USA, aged 90 years. Adrian Kantrowitz may be most famous as the man who was poised to do the world’s ﬁrst heart transplantation, and wound up doing the second. But the devices he developed for treating heart failure have helped millions of people worldwide and will be his lasting legacy. “He’s probably saved more lives than just about anybody else in the heart business”, says Larry Stephenson, professor in chief of cardiothoracic surgery at Wayne State University, Detroit, MI, USA. Kantrowitz ﬁrst tackled the problem of helping the heart pump as a surgical resident, and continued to wrestle with how to assist failing hearts mechanically up to the end of his life. As a resident at Monteﬁore Medical Center in the Bronx, NY, USA, fresh out of a short stint as an Army medic in World War II, Kantrowitz was given 6 months and US$300 to research anything he chose. He decided to investigate whether it would be possible to bypass the mitral valve by putting a cannula into the veins leading to the heart, diverting the blood to a pump, and then returning it to the aorta. The 60th animal to undergo the procedure survived. Soon after publishing his results, Kantrowitz decided that he wanted to be a heart surgeon. Having started out in general surgery, he soon learned about a technique for repairing mitral valve stenosis, in which the surgeon placed a ﬁnger, “blind”, into the valve and pushed it open. “That 24

then really started heart surgery”, Kantrowitz said in 1999. “My ﬁrst year in practice, I must have done 100 of them.” Open heart surgery soon followed. In 1955, Kantrowitz was named chief of surgery at Maimonides Hospital in Brooklyn, NY, USA, where he operated and did research with the help of a $3 million grant from the National Institutes of Health (NIH). In 1961, as an oﬀshoot of his research on heart pacing, he used electrodes to help a paralysed person stand up and sit down under computer control. In 1966, he implanted a left-ventricular device in a patient with heart failure for the ﬁrst time. The next year, he invented the intra-aortic balloon pump. The device, which came into widespread use in the early 1980s, has now been used in more than 3 million people. In 1967, after an aborted attempt the previous year, Kantrowitz and his colleagues were given approval to do a heart transplantation in a human baby and had found a recipient and an anencephalic newborn donor. Just before the surgery was to take place, word came that Christiaan Barnard had done the world’s ﬁrst heart transplantation. Kantrowitz and his colleagues proceeded with their own operation shortly afterwards. The infant died 6·5 h after the operation. The next month, Kantrowitz did another heart transplantation—the ﬁfth in the world—this time in an adult, who also died within hours. As worldwide controversy raged over the ethics of heart transplantation, which after 1 year and 101 procedures still seemed of questionable beneﬁt, Kantrowitz moved to Sinai Hospital in Detroit, where he also taught at Wayne State University, taking his team of engineers, surgeons, and nurses with him—along with his NIH funding. Kantrowitz reﬁned the devices he invented for the rest of his life. Just before his death, he won Food and Drug Administration approval for a second pilot trial of a version of his balloon pump for long-term and ambulatory use. “That’s what he had in mind from the beginning”, says his wife of 60 years, Jean, who worked with him to help administer his NIH grants. The way he saw it, she adds, “the heart is a pump, that’s all it is, you don’t have to take it out. You just leave it there and a little bit of help will do a lot for the patient.” “Adrian Kantrowitz was a very, very nice man, and a real inspiration for guys like me”, says Stephen Lahey, chief of cardiac and thoracic surgery at Maimonides Medical Center. Lahey likens Kantrowitz and other early heart surgeons to the ﬁrst astronauts, “They were way out on the fringe and pushing the technological envelope and it was a very diﬃcult job, because most of the people they operated on died. The psychological toll on these guys was incredible.” Kantrowitz is survived by his wife; his three children, Niki, a cardiologist, Lisa, a radiologist, and Allen, a neurologist; and nine grandchildren.

Anne Harding [email protected]

www.thelancet.com Vol 373 January 3, 2009

Correspondence

In 2003, the potential risks to children arising from information being held in diﬀerent locations led Lord Laming to recommend that “within a given location, health professionals should work from a single set of records for each child”.1 In 2004, British Medical Association guidance on medical note-keeping2 recommended that there should be clear means of identifying records of children about whom there were child protection concerns, usually with the consent of the parent. One of the factors that led to the UK General Medical Council’s decision to erase paediatrician David Southall’s name from the medical register was the keeping of children’s records separately from the main hospital notes. Recognising that, within our own organisations, information on children is still held in diﬀerent places, we sent questionnaires to 135 consultant paediatricians in the UK. We received responses from 32. The questions included whether for any particular child: (a) medical records including child protection records were kept together; (b) reasons why this would not be so; (c) whether notes were tagged to indicate that there were separate records; and (d) whether permission to keep records separately was obtained from parents. Only ﬁve organisations had completely combined records including child protection records, and merging had only occurred since the Laming report in 2003.1 Tertiary records were often held separately—eg, nephrology, oncology, paediatric intensive-care unit, and sleep laboratory records. Reasons given for keeping separate records were many but included maintaining conﬁdentiality, having diﬀerent locations within the same Trust, convenience of access, www.thelancet.com Vol 373 January 3, 2009

historical processes, and shortage of administration staﬀ. In 11 organisations, child protection notes were in the main records; ﬁve always kept the notes separately. In other organisations, records could be in the hospital or community notes, depending on where the child had been seen, and some kept case conference minutes and legal documents separately. If children were seen or examined at the request of police or social care, most units created hospital or community records for them, but six did not. Where tagging might be applicable, there was again variation, including no tagging, inconsistent use of written methods in the notes, and electronic tagging. No permission was sought anywhere. Several responses indicated that advice on consistent practice across the UK would be welcome. Some organisations had grappled with single records and were struggling. In conclusion, achieving a single clinical record for a particular child in the UK is not straightforward and the all-encompassing electronic record is still some years away. We all undertake child protection work to a greater or lesser extent. PF and MC prepared reports on the topic of keeping separate ﬁles for the General Medical Council hearing concerning David Southall, but were not called to give evidence. MC is a member of Professionals Against Child Abuse.

*Margaret Crawford, Peter Fleming, Nadeem Moghal [email protected] Pilgrim Hospital, United Lincolnshire Hospitals Trust, Boston PE21 9QS, UK (MC); Department of Community Based Medicine, University of Bristol, Bristol, UK (PF); and Department of Paediatric Nephrology, Royal Victoria Inﬁrmary, Newcastle Hospitals NHS Foundation Trust, Newcastle-upon-Tyne, UK (NM) 1

2

The Victoria Climbie Inquiry. Report of an inquiry by Lord Laming. London: HMSO, 2003. http://www.victoria-climbie-inquiry.org.uk/ ﬁnreport/ﬁnreport.htm (accessed Dec 5, 2008). British Medical Association. Doctors’ responsibilities in child protection cases: guidance from the ethics department. London: BMA, 2004. http://www.bma.org.uk/ap.nsf/ content/childprotection#Summary (accessed Dec 5, 2008).

Association between cerebral palsy and erythromycin In the ORACLE II follow-up study (Oct 11, p 1310),1 Sara Kenyon and colleagues report increased functional impairment and cerebral palsy among children who were exposed to erythromycin while their mothers were in preterm labour. We would like to draw attention to several key features of this trial that could aﬀect the conclusions drawn. First, only a minority of women were actually in preterm labour and no objective criterion was applied to stratify women on the basis of infection. It is clearly possible that a disproportionate number of women with infection were allocated to the treatment groups. Second, neonates in the treatment groups might have been more immature than those in the control group and had resulting poorer outcomes irrespective of erythromycin exposure. Kenyon and colleagues should provide data on the distribution of immaturity in the diﬀerent groups. Additionally, we are disappointed that though a Bonferroni correction is mentioned in the text, it has not been applied to the dataset. Finally, Kenyon and colleagues strengthen their argument by providing parent-reported data on functional impairment. Clearly these data are unreliable because close to 40% of children age 7 years were reported as being neurologically impaired yet with normal educational achievements. Multi-Attribute Health Status scales have been known to overestimate quality-of-life indices,2 and we suspect that this has happened here. The role of infection and inﬂammation in preterm labour and its poor outcome3,4 is clearly established. We fear that this study will add to the climate generated by the PREMET study5 and could therefore lead to reluctance in prescribing and using antibiotics.

The printed journal includes an image merely for illustration PA Photos

The fragmented clinical record: a risk to at-risk children

Lord Laming

Submissions should be made via our electronic submission system at http://ees.elsevier.com/ thelancet/

25

Correspondence

In 2003, the potential risks to children arising from information being held in diﬀerent locations led Lord Laming to recommend that “within a given location, health professionals should work from a single set of records for each child”.1 In 2004, British Medical Association guidance on medical note-keeping2 recommended that there should be clear means of identifying records of children about whom there were child protection concerns, usually with the consent of the parent. One of the factors that led to the UK General Medical Council’s decision to erase paediatrician David Southall’s name from the medical register was the keeping of children’s records separately from the main hospital notes. Recognising that, within our own organisations, information on children is still held in diﬀerent places, we sent questionnaires to 135 consultant paediatricians in the UK. We received responses from 32. The questions included whether for any particular child: (a) medical records including child protection records were kept together; (b) reasons why this would not be so; (c) whether notes were tagged to indicate that there were separate records; and (d) whether permission to keep records separately was obtained from parents. Only ﬁve organisations had completely combined records including child protection records, and merging had only occurred since the Laming report in 2003.1 Tertiary records were often held separately—eg, nephrology, oncology, paediatric intensive-care unit, and sleep laboratory records. Reasons given for keeping separate records were many but included maintaining conﬁdentiality, having diﬀerent locations within the same Trust, convenience of access, www.thelancet.com Vol 373 January 3, 2009

historical processes, and shortage of administration staﬀ. In 11 organisations, child protection notes were in the main records; ﬁve always kept the notes separately. In other organisations, records could be in the hospital or community notes, depending on where the child had been seen, and some kept case conference minutes and legal documents separately. If children were seen or examined at the request of police or social care, most units created hospital or community records for them, but six did not. Where tagging might be applicable, there was again variation, including no tagging, inconsistent use of written methods in the notes, and electronic tagging. No permission was sought anywhere. Several responses indicated that advice on consistent practice across the UK would be welcome. Some organisations had grappled with single records and were struggling. In conclusion, achieving a single clinical record for a particular child in the UK is not straightforward and the all-encompassing electronic record is still some years away. We all undertake child protection work to a greater or lesser extent. PF and MC prepared reports on the topic of keeping separate ﬁles for the General Medical Council hearing concerning David Southall, but were not called to give evidence. MC is a member of Professionals Against Child Abuse.

*Margaret Crawford, Peter Fleming, Nadeem Moghal [email protected] Pilgrim Hospital, United Lincolnshire Hospitals Trust, Boston PE21 9QS, UK (MC); Department of Community Based Medicine, University of Bristol, Bristol, UK (PF); and Department of Paediatric Nephrology, Royal Victoria Inﬁrmary, Newcastle Hospitals NHS Foundation Trust, Newcastle-upon-Tyne, UK (NM) 1

2

The Victoria Climbie Inquiry. Report of an inquiry by Lord Laming. London: HMSO, 2003. http://www.victoria-climbie-inquiry.org.uk/ ﬁnreport/ﬁnreport.htm (accessed Dec 5, 2008). British Medical Association. Doctors’ responsibilities in child protection cases: guidance from the ethics department. London: BMA, 2004. http://www.bma.org.uk/ap.nsf/ content/childprotection#Summary (accessed Dec 5, 2008).

Association between cerebral palsy and erythromycin In the ORACLE II follow-up study (Oct 11, p 1310),1 Sara Kenyon and colleagues report increased functional impairment and cerebral palsy among children who were exposed to erythromycin while their mothers were in preterm labour. We would like to draw attention to several key features of this trial that could aﬀect the conclusions drawn. First, only a minority of women were actually in preterm labour and no objective criterion was applied to stratify women on the basis of infection. It is clearly possible that a disproportionate number of women with infection were allocated to the treatment groups. Second, neonates in the treatment groups might have been more immature than those in the control group and had resulting poorer outcomes irrespective of erythromycin exposure. Kenyon and colleagues should provide data on the distribution of immaturity in the diﬀerent groups. Additionally, we are disappointed that though a Bonferroni correction is mentioned in the text, it has not been applied to the dataset. Finally, Kenyon and colleagues strengthen their argument by providing parent-reported data on functional impairment. Clearly these data are unreliable because close to 40% of children age 7 years were reported as being neurologically impaired yet with normal educational achievements. Multi-Attribute Health Status scales have been known to overestimate quality-of-life indices,2 and we suspect that this has happened here. The role of infection and inﬂammation in preterm labour and its poor outcome3,4 is clearly established. We fear that this study will add to the climate generated by the PREMET study5 and could therefore lead to reluctance in prescribing and using antibiotics.

The printed journal includes an image merely for illustration PA Photos

The fragmented clinical record: a risk to at-risk children

Lord Laming

Submissions should be made via our electronic submission system at http://ees.elsevier.com/ thelancet/

25

Correspondence

We declare that we have no conﬂict of interest.

*Mallika Azizia, Momin Ahmed [email protected] University Hospital Basildon, East of England Deanery, London CB21 5EE, UK (MaA); and Institute of Cell and Molecular Sciences, London, UK (MoA) 1

2

3

4

5

Kenyon S, Pike K, Jones D, et al. Childhood outcomes after prescription of antibiotics to pregnant women with spontaneous preterm labour: 7-year follow-up of the ORACLE II trial. Lancet 2008; 372: 1319–27. Saigal S, Rosenbaum PL, Feeny D, et al. Parental perspectives of the health status and health-related quality of life of teen-aged children who were extremely low birth weight and term controls. Pediatrics 2000; 105 (3 pt 1): 569–74. Azizia M, Lloyd J, Allen M, Klein N, Peebles D. Low cord blood monocytes MHC class II expression is associated with sepsis in term and preterm neonates. Arch Dis Child Fetal Neonatal Edn 2008; 93 (suppl 1): F6–F8. Lloyd J, Allen M, Azizia M, Klein N, Peebles D. Monocyte major histocompatibility complex class II expression in term and preterm labor. Obstet Gynecol 2007; 110: 1335–42. Shennan A, Crawshaw S, Briley A, et al. A randomised controlled trial of metronidazole for the prevention of preterm birth in women positive for cervicovaginal fetal ﬁbronectin: the PREMET Study. BJOG 2006; 113: 65–74.

Authors’ reply

Getty Images

Mallika Azizia and Momin Ahmed are correct that most of the women in the ORACLE II trial (64%) gave birth after 37 weeks, but this reﬂects the clinical situation in which the diagnosis of preterm labour remains imperfect. At the point at which these women were recruited they showed no obvious signs of infection and markers of infection collected after birth do not seem to be unequally distributed across the treatment groups, as shown in table 1 of the paper. That table also shows no diﬀerences in gestational age at birth for each treatment. Assessment of causality demands more than mechanical application of Bonferroni corrections for multiplicity. Chance is one possible explanation for an observed association such as that between cerebral palsy and either erythromycin or co-amoxiclav. However, we explored the observed excess as fully as possible, and sought to follow Hill’s approach1 to assessing causality. We brieﬂy indicated other characteristics of the association that might suggest 26

it is not just a chance ﬁnding. We cannot be certain about this, but, as we wrote originally, it would be unwise to dismiss it as chance. We do not present data as to the children’s neurological status, but parental reports of the children’s functional impairment, as derived from the Mark III classiﬁcation system.2 This was intended as further exploration not to strengthen the argument.

high incidence of adverse pregnancy outcome, suggesting that whatever damage is done by infection occurs early and persists. It follows that, if antibiotics are to be eﬀective in preventing preterm birth of infectious cause (which in itself is associated with cerebral palsy), these must be used early in pregnancy before inﬂammatory tissue damage has occurred.

I declare that I have no conﬂict of interest.

Ronald F Lamont

*Sara Kenyon, on behalf of the authors

[email protected]

[email protected]

Imperial College, London and Northwick Park Institute of Medical Research, Harrow HA1 3UJ, UK

Reproductive Sciences Section, Cancer Studies and Molecular Medicine, University of Leicester, 22–28 Princess Road West, Leicester LE1 6TP, UK 1

2

Hill AB. The environment and disease: association or causation? Proc R Soc Med 1965; 58: 295–300. Saigal S, Rosenbaum P, Stoskupf B, et al. Comprehensive assessment of the health status of extremely low birthweight children at eight years of age: comparisons with a reference group. J Pediatr 1994; 125: 411–17.

Using proxy information, Sara Kenyon and colleagues1 show that erythromycin and co-amoxiclav given to women in spontaneous preterm labour might result in more harm than good. With no objective evidence of abnormal genital-tract colonisation, these antibiotics might be causing harm by decimating normal ﬂora rather than correcting abnormal ﬂora. By contrast, even in a low-risk population, the appropriate antibiotics (clindamycin is active against most organisms that cause bacterial vaginosis) given to appropriate women (those at risk of preterm birth because of a proven imbalance of genital-tract ﬂora) early in pregnancy has shown a signiﬁcant 40–60% reduction in the incidence of preterm birth.2–5 The earlier in pregnancy at which spontaneous preterm birth occurs, the more likely this is to be due to an abnormal trigger such as infection. The earlier in pregnancy at which abnormal genital colonisation is detected, the greater is the risk of late miscarriage and early preterm birth. Even if abnormal genital-tract colonisation in early pregnancy resolves, there is still a

I declare that I have no conﬂict of interest.

1

2

3

4

5

Kenyon S, Pike K, Jones DR, et al. Childhood outcomes after prescription of antibiotics to pregnant women with spontaneous preterm labour: 7 year follow up of the ORACLE II trial. Lancet 2008; 372: 1319–27. Lamont RF, Duncan S, Mandal D, Bassett P. Intravaginal clindamycin to reduce preterm birth in women with abnormal genital tract ﬂora. Obstet Gynecol 2003; 101: 516–22. Ugwumadu A, Manyonda I, Reid F, Hay P. Eﬀect of early oral clindamycin on late miscarriage and preterm delivery in asymptomatic women with abnormal vaginal ﬂora and bacterial vaginosis: a randomised clinical trial. Lancet 2003; 361: 983–88. Kiss H, Petricevic L, Husslein P. Prospective randomised controlled trial of an infection screening programme to reduce the rate of preterm delivery. BMJ 2004; 329: 371–75. Larsson P-G, Carlsson B, Jakobsson T, Forsum U. Late miscarriage and preterm birth after treatment with clindamycin: a randomised consent design study according to Zelen. BJOG 2006; 113: 629–37.

Treatment of extensively drug-resistant tuberculosis Salmaan Keshavjee and colleagues (Oct 18, p 1403)1 achieved a 48% cure or treatment completion rate for extensively drug-resistant tuberculosis (XDRTB), despite the fact that 96%, 46%, and around 60% of the patients with XDR-TB were resistant to the second-line antitubercular drugs oﬂoxacin, capreomycin, and ethionamide, respectively. It would be useful for readers if Keshavjee and colleagues could provide details of the antitubercular regimens, including drug doses, since the eﬃcacy and adverse eﬀects of www.thelancet.com Vol 373 January 3, 2009

Correspondence

We declare that we have no conﬂict of interest.

*Mallika Azizia, Momin Ahmed [email protected] University Hospital Basildon, East of England Deanery, London CB21 5EE, UK (MaA); and Institute of Cell and Molecular Sciences, London, UK (MoA) 1

2

3

4

5

Kenyon S, Pike K, Jones D, et al. Childhood outcomes after prescription of antibiotics to pregnant women with spontaneous preterm labour: 7-year follow-up of the ORACLE II trial. Lancet 2008; 372: 1319–27. Saigal S, Rosenbaum PL, Feeny D, et al. Parental perspectives of the health status and health-related quality of life of teen-aged children who were extremely low birth weight and term controls. Pediatrics 2000; 105 (3 pt 1): 569–74. Azizia M, Lloyd J, Allen M, Klein N, Peebles D. Low cord blood monocytes MHC class II expression is associated with sepsis in term and preterm neonates. Arch Dis Child Fetal Neonatal Edn 2008; 93 (suppl 1): F6–F8. Lloyd J, Allen M, Azizia M, Klein N, Peebles D. Monocyte major histocompatibility complex class II expression in term and preterm labor. Obstet Gynecol 2007; 110: 1335–42. Shennan A, Crawshaw S, Briley A, et al. A randomised controlled trial of metronidazole for the prevention of preterm birth in women positive for cervicovaginal fetal ﬁbronectin: the PREMET Study. BJOG 2006; 113: 65–74.

Authors’ reply

Getty Images

Mallika Azizia and Momin Ahmed are correct that most of the women in the ORACLE II trial (64%) gave birth after 37 weeks, but this reﬂects the clinical situation in which the diagnosis of preterm labour remains imperfect. At the point at which these women were recruited they showed no obvious signs of infection and markers of infection collected after birth do not seem to be unequally distributed across the treatment groups, as shown in table 1 of the paper. That table also shows no diﬀerences in gestational age at birth for each treatment. Assessment of causality demands more than mechanical application of Bonferroni corrections for multiplicity. Chance is one possible explanation for an observed association such as that between cerebral palsy and either erythromycin or co-amoxiclav. However, we explored the observed excess as fully as possible, and sought to follow Hill’s approach1 to assessing causality. We brieﬂy indicated other characteristics of the association that might suggest 26

it is not just a chance ﬁnding. We cannot be certain about this, but, as we wrote originally, it would be unwise to dismiss it as chance. We do not present data as to the children’s neurological status, but parental reports of the children’s functional impairment, as derived from the Mark III classiﬁcation system.2 This was intended as further exploration not to strengthen the argument.

high incidence of adverse pregnancy outcome, suggesting that whatever damage is done by infection occurs early and persists. It follows that, if antibiotics are to be eﬀective in preventing preterm birth of infectious cause (which in itself is associated with cerebral palsy), these must be used early in pregnancy before inﬂammatory tissue damage has occurred.

I declare that I have no conﬂict of interest.

Ronald F Lamont

*Sara Kenyon, on behalf of the authors

[email protected]

[email protected]

Imperial College, London and Northwick Park Institute of Medical Research, Harrow HA1 3UJ, UK

Reproductive Sciences Section, Cancer Studies and Molecular Medicine, University of Leicester, 22–28 Princess Road West, Leicester LE1 6TP, UK 1

2

Hill AB. The environment and disease: association or causation? Proc R Soc Med 1965; 58: 295–300. Saigal S, Rosenbaum P, Stoskupf B, et al. Comprehensive assessment of the health status of extremely low birthweight children at eight years of age: comparisons with a reference group. J Pediatr 1994; 125: 411–17.

Using proxy information, Sara Kenyon and colleagues1 show that erythromycin and co-amoxiclav given to women in spontaneous preterm labour might result in more harm than good. With no objective evidence of abnormal genital-tract colonisation, these antibiotics might be causing harm by decimating normal ﬂora rather than correcting abnormal ﬂora. By contrast, even in a low-risk population, the appropriate antibiotics (clindamycin is active against most organisms that cause bacterial vaginosis) given to appropriate women (those at risk of preterm birth because of a proven imbalance of genital-tract ﬂora) early in pregnancy has shown a signiﬁcant 40–60% reduction in the incidence of preterm birth.2–5 The earlier in pregnancy at which spontaneous preterm birth occurs, the more likely this is to be due to an abnormal trigger such as infection. The earlier in pregnancy at which abnormal genital colonisation is detected, the greater is the risk of late miscarriage and early preterm birth. Even if abnormal genital-tract colonisation in early pregnancy resolves, there is still a

I declare that I have no conﬂict of interest.

1

2

3

4

5

Kenyon S, Pike K, Jones DR, et al. Childhood outcomes after prescription of antibiotics to pregnant women with spontaneous preterm labour: 7 year follow up of the ORACLE II trial. Lancet 2008; 372: 1319–27. Lamont RF, Duncan S, Mandal D, Bassett P. Intravaginal clindamycin to reduce preterm birth in women with abnormal genital tract ﬂora. Obstet Gynecol 2003; 101: 516–22. Ugwumadu A, Manyonda I, Reid F, Hay P. Eﬀect of early oral clindamycin on late miscarriage and preterm delivery in asymptomatic women with abnormal vaginal ﬂora and bacterial vaginosis: a randomised clinical trial. Lancet 2003; 361: 983–88. Kiss H, Petricevic L, Husslein P. Prospective randomised controlled trial of an infection screening programme to reduce the rate of preterm delivery. BMJ 2004; 329: 371–75. Larsson P-G, Carlsson B, Jakobsson T, Forsum U. Late miscarriage and preterm birth after treatment with clindamycin: a randomised consent design study according to Zelen. BJOG 2006; 113: 629–37.

Treatment of extensively drug-resistant tuberculosis Salmaan Keshavjee and colleagues (Oct 18, p 1403)1 achieved a 48% cure or treatment completion rate for extensively drug-resistant tuberculosis (XDRTB), despite the fact that 96%, 46%, and around 60% of the patients with XDR-TB were resistant to the second-line antitubercular drugs oﬂoxacin, capreomycin, and ethionamide, respectively. It would be useful for readers if Keshavjee and colleagues could provide details of the antitubercular regimens, including drug doses, since the eﬃcacy and adverse eﬀects of www.thelancet.com Vol 373 January 3, 2009

Correspondence

We declare that we have no conﬂict of interest.

*Mallika Azizia, Momin Ahmed [email protected] University Hospital Basildon, East of England Deanery, London CB21 5EE, UK (MaA); and Institute of Cell and Molecular Sciences, London, UK (MoA) 1

2

3

4

5

Kenyon S, Pike K, Jones D, et al. Childhood outcomes after prescription of antibiotics to pregnant women with spontaneous preterm labour: 7-year follow-up of the ORACLE II trial. Lancet 2008; 372: 1319–27. Saigal S, Rosenbaum PL, Feeny D, et al. Parental perspectives of the health status and health-related quality of life of teen-aged children who were extremely low birth weight and term controls. Pediatrics 2000; 105 (3 pt 1): 569–74. Azizia M, Lloyd J, Allen M, Klein N, Peebles D. Low cord blood monocytes MHC class II expression is associated with sepsis in term and preterm neonates. Arch Dis Child Fetal Neonatal Edn 2008; 93 (suppl 1): F6–F8. Lloyd J, Allen M, Azizia M, Klein N, Peebles D. Monocyte major histocompatibility complex class II expression in term and preterm labor. Obstet Gynecol 2007; 110: 1335–42. Shennan A, Crawshaw S, Briley A, et al. A randomised controlled trial of metronidazole for the prevention of preterm birth in women positive for cervicovaginal fetal ﬁbronectin: the PREMET Study. BJOG 2006; 113: 65–74.

Authors’ reply

Getty Images

Mallika Azizia and Momin Ahmed are correct that most of the women in the ORACLE II trial (64%) gave birth after 37 weeks, but this reﬂects the clinical situation in which the diagnosis of preterm labour remains imperfect. At the point at which these women were recruited they showed no obvious signs of infection and markers of infection collected after birth do not seem to be unequally distributed across the treatment groups, as shown in table 1 of the paper. That table also shows no diﬀerences in gestational age at birth for each treatment. Assessment of causality demands more than mechanical application of Bonferroni corrections for multiplicity. Chance is one possible explanation for an observed association such as that between cerebral palsy and either erythromycin or co-amoxiclav. However, we explored the observed excess as fully as possible, and sought to follow Hill’s approach1 to assessing causality. We brieﬂy indicated other characteristics of the association that might suggest 26

it is not just a chance ﬁnding. We cannot be certain about this, but, as we wrote originally, it would be unwise to dismiss it as chance. We do not present data as to the children’s neurological status, but parental reports of the children’s functional impairment, as derived from the Mark III classiﬁcation system.2 This was intended as further exploration not to strengthen the argument.

high incidence of adverse pregnancy outcome, suggesting that whatever damage is done by infection occurs early and persists. It follows that, if antibiotics are to be eﬀective in preventing preterm birth of infectious cause (which in itself is associated with cerebral palsy), these must be used early in pregnancy before inﬂammatory tissue damage has occurred.

I declare that I have no conﬂict of interest.

Ronald F Lamont

*Sara Kenyon, on behalf of the authors

[email protected]

[email protected]

Imperial College, London and Northwick Park Institute of Medical Research, Harrow HA1 3UJ, UK

Reproductive Sciences Section, Cancer Studies and Molecular Medicine, University of Leicester, 22–28 Princess Road West, Leicester LE1 6TP, UK 1

2

Hill AB. The environment and disease: association or causation? Proc R Soc Med 1965; 58: 295–300. Saigal S, Rosenbaum P, Stoskupf B, et al. Comprehensive assessment of the health status of extremely low birthweight children at eight years of age: comparisons with a reference group. J Pediatr 1994; 125: 411–17.

Using proxy information, Sara Kenyon and colleagues1 show that erythromycin and co-amoxiclav given to women in spontaneous preterm labour might result in more harm than good. With no objective evidence of abnormal genital-tract colonisation, these antibiotics might be causing harm by decimating normal ﬂora rather than correcting abnormal ﬂora. By contrast, even in a low-risk population, the appropriate antibiotics (clindamycin is active against most organisms that cause bacterial vaginosis) given to appropriate women (those at risk of preterm birth because of a proven imbalance of genital-tract ﬂora) early in pregnancy has shown a signiﬁcant 40–60% reduction in the incidence of preterm birth.2–5 The earlier in pregnancy at which spontaneous preterm birth occurs, the more likely this is to be due to an abnormal trigger such as infection. The earlier in pregnancy at which abnormal genital colonisation is detected, the greater is the risk of late miscarriage and early preterm birth. Even if abnormal genital-tract colonisation in early pregnancy resolves, there is still a

I declare that I have no conﬂict of interest.

1

2

3

4

5

Kenyon S, Pike K, Jones DR, et al. Childhood outcomes after prescription of antibiotics to pregnant women with spontaneous preterm labour: 7 year follow up of the ORACLE II trial. Lancet 2008; 372: 1319–27. Lamont RF, Duncan S, Mandal D, Bassett P. Intravaginal clindamycin to reduce preterm birth in women with abnormal genital tract ﬂora. Obstet Gynecol 2003; 101: 516–22. Ugwumadu A, Manyonda I, Reid F, Hay P. Eﬀect of early oral clindamycin on late miscarriage and preterm delivery in asymptomatic women with abnormal vaginal ﬂora and bacterial vaginosis: a randomised clinical trial. Lancet 2003; 361: 983–88. Kiss H, Petricevic L, Husslein P. Prospective randomised controlled trial of an infection screening programme to reduce the rate of preterm delivery. BMJ 2004; 329: 371–75. Larsson P-G, Carlsson B, Jakobsson T, Forsum U. Late miscarriage and preterm birth after treatment with clindamycin: a randomised consent design study according to Zelen. BJOG 2006; 113: 629–37.

Treatment of extensively drug-resistant tuberculosis Salmaan Keshavjee and colleagues (Oct 18, p 1403)1 achieved a 48% cure or treatment completion rate for extensively drug-resistant tuberculosis (XDRTB), despite the fact that 96%, 46%, and around 60% of the patients with XDR-TB were resistant to the second-line antitubercular drugs oﬂoxacin, capreomycin, and ethionamide, respectively. It would be useful for readers if Keshavjee and colleagues could provide details of the antitubercular regimens, including drug doses, since the eﬃcacy and adverse eﬀects of www.thelancet.com Vol 373 January 3, 2009

Correspondence

We declare that we have no conﬂict of interest.

*Mallika Azizia, Momin Ahmed [email protected] University Hospital Basildon, East of England Deanery, London CB21 5EE, UK (MaA); and Institute of Cell and Molecular Sciences, London, UK (MoA) 1

2

3

4

5

Kenyon S, Pike K, Jones D, et al. Childhood outcomes after prescription of antibiotics to pregnant women with spontaneous preterm labour: 7-year follow-up of the ORACLE II trial. Lancet 2008; 372: 1319–27. Saigal S, Rosenbaum PL, Feeny D, et al. Parental perspectives of the health status and health-related quality of life of teen-aged children who were extremely low birth weight and term controls. Pediatrics 2000; 105 (3 pt 1): 569–74. Azizia M, Lloyd J, Allen M, Klein N, Peebles D. Low cord blood monocytes MHC class II expression is associated with sepsis in term and preterm neonates. Arch Dis Child Fetal Neonatal Edn 2008; 93 (suppl 1): F6–F8. Lloyd J, Allen M, Azizia M, Klein N, Peebles D. Monocyte major histocompatibility complex class II expression in term and preterm labor. Obstet Gynecol 2007; 110: 1335–42. Shennan A, Crawshaw S, Briley A, et al. A randomised controlled trial of metronidazole for the prevention of preterm birth in women positive for cervicovaginal fetal ﬁbronectin: the PREMET Study. BJOG 2006; 113: 65–74.

Authors’ reply

Getty Images

Mallika Azizia and Momin Ahmed are correct that most of the women in the ORACLE II trial (64%) gave birth after 37 weeks, but this reﬂects the clinical situation in which the diagnosis of preterm labour remains imperfect. At the point at which these women were recruited they showed no obvious signs of infection and markers of infection collected after birth do not seem to be unequally distributed across the treatment groups, as shown in table 1 of the paper. That table also shows no diﬀerences in gestational age at birth for each treatment. Assessment of causality demands more than mechanical application of Bonferroni corrections for multiplicity. Chance is one possible explanation for an observed association such as that between cerebral palsy and either erythromycin or co-amoxiclav. However, we explored the observed excess as fully as possible, and sought to follow Hill’s approach1 to assessing causality. We brieﬂy indicated other characteristics of the association that might suggest 26

it is not just a chance ﬁnding. We cannot be certain about this, but, as we wrote originally, it would be unwise to dismiss it as chance. We do not present data as to the children’s neurological status, but parental reports of the children’s functional impairment, as derived from the Mark III classiﬁcation system.2 This was intended as further exploration not to strengthen the argument.

high incidence of adverse pregnancy outcome, suggesting that whatever damage is done by infection occurs early and persists. It follows that, if antibiotics are to be eﬀective in preventing preterm birth of infectious cause (which in itself is associated with cerebral palsy), these must be used early in pregnancy before inﬂammatory tissue damage has occurred.

I declare that I have no conﬂict of interest.

Ronald F Lamont

*Sara Kenyon, on behalf of the authors

[email protected]

[email protected]

Imperial College, London and Northwick Park Institute of Medical Research, Harrow HA1 3UJ, UK

Reproductive Sciences Section, Cancer Studies and Molecular Medicine, University of Leicester, 22–28 Princess Road West, Leicester LE1 6TP, UK 1

2

Hill AB. The environment and disease: association or causation? Proc R Soc Med 1965; 58: 295–300. Saigal S, Rosenbaum P, Stoskupf B, et al. Comprehensive assessment of the health status of extremely low birthweight children at eight years of age: comparisons with a reference group. J Pediatr 1994; 125: 411–17.

Using proxy information, Sara Kenyon and colleagues1 show that erythromycin and co-amoxiclav given to women in spontaneous preterm labour might result in more harm than good. With no objective evidence of abnormal genital-tract colonisation, these antibiotics might be causing harm by decimating normal ﬂora rather than correcting abnormal ﬂora. By contrast, even in a low-risk population, the appropriate antibiotics (clindamycin is active against most organisms that cause bacterial vaginosis) given to appropriate women (those at risk of preterm birth because of a proven imbalance of genital-tract ﬂora) early in pregnancy has shown a signiﬁcant 40–60% reduction in the incidence of preterm birth.2–5 The earlier in pregnancy at which spontaneous preterm birth occurs, the more likely this is to be due to an abnormal trigger such as infection. The earlier in pregnancy at which abnormal genital colonisation is detected, the greater is the risk of late miscarriage and early preterm birth. Even if abnormal genital-tract colonisation in early pregnancy resolves, there is still a

I declare that I have no conﬂict of interest.

1

2

3

4

5

Kenyon S, Pike K, Jones DR, et al. Childhood outcomes after prescription of antibiotics to pregnant women with spontaneous preterm labour: 7 year follow up of the ORACLE II trial. Lancet 2008; 372: 1319–27. Lamont RF, Duncan S, Mandal D, Bassett P. Intravaginal clindamycin to reduce preterm birth in women with abnormal genital tract ﬂora. Obstet Gynecol 2003; 101: 516–22. Ugwumadu A, Manyonda I, Reid F, Hay P. Eﬀect of early oral clindamycin on late miscarriage and preterm delivery in asymptomatic women with abnormal vaginal ﬂora and bacterial vaginosis: a randomised clinical trial. Lancet 2003; 361: 983–88. Kiss H, Petricevic L, Husslein P. Prospective randomised controlled trial of an infection screening programme to reduce the rate of preterm delivery. BMJ 2004; 329: 371–75. Larsson P-G, Carlsson B, Jakobsson T, Forsum U. Late miscarriage and preterm birth after treatment with clindamycin: a randomised consent design study according to Zelen. BJOG 2006; 113: 629–37.

Treatment of extensively drug-resistant tuberculosis Salmaan Keshavjee and colleagues (Oct 18, p 1403)1 achieved a 48% cure or treatment completion rate for extensively drug-resistant tuberculosis (XDRTB), despite the fact that 96%, 46%, and around 60% of the patients with XDR-TB were resistant to the second-line antitubercular drugs oﬂoxacin, capreomycin, and ethionamide, respectively. It would be useful for readers if Keshavjee and colleagues could provide details of the antitubercular regimens, including drug doses, since the eﬃcacy and adverse eﬀects of www.thelancet.com Vol 373 January 3, 2009

Correspondence

the drugs go hand in hand with the administered doses. Moreover, levels of ﬂuoroquinolones above mutant prevention concentration also prevent selection of resistant mutants.2 Nevertheless, the facts enforce the need for long-term follow-up of such patients for chances of relapse. Extrapolating the results of the study and considering the data on the presence of multiple tubercular strains in sputum isolates from an individual patient,3 further trials are required to reassess the degree of reliance on drug susceptibility tests, especially the in context of XDR-TB. Linezolid, a new drug in the oxazolidinone class, and moxiﬂoxacin are promising additions to the antitubercular drug armamentarium.4,5 The outcome of Keshavjee and colleagues’ study might have been very diﬀerent if these new agents had been incorporated in the treatment algorithm.

TB patients with treatment failure had a signiﬁcantly shorter duration of treatment (median 10·9 months) than nonXDR-TB patients (median 18·2 months, p=0·02), a ﬁnding that could not be explained by lower adherence to treatment, since there was no diﬀerence between the groups in that regard.1 Keshajvee and colleagues, taking into account previous knowledge on this hard-to-treat infection,2,3 state that longer treatments are necessary to cure XDR-TB.1 Our main concern is that duration of treatment was not included in the multivariate model. Thus, we do not know whether XDR-TB was really independently and adversely associated with treatment success, as shown in the ﬁnal model (adjusted odds ratio 0·41, 95% CI 0·19–0·91), or whether it might be confounded by the fact that those with treatment failure were treated for a signiﬁcantly shorter period.

We declare that we have no conﬂict of interest.

We declare that we have no conﬂict of interest.

*Deepak Aggarwal, Prasanta R Mohapatra, Ashok K Janmeja

*Jussara Munareto Silva, Sandra C Fuchs, Nêmora T Barcellos, Alexandre P Zavascki

[email protected] Government Medical College and Hospital, Chandigarh 160030, India 1

2

3

4

5

Keshavjee S, Gelmanova IY, Farmer PE, et al. Treatment of extensively drug-resistant tuberculosis in Tomsk, Russia: a retrospective cohort study. Lancet 2008; 372: 1403–09. Almeida D, Nuermberger E, Tyagi S, Bishai WR, Grosset J. In vivo validation of the mutant selection window hypothesis with moxiﬂoxacin in a murine model of tuberculosis. Antimicrob Agents Chemother 2007; 51: 4261–66. Braden CR, Morlock GP, Woodley CL, et al. Simultaneous infection with multiple strains of Mycobacterium tuberculosis. Clin Infect Dis 2001; 33: e42–47. Condos R, Hadgiangelis N, Leibert E, Jacquette G, Harkin T, Rom WN. Case series report of a linezolid-containing regimen for extensively drug-resistant tuberculosis. Chest 2008; 134: 187–92. Mitnick CD, Shin SS, Seung KJ, et al. Comprehensive treatment of extensively drug-resistant tuberculosis. N Engl J Med 2008; 359: 563–74.

Salmaan Keshavjee and colleagues1 compared treatment outcomes of patients with extensively drugresistant tuberculosis (XDR-TB) and non-XDR-TB and investigated factors associated with treatment failure. XDRwww.thelancet.com Vol 373 January 3, 2009

[email protected] Medical Sciences Post-graduate Program, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil (JMS); Department of Social Medicine, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil (SCF); Hospital Sanatório Partenon, Health State Secretariat, Porto Alegre, Brazil (NTB); and Infectious Diseases Unit, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil (APZ) 1

2

3

Keshavjee S, Gelmanova IY, Farmer PE, et al. Treatment of extensively drug resistant tuberculosis in Tomsk, Russia: a retrospective cohort study. Lancet 2008; 372: 1403–09. Madariaga MG, Lalloo UG, Swindells S. Extensively drug-resistant tuberculosis. Am J Med 2008; 121: 835–44. Mitnick CD, Shin SS, Seung KJ, et al. Comprehensive treatment of extensively drug-resistant tuberculosis. N Engl J Med 2008; 359: 563–74.

Authors’ reply We agree with Deepak Aggarwal and colleagues that it would be useful to report details of antituberculosis regimens, including doses and adverse events. Although this was beyond the scope of our paper, Mitnick and colleagues1 reported details of speciﬁc regimens (although not adverse

events) in the online appendix of their paper. In Tomsk, all patients received the highest doses of second-line antituberculosis medications that they could tolerate in regimens formulated by use of the algorithm described by Mukherjee and colleagues.2 We agree that further analysis of the associations between doses, eﬃcacy, and adverse events is very important. The small number of patients with extensively drug-resistant tuberculosis (XDR-TB) in our report constrained our ability to examine these associations. What we did see is that the overall incidence of adverse events was not higher among XDR-TB patients, and we do not have strong clinical reasons to suspect that it would be. As for the ideal drugs to use for XDR-TB patients, one study in human beings3 suggests that later-generation ﬂuoroquinolones have greater antituberculosis activity than earlier ones; this improvement might contribute to better outcomes. Tomsk patients did not receive third-line antituberculosis drugs such as clofazimine or linezolid. Questions about the best drug combinations and doses for XDR-TB highlight the urgent need for clinical trials to optimise current treatment approaches.4 Indeed, scientists and activists have called attention to this need.5 Jussara Silva and colleagues raise the crucial question of whether treatment duration itself should be regarded as an exposure variable in the multivariable model used in our analysis. Although we agree that the shorter treatment duration in XDR-TB patients was a notable ﬁnding, we felt that it was more likely to be a consequence of poor clinical response and not a contributing factor to poor outcome. Tomsk physicians weighed up publichealth considerations and the likelihood of treatment response, and only stopped treatment when they felt that the patient had clinically declared himself or herself as a treatment failure. Treatment was discontinued well after the median culture conversion time to 27

Correspondence

the drugs go hand in hand with the administered doses. Moreover, levels of ﬂuoroquinolones above mutant prevention concentration also prevent selection of resistant mutants.2 Nevertheless, the facts enforce the need for long-term follow-up of such patients for chances of relapse. Extrapolating the results of the study and considering the data on the presence of multiple tubercular strains in sputum isolates from an individual patient,3 further trials are required to reassess the degree of reliance on drug susceptibility tests, especially the in context of XDR-TB. Linezolid, a new drug in the oxazolidinone class, and moxiﬂoxacin are promising additions to the antitubercular drug armamentarium.4,5 The outcome of Keshavjee and colleagues’ study might have been very diﬀerent if these new agents had been incorporated in the treatment algorithm.

TB patients with treatment failure had a signiﬁcantly shorter duration of treatment (median 10·9 months) than nonXDR-TB patients (median 18·2 months, p=0·02), a ﬁnding that could not be explained by lower adherence to treatment, since there was no diﬀerence between the groups in that regard.1 Keshajvee and colleagues, taking into account previous knowledge on this hard-to-treat infection,2,3 state that longer treatments are necessary to cure XDR-TB.1 Our main concern is that duration of treatment was not included in the multivariate model. Thus, we do not know whether XDR-TB was really independently and adversely associated with treatment success, as shown in the ﬁnal model (adjusted odds ratio 0·41, 95% CI 0·19–0·91), or whether it might be confounded by the fact that those with treatment failure were treated for a signiﬁcantly shorter period.

We declare that we have no conﬂict of interest.

We declare that we have no conﬂict of interest.

*Deepak Aggarwal, Prasanta R Mohapatra, Ashok K Janmeja

*Jussara Munareto Silva, Sandra C Fuchs, Nêmora T Barcellos, Alexandre P Zavascki

[email protected] Government Medical College and Hospital, Chandigarh 160030, India 1

2

3

4

5

Keshavjee S, Gelmanova IY, Farmer PE, et al. Treatment of extensively drug-resistant tuberculosis in Tomsk, Russia: a retrospective cohort study. Lancet 2008; 372: 1403–09. Almeida D, Nuermberger E, Tyagi S, Bishai WR, Grosset J. In vivo validation of the mutant selection window hypothesis with moxiﬂoxacin in a murine model of tuberculosis. Antimicrob Agents Chemother 2007; 51: 4261–66. Braden CR, Morlock GP, Woodley CL, et al. Simultaneous infection with multiple strains of Mycobacterium tuberculosis. Clin Infect Dis 2001; 33: e42–47. Condos R, Hadgiangelis N, Leibert E, Jacquette G, Harkin T, Rom WN. Case series report of a linezolid-containing regimen for extensively drug-resistant tuberculosis. Chest 2008; 134: 187–92. Mitnick CD, Shin SS, Seung KJ, et al. Comprehensive treatment of extensively drug-resistant tuberculosis. N Engl J Med 2008; 359: 563–74.

Salmaan Keshavjee and colleagues1 compared treatment outcomes of patients with extensively drugresistant tuberculosis (XDR-TB) and non-XDR-TB and investigated factors associated with treatment failure. XDRwww.thelancet.com Vol 373 January 3, 2009

[email protected] Medical Sciences Post-graduate Program, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil (JMS); Department of Social Medicine, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil (SCF); Hospital Sanatório Partenon, Health State Secretariat, Porto Alegre, Brazil (NTB); and Infectious Diseases Unit, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil (APZ) 1

2

3

Keshavjee S, Gelmanova IY, Farmer PE, et al. Treatment of extensively drug resistant tuberculosis in Tomsk, Russia: a retrospective cohort study. Lancet 2008; 372: 1403–09. Madariaga MG, Lalloo UG, Swindells S. Extensively drug-resistant tuberculosis. Am J Med 2008; 121: 835–44. Mitnick CD, Shin SS, Seung KJ, et al. Comprehensive treatment of extensively drug-resistant tuberculosis. N Engl J Med 2008; 359: 563–74.

Authors’ reply We agree with Deepak Aggarwal and colleagues that it would be useful to report details of antituberculosis regimens, including doses and adverse events. Although this was beyond the scope of our paper, Mitnick and colleagues1 reported details of speciﬁc regimens (although not adverse

events) in the online appendix of their paper. In Tomsk, all patients received the highest doses of second-line antituberculosis medications that they could tolerate in regimens formulated by use of the algorithm described by Mukherjee and colleagues.2 We agree that further analysis of the associations between doses, eﬃcacy, and adverse events is very important. The small number of patients with extensively drug-resistant tuberculosis (XDR-TB) in our report constrained our ability to examine these associations. What we did see is that the overall incidence of adverse events was not higher among XDR-TB patients, and we do not have strong clinical reasons to suspect that it would be. As for the ideal drugs to use for XDR-TB patients, one study in human beings3 suggests that later-generation ﬂuoroquinolones have greater antituberculosis activity than earlier ones; this improvement might contribute to better outcomes. Tomsk patients did not receive third-line antituberculosis drugs such as clofazimine or linezolid. Questions about the best drug combinations and doses for XDR-TB highlight the urgent need for clinical trials to optimise current treatment approaches.4 Indeed, scientists and activists have called attention to this need.5 Jussara Silva and colleagues raise the crucial question of whether treatment duration itself should be regarded as an exposure variable in the multivariable model used in our analysis. Although we agree that the shorter treatment duration in XDR-TB patients was a notable ﬁnding, we felt that it was more likely to be a consequence of poor clinical response and not a contributing factor to poor outcome. Tomsk physicians weighed up publichealth considerations and the likelihood of treatment response, and only stopped treatment when they felt that the patient had clinically declared himself or herself as a treatment failure. Treatment was discontinued well after the median culture conversion time to 27

Correspondence

the drugs go hand in hand with the administered doses. Moreover, levels of ﬂuoroquinolones above mutant prevention concentration also prevent selection of resistant mutants.2 Nevertheless, the facts enforce the need for long-term follow-up of such patients for chances of relapse. Extrapolating the results of the study and considering the data on the presence of multiple tubercular strains in sputum isolates from an individual patient,3 further trials are required to reassess the degree of reliance on drug susceptibility tests, especially the in context of XDR-TB. Linezolid, a new drug in the oxazolidinone class, and moxiﬂoxacin are promising additions to the antitubercular drug armamentarium.4,5 The outcome of Keshavjee and colleagues’ study might have been very diﬀerent if these new agents had been incorporated in the treatment algorithm.

TB patients with treatment failure had a signiﬁcantly shorter duration of treatment (median 10·9 months) than nonXDR-TB patients (median 18·2 months, p=0·02), a ﬁnding that could not be explained by lower adherence to treatment, since there was no diﬀerence between the groups in that regard.1 Keshajvee and colleagues, taking into account previous knowledge on this hard-to-treat infection,2,3 state that longer treatments are necessary to cure XDR-TB.1 Our main concern is that duration of treatment was not included in the multivariate model. Thus, we do not know whether XDR-TB was really independently and adversely associated with treatment success, as shown in the ﬁnal model (adjusted odds ratio 0·41, 95% CI 0·19–0·91), or whether it might be confounded by the fact that those with treatment failure were treated for a signiﬁcantly shorter period.

We declare that we have no conﬂict of interest.

We declare that we have no conﬂict of interest.

*Deepak Aggarwal, Prasanta R Mohapatra, Ashok K Janmeja

*Jussara Munareto Silva, Sandra C Fuchs, Nêmora T Barcellos, Alexandre P Zavascki

[email protected] Government Medical College and Hospital, Chandigarh 160030, India 1

2

3

4

5

Keshavjee S, Gelmanova IY, Farmer PE, et al. Treatment of extensively drug-resistant tuberculosis in Tomsk, Russia: a retrospective cohort study. Lancet 2008; 372: 1403–09. Almeida D, Nuermberger E, Tyagi S, Bishai WR, Grosset J. In vivo validation of the mutant selection window hypothesis with moxiﬂoxacin in a murine model of tuberculosis. Antimicrob Agents Chemother 2007; 51: 4261–66. Braden CR, Morlock GP, Woodley CL, et al. Simultaneous infection with multiple strains of Mycobacterium tuberculosis. Clin Infect Dis 2001; 33: e42–47. Condos R, Hadgiangelis N, Leibert E, Jacquette G, Harkin T, Rom WN. Case series report of a linezolid-containing regimen for extensively drug-resistant tuberculosis. Chest 2008; 134: 187–92. Mitnick CD, Shin SS, Seung KJ, et al. Comprehensive treatment of extensively drug-resistant tuberculosis. N Engl J Med 2008; 359: 563–74.

Salmaan Keshavjee and colleagues1 compared treatment outcomes of patients with extensively drugresistant tuberculosis (XDR-TB) and non-XDR-TB and investigated factors associated with treatment failure. XDRwww.thelancet.com Vol 373 January 3, 2009

[email protected] Medical Sciences Post-graduate Program, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil (JMS); Department of Social Medicine, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil (SCF); Hospital Sanatório Partenon, Health State Secretariat, Porto Alegre, Brazil (NTB); and Infectious Diseases Unit, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil (APZ) 1

2

3

Keshavjee S, Gelmanova IY, Farmer PE, et al. Treatment of extensively drug resistant tuberculosis in Tomsk, Russia: a retrospective cohort study. Lancet 2008; 372: 1403–09. Madariaga MG, Lalloo UG, Swindells S. Extensively drug-resistant tuberculosis. Am J Med 2008; 121: 835–44. Mitnick CD, Shin SS, Seung KJ, et al. Comprehensive treatment of extensively drug-resistant tuberculosis. N Engl J Med 2008; 359: 563–74.

Authors’ reply We agree with Deepak Aggarwal and colleagues that it would be useful to report details of antituberculosis regimens, including doses and adverse events. Although this was beyond the scope of our paper, Mitnick and colleagues1 reported details of speciﬁc regimens (although not adverse

events) in the online appendix of their paper. In Tomsk, all patients received the highest doses of second-line antituberculosis medications that they could tolerate in regimens formulated by use of the algorithm described by Mukherjee and colleagues.2 We agree that further analysis of the associations between doses, eﬃcacy, and adverse events is very important. The small number of patients with extensively drug-resistant tuberculosis (XDR-TB) in our report constrained our ability to examine these associations. What we did see is that the overall incidence of adverse events was not higher among XDR-TB patients, and we do not have strong clinical reasons to suspect that it would be. As for the ideal drugs to use for XDR-TB patients, one study in human beings3 suggests that later-generation ﬂuoroquinolones have greater antituberculosis activity than earlier ones; this improvement might contribute to better outcomes. Tomsk patients did not receive third-line antituberculosis drugs such as clofazimine or linezolid. Questions about the best drug combinations and doses for XDR-TB highlight the urgent need for clinical trials to optimise current treatment approaches.4 Indeed, scientists and activists have called attention to this need.5 Jussara Silva and colleagues raise the crucial question of whether treatment duration itself should be regarded as an exposure variable in the multivariable model used in our analysis. Although we agree that the shorter treatment duration in XDR-TB patients was a notable ﬁnding, we felt that it was more likely to be a consequence of poor clinical response and not a contributing factor to poor outcome. Tomsk physicians weighed up publichealth considerations and the likelihood of treatment response, and only stopped treatment when they felt that the patient had clinically declared himself or herself as a treatment failure. Treatment was discontinued well after the median culture conversion time to 27

Correspondence

avoid unnecessary toxic eﬀects and the potential for selection of pan-resistant strains. We conclude that use of aggressive second-line antituberculosis treatment regimens from the outset— with proper management of adverse events—can lead to positive outcomes in a signiﬁcant proportion of patients. Treatment duration is another timely question that requires rigorous examination through clinical trials. SK, IG, and SS received partial salary support or travel support from the Bill & Melinda Gates Foundation and from the Eli Lilly Foundation. SK and SS received salary support from the Frank Hatch Fellowships in Global Health Equity at the Brigham & Women’s Hospital. SS received additional salary support from the Infectious Disease Society of America, the Heiser Foundation, and the US National Institutes of Health. SK received travel and research funding from the John D and Catherine T MacArthur Foundation.

*Salmaan Keshavjee, Irina Gelmanova, Sid Atwood, Sonya Shin [email protected] Department of Global Health and Social Medicine, Harvard Medical School, 641 Huntington Avenue, Boston, MA 02115, USA (SK); Division of Global Health Equity, Brigham and Women’s Hospital, Boston, MA, USA (SK, SA, SS); and Partners In Health, Boston, MA, USA (IG) 1

2

3

4

5

Panos Pictures

The printed journal includes an image merely for illustration See Department of Error page 30

28

Mitnick CD, Shin SS, Seung KJ, et al. Comprehensive treatment of extensively drug-resistant tuberculosis. N Engl J Med 2008; 359: 563–74. Mukherjee JS, Rich ML, Socci AR, et al. Programmes and principles in treatment of multidrug-resistant tuberculosis. Lancet 2004; 363: 474–81. Johnson JL, Hadad DJ, Boom WH, et al. Early and extended early bactericidal activity of levoﬂoxacin, gatiﬂoxacin and moxiﬂoxacin in pulmonary tuberculosis. Int J Tuberc Lung Dis 2006; 10: 605–12. Mitnick CD, Castro KG, Harrington M, Sacks LV, Burman W. Randomized trials to optimize treatment of multidrug-resistant tuberculosis. PLoS Med 2007; 4: e292. The Cambridge Declaration: towards clinical trials for drug-resistant tuberculosis. Cambridge, MA, USA; June 12, 2008. http://www.ghdonline. org/drtb/resource/the-cambridge-declarationtowards-clinical-trials-/ (accessed Dec 3, 2008).

Unsexy tuberculosis XDR-TB—extensively drug-resistant tuberculosis. The acronym makes one cringe with fear. Here we have an increasingly prevalent disease that is potentially as lethal as any weapon

of terrorism. But what is it really, and what is the secret to conquering it? First described in 2006 in KwaZulu Natal, South Africa,1 XDR-TB killed 52 of the 53 HIV-infected people it aﬀected. In 2007, when Andrew Speaker was diagnosed in Atlanta, GA, USA and placed on (but did not heed) a “no-ﬂy” order on the basis of his potential for disease transmission, the world’s press reacted with almost unprecedented concern. As reports of cohorts of patients followed from reputable centres,2,3 with cure rates similar to survival rates for tuberculosis in the “pre-drug” era, the Stop TB Department of WHO published an emergency response plan appropriately calling for integration of global resistant and sensitive tuberculosis control programmes.4 Cries for new drugs, diagnostics, and vaccines soon followed. Although all of this concern and attention is appropriate, it seems to skip the more important message. Certainly XDR-TB or MDR-TB (multiple-drug-resistant tuberculosis) is frightening. Certainly we need new drugs and diagnostics and the capacity to allow us to address the speciﬁcs of drug-resistant tuberculosis. But the key aspect that has been largely forgotten in all of the hype is that XDR-TB, MDR-TB, and drug-sensitive tuberculosis are all the same disease. The only diﬀerence is that MDRTB is drug-sensitive tuberculosis modiﬁed by inappropriate treatment or drug taking, and XDR-TB is MDRTB thus modiﬁed. In other words, every person with MDR-TB or XDRTB was not treated properly, did not take their drugs properly, or were infected by somebody who was not treated properly or did not take their medicines properly. In 2000, I coauthored an editorial5 pointing out that, with the growing acceptance of the necessity for prescribing MDR-TB drug therapy (then called DOTS Plus), this far more “sexy” aspect of tuberculosis care and control would distract attention

and resources away from basic tuberculosis control and therefore interfere with solving the problem. I am afraid that that is exactly what is being seen now. We need to recognise that there are more than 9 000 000 new active drugsensitive cases of tuberculosis globally that could be feeding drug resistance. It might be a less sexy concept, but they all must be appropriately treated with current strategies (as well as new diagnostics, drugs, vaccines, and proper infection control measures) to avoid preventable MDR-TB and XDR-TB, which are always lurking. Preventing active, drug-sensitive tuberculosis, or treating it properly, should be everybody’s priority: it is the only way to prevent MDR-TB and XDR-TB. I declare that I have no conﬂict of interest.

Lee B Reichman [email protected] New Jersey Medical School, Global Tuberculosis Institute, Newark, NJ 07101, USA 1

2

3

4

5

Gandhi NR, Moll A, Sturm AW, et al. Extensively drug-resistant tuberculosis as a cause of death in patients co-infected with tuberculosis and HIV in a rural area of South Africa. Lancet 2006; 368: 1575–80. Mitnick CD, Shin SS, Seung KJ, et al. Comprehensive treatment of extensively drug-resistant tuberculosis. N Engl J Med 2008; 359: 563–74. Keshavjee S, Gelmanova IY, Farmer PE, et al. Treatment of extensively drug-resistant tuberculosis in Tomsk, Russia: a retrospective cohort study. Lancet 2008; 372: 1403–09. Stop TB Partnership. The global MDR-TB and XDR-TB response plan 2007–2008. Geneva: World Health Organization, 2007. Lambregts-Van Weezenbeek K, Reichman LB. DOTS and DOTS-Plus: what’s in a name. Int J Tuberc Lung Dis 2000; 4: 995–96.

Intermittent preventive treatment with antimalarial drugs The Article by Siân Clarke and colleagues (July 12, p 127)1 provides valuable information about the eﬃcacy of intermittent preventive treatment (IPT) of malaria on health and education in schoolchildren in Kenya. We have a few concerns, however. www.thelancet.com Vol 373 January 3, 2009

Correspondence

avoid unnecessary toxic eﬀects and the potential for selection of pan-resistant strains. We conclude that use of aggressive second-line antituberculosis treatment regimens from the outset— with proper management of adverse events—can lead to positive outcomes in a signiﬁcant proportion of patients. Treatment duration is another timely question that requires rigorous examination through clinical trials. SK, IG, and SS received partial salary support or travel support from the Bill & Melinda Gates Foundation and from the Eli Lilly Foundation. SK and SS received salary support from the Frank Hatch Fellowships in Global Health Equity at the Brigham & Women’s Hospital. SS received additional salary support from the Infectious Disease Society of America, the Heiser Foundation, and the US National Institutes of Health. SK received travel and research funding from the John D and Catherine T MacArthur Foundation.

*Salmaan Keshavjee, Irina Gelmanova, Sid Atwood, Sonya Shin [email protected] Department of Global Health and Social Medicine, Harvard Medical School, 641 Huntington Avenue, Boston, MA 02115, USA (SK); Division of Global Health Equity, Brigham and Women’s Hospital, Boston, MA, USA (SK, SA, SS); and Partners In Health, Boston, MA, USA (IG) 1

2

3

4

5

Panos Pictures

The printed journal includes an image merely for illustration See Department of Error page 30

28

Mitnick CD, Shin SS, Seung KJ, et al. Comprehensive treatment of extensively drug-resistant tuberculosis. N Engl J Med 2008; 359: 563–74. Mukherjee JS, Rich ML, Socci AR, et al. Programmes and principles in treatment of multidrug-resistant tuberculosis. Lancet 2004; 363: 474–81. Johnson JL, Hadad DJ, Boom WH, et al. Early and extended early bactericidal activity of levoﬂoxacin, gatiﬂoxacin and moxiﬂoxacin in pulmonary tuberculosis. Int J Tuberc Lung Dis 2006; 10: 605–12. Mitnick CD, Castro KG, Harrington M, Sacks LV, Burman W. Randomized trials to optimize treatment of multidrug-resistant tuberculosis. PLoS Med 2007; 4: e292. The Cambridge Declaration: towards clinical trials for drug-resistant tuberculosis. Cambridge, MA, USA; June 12, 2008. http://www.ghdonline. org/drtb/resource/the-cambridge-declarationtowards-clinical-trials-/ (accessed Dec 3, 2008).

Unsexy tuberculosis XDR-TB—extensively drug-resistant tuberculosis. The acronym makes one cringe with fear. Here we have an increasingly prevalent disease that is potentially as lethal as any weapon

of terrorism. But what is it really, and what is the secret to conquering it? First described in 2006 in KwaZulu Natal, South Africa,1 XDR-TB killed 52 of the 53 HIV-infected people it aﬀected. In 2007, when Andrew Speaker was diagnosed in Atlanta, GA, USA and placed on (but did not heed) a “no-ﬂy” order on the basis of his potential for disease transmission, the world’s press reacted with almost unprecedented concern. As reports of cohorts of patients followed from reputable centres,2,3 with cure rates similar to survival rates for tuberculosis in the “pre-drug” era, the Stop TB Department of WHO published an emergency response plan appropriately calling for integration of global resistant and sensitive tuberculosis control programmes.4 Cries for new drugs, diagnostics, and vaccines soon followed. Although all of this concern and attention is appropriate, it seems to skip the more important message. Certainly XDR-TB or MDR-TB (multiple-drug-resistant tuberculosis) is frightening. Certainly we need new drugs and diagnostics and the capacity to allow us to address the speciﬁcs of drug-resistant tuberculosis. But the key aspect that has been largely forgotten in all of the hype is that XDR-TB, MDR-TB, and drug-sensitive tuberculosis are all the same disease. The only diﬀerence is that MDRTB is drug-sensitive tuberculosis modiﬁed by inappropriate treatment or drug taking, and XDR-TB is MDRTB thus modiﬁed. In other words, every person with MDR-TB or XDRTB was not treated properly, did not take their drugs properly, or were infected by somebody who was not treated properly or did not take their medicines properly. In 2000, I coauthored an editorial5 pointing out that, with the growing acceptance of the necessity for prescribing MDR-TB drug therapy (then called DOTS Plus), this far more “sexy” aspect of tuberculosis care and control would distract attention

and resources away from basic tuberculosis control and therefore interfere with solving the problem. I am afraid that that is exactly what is being seen now. We need to recognise that there are more than 9 000 000 new active drugsensitive cases of tuberculosis globally that could be feeding drug resistance. It might be a less sexy concept, but they all must be appropriately treated with current strategies (as well as new diagnostics, drugs, vaccines, and proper infection control measures) to avoid preventable MDR-TB and XDR-TB, which are always lurking. Preventing active, drug-sensitive tuberculosis, or treating it properly, should be everybody’s priority: it is the only way to prevent MDR-TB and XDR-TB. I declare that I have no conﬂict of interest.

Lee B Reichman [email protected] New Jersey Medical School, Global Tuberculosis Institute, Newark, NJ 07101, USA 1

2

3

4

5

Gandhi NR, Moll A, Sturm AW, et al. Extensively drug-resistant tuberculosis as a cause of death in patients co-infected with tuberculosis and HIV in a rural area of South Africa. Lancet 2006; 368: 1575–80. Mitnick CD, Shin SS, Seung KJ, et al. Comprehensive treatment of extensively drug-resistant tuberculosis. N Engl J Med 2008; 359: 563–74. Keshavjee S, Gelmanova IY, Farmer PE, et al. Treatment of extensively drug-resistant tuberculosis in Tomsk, Russia: a retrospective cohort study. Lancet 2008; 372: 1403–09. Stop TB Partnership. The global MDR-TB and XDR-TB response plan 2007–2008. Geneva: World Health Organization, 2007. Lambregts-Van Weezenbeek K, Reichman LB. DOTS and DOTS-Plus: what’s in a name. Int J Tuberc Lung Dis 2000; 4: 995–96.

Intermittent preventive treatment with antimalarial drugs The Article by Siân Clarke and colleagues (July 12, p 127)1 provides valuable information about the eﬃcacy of intermittent preventive treatment (IPT) of malaria on health and education in schoolchildren in Kenya. We have a few concerns, however. www.thelancet.com Vol 373 January 3, 2009

Correspondence

avoid unnecessary toxic eﬀects and the potential for selection of pan-resistant strains. We conclude that use of aggressive second-line antituberculosis treatment regimens from the outset— with proper management of adverse events—can lead to positive outcomes in a signiﬁcant proportion of patients. Treatment duration is another timely question that requires rigorous examination through clinical trials. SK, IG, and SS received partial salary support or travel support from the Bill & Melinda Gates Foundation and from the Eli Lilly Foundation. SK and SS received salary support from the Frank Hatch Fellowships in Global Health Equity at the Brigham & Women’s Hospital. SS received additional salary support from the Infectious Disease Society of America, the Heiser Foundation, and the US National Institutes of Health. SK received travel and research funding from the John D and Catherine T MacArthur Foundation.

*Salmaan Keshavjee, Irina Gelmanova, Sid Atwood, Sonya Shin [email protected] Department of Global Health and Social Medicine, Harvard Medical School, 641 Huntington Avenue, Boston, MA 02115, USA (SK); Division of Global Health Equity, Brigham and Women’s Hospital, Boston, MA, USA (SK, SA, SS); and Partners In Health, Boston, MA, USA (IG) 1

2

3

4

5

Panos Pictures

The printed journal includes an image merely for illustration See Department of Error page 30

28

Mitnick CD, Shin SS, Seung KJ, et al. Comprehensive treatment of extensively drug-resistant tuberculosis. N Engl J Med 2008; 359: 563–74. Mukherjee JS, Rich ML, Socci AR, et al. Programmes and principles in treatment of multidrug-resistant tuberculosis. Lancet 2004; 363: 474–81. Johnson JL, Hadad DJ, Boom WH, et al. Early and extended early bactericidal activity of levoﬂoxacin, gatiﬂoxacin and moxiﬂoxacin in pulmonary tuberculosis. Int J Tuberc Lung Dis 2006; 10: 605–12. Mitnick CD, Castro KG, Harrington M, Sacks LV, Burman W. Randomized trials to optimize treatment of multidrug-resistant tuberculosis. PLoS Med 2007; 4: e292. The Cambridge Declaration: towards clinical trials for drug-resistant tuberculosis. Cambridge, MA, USA; June 12, 2008. http://www.ghdonline. org/drtb/resource/the-cambridge-declarationtowards-clinical-trials-/ (accessed Dec 3, 2008).

Unsexy tuberculosis XDR-TB—extensively drug-resistant tuberculosis. The acronym makes one cringe with fear. Here we have an increasingly prevalent disease that is potentially as lethal as any weapon

of terrorism. But what is it really, and what is the secret to conquering it? First described in 2006 in KwaZulu Natal, South Africa,1 XDR-TB killed 52 of the 53 HIV-infected people it aﬀected. In 2007, when Andrew Speaker was diagnosed in Atlanta, GA, USA and placed on (but did not heed) a “no-ﬂy” order on the basis of his potential for disease transmission, the world’s press reacted with almost unprecedented concern. As reports of cohorts of patients followed from reputable centres,2,3 with cure rates similar to survival rates for tuberculosis in the “pre-drug” era, the Stop TB Department of WHO published an emergency response plan appropriately calling for integration of global resistant and sensitive tuberculosis control programmes.4 Cries for new drugs, diagnostics, and vaccines soon followed. Although all of this concern and attention is appropriate, it seems to skip the more important message. Certainly XDR-TB or MDR-TB (multiple-drug-resistant tuberculosis) is frightening. Certainly we need new drugs and diagnostics and the capacity to allow us to address the speciﬁcs of drug-resistant tuberculosis. But the key aspect that has been largely forgotten in all of the hype is that XDR-TB, MDR-TB, and drug-sensitive tuberculosis are all the same disease. The only diﬀerence is that MDRTB is drug-sensitive tuberculosis modiﬁed by inappropriate treatment or drug taking, and XDR-TB is MDRTB thus modiﬁed. In other words, every person with MDR-TB or XDRTB was not treated properly, did not take their drugs properly, or were infected by somebody who was not treated properly or did not take their medicines properly. In 2000, I coauthored an editorial5 pointing out that, with the growing acceptance of the necessity for prescribing MDR-TB drug therapy (then called DOTS Plus), this far more “sexy” aspect of tuberculosis care and control would distract attention

and resources away from basic tuberculosis control and therefore interfere with solving the problem. I am afraid that that is exactly what is being seen now. We need to recognise that there are more than 9 000 000 new active drugsensitive cases of tuberculosis globally that could be feeding drug resistance. It might be a less sexy concept, but they all must be appropriately treated with current strategies (as well as new diagnostics, drugs, vaccines, and proper infection control measures) to avoid preventable MDR-TB and XDR-TB, which are always lurking. Preventing active, drug-sensitive tuberculosis, or treating it properly, should be everybody’s priority: it is the only way to prevent MDR-TB and XDR-TB. I declare that I have no conﬂict of interest.

Lee B Reichman [email protected] New Jersey Medical School, Global Tuberculosis Institute, Newark, NJ 07101, USA 1

2

3

4

5

Gandhi NR, Moll A, Sturm AW, et al. Extensively drug-resistant tuberculosis as a cause of death in patients co-infected with tuberculosis and HIV in a rural area of South Africa. Lancet 2006; 368: 1575–80. Mitnick CD, Shin SS, Seung KJ, et al. Comprehensive treatment of extensively drug-resistant tuberculosis. N Engl J Med 2008; 359: 563–74. Keshavjee S, Gelmanova IY, Farmer PE, et al. Treatment of extensively drug-resistant tuberculosis in Tomsk, Russia: a retrospective cohort study. Lancet 2008; 372: 1403–09. Stop TB Partnership. The global MDR-TB and XDR-TB response plan 2007–2008. Geneva: World Health Organization, 2007. Lambregts-Van Weezenbeek K, Reichman LB. DOTS and DOTS-Plus: what’s in a name. Int J Tuberc Lung Dis 2000; 4: 995–96.

Intermittent preventive treatment with antimalarial drugs The Article by Siân Clarke and colleagues (July 12, p 127)1 provides valuable information about the eﬃcacy of intermittent preventive treatment (IPT) of malaria on health and education in schoolchildren in Kenya. We have a few concerns, however. www.thelancet.com Vol 373 January 3, 2009

Correspondence

The average prevalence of anaemia at 12 months was 6·3% in the IPT group and 12·6% in the placebo group (risk ratio 0·52, 95% CI 0·29–0·93; p=0·028).1 However, the prevalence of anaemia at baseline was 3% higher in the placebo group than the IPT group.2 This diﬀerence should have been adjusted for. Widespread use of IPT antibiotics can introduce selection pressure for resistance to antimalarial drugs in the community. Bell and colleagues2 in Malawi reported that the prevalence of the pfmdr1 86Y mutation was higher after prophylaxis with sulfadoxinepyrimethamine plus amodiaquine than after sulfadoxine-pyrimethamine alone (p=0·002); amodiaquine exerted selective pressure for resistanceassociated mutations many weeks after treatment.3 Additionally, prophylaxis with sulfadoxine-pyrimethamine plus amodiaquine has been associated with striking declines in neutrophil counts compared with other IPT regimens.2 Has there been any emergence of amodiaquine resistance in the malarial parasite due to IPT? Are there any data on the children’s neutrophil counts and any associated ill eﬀects? There are about 500 million cases of malaria and 1 million deaths from malaria each year, mainly in poor populations in tropical and subtropical areas where the temperature and rainfall are most suitable for the development of the malaria-causing Plasmodium parasites in Anopheles mosquitoes.3 We feel that concepts other than IPT are required to achieve eradication here. Transmission blocking, live attenuated parasite vaccines, and vector targeting are potential new candidates. We declare that we have no conﬂict of interest.

*Dewan Sakhawat Billal, Muneki Hotomi, Noboru Yamanaka [email protected] Division of Infection and Immunity Research Center, Department of Otolaryngology-Head and Neck Surgery, Wakayama Medical University, Wakayama 641-8509, Japan

www.thelancet.com Vol 373 January 3, 2009

1

2

3

Clarke SE, Jukes MC, Njagi JK, et al. Eﬀect of intermittent preventive treatment of malaria on health and education in schoolchildren: a cluster-randomised, double-blind, placebo-controlled trial. Lancet 2008; 372: 127–38. Bell DJ, Nyirongo SK, Mukaka M, et al. Sulfadoxine-pyrimethamine-based combinations for malaria: a randomised blinded trial to compare eﬃcacy, safety and selection of resistance in Malawi. PLoS ONE 2008; 3: e1578. Greenwood BM, Fidock DA, Kyle DE, et al. Malaria: progress, perils, and prospects for eradication. J Clin Invest 2008; 118: 1266–76.

Authors’ reply Dewan Billal and colleagues highlight two important issues for intermittent preventive treatment (IPT) with sulfadoxine-pyrimethamine and amodiaquine, namely the potential selection of drug-resistant parasites and the safety of amodiaquine. In our study, we compared the prevalence of mutations in the pfdhfr and pfdhps genes, associated with resistance to sulfadoxine-pyrimethamine, among 537 children who were parasite-positive 12 months after the intervention. No diﬀerences between the intervention and placebo groups were evident, but the prevalence of resistance-associated alleles was already close to 100% in both groups. The prevalence of pfdhps alleles associated with sulfadoxine-pyrimethamine sensitivity also did not differ; however, wild-type alleles of pfdhfr were signiﬁcantly less prevalent in the IPT group: 6% versus 18% (p=0·03) and 20% versus 42% (p=0·001) for codons 51-N and 59-C, respectively. It is possible, therefore, that selection against wild-type parasites could lead to a gradual increase in resistance within the parasite population over time, although such risks could be balanced by the small proportion of children still infected after treatment (<5%) and acquired immunity among older children. Resistance to sulfadoxine-pyrimethamine is enhanced by its slow elimination from the body, the prophylactic eﬀect of which might also explain its eﬃcacy.1 Use of more rapidly eliminated drug combinations for IPT would decrease selection for resistance, but might confer limited prophylactic beneﬁt. Moreover, there seems wide-

spread consensus that drugs used for IPT should diﬀer from those used for ﬁrst-line treatment, thereby ruling out the use of artemisinin combination therapies for IPT. Unfortunately, our trial design did not permit the collection of serial neutrophil counts routinely on all participants, and we cannot comment on this from our data. We recognise the need for further studies on the safety of diﬀerent drug combinations for IPT among schoolchildren, which we are currently investigating in a study in Uganda. Billal and colleagues also note that anaemia was more prevalent in the control group at baseline. However, this diﬀerence was small and the published results were fully adjusted for nutritional status and other potential confounders (anaemia: risk ratio 0·52, 95% CI 0·29–0·93; p=0·028). Analysis was also repeated for 2471 children for whom baseline haemoglobin data were available, and this yielded similar results (risk ratio 0·44, 0·20–0·94; p=0·036). Therefore, we regard our results as providing convincing evidence that the observed eﬀect of IPT on anaemia cannot be attributed to a slight imbalance between study groups at baseline, and were mainly caused by reductions in malaria parasitaemia. Several scientiﬁc and operational issues on the school-based delivery of IPT remain to be investigated, but we believe the insidious eﬀect of asymptomatic malaria parasitaemia on the health and education of schoolchildren in Africa justiﬁes its investigation further. We declare we have no conﬂict of interest.

*Siân Clarke, J Kiambo Njagi, Bonnie Cundill, Rachel Hallett, Simon Brooker [email protected] London School of Hygiene and Tropical Medicine, Keppel St, London, WC1E 7HT, UK (SC, BC, RH, SB); Division of Malaria Control, Ministry of Health, Kenya (JKN); and KEMRI-Wellcome Trust Collaborative Programme, Nairobi, Kenya (SB) 1

White NJ. Intermittent presumptive treatment for malaria. PLoS Med 2005; 2: e3.

29

Correspondence

The average prevalence of anaemia at 12 months was 6·3% in the IPT group and 12·6% in the placebo group (risk ratio 0·52, 95% CI 0·29–0·93; p=0·028).1 However, the prevalence of anaemia at baseline was 3% higher in the placebo group than the IPT group.2 This diﬀerence should have been adjusted for. Widespread use of IPT antibiotics can introduce selection pressure for resistance to antimalarial drugs in the community. Bell and colleagues2 in Malawi reported that the prevalence of the pfmdr1 86Y mutation was higher after prophylaxis with sulfadoxinepyrimethamine plus amodiaquine than after sulfadoxine-pyrimethamine alone (p=0·002); amodiaquine exerted selective pressure for resistanceassociated mutations many weeks after treatment.3 Additionally, prophylaxis with sulfadoxine-pyrimethamine plus amodiaquine has been associated with striking declines in neutrophil counts compared with other IPT regimens.2 Has there been any emergence of amodiaquine resistance in the malarial parasite due to IPT? Are there any data on the children’s neutrophil counts and any associated ill eﬀects? There are about 500 million cases of malaria and 1 million deaths from malaria each year, mainly in poor populations in tropical and subtropical areas where the temperature and rainfall are most suitable for the development of the malaria-causing Plasmodium parasites in Anopheles mosquitoes.3 We feel that concepts other than IPT are required to achieve eradication here. Transmission blocking, live attenuated parasite vaccines, and vector targeting are potential new candidates. We declare that we have no conﬂict of interest.

*Dewan Sakhawat Billal, Muneki Hotomi, Noboru Yamanaka [email protected] Division of Infection and Immunity Research Center, Department of Otolaryngology-Head and Neck Surgery, Wakayama Medical University, Wakayama 641-8509, Japan

www.thelancet.com Vol 373 January 3, 2009

1

2

3

Clarke SE, Jukes MC, Njagi JK, et al. Eﬀect of intermittent preventive treatment of malaria on health and education in schoolchildren: a cluster-randomised, double-blind, placebo-controlled trial. Lancet 2008; 372: 127–38. Bell DJ, Nyirongo SK, Mukaka M, et al. Sulfadoxine-pyrimethamine-based combinations for malaria: a randomised blinded trial to compare eﬃcacy, safety and selection of resistance in Malawi. PLoS ONE 2008; 3: e1578. Greenwood BM, Fidock DA, Kyle DE, et al. Malaria: progress, perils, and prospects for eradication. J Clin Invest 2008; 118: 1266–76.

Authors’ reply Dewan Billal and colleagues highlight two important issues for intermittent preventive treatment (IPT) with sulfadoxine-pyrimethamine and amodiaquine, namely the potential selection of drug-resistant parasites and the safety of amodiaquine. In our study, we compared the prevalence of mutations in the pfdhfr and pfdhps genes, associated with resistance to sulfadoxine-pyrimethamine, among 537 children who were parasite-positive 12 months after the intervention. No diﬀerences between the intervention and placebo groups were evident, but the prevalence of resistance-associated alleles was already close to 100% in both groups. The prevalence of pfdhps alleles associated with sulfadoxine-pyrimethamine sensitivity also did not differ; however, wild-type alleles of pfdhfr were signiﬁcantly less prevalent in the IPT group: 6% versus 18% (p=0·03) and 20% versus 42% (p=0·001) for codons 51-N and 59-C, respectively. It is possible, therefore, that selection against wild-type parasites could lead to a gradual increase in resistance within the parasite population over time, although such risks could be balanced by the small proportion of children still infected after treatment (<5%) and acquired immunity among older children. Resistance to sulfadoxine-pyrimethamine is enhanced by its slow elimination from the body, the prophylactic eﬀect of which might also explain its eﬃcacy.1 Use of more rapidly eliminated drug combinations for IPT would decrease selection for resistance, but might confer limited prophylactic beneﬁt. Moreover, there seems wide-

spread consensus that drugs used for IPT should diﬀer from those used for ﬁrst-line treatment, thereby ruling out the use of artemisinin combination therapies for IPT. Unfortunately, our trial design did not permit the collection of serial neutrophil counts routinely on all participants, and we cannot comment on this from our data. We recognise the need for further studies on the safety of diﬀerent drug combinations for IPT among schoolchildren, which we are currently investigating in a study in Uganda. Billal and colleagues also note that anaemia was more prevalent in the control group at baseline. However, this diﬀerence was small and the published results were fully adjusted for nutritional status and other potential confounders (anaemia: risk ratio 0·52, 95% CI 0·29–0·93; p=0·028). Analysis was also repeated for 2471 children for whom baseline haemoglobin data were available, and this yielded similar results (risk ratio 0·44, 0·20–0·94; p=0·036). Therefore, we regard our results as providing convincing evidence that the observed eﬀect of IPT on anaemia cannot be attributed to a slight imbalance between study groups at baseline, and were mainly caused by reductions in malaria parasitaemia. Several scientiﬁc and operational issues on the school-based delivery of IPT remain to be investigated, but we believe the insidious eﬀect of asymptomatic malaria parasitaemia on the health and education of schoolchildren in Africa justiﬁes its investigation further. We declare we have no conﬂict of interest.

*Siân Clarke, J Kiambo Njagi, Bonnie Cundill, Rachel Hallett, Simon Brooker [email protected] London School of Hygiene and Tropical Medicine, Keppel St, London, WC1E 7HT, UK (SC, BC, RH, SB); Division of Malaria Control, Ministry of Health, Kenya (JKN); and KEMRI-Wellcome Trust Collaborative Programme, Nairobi, Kenya (SB) 1

White NJ. Intermittent presumptive treatment for malaria. PLoS Med 2005; 2: e3.

29

Correspondence

Computerised clinical decision support in rural China

Getty Images

As you note in your Nov 8 Editorial (p 1608),1 a new draft of the Chinese health-care system reform plan was published on Oct 14, 2008. The government understands that improving the quality of health care in rural areas is vital. The low quality of the health-care service in rural areas is largely due to the low educational level of many physicians in those areas. Data from the Ministry of Health2 indicate that, in China as a whole, more than 60% of medical practitioners (excluding assistant medical practitioners) have 3 years or fewer of professional training, and around 30% of the whole medical force has only a senior highschool level of education. The average level of education is even lower in rural areas.3 It would take too long to retrain most of these doctors to 21st-century standards. However, with extra funding, it might be possible to boost their performance by equipping them with a computerised clinical-decision support system. Such systems have greatly improved over the past 30 years.4,5 A system tailored to rural doctors could work as an electronic manual, allowing physicians to follow diagnostic algorithms, and providing updated treatment plans and prevention guidelines—something that rural doctors might never have learned before. A tailored computerised support system for rural doctors in China could be developed with today’s advanced technology, within a short time, and at low cost. China has acknowledged problems with its health service, and, by leapfrogging ahead to computerassisted diagnosis and treatment, might be able to overcome some of its present diﬃculties. We declare that we have no conﬂict of interest.

30

*Wenbin Liang, Colin W Binns, Andy H Lee [email protected] School of Public Health, Curtin University of Technology, Perth, WA 6102, Australia 1 2

3

4

5

The Lancet. China unveils plans for health-care reform. Lancet 2008; 372: 1608. Ministry of Health of the People’s Republic of China. Percentages of medical professionals by sex, age, qualiﬁcation and title in 2005. China health statistics yearbook 2007. Beijing: Ministry of Health of the People’s Republic of China, 2007. Wang J, Kushner K, Frey JJ 3rd, Ping Du X, Qian N. Primary care reform in the Peoples’ Republic of China: implications for training family physicians for the world’s largest country. Fam Med 2007; 39: 639–43. Berner ES, Webster GD, Shugerman AA, et al. Performance of four computer-based diagnostic systems. N Engl J Med 1994; 330: 1792–96. Garg AX, Adhikari NK, McDonald H, et al. Eﬀects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. JAMA 2005; 293: 1223–38.

26%. China also has the world’s most rapid growth rate of resistance (22% average growth in a study spanning 1994 to 2000).5 The faceless threat of antibiotic resistance is likely to be one of the greatest challenges to global health during the 21st century, with a direct eﬀect on health indicators in lowincome, middle-income, and highincome countries. It is positive that China’s new health system reform suggests a pharmaceutical policy that includes a strategy for rational drug use. What happens in China matters for the world. We declare that we have no conﬂict of interest.

*Andreas Heddini, Otto Cars, Sun Qiang, Göran Tomson [email protected]

Antibiotic resistance in China—a major future challenge Longde Wang and colleagues (Nov 1, p 1697)1 outline the future challenges posed by infectious diseases in the Chinese context. Infectious diseases remain a major problem in China today and Wang and colleagues provide a comprehensive review of current and emerging infectious diseases and their control. Strikingly, however, the increasing threat of antibiotic resistance is only brieﬂy mentioned. The situation with respect to overuse of antibiotics and antibiotic resistance in China is severe.2 Several factors are involved, including a health system with strong ﬁnancial incentives for drug prescribing.3 Around 75% of patients with seasonal inﬂuenza are estimated to be prescribed antibiotics, and the rate of antibiotic prescription to inpatients is 80%.4 In a study of resistance patterns of several common bacteria in China in 1999 and 2001,5 the mean prevalence of resistance among hospital-acquired infections was as high as 41% and that among community-acquired infections was

Centre for Microbiological Preparedness, Swedish Institute for Infectious Disease Control, Nobels väg 18, SE-171 82 Solna, Sweden (AH); Action on Antibiotic Resistance, Uppsala University, Uppsala, Sweden (OC); Center for Health Management and Policy, Shandong University, Jinan, China (SQ); Division of International Health, Department of Public Health Sciences, Karolinska Institutet, Stockholm, Sweden (GT) 1

2

3

4

5

Wang L, Wang Y, Jin S, et al. Emergence and control of infectious diseases in China. Lancet 2008; 372: 1598–605. Reynolds L, McKee M. Factors inﬂuencing antibiotic prescribing in China: an exploratory analysis. Health Policy 2008; published online Oct 13. DOI: 10.1016/ j.healthpol.2008.09.002. Sun Q, Santoro MA, Meng Q, Liu C, Eggleston K. Pharmaceutical policy in China. Health Aﬀ 2008; 27: 1042–50. Zheng Y, Zhou Z. The root causes of the abuse of antibiotics, harm and the rational use of the strategy. Hospital Management Forum 2007; 123: 23–27 (in Chinese). Zhang R, Eggleston K, Rotimi V, Zeckhauser RJ. Antibiotic resistance as a global threat: evidence from China, Kuwait and the United States. Global Health 2006; 2: 6.

Department of Error Clarke SE, Jukes MC, Njagi JK, et al. Eﬀect of intermittent preventive treatment of malaria on health and education in schoolchildren: a clusterrandomised, double-blind, placebo-controlled trial. Lancet 2008; 372: 127–38—In this Article (July 12), in paragraph four of the Results section (p 134), the second sentence should have read: “The eﬀect size was 0·18 (95% CI 0·003–0·35) in the counting sounds test and 0·48 (0·09–0·88) in the code transmission test.”

www.thelancet.com Vol 373 January 3, 2009

Correspondence

Computerised clinical decision support in rural China

Getty Images

As you note in your Nov 8 Editorial (p 1608),1 a new draft of the Chinese health-care system reform plan was published on Oct 14, 2008. The government understands that improving the quality of health care in rural areas is vital. The low quality of the health-care service in rural areas is largely due to the low educational level of many physicians in those areas. Data from the Ministry of Health2 indicate that, in China as a whole, more than 60% of medical practitioners (excluding assistant medical practitioners) have 3 years or fewer of professional training, and around 30% of the whole medical force has only a senior highschool level of education. The average level of education is even lower in rural areas.3 It would take too long to retrain most of these doctors to 21st-century standards. However, with extra funding, it might be possible to boost their performance by equipping them with a computerised clinical-decision support system. Such systems have greatly improved over the past 30 years.4,5 A system tailored to rural doctors could work as an electronic manual, allowing physicians to follow diagnostic algorithms, and providing updated treatment plans and prevention guidelines—something that rural doctors might never have learned before. A tailored computerised support system for rural doctors in China could be developed with today’s advanced technology, within a short time, and at low cost. China has acknowledged problems with its health service, and, by leapfrogging ahead to computerassisted diagnosis and treatment, might be able to overcome some of its present diﬃculties. We declare that we have no conﬂict of interest.

30

*Wenbin Liang, Colin W Binns, Andy H Lee [email protected] School of Public Health, Curtin University of Technology, Perth, WA 6102, Australia 1 2

3

4

5

The Lancet. China unveils plans for health-care reform. Lancet 2008; 372: 1608. Ministry of Health of the People’s Republic of China. Percentages of medical professionals by sex, age, qualiﬁcation and title in 2005. China health statistics yearbook 2007. Beijing: Ministry of Health of the People’s Republic of China, 2007. Wang J, Kushner K, Frey JJ 3rd, Ping Du X, Qian N. Primary care reform in the Peoples’ Republic of China: implications for training family physicians for the world’s largest country. Fam Med 2007; 39: 639–43. Berner ES, Webster GD, Shugerman AA, et al. Performance of four computer-based diagnostic systems. N Engl J Med 1994; 330: 1792–96. Garg AX, Adhikari NK, McDonald H, et al. Eﬀects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. JAMA 2005; 293: 1223–38.

26%. China also has the world’s most rapid growth rate of resistance (22% average growth in a study spanning 1994 to 2000).5 The faceless threat of antibiotic resistance is likely to be one of the greatest challenges to global health during the 21st century, with a direct eﬀect on health indicators in lowincome, middle-income, and highincome countries. It is positive that China’s new health system reform suggests a pharmaceutical policy that includes a strategy for rational drug use. What happens in China matters for the world. We declare that we have no conﬂict of interest.

*Andreas Heddini, Otto Cars, Sun Qiang, Göran Tomson [email protected]

Antibiotic resistance in China—a major future challenge Longde Wang and colleagues (Nov 1, p 1697)1 outline the future challenges posed by infectious diseases in the Chinese context. Infectious diseases remain a major problem in China today and Wang and colleagues provide a comprehensive review of current and emerging infectious diseases and their control. Strikingly, however, the increasing threat of antibiotic resistance is only brieﬂy mentioned. The situation with respect to overuse of antibiotics and antibiotic resistance in China is severe.2 Several factors are involved, including a health system with strong ﬁnancial incentives for drug prescribing.3 Around 75% of patients with seasonal inﬂuenza are estimated to be prescribed antibiotics, and the rate of antibiotic prescription to inpatients is 80%.4 In a study of resistance patterns of several common bacteria in China in 1999 and 2001,5 the mean prevalence of resistance among hospital-acquired infections was as high as 41% and that among community-acquired infections was

Centre for Microbiological Preparedness, Swedish Institute for Infectious Disease Control, Nobels väg 18, SE-171 82 Solna, Sweden (AH); Action on Antibiotic Resistance, Uppsala University, Uppsala, Sweden (OC); Center for Health Management and Policy, Shandong University, Jinan, China (SQ); Division of International Health, Department of Public Health Sciences, Karolinska Institutet, Stockholm, Sweden (GT) 1

2

3

4

5

Wang L, Wang Y, Jin S, et al. Emergence and control of infectious diseases in China. Lancet 2008; 372: 1598–605. Reynolds L, McKee M. Factors inﬂuencing antibiotic prescribing in China: an exploratory analysis. Health Policy 2008; published online Oct 13. DOI: 10.1016/ j.healthpol.2008.09.002. Sun Q, Santoro MA, Meng Q, Liu C, Eggleston K. Pharmaceutical policy in China. Health Aﬀ 2008; 27: 1042–50. Zheng Y, Zhou Z. The root causes of the abuse of antibiotics, harm and the rational use of the strategy. Hospital Management Forum 2007; 123: 23–27 (in Chinese). Zhang R, Eggleston K, Rotimi V, Zeckhauser RJ. Antibiotic resistance as a global threat: evidence from China, Kuwait and the United States. Global Health 2006; 2: 6.

Department of Error Clarke SE, Jukes MC, Njagi JK, et al. Eﬀect of intermittent preventive treatment of malaria on health and education in schoolchildren: a clusterrandomised, double-blind, placebo-controlled trial. Lancet 2008; 372: 127–38—In this Article (July 12), in paragraph four of the Results section (p 134), the second sentence should have read: “The eﬀect size was 0·18 (95% CI 0·003–0·35) in the counting sounds test and 0·48 (0·09–0·88) in the code transmission test.”

www.thelancet.com Vol 373 January 3, 2009

Correspondence

Computerised clinical decision support in rural China

Getty Images

As you note in your Nov 8 Editorial (p 1608),1 a new draft of the Chinese health-care system reform plan was published on Oct 14, 2008. The government understands that improving the quality of health care in rural areas is vital. The low quality of the health-care service in rural areas is largely due to the low educational level of many physicians in those areas. Data from the Ministry of Health2 indicate that, in China as a whole, more than 60% of medical practitioners (excluding assistant medical practitioners) have 3 years or fewer of professional training, and around 30% of the whole medical force has only a senior highschool level of education. The average level of education is even lower in rural areas.3 It would take too long to retrain most of these doctors to 21st-century standards. However, with extra funding, it might be possible to boost their performance by equipping them with a computerised clinical-decision support system. Such systems have greatly improved over the past 30 years.4,5 A system tailored to rural doctors could work as an electronic manual, allowing physicians to follow diagnostic algorithms, and providing updated treatment plans and prevention guidelines—something that rural doctors might never have learned before. A tailored computerised support system for rural doctors in China could be developed with today’s advanced technology, within a short time, and at low cost. China has acknowledged problems with its health service, and, by leapfrogging ahead to computerassisted diagnosis and treatment, might be able to overcome some of its present diﬃculties. We declare that we have no conﬂict of interest.

30

*Wenbin Liang, Colin W Binns, Andy H Lee [email protected] School of Public Health, Curtin University of Technology, Perth, WA 6102, Australia 1 2

3

4

5

The Lancet. China unveils plans for health-care reform. Lancet 2008; 372: 1608. Ministry of Health of the People’s Republic of China. Percentages of medical professionals by sex, age, qualiﬁcation and title in 2005. China health statistics yearbook 2007. Beijing: Ministry of Health of the People’s Republic of China, 2007. Wang J, Kushner K, Frey JJ 3rd, Ping Du X, Qian N. Primary care reform in the Peoples’ Republic of China: implications for training family physicians for the world’s largest country. Fam Med 2007; 39: 639–43. Berner ES, Webster GD, Shugerman AA, et al. Performance of four computer-based diagnostic systems. N Engl J Med 1994; 330: 1792–96. Garg AX, Adhikari NK, McDonald H, et al. Eﬀects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. JAMA 2005; 293: 1223–38.

26%. China also has the world’s most rapid growth rate of resistance (22% average growth in a study spanning 1994 to 2000).5 The faceless threat of antibiotic resistance is likely to be one of the greatest challenges to global health during the 21st century, with a direct eﬀect on health indicators in lowincome, middle-income, and highincome countries. It is positive that China’s new health system reform suggests a pharmaceutical policy that includes a strategy for rational drug use. What happens in China matters for the world. We declare that we have no conﬂict of interest.

*Andreas Heddini, Otto Cars, Sun Qiang, Göran Tomson [email protected]

Antibiotic resistance in China—a major future challenge Longde Wang and colleagues (Nov 1, p 1697)1 outline the future challenges posed by infectious diseases in the Chinese context. Infectious diseases remain a major problem in China today and Wang and colleagues provide a comprehensive review of current and emerging infectious diseases and their control. Strikingly, however, the increasing threat of antibiotic resistance is only brieﬂy mentioned. The situation with respect to overuse of antibiotics and antibiotic resistance in China is severe.2 Several factors are involved, including a health system with strong ﬁnancial incentives for drug prescribing.3 Around 75% of patients with seasonal inﬂuenza are estimated to be prescribed antibiotics, and the rate of antibiotic prescription to inpatients is 80%.4 In a study of resistance patterns of several common bacteria in China in 1999 and 2001,5 the mean prevalence of resistance among hospital-acquired infections was as high as 41% and that among community-acquired infections was

Centre for Microbiological Preparedness, Swedish Institute for Infectious Disease Control, Nobels väg 18, SE-171 82 Solna, Sweden (AH); Action on Antibiotic Resistance, Uppsala University, Uppsala, Sweden (OC); Center for Health Management and Policy, Shandong University, Jinan, China (SQ); Division of International Health, Department of Public Health Sciences, Karolinska Institutet, Stockholm, Sweden (GT) 1

2

3

4

5

Wang L, Wang Y, Jin S, et al. Emergence and control of infectious diseases in China. Lancet 2008; 372: 1598–605. Reynolds L, McKee M. Factors inﬂuencing antibiotic prescribing in China: an exploratory analysis. Health Policy 2008; published online Oct 13. DOI: 10.1016/ j.healthpol.2008.09.002. Sun Q, Santoro MA, Meng Q, Liu C, Eggleston K. Pharmaceutical policy in China. Health Aﬀ 2008; 27: 1042–50. Zheng Y, Zhou Z. The root causes of the abuse of antibiotics, harm and the rational use of the strategy. Hospital Management Forum 2007; 123: 23–27 (in Chinese). Zhang R, Eggleston K, Rotimi V, Zeckhauser RJ. Antibiotic resistance as a global threat: evidence from China, Kuwait and the United States. Global Health 2006; 2: 6.

Department of Error Clarke SE, Jukes MC, Njagi JK, et al. Eﬀect of intermittent preventive treatment of malaria on health and education in schoolchildren: a clusterrandomised, double-blind, placebo-controlled trial. Lancet 2008; 372: 127–38—In this Article (July 12), in paragraph four of the Results section (p 134), the second sentence should have read: “The eﬀect size was 0·18 (95% CI 0·003–0·35) in the counting sounds test and 0·48 (0·09–0·88) in the code transmission test.”

www.thelancet.com Vol 373 January 3, 2009

Articles

Second-generation versus ﬁrst-generation antipsychotic drugs for schizophrenia: a meta-analysis Stefan Leucht, Caroline Corves, Dieter Arbter, Rolf R Engel, Chunbo Li, John M Davis

Summary Background Because of the debate about whether second-generation antipsychotic drugs are better than ﬁrst-generation antipsychotic drugs, we did a meta-analysis of randomised controlled trials to compare the eﬀects of these two types of drugs in patients with schizophrenia. Methods We compared nine second-generation antipsychotic drugs with ﬁrst-generation drugs for overall eﬃcacy (main outcome), positive, negative and depressive symptoms, relapse, quality of life, extrapyramidal side-eﬀects, weight gain, and sedation. Findings We included 150 double-blind, mostly short-term, studies, with 21 533 participants. We excluded open studies because they systematically favoured second-generation drugs. Four of these drugs were better than ﬁrst-generation antipsychotic drugs for overall eﬃcacy, with small to medium eﬀect sizes (amisulpride −0·31 [95% CI −0·44 to −0·19, p<0·0001], clozapine −0·52 [−0·75 to −0·29, p<0·0001], olanzapine −0·28 [−0·38 to −0·18, p<0·0001], and risperidone −0·13 [−0·22 to −0·05, p=0·002]). The other second-generation drugs were not more eﬃcacious than the ﬁrst-generation drugs, even for negative symptoms. Therefore eﬃcacy on negative symptoms cannot be a core component of atypicality. Second-generation antipsychotic drugs induced fewer extrapyramidal side-eﬀects than did haloperidol (even at low doses). Only a few have been shown to induce fewer extrapyramidal side-eﬀects than low-potency ﬁrst-generation antipsychotic drugs. With the exception of aripiprazole and ziprasidone, second-generation antipsychotic drugs induced more weight gain, in various degrees, than did haloperidol but not than low-potency ﬁrst-generation drugs. The second-generation drugs also diﬀered in their sedating properties. We did not note any consistent eﬀects of moderator variables, such as industry sponsorship, comparator dose, or prophylactic antiparkinsonian medication. Interpretation Second-generation antipsychotic drugs diﬀer in many properties and are not a homogeneous class. This meta-analysis provides data for individualised treatment based on eﬃcacy, side-eﬀects, and cost. Funding National Institute of Mental Health.

Introduction The high costs of second-generation (atypical) antipsychotic drugs, with $7·5 billion sales in the USA in 2003,1 has led to a continuing debate about their beneﬁts compared with ﬁrst-generation compounds. Limitations of previous reviews2,3 were that they analysed only one global eﬃcacy outcome, even though the main advantage of secondgeneration antipsychotic drugs is claimed to be their broad eﬃcacy spectrum. In particular, these drugs are thought to improve negative symptoms, depression, and quality of life more than do conventional antipsychotic drugs. Improved eﬃcacy for these problems is thought to be a major characteristic of the atypicality of second-generation antipsychotic drugs, in addition to a reduction in extrapyramidal side-eﬀects. In previous meta-analyses (apart from Cochrane reviews), side-eﬀects were not assessed thoroughly, even though they are important criteria in drug choice. Furthermore, the number of randomised controlled trials in which antipsychotic drugs were assessed is continually increasing, making new meta-analyses necessary. We present a meta-analysis of randomised controlled trials to compare the eﬀects of second-generation antipsychotic drugs with ﬁrstwww.thelancet.com Vol 373 January 3, 2009

generation antipsychotic drugs on several outcomes in patients with schizophrenia.

Methods Search

Lancet 2009; 373: 31–41 Published Online December 5, 2008 DOI:10.1016/S01406736(08)61764-X See Comment page 4 Department of Psychiatry and Psychotherapy, Technische Universität München, Munich, Germany (S Leucht MD, C Corves MSc, D Arbter MD); Department of Psychiatry and Psychotherapy, LudwigMaximilians-University, Munich, Germany (Prof R R Engel PhD, C Corves); Department of Biological Psychiatry, Shanghai Mental Health Centre, Shanghai Jiaotong University, China (C Li MD); Department of Psychiatry, Medical School, Tongji University, Shanghai, China (C Li); Psychiatric Institute, University of Illinois, Chicago, IL, USA (Prof J M Davis MD); and Maryland Psychiatric Research Center, Baltimore, MD, USA (Prof J M Davis) Correspondence to: PD Dr Stefan Leucht, Klinik und Poliklinik für Psychiatrie und Psychotherapie, Technische Universität München, Klinikum rechts der Isar, Ismaningerstr 22, 81675 Munich, Germany [email protected]

We searched (without language restrictions) the register of the Cochrane Schizophrenia Group,4 US Food and Drugs Administration website, and previous reviews2–4 for randomised controlled trials in which oral formulations of second-generation antipsychotic drugs (amisulpride, aripiprazole, clozapine, olanzapine, quetiapine, risperidone, sertindole, ziprasidone, and zotepine) were compared with ﬁrst-generation antipsychotic drugs for the treatment of schizophrenia or related disorders (schizoaﬀective, schizophreniform, or delusional disorder, and irrespective of the diagnostic criterion used). We started the search in August, 2005, and searched Medline up to October, 2006. The Cochrane Schizophrenia Group register is compiled with regular methodical searches of ten electronic databases, and supplemented with manual searching of relevant journals and conference proceedings.4 We included only those studies meeting quality criteria A (adequate randomisation) and B (usually 31

Articles

Data extraction and outcome variables

SGA better

SGA worse

Overall symptoms

Positive symptoms

Hedges’ g 1·0 Double-blind studies 0·9 Non-double-blind studies 0·8 0·7 0·6 0·5 0·4 0·3 0·2 0·1 0 –0·1 –0·2 –0·3 –0·4 –0·5 –0·6 –0·7 –0·8 –0·9 –1·0

Negative symptoms

RR

Antiparkinsonian medication

Sedation

10

1

Weight gain Hedges’ g 1·0 0·9 0·8 0·7 0·6 0·5 0·4 0·3 0·2 0·1

0·1

Figure 1: Non double-blind studies favour second-generation antipsychotic drugs Data are Hedges’ g (95% CI) and relative risk (RR; 95% CI). Similar results were obtained after correction for diﬀerences in eﬃcacy and side-eﬀects of the drugs. SGA=second-generation antipsychotic drug. Number of double-blind studies

Number of open- Q label/single-blind studies

p value*

Overall symptoms

127

49

1·57

0·2110

Positive symptoms

81

36

1·37

0·2414

Negative symptoms

101

41

11·98

0·0005

Antiparkinsonian medication

87

17

5·48

0·0192

Sedation

69

18

4·05

0·0441

Weight gain

44

6

0·38

0·540

*For diﬀerence between double-blind and open-label or single-blind studies.

Table 1: Non double-blind studies favour second-generation antipsychotic drugs

stated as randomised without details) according to the Cochrane handbook.5 For ﬁxed-dose studies, we selected only those with optimum doses of second-generation antipsychotic drugs as reported in dose-ﬁnding studies (amisulpride 50–300 mg per day for predominantly negative symptoms and 400–800 mg per day for positive symptoms, aripiprazole 10–30 mg per day, olanzapine 10–20 mg per day, quetiapine >250 mg per day, risperidone 4–6 mg per day, sertindole 16–24 mg per day, and ziprasidone 120–160 mg per day). Note that if we had used an increased threshold dose of quetiapine, the eﬃcacy would have been reduced because 750 mg per day was the least eﬀective dose6 in the only relevant study. For the Clinical Antipsychotic Trials of Intervention Eﬀectiveness (CATIE) study,7 we used the Positive and Negative Syndrome Scale (PANSS) total score and quality-of-life score, since these alone were available for all patients without tardive dyskinesia.8,9 We included studies in which medications were allowed to be switched between groups.10–12 Inclusion or exclusion of these studies and other CATIE7 results had no important eﬀect on the outcomes. 32

Two reviewers (DA, CL, SL) independently extracted all data. We contacted ﬁrst authors (if address was available) and all second-generation antipsychotic drug manufacturers for missing data. We assessed the mean overall change in symptoms, with the following order: change in PANSS total score from baseline, if not available then the change in the Brief Psychiatric Rating Scale (BPRS), and then values of these scales at study endpoint, all based on intention-to-treat datasets whenever available. We similarly analysed negative, positive, and depressive symptoms, and overall quality of life; and we analysed dichotomous-outcome responder rates for number needed to treat (NNT), number needed to harm (NNH), and relapse rates. A 50% reduction from baseline in PANSS or BPRS scores, or a score of much improved on the Clinical Global Impression Scale, were the a-priori chosen cutoﬀs;4,13 however, when these were not available, we applied the authors’ deﬁnitions of response. We analysed weight gain and sedation. For extrapyramidal side-eﬀects, the main outcome was use of antiparkinsonian medication; in comparisons with low-potency ﬁrst-generation antipsychotic drugs, use of antiparkinsonian medication was so rarely reported so we used at least one extrapyramidal side-eﬀect as the outcome in such studies. In meta-regression analyses, in which we assessed the eﬀect of prophylactic antiparkinsonian medication on diﬀerences in extrapyramidal side-eﬀects, the results of the extrapyramidal side-eﬀect rating scales were the dependent variable.

Meta-analytical calculations For continuous data, we used the standardised mean diﬀerence Hedges’ adjusted g. Unreported SD values were calculated from other statistics or from the average of the other studies. Standard inverse of the variance weighting was used when we pooled the studies. We did not apply weighting for study quality, because determination of how much weight to assign to diﬀerent quality criteria has no empirical basis.5 For dichotomous data, we applied a once randomised–analysed endpoint assessment, calculating relative risks (RR) primarily, risk diﬀerences, and NNT or NNH. Since considerable heterogeneity exists in some analyses according to the I² statistics,14 we applied the Der-Simonian and Laird15 random-eﬀects model throughout. We compared double-blind studies with open-label or single-blind studies and noted that the open-label and single-blind studies systematically favoured the secondgeneration antipsychotic drugs. We therefore based all subsequent analyses on double-blind studies. With random-eﬀect restricted maximum-likelihood metaregression or sensitivity analyses, or both, we assessed industry sponsorship, chronicity, study duration, western versus Oriental (mainly Chinese) studies, comparator dose, diﬀerences in extrapyramidal side-eﬀects between second-generation and ﬁrst-generation antipsychotic www.thelancet.com Vol 373 January 3, 2009

Articles

–0·7 –0·6 –0·5 –0·4 –0·3 –0·2 –0·1 0 0·1 0·2 0·3 0·4 –0·6 –0·5 –0·4 –0·3 –0·2 –0·1 0 0·1 0·2 0·3 0·4

ne Zo

te

pi

e id as pr Zi

Se

rti

nd

ol

on

e

e on id er sp Ri

et Qu

an

za

pi

iap in e

ne

e Ol

ap oz Cl

zo Ar

ip

ip

ra

lp isu Am

in

le

e

–0·9 –0·8 –0·7 –0·6 –0·5 –0·4 –0·3 –0·2 –0·1 0 0·1 0·2 0·3

rid

Depression SGA worse SGA better

Negative symptoms SGA worse SGA better

Positive symptoms SGA worse SGA better

Overall symptoms SGA worse SGA better

Hedges’ g –0·7 –0·6 –0·5 –0·4 –0·3 –0·2 –0·1 0 0·1 0·2

Figure 2: Second-generation versus ﬁrst-generation antipsychotic drugs—eﬃcacy in various domains Data are Hedges’ g (95% CI). Note that the results are signiﬁcant at p<0·05 if the 95% CIs do not overlap the x axis. SGA=second-generation antipsychotic drug.

drugs, prophylactic antiparkinsonian medication, and haloperidol versus low-potency comparator drug (deﬁned as less or equipotent to chlorpromazine) as potential moderators.16 With the last four moderators, we tested the hypothesis that extrapyramidal side-eﬀects induced by ﬁrst-generation antipsychotic drugs might mimic symptoms of schizophrenia and falsely suggest that second-generation drugs are better.1,2,17 We analysed the eﬀects of comparator-drug dose with the following cutoﬀs: haloperidol 12 mg per day2,16 or 7·5 mg per day (an adequate dose according to a Cochrane review),18 and, for low-potency ﬁrst-generation antipsychotic drugs, 600 mg per day17 chlorpromazine equivalents.16,17 We assessed publication bias with funnel plots.5 We did calculations with Comprehensive Meta-Analysis (version 2.2.034)19 and Stata (version 7.0). Two-sided α was set at p<0·05. We did not adjust signiﬁcance levels for multiple testing. Note that we did the sensitivity www.thelancet.com Vol 373 January 3, 2009

analyses to assess the robustness of results and not to gather signiﬁcant results.

Role of the funding source The sponsor had no inﬂuence on design, analysis, interpretation, and writing of the report. The corresponding author had full access to all the data in the study and had ﬁnal responsibility for the decision to submit for publication.

Results Search Our search yielded 4166 citations. Of 411 inspected, we excluded 107 studies for reasons of inadequate randomisation (n=50), no appropriate intervention or control group (n=29), inappropriate participants (n=2), no usable data (n=24), presentation of a subgroup only (n=1), and very short duration (ie, 5 days; n=1). Another 65 open 33

Articles

See Online for webtables 1–8 and webﬁgures 1–11

or single-blind studies were excluded after the absence of double blind was detected as a bias. We included a total of 239 publications of 150 double-blind studies with 21 533 participants. Haloperidol was the comparator drug in 95 studies, chlorpromazine in 28, perphenazine in ﬁve, ﬂuphenazine Number of studies

Number of participants

Hedges’ g (95% CI)

p value

<0·0001

Amisulpride Overall symptoms

13

1017

−0·31 (−0·44 to −0·19)

Positive symptoms

4

703

−0·22 (−0·37 to −0·06)

0·005

Negative symptoms

10

929

−0·27 (−0·40 to −0·14)

0·0001

9

900

−0·37 (−0·51 to −0·24)

<0·0001

Overall symptoms

5

2049

−0·05 (−0·14 to 0·05)

0·326

Positive symptoms

4

1983

0·03 (−0·06 to 0·12)

0·508

Negative symptoms

5

2049

−0·09 (−0·19 to 0·01)

0·079

Depression

1

1278

−0·12 (−0·24 to −0·01)

0·040

Depression Aripiprazole

Clozapine Overall symptoms

23

1997

−0·52 (−0·75 to −0·29)

<0·0001

Positive symptoms

10

1080

−0·36 (−0·56 to −0·16)

<0·0001

Negative symptoms

17

1603

−0·27 (−0·42 to −0·13)

<0·0001

6

426

−0·51 (−0·87 to −0·14)

0·006

Overall symptoms

28

4966

−0·28 (−0·38 to −0·18)

<0·0001

Positive symptoms

24

4189

−0·15 (−0·21 to −0·09)

<0·0001

Negative symptoms

24

4187

−0·32 (−0·47 to −0·16)

<0·0001

Depression

12

2893

−0·27 (−0·35 to −0·19)

<0·0001

Overall symptoms

11

2412

0·04 (−0·04 to 0·12)

0·308

Positive symptoms

9

1742

0·14 (0·03 to 0·26)

0·013

Negative symptoms

10

1926

0 (−0·09 to 0·09)

0·928

4

442

−0·23 (−0·41 to −0·04)

0·016 0·002

Depression Olanzapine

Quetiapine

Depression Risperidone Overall symptoms

34

4173

−0·13 (−0·22 to 0·05)

Positive symptoms

28

3286

−0·13 (−0·20 to −0·05)

0·001

Negative symptoms

30

3455

−0·13 (−0·21 to −0·06)

<0·0001

Depression

11

1611

−0·10 (−0·23 to 0·03)

0·145

Sertindole Overall symptoms

4

1344

0·02 (−0·13 to 0·16)

0·836

Positive symptoms

3

1145

0·17 (−0·03 to 0·36)

0·089

Negative symptoms

4

1198

−0·11 (−0·22 to 0·01)

0·068

Depression

2

574

−0·04 (−0·22 to 0·14)

0·680

Overall symptoms

5

980

0·04 (−0·08 to 0·17)

0·483

Positive symptoms

4

728

0·03 (−0·20 to 0·26)

0·813

Negative symptoms

3

691

−0·09 (−0·29 to 0·11)

0·384

Depression

3

691

0·01 (−0·14 to 0·16)

0·910

Overall symptoms

15

1125

−0·10 (−0·27 to 0·06)

0·212

Positive symptoms

2

192

0·12 (−0·16 to 0·40)

0·409

Negative symptoms

5

450

−0·23 (−0·46 to 0)

0·050

Depression

2

134

−0·14 (−0·48 to 0·20)

0·413

Ziprasidone

Zotepine

Table 2: Second-generation versus ﬁrst-generation antipsychotic drugs—eﬃcacy in various domains

34

in four, ﬂupenthixol and perazine in three each, thioridazine and levomepromazine in two each, and all other drugs (clopenthixol, zuclopenthixol, mosapramine, tiothixene, clocapramine, triﬂuoperazine, periciacine, and any ﬁrst-generation antipsychotic drugs) in one each. 35 studies were of Oriental origin; in ﬁve studies, the ﬁrst episode of schizophrenia was assessed; 121 (81%) studies lasted 12 weeks or less; 17 (11%) lasted up to 6 months; and 12 (8%) were longer than 6 months. The mean duration of illness was 11·8 years (SD 7·7) and mean age of patients was 36·2 years (7·1; webtable 1).

Outcomes Figures 1–7 and tables 1–7 summarise the ﬁndings. Webtables 2–4 show detailed statistics, meta-regressions, and sensitivity analyses; webﬁgures 1–10 show forestplots; webﬁgure 11 shows the funnel-plots, webtable 5 shows further results and discussions on comparator dose; webtable 6 shows prophylactic antiparkinsonian medications; webtable 7 shows industry-sponsorship; and webtable 8 shows eﬃcacy versus eﬀectiveness research.

Eﬀects of blinding Open-label and single-blind studies yielded signiﬁcantly higher eﬀect sizes than did double-blind studies in several domains of eﬃcacy and tolerability (ﬁgure 1; table 1). Further eﬀects of the absence of masking were noted for single second-generation drugs—eg, in the overall eﬃcacy of olanzapine (p=0·040) and quetiapine (p=0·009).

Overall eﬃcacy Five second-generation antipsychotic drugs (aripiprazole, quetiapine, sertindole, ziprasidone, and zotepine) were not signiﬁcantly diﬀerent from ﬁrst-generation antipsychotic drugs in their eﬀects on overall symptoms (ﬁgure 2; table 2). Four second-generation antipsychotic drugs—ie, amisulpride, clozapine, olanzapine, and risperidone—were more eﬃcacious (Hedges’ g −0·13 to −0·52) than ﬁrst-generation drugs (ﬁgure 2; table 2). The NNT for one additional responder was between 6 (95% CI 4–10) for amisulpride and 15 (9–36) for risperidone (webtable 4).

Speciﬁc psychopathology These four second-generation antipsychotic drugs were also more eﬃcacious than ﬁrst-generation drugs for treatment of positive and negative symptoms (ﬁgure 2; table 2). Importantly for the notion of atypicality, the other ﬁve second-generation antipsychotic drugs (ie, aripiprazole, quetiapine, sertindole, ziprasidone, and zotepine) were not more eﬀective than ﬁrst-generation drugs for treatment of negative symptoms. The drugs were also no more eﬃcacious than ﬁrst-generation antipsychotic drugs for positive symptoms, and quetiapine was less eﬃcacious. www.thelancet.com Vol 373 January 3, 2009

Articles

The pattern for depression was somewhat diﬀerent— ie, amisulpride, clozapine, olanzapine, and aripiprazole and quetiapine, were signiﬁcantly better than ﬁrstgeneration drugs, whereas risperidone was not.

sedating than was haloperidol, whereas aripiprazole (33 [20–1011]) was signiﬁcantly less sedating (ﬁgure 6; table 6). By contrast, compared with low-potency ﬁrstgeneration antipsychotic drugs, only clozapine (13 [7–220]) was signiﬁcantly more sedating (ﬁgure 6; table 6).

Relapse Relapse was reported in only 14 long-term studies. Olanzapine (four studies, 1008 participants, RR 0·67 [0·49–0·92], NNT 17 [8–100]), risperidone (5, 1174, 0·74 [0·63–0·87], 11 [7–33]), and sertindole (1, 282, 0·17 [0·04–0·73], 14 [8–50]) proved to be signiﬁcantly better than ﬁrst-generation antipsychotic drugs; amisulpride, aripiprazole, and clozapine showed no signiﬁcant diﬀerence (webtable 2). No studies were available for the other second-generation antipsychotic drugs. For quetiapine, in a large unpublished study, no diﬀerence compared with haloperidol (n=301)20 was reported, but the data necessary for meta-analytical calculations were not presented.

Eﬀects of comparator dose We did not note a clear pattern of comparator-drug dose aﬀecting the eﬃcacy of second-generation antipsychotic drugs, and the few signiﬁcant diﬀerences between studies with haloperidol at more or less than 12 mg per day or 7·5 mg per day (or chlorpromazine 600 mg equivalents for low-potency ﬁrst-generation drugs) were contradictory. Figure 7 and table 7 show results based on the haloperidol cutoﬀ of 12 mg per day. Haloperidol was given to participants at less than or equal to 7·5 mg per day in only 12 studies (webtable 5). Higher haloperidol doses usually induced more extrapyramidal side-eﬀects than did lower doses, but the eﬀects

Quality of life Quality of life was reported in only 17 studies. Only amisulpride, clozapine, and sertindole were better than ﬁrstgeneration antipsychotic drugs (ﬁgure 3; table 3). In three further olanzapine studies, no signiﬁcant diﬀerence was reported for the related idea of patients’ attitude towards treatment (n=171, −0·36 [95% CI −0·90 to 0·21, p=0·21]).

–2

–1

Hedges’ g

Sedation Clozapine (NNH 5 [3–14]), quetiapine (13 [8–20]), and zotepine (NNH not signiﬁcant)] were signiﬁcantly more www.thelancet.com Vol 373 January 3, 2009

ne pi te Zo

as id pr Zi

Se rti n

do

on

e

le

e id er sp Ri

Qu et

iap

in

on

e

ne za pi

e Ol an

in oz Cl

Figure 3: Quality of life Data are Hedges’ g (95% CI). SGA=second-generation antipsychotic drug. Number of studies

Weight gain Amisulpride, clozapine, olanzapine, quetiapine, risperidone, sertindole, and zotepine were associated with signiﬁcantly more weight gain than was haloperidol, whereas aripiprazole and ziprasidone were not (ﬁgure 5; table 5). We did not note a signiﬁcant diﬀerence between second-generation antipsychotic drugs and low-potency ﬁrst-generation drugs (ﬁgure 5; table 5).

ap

ra zo Ar ip ip

lp rid

e

le

1

isu

All second-generation antipsychotic drugs were associated with much fewer extrapyramidal side-eﬀects than haloperidol. NNT was between 2 for clozapine and 5 for zotepine (ﬁgure 4; table 4). However, with the exception of clozapine, olanzapine, and risperidone, second-generation drugs have not been shown to be better than low-potency ﬁrst-generation antipsychotic drugs, and we noted a robust superiority based on more than two studies only for clozapine (ﬁgure 4; table 4).

SGA worse

Extrapyramidal side-eﬀects

0

Am

According to textbooks, high-potency and low-potency ﬁrst-generation antipsychotic drugs are equally eﬃcacious, but diﬀer in side-eﬀects.21 Therefore, we have presented the tolerability results separately for haloperidol and low-potency comparator drugs.

SGA better

Side-eﬀects

Number of participants

Hedges’ g (95% CI)

p value

Amisulpride

1

194

−0·31 (−0·60 to −0·03)

0·030

Aripiprazole

1

206

0·06 (−0·22 to 0·33)

0·683

Clozapine

1

311

−0·24 (−0·46 to −0·01)

0·039

Olanzapine

5

1450

−0·07 (−0·23 to 0·09)

0·398

Quetiapine

2

166

0·12 (−0·18 to 0·43)

0·432

Risperidone

4

330

−0·02 (−0·23 to 0·20)

0·887

Sertindole

1

105

−0·44 (−0·83 to −0·05)

0·027

Ziprasidone

1

72

0·03 (−0·43 to 0·49)

0·905

Zotepine

1

122

−0·27 (−0·63 to 0·09)

0·138

Table 3: Quality of life

35

Articles

Number of Number of Relative risk studies participants (95% CI)

SGA versus haloperidol: use of antiparkinsonian medication RR 1·0 SGA better

SGA versus haloperidol*

0·1

0·01

SGA better

SGA worse

SGA versus low-potency FGA: at least one extrapyramidal side-eﬀect RR 10

Amisulpride

8

783

0·58 (0·45–0·76)

<0·0001

Aripiprazole

4

1794

0·45 (0·32–0·64)

<0·0001

Clozapine

3

162

0·17 (0·03–0·88)

0·035

Olanzapine

12

3670

0·39 (0·30–0·51)

<0·0001

Quetiapine

5

1167

0·43 (0·25–0·74)

0·002

Risperidone

21

2738

0·61 (0·52–0·72)

<0·0001

Sertindole

4

1472

0·36 (0·29–0·45)

<0·0001

Ziprasidone

3

501

0·50 (0·26–0·96)

0·037

Zotepine

4

398

0·59 (0·44–0·79)

<0·0001

1·00 (0·70–1·43)

1·000

SGA versus low-potency FGA†

1·0

0·1

ne te Zo

on

Amisulpride

1

30

Aripiprazole

··

··

Clozapine

11

775

Olanzapine

2

152

0·53 (0·32–0·89)

0·016

Quetiapine

2

422

0·66 (0·19–2·23)

0·503

Risperidone

2

108

0·47 (0·22–0·99)

0·046

Sertindole

··

··

Ziprasidone*

1

306

1·13 (0·91–1·41)

0·252

Zotepine

5

322

1·04 (0·76–1·42)

0·801

·· 0·66 (0·48–0·91)

··

·· 0·010

··

Zi

pr as id

pi

e*

e ol nd rti Se

on Ri sp er id

Qu

et

iap

in

e

e

ne pi za an Ol

ap oz Cl

zo ra ip ip Ar

in

le

e rid lp isu Am

e

0·01

Figure 4: Extrapyramidal side-eﬀects Data are relative risk (RR; 95% CI). SGA=second-generation antipsychotic drug. FGA=ﬁrst-generation antipsychotic drug. *Use of antiparkinsonian medication.

were small and not always consistent (ﬁgure 7; table 7; webtable 5). Only higher doses of low-potency ﬁrstgeneration antipsychotic drugs than 600 mg per day produced more extrapyramidal side-eﬀects than did clozapine, the only drug with enough studies for assessment.

Prophylactic antiparkinsonian medication In 11 studies of clozapine, olanzapine, or risperidone, prophylactic antiparkinsonian medications were used by participants in the ﬁrst-generation antipsychotic drugs’ groups. Only one meta-regression analysis (clozapine for negative symptoms) was signiﬁcant (webtable 3). The eﬃcacy eﬀect sizes were in the same range as those in the overall analysis, but the statistical signiﬁcance was inconsistent and absent for risperidone. Clozapine and olanzapine induced signiﬁcantly fewer extrapyramidal side-eﬀects than did ﬁrst-generation antipsychotic drugs despite prophylactic antiparkinsonian medication, but the eﬀect size was relatively small. Risperidone showed no diﬀerence in these side-eﬀects compared with ﬁrstgeneration antipsychotic drugs combined with prophylactic antiparkinsonian medication (webtable 6).

Industry sponsorship There were enough non-industry sponsored studies for only clozapine, olanzapine, quetiapine, and risperidone. The only signiﬁcant diﬀerence between sponsored and non-sponsored studies was noted for the eﬀect of 36

p value

SGA=second-generation antipsychotic drug. FGA=ﬁrst-generation antipsychotic drug. *Use of antiparkinsonian medication. †At least one extrapyramidal side-eﬀect.

Table 4: Extrapyramidal side-eﬀects

clozapine on positive symptoms (webtable 3). Nevertheless, when industry-sponsored studies were excluded in a sensitivity analysis, the eﬃcacy of this drug was reduced (eg, an eﬀect size of −0·22 for overall symptoms compared with −0·52 when all studies were included) but still signiﬁcant. Risperidone was not signiﬁcantly more eﬃcacious than ﬁrst-generation antipsychotic drugs for the overall change in symptoms when industrysponsored studies were excluded. The results for olanzapine and quetiapine were unchanged by sponsorship (webtable 2; webtable 7). Other moderators did not aﬀect the results in a uniform direction, and most sensitivity analyses were consistent with the main results (webtable 2; webtable 3). Funnel plots did not show a potential publication bias (webﬁgure 11). Webtable 8 compares the results of eﬃcacy and eﬀectiveness studies.

Discussion Four second-generation antipsychotic drugs—amisulpride, clozapine, olanzapine, and risperidone—were more eﬃcacious than ﬁrst-generation drugs in the main domains (overall change in symptoms, and positive and negative symptoms). The other ﬁve second-generation antipsychotic drugs were only as eﬃcacious as ﬁrst-generation antipsychotic drugs, even in terms of negative symptoms. Second-generation antipsychotic drugs caused fewer extrapyramidal side-eﬀects than did haloperidol, even www.thelancet.com Vol 373 January 3, 2009

Articles

Zotepine

3

321

2·7 (1·7 to 3·7)

<0·0001

0·3 (–3·6 to 4·2)

0·881

SGA versus low-potency FGA Amisulpride

1

30

Aripiprazole

··

··

Clozapine

3

232

Olanzapine

··

··

Quetiapine

1

201

·· 0·3 (–1·6 to 2·2) ·· 0·5 (–1·0 to 2·0)

·· 0·753 ·· 0·518

··

··

··

··

··

··

··

··

Ziprasidone

1

307

−1·1 (–2·3 to 0·2)

0·087

Zotepine

1

106

1·0 (–0·9 to 2·9)

0·306

SGA worse

Am

Risperidone Sertindole

FGA=ﬁrst-generation antipsychotic drug. SGA=second-generation antipsychotic drug.

Table 5: Weight gain

www.thelancet.com Vol 373 January 3, 2009

pi ne

0·887

te

0·1 (–1·2 to 1·3)

Zo

301

ne

1

id o

0·040

Ziprasidone

e

<0·0001

Zi pr as

3·3 (0·2 to 6·4)

ol

1·7 (0·9 to 2·4)

779

rti nd

1366

2

Se

9

Sertindole

on e

Risperidone

SGA versus low-potency FGA kg 4·0 3·5 3·0 2·5 2·0 1·5 1·0 0·5 0 –0·5 –1·0 –1·5 –2·0 –2·5 –3·3 –3·5 –4·0 er id

<0·0001

Ri sp

<0·0001

1·4 (0·7 to 2·1)

pi ne

3·3 (2·2 to 4·4)

945

et ia

2952

3

Qu

9

Quetiapine

ne

Olanzapine

pi

<0·0001

an za

3·4 (2·0 to 4·9)

Ol

170

e

3

in

0·071

Clozapine

Cl oz ap

0·012

e

0·6 (–0·1 to 1·2)

az ol

0·9 (0·2 to 1·6)

1598

e

373

2

Ar ip ip r

2

Aripiprazole

rid

Amisulpride

lp

SGA versus haloperidol

isu

p value

SGA better

Number of Mean weightparticipants gain diﬀerence (kg; 95% CI)

SGA versus haloperidol kg 7·0 6·5 6·0 5·5 5·0 4·5 4·0 3·5 3·0 2·5 2·0 1·5 1·0 0.5 0 –0·5 –1·0 –1·5 –2·0

SGA worse

Number of studies

With respect to the magnitude of the eﬃcacy eﬀect sizes, the superiority of the more eﬃcacious second-generation antipsychotic drugs was only small to medium according to Cohen’s classiﬁcation.27 For perspective, the pooled eﬀect size in another review comparing second-generation antipsychotic drugs with placebo was −0·51 and the NNT was 6.28 Diﬀerences, such as higher dropout rates in the placebo-controlled trials29 than in the active-comparatordrug-controlled trials make it impossible for us to say that the eﬃcacy of clozapine doubles the eﬃcacy compared with placebo (ie, the eﬀect size of antipsychotic drugs vs placebo is 0·5128 and the eﬀect size of clozapine vs ﬁrst-generation antipsychotic drugs is 0·52). However, schizophrenia usually aﬄicts patients for life and even a small beneﬁt could be important. In this study, second-generation antipsychotic drugs induced fewer extrapyramidal side-eﬀects than did haloperidol, and most of them even when haloperidol was used at doses less than 7·5 mg per day. In individual

SGA better

when the haloperidol dose was less than 7·5 mg per day; however, a diﬀerence between most second-generation antipsychotic drugs and low-potency ﬁrst-generation antipsychotics has not been shown. Most second-generation drugs (except aripiprazole and ziprasidone) induced more weight gain than did high-potency but not low-potency ﬁrst-generation antipsychotic drugs. Many companies claimed that improved eﬃcacy in negative symptoms is a core characteristic of atypicality.22 Our meta-analysis does not conﬁrm this common notion because the eﬀects of some second-generation antipsychotic drugs were not signiﬁcant compared with those of ﬁrst-generation drugs. The most eﬃcacious drugs were better in all eﬃcacy domains, whereas the others ones were only as eﬃcacious as ﬁrst-generation antipsychotic drugs, although the eﬀect sizes for negative symptoms were often larger than those for positive symptoms. The ﬁndings for depression were diﬀerent; risperidone did not seem to be better than ﬁrst-generation drugs, whereas aripiprazole and quetiapine were, consistent with evidence of their eﬀectiveness in major depression.23,24 Quality of life was reported in only very few studies; if a superiority of second-generation drugs was noted, the eﬀect size was in the same range as that for eﬃcacy. In another meta-analysis, second-generation antipsychotic drugs were better for global cognitive functioning (eﬀect size −0·24).25 Clozapine has been shown to reduce suicidality more than does olanzapine.26

Figure 5: Weight gain Data are mean weight-gain diﬀerence (kg; 95% CI). FGA=ﬁrst generation antipsychotic drug. SGA=second-generation antipsychotic drug.

37

Articles

SGA better

SGA worse

SGA versus haloperidol: number of patients with sedation RR 10

1·0

0·1

1·0

ne

e Zi

pr

Zo

as

te

id

pi

on

e ol nd Se

rti

Ri sp er id

iap et Qu

Ol

on

in

e

e

ne za an

ap Cl

oz

pr Ar ip i

pi

in

e ol az

rid lp isu Am

e

0·1 e

SGA better SGA worse

SGA versus low-potency FGA: number of patients with sedation RR 10

Figure 6: Sedation Data are relative risk (RR; 95% CI). FGA=ﬁrst-generation antipsychotic drug. SGA=second-generation antipsychotic drug.

Number of studies

Number of patients with sedation

Relative risk (95% CI)

p value

SGA versus haloperidol Amisulpride

4

490

0·69 (0·15–3·13)

0·634

Aripiprazole

2

1602

0·65 (0·45–0·95)

0·024

Clozapine

6

655

1·50 (1·01–2·23)

0·043

Olanzapine

6

2767

0·95 (0·82–1·10)

0·507

Quetiapine

4

970

2·07 (1·01–4·27)

0·047

Risperidone

15

2194

0·86 (0·70–1·05)

0·137

Sertindole

3

1127

0·77 (0·44–1·34)

0·360

Ziprasidone

1

301

1·59 (0·82–3·08)

0·169

Zotepine

3

221

1·86 (1·04–3·33)

0·037

SGA versus low-potency FGA Amisulpride

··

··

··

Aripiprazole

··

··

··

Clozapine

9

928

1·32 (1·10–1·59)

·· ·· 0·003

Olanzapine

1

84

0·68 (0·41–1·12)

0·132

Quetiapine

3

659

0·49 (0·23–1·03)

0·061

2·59 (0·29–22·94)

0·393

Risperidone

2

108

Sertindole

··

··

Ziprasidone

1

306

0·67 (0·44–1·01)

0·055

Zotepine

2

146

1·09 (0·69–1·73)

0·719

··

··

FGA=ﬁrst-generation antipsychotic drug. SGA=second-generation antipsychotic drug.

Table 6: Sedation

studies, treatment with haloperidol 3–4 mg per day resulted in more extrapyramidal side-eﬀects than with risperidone and sertindole.30,31 Only a few secondgeneration antipsychotic drugs have been shown to be better than low-potency ﬁrst-generation antipsychotic 38

drugs.17 A limitation of all comparisons with low-potency ﬁrst-generation drugs is the smaller evidence base than that for high-potency ﬁrst-generation drugs. Compared with haloperidol (but not low-potency ﬁrst-generation antipsychotic drugs), clozapine, olanzapine, sertindole, and zotepine induced the most weight gain, quetiapine and risperidone caused intermediate weight gain, amisulpride had little eﬀect, and aripiprazole and ziprasidone had no signiﬁcant eﬀect. Weight gain is time dependent but most studies were short-term. Nevertheless, the hierarchy is similar to that reported by Allison and colleagues.32 Clozapine, quetiapine, and zotepine were more sedating, and aripiprazole was less sedating than was haloperidol, whereas some second-generation drugs might be less sedating than are low-potency ﬁrstgeneration drugs. Concomitant use of benzodiazepines in the studies should not be ignored. Although sedation is sometimes transient, it is an important side-eﬀect, and more data are needed. The fact that absence of masking can bias the results is important because previous meta-analyses included both open-label and double-blind studies (all Cochrane reviews4 and others2,3,33). We did not note a consistent pattern of other moderators aﬀecting the results. This inconsistency supports the notion that the secondgeneration antipsychotic drugs are a heterogeneous group of drugs. However, the meta-regressions and sensitivity analyses were hampered by missing data in the predictor matrix (rarely were all outcomes in a study reported) and often by the small numbers of studies. Although the comparator-drug dose had some eﬀects on extrapyramidal side-eﬀects, a consistent eﬀect on eﬃcacy was not noted. The optimum haloperidol dose is still not known, which is a problem when it is used. In one study, about 3 mg per day was suﬃcient,34 whereas in another study the eﬃcacy increased with doses up to 20 mg per day;35 and the American Psychiatric Association guideline recommends a broad range of 5–20 mg per day.16 Use of low-potency ﬁrst-generation antipsychotic drugs does not solve all problems, because these drugs induce weight gain and cause sedation. Whether prophylactic antiparkinsonian medication can reverse the superiority in eﬃcacy of second-generation drugs cannot be shown with certainty. The eﬀects were inconsistent, and prophylactic antiparkinsonian medication was used in only 11 studies with three secondgeneration antipsychotic drugs. Although the prophylactic antiparkinsonian drug reduced the diﬀerences in extrapyramidal eﬀects, signiﬁcance was maintained for clozapine and olanzapine. Use of prophylactic antiparkinsonian drugs warrants further investigation; guidelines about their use are ambivalent.16,36 The advantages of these drugs are the avoidance of extrapyramidal side-eﬀects that can also mimic negative symptoms; disadvantages are that many patients will not have these side-eﬀects, and antiparkinsonian drugs cause anticholinergic side-eﬀects. www.thelancet.com Vol 373 January 3, 2009

Articles

Hedges’ g SGA less eﬀective SGA more eﬀective

Eﬀects on overall symptoms –1·0 –0·9 –0·8 –0·7 –0·6 –0·5 –0·4 –0·3 –0·2 –0·1 0 0·1 0·2 0·3 0·4 0·5 0·6 0·7 0·8 0·9 1

Haloperidol ≤12 mg per day Haloperidol >12 mg per day

0·1

*

1

ne

e Zo

te

pi

on id as pr Zi

nd ol e rti

e Ri

sp

er

id

on

in Qu

et

iap

pi za an Ol

Se

e

ne

e in ap oz Cl

ip ip Ar

isu Am

ra

rid

zo

le

e†

–0·1

lp

Relative risk SGA fewer extrapyramidal side-eﬀects

Eﬀects on extrapyramidal side-eﬀects (use of antiparkinsonian medication)

Figure 7: Eﬀects of haloperidol dose on eﬀect sizes of overall symptoms and extrapyramidal side-eﬀects Data are Hedge’s g (95% CI) or relative risk (95% CI). SGA=second-generation antipsychotic drug. >12 mg per day=mean eﬀect size of studies with haloperidol dose greater than 12 mg per day. ≤12 mg per day=mean eﬀect size of studies with haloperidol dose less than or equal to 12 mg per day. *The reversed dose eﬀect on extrapyramidal side-eﬀects in the quetiapine studies can be explained by two outliers (for explanation see webtable 5). †All studies had haloperidol dose more than 12 mg per day.

We could not ﬁnd a consistent eﬀect on eﬃcacy of sponsoring by industry because the results of olanzapine and quetiapine were largely unchanged. The reasons for possible sponsorship eﬀects in clozapine and risperidone studies need to be assessed in more detail. They could relate to diﬀerences in questions addressed, ﬂawed or diﬀerent designs, or selective publication of positive studies by industry.37 We have noted systematic bias in the reporting of results by industry with masked ratings of abstracts.38 We discuss our results in the context of the eﬀectiveness studies CATIE7 and Cost Utility of the Latest Antipsychotic Drugs in Schizophrenia Study (CUtLASS).12 In phase I of the CATIE study,7 olanzapine treatment resulted in the lowest discontinuation rate (all-cause and ineﬃcacy) but the largest weight gain. In phase II, clozapine was more eﬀective than the other second-generation antipsychotic drugs.39 Treatment with the ﬁrst-generation antipsychotic drug perphenazine resulted in the highest discontinuation rate because of extrapyramidal side-eﬀects, but was not diﬀerent in scale-derived extrapyramidal side-eﬀects.39 The eﬀects of second-generation drugs were not better than those of perphenazine on PANSS total score,8 cognition,40 cost,8 quality of life,9 and psychosocial functioning.9 Although some of the CATIE7 results are compatible with our ﬁndings (a detailed comparison is provided in www.thelancet.com Vol 373 January 3, 2009

Haloperidol >12 mg per day

Haloperidol ≤12 mg per day

p value*

Overall symptoms (Hedges’ g, 95% CI) Amisulpride†

··

··

··

Aripiprazole

0·30 (–0·18 to 0·78)

–0·08 (–0·18 to 0·02)

0·14

Clozapine

–0·20 (–0·36 to –0·04)

–0·67 (–1·06 to –0·27)

0·0053

Olanzapine

–0·35 (–0·60 to –0·10)

–0·21 (–0·29 to –0·14)

0·30

Quetiapine

–0·15 (–0·38 to 0·07)

0·08 (–0·06 to 0·22)

0·08

Risperidone

–0·31 (–0·45 to –0·17)

–0·09 (–0·19 to 0·01)

0·0124

Sertindole

0·21 (–0·02 to 0·43)

–0·04 (–0·17 to 0·09)

0·06

Ziprasidone

0·21 (–0·07 to 0·49)

–0·06 (–0·32 to 0·20)

0·16

Zotepine

–0·01 (–0·28 to 0·26)

–0·07 (–0·67 to 0·52)

0·85

Extrapyramidal side-eﬀects (use of antiparkinsonian medication; relative risk, 95% CI) Amisulpride

··

··

··

Aripiprazole

0·32 (0·14 to 0·69)

0·48 (0·32 to 0·72)

0·35 0·81

Clozapine

0·05 (0 to 0·75)

0·07 (0·01 to 0·49)

Olanzapine

0·25 (0·14 to 0·44)

0·47 (0·34 to 0·65)

0·05

Quetiapine

0·75 (0·60 to 0·95)

0·33 (0·20 to 0·55)

0·0040 0·0080

Risperidone

0·43 (0·29 to 0·63)

0·74 (0·67 to 0·81)

Sertindole

0·34 (0·23 to 0·49)

0·37 (0·27 to 0·51)

0·72

Ziprasidone

0·36 (0·23 to 0·57)

0·79 (0·62 to 1·02)

0·0034

Zotepine

0·60 (0·42 to 0·85)

0·40 (0·09 to 1·75)

0·60

*For comparison of the eﬀect sizes of studies with haloperidol dose less than and more than 12 mg per day.

Table 7: Eﬀects of haloperidol dose on the eﬀect sizes of overall symptoms and extrapyramidal side-eﬀects

39

Articles

webtable 8), those of the CUtLASS12 did not show any diﬀerences between second-generation and ﬁrstgeneration drugs.41 CATIE7 and CUtLASS12 addressed diﬀerent questions with diﬀerent designs. Most previous studies addressed pure eﬃcacy and safety, whereas in the CATIE7 and CUtLASS12 studies the investigators focused on real-world eﬀectiveness. In these studies, broader inclusion criteria were applied and use of more concomitant medication was allowed than in eﬃcacy studies; in the CUtLASS12 study, the doctors could choose from among the diﬀerent ﬁrst-generation and second-generation antipsychotic drugs, and even switch between drug groups. Both study types have strengths and weaknesses. A strength of CATIE7 and CUtLASS12 was the use of comparator drugs that are less potent than haloperidol. Sulpiride, initially used by 50% of the participants in the CUtLASS12 study’s ﬁrst-generation antipsychotic drug group, might induce fewer extrapyramidal side-eﬀects than do other ﬁrst-generation drugs.42 A major limitation of our metaanalysis is that haloperidol was the comparator drug in most of the studies, and the number of studies of mid-potency ﬁrst-generation drugs was insuﬃcient. Results of the CATIE7 and CUtLASS12 studies suggest that mid-potency ﬁrst-generation drugs would have been more appropriate, because they are less likely to cause extrapyramidal side-eﬀects (early work has suggested that perphenazine causes fewer dystonias than does ﬂuphenazine),43 and they are not associated with sedation or weight gain. In our database, we did not note a diﬀerence in the use of antiparkinsonian medication between patients given thiothixene and zotepine in the only available study,44 and in one of two perphenazinecontrolled studies (with high-dose risperidone 5–15 mg).45 In the other perphenazine-controlled study, only a 10% diﬀerence in use of antiparkinsonian medication compared with aripiprazole was noted.46 But to conclude from CATIE and CUtLASS that all antipsychotics are the same and thus to let psychiatrists revert to old bad habits, such as the widespread use of high-dose haloperidol (and not sulpiride or perphenazine) as the primary ﬁrstgeneration antipsychotic drug in many industrialised countries47–49 would not help patients, and there are problems with low-potency ﬁrst-generation drugs as well. The second-generation drugs are expensive, and costeﬀectiveness has not been proven.8,41,50 Public institutions could save costs by funding studies to accurately deﬁne selected old compounds, because they were not rigorously studied at the time they were introduced.51 Because the second-generation antipsychotic drugs diﬀer in many properties, including eﬃcacy, side-eﬀects, cost (some are now generic), and pharmacology (amisulpride is not a serotonin receptor blocker), they do not form a homogeneous class and neither do ﬁrstgeneration antipsychotic drugs. Improper generalisation creates confusion and as a result the classiﬁcation might be abandoned. 40

This meta-analysis provides data that clinicians could use for individualised treatment of patients with schizophrenia based on eﬃcacy, side-eﬀects, and cost of antipsychotic drugs. Contributors SL contributed to designing the study, quality assessment of single-drug studies, data extraction, statistical analysis, and writing of the report. CC contributed to statistical analysis and writing of the report. DA contributed to quality assessment of single-drug studies, data extraction, and writing of the report. RE and JD contributed to designing the study, statistical analysis, and writing of the report. CL contributed to quality assessment of single-drug studies and data extraction. Conﬂict of interest statement SL has received speaker and consultancy honoraria from Sanoﬁ-Aventis, BMS, Lilly, Janssen, Lundbeck, and Pﬁzer. The other authors declare that they have no conﬂict of interest. Acknowledgments A small proportion of JMD’s work was supported by a National Institute of Mental Health grant (1-P01MH68580-01-CFDA-#93.242); the authors did most of the work in their own time.We are greatly indebted to the Cochrane Schizophrenia Group; without access to their register of randomised controlled trials, this review would not have been possible. We thank Köksal Alptekin, Michael Berk, Robert Buchanan, Chen Jindong, Robert Conley, Nancy Covell, Lieuwe de Haan, Eduardo Ponde de Sena, Robin Emsley, Susan Essock, Bernd Gallhofer, Alan Gelenberg, Jes Gerlach, Donald Goﬀ, Adolph Heck, Chen-Jee Hong, Matti Huttunen, Marek Jarema, Jeanette Jerrell, Deanna Kelly, Eckhard Klieser, Li Zhongyi, Nicolai Malykhin, Sung Kil Min, Ann Mortimer, Robert Rosenheck, Ronald See, Bilgen Taneli, Donna Wirshing, and Zhang Hong Yan for sending us further information on their studies; AstraZeneca, Astellas, Bristol-Myers Squibb, Eli Lilly, Lundbeck, and Sanoﬁ-Aventis for providing unpublished data; and Michiaki Taniguchi, Makoto Wada, Toshiyuki Watanabe, and Rong Xie for translating Japanese and Chinese articles. References 1 Rosenheck RA. Eﬀectiveness versus eﬃcacy of second-generation antipsychotics: haloperidol without anticholinergics as a comparator. Psychiatr Serv 2005; 56: 85–92. 2 Geddes J, Freemantle N, Harrison P, Bebbington P. Atypical antipsychotics in the treatment of schizophrenia: systematic overview and meta-regression analysis. BMJ 2000; 321: 1371–76. 3 Davis JM, Chen N, Glick ID. A meta-analysis of the eﬃcacy of second-generation antipsychotics. Arch Gen Psychiatry 2003; 60: 553–64. 4 Adams CE, Coutinho E, Davis JM, et al. Cochrane Schizophrenia Group. The Cochrane Library. Chichester: John Wiley and Sons, 2006. 5 Higgins JPT, Green S. Cochrane Handbook for Systematic Reviews of Interventions 4.2.5. In: The Cochrane Library. Chichester: Wiley and Sons, 2005. 6 Arvanitis LA, Miller BG, Seroquel trial 13 study group. Multiple ﬁxed doses of “Seroquel” (quetiapine) in patients with acute exacerbation of schizophrenia: a comparison with haloperidol and placebo. Biol Psychiatry 1997; 42: 233–46. 7 Lieberman JA, Stroup TS, McEvoy JP, et al. Eﬀectiveness of antipsychotic drugs in patients with chronic schizophrenia. N Engl J Med 2005; 353: 1209–23. 8 Rosenheck RA, Leslie DL, Sindelar J, et al. Cost-eﬀectiveness of second-generation antipsychotics and perphenazine in a randomized trial of treatment of chronic schizophrenia. Am J Psychiatry 2006; 163: 2080–89. 9 Swartz MS, Perkins DO, Stroup TS, et al. Eﬀects of antipsychotic medications on psychosocial functioning in patients with chronic schizophrenia: Findings from the NIMH CATIE Study. Am J Psychiatry 2007; 164: 428–36. 10 Rosenheck R, Cramer J, Xu WC, et al. A comparison of clozapine and haloperidol in hospitalized patients with refractory schizophrenia. N Engl J Med 1997; 337: 809–15.

www.thelancet.com Vol 373 January 3, 2009

Articles

11

12

13

14 15 16

17

18

19

20 21

22

23

24

25

26

27 28

29

30

31

Rosenheck R, Perlick D, Bingham S, et al. Eﬀectiveness and cost of olanzapine and haloperidol in the treatment of schizophrenia— a randomized controlled trial. JAMA 2003; 290: 2693–702. Jones PB, Barnes TRE, Davies L, et al. Randomized controlled trial of the eﬀect on quality of life of second- vs ﬁrst-generation antipsychotic drugs in schizophrenia – cost utility of the latest antipsychotic drugs in schizophrenia study (CUtLASS 1). Arch Gen Psychiatry 2006; 63: 1079–86. Leucht S, Davis JM, Engel RR, Kane JM, Wagenpfeil S. Deﬁning ‘response’ in antipsychotic drug trials: recommendations for the use of scale-derived cutoﬀs. Neuropsychopharmacology 2007; 32: 1903–10. Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003; 327: 557–60. Der-Simonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986; 7: 177–88. Lehman AF, Lieberman JA, Dixon LB, et al. Practice guideline for the treatment of patients with schizophrenia, 2nd edn. Am J Psychiatry 2004; 161: 1–56. Leucht S, Wahlbeck K, Hamann J, Kissling W. New generation antipsychotics versus low-potency conventional antipsychotics: a systematic review and meta-analysis. Lancet 2003; 361: 1581–89. Waraich P, Adams C, Hammill K, Marti J, Roque M. Haloperidol dose for the acutely ill phase of schizophrenia. Cochrane Database Syst Rev 2002; 2: CD001951. Borenstein M, Hedges LV, Higgins JPT, Rothstein H. Comprehensive meta-analysis version 2, 2006. http://www.meta-analysis. com (accessed Nov 24, 2008). Mosholder AD. Review and evaluation of clinical data. Application information. NDA 20-639. Food and Drug Administration 1997. Davis JM, Barter JT, Kane JM. Antipsychotic drugs. In: Kaplan HJ, Saddock BJ, eds. Comprehensive textbook of psychiatry. 5th edn. Baltimore: Williams and Wilkins, 1989: 1591–626. Sernyak M, Rosenheck R. Experience of VA psychiatrists with pharmaceutical detailing of antipsychotic medications. Psychiatr Serv 2007; 58: 1292–96. Berman RM, Marcus RN, Swanink R, et al. The eﬃcacy and safety of aripiprazole as adjunctive therapy in major depressive disorder: A multicenter, randomized, double-blind, placebo-controlled study. J Clin Psychiatry 2007; 68: 843–53. Calabrese JR, Keck PE, Macfadden W, et al. A randomized, double-blind, placebo-controlled trial of quetiapine in the treatment of bipolar I or II depression. Am J Psychiatry 2005; 162: 1351–60. Woodward ND, Purdon SE, Meltzer HY, Zald DH. A meta-analysis of neuropsychological change to clozapine, olanzapine, quetiapine and risperidone in schizophrenia. Int J Neuropsychopharmacol 2005; 8: 1–16. Meltzer HY, Alphs L, Green AI, et al. Clozapine treatment for suicidality in schizophrenia: International Suicide Prevention Trial (InterSePT). Arch Gen Psychiatry 2003; 60: 82–91. Cohen J. Statistical power analysis for the behavioral sciences. 2nd edn. Hillsdale, New Jersey: Lawrence Erlbaum Associates, 1988. Leucht S, Arbter D, Engel RR, Kissling W, Davis JM. How eﬀective are second-generation antipsychotic drugs? A meta-analysis of placebo-controlled trials. Mol Psychiatry 2008; published online Jan 08. DOI:10.1038/sj.mp.4002136. Kemmler G, Hummer M, Widschwendter C, Fleischhacker W. Dropout rates in placebo-controlled and active-control clinical trials of antipsychotic drugs: a meta-analysis. Arch Gen Psychiatry 2005; 62: 1305–12. Schooler N, Rabinowitz J, Davidson M, et al. Risperidone and haloperidol in ﬁrst-episode psychosis: a long-term randomized trial. Am J Psychiatry 2005; 162: 947–53. Zimbroﬀ DL, Kane JM, Tamminga CA, et al. Controlled, dose response study of sertindole and haloperidol in the treatment of schizophrenia. Am J Psychiatry 1997; 154: 782–91.

www.thelancet.com Vol 373 January 3, 2009

32

33

34 35

36

37

38

39

40

41

42 43 44

45

46

47

48

49

50

51

Allison DB, Mentore JL, Heo M, et al. Antipsychotic-induced weight gain: a comprehensive research synthesis. Am J Psychiatry 1999; 156: 1686–96. Leucht S, Pitschel-Walz G, Engel R, Kissling W. Amisulpride— an unusual atypical antipsychotic. A meta-analysis of randomized controlled trials. Am J Psychiatry 2002; 159: 180–90. McEvoy JP, Hogarty GE, Steingard S. Optimal dose of neuroleptic in acute schizophrenia. Arch Gen Psychiatry 1991; 48: 740–45. Van Putten T, Marder SR, Mintz J. A controlled dose comparison of haloperidol in newly admitted schizophrenic patients. Arch Gen Psychiatry 1990; 47: 754–58. WHO. Prophylactic use of anticholinergics in patients on long-term neuroleptic treatment. A consensus statement. World Health Organization heads of centres collaborating in WHO co-ordinated studies on biological aspects of mental illness. Br J Psychiatry 1990; 156: 412–14. Turner EH, Matthews AM, Linardatos E, Tell RA, Rosenthal R. Selective publication of antidepressant trials and its inﬂuence on apparent eﬃcacy. N Engl J Med 2008; 358: 252–60. Heres S, Davis J, Maino K, Jetzinger E, Kissling W, Leucht S. Why olanzapine beats risperidone, risperidone beats quetiapine, and quetiapine beats olanzapine: an exploratory analysis of head-to-head comparison studies of second-generation antipsychotics. Am J Psychiatry 2006; 163: 185–94. McEvoy JP, Lieberman JA, Stroup TS, et al. Eﬀectiveness of clozapine versus olanzapine, quetiapine, and risperidone in patients with chronic schizophrenia who did not respond to prior atypical antipsychotic treatment. Am J Psychiatry 2006; 163: 600–10. Keefe RSE, Bilder RM, Davis SM, et al. Neurocognitive eﬀects of antipsychotic medications in patients with chronic schizophrenia in the CATIE trial. Arch Gen Psychiatry 2007; 64: 633–47. Davies LM, Lewis S, Jones PB, et al. Cost-eﬀectiveness of ﬁrst- v. second-generation antipsychotic drugs: results from a randomised controlled trial in schizophrenia responding poorly to previous therapy. Br J Psychiatry 2007; 191: 14–22. Soares BGO, Fenton M, Chue P. Sulpiride for schizophrenia. Cochrane Database Syst Rev 1999; 1: CD001162. Klein DF, Davis JM. Diagnosis and drug treatment of psychiatric disorders. Baltimore: Williams and Wilkins, 1969. Sarai K, Okada M. Comparison of eﬃcacy of zotepine and thiothixene in schizophrenia in a double-blind study. Pharmacopsychiatry 1987; 20: 38–46. Hoyberg OJ, Fensbo C, Remvig J, Lingjaerde O, Sloth-Nielsen M, Salvesen I. Risperidone versus perphenazine in the treatment of chronic schizophrenic patients with acute exacerbations. Acta Psychiatr Scand 1993; 88: 395–402. Kane JM, Meltzer HY, Carson WH, McQuade RD, Marcus RN, Sanchez R. Aripiprazole for treatment-resistant schizophrenia: results of a multicenter, randomized, double-blind, comparison study versus perphenazine. J Clin Psychiatry 2007; 68: 213–23. Lohse MJ, Lorenzen A, Müller-Oerlinghausen B. Psychopharmaka. In: Schwabe U, Pfaﬀrath D, eds. Arzneimittelverordnungsreport. Heidelberg: Springer, 2005: 820–64. Paton C, Lelliott P, Harrington M, Okocha C, Sensky T, Duﬀett R. Patterns of antipsychotic and anticholinergic prescribing for hospital inpatients. J Psychopharmacol 2003; 17: 223–29. Kaye JA, Bradbury BD, Jick H. Changes in antipsychotic drug prescribing by general practitioners in the United Kingdom from 1991 to 2000: a population-based observational study. Br J Clin Pharmacol 2003; 56: 569–75. Polsky D, Doshi JA, Bauer MS, Glick HA. Clinical trial-based cost-eﬀectiveness analyses of antipsychotic use. Am J Psychiatry 2006; 163: 2047–56. Hartung B, Wada M, Laux G, Leucht S. Perphenazine for schizophrenia. Cochrane Database Syst Rev 2005; 1: CD003443.

41

Articles

Outpatient management of patients with low-risk upper-gastrointestinal haemorrhage: multicentre validation and prospective evaluation A J Stanley, D Ashley, H R Dalton, C Mowat, D R Gaya, E Thompson, U Warshow, M Groome, A Cahill, G Benson, O Blatchford, W Murray

Summary Lancet 2009; 373: 42–47 Published Online December 15, 2008 DOI:10.1016/S01406736(08)61769-9 See Comment page 5 Gastrointestinal Units, Glasgow Royal Inﬁrmary, Glasgow, UK (A J Stanley MD, D R Gaya MBChB, A Cahill MB, G Benson BSc, W Murray MD); University Hospital of NorthTees, Stockton, UK (D Ashley BMBS, E Thompson MBChB); Royal Cornwall Hospital, Truro, UK (H R Dalton DPhil, U Warshow MBBS); Ninewells Hospital, Dundee, UK (C Mowat MD, M Groome MBChB); and Public Health Department, Greater Glasgow & Clyde, Glasgow, UK (O Blatchford MD) Correspondence to: Dr A J Stanley, Glasgow Royal Inﬁrmary, Department of Gastroenterology, Wards 8 and 9, Castle Street, Glasgow G4 OSF, UK [email protected]

Background Upper-gastrointestinal haemorrhage is a frequent reason for hospital admission. Although most risk scoring systems for this disorder incorporate endoscopic ﬁndings, the Glasgow-Blatchford bleeding score (GBS) is based on simple clinical and laboratory variables; a score of 0 identiﬁes low-risk patients who might be suitable for outpatient management. We aimed to evaluate the GBS then assess the eﬀect of a protocol based on this score for non-admission of low-risk individuals. Methods Our study was undertaken at four hospitals in the UK. We calculated GBS and admission (pre-endoscopy) and full (post-endoscopy) Rockall scores for consecutive patients presenting with upper-gastrointestinal haemorrhage. With receiver-operating characteristic (ROC) curves, we compared the ability of these scores to predict either need for clinical intervention or death. We then prospectively assessed at two hospitals the introduction of GBS scoring to avoid admission of low-risk patients. Findings Of 676 people presenting with upper-gastrointestinal haemorrhage, we identiﬁed 105 (16%) who scored 0 on the GBS. For prediction of need for intervention or death, GBS (area under ROC curve 0·90 [95% CI 0·88–0·93]) was superior to full Rockall score (0·81 [0·77–0·84]), which in turn was better than the admission Rockall score (0·70 [0·65–0·75]). When introduced into clinical practice, 123 patients (22%) with upper-gastrointestinal haemorrhage were classiﬁed as low risk, of whom 84 (68%) were managed as outpatients without adverse events. The proportion of individuals with this condition admitted to hospital also fell (96% to 71%, p<0·00001). Interpretation The GBS identiﬁes many patients presenting to general hospitals with upper-gastrointestinal haemorrhage who can be managed safely as outpatients. This score reduces admissions for this condition, allowing more appropriate use of in-patient resources. Funding None.

Introduction Upper-gastrointestinal haemorrhage is a frequent cause of acute admission to hospital, with an incidence in the UK of 103–172 per 100 000 adults per year.1,2 The severity of the disorder varies from mild coﬀee-ground vomiting to exsanguination. However, most patients do not need endoscopic treatment, surgery, or blood transfusion and do not rebleed or die.1,3 Individuals presenting with upper-gastrointestinal haemorrhage have traditionally been admitted for a period of observation, with or without endoscopy. Admission and endoscopy on the next available list is recommended in the 2002 British Society of Gastroenterology guideline for people with mild-tomoderate upper-gastrointestinal haemorrhage,4 although very low-risk young people with a minor bleed and without haemodynamic compromise can be discharged without endoscopy. We know from our experience and in other hospitals that some clinicians use their judgment informally to avoid admittance of individuals they view as being at low risk. However, objective identiﬁcation of such patients with clinical conﬁdence is sometimes diﬃcult.

42

Several risk assessment and scoring systems for uppergastrointestinal haemorrhage have been developed in an attempt to stratify risk for poor outcome.2,5–12 However, most, including the widely used Rockall score,3 include endoscopic ﬁndings; therefore, many patients are kept in hospital until this procedure is undertaken. Although many hospitals in the UK have an emergency endoscopy rota, this facility is usually for individuals with major haemorrhage only, with others waiting until the next day or longer for a semi-elective procedure. Furthermore, nonemergency endoscopy is unavailable at weekends in many hospitals. An abbreviated pre-endoscopy admission Rockall score, which excludes endoscopic ﬁndings, is sometimes used, but this measure has not been fully validated.3 In a previous report from Glasgow, UK, logistic regression was used to derive the Glasgow-Blatchford bleeding score (GBS; table 1), which is used to predict either a patient’s need for hospital-based intervention (blood transfusion, endoscopic treatment, or surgery) or death.5 The score was derived from data of 1748 people presenting with upper-gastrointestinal haemorrhage but was only validated locally in a few aﬀected individuals presenting to three Glasgow hospitals, not including the www.thelancet.com Vol 373 January 3, 2009

Articles

Glasgow Royal Inﬁrmary. It is based on simple variables from a patient’s history, examination, and laboratory results. A GBS score of 0 fulﬁls low-risk criteria (panel), which seems to identify people at very low (0·5%) risk of needing intervention, as described above.5 The aim of our study was to assess and externally validate the GBS in four large general hospitals in Scotland and England. We also prospectively looked at the eﬀect of the introduction of GBS low-risk criteria on accident and emergency (A&E) departments, with the intention to avoid admission for patients assessed as low risk.

Score value Blood urea (mmol/L) 6·5–7·9

2

8·0–9·9

3

10·0–25·0

4

>25·0

6

Haemoglobin for men (g/L) 120–129

1

100–119

3

<100

6

Haemoglobin for women (g/L)

Methods

100–119

1

Data collection

<100

6

We divided our study into two phases. In phase one, we obtained data prospectively from consecutive patients presenting with upper-gastrointestinal haemorrhage over a 12-month period at Royal Cornwall Hospital, Truro, for 6 months at Glasgow Royal Inﬁrmary, Glasgow, and over 3 months at Ninewells Hospital, Dundee, and retrospectively for 3 months at University Hospital of North-Tees, Stockton. We deﬁned upper-gastrointestinal haemorrhage as haematemesis, coﬀee-ground vomit, or melaena. We excluded inpatients with the disorder; nasogastric lavage was not undertaken routinely. A speciﬁc junior doctor or research nurse at every site obtained data, which included patients’ characteristics, any history of melaena, syncope, cardiac failure, or liver disease, haemodynamic and laboratory variables, endoscopic ﬁndings (if undertaken), and length of inpatient stay. They also recorded outcome data in the form of interventions (blood transfusion, endoscopic treatment, or surgery) or death. In phase two of our study, we used GBS low-risk criteria (GBS=0) in A&E departments at Glasgow and Stockton to identify patients with upper-gastrointestinal haemorrhage for whom admission could be avoided. We did not admit individuals meeting these criteria unless necessary for other reasons. All Glasgow patients who were not admitted were oﬀered outpatient endoscopy, as were those older than 50 years in Stockton (or younger patients at the discretion of the clinician). We followed up aﬀected individuals who failed to attend for endoscopy either at a clinic or by discussion with their family doctor at least 6 months later, in conjunction with case-note review. We obtained phase two data prospectively in consecutive patients presenting to A&E departments for 1 year at Glasgow and for 3 months at Stockton. We assessed outcomes and compared admission numbers and inpatient stay for low-risk people between phase one and phase two for these two centres. All described analyses were prespeciﬁed. Each hospital viewed this assessment as an evaluation of service delivery rather than research, since non-admission of low-risk patients is not a novel practice and no additional data were gathered. In particular, no allocation to intervention groups took place and randomisation was not done. Therefore, we did not

Systolic blood pressure (mm Hg)

www.thelancet.com Vol 373 January 3, 2009

100–109

1

90–99

2

<90

3

Other markers Pulse ≥100/min

1

Presentation with melaena

1

Presentation with syncope

2

Hepatic disease*

2

Cardiac failure†

2

*Known history, or clinical and laboratory evidence, of chronic or acute liver disease. †Known history, or clinical and echocardiographic evidence, of cardiac failure.

Table 1: Admission risk markers for GBS5

need to obtain ethics approval or informed consent. Our report follows STROBE guidelines. Our data have been presented in part and published as abstracts.

Statistical analysis We used the SPSS statistical package for data analysis (version 16 for Windows). Data are presented as median values with IQRs, unless otherwise stated. When necessary, we calculated exact Poisson CIs. We compared the GBS with admission (pre-endoscopy) and full (postendoscopy) Rockall scores to predict intervention or death, by calculation of areas under receiver-operator characteristic (ROC) curves and 95% CIs. We used the Mann-Whitney U test and χ² test to compare medians and proportions, respectively.

Role of the funding source No funding was received for this study. All doctors had access to their local hospital data, and AJS and OB had

Panel: Low-risk criteria of GBS • • • • •

Urea <6·5 mmol/L Haemoglobin ≥130 g/L (men) or ≥120 g/L (women) Systolic blood pressure ≥110 mm Hg Pulse <100 beats per min Absence of melaena, syncope, cardiac failure, or liver disease

43

Articles

full access to all combined data in the study. The ﬁnal decision to submit the manuscript for publication was made by AJS, DA, HRD, CM, OB, and WM.

Patients (n=676) Site Glasgow

211 (31%)

Truro

232 (34%)

Results

Stockton

123 (18%)

Dundee

110 (16%)

From the four study centres, a total of 676 patients were included in phase one. Table 2 outlines demographic characteristics and outcomes for these people. 19 individuals had data missing for measurement of admission Rockall score and 27 had omissions for GBS. Of those with complete data, GBS was 0 (low-risk criteria met) in 105 (16%) and admission Rockall score was 0 in 184 (28%). The GBS low-risk group consisted of 27 people (12%) from Truro, 17 (17%) from Stockton, 36 (17%) from Glasgow, and 25 (23%) from Dundee. Median age of patients in the low-risk group was signiﬁcantly lower than that of the other individuals with complete data (41 [IQR 28–55] vs 64 [48–78] years; p<0·0001). Of the 105 low-risk patients, 22 (21%) were older than 60 years and 14 (13%) were older than 70 years. No interventions and no deaths were recorded in the low-risk group identiﬁed by a GBS of 0. However, one death and 44 interventions (21 endoscopic or surgical and 23 transfusions) were noted for 32 (17%) people with an admission Rockall score of 0. Figure 1 shows interventions or death for admission Rockall score and GBS. By ROC curve comparison of the 647 patients with full data for both scores, GBS was superior to admission Rockall score for prediction of intervention or death (area under the curve 0·92 [95% CI 0·90–0·94] vs 0·72 [0·68–0·76]; ﬁgure 2). Table 3 shows endoscopic ﬁndings for 485 patients who underwent the procedure in phase one. 467 of these had complete data available for measurement of full and admission Rockall scores and GBS. By ROC curve analysis, the GBS was superior to full Rockall score for prediction of intervention or death (area under the curve 0·90 [95% CI 0·88–0·93] vs 0·81 [0·77–0·84]), which was in turn superior to admission Rockall score (0·70 [0·65–0·75]; ﬁgure 3). Table 4 presents individual data for the four study centres. In phase two, GBS low-risk criteria (GBS=0) were used to assess 491 consecutive patients presenting to A&E at Glasgow and 81 at Stockton. Overall, 123 (22%) individuals were identiﬁed as low risk, with 84 (68%) of this group not admitted (table 5). Low-risk patients not admitted were younger than those who were (median age 30 [IQR 21–42] vs 37 [30–55] years; p=0·005). Only 23 (40%) people oﬀered outpatient endoscopy attended for their planned procedure. Endoscopic ﬁndings showed no malignant disease, varices, or ulcers and no need for intervention in any patient. One individual died from disseminated (non-upper gastrointestinal) malignant disease 2 months after endoscopy had indicated gastritis only. Of the low-risk group who failed to attend for endoscopy, case-note review and consultation with the patient and family doctor clariﬁed that none had been readmitted with upper-gastrointestinal haemorrhage or

Age (years; median [IQR])*

62 (43–76)

Sex† Men

416 (62%)

Women

256 (38%)

Outcomes Endoscopic or surgical procedure

137 (20%)

Blood transfusion

175 (26%)

Hospital stay (days; median [IQR])

4 (1–7)

In-hospital mortality

30 (4%)

Data are number of patients (%), unless otherwise stated. *Age unknown for one patient. †Sex not recorded for four patients.

Table 2: Patients’ demographics and outcomes in phase one

200

Patients not needing intervention and alive Patients needing interventions or died

180

Number of patients

160 140 120 100 80 60 40 20 0 0

1

2 3 4 5 Rockall admission (pre-endoscopy) score

6

7

120

Number of patients

100

80

60

40

20

0 0

1

2

3

4

5

6

7

8

9 10 GBS

11

12

13

14

Figure 1: Need for intervention or death by score for all four centres in phase one

44

15

16

17

18

19

www.thelancet.com Vol 373 January 3, 2009

Articles

haemorrhage was 2 days for both phases (phase one IQR 1–6, phase two 0–7; p=0·2). Mean bed-days per patient presenting with upper-gastrointestinal haemorrhage in phases one and two were 6·2 (SD 11·8) and 5·0 (7·6), respectively.

1 0·9 0·8

Discussion 0·6 0·5 0·4 0·3 0·2 GBS Rockall admission (pre-endoscopy) score

0·1 0 0

0·2

0·4

0·6 1-speciﬁcity

0·8

1

Figure 2: ROC curve comparison of GBS and admission Rockall score for prediction of need for intervention or death for all four centres in phase one (n=647)

had died after a minimum of 6 months follow-up. Therefore, of the 123 patients meeting low-risk criteria (GBS=0), none (95% CI 0–3%) needed any intervention related to their disorder. By comparison of data from phase one and two, a reduction was noted in the proportion of patients presenting with upper-gastrointestinal haemorrhage who were admitted, from 319 (96%) to 405 (71%, p<0·0001; table 5). Median hospital stay for admitted patients rose between the two phases, from 3 (IQR 1–6) to 4 (2–9) days (p<0·0001), although the median hospital stay for all patients who presented with upper-gastrointestinal GBS=0 (n=66)

GBS>0 (n=419)

Normal/hiatus hernia

37 (56%)

100 (24%)

Oesophagitis

12 (18%)

73 (17%)

Gastritis

6 (9%)

70 (17%)

Duodenitis

9 (14%)

33 (8%)

Mallory-Weiss tear

3 (5%)

17 (4%)

Barrett’s oesophagus

2 (3%)

11 (3%)

Dieulafoy’s erosion

0

2 (<1%)

Duodenal ulcer

0

67 (16%)

Gastric ulcer

0

41 (10%)

Varices

0

30 (7%)

Arteriovenous malformation

0

10 (2%)

Upper-gastrointestinal cancer

0

19 (5%)

Other

1 (2%)*

11 (3%)†

Data are number of ﬁndings (% patients). Some patients had more than one endoscopic ﬁnding. *Oesophageal diverticulum. †Oesophageal candidiasis (n=3) and one each of: herpes oesophagitis, oesophageal diverticulum, duodenal diverticulum, gastric polyp, Schatzki’s ring, bleeding vessel at surgical gastrojejunostomy anastomosis, intraperitoneal bleed, oesophagogastric junction erosion.

Table 3: Endoscopic ﬁndings in phase one, by GBS

www.thelancet.com Vol 373 January 3, 2009

Our ﬁndings show that simple GBS low-risk criteria can identify a signiﬁcant proportion of individuals presenting with upper-gastrointestinal haemorrhage who are suitable for outpatient management. Furthermore, use of these criteria in A&E departments leads to a reduction in admissions for this disorder, with no apparent deleterious eﬀects on patients’ care. Although most scoring systems for upper-gastrointestinal haemorrhage incorporate endoscopic ﬁndings, outcomes of an audit by the British Society of Gastroenterology indicated that only 50% of people have endoscopy within 24 h.13 Workers on the audit also reported that only 55% of hospitals have a consultant on-call rota for out-of-hours endoscopy.14 Many patients with upper-gastrointestinal haemorrhage are admitted under general doctors who might feel uncomfortable about discharging them without endoscopy (or further observation). Indeed, in phase one of our study, many individuals were admitted for observation without inpatient endoscopy. A validated non-endoscopic scoring system to risk-stratify these patients could allow triage of low-risk individuals to outpatient management on attendance at A&E departments. In our study, GBS low-risk criteria identiﬁed more than 15% of patients presenting with upper-gastrointestinal haemorrhage in whom outpatient management seems safe. This proportion is similar to that reported in the UK of individuals meeting US endoscopic-based criteria for outpatient management.7,15 The variation across our study 1 0·9 0·8 0·7 Sensitivity

Sensitivity

0·7

0·6 0·5 0·4 0·3 0·2

GBS Rockall admission (pre-endoscopy) score Rockall full (post-endoscopy) score

0·1 0 0

0·2

0·4

0·6 1-speciﬁcity

0·8

1

Figure 3: ROC curve comparison of GBS and full and admission Rockall scores for prediction of need for intervention or death for all four centres in phase one (n=467)

45

Articles

Total patients

Interventions or deaths

Area under ROC curve (95% CI)

Glasgow GBS

208

50

Admission Rockall

208

50

0·67 (0·58–0·76)

79

44

0·68 (0·56–0·80)

Full Rockall

0·95 (0·92–0·98)

Stockton GBS

101

33

0·85 (0·78–0·93)

Admission Rockall

108

34

0·71 (0·60–0·81)

74

27

0·69 (0·56–0·81)

GBS

108

44

0·96 (0·93–0·99)

Admission Rockall

109

45

0·79 (0·70–0·88)

89

35

0·96 (0·92–0·99)

Full Rockall Dundee

Full Rockall Truro GBS

232

99

0·91 (0·87–0·94)

Admission Rockall

232

99

0·73 (0·66–0·80)

Full Rockall

232

99

0·83 (0·78–0·88)

GBS

649

226

0·92 (0·90–0·94)

Admission Rockall

657

228

0·72 (0·68–0·76)

Full Rockall

474

205

0·80 (0·76–0·84)

All sites

Data are number of patients, unless otherwise stated.

Table 4: Comparison of GBS and admission and full Rockall scores for prediction of intervention or death across the four study sites

sites in the proportion of people having a GBS score of 0 could indicate local population and referral diﬀerences. Our results showed the GBS to be superior to both full and admission Rockall scores for prediction of need for blood transfusion, endoscopic treatment, or surgery, or death. We accept that comparison between GBS and full Rockall score is restricted to individuals who underwent endoscopy, excluding several who probably had fairly minor bleeds that the clinician judged did not need the procedure. However, the area under the ROC curve for GBS fell when we excluded this group of patients, suggesting that this comparison could underestimate this scoring system. In a report from Taiwan, researchers also noted the GBS to be superior to both Rockall scores for prediction of patients with high-risk upper-gastrointestinal haemorrhage.16 Before introduction (phase one; n=334) Age (years; median [IQR])

54 (37–72)

Low-risk patients (GBS=0)

53 (16%)

After introduction (phase two; n=572) 52 (35–68) 123 (22%)

Interventions in low-risk group

0

Low-risk patients not admitted

3 (6%)

84 (68%)*

15 (4%)

167 (29%)*

Total number with upper-gastrointestinal haemorrhage not admitted

0

Data are number of patients (%), unless otherwise stated. *p<0·0001.

Table 5: Comparison of data before and after introduction of GBS low-risk criteria into clinical practice

46

In our study, an admission Rockall score of 0 indicated more people presenting with upper-gastrointestinal haemorrhage than did a GBS of 0. However, 17% of patients identiﬁed with the admission Rockall score needed hospital-based intervention compared with none with the GBS, and one patient judged low risk with the admission Rockall score died. On clinical introduction of GBS low-risk criteria in the second phase of our study, we were able to avoid admission of most patients who met the criteria (GBS=0). A few lowrisk individuals were admitted for other reasons, including alcohol withdrawal and poor social circumstances. A limitation of our study is that many people did not attend for planned outpatient endoscopy. However, none of the low-risk patients with upper-gastrointestinal haemorrhage who were not admitted needed any relevant intervention and no upper-gastrointestinal malignant disease was detected on follow-up. Introduction of these low-risk criteria led to a signiﬁcant reduction in the proportion of patients presenting with upper-gastrointestinal haemorrhage who were admitted. The rise in the length of stay for patients actually admitted with the disorder is probably accounted for by the reduction in admission numbers of those with minor bleeds. Although we accept that the change in length of stay was not signiﬁcant in these skewed data, the reduction in average bed-days per patient of 1·2 is perhaps more meaningful. Taking into account 2005 population data from the UK Oﬃce for National Statistics for individuals older than 15 years and the reported incidence of upper-gastrointestinal haemorrhage in the UK, this reduction could be between 60 000 and 100 000 bed-days per year if these results were replicated in all UK hospitals.1,2 However, we acknowledge this is speculative and variations exist in incidence and management of upper-gastrointestinal haemorrhage across the UK. Unlike most other risk scores, age is not a component in the GBS. Stepwise logistic regression had previously conﬁrmed that age was not an important predictor of need for intervention after other variables were taken into account.5 A ﬁfth of patients who met GBS low-risk criteria in the four centres were older than 60 years. Researchers on a large Canadian study reported that age was not an independent predictor of rebleeding.17 They also noted that a modiﬁed GBS (because of non-recording of syncope or serum urea concentrations) was strongly associated with rebleeding, death, and endoscopic stigmata of bleeding, and was superior to the admission Rockall in prediction of these outcomes. Cameron and colleagues described another nonendoscopic risk stratiﬁcation with 14 clinical and laboratory variables.9 However, this complex score identiﬁed only 6% of patients with upper-gastrointestinal haemorrhage as low risk who might be suitable for outpatient management. Workers on an American study assessed an artiﬁcial neural network for prediction of www.thelancet.com Vol 373 January 3, 2009

Articles

endoscopic ﬁndings and need for endoscopic treatment in individuals with upper-gastrointestinal haemorrhage.18 Their network was superior to the admission Rockall score and similar to the full Rockall score. However, it required input of 27 patient’s variables and computer software was needed for analysis. Furthermore, it was not assessed in a truly unselected cohort presenting with the disorder. However, future studies to compare this network with the GBS would be useful. Although many risk models for upper-gastrointestinal haemorrhage use rebleeding or death as endpoints, need for hospital-based intervention seems a logical way to assess this disorder in the era of increased outpatient management. Moreover, costs associated with management of upper-gastrointestinal haemorrhage are mostly for hospital admission.19 Risk-stratiﬁcation of these patients is analogous to that which already takes place for other frequent medical disorders, including deep venous thrombosis, chest infection, and chest pain.20–22 Groups from Los Angeles, USA, and Tokyo, Japan, have suggested a 100% negative predictive value for rebleeding or death, and need for intervention, respectively, with GBS low-risk criteria.23,24 The GBS is based on simple clinical and laboratory variables; therefore, aﬀected individuals can be assessed quickly in A&E departments or at a clinical decision unit. We suggest that further assessment of the GBS as a method to identify low-risk people for outpatient management is undertaken in diﬀerent populations, for whom both incidence and pathology could have some eﬀect. Contributors The study was designed by AJS, DA, HRD, CM, OB, and WM. Data were obtained by DRG, AC, ET, UW, MG, and GB. Data were analysed by OB and AJS. The report was written by AJS, and all authors approved the ﬁnal version. Conﬂict of interest statement We declare that we have no conﬂict of interest. References 1 Rockall TA, Logan RFA, Devlin HB, Northﬁeld TC. Incidence of and mortality from acute upper gastrointestinal haemorrhage in the UK. BMJ 1995; 311: 222–26. 2 Blatchford O, Davidson LA, Murray WR, Blatchford M, Pell J. Acute upper gastrointestinal haemorrhage in west of Scotland: case ascertainment study. BMJ 1997; 315: 510–14. 3 Rockall TA, Logan RFA, Devlin HB, Northﬁeld TC. Risk assessment after acute upper gastrointestinal haemorrhage. Gut 1996; 38: 316–21. 4 British Society of Gastroenterology endoscopic committee. Non-variceal upper gastrointestinal haemorrhage: guidelines. Gut 2002; 51 (suppl 4): iv1–6. 5 Blatchford O, Murray WR, Blatchford M. A risk score to predict need for treatment for upper gastrointestinal haemorrhage. Lancet 2000; 356: 1318–21.

www.thelancet.com Vol 373 January 3, 2009

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

Cipolletta L, Bianco MA, Rotondano G, Marmo R, Piscopo R. Outpatient management for low-risk nonvariceal upper GI bleeding: a randomized controlled trial. Gastrointest Endosc 2002; 55: 1–5. Longstreth GF, Feitelberg SP. Outpatient care of selected patients with acute non-variceal upper gastrointestinal haemorrhage. Lancet 1995; 345: 108–11. Longstreth GF, Feitelberg SP. Successful outpatient management of acute upper gastrointestinal hemorrhage: use of practice guidelines in a large patient series. Gastrointest Endosc 1998; 47: 219–22. Cameron EA, Pratap JN, Sims TJ, et al. Three year prospective validation of a pre-endoscopic risk stratiﬁcation in patients with acute upper-gastrointestinal haemorrhage. Eur J Gastro Hepatol 2002; 14: 497–501. Hay JA, Maldonado L, Weingarten SR, Ellrodt AG. Prospective evaluation of a clinical guideline recommending hospital length of stay in upper gastrointestinal tract hemorrhage. JAMA 1997; 278: 2151–56. Imperiale TF, Dominitz JA, Provencale DT, et al. Predicting poor outcome from acute upper gastrointestinal hemorrhage. Arch Intern Med 2007; 167: 1291–96. Saeed ZA, Winchester CB, Michaeletz PA, Woods KL, Graham DY. A scoring system to predict rebleeding after endoscopic therapy of non-variceal upper gastrointestinal hemorrhage. Am J Gastroenterol 1993; 88: 1842–49. Hearnshaw SA, Lowe D, Logan RF, Murphy MF, Travis SPL, Palmer KR. Preliminary results of the UK comparative audit of over 6500 patients with acute upper gastrointestinal bleeding. Gut 2008; 57 (suppl I): A12. Hearnshaw SA, Lowe D, Logan RF, Murphy MF, Travis SPL, Palmer KR. Preliminary results of the UK comparative audit of acute upper gastrointestinal bleeding. Gut 2008; 57 (suppl I): A8. Packham CJ, Rockall TA, Logan RFA. Outpatient care for selected patients with acute upper gastrointestinal bleeding. Lancet 1995; 345: 659–60. Chen IC, Hung MS, Chiu TF, Chen JC, Hsiao CT. Risk scoring systems to predict need for clinical intervention for patients with non-variceal upper gastrointestinal tract bleeding. Am J Emerg Med 2007; 25: 774–79. Romagnuolo J, Barkun AN, Enns R, Armstrong D, Gregor J. Simple clinical predictors may obviate urgent endoscopy in selected patients with nonvariceal upper gastrointestinal tract bleeding. Arch Intern Med 2007; 167: 265–70. Das A, Ben-Menachem T, Farooq FT, et al. Artiﬁcial neural network as a predictive instrument in patients with acute non-variceal upper gastrointestinal hemorrhage. Gastroenterology 2008; 134: 65–74. Gralnek IM. Outpatient management of “low-risk” nonvariceal upper GI hemorrhage: are we ready to put evidence into practice? Gastrointest Endosc 2002; 55: 131–34. Baron RM, Goldhaber SZ. Deep venous thrombosis: early discharge strategies and outpatient management. J Throm Thrombolysis 1999; 7: 113–22. Lim WS, van der Eerden MM, Laing R, et al. Deﬁning community acquired pneumonia severity on presentation to hospital: an international derivation and validation study. Thorax 2003; 58: 377–82. Antman EM, Cohen M, Bernink PJ, et al. The TIMI risk score for unstable angina/non-ST elevation MI: a method for prognostication and therapeutic decision making. JAMA 2000; 284: 835–42. Gralnek IM, Dulai GS. Incremental value of upper endoscopy for triage of patients with acute non-variceal upper-GI hemorrhage. Gastrointest Endosc 2004; 60: 9–14. Masaoka T, Suzuki H, Hori S, Aikawa N, Hibi T. Blatchford scoring system is a useful scoring system for detecting patients with upper gastrointestinal bleeding who do not need endoscopic intervention. J Gastro Hepatol 2007; 22: 1404–08.

47

Articles

Universal voluntary HIV testing with immediate antiretroviral therapy as a strategy for elimination of HIV transmission: a mathematical model Reuben M Granich, Charles F Gilks, Christopher Dye, Kevin M De Cock, Brian G Williams

Summary Lancet 2009; 373: 48–57 Published Online November 26, 2008 DOI:10.1016/S01406736(08)61697-9 See Comment pages 7 and 9 Department of HIV/AIDS (R M Granich MD, Prof C F Gilks DPhil, Prof K M De Cock MD) and Stop TB Department (Prof C Dye DPhil, B G Williams PhD), WHO, Geneva, Switzerland Correspondence to: Reuben Granich, Antiretroviral Treatment and HIV Care Unit, Department of HIV/AIDS, WHO, Avenue Appia 20, CH-1211, Geneva 27, Switzerland [email protected]

Background Roughly 3 million people worldwide were receiving antiretroviral therapy (ART) at the end of 2007, but an estimated 6·7 million were still in need of treatment and a further 2·7 million became infected with HIV in 2007. Prevention eﬀorts might reduce HIV incidence but are unlikely to eliminate this disease. We investigated a theoretical strategy of universal voluntary HIV testing and immediate treatment with ART, and examined the conditions under which the HIV epidemic could be driven towards elimination. Methods We used mathematical models to explore the eﬀect on the case reproduction number (stochastic model) and long-term dynamics of the HIV epidemic (deterministic transmission model) of testing all people in our test-case community (aged 15 years and older) for HIV every year and starting people on ART immediately after they are diagnosed HIV positive. We used data from South Africa as the test case for a generalised epidemic, and assumed that all HIV transmission was heterosexual. Findings The studied strategy could greatly accelerate the transition from the present endemic phase, in which most adults living with HIV are not receiving ART, to an elimination phase, in which most are on ART, within 5 years. It could reduce HIV incidence and mortality to less than one case per 1000 people per year by 2016, or within 10 years of full implementation of the strategy, and reduce the prevalence of HIV to less than 1% within 50 years. We estimate that in 2032, the yearly cost of the present strategy and the theoretical strategy would both be US$1·7 billion; however, after this time, the cost of the present strategy would continue to increase whereas that of the theoretical strategy would decrease. Interpretation Universal voluntary HIV testing and immediate ART, combined with present prevention approaches, could have a major eﬀect on severe generalised HIV/AIDS epidemics. This approach merits further mathematical modelling, research, and broad consultation. Funding None.

Introduction 25 years after the discovery of HIV,1 control of the HIV epidemic remains elusive and some have called for a re-examination of the approach to control this virus.2 Development of an eﬀective HIV-1 vaccine remains a remote possibility, and trials of vaginal microbicides have not shown any protective beneﬁt.3,4 Where HIV transmission is mainly heterosexual, male circumcision can reduce adult heterosexual HIV transmission, but only by about 40% at the overall population level.5,6 A call has been made to focus prevention interventions on high-risk populations7 and to expand programmes for female sex workers and their clients,8 and injecting drug users.9 Few people are aware of their HIV status, and rapid expansion of voluntary HIV testing and counselling has been recommended by WHO.10 About 3 million people worldwide had been given antiretroviral therapy (ART) at the end of 200710 but an estimated 6·7 million were still in need of it and a further 2·7 million were infected with HIV in 2007.10–12 At present there is inadequate evidence for WHO to provide guidance on the role of ART for people living 48

with HIV as a strategy to prevent further sexual transmission. To control the HIV/AIDS epidemic, infectious individuals would have to be rendered noninfectious, or susceptible people protected from infection. Vertical transmission of HIV can be eliminated by testing of mothers and blocking of transmission through the use of antiretroviral drugs, accompanied by elective caesarean section and the use of replacement infant feeding.13 Although increasing emphasis is being placed on positive prevention14,15 and provider-initiated HIV testing and counselling,16 no large-scale studies have been undertaken of the eﬀect of diagnosing all HIV-positive people early and treating them immediately. Present guidelines suggest that ART should be started when infected people reach speciﬁc immunological or clinically-deﬁned stages of disease to keep subsequent morbidity and mortality in individual patients to a minimum.17 Wherever ART has been implemented it has had a substantial and rapid eﬀect on survival for individuals and within populations.18 The eﬀect of treatment on transmission and the possible public-health beneﬁts have, with some exceptions, received less www.thelancet.com Vol 373 January 3, 2009

A

B

0·15

1·0 0·8 Proportion surviving

Proportion of HIV-negative men

Articles

0·10

0·05

0·4 0·2

0

0 500

0

1000 1500 CD4+ cells per μL

2000

0

2500

C

5

10 15 Time since infection (years)

20

25

D

10

1·0

Final phase c

d

6 4

e Chronic phase a

2

Cumulative infections

b Acute phase

8 Relative infectivity

0·6

0·8 0·6 0·4 0·2 0

0 0

2

4 6 Time since infection (years)

8

10

0

5

10 15 Time since infection (years)

20

Figure 1: Theoretical basis for the stochastic and deterministic transmission models (A) The observed distribution of CD4+ cells in HIV-negative men in South Africa.29,30 Error bars indicate 95% CIs. (B) The Weibull survival curve, with HIV infection used in the model on the basis of data from the CASCADE study.30 (C) Schematic diagram of the change in infectivity with time in a person who survives for 10 years. The infectivity during all three phases (a−c) can be varied, as can the duration of the acute and ﬁnal phases (d and e). (D) The cumulative distribution of the time between the infection of the index case and the infection of a secondary case for the baseline parameters as calculated from the model. The initial rapid increase (from 0% to 10%) is the result of the relatively high infectivity during the acute phase.

attention.19–22 The use of ART can reduce the plasma viral load by up to six orders of magnitude,23 and several investigators have assessed the eﬀect of ART on transmission.19,20,24 However, the high viral load during the acute phase of infection, the long duration of infectiousness, the present policy of limiting costly and potentially toxic ART to people whose immune systems are severely compromised, and low coverage can reduce the extent to which the use of ART reduces transmission. Despite substantial eﬀorts to expand access to voluntary HIV testing, nearly 80% of HIV-infected adults in sub-Saharan Africa are unaware of their status and more than 90% do not know whether their partners are infected with HIV.10,25 Present approaches to HIV testing, prevention, and treatment are unlikely to bring about a rapid reduction in HIV incidence, and demand for treatment in countries that are most heavily aﬀected will continue to grow. Reduction in HIV incidence, with the goal of eventual elimination, would require that the case reproduction number R0—the number of secondary infections resulting from one primary infection in an otherwise susceptible population—is reduced to and kept below 1.26 A potential shift in strategy is to diagnose all HIV-infected people as soon as possible after infection www.thelancet.com Vol 373 January 3, 2009

and provide them with immediate ART. In considering the use of ART to eliminate transmission, we focused on two questions: how often would people have to be tested and how soon after testing positive should they start ART? In this hypothetical modelling exercise, we examined a strategy of universal voluntary HIV testing and immediate treatment with ART in the context of a generalised heterosexual epidemic of the same intensity as in southern Africa, and examined the conditions under which the epidemic could be driven towards elimination. The results have potential implications for HIV prevention that require broad consultation.

Methods Study design To establish a hypothetical HIV epidemic, we relied on available data from South Africa as the test case for a generalised HIV epidemic, representing 17% of all people living with HIV.11 Our hypothetical test case assumed, as in South Africa, that almost all transmission was heterosexual rather than homosexual. This assumption holds true for South Africa and similar settings, and was supported by the observation that the prevalence of HIV in South African men is estimated to be 58% of the 49

Articles

βN

λ=λ oe–αPn P = NI I =Σi (Ii +Ai) J =Σi (Ii +ε Ai) μ μ μ μ μ λ SJ/N ρ ρ ρ ρ S I1 I2 I3 I4 D τ φ τ φ τ φ τ φ A1 σ A2 σ A3 σ A4 σ D μ

For more on both models see http://www.who.int/hiv/topics/ treatmentasprevention

μ

μ

μ

Stochastic model

Figure 2: Transmission model for HIV infection and antiretroviral therapy (ART) provision N represents population aged 15 years and above. People enter into the susceptible class (S) at a rate βN, become infected at a rate λSJ/N, progress through four stages of HIV (Ii, i=1–4) at a rate between each stage, and then die (D). The background mortality rate is μ and people are tested at a rate τ. If they are tested and put onto ART, they move to the corresponding ART box Ai (i=1–4), where they progress through four stages at a rate and then die. The term governing transmission contains the factor J α (Ii+εAi) where ε allows for the fact that people receiving ART are less infectious than are those who are not. They might also stop treatment or the treatment might become ineffective, in which case they return to the corresponding non-ART state at a rate . To allow for heterogeneity in sexual behaviour and for the observed steady state prevalence of HIV, we let the transmission decrease with the prevalence, P. If n=1, the decrease is exponential; if n=∞, the decrease is a step function. Both have been used in previous models.5,29

6 months

8 months

Average time between tests

1

2 1 year 3

4 2 years 5 4 years 10 years 0

500 1000 1500 CD4+ cells per μL at start of treatment

2000

Figure 3: Relation between HIV testing frequency, CD4+ cell count, and R0 R0 (the number of secondary infections resulting from one primary infection in an otherwise susceptible population) plotted against the CD4+ cell count at which treatment starts for diﬀerent frequencies of HIV testing (average time between HIV tests represented in years and months). Numbers in circles represent R0 values. Green shading: R0<1; yellow: 1
prevalence in women.27 Furthermore, we assumed that intravenous drug use does not contribute substantially to the overall rates of infection.28 We used a stochastic model to explore the eﬀect of various treatment strategies and model parameters on R0, allowing transmission to vary between the acute and chronic phase of HIV infection. To set an elimination 50

target for the purposes of this study, we deﬁned HIV elimination as a reduction in incidence to less than one case per 1000 people per year. We used a deterministic transmission model to explore the eﬀect of various HIV testing and treatment strategies on the long-term dynamics of the epidemic. The models were programmed in Visual Basic (Microsoft 2002).

In the stochastic model, we chose an initial value for a person’s CD4+ cell count and a survival time, which were taken from population distributions derived from available data. For the pre-infection distribution of CD4+ cell counts (ﬁgure 1A) we used data from a survey in Orange Farm, South Africa.29 We assumed that the CD4+ cell count decreased by 25% immediately after infection and linearly thereafter.29 Every month, the index case could infect another person, be tested for HIV, or, if HIV positive, start treatment at a preset CD4+ cell count. We assumed that survival after infection with HIV, without ART, would follow a Weibull distribution29 (ﬁgure 1B) with a mean of 11 (SD 0·5) years.31 The acute phase would last for 2 months, during which time the infectivity would be ten times higher than in the chronic phase (ﬁgure 1C).32,33 The ﬁnal phase would last for 5% of the survival time without ART, during which time the infectivity would be ﬁve times higher than in the chronic phase.33 If a person was tested and if their CD4+ cell count was less than a speciﬁc threshold, they would be given ART. Once receiving ART, we assumed that their infectiousness fell to 1% of their value before treatment on the basis of the relation between plasma viral load and ART and estimated decreases in viral load for people receiving ART.23,34 We started the model without ART by adjusting the chronic-phase infectivity to obtain a value of R0 equal to 7 and a doubling time of 1·25 years to match the data describing the epidemic trend in South Africa.29 The parameters in ﬁgure 1C could all be varied to examine the eﬀect of introducing ART at various CD4+ cell counts on the values of R0, to calculate the time from the infection of the index case to the infection of each secondary case, and to assess the cumulative number of infections over time (ﬁgure 1D). The model would provide the mean and frequency distribution of each of these statistics, which were obtained over many runs (ﬁgure 1D).

Deterministic transmission model The deterministic transmission model in ﬁgure 2 had four compartments for people infected with HIV to give a close ﬁt to the observed Weibull survival distribution (ﬁgure 1B). The compartments in the transmission model simulated the progression from infection to AIDS (and were unrelated to the acute and ﬁnal phases in the stochastic model). To calibrate the model we used the prevalence of HIV in South African adults aged 15 years and older, which www.thelancet.com Vol 373 January 3, 2009

Articles

A

D

HIV No intervention ART CD4+ when count <350 cells per μL Universal voluntary HIV testing and immediate ART

Mortality per year

Prevalence per year

Incidence per year

0·020 0·015

ART

0·005 0·004 0·003

0·010 0·002 0·005

0·001

0

0

B

E

0·15

0·15

0·10

0·10

0·05

0·05

0

0

C

F

0·010

0·010

0·005

0·005

0 1980

2000

2020

2040

Year

0 1980

2000

2020

2040

Year

Figure 4: HIV incidence (A), prevalence (B), and mortality (C) and the incidence (D), prevalence (E), and mortality (F) of people placed on antiretroviral therapy (ART) Data are proportions of adolescents and adults aged 15 years and older. Left panels, A−C, show HIV incidence, HIV prevalence, and mortality under the three scenarios. In B, data points with error bars give estimated prevalence of infection in adults derived from South Africa antenatal clinic data.35 Right panels, D−F, show incidence of ART—ie, the rate at which people start ART—the proportion of people receiving ART, and the mortality in those receiving ART.

was obtained by scaling data from the national antenatal clinic survey35 to the UNAIDS estimate for adults in 2005.36 To obtain the best ﬁt we varied the timing of the epidemic, the transmission parameter (λ), and the parameters that account for heterogeneity in sexual behaviour (α and n) (ﬁgure 2). A proportion of people oﬀered ART might refuse treatment or fail treatment within the ﬁrst few weeks; thus a proportion of people in the model did not start treatment. To allow for subsequent drop out because of logistical and other challenges such as adherence to treatment, toxic eﬀects of drugs, and drug resistance, individuals stopped taking ART at a rate per head per year (ﬁgure 2). We used failure, refusal, and retention rates that have been achieved in the national programme in Malawi.37 Data www.thelancet.com Vol 373 January 3, 2009

from the national ART programme in Malawi37 suggest that up to 8% of people drop out immediately or soon after starting treatment and then at between 1% and 3% per year, excluding those who die. We assumed a long-term drop-out rate of 1·5% per year. We assumed that ﬁrst-line treatment would fail in 3% of individuals every year and that those individuals would be identiﬁed and put on second-line treatment. Although their prognosis on second-line drugs would then be the same as for those on ﬁrst-line drugs, this assumption has important implications for the cost of treatment. When we introduced the intervention, we assumed that coverage would increase logistically to 50% by 2012 and 90% by 2016. We examined a strategy of universal voluntary HIV testing and immediate ART in which all adults in our 51

Articles

Deaths (thousands)

Deaths averted (thousands)

Neither strategy 2015

269

··

2030

263

··

2050 2008–50

263

··

11 078

··

ART started when CD4+ count <350 cells per μL 2015

193

76

2030

202

61

2050

210

53

8658

2419

2008–50

ART started when CD4+ count <350 cells per μL and universal voluntary HIV testing/immediate ART 2015

165

2030

76

187

2050

17

246

3879

7199

2008–50

104

ART started when CD4+ count <350 cells per μL, universal voluntary HIV testing/immediate ART, and other adult prevention strategies 2015

164

105

2030

72

191

2050

12

251

3727

7350

2008–50 ART=antiretroviral therapy.

Table: Estimated number of AIDS-related deaths for the years 2015, 2030, 2050, and 2008–50 with diﬀerent strategies

test-case community accepted being tested for HIV once a year on average, and all HIV-infected people started ART as soon as they were diagnosed HIV positive (ie, irrespective of clinical stage or CD4+ cell count). The sensitivity and speciﬁcity of the present generation of HIV tests are near 100% and result in a high predictive value, especially when diﬀerent tests are used together as part of a quality-controlled algorithm.38,39 Therefore, we assumed 100% sensitivity and speciﬁcity for the modelling exercise. We compared the theoretical strategy of universal voluntary HIV testing and immediate ART with a strategy in which ART was started with a CD4+ count less than 350 cells per μL on the basis of optimum implementation of present guidelines from the Southern African Clinicians Society40 in which we assumed that all infected individuals would have presented to a health service before their CD4+ count fell to 350 cells per μL and would start ART at that threshold. For this initial comparison, we focused on the eﬀect of ART and excluded the eﬀect of other prevention interventions. We used the deterministic transmission model to compare the eﬀect of the two interventions on the incidence, prevalence, and mortality of HIV-positive people who would be, and would not be, receiving ART. We then used the deterministic model to examine the time to elimination after implementation of the 52

theoretical strategy. We combined yearly universal voluntary HIV testing and immediate ART for all people who tested HIV positive irrespective of CD4+ cell count or clinical stage with a full package of relevant adult prevention interventions. Particularly in the early stages of the programme, this strategy could reach and place on ART people who would not have been tested for HIV and those with CD4+ counts greater than 350 cells per μL. We also assumed that other adult interventions for prevention of HIV would contribute to reductions in transmission. These interventions could include male circumcision, behaviour-change programmes, condom promotion, and treatment of curable sexually transmitted infections. We assumed that these other interventions together would reduce transmission by 40% and would be rolled out at the same rate as the ART programmes.

Cost analysis To estimate the funding needed to implement the two strategies in our test-case scenario of a severe generalised epidemic, we assumed a cost for ﬁrst-line drugs—including drug delivery, HIV and laboratory testing, and patient management—of US$727 (range $290–1163) for ﬁrst-line drugs and $3290 ($2497–4083) for second-line drugs.41–43 We assumed that 30% of treatment costs were for antiretroviral drugs43 and that 3% of people on ﬁrst-line drugs would fail and need second-line drugs every year. We also incorporated available data suggesting that costs for HIV testing range from 0·5–3% of the per person ﬁrst-line ART costs and 0·1–0·6% of the second-line ART costs (Mermin J, Centers for Disease Control and Prevention Kenya, Coordinating Oﬃce for Global Health, CDC, Nairobi, Kenya, personal communication). To compare the cost of the theoretical strategy for the hypothetical case, we assumed that 17% or $2·87 billion of estimated yearly global funding for HIV/AIDS up to 2008, and 17% or $8·84 billion of UNAIDS projected yearly needs for universal access up to 2015, would be available for HIV/AIDS control as described for South Africa.36,44 UNAIDS universal-access estimates for prevention, care, and treatment include treatment for 13·7 million people by 2010.35,44 Cost calculation results apply to this hypothetical scenario analogous to South Africa, unless otherwise speciﬁed.

Role of the funding source There was no funding source for this study. All authors had access to the data and agreed to submit for publication.

Results Figure 3 shows results from the stochastic model as contours of R0 plotted against the CD4+ cell count at which ART would start and the frequency of testing. To reduce R0 to less than 1 (green area), adolescents and adults would need to be tested at least once per year and started on ART when their CD4+ count is greater than www.thelancet.com Vol 373 January 3, 2009

Articles

www.thelancet.com Vol 373 January 3, 2009

Present phase

Roll-out

Elimination phase

Prevalence (per total population)

0·20

0·020 Incidence of HIV in people not yet started on ART Prevalence of HIV in people not yet started on ART Mortality of HIV in people not yet started on ART Prevalence of people on ART Mortality of people on ART

0·15

0·015

0·10

0·010

0·05

0·005

0 1980

1990

2000

2010

Year

2020

2030

2040

Incidence and mortality per year

0 2050

Figure 5: Time trends resulting from application of universal voluntary HIV testing and immediate ART strategy for people who test HIV positive, in combination with other adult prevention interventions that reduce incidence by 40% The programme implementation start date is arbitrarily set as immediate, with coverage increasing logistically to 50% by 2012 and 90% by 2016. The parameters are τ=1·0 per year; =0·08; =0·015 per year; and η=0·10. See ﬁgure 2 legend for description of these variables.

17% of available global funding for HIV prevention, care, and treatment 17% of the UNAIDS estimate of funds needed to achieve universal access by 2010 and sustain it to 2015 Universal voluntary HIV testing and immediate ART ART when CD4+ count <350 cells per μL

8

US$ billions per year

900 cell per μL. In South Africa, the average value of the CD4+ count immediately after seroconversion is about 884 cells per μL,29 so most adolescents and adults would need to start ART as soon as they are diagnosed with HIV to ensure that R0 stays below 1. Figure 3 also shows that if adolescents and adults are tested on average once a year, starting ART at a CD4+ count of 200 cells per μL could reduce R0 to 4, starting at 350 cells per μL could reduce it to 3, and starting at 500 cells per μL could reduce it to 2·5. Although these other strategies could have a substantial eﬀect on transmission, morbidity, and mortality, and are in agreement with previous studies,22,24,45 they would not be suﬃcient to reduce R0 to less than 1 and move the epidemic towards elimination. The stochastic model shows that testing all adolescents and adults at least 15 years old once a year, on average, and starting individuals on ART as soon as they test positive for HIV would reduce R0 below 1 and eventually eliminate HIV. The model allows for a high level of concurrency and for a much higher infectiousness during the acute phase than during the chronic phase.46 The strategy of starting ART when CD4+ count is less than 350 cells per μL approximates the optimum implementation of many national guidelines for HIV treatment, including those recommended for South Africa.40 In countries most heavily aﬀected by HIV, most people are not aware of their HIV status,10 and the median CD4+ count when individuals start ART is often much lower than 350 cells per μL.47 Figure 4 shows the deterministic transmission model of the two strategies—universal HIV testing and immediate ART strategy versus starting ART when the CD4+ count is less than 350 cells per μL. The strategy of starting ART when CD4+ count is less than 350 cells per μL could reduce both HIV incidence (ﬁgure 4A) and prevalence (ﬁgure 4B) by roughly 30% but could give a much greater reduction in mortality (ﬁgure 4C) since everyone, apart from those who refuse treatment, could start ART when their CD4+ cell count reached this level. The rate at which adolescents and adults start ART could increase to about 0·08% per year by 2014 (ﬁgure 4D), the prevalent number receiving ART could increase to about 5% of the adult population by 2020 (ﬁgure 4E), and the mortality of people receiving ART could increase to about 0·8% per year by 2020 (ﬁgure 4F). By contrast, the theoretical strategy of yearly universal voluntary HIV testing and immediate ART could reduce HIV incidence, prevalence, and mortality to about one case per ﬁve thousand adolescents and adults per year, since by 2016 most infected people could be receiving ART (ﬁgure 4). The number of people starting ART and the number already receiving ART could initially be much greater than with the strategy of starting ART when CD4+ count is less than 350 cells per μL (ﬁgure 4), but values decrease rapidly as transmission is interrupted. By 2016, fewer people could be starting ART and by 2030 fewer people could be receiving ART than with the

6

4

2

0 2000

2010

2020

2030

2040

2050

Year

Figure 6: Yearly cost of the two strategies compared with available and projected funding for HIV/AIDS for the test-case country Heavy lines correspond to expected treatment costs, light lines to high and low estimates of treatment costs for the hypothetical epidemic representing 17% of worldwide HIV prevalence, as described in the methods section. Blue line: 17% of available global funding for HIV prevention, care and treatment;44 brown line: 17% of the UNAIDS estimate of funds needed to achieve universal access by 2010 and to sustain it to 2015.36

strategy of starting ART when CD4+ count is less than 350 cells per μL (ﬁgure 4). Mortality could be much the same for both strategies until 2016, when the number of deaths per year could fall rapidly with the theoretical strategy of universal voluntary HIV testing and immediate ART (ﬁgure 4). We estimated a substantial diﬀerence in number of lives saved between the strategies (table). The number of HIV deaths between 2008 and 2050 could be about 53

Articles

11 million without ART, about 8·7 million with the strategy of starting ART when CD4+ count is less than 350 cells per μL, and 3·9 million with the theoretical strategy of universal voluntary HIV testing and immediate ART combined with the CD4+ count less than 350 cells per μL strategy. The combined strategy could reduce the number of HIV deaths up to 2050 by 55% compared with the strategy of starting ART when CD4+ count is less than 350 cells per μL, at which time the mortality could be only 0·04% per year and falling, rather than 0·6% per year and rising. Figure 5 shows the time to elimination after implementation of the theoretical strategy in combination with other adult prevention interventions. The implementation start date of the programme is arbitrary and the time to elimination is dependent on achievement of programme scale-up. With the assumption of an immediate start date and full programme scale-up by 2016, HIV transmission could switch from the present endemic phase, in which most people living with HIV would not be receiving ART, to the elimination phase, in which most would be receiving ART, at around 2010. By 2020, mortality in people not receiving ART could have fallen to about two per thousand adolescents and adults per year. Once the epidemic is in the elimination phase, the focus of control eﬀorts could also change. Before 2010, the focus could be on ensuring that HIV testing is widely done and ART is immediately available. After 2010, there could be an increasing need to ensure that people who are now receiving ART are fully adherent, that switching to second-line therapy is prompt and eﬃcient, that sexual partners, in particular, are monitored for evidence of secondary HIV infections, and that population-based drug resistance is monitored.48 We used the model to compare the cost of the theoretical strategy of universal voluntary HIV testing and immediate ART and the strategy of starting ART when CD4+ count is less than 350 cells per μL (ﬁgure 6). Initially, the cost of the universal voluntary HIV testing and immediate ART strategy is greatest—in 2015, the cost of this theoretical strategy is almost three times more than that of the strategy of starting ART when CD4+ count is less than 350 cells per μL. However, after 2015, the yearly cost of the strategy of universal voluntary HIV testing and immediate ART falls and is less than the cost of the strategy of starting ART when CD4+ count is less than 350 cells per μL after 2030. As the yearly costs for the strategy of universal voluntary HIV testing and immediate ART fall, the costs for the strategy of starting ART when CD4+ count is less than 350 cells per μL could continue to rise as more people need ART. The funding needed to implement the theoretical strategy for an epidemic of South African-type severity peaks in 2015 at $3·4 billion per year (range $2·2 billion–$5·3 billion). Although the initial yearly cost of the theoretical strategy is higher than the present 54

strategy, it is well within UNAIDS projections of the $8·84 billion needed every year for universal access to prevention, care, and treatment in a South African-type situation in 2015. In the long term, costs could reduce to very low amounts as progress towards elimination is achieved.

Discussion The results show that universal voluntary HIV testing once a year of all people older than 15 years, combined with immediate ART after diagnosis, could bring about a phase change in the nature of the epidemic. Instead of dealing with the constant pressure of newly infected people, mortality could decrease rapidly and the epidemic could begin to resemble a concentrated epidemic with particular populations remaining at risk. The focus of control would switch from making ART available to people with greatest need to providing support and services for those who are receiving ART. Transmission could be reduced to low levels and the epidemic could go into a steady decrease towards elimination as those receiving ART grow older and die. Although other prevention interventions, alone or in combination, could substantially reduce HIV incidence, our model suggests that only universal voluntary HIV testing and immediate initiation of ART could reduce transmission to the point at which elimination might be feasible by 2020 for a generalised epidemic, such as that in South Africa. This analysis lends support to, and extends, earlier analyses suggesting that rapid scale-up of conventional ART approaches could greatly reduce mortality49 and have a substantial eﬀect on HIV incidence.19,20 However, other studies that examined scaling up ART on the basis of present HIV testing access and clinical eligibility criteria for ART did not show the reduction of R0 to less than 1, which could be seen with the theoretical strategy of universal voluntary HIV testing and immediate ART.22,45 The main restrictions in this modelling exercise relate to need for much better data, especially for programmatic aspects of the intervention. Better data are needed for the acceptability and uptake of universal voluntary HIV testing, the infectiousness of people receiving ART, adherence, behaviour change after starting ART, and rates of emergence of resistance. Although we undertook a preliminary costing exercise, a full economic analysis of the proposed strategy could further improve our understanding of the economic implications of the theoretical strategy. Several trials are in progress, and many of these data could become available. A trial of our theoretical strategy is technically feasible, especially in view of the expected rapid eﬀect on incidence. Such a trial would address the assumptions in the model concerning practical aspects of implementing the strategy.50,51 The studied strategy would pose implementation challenges. It could initially be costly and labour www.thelancet.com Vol 373 January 3, 2009

Articles

intensive, and diﬃculties could arise from drug resistance or adverse events related to medication. It would be necessary to provide consistent, secure access to HIV rapid-test kits and ﬁrst-line and second-line antiretroviral drugs, to ensure high levels of adherence, and to monitor the programme carefully, especially when most people with HIV are receiving ART. The sensitivity and speciﬁcity of the present generation of HIV tests are near 100% and result in a high positive predictive value, particularly when diﬀerent tests are used together as part of a quality-controlled algorithm. However, the scale-up of HIV testing to millions of people means that even a small false-positivity rate could lead to many people being falsely diagnosed as HIV positive.38,39 Similar to HIV testing eﬀorts at present, this theoretical strategy will require substantial attention to the testing algorithm and quality assurance. The behavioural implications of a large cohort of people receiving ART on the community are largely unknown, and particular attention would have to be given to the sexual partners of people on this treatment. We have assumed a moderately paced scale-up that reaches full coverage by 2016, and we used rates of failure, refusal, and retention that have been achieved in the national programme in Malawi.37 If the intervention is scaled up more slowly or started at a later date, the long-term eﬀect would be the same but delayed. Universal access implies universal knowledge of serostatus, but even without early HIV testing, all HIV-infected people will eventually need ART for clinically progressive, ultimately fatal immunodeﬁciency. The question of when to start treatment is dominating therapeutic discussions, with frequent calls for earlier initiation of ART.52,53 In this context, the theoretical strategy of universal voluntary HIV testing and immediate ART has some programmatic advantages. It avoids some of the major operational diﬃculties of ART programmes in which every patient needs to be assessed for eligibility with CD4+ cell counts and possibly viral load measurements to establish whether deﬁned immunological or clinical thresholds, or both, for starting treatment have been achieved. Implementation of ART is simpler based solely on a positive HIV test, when most infected people are well with fairly well preserved immune systems and before many clients are lost to follow-up. The best possible drug regimen in this theoretical strategy should emphasise simplicity, low toxic eﬀects, and ease of adherence. There will be economies of scale, allowing use of more expensive but more convenient regimens than are used in Africa at present. The supply chain for drugs could beneﬁt from the predictable scale-up that relies on standard ﬁrst-line and second-line therapy with newer regimens that are simple to use. Since increasing numbers of people living with HIV access ART, the substantial reduction in morbidity is likely to reduce the burden on the health system, freeing human resources and capacity. www.thelancet.com Vol 373 January 3, 2009

Studies in Malawi and elsewhere have shown high adherence to ART, and we have no reason to believe that adherence would be reduced in special programmes,37,54 if adequate patient support and treatment literacy is provided. Studies suggest that the development and transmission of drug-resistant strains have been restricted despite the rapid scale-up of treatment.48 Universal voluntary HIV testing and immediate ART could raise concerns around human rights and coercion but appropriate training, engagement with the community, and supervision should keep problems to a minimum, and beneﬁts from reduced transmission should greatly outweigh adverse results. Universal voluntary testing and treatment could reduce HIV-associated stigma and could substantially reduce the incidence of AIDS-related disease and death, including that from tuberculosis.55 Maximising the prevention potential of ART also has important ﬁnancial implications. ART is still expensive, and concern exists over the long-term sustainability of treatment, especially if treatment is started earlier in the course of infection with less toxic but more expensive regimens. From 1996 to 2007, the yearly funding for HIV prevention, care, and treatment in low-income and middle-income countries increased 33-fold, reaching $10 billion in 2007.36 However, according to UNAIDS, even this amount of expenditure falls short of the estimated need. To achieve universal access—including treatment for 13·7 million people by 2010—ﬁnancial resources have to nearly quadruple by 2010 from 2007, reaching $41−58 billion by 2015.36 Universal voluntary HIV testing and immediate ART would entail a substantial, front-loaded investment. A full economic analysis is beyond the scope of this Article, but by changing the fundamental dynamics of the epidemic, our preliminary cost calculations suggest that there could be substantial yearly and long-term cost savings between now and 2050, by which time HIV infections could be reduced to very low and manageable amounts. Our modelling suggests that the cost of implementing the new intervention for the test-case country is much less than what UNAIDS projected for universal access to prevention, care, and treatment. Our model suggests that massive scale-up of universal voluntary HIV testing with immediate initiation of ART could nearly stop transmission and drive HIV into an elimination phase in a high-burden setting within 1–2 years of reaching 90% of programme coverage. As with all prevention interventions for HIV, this approach should not be viewed independently of other methods of prevention. Expert evaluation and consultation with all stakeholders, including community representatives, are needed to further assess this theoretical approach and to deﬁne the role of ART in the prevention and control of HIV/AIDS. 55

Articles

Contributors RMG, CFG, CD, KMDC, and BGW participated in developing the conceptual framework, and in analysis, drafting, and approval of the ﬁnal version of the report. Conﬂict of interest statement We declare that we have no conﬂict of interest. Acknowledgments The opinions and statements in this Article are those of the authors and do not represent the oﬃcial policy, endorsement, or views of WHO. References 1 Barre-Sinoussi F, Chermann JC, Rey F, et al. Isolation of a T-lymphotropic retrovirus from a patient at risk for acquired immune deﬁciency syndrome (AIDS). Science 1983; 220: 868–71. 2 Horton R, Das P. Putting prevention at the forefront of HIV/AIDS. Lancet 2008; 372: 421–22. 3 Desrosiers R. Scientiﬁc obstacles to an eﬀective HIV vaccine. Presented at 15th Conference on Retroviruses and Opportunistic Infections (CROI) Boston, USA; Feb 3–6, 2008, Abstract 91. http://www.retroconference.org/2008/Abstracts/33477.htm (accessed Nov 13, 2008). 4 Weiss RA. Special anniversary review: twenty-ﬁve years of human immunodeﬁciency virus research: successes and challenges. Clin Exp Immunol 2008; 152: 201–10. 5 Williams BG, Lloyd-Smith JO, Gouws E, et al. The potential impact of male circumcision on HIV in Sub-Saharan Africa. PLoS Med 2006; 3: e262. 6 Auvert B, Taljaard D, Lagarde E, Sobngwi-Tambekou J, Sitta R, Puren A. Randomized, controlled intervention trial of male circumcision for reduction of HIV infection risk: the ANRS 1265 Trial. PLoS Med 2005; 2: e298. 7 Potts M, Halperin DT, Kirby D, et al. Public health. Reassessing HIV prevention. Science 2008; 320: 749–50. 8 Ghys PD, Diallo MO, Ettiegne-Traore V, et al. Eﬀect of interventions to control sexually transmitted disease on the incidence of HIV infection in female sex workers. AIDS 2001; 15: 1421–31. 9 Des Jarlais DC, Hagan H, Friedman SR, et al. Maintaining low HIV seroprevalence in populations of injecting drug users. JAMA 1995; 274: 1226–31. 10 WHO. Towards universal access: scaling up priority HIV/AIDS interventions in the health sector. Geneva: World Health Organization, 2008. http://www.who.int/hiv/pub/towards_ universal_access_report_2008.pdf (accessed Nov 13, 2008). 11 UNAIDS. 2008 report on the global AIDS epidemic. UNAIDS Report, 2008. http://data.unaids.org/globalreport/2008/jc1510_ 2008Globalreport_en.zip (accessed Nov 13, 2008). 12 UNAIDS. Executive summary: 2008 report on the global AIDS epidemic. Geneva: UNAIDS, 2008. http://data.unaids. org/pub/GlobalReport/2008/JC1511_GR08_ExecutiveSummary_ en.pdf (accessed Nov 13, 2008). 13 Achievements in public health. Reduction in perinatal transmission of HIV infection—United States, 1985–2005. MMWR Morb Mortal Wkly Rep 2006; 55: 592–97. 14 Bunnell R, Mermin J, De Cock KM. HIV prevention for a threatened continent: implementing positive prevention in Africa. JAMA 2006; 296: 855–58. 15 WHO. Essential prevention and care interventions for adults and adolescents living with HIV in resource-limited settings. Geneva: World Health Organization, 2008. http://www.who.int/hiv/pub/ prev_care/OMS_EPP_AFF_en.pdf (accessed Nov 13, 2008). 16 WHO. Guidance on Provider-Initiated HIV Testing and Counseling in Health Facilities. Geneva: World Health Organization, 2007. http://whqlibdoc.who.int/publications/2007/9 789241595568_eng.pdf (accessed Nov 13, 2008). 17 Gilks CF, Crowley S, Ekpini R, et al. The WHO public-health approach to antiretroviral treatment against HIV in resource-limited settings. Lancet 2006; 368: 505–10. 18 Jahn A, Floyd S, Crampin AC, et al. Population-level eﬀect of HIV on adult mortality and early evidence of reversal after introduction of antiretroviral therapy in Malawi. Lancet 2008; 371: 1603–11. 19 Velasco-Hernandez JX, Gershengorn HB, Blower SM. Could widespread use of combination antiretroviral therapy eradicate HIV epidemics? Lancet Infect Dis 2002; 2: 487–93.

56

20

21

22 23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40 41

42

Montaner JS, Hogg R, Wood E, et al. The case for expanding access to highly active antiretroviral therapy to curb the growth of the HIV epidemic. Lancet 2006; 368: 531–36. Bunnell R, Ekwaru JP, Solberg P, et al. Changes in sexual behavior and risk of HIV transmission after antiretroviral therapy and prevention interventions in rural Uganda. AIDS 2006; 20: 85–92. Baggaley RF, Garnett GP, Ferguson NM. Modelling the impact of antiretroviral use in resource-poor settings. PLoS Med 2006; 3: e124. Kilby JM, Lee HY, Hazelwood JD, et al. Treatment response in acute/early infection versus advanced AIDS: equivalent ﬁrst and second phases of HIV RNA decline. AIDS 2008; 22: 957–62. Lima VD, Johnston K, Hogg RS, et al. Expanded access to highly active antiretroviral therapy: a potentially powerful strategy to curb the growth of the HIV epidemic. J Infect Dis 2008; 198: 59–67. Bunnell R, Opio A, Musinguzi J, et al. HIV transmission risk behavior among HIV-infected adults in Uganda: results of a nationally representative survey. AIDS 2008; 22: 617–24. May RM, Anderson RM. The transmission dynamics of human immunodeﬁciency virus (HIV). Philos Trans R Soc Lond B Biol Sci 1988; 321: 565–607. Williams BG, Gouws E, Colvin M, Sitas F, Ramjee G, Karim SSA. Patterns of infection: using age prevalence data to understand the epidemic of HIV in South Africa. S Afr J Sci 2000; 96: 305–12. Shisana O, Rehle T, Simbayi CC, Parker W, Zuma K, Bhana A. South African national HIV prevalence, HIV incidence, behavioural and communication survey, 2005. Cape Town: Human Sciences Research Council, 2005. Williams BG, Korenromp EL, Gouws E, Schmid GP, Auvert B, Dye C. HIV infection, antiretroviral therapy, and CD4+ cell count distributions in African populations. J Infect Dis 2006; 194: 1450–58. CASCADE. Survival after introduction of HAART in people with known duration of HIV-1 infection. The CASCADE Collaboration. Concerted Action on SeroConversion to AIDS and Death in Europe. Lancet 2000; 355: 1158–59. Ghys PD, Zaba B, Prins M. Survival and mortality of people infected with HIV in low and middle income countries: results from the extended ALPHA network. AIDS 2007; 21 (suppl 6): S1–4. Wawer MJ, Gray RH, Sewankambo NK, et al. Rates of HIV-1 transmission per coital act, by stage of HIV-1 infection, in Rakai, Uganda. J Infect Dis 2005; 191: 1403–09. Pilcher CD, Tien HC, Eron JJ Jr, et al. Brief but eﬃcient: acute HIV infection and the sexual transmission of HIV. J Infect Dis 2004; 189: 1785–92. Quinn TC, Wawer MJ, Sewankambo N, et al. Viral load and heterosexual transmission of human immunodeﬁciency virus type 1. Rakai Project Study Group. N Engl J Med 2000; 342: 921–29. Anonymous. National HIV and syphilis antenatal sero-prevalence survey in South Africa 2007. Pretoria: Directorate: Health Systems Research, Research Coordination and Epidemiology, 2008. http:// www.doh.gov.za (accessed Nov 13, 2008). UNAIDS. Financial resources required to achieve universal access to HIV prevention, treatment, care and support. Geneva: UNAIDS, 2007. http://data.unaids.org/pub/Report/2007/20070925_ advocacy_grne2_en.pdf (accessed Nov 13, 2008). Yu JK, Chen SC, Wang KY, et al. True outcomes for patients on antiretroviral therapy who are “lost to follow-up” in Malawi. Bull World Health Organ 2007; 85: 550–54. Plate DK. Evaluation and implementation of rapid HIV tests: the experience in 11 African countries. AIDS Res Hum Retroviruses 2007; 23: 1491–98. Wright RJ, Stringer JS. Rapid testing strategies for HIV-1 serodiagnosis in high-prevalence African settings. Am J Prev Med 2004; 27: 42–48. Anonymous. Southern African Clinicians Society guidelines: antiretroviral therapy in adults. South Afr J HIV Med 2008; 29: 18–31. Clinton Foundation. Antiretroviral (ARV) Price List, 2008. http://www.clintonfoundation.org/download/?guid=62e82ddc98de-102b-be34-001143e0d9b6 (accessed Nov 13, 2008). Médecins Sans Frontières. Untangling the web of ARV price reductions. Médecins Sans Frontières report, 2007. http://www. msfaccess.org/ﬁleadmin/user_upload/diseases/hiv-aids/ Untangling_the_Web/UTW10_RSep_horizontal.pdf (accessed Nov 13, 2008).

www.thelancet.com Vol 373 January 3, 2009

Articles

43

44 45

46 47

48

PEPFAR. The power of partnerships: the President’s emergency plan for AIDS Relief Third Annual Report to Congress”. Oﬃce of the Global AIDS Coordinator report, 2007. http://www.pepfar. gov/documents/organization/81019.pdf (accessed Nov 13, 2008). Cohen J. HIV/AIDS. The great funding surge. Science 2008; 321: 512–19. Gray RH, Li X, Wawer MJ, et al. Stochastic simulation of the impact of antiretroviral therapy and HIV vaccines on HIV transmission; Rakai, Uganda. AIDS 2003; 17: 1941–51. Abu-Raddad LJ, Longini IM Jr. No HIV stage is dominant in driving the HIV epidemic in sub-Saharan Africa. AIDS 2008; 22: 1055–61. Nash D. Long-term CD4 response to potent ART among ART-naïve patients in several low-income countries. Presented at 15th Conference on Retroviruses and Opportunistic Infections (CROI), Boston, USA, 2008; Abstract 126. http://www. retroconference.org/2008/Abstracts/31581.htm (accessed Nov 13, 2008). Bennett DE, Myatt M, Bertagnolio S, Sutherland D, Gilks CF. Recommendations for surveillance of transmitted HIV drug resistance in countries scaling up antiretroviral treatment. Antivir Ther 2008; 13 (suppl 2): 25–36.

www.thelancet.com Vol 373 January 3, 2009

49

50

51

52 53

54

55

Walensky RP, Wood R, Weinstein MC, et al. Scaling up antiretroviral therapy in South Africa: the impact of speed on survival. J Infect Dis 2008; 197: 1324–32. Hallett TB, Garnett GP, Mupamberiyi Z, Gregson S. Measuring eﬀectiveness in community randomized trials of HIV prevention. Int J Epidemiol 2008; 37: 77–87. Hallett TB, White PJ, Garnett GP. Appropriate evaluation of HIV prevention interventions: from experiment to full-scale implementation. Sex Transm Infect 2007; 83 (suppl 1): i55–60. Hirsch MS. Initiating therapy: when to start, what to use. J Infect Dis 2008; 197 (suppl 3): S252–60. Holmberg SD, Palella FJ Jr., Lichtenstein KA, Havlir DV. The case for earlier treatment of HIV infection. Clin Infect Dis 2004; 39: 1699–704. Brinkhof MW, Dabis F, Myer L, et al. Early loss of HIV-infected patients on potent antiretroviral therapy programmes in lower-income countries. Bull World Health Organ 2008; 86: 559–67. Corbett EL, Marston B, Churchyard GJ, De Cock KM. Tuberculosis in sub-Saharan Africa: opportunities, challenges, and change in the era of antiretroviral treatment. Lancet 2006; 367: 926–37.

57

Articles

Insecticide-treated net coverage in Africa: mapping progress in 2000–07 Abdisalan M Noor, Juliette J Mutheu, Andrew J Tatem, Simon I Hay, Robert W Snow

Summary Lancet 2009; 373: 58–67 Published Online November 18, 2008 DOI:10.1016/S01406736(08)61596-2 See Comment page 11 Malaria Public Health and Epidemiology Group, Centre for Geographic Medicine, KEMRI–University of Oxford–Wellcome Trust Collaborative Programme, Nairobi, Kenya (A M Noor PhD, J J Mutheu BBS, A J Tatem PhD, S I Hay DPhil, Prof R W Snow PhD); Centre for Tropical Medicine, Nuﬃeld Department of Clinical Medicine, University of Oxford, CCVTM, Oxford, UK (A M Noor, R W Snow); and Spatial Ecology and Epidemiology Group, Tinbergen Building, Department of Zoology, University of Oxford, Oxford, UK (A J Tatem, S I Hay) Correspondence to: Dr Abdisalan M Noor, Malaria Public Health and Epidemiology Group, Centre for Geographic Medicine, KEMRI–University of Oxford–Wellcome Trust Collaborative Programme, PO Box 43640-00100, Nairobi, Kenya [email protected]

58

Background Insecticide-treated bednets (ITNs) provide a means to improve child survival across Africa. Sales ﬁgures of these nets and survey coverage data presented nationally mask inequities in populations at biological and economic risk, and do not allow for precision in the estimation of unmet commodity needs. We gathered subnational ITN coverage sample survey data from 40 malaria-endemic countries in Africa between 2000 and 2007. Methods We computed the projected ITN coverage among children aged less than 5 years for age-adjusted population data that were stratiﬁed according to malaria transmission risks, proximate determinants of poverty, and methods of ITN delivery. Findings In 2000, only 1·7 million (1·8%) African children living in stable malaria-endemic conditions were protected by an ITN and the number increased to 20·3 million (18·5%) by 2007 leaving 89·6 million children unprotected. Of these, 30 million were living in some of the poorest areas of Africa: 54% were living in only seven countries and 25% in Nigeria alone. Overall, 33 (83%) countries were estimated to have ITN coverage of less than 40% in 2007. On average, we noted a greater increase in ITN coverage in areas where free distribution had operated between survey periods. Interpretation By mapping the distribution of populations in relation to malaria risk and intervention coverage, we provide a means to track the future requirements for scaling up essential disease-prevention strategies. The present coverage of ITN in Africa remains inadequate and a focused eﬀort to improve distribution in selected areas would have a substantial eﬀect on the continent’s malaria burden. Funding Wellcome Trust.

Introduction

Methods

Although international donor funding for malaria control in Africa has increased since 2002,1,2 funding remains inadequate2,3 and our understanding of how increased ﬁnancial resources have inﬂuenced equitable and targeted coverage of key malaria control strategies across Africa is incomplete. Insecticide-treated bednets (ITNs) are one of the most important methods for achievement of the Millennium Development Goal 6 target to reduce child mortality by 2015.4 Progress toward the Roll Back Malaria target of 80% ITN coverage among vulnerable groups is reported by the Global Fund for AIDS, Tuberculosis and Malaria (GFATM), WHO, and UNICEF as yearly increases in ITN procurement, and coverage data from national sample surveys summarised nationally.5,6 National survey data represent the most precise benchmark of progress toward internationally agreed targets. Deﬁnition of biological and economic vulnerability against intervention coverage targets subnationally, however, is central to the appropriateness of scaling up intervention coverage. In this report, we present a subnational analysis of temporal changes in ITN coverage among African children that also quantiﬁes the risks of Plasmodium falciparum transmission and proximate determinants of poverty.

ITN coverage data The main sources of ITN coverage data were national-household cluster-sample surveys undertaken as part of multiple indicators cluster surveys,7 demographic and health surveys,8 and national sample surveys—referred to as malaria indicator surveys—in countries with GFATM or bilateral donor funding. The multistage sampling design from ﬁrst-level administration (eg, province, state, or governorate) to nationalcensus-deﬁned enumeration clusters is common to all these surveys, and sample sizes are determined so as to provide precision in health and population indicators at the ﬁrst-level administrative unit (ADMIN1). We have reconstructed information from survey reports, websites, and other published sources on the numbers of children aged less than 5 years; number of these children reported sleeping on the night before the survey under a net that was treated in the past 6 months or that was a long-lasting treated net; and dates of the survey and the ﬁrst-level sampling geographical extent reported in each survey. We have selected two periods of ITN coverage data by choosing national surveys undertaken as close as possible to 2000 and 2007. ITN coverage data were not available for Botswana, Cape Verde, Reunion, Gabon, and Liberia in either period. For Comoros, ITN data were only www.thelancet.com Vol 373 January 3, 2009

Articles

available for 2000. These six countries are therefore not analysed further and represent only 1·0% (1·18 million) of the total childhood population in malaria-endemic Africa and 0·9% (1·03 million) of children exposed to stable P falciparum in 2007. For early multiple indicators cluster surveys and demographic and health surveys that predated a standardised malaria module,9 some ITN coverage data were reported in a non-standard format—eg, proportion of households in which all or some children slept under a net or ITN (Mali, 2001; Tanzania and Zimbabwe, 1999);

proportion of nets used by children that were ITN (Uganda, 2000–01); proportion of women aged 15–49 years who were sleeping under a net (Mozambique, 2003); or were reported for some ADMIN1 units and not others (Sudan, 1999). In each case we made an informed decision on the likely association between the reported indicators and the proportion of children sleeping under an ITN. These adjusted survey data indicate very low ITN coverage (mean 2·6% [SD 2·4]) and thus absolute errors resulting from these assumptions are likely to be small.

A

B

C

D

Figure 1: Availability of insecticide-treated bednet (ITN) data for two study periods, delivery methods, poverty mapping, and risk of malaria National boundaries are shown in black and ﬁrst-level administrative boundaries are shown in white. (A) First-level administrative units used to deﬁne ITN use between 2000 and 2007. Countries shown in grey are those where no baseline or matched follow-up data were available (Botswana, Cape Verde, Comoros, Gabon, Liberia, and Reunion) or that were not at risk of Plasmodium falciparum malaria (Algeria, Egypt, Lesotho, Libya, Morocco, Tunisia, and seven provinces in South Africa). (B) Main delivery methods adopted by countries after 2000 and before follow-up national surveys as indicated in table 1. Light green is cost recovery through public sector or subsidised private or public sector; middle green is highly subsidised routine distribution through public sector; and dark green is free mass campaigns, either localised or nationally, or routine free distribution through public sector. (C) Poverty map showing the least poor quintile (light blue), the two moderately poor quintiles (middle blue), and the two poorest quintiles (dark blue), based on the mean brightness of night-time lights in the ﬁrst-level administration unit. (D) Map of malaria showing areas of no malaria risk (white) and those under unstable (light pink) and stable (dark pink) transmission.12

www.thelancet.com Vol 373 January 3, 2009

59

60

4

6

13

1

10

2

8

4

11

2

11

3

6

11

7

10

8

4

8

4

3

7

5

Angola

Benin

Burkina Faso

Burundi

Cameroon

CAR

Chad

Congo

Côte d’Ivoire

Djibouti

DRC

Equatorial Guinea

Eritrea

Ethiopia

Gambia

Ghana

Guinea

Guinea Bissau

Kenya

Madagascar

Malawi

Mali

Mauritania

Estimated

DHS (2001)

DHS (2000)

MICS (2000)

MICS (2000)

MICS (2000)

Estimated

DHS (2003)

MICS (2000)

DHS (2000)

DHS (2002)

MICS (2000)

MICS (2001)

Estimated

MICS (2000)

Estimated

MICS (2000)

MICS (2000)

MICS (2000)

MICS (2000)

DHS (2003)

DHS (2001)

MICS (2001)

Estimated

January to May

July to November

August to October

September to October

April to May

Estimated

July to October

May to June

February to May

March to July

September to November

April to October

Estimated

June to August

Estimated

May to September

April to December

July to August

April to May

January to July

May to June

March to October

Source (year) Months

ADMIN1 2000–03 (baseline)

··

11143

10 559

6561

7116

5856

··

3593

3620

11 771

5935

3033

9954

..

7952

··

5384

14 268

3385

3325

9894

4689

5663

Children <5 years

<2%†

557 (5%)

317 (3%)

59 (<1%)

213 (3%)

410 (7%)

<1%†

144 (4%)

543 (15%)

0

237 (4%)

30 (1%)

100 (1%)

<1%†

80 (1%)

<5%†

54 (1%)

285 (2%)

34 (1%)

33 (1%)

198 (2%)

328 (7%)

113 (2%)

ITN use

February to May

May to August

July to October

August to October

July to November

July to December

October to December

May to June

September

April to June

July to November

November to April

Months

April to September

July to November

April to May

June to July

May to June

February to June

August to October

December to March

Octoberto December

DHS (2003–04) August to February

DHS (2006)

MICS (2006)

Other (2008)15

MIS (2007)

MICS (2006)

DHS (2005)

MICS (2006)

MICS (2005–06)

MIS (2007))

Other (2004)§14 October to November

Other (2007)‡

DHS (2007)

MICS (2006)

MICS (2006)

DHS (2005)

DHS (2004)

MICS (2006)

MICS (2006)

Other (2007)13

MICS (2006)

DHS (2006)

MIS (2006–07)

Source (year)

2004–07 (follow-up)

3239

13 359

19 628

3355

6783

5845

6839

3467

6543

5225

1266

3456

8913

2245

8650

4340

6206

9585

6361

2845

5384

15 941

2739

65 (2%)

3607 (27%)

4907 (25%)

2047 (61%)

2985 (44%)

2280 (39%)

<1%†

763 (22%)

3207 (49%)

1724 (33%)

608 (48%)

346 (10%)

980 (11%)

22 (1%)

519 (6%)

260 (6%)

310 (5%)

1438 (15%)

827 (13%)

910 (32%)

538 (10%)

3188 (20%)

493 (18%)

Children ITN use <5 years

3%

31%

28%

61%

44%

45%

1%

27%

58%

33%

85%

10%

11%

2%

7%

7%

7%

16%

15%

32%

13%

22%

19%

ITN use

Projection for July, 2007

CRps

··

··

··

··

··

··

··

··

··

Spps

··

CRps

··

CRps

··

··

CRps

CRps

··

··

··

··

2000

··

··

··

··

··

··

··

··

Spps

··

RFD§

CRps

··

··

··

··

··

··

··

··

CRps

Spps

··

2001

··

Spps

HSrp

Spps

Spps

··

··

RFD

··

··

RFD§

··

··

··

··

··

··

··

··

··

··

··

Spps

2002

RFD

··

··

··

··

··

Spps

FMCL

··

Spps

RFD§

··

Spps

··

··

··

··

··

··

Spps

··

··

··

2003

··

··

··

··

··

··

Spps

··

FMCL

··

FMCN

RFD§

··

··

··

··

Spps

··

··

··

RFD, FMCL

Spps

··

HSrp, RFD

2005

··

··

FMCN

HSrp, FMCN

FMCN

··

··

FMCN*

··

FMCN

RFD§

FMCN

FMCL

FMCN*

··

..

HSrp

··

RFD*

FMCN

··

··

FMCN

2006–07

(Continues on next page)

HSrp

··

··

HSrp

··

HSrp

FMCL

··

··

RFD§

··

··

··

··

··

··

··

CRps, RFD

··

··

HSrp

··

2004

Start year and modality of ITN scale-up

Articles

www.thelancet.com Vol 373 January 3, 2009

www.thelancet.com Vol 373 January 3, 2009

9

10

Zambia

Zimbabwe

DHS (1999)

DHS (2001– 02)

DHS (2000– 01)

MICS (2000)

DHS (1999)

MICS (2000)

MICS (2000)¦¦

Other (2001)19

MICS (1999)

MICS (2000)

MICS (2000)

MICS (2000)

MICS (2000)

DHS (2003)

MICS (2000)

DHS (2000)

DHS (2003)

August to November

November to May

September to March

August to September

September to November

July to September

July to August

..

November to December

March to May

May to July

August to September

July to October

March to August

April to August

September to December

August to December

3269

5786

8049

3126

3011

3508

23 297

··

4487

2686

9033

2185

3153

5861

5080

3785

10 648

Children <5 years

<3%†

58 (1%)

<1%†

56 (2%)

60 (2%)

21 (<1%)

443 (2%)

1%†

54 (1%)

54 (2%)

181 (2%)

503 (23%)

126 (4%)

57 (1%)

51 (1%)

114 (3%)

532 (5%)

ITN use

June to July

Months

March to June

··

August to September

October to November

November to December

May to July

February to July

March to April

January to June

May to June

May to October

May to June

June to August

DHS (2005–06) August to March

MIS (2006)

DHS (2006)

MICS (2006)

Other (2006– 07)**

DHS (2006–07) July to February

Other (2006)16

PC (2007)¶

MICS (2006)

MICS (2005)

MIS (2006)

MICS (2006)

DHS (2005)

MICS (2007)

DHS (2006)

DHS (2006–07) October to March

MIS (2007)

Source (year)

2004–07 (follow-up)

5751

2700

8402

4073

8608

3268

2363

··

6304

173 (3%)

621 (23%)

840 (10%)

1548 (38%)

2324 (27%)

20 (<1%)

189 (8%)

<5%†

567 (9%)

262 (5%)

749 (16%)

4683

5245

1313 (42%)

979 (13%)

579 (3%)

576 (7%)

600 (11%)

363 (7%)

ITN use

3125

7534

16 549

8236

5453

5187

Children <5 years

4%

29%

11%

45%

33%

1%

10%

5%

10%

6%

17%

46%

18%

4%

8%

12%

7%

ITN use

Projection for July, 2007

··

··

CRps

··

··

··

··

··

··

CRps

··

CRps

··

··

Spps

··

··

2000

··

··

··

··

··

··

··

··

··

··

··

··

··

··

··

··

··

2001

··

··

··

··

··

··

··

··

··

··

Spps

··

··

··

··

··

··

2002

··

Spps FMCL

Spps

··

··

Spps

··

··

··

··

··

··

Spps

··

··

Spps

Spps

2003

··

HSrp

··

FMCN

Spps

··

CRps

··

Spps

··

··

··

··

··

··

FMCL

··

2004

Start year and modality of ITN scale-up

Spps

FMCN

··

··

··

··

FMCL

··

FMCL

··

··

FMCL

··

CRps HSrp

FMCN

··

FMCL

2005

HSrp

FMCN*

··

··

HSrp, FMCL

··

FMCL

··

··

FMCN*

HSrp

··

RFD, FMCN*

··

FMCN

··

··

2006–07

Table 1: National summary of insecticide-treated bednet (ITN) coverage data sources for baseline, follow-up, and target periods, and evolution of ITN distribution mechanisms for 40 African malaria-endemic countries

The recorded ITN distribution mechanisms are cost recovery through public sector (CRps), subsidised private or public sector (Spps), highly subsidised routine distribution through public sector (HSrp), routine free distribution through public sector (RFD), localised free mass campaigns (FMCL), and national free mass campaigns (FMCN); and are summarised in ﬁgure 1B. Botswana, Cape Verde, Reunion, Comoros, Gabon, and Liberia were not included because they had no survey data for baseline or follow-up years. ADMIN1=ﬁrst-level administrative unit. CAR=Central African Republic. DHS=demographic and health surveys. DRC=Democratic Republic of Congo. MIC=multiple indicators cluster surveys. MIS=malaria indicator surveys. PC=personal communication. *Countries where free mass campaigns of long-lasting treated nets were undertaken after the follow-up survey are Djibouti in November, 2006; Ghana in November, 2006; Rwanda in September, 2006; Sierra Leone in November, 2006; Cameroon in 2007; Mali in 2007; and Zambia in 2007. †Values are estimates because survey data were not available for the country during a particular period. ‡Derived from household sample survey data from 2007 (Kleinschmidt I, London School of Hygiene and Tropical Medicine, personal communication). §In Eritrea data for four of six ADMIN1 units were obtained from a published source14 and an averaged assumption was imputed for the remaining two on the basis of neighbouring ADMIN1 coverage. Eritrea uses regular free distribution through the routine public sector as the main channel to deliver ITNs and is classiﬁed with free mass campaign in ﬁgure 1B. ¶For the two malaria endemic provinces in South Africa, we have assumed that ITN is less than 5% in 2007 after communications with provincial malaria control programme managers in both areas who indicated that over the past 5 years the focus of malaria vector control has been on indoor residual spraying rather than ITN distribution (Kruger P, Limpopo Malaria Control Programme, personal communication). ||In Sudan, baseline ADMIN1 data were available only for north Sudan (n=16). For south Sudan, ITN coverage data for the ADMIN1 units (n=10) were estimated from the means of nearest neighbours (n=5); or were assigned the national average of 1·9% from the MICS in 2000 (n=5). **Tanzania data were obtained from a survey on the national ITN voucher scheme undertaken in 200617 and sample surveys undertaken on the islands of Zanzibar and Pemba in 2007.18

4

2

South Africa

Uganda

3

Somalia

5

4

Sierra Leone

Togo

10

Senegal

9

2

São Tomé and Principe

Tanzania

12

Rwanda

4

6

Nigeria

Swaziland

8

Niger

26

13

Namibia

Sudan

11

Mozambique

(Continued from previous page)

Source (year) Months

ADMIN1 2000–03 (baseline)

Articles

61

Articles

The reference ADMIN1 digital boundaries for Africa were obtained through a combination of data from the UN Geographic Information Working Group,10 and the Food and Agriculture Organisation.11 These boundary units matched the reported information on ITN for 34 of 40 national survey reports assembled for the baseline period (2000–03) and 30 of 40 national surveys for the follow-up period (2004–07). For Angola, Benin, Chad, Djibouti, Guinea Bissau, Mali, Mauritania, Madagascar, Nigeria, and Uganda, non-standard ADMIN1 units were reported in the national sample surveys and these were digitised with ArcGIS (version 9.1) to replace existing ADMIN1 boundaries and thus create one ITN spatial reporting surface. In Angola, Central African Republic, Somalia, Madagascar, Equatorial Guinea, Burundi, and Tanzania, baseline and follow-up national surveys were sampled and presented at diﬀerent administrative resolutions between surveys. These were reconciled to the largest unit reported in either type of survey to maintain the integrity of the sample-size precision and represent temporally congruent units. Figure 1A shows the 286 spatial units used to deﬁne ITN use between 2000 and 2007 in the 40 countries where analysis was undertaken. We computed the absolute diﬀerence in ITN use between baseline and follow-up surveys for each ADMIN1. We then computed the number of months between surveys and obtained the monthly change in ITN, which we used as our ITN growth rate. We used this growth rate to project all ADMIN1 survey data described during the second-period survey to just one time reference in July, 2007. Since 2003, a wide range of approaches was used to improve the delivery of ITN to young children across Africa. These approaches can be classiﬁed broadly as cost recovery through the public sector, subsidised private or public sector, highly subsidised routine distribution through the public sector, routine free distribution through the public sector, and free mass campaigns; and are localised within speciﬁc ADMIN1 areas or nationally. We identiﬁed the adoption of these varied approaches by individual countries at diﬀerent times between 2000 and 2007 through various literature and web searches. Table 1 shows the national adoption of each strategy and ﬁgure 1B summarises this information spatially according to the distribution methods post-baseline and pre-follow-up survey.

Deﬁnition of poverty Roll Back Malaria targets emphasise the need to target speciﬁcally those population groups living in the two poorest quintiles in areas of biological vulnerability.19 The deﬁnition of poverty is fraught with diﬃculties and most health indicators are expressed against composites of household assets as measures of economic vulnerability but are diﬃcult to compare between countries. A spatially consistent and simpler alternative is to use a proxy for poverty—ie, illumination from night-time lights that are 62

seen from earth-orbiting satellites. Data for these lights have been used as a surrogate for economic vulnerability and poverty mapping in North America, Europe, and globally20 and were shown to be highly discriminatory in the separation of the most and least poor administrative areas in Africa.21 Here we use operational linescan system night-time lights gridded data that are produced by the US Defence Meteorological Satellite Program and the National Oceanic and Atmospheric Administration’s National Geophysical Data Center. We downloaded global night-time light data at about 1 km×1 km spatial resolution for the year 2000 from the National Geophysical Data Center-US Defence Meteorological Satellite Program website20 in raster-grid format and extracted data for Africa. We computed the mean brightness of the light pixels for each ITN ADMIN1 and then ranked them across Africa into quintiles with the most economically vulnerable represented by the two lowest quintiles of night-time lights (ﬁgure 1C).

Children at risk of P falciparum transmission We used recently completed work on the limits of stable P falciparum transmission.12 We assumed no transmission when medical intelligence from international travel advisories or national malaria-control programmes stated no malaria risk or when temperature was too low for sporogony to complete within the average lifespan of the local dominant vector species. Unstable malaria in Africa refers to areas where transmission is plausible biologically but limited by the eﬀects of aridity on anopheline adult and larval survival, and the clinical incidence is less than one case per 10 000 population per year. A deﬁnition of stable malaria was assumed to be a minimum of one clinical case per 10 000 population per year in a particular administrative unit, similar to rules used during the global malaria eradication programme. Within this range of stable transmission, conditions of transmission intensity vary enormously but cover all those in Africa where ITN is recommended as a key malaria prevention strategy.22 The three classiﬁcations of malaria risk are shown in ﬁgure 1D.

Deﬁnition of projected population estimates for 2007 The Global Rural Urban Mapping Project (version alpha) provides gridded population counts and population density estimates for the years 1990, 1995, and 2000, both adjusted and unadjusted to the UN national population estimates.23 We projected the adjusted population counts for the year 2000 forward to create seven further population count surfaces for each year from 2001 to 2007 by applying national, medium variant, intercensal growth rates by country, with methods previously described.24 We then stratiﬁed these population counts nationally by age group using UN-deﬁned24 population age structures to obtain population count surfaces for children younger than 5 years for each year from 2000 to 2007. We used these population maps in combination with the ADMIN1 www.thelancet.com Vol 373 January 3, 2009

Articles

A

B

São Tomé & Principe

São Tomé & Principe

The Gambia

The Gambia

Rwanda

Rwanda

Swaziland

Swaziland

Burundi

Burundi

C Country boundary First-level administrative unit boundary Proportion of children younger than 5 years sleeping under an insecticide-treated bed net 0 0·1–4·9% 5·0–9·9% 10·0–14·9% 15·0–19·9% São Tomé & Principe

20·0–29·9% 30·0–39·9% 40·0–49·9% 50·0–59·9%

The Gambia

60·0–79·9% 80·0–100% No data/malaria risk

Rwanda

Swaziland

Burundi

Figure 2: Insecticide-treated bednet (ITN) coverage among children under the age of 5 years reported in (A) 1999–2003 (baseline) and (B) 2004–07 (follow-up), and (C) projected to July, 2007 (target period)

boundaries to extract the numbers of children less than 5 years of age living with no risk, or unstable and stable P falciparum transmission risk for each reconstituted ADMIN1 polygon in ArcView (version 3.2).

of the report. The corresponding author had full access to all the data in the study and had ﬁnal responsibility for the decision to submit for publication.

Results Role of the funding source The sponsor of the study had no role in study design, data collection, analysis, interpretation of data, or writing www.thelancet.com Vol 373 January 3, 2009

Table 1 shows that data were obtained from multiple indicators cluster surveys (n=21), demographic and health surveys (n=14), or alternative sources (n=5), and 63

Articles

Proportion of children, younger than 5 years, using insecticide-treated bed nets in 2007 (%)

100

80

60

40

20

0 Free distribution

Routine subsidised delivery

Full cost recovery

Figure 3: Insecticide-treated bednet (ITN) use among children under the age of 5 years in 2007 by the main country ITN delivery mechanism The box indicates the IQR (25% and 75%); the thick line within the box represents the median; and the error bars represent the 2·5% and 97·5% centiles; and outliers are plotted as circles outside this range. ITN distribution was free (n=117), moderately or highly subsidised (n=117), and full cost recovery (n=50). The two malaria-endemic ﬁrst-level administration units (ADMIN1) in South Africa (KwaZulu Natal and Mpumalanga) are not included. Median ITN coverage among children less than 5 years of age was higher in ADMIN1 without free distribution than in those with routine subsidised delivery (ANOVA p=0·05) or full cost recovery (p=0·02).

used to deﬁne ITN use in the 286 ADMIN1 areas between 1999 and 2003, with median ITN use among children aged less than 5 years of 3·04% (IQR 0·50–3·71). We used 2003 demographic and health survey data for Burkina Faso, Ghana, Mozambique, and Nigeria because there were no earlier nationally representative surveys with ITN data within the baseline period. Although there were no national sample survey data close to 2000, data gathered between 2005 and 2006 showed that ITN coverage among children less than 5 years of age was between 0 and less than 5% for Mauritania, Congo, Djibouti, and Guinea. With such a low coverage, a reasonable assumption is that a negligible change in ITN coverage occurred between baseline and our reference year of 2007. For the two malaria-endemic provinces of South Africa (KwaZulu Natal and Mpumalanga), we used data from a Roll Back Malaria report in 200125 as the baseline. In the surveyed countries, 94 million children were likely to be living in areas of stable malaria-endemic risks in 2000. Figure 2A shows almost universally poor ITN use across these countries in 1999–2003, with only 50 (17%) of the ADMIN1 areas surveyed on the continent during this period showing more than 5% of children using an ITN. Additionally, 1·7 million (1·8%) children were protected by an ITN in areas of stable P falciparum malaria in 2000. Between 2004 and 2007, we obtained data from 13 demographic and health surveys, 14 multiple indicators cluster surveys, and 13 other national surveys and personal communications (table 1).13–18,26–31 Figure 2B shows the ADMIN1 ITN use distribution across countries 64

reporting data after 2003. The average duration between the data shown in ﬁgure 2B and ﬁgure 2A was 5·3 years. In 2004–07, 205 of 286 (72%) administrative polygons reported childhood ITN use to be greater than 5%, 109 (38%) reported ITN use in excess of 20%, 11 (4%) above 60%, and three (1%; one in Ethiopia and two in Madagascar) had reached the Roll Back Malaria target of 80%. The fastest yearly growth in ITN usage estimates between 2000 and 2007 was in Eritrea, Madagascar, Ghana, Togo, Kenya, Gambia, Guinea Bissau, Zambia, Ethiopia, and Burundi—all are countries that had promoted the delivery of free nets through mass campaigns between survey periods, except Eritrea, which relied on free distribution through the routine public-health system (ﬁgure 1B). The median reported ITN use for 2004–07 for ADMIN1, where the main delivery channel was free distribution (national and local), was 25·2% (n=117, IQR 11·2–40·3) compared with 14·1% (n=117, 4·2–24·4) for ADMIN1 areas where no free mass campaigns had been implemented, but where subsidised and heavily subsidised delivery had been promoted. Unsurprisingly, areas without free or subsidised programmes, relying mainly on full-cost recovery mechanisms of ITN delivery, had the lowest ITN coverage during the period of observation (n=50, median 3·9%, 1·8–7·4; ﬁgure 3). The coverage data shown in ﬁgure 2B indicate varying periods after 2004 and we have standardised the estimates of ITN use to the base year 2007 by use of expected ITN use growth rates per ADMIN1 for the 40 countries reporting in 2007. Mean projection periods were 14 months from the reported follow-up survey through to July, 2007 (ﬁgure 2C). These estimates of adapted ITN usage suggest that 218 (76%) of 286 areas had childhood ITN use reported as greater than 5% in 2007, 20 (7%) administrative areas reported ITN use in excess of 60%, and 10 (3%) of the ADMIN1 areas had reached the Roll Back Malaria ITN use target of at least 80% (ﬁgure 2C). With the projected ITN use data adjusted to childhood populations at risk of stable P falciparum transmission, we estimate that 20·3 million (18·5%) children younger than 5 years were protected by an ITN in 2007, whereas 89·6 million (81·5%) remained unprotected, of which 30 million unprotected children lived in the poorest areas of stable malaria-endemic Africa (table 2). Among communities in unstable transmission areas or areas at no risk of transmission, similar proportions of children were likely to have been protected by an ITN (table 2). On a continental scale, ITN coverage in 2007 showed equity in relation to proximate determinants of poverty without signiﬁcant diﬀerences in ITN coverage between children living in the most poor areas compared with those in the least poor (21% vs 16% [table 2]; ANOVA p=0·275) with similar observations across all malaria risk classes (data not shown). Biological equity scaled less well, however, with more children protected in areas of no or www.thelancet.com Vol 373 January 3, 2009

Articles

No P falciparum risk (N=11·6 million)

Unstable P falciparum risk (N=2·2 million)

Stable P falciparum risk (N=109·9 million)

Total (N=123·7 million)

Least poor (highest NTL quintile, N=22 500 000) Children protected by ITN ADMIN1 with ITN use

441 000 (18·2%) 40 (13·9%, 4·4–38·4)

1 200 000 (16·8%) 11 (5·05%, 2·2–19·0)

3 080 000 (15·9%) 57 (15·6%, 4·7–32·5)

3 640 000 (16·2%) 57 (15·6%, 4·7–32·4)

Moderately poor (NTL quintiles 2 and 3, N=58 700 000) Children protected by ITN ADMIN1 with ITN use

1 406 000 (25·9%) 61 (17·3%, 5·1–34·0)

83 000 (30·3%) 24 (36·5%, 20·9–50·0)

9 779 000 (18·5%) 112 (21.4%, 6·1–37·5)

11 268 000 (19·2%) 114 (19·6%, 6·0–37·1)

Most poor (lowest NTL quintiles 4 and 5, N=42 500 000) Children protected by ITN ADMIN1 with ITN use

1 124 000 (29·8%) 62 (13·6%, 6·3–32·3)

319 000 (27·3%) 52 (11·1%, 2·5–31·1)

7 484 000 (19·9%) 113 (12·5%, 6·1–29·9)

8 927 000 (21·0%) 116 (12·3%, 5·7–29·8)

Total Children protected by ITN ADMIN1 with ITN use

2 972 000 (25·5%) 163 (15·6%, 5·1–35·2)

522 000 (24·2%) 87 (16·6%, 4·5–35·6)

20 343 000 (18·5%) 282 (15·0%, 5·6–32·9)

23 837 000 (19·2%) 286 (14·3%, 5·1–32·6)

Data are number (%) or number (median, IQR). The P falciparum-endemic countries not included are Botswana, Cape Verde, Comoros, Gabon, Liberia, and Reunion, and represent 1·0% of children in all malaria-endemic countries of Africa in 2007. ADMIN1=ﬁrst-level administrative units. NTL=night-time lights.

Table 2: Children aged less than 5 years who were protected by an insecticide-treated bednet (ITN) in 2007 according to classes of Plasmodium falciparum and poverty risks across 40 malaria-endemic countries

unstable risk compared with areas with stable endemic malaria (25·3% vs 18·5% [derived from data in table 2]; p=0·032).

Discussion We estimate that only about a ﬁfth of children at risk of stable malaria transmission were protected by an ITN in 2007. Conversely, nearly 90 million African children living under conditions of stable malaria transmission have been neglected by the calls for rapid scaling up of ITN coverage made by the Roll Back Malaria movement in 2000.32 This neglect comes at a time when these same agencies and international partners are calling for elimination33 and a malaria-free world.34 A large proportion of these unprotected children live in some of the poorest parts of Africa, but the diﬀerences in ITN use between spatially deﬁned areas of least and most poor seem less obvious than has been reported before intervention coverage was scaled up (table 2). The procurement of more than 60 million long-lasting treated nets35 and the reports of rapid scaling up of ITN delivery in a few select countries have been hailed as a clear indication of progress toward Roll Back Malaria and Millennium Development Goals’ targets.5,6 Reports of sales ﬁgures or percentage coverage changes from selective examples mask underlying inequities in ITN use on a continental scale, adjusted for actual populations at risk of developing malaria. Almost 25 years have elapsed since the ﬁrst clinical trials of ITN were completed in Africa and more than 10 years since large-scale clinical trial evidence was provided on the contribution of ITN as a major method to reduce childhood mortality across most malaria-endemic settings in Africa. Why then is coverage so poor in 2007? An analysis of international donor funding in relation to populations exposed to stable transmission highlights huge disparities www.thelancet.com Vol 373 January 3, 2009

and inadequacies in malaria funding across Africa,2 which must contribute to inabilities to scale up coverage. Some controversy remains about the best approaches to ITN delivery.36,37 In this report, we have shown that the areas of Africa that have promoted free ITN distribution (ﬁgure 1B) have overall achieved more rapid progress than those that rely on cost recovery (21% lower median coverage) or routine subsidised public-sector promotion (11% lower median coverage). Fortunately, increasing numbers of countries are complementing existing delivery strategies with free distributions as national or localised strategies after the period of observations reported here, and current ITN coverage in these countries might be higher than our projected estimates. In some cases, biological vulnerability has scaled up with diﬀerences in ITN use within a country, notably Angola, Eritrea, Kenya, Madagascar, and Zambia, and less strategically elsewhere, notably Sudan. National ITN coverage was less than 15% in 2007 in 13 countries, including seven countries (Nigeria, Demographic Republic of Congo, Uganda, Sudan, Mozambique, Côte d’Ivoire, and Cameroon) that account for 53·5 million (48·6%) of all children (110 million) in Africa living under conditions of stable malaria transmission and 48·3 million (54%) of all unprotected children (89·6 million) in these transmission areas. Nigeria alone accounts for 22·2 million (25%) of all African children (89·6 million) living under conditions of stable malaria transmission who were not protected by an ITN in 2007. A focus of attention on these areas in Africa must be seen as a priority if health eﬀects at the continental level are to be realised by 2015. Such approaches to mapping intervention coverage and risk come with caveats and opportunities for improvement. We have attempted to standardise the ITN coverage to 1 year (2007) of assessment using 65

Articles

For the Malaria Atlas Project see http://www.map.ox.ac.uk

subnational resolution estimates of ITN use growth rates. For the most part, this process required minor extrapolations, but in a few countries these estimates would have been aﬀected by the timing of the follow-up surveys (table 1). More regular survey data corresponding to changes in delivery modalities is central for improvement of the precision of such temporal interpolation. Additionally, standardised information on coverage of other vector control strategies, such as indoor residual spraying, needs to be generated at the same resolution as ITN to measure the combined eﬀect of these complementary strategies. To generate this information, we need more investment in measurement of progress than is currently available to countries and should be redressed if international agencies are serious about an analysis of whether money is spent where it should be to achieve the intended goals.38 The available evidence suggests that ITNs are similarly eﬀective under a wide range of transmission intensities in averting new infections,22 but the subsequent public-health eﬀect varies,39 and deaths and disease events averted will be highest in communities exposed to high transmission. To assess the public-health impact with changes in ITN coverage will require a more detailed mapping of malaria risk and the eﬀect of seasonality on ITN use at a continental scale. Although not presently available, this work is ongoing as part of the Malaria Atlas Project40 and will provide a more informed map of biological risk with which to plan and assess resource allocation. Deﬁnition of vulnerability and unmet need is central to eﬀective investment strategies by the donor community. Mapping risks, target populations, vulnerability, and coverage provides a means to redress deﬁciencies in the international calls for 80% coverage of ITN by 2015. These targets remain elusive across vast areas of Africa. Increased funding and more informed use of this funding is desperately needed to protect more children in the most vulnerable and most populated areas of Africa. Contributors AMN was responsible for the design of the study, gathering, collation, preparation, analysis, and interpretation of the data, and produced the ﬁnal report. JJM was responsible for gathering and checking the data. AJT provided population projection data. SIH was responsible for interpretation and preparation of the ﬁnal report. RWS was responsible for the conception, overall scientiﬁc management, analysis, interpretation of data, and preparation of the ﬁnal report. Conﬂict of interest statement We declare that we have no conﬂict of interest. Acknowledgements AMN (Research Training Fellowship, 081829), SIH (Senior Research Fellowship, 079091), and RWS (Principal Research Fellowship, 079081) are supported by the Wellcome Trust. This work forms part of the output of the Malaria Atlas Project, mainly funded by the Wellcome Trust, UK. We thank Priscilla Gikandi and Victor Alegana for their assistance during data abstraction and digitisation of administrative boundaries; Alfredo Fort of Opinion Research Corporation-Macro International for his assistance in connecting the authors with national survey teams; Pete Gething and Emelda Okiro for comments on earlier versions of this report. This report is published with the permission of the Director of KEMRI.

66

References 1 Grabowsky M. The billion-dollar question. Nature 2008; 451: 1051–52. 2 Snow RW, Guerra CA, Mutheu J, Hay SI. International funding for malaria control in relation to populations at risk of stable Plasmodium falciparum transmission. PLoS Med 2008; 7: e142. 3 Teklehaimanot A, McCord GC, Sachs JD. Scaling up malaria control in Africa: an economic and epidemiological assessment. Am J Trop Med Hyg 2007; 77: 138–44. 4 UN. Road map towards the implementation of the United Nations millennium declaration. United Nations General Assembly 2001, A/56/326. http://www.un.org/ (accessed July 17, 2007). 5 UNICEF. Malaria and Children. Progress in Intervention Coverage. New York: The United Nations Children’s Fund (UNICEF), 2007. http://www.unicef.org/health/ﬁles/ Malaria_Oct6_for_web(1).pdf (accessed Jan 28, 2008). 6 Roll Back Malaria, World Health Organization and United Nations Children Fund. World Malaria Report. Geneva, Switzerland, 2008. 7 UNICEF, Multiple Cluster Indicator Surveys. http:///www. childinfo.org/mics.html (accessed Sept 10, 2007). 8 Measure-DHS. Demographic and health surveys. http://www. measuredhs.com/accesssurveys/ (accessed July 15, 2007). 9 RBM-MERG. Monitoring and Evaluation Resource Group. Core Household Questionnaire. http://www.rbm.who.int/partnership/ wg/wg_monitoring/docs/mis2005/cc1.pdf (accessed Aug 5, 2008). 10 UNGIWG-SALB. Second Administrative Level Boundaries. http:// www.who.int/whosis/database/gis/salb/salb_PO.htm (accessed July 17, 2008). 11 FAO-GAUL. http://www.fao.org/geonetwork/srv/cn/metadata. show?id=12691 (accessed July 12, 2008). 12 Guerra CA, Gikandi PW, Tatem AJ, et al. The limits and intensity of Plasmodium falciparum transmission: implications for malaria control and elimination worldwide. PLoS Med 2008; 5: e38. 13 Population Services International. Déterminants de l’utilisation des moustiquaires imprégnées au Burundi. République du Burundi, Bujumbura, 2007. 14 Nyarango P, Gebremeskel T, Mebrahtu G, et al. A steep decline of malaria morbidity and mortality trends in Eritrea between 2000 and 2004: the eﬀect of combination of control methods. Malaria J 2006; 5: 33. 15 Kulkarni M, Desrochers R, Goodson J, et al. Evaluation de l’appropriation et l’utilisation des moustiquaires imprégnées d’insecticide à longue durée (MIDs) au Madagascar six mois après la campagne intégrée rougeole/paludisme d’octobre 2007. Rapport Final. Ottawa and Atlanta: HealthBridge and CDC Etats-Unis, 2008. 16 Federal Ministry of Health. Sudan Household Health Survey. Khartoum, Sudan: Federal Ministry of Health, 2007. 17 Hanson K, Marchant T, Mponda H, et al. Monitoring and evaluation of the Tanzania national voucher scheme (TNVS): report on 2006 TNVS household, facility services and facility users surveys (a comparison between baseline and 12 month follow-up). Ifakara Health Research and Development Centre and London School of Hygiene and Tropical Medicine, 2007. 18 Ministry of Health and Social Welfare, Zanzibar. Zanzibar national malaria indicator survey 2007. Zanzibar, Ministry of Health, Zanzibar National Malaria Control Programme, 2008. 19 RBM. Global strategic plan: Roll Back Malaria 2005–2015. Roll Back Malaria Partnership, World Health Organization: Geneva, Switzerland, 2005. 20 DMSP. Nighttime lights of the world – change pair (version 1). http://www.ngdc.noaa.gov/dmsp/download_world_change_pair. html (accessed July 12, 2008). 21 Noor AM, Alegana VA, Gething PW, et al. Using remotely sensed night-time light as a proxy for poverty in Africa. Popul Health Metr 2008; 6: 5. 22 Lengeler C. Insecticide-treated bed nets and curtains for preventing malaria. Cochrane Database Syst Rev 2004; 2: CD000363. 23 Center for International Earth Science Information Network (CIESIN), Columbia University, International Food Policy Research Institute (IFPRI), The World Bank, Centro International de Agricultura Tropical (CIAT) (2007). Global Rural Urban Mapping Project (GRUMP) alpha: Gridded population of the world, version 2, with urban reallocation (GPW-UR). http://sedac.ciesin. columbia.edu/gpw (accessed Sept 8, 2008).

www.thelancet.com Vol 373 January 3, 2009

Articles

24

25

26 27

28 29 30 31

UNPD. World population prospects: the 2006 revision population database. New York: United Nations Population Division (UNDP). http://esa.un.org/unpp/ (accessed Sept 8, 2008). RBM. The Abuja declaration and the plan of action. An extract from the African Summit on Roll Back Malaria, Abuja, April 25, 2000. (WHO/CDS/RBM/2000.17). Geneva: Roll Back Malaria (RBM), 2001. Ministry of Health Angola. Angola National Malaria Indicator Survey 2006-7. Luanda, Angola: Ministry of Health, 2007. Ministry of Health Ethiopia. Ethiopia National Malaria Indicator Survey 2007. Addis Ababa, Ethiopia: Ministry of Health; September, 2008. http://www.moh.gov.et/index.php?option=com_ events&task=view_detail&agid=1&year=2008&month=09&day=02& Itemid=1&catids=29 (accessed Sept 2, 2008). Ministry of Health Kenya. Kenya national malaria indicator survey 2007. Nairobi, Kenya: Ministry of Health, 2008. Ministry of Health Mozambique. Mozambique national malaria indicator survey 2007. Maputo, Mozambique: Ministry of Health, 2007. Ministere de la Sante. Enquête Nationale sur le Paludisme au Sénégal 2006. Dakar, Senegal: Ministere de la Sante, 2007. Ministry of Health Zambia. Zambia National Malaria Indicator Survey 2006. Lusaka, Zambia: Ministry of Health, 2006.

www.thelancet.com Vol 373 January 3, 2009

32 33 34 35

36

37 38 39 40

WHO. The African Summit on Roll Back Malaria, Abuja, April 25, 2000. Geneva, World Health Organization, 2000. Roberts L, Enserink M. Did they really say...eradication? Science 2007; 318: 1544–45. RBM. The global malaria action plan. Roll Back Malaria partnership. Geneva: World Health Organization, 2008. UNICEF. Bednets: supply update. United Nations Children’s Fund (UNICEF) supply division. http://www.rbm.who.int/docs/ llin2008/05.pdf (accessed Sept 4, 2008). Lengeler C, Grabowsky M, McGuire D, de Savigny D. Quick wins versus sustainability: options for the up-scaling of insecticide-treated nets. Am J Trop Med Hyg 2007; 77: 222–26. Teklehaimanot A, Sachs JD, Curtis C. Malaria control needs mass distribution of insecticidal bednets. Lancet 2007; 369: 2143–46. Snow RW. The invisible victims. Nature 2004; 430: 934–35. Snow RW, Marsh K. The consequences of reducing Plasmodium falciparum transmission in Africa. Adv Parasitol 2002; 52: 235–64. Hay SI, Snow RW. The Malaria Atlas Project: developing global maps of malaria risk. PLoS Med; 3: e473.

67

Series

Child Maltreatment 1 Burden and consequences of child maltreatment in high-income countries Ruth Gilbert, Cathy Spatz Widom, Kevin Browne, David Fergusson, Elspeth Webb, Staﬀan Janson Lancet 2009; 373: 68–81 Published Online December 3, 2008 DOI:10.1016/S01406736(08)61706-7 This is the ﬁrst in a Series of four papers about child maltreatment Centre for Evidence-Based Child Health and MRC Centre of Epidemiology for Child Health, UCL Institute of Child Health, London, UK (Prof R Gilbert MD); Psychology Department, John Jay College, City University of New York, NY, USA (Prof C Spatz Widom PhD); Institute of Work, Health and Organisations, University of Nottingham, Nottingham, UK (Prof K Browne PhD); WHO Collaborating Centre on Child Care and Protection, University of Birmingham, Birmingham, UK (Prof K Browne); Christchurch Health and Development Study, Department of Psychological Medicine, Christchurch School of Medicine and Health Sciences, Christchurch, New Zealand (D Fergusson PhD); Department of Child Health, School of Medicine, Cardiﬀ University, Cardiﬀ, UK (E Webb FRCPCH); and Department of Public Health, Karlstad University, Karlstad, Sweden (Prof S Janson DM) Correspondence to: Prof Ruth Gilbert, Centre for Evidence-based Child Health and MRC Centre of Epidemiology for Child Health, UCL Institute of Child Health, 30 Guilford Street, London WC1 1EH, UK [email protected]

68

Child maltreatment remains a major public-health and social-welfare problem in high-income countries. Every year, about 4–16% of children are physically abused and one in ten is neglected or psychologically abused. During childhood, between 5% and 10% of girls and up to 5% of boys are exposed to penetrative sexual abuse, and up to three times this number are exposed to any type of sexual abuse. However, oﬃcial rates for substantiated child maltreatment indicate less than a tenth of this burden. Exposure to multiple types and repeated episodes of maltreatment is associated with increased risks of severe maltreatment and psychological consequences. Child maltreatment substantially contributes to child mortality and morbidity and has longlasting eﬀects on mental health, drug and alcohol misuse (especially in girls), risky sexual behaviour, obesity, and criminal behaviour, which persist into adulthood. Neglect is at least as damaging as physical or sexual abuse in the long term but has received the least scientiﬁc and public attention. The high burden and serious and long-term consequences of child maltreatment warrant increased investment in preventive and therapeutic strategies from early childhood.

Introduction Maltreatment of children by their parents or other caregivers is a major public-health and social-welfare problem in high-income countries. It is common and can cause death, serious injury, and long-term consequences that aﬀect the child’s life into adulthood, their family, and society in general. The 2006 WHO report on prevention of child maltreatment1 drew attention to the need for this topic to achieve the prominence and investment in prevention and epidemiological monitoring that is given to other serious public-health concerns with lifelong consequences aﬀecting children—such as HIV/AIDS, smoking, and obesity—and it recommended expansion of the scientiﬁc evidence base for the magnitude, eﬀects, and preventability of the problem. This Series of four papers critically assesses this expanding evidence base

Key messages • A substantial minority of children in high-income countries are maltreated by their caregivers • Repeated abuse and high levels of neglect mean that for many children maltreatment is a chronic condition • Parental poverty, low educational achievement, and mental illness are often associated with child maltreatment • Child maltreatment has longlasting eﬀects on mental health, drug and alcohol problems, risky sexual behaviour, obesity, and criminal behaviour, from childhood to adulthood • Neglect is at least as damaging as physical or sexual abuse in the long term, but has received the least scientiﬁc and public attention • The high burden and serious, longlasting consequences of child maltreatment warrant increased investment in preventive and therapeutic strategies from early childhood

with the aim of informing policy and practice relating to child maltreatment. We focus mainly on high-income countries and eastern European countries that are in economic transition, since the problem and systems for response diﬀer in low-income and many middle-income countries. In this ﬁrst paper of the Series, we aim to quantify the magnitude of the problem, its determinants, and consequences. The second charts the evidence underpinning recognition and response by professional agencies dealing with children. The third assesses what works for prevention of child maltreatment and associated impairment, and the ﬁnal paper discusses how consideration of children’s rights could enable a more coherent and eﬀective approach to child maltreatment.

Burden of child maltreatment and deﬁnitions Child maltreatment encompasses any acts of commission or omission by a parent or other caregiver that result in harm, potential for harm, or threat of harm to a child

Search strategy and selection criteria We did a comprehensive search of PubMed, Psychinfo, and Education Resources Information Center (ERIC) for any systematic reviews or overviews related to child maltreatment published after 2000 (to June, 2008) and then scrutinised reference lists of relevant studies. We also searched PubMed, ERIC, and Psychinfo using additional synonyms and indexing terms speciﬁc to each outcome. Searches on PubMed were enhanced with the related articles facility for selected studies. Recent psychological abstracts, child abuse and neglect abstracts, and criminal justice abstracts were also searched. We searched websites posted by governments or major advocacy bodies on child maltreatment for reports on incidence and prevalence rates.

www.thelancet.com Vol 373 January 3, 2009

Series

Deﬁnition

Comment

Child Any act of commission or omission by a parent or other caregiver maltreatment* that results in harm, potential for harm, or threat of harm to a child. Harm does not need to be intended

In the USA, 82% of substantiated cases were perpetrated by parents or other caregivers3

Physical abuse*

Intentional use of physical force or implements against a child that results in, or has the potential to result in, physical injury

Includes hitting, kicking, punching, beating, stabbing, biting, pushing, shoving, throwing, pulling, dragging, shaking, strangling, smothering, burning, scalding, and poisoning. 77% of perpetrators were parents according to US ﬁgures for substantiated physical abuse3

Sexual abuse*

Any completed or attempted sexual act, sexual contact, or non-contact sexual interaction with a child by a caregiver†

Penetration: between mouth, penis, vulva, or anus of the child and another individual. Contact: intentional touching directly or through clothing of genitalia, buttocks, or breasts (excluding contact required for normal care). Non-contact: exposure to sexual activity, ﬁlming, or prostitution. For substantiated cases in the USA in 2006, 26% of perpetrators were parents and 29% a relative other than a parent.3 Parents form a smaller percentage (3–5%) of perpetrators of self-reported sexual abuse4

Psychological (or emotional) abuse*

Intentional behaviour that conveys to a child that he/she is worthless, ﬂawed, unloved, unwanted, endangered, or valued only in meeting another’s needs. In the UK, the deﬁnition includes harmful parent–child interactions which are unintentional: “the persistent emotional ill-treatment of a child such as to cause severe and persistent adverse eﬀects on the child’s emotional development”5

Can be continual or episodic—eg, triggered by substance misuse. Can include blaming, belittling, degrading, intimidating, terrorising, isolating, or otherwise behaving in a manner that is harmful, potentially harmful, or insensitive to the child’s developmental needs, or can potentially damage the child psychologically or emotionally. Witnessing intimate-partner violence can be classiﬁed as exposure to psychological abuse. 81% of substantiated cases in the USA were perpetrated by parents3

Neglect*

Failure to meet a child’s basic physical, emotional, medical/dental, or educational needs; failure to provide adequate nutrition, hygiene, or shelter; or failure to ensure a child’s safety

Includes failure to provide adequate food, clothing, or accommodation; not seeking medical attention when needed; allowing a child to miss large amounts of school; and failure to protect a child from violence in the home or neighbourhood or from avoidable hazards. Parents make up 87% of perpetrators of substantiated cases in the USA3

Intimatepartner violence

Any incident of threatening behaviour, violence, or abuse (psychological, physical, sexual, ﬁnancial, or emotional) between adults who are, or have been, intimate partners or family members, irrespective of sex or sexuality

Most frequently the perpetrator is the man in heterosexual couples, but there is growing recognition of violence inﬂicted by women. One community survey reported unanimous agreement that punching, slapping, or forcing a partner to have sex should be regarding as intimate-partner violence, but there was less consensus about emotional or economic abuse

*Deﬁnitions are based on Centers for Disease Control and Prevention report 2008, with modiﬁcations in italics.2 †Includes substitute caregivers in a temporary custodial role (eg, teachers, coaches, clergy, and relatives).

Table 1: Deﬁnitions of child maltreatment

(usually interpreted as up to 18 years of age), even if harm is not the intended result.2 Four forms of maltreatment are widely recognised: physical abuse; sexual abuse; psychological abuse, sometimes referred to as emotional abuse; and neglect. Increasingly, witnessing intimate-partner violence is also regarded as a form of child maltreatment. Consensus deﬁnitions place responsibility for safeguarding children from maltreatment on all caregivers, including teachers, trainers, or child minders (table 1).2 In practice, however, 80% or more of maltreatment is perpetrated by parents or parental guardians, apart from sexual abuse, which is mostly perpetrated by acquaintances or other relatives (table 1). Reliable measurement of the frequency and severity of child maltreatment is not straightforward. We review three types of studies that measure the frequency of maltreatment. The ﬁrst two types are community studies based on self-reports from victims who are old enough to comply with surveys, or studies based on parents reporting severe physical punishment or patterns of care. The third type involves oﬃcial statistics from agencies investigating victims (eg, child-protection services) or police (investigating victims and oﬀenders). All these measures have biases and inconsistencies: thus the prevalence ﬁgures in panel 1 are presented as a range of estimates. Despite the uncertainty of these estimates, the gap between the low rates of maltreatment substantiated by child-protection agencies and the ten-fold higher rates reported by victims or parents www.thelancet.com Vol 373 January 3, 2009

underlines the fact that only a few children who are maltreated receive oﬃcial attention.25–27 Studies that have linked self-reports to oﬃcial statistics for child protection provide direct evidence of under-reporting to agencies. One study reported evidence of contact with childprotection services in only 5% of children who were physically abused and 8% of those sexually abused.26 Another showed that even children who were being monitored by agencies reported four to six times more episodes of abuse than did oﬃcial records.28 The discrepancies between oﬃcial statistics and community studies are even more substantial when examined by age at maltreatment. National statistics from child-protection agencies in the UK and USA show an inverse relation between rate of reports and age for all categories of maltreatment apart from sexual abuse, which is stable across the age range.3,7 Opposite trends have been noted for self-report or parent-report studies in the UK and USA for physical, sexual, or psychological abuse, whereas the prevalence of neglect seems to remain relatively constant.20,27,29 Explanations for these diverging trends include increased risks of under-reporting by parents of younger children, and underdetection of maltreatment by child-protection agencies in older children. Although self-reports or parent reports are probably closer to the true, unobserved rate of maltreatment than are oﬃcial reports to agencies, they might still be underestimates. Biases in self-reports of sexual abuse have been investigated, although problems such as 69

Series

Panel 1: Burden of maltreatment—prevalence of maltreatment in the past year per child population or cumulative prevalence during childhood Agency reports UK (England) • 1·50% of children were estimated to have been referred to social services for abuse (excluding neglect and intimate-partner violence);6 the rate for all social welfare referrals for children (<18 years) in 2007 was 4·96% per year7 • 0·84% of all social welfare referrals were estimated to have been investigated for abuse;6 2·77% of children were investigated in 2007 • 0·30% of children started on a child-protection plan in 2007 (previously child protection registration);7 reports according to primary reason were: neglect 44%, physical abuse 15%, multiple 10%, psychological abuse 23%, and sexual abuse 7% USA • 4·78% of children were investigated in 20063 • 1·21% of children were substantiated in 2006; primary reasons were: neglect 60%, physical abuse 10%, multiple 12%, psychological abuse/unknown 11%, and sexual abuse 7% Canada • 2·15% of children were investigated in 20038 • 0·47% of children remained suspicious8 • 0·97% of children were substantiated; primary reasons were: neglect 38%, physical abuse 23%, psychological abuse 23%, and sexual abuse 9% Australia • 3·34% of children were referred in 2002–039 • 0·68% of children were substantiated; primary reasons were: neglect 34%, physical abuse 28%, psychological abuse 34%, and sexual abuse 10% Self-reported maltreatment or parent-reported perpetration Physical abuse • 3·7–16·3% (5–35% cumulative) of children per year experienced severe parental violence or worse, which is likely to place child at risk of harm; typically included studies classiﬁed hitting with ﬁst/object, kicking, biting, threatening/using a knife/weapon as severe violence (review includes studies in UK, USA, New Zealand, Finland, Italy, and Portugal);10,11 slapping, hitting, and grabbing were classiﬁed as minor violence and are not counted in the ﬁgures shown here • 12·2–29·7% is the yearly prevalence of physical abuse for Macedonia, Moldova, Latvia, and Lithuania12 • 24–29% is the cumulative prevalence of physical abuse for Siberia, Russia, and Romania13,14 Psychological abuse • 10·3% is the yearly prevalence of psychological abuse (verbal abuse by adults or told not wanted; US study)15 • 4–9% is the cumulative prevalence based on categories consistent with severe emotional abuse (studies in Sweden, USA, and UK)16–18 • 12·5–33·3% is the yearly prevalence of severe or moderate psychological abuse reported for four eastern European states (Macedonia, Latvia, Lithuania, and Moldova)12 (Continues on next page)

forgetting, denial, misunderstanding, and embarrassment also apply to other forms of maltreatment.30 All these problems are likely to lead to the under-reporting rather than over-reporting of sexual abuse of children.25,31,32 Test-retest studies have shown modest to moderate agreement between successive self-reports by young 70

adults of sexual or physical abuse several years later (κ coeﬃcient 0·4–0·6) and good agreement is shown for all types of victimisation several weeks later.25,27,33 One study using latent class methods estimated that reported rates of child sexual abuse were roughly half the true but non-observed rate.25 Studies measuring physical abuse in young children use parent reports of physical violence, whereas parent or adolescent self-reports can be used in older children to yield similar estimates.25,27 Comparison between oﬃcial statistics and parent-report studies within a country suggest that only a small proportion of these cases are investigated by child-protection services (panel 1). For example, a systematic review in the UK estimated that around one in 30 children who were physically abused by parents (yearly prevalence 9%) were investigated by social-welfare services responsible for child protection, and only one in 250 children who were physically abused were monitored in accordance with a child-protection plan.10 Measurement of sexual abuse relies on retrospective self-report studies of episodes that are recalled years later by adolescents or adults. Between 5% and 10% of girls and 1% to 5% of boys are exposed to penetrative sexual abuse during childhood, although ﬁgures that include any form of sexual abuse are much higher (panel 1). These estimates are supported by results of a meta-analysis of worldwide studies of variable quality and methodologies,20 but they probably give a lower limit of the true rate of sexual abuse because of underreporting. Few studies have examined the prevalence of psychological abuse. Results from large population-based, self-report studies in the UK and USA showed that 8–9% of women and about 4% of men reported exposure to severe psychological abuse during childhood.16,17 Similar ﬁgures have been recorded for psychological abuse in the past year in boys and girls (10·3%).15 Higher rates have been reported in eastern Europe by similar measures (panel 1).12 Measurement of neglect in the community is diﬃcult, partly because there are many aspects of omission or lack of provision of care that are harmful or could place a child at risk of harm.34 UK and US studies noted that between 1·4% and 10·1% of children or their mothers reported persistent absence of care or instances in which a child was hurt because of insuﬃcient supervision (panel 1). Neglect has received little attention from self-report and parent-report studies despite being the most frequent category of child maltreatment recorded by child-protection agencies (panel 1).3,7 Children who witness intimate-partner violence can be harmed psychologically by witnessing the experience or by being caught up in the violence. The reported prevalence of witnessing intimate-partner violence during childhood ranges from 8–10% in Swedish children aged 15–16 years, who were surveyed in 2000 and 2006, www.thelancet.com Vol 373 January 3, 2009

Series

to 24% reported in a survey of 8600 adult members of a US health maintenance organisation.18,24 The risk of other forms of child maltreatment for witnesses of intimatepartner violence is 30–60%.35,36 Children who are exposed to one type of maltreatment are often exposed to other types on several occasions or continuously. How frequently this abuse occurs is underestimated by oﬃcial reports because recording of more than one type of maltreatment is often discouraged by child-protection agencies and oﬃcial reports often do not capture the chronology of exposure over time. However, retrospective self-report studies consistently show that some children are exposed to more than one type of maltreatment.3,7,15,16,37 This pattern is emphasised by detailed examination of narratives in US childprotection reports of 519 cases of maltreatment, in which high rates of multiple types of maltreatment were reported (36–91% depending on the classiﬁcation used) with emotional abuse rarely occurring alone (1·2%).38 Exposure to multiple types of abuse contributes to high rates of repeated referrals to child-protection services—eg, 22% of children with substantiated maltreatment in the US were re-reported within 24 months,39 with similar rates in the UK (24% within 27 months) and in eight European countries (7–33%).40–42 Factors that consistently aﬀect re-reporting to agencies are primarily ongoing risk factors in the child (such as disability or chronic medical disorders), in the parent (such as alcohol misuse), indices of social adversity (such as low income, contact with services), and multiple or chronic maltreatment, particularly neglect.43 Re-report can also indicate increased surveillance.27,39,42–46 Much less is known from self-report studies about patterns of maltreatment for more than one child in a family. However, an analysis of child-protection referrals in the UK showed that maltreatment was restricted to one speciﬁc child, who was more likely to be abused physically or sexually, in 44% of 310 index cases. Referrals of multiple siblings (56% of cases) were linked to neglect or psychological abuse. Parental diﬃculties and family stressors—such as family conﬂict and separation, drug or alcohol misuse, or family criminality—were associated with maltreatment of all children in the family (37%).47 Throughout childhood, maltreatment by parents or other caregivers merges with other forms of victimisation. In a nationally representative study, Finkelhor and colleagues27,48 noted that the 22% of children aged 2–17 years who had four or more types of victimisation in the previous year—including physical, sexual, or psychological abuse; neglect; or exposure to crime, assault, witnessing intimate-partner violence; or peer or sibling victimisation—were much more likely to be victimised the following year than were those who had fewer types of victimisation, and to have the most serious victimisations and most serious psychological symptomology. Evidence from several studies suggests www.thelancet.com Vol 373 January 3, 2009

(Continued from previous page) Sexual abuse • Cumulative prevalence of any sexual abuse: 15–30% for girls and 5–15% for boys; cumulative prevalence of penetrative sexual abuse: 5–10% for girls and 1–5% for boys (any sexual abuse includes non-contact, contact, or penetrative abuse); ﬁgures are taken from population-based studies in developed countries (Australia, New Zealand, Canada, and USA)4,19 • Similar results were derived in a meta-analysis by Andrews and colleagues20 of studies worldwide (93 for boys and 143 for girls): estimates of childhood prevalence rates were: non-contact sexual abuse (3·1% boys, 6·8% girls); contact sexual abuse (3·7% boys, 13·2% girls); penetrative sexual abuse (1·9% boys, 5·3% girls); and any sexual abuse (8·7% boys, 25·3% girls) Neglect • 1·4–15·4% is the incidence15,21 (6–11·8% cumulative childhood prevalence17,22) of persistent absence of care or provision likely to place a child at risk of harm (eg, not enough food, no medical care when needed, no safe place to stay,15 serious absence of care,17 or in maternal reports—child hurt because of lack of supervision,21 self-report and maternal-report studies from USA and UK) Witnessing intimate-partner violence* • 10–20% is the yearly prevalence estimates based on a review of US community studies by Carlson.23 Few recent studies have been undertaken • 8–25% is the childhood prevalence of witnessing intimate-partner violence—cross-sectional surveys of adolescents and adults18,24 *This category is not included in child-protection reports, therefore not listed in ﬁrst part of panel.

that children who are exposed to one type of maltreatment are at high risk of other types and of repeated exposure over time, and that the frequency of exposure is correlated with the severity of maltreatment.16,24,48,49 For a few children, maltreatment is a chronic condition, not an event.

Determinants of maltreatment Characteristics of the victim Understanding what characteristics of parent–child relationships place children at increased risk of maltreatment within a family is complex and beyond the scope of this review. Girls have a higher risk of being sexually abused than do boys, although rates of other types of maltreatment are similar for both sexes in high-income countries.3,7,20,50 In low-income countries, girls are at higher risk for infanticide, sexual abuse, and neglect, whereas boys seem to be at greater risk of harsh physical punishment.51 Disabled children are at increased risk of maltreatment, although whether their disability is a cause or consequence is uncertain.52–54 A record-linkage study in the USA showed a cumulative prevalence of any maltreatment in 9% of non-disabled children and in 31% of disabled children.52 The overall prevalence of any recorded disability was 8%, but a quarter of all maltreated children had a disability.

Characteristics of the parents and community Identiﬁcation of the separate eﬀects of parental characteristics on the risk of child maltreatment is challenging 71

Series

100

Parent attitude Child attitude Parent-reported violence Child-reported parental violence

90 Proportion of responders (%)

80 70 60 50 40 30 20 10 0 1965

1968

1971

1980 Year of survey

1994

2000

2006

Figure: Time trends in parental violence towards children in Sweden Parental attitudes are based on nationally representative interview surveys (1965, 1968, 1971, 1980) and questionnaire surveys (1994, 2000, 2006). Child attitudes are based on questionnaire surveys of schoolchildren aged 13 and 16 years in 1994, 2000, and 2006. Responses are to the question “Is it right to punish your child physically (including a box on the ear) if they have made you angry” (for children “Is it OK for your parents to hit you if you have made them angry?”). Parental violence is based on parent-reported physical punishment in the past year and child reports on parental violence in preschool years.18

Panel 2: Prevalence of abuse in residential care institutions

See Online for webﬁgures 1 and 2

About 1·3 million children (aged 0–17 years) are in social-care facilities within 20 countries in eastern Europe and the former Soviet Union.71 Physical and sexual abuse by caregivers and peers in these institutions seems to be common.72 In 2000, an anonymous questionnaire study of 3164 children in residential care aged 7–18 years (8% of all children in residential care in Romania) showed that 38% reported severe physical punishment or beatings, usually by residential care staﬀ (in 77% of cases).73 A ﬁfth of respondents (roughly half were boys) claimed to have been blackmailed or coerced into sexual activity, and a further 4% claimed that they were constrained to have sex. The reported perpetrators of these acts of sexual abuse were older residents of the same sex (50%), older residents of the opposite sex (12%), institutional staﬀ (1·3%) oﬀending inside the institution, as well as relatives (4%), other young people (3%), and adults (1%) oﬀending outside the institution. 29% of respondents would not identify their perpetrator. Public scandals involving the sexual exploitation of children in residential care by their carers occur worldwide, with recent examples in Belgium, Portugal, UK, and Ireland.51 However, the consistency of the problem across residential care homes in Romania suggests endemic abuse, which, given that 1·9% of children are in residential care at any one time in that country, represents a major public-health problem.73

since many factors are inextricably clustered. Poverty, mental-health problems, low educational achievement, alcohol and drug misuse, and exposure to maltreatment as a child are strongly associated with parents maltreating their children. The extent to which each of these risk 72

factors is causally related to the occurrence of maltreatment is hard to establish. Risk factors might aﬀect the child diﬀerently depending on the type of maltreatment and might also be linked to the adverse consequences of maltreatment. The ecological model conceptualises maltreatment as multiply determined by forces at work in the individual, in the family, and in the community and culture, and suggests that these determinants modify each other. Thus, parental risk factors can be modiﬁed by the environment and community.55 Nevertheless, some relationships can be generalised. First, income and parental education are risk factors for child maltreatment, although their importance varies with the type of maltreatment.17,22,43,56,57 Second, socioeconomic inequalities are especially steep for deaths from child abuse.58 Third, in the USA, there is controversy about the extent to which ethnic diﬀerences in allegations and substantiation of maltreatment, and in deaths from injury due to maltreatment, are explained by sociodemographic characteristics.48,59–61 However, ethnic diﬀerences in the overall risk of maltreatment are largely explained by sociodemographic characteristics, apart from for children of mixed or multiracial heritage who have an increased risk.22 Fourth, although a clear pathway exists by which parental drug and alcohol problems can cause child maltreatment in individual families, evidence for a causal link within populations is less certain. However, substance misuse is undoubtedly a common factor in incidents involving both spouse and child maltreatment.62 Last, the community environment seems to have a small to moderate eﬀect in addition to family and individual characteristics. A UK cohort study63 reported that individual strengths distinguished resilient from non-resilient children who were exposed to physical abuse under conditions of low but not high family and neighbourhood stress, which was manifested by high crime and low social cohesion, and informal social control. Similarly, a systematic review64 reported that 10% of the variation in child health and adolescent outcomes, including maltreatment, was explained by neighbourhood socioeconomic status and social climate.

Changes over time Evidence suggests that physical and sexual abuse are decreasing in some settings. In the USA, substantiated reports of sexual and physical abuse have fallen by around 50% from the mid-1990s to 2005 (webﬁgure 1),27,50,65 with a similar trend in England (webﬁgure 2).7 These decreases are probably accurate estimates since they are present across both types of abuse with no preponderance of equivocal cases. No analysis of trends in Europe has been done, despite clear evidence, at least in Sweden, of a reduction in acceptance and occurrence of parental violence towards children since the 1960s (ﬁgure).18 Further research is needed to conﬁrm these trends that emphasise the www.thelancet.com Vol 373 January 3, 2009

Series

predominance and continuing problem of neglect and the rise in recognition of psychological abuse, which is often associated with other forms of family violence (webﬁgures 1 and 2).

Prospective studies*

Retrospective studies*

Education and employment Low educational achievement

Moderate

Weak

Low skilled employment

Moderate

Lacking

Diﬀerences between countries

Mental health

Comparisons of the prevalence or incidence of maltreatment between diﬀerent countries need parentreport or self-report studies using similar survey methods. Few such studies have been published. 30 years ago, Gelles and Edfeldt66 reported a 5% higher prevalence of physical abuse in the past year in the USA than in Sweden when the same instrument was used. A meta-regression of self-report studies20 indicates higher rates of sexual abuse in the USA than in Europe (22% vs 15%), although diﬀerences might be partly due to less sensitive survey methods in the European studies. The agency reports for diﬀerent countries in panel 1 are diﬃcult to compare since they reﬂect diﬀerent systems and thresholds. Child maltreatment is a particular concern in the newly independent eastern and central European states, where the economic transition in the past 15 years has been associated with substantial rises in premature adult mortality (panel 1).67,68 Although data are scarce, a questionnaire survey of children aged 10–14 years (n=1145) in Macedonia, Latvia, Lithuania, and Moldova recorded the lowest yearly prevalence rates of severe and moderate psychological abuse and physical abuse in Macedonia (18% and 12%, respectively) and the highest in Moldova (43% and 29%, respectively).12 Abuse was higher in rural areas than in urban areas, and was associated with parental overuse of alcohol.12 Other studies report similar rates of child sexual abuse to those in western Europe.13,69 As in western Europe, by far the greatest problem is neglect. The WHO national prevalence study of child maltreatment in Romanian families showed that physical neglect was reported by 46% of adolescents surveyed, emotional neglect by 44%, and educational neglect by 34%.13 These rates are much higher than are those in western Europe.41 A WHO study in Samara, Russia, reported that the identiﬁcation of neglect by health and social services is seven times more common than is identiﬁcation of physical abuse.70 In two-thirds of all cases of maltreatment, the parents were recorded as alcoholic. The usual response to such cases in 2002 was to place the child into residential or foster care. However, the chances of physical and sexual abuse in residential care are even higher than in family-based care (panel 2).

Behaviour problems as child/adolescent

Strong

Strong

Post-traumatic stress disorder

Strong

Strong

Depression

Moderate

Strong

Attempted suicide

Moderate

Strong

Death from child maltreatment The most tragic manifestation of the burden of child maltreatment is the thousands of child deaths every year due to deliberate killing (homicide) or neglect (manslaughter). WHO estimated that 155 000 deaths in children younger than 15 years occur worldwide every www.thelancet.com Vol 373 January 3, 2009

Self-injurious behaviour

Weak

Weak

Alcohol problems

Moderate

Strong

Drug misuse/dependence

Weak

Strong

Physical health and sexual behaviour Prostitution/sex trading

Moderate

Strong

Teenage pregnancy

Inconsistent

Strong

Promiscuity

No eﬀect

Strong

General adult health

Lacking

Moderate

Chronic pain in adulthood

No eﬀect

Weak

Obesity

Strong

Weak

Health-care use/costs

Lacking

Moderate

Quality of life

Lacking

Lacking

Strong

Strong

Aggression, violence, criminality Criminal behaviour

*Refers to ascertainment of maltreatment. The classiﬁcation indicates consensus about the ﬁndings from included studies and are broadly consistent with the following criteria: strong=evidence of a signiﬁcant eﬀect after adjustment for confounders; moderate=evidence of a signiﬁcant but small eﬀect, or of a stronger eﬀect that is reduced after adjustment for confounders or highly likely to be confounded; weak=evidence of an eﬀect based on methodologically problematic studies or associations that do not persist after adjustment, but consistently favour a positive eﬀect; inconsistent=eﬀect qualitatively diﬀerent across studies (ie, positive and negative or no associations); lacking=no studies addressing this question.

Table 2: Summary of review ﬁndings on consequences of child maltreatment—evidence for an association in prospective and retrospective studies

year as a result of abuse or neglect, which is 0·6% of all deaths and 12·7% of deaths due to any injury.51 Only a third of these deaths are classiﬁed as homicide. Furthermore, substantial under-reporting occurs because of insuﬃcient routine investigations and post-mortem examinations of child deaths in most countries.74 The biological parents are responsible for four-ﬁfths of cases, and step-parents are to blame for most of the remaining cases (15% of the total).74 Child homicide occurs most frequently during infancy—in the UK, 35% of child homicide victims (<16 years) are younger than 1 year.74,75 In infancy, homicide is equally likely to be perpetrated by the mother and the father; however, for older children, the perpetrator is usually a man.75 Large diﬀerences in infant homicide rates exist between high-income countries, with the highest rates recorded in the USA and lowest in Scandinavia and southern Europe.76 An analysis of infant homicide rates between 1945 and 1998 in 39 countries conﬁrmed previously reported associations between infant homicide and higher rates of female participation in the workforce and income inequalities.77 73

Series

See Online for webﬁgure 3

According to WHO estimates, rates of death in children younger than 15 years due to homicide or manslaughter in central and eastern Europe and the newly independent states of the former Soviet Union are consistently higher than in the western European countries of the EU (webﬁgure 3). The peak incidence from 1993 to 2003 coincided with the period of economic and political transition when community services were severely disrupted.68 Despite improvement over the past 30 years in child protection in western European countries and the USA, there has been very little decrease in the rate of child homicides.78,79

Long-term consequences of child maltreatment Since groundbreaking work in the early 1970s drew attention to the battered child syndrome, research designed to quantify the long-term consequences of child maltreatment has grown.80 Here we summarise the evidence for associations between diﬀerent types of maltreatment and outcomes related to education, mental health, physical health, and violence or criminal behaviour. Findings from cohort studies that prospectively ascertained whether children were maltreated or not, and which followed up these children over time to identify later outcomes, are contrasted with more diverse work of cohort and cross-sectional studies that measure maltreatment retrospectively, usually on the basis of self-reporting in adolescence or adulthood. Since we are interested in the consequences of child maltreatment, we want to assess causality. Thus, the strengths of prospective studies include the temporal ordering of maltreatment and subsequent outcomes, objective measurement of maltreatment, avoidance of recall bias, minimisation of selective inclusion of participants on the basis of the outcome, and the opportunity to adjust for social and individual confounding factors as they occur. All these factors are weaknesses of studies using retrospective measurement of maltreatment, especially since the temporal ordering of maltreatment and outcomes cannot be reliably established. Recall bias is also a concern, with ambiguity about whether consequences are due to the actual abuse experience, aftermath of the abuse experience, or a person’s cognitive appraisal of the experience. However, studies that use only oﬃcial cases of child maltreatment might detect only the few maltreated children who come to professional attention, who might diﬀer in some ways from other maltreated children and whose outcomes could also be diﬀerent. The problem of representativeness, which can distort the prevalence and eﬀect size, is reduced for population-based longitudinal cohort studies. The validity of various methods of assessing and studying maltreatment is a source of ongoing debate.81,82 Our analysis endeavours to draw on the strengths of prospective and retrospective studies and, when available, on ﬁndings from systematic reviews (table 2). 74

Education and employment Child maltreatment is associated with long-term deﬁcits in educational achievement. Prospective longitudinal studies have consistently shown that maltreated children have lower educational achievement than do their peers, and are more likely to receive special education83–86 (Jonson-Reid and colleagues83 found that 24% of maltreated children received special education at a mean age of 8 years, compared with 14% of children with no maltreatment record). The diﬀerences are substantial— eg, only 42% of the maltreated children completed high school compared with two-thirds of community-matched controls.85 Another prospective study showed that decreases in school attendance and school performance were related to the timing of maltreatment, and were cumulative.87 Most of these associations persisted after adjustment for family and social characteristics (eg, ethnic origin, age, sex, and socioeconomic status), as seen in some but not all studies. A longitudinal population-based cohort study in New Zealand,86 with retrospective ascertainment of child maltreatment, conﬁrmed these reduced levels of educational achievement in adults who had been physically or sexually abused (eg, 6–10% of abused children attained a university degree compared with 28% of those not abused) but such diﬀerences were largely explained by social, parental, and individual characteristics. Exposure of children to intimate-partner violence also seems to be linked to low educational achievement, but the extent to which this factor is independent of other forms of child maltreatment is unclear.88 Although the risk of underachievement in education is clearly high in children who are maltreated, evidence for a causal link is mixed. Studies are needed from outside the USA to help quantify the extent of this burden in diﬀerent educational settings. Maltreatment has longlasting economic consequences for aﬀected individuals.89 In a prospective study of court documented cases of childhood maltreatment and community-matched controls, signiﬁcantly more of the abused and neglected individuals were in menial and semi-skilled occupations than were controls (62% vs 45%) at 29 years of age, and fewer had remained in employment during the past 5 years (41% vs 58%). Further research is needed to examine the eﬀect of child maltreatment on economic productivity throughout life and in diﬀerent settings.

Mental-health outcomes Child maltreatment increases the risk of behaviour problems, including internalising (anxiety, depression) and externalising (aggression, acting out) behaviour.84,90–95 Children who witness intimate-partner violence are at increased risk of behaviour problems, but whether this factor is independent of other forms of maltreatment is contentious.88,96,97 Behaviour problems in childhood seem to be strongly determined by early timing of maltreatment, www.thelancet.com Vol 373 January 3, 2009

Series

although whether early physical or psychological abuse, or neglect, is most damaging at this age is unclear.90,98 Behaviour problems that arise later in adolescence might be related most strongly to maltreatment during adolescence.91 Consistent evidence suggests a cumulative eﬀect of diﬀerent types of maltreatment on later behaviour problems,91,99 with one group concluding “there is no point beyond which services for children are hopeless…every risk factor we can reduce matters”.99 Maltreated children have a moderately increased risk of depression in adolescence and adulthood (adjusted odds ratios ranging from 1·3 to 2·4), which only partly reﬂects the family context in which maltreatment occurs.84,91,92,95,100–103 Because depression is common and serious—around a quarter to a third of maltreated children meet criteria for major depression by their late 20s (with use of criteria from the Diagnostic and Statistical Manual of Mental Disorders [DSM])92,102,104—this association represents a substantial burden. For many aﬀected individuals, the onset of depression begins in childhood, reinforcing the need for early intervention in the lives of these abused and neglected children, before symptoms of depression cascade into other spheres of functioning.91,102 Depression is associated with neglect and physical and sexual abuse, with no clear evidence for a speciﬁc eﬀect of any particular type of maltreatment. Some investigators have shown a dose response, with depression more likely with harsh or severe physical abuse than with less severe forms of maltreatment.20,92 Evidence suggests that child maltreatment increases the risk of post-traumatic stress disorder, which, by deﬁnition, develops after a terrifying event or ordeal. Symptoms include recurrent intrusion of frightening thoughts and memories, sleep diﬃculties, and detached or numb feelings, which can substantially aﬀect a person’s ability to function. Prospective and retrospective studies consistently show associations between physical or sexual abuse or neglect and post-traumatic stress disorder in adolescents and adults, which persist after controlling for family and child characteristics that are correlated with maltreatment.20,84,95,105–108 These eﬀects can be longlasting. One prospective study105 of children who were maltreated before 12 years of age and assessed at 29 years reported that 23% of people who were sexually abused, 19% of those physically abused, and 17% of those neglected, had a present diagnosis of post-traumatic stress disorder (with use of DSM-III criteria) compared with 10% of controls, and lifetime risks of this disorder were much higher in cases than in controls. However, family, individual, and lifestyle variables, such as having a parent who is an alcoholic or has been arrested, also increased the risk of post-traumatic stress disorder. A meta-analysis20 of studies of children who have been sexually abused suggests a dose-response eﬀect, with higher risks associated with penetrative sexual abuse than with contact or non-contact abuse. www.thelancet.com Vol 373 January 3, 2009

Evidence for an association between childhood maltreatment and adult psychosis is inconclusive.109–111 No clear link between personality disorder and maltreatment has been noted,89 although one prospective study101 showed an increased risk of personality disorder in maltreated children including those exposed to verbal abuse, which was independent of physical or sexual abuse or neglect. These ﬁndings emphasise the need for further research into the eﬀects of psychological abuse. Consistent evidence suggests that both physical abuse and sexual abuse are associated with a doubling of the risk of attempted suicide for young people who are followed up into their late 20s. For physical and sexual abuse, these eﬀects persist after adjustment for confounding family and individual variables,89,92 but for neglect, these eﬀects are mainly explained by family context.100 According to cross-sectional studies, the risk of attempted suicide increases with the accumulation of multiple adversities, including repeated maltreatment and witnessing intimate-partner violence.112,113 The risk of attempted suicide can be very high in young people. Widom and colleagues89 reported lifetime rates of 19% in 29-year-old adults who were abused or neglected as children compared with 8% of community-matched controls, whereas a population-based cohort in New Zealand reported suicide attempts by 11–21% of young adults or adolescents who were exposed to severe physical abuse or penetrative sexual abuse compared with 1–3% of controls.92 Similar rates have been reported in a systematic review of ten studies114 and one prospective study in New York, which showed that 6% of adolescents who were abused made multiple suicide attempts.100 The hypothesis that children who have been sexually abused use self-injurious behaviour (such as cutting) as a maladaptive coping mechanism is only weakly supported by a systematic review of 45 retrospective studies.115 By contrast, a prospective study reported a strong association with sexual abuse but no association with physical abuse or neglect.116 Converging evidence from prospective and retrospective studies suggests that child maltreatment increases the risk of alcohol problems in adolescence and adulthood. These eﬀects are moderate and persist in some but not all studies after adjustment for family characteristics and parental alcohol use.20,22,91,92,102,117–119 On the basis of results from a prospective study with follow up at 29 and 39 years of age,102,117 and from a systematic review of 224 studies,119 the association with alcohol problems, at least in adulthood, is conﬁned to girls. These ﬁndings emphasise the need for interventions for girls and young women to prevent the development of alcohol problems and the associated health, safety, and social problems that excessive drinking in women can cause. For example, problem drinking in women increases the risk of fetal alcohol syndrome and might aﬀect their ability to look after a child.120 75

Series

See Online for webappendix

The link between child maltreatment and drug dependency is not straightforward.22,84,91,92,121 One prospective study122 reported that individuals who were maltreated in childhood were no more likely to have a diagnosis of drug dependency by the age of 29 years than were community controls. However, when a diﬀerent measure of drug use is used, individuals who were abused and neglected were at increased risk for present illicit drug use at roughly 40 years of age.121 Investigators of this study speculated that although individuals who had experienced neglect or abuse would mature out of drug use, abused and neglected individuals might continue in a problematic drug-use trajectory. Crosssectional studies indicate that exposure to multiple forms of abuse and other childhood adversities, including witnessing intimate-partner violence, leads to a cumulative increase in the risk of self-reported alcohol or drug misuse in adulthood.123,124 Overall, the burden of mental ill health resulting from child maltreatment is substantial. A New Zealand cohort study92 estimated that physical abuse accounted for 5% of mental disorders and sexual abuse for 13%, after taking account of the family context in which maltreatment occurs. How exposure to maltreatment of diﬀerent types, at diﬀerent developmental stages, leads to adverse mental-health outcomes is complex, although early and cumulative maltreatment seem to be particularly harmful to the development of the brain.125,126 The webappendix summarises the evidence for biological mechanisms that link child maltreatment and later outcomes.

Physical-health outcomes Four very diﬀerent prospective longitudinal studies127–130 have reported strong associations between physical abuse, neglect, and sexual abuse and obesity, which persist after accounting for family characteristics and individual risk factors, such as childhood obesity. Large diﬀerences in the magnitude of this association between studies (adjusted odds ratios range from 1·3 to 9·8)129,130 probably indicate diﬀerences in exposure and outcome measures and analyses. Retrospective studies also suggest an association between child sexual abuse and eating disorders (eg, bulimia and anorexia), but there is less information about other forms of maltreatment.131 Several large cross-sectional studies have reported relations between multiple child adversities, including child maltreatment, and a range of health outcomes in adulthood (eg, ischaemic heart disease, cancer, chronic lung disease, skeletal fractures, and liver disease), albeit with little adjustment for lifetime confounders.132,133 Abnormally overt or intrusive sexualised behaviour is a common problem in preteen children who are exposed to sexual abuse.134 However, sexualised behaviour is not speciﬁc to child sexual abuse and has been associated with physical abuse, characteristics of family adversity, 76

coercive parenting, child behaviour, and modelling of sexual behaviour.135 Most studies that have examined the relation between child maltreatment and sexual behaviour in adolescence and adulthood have focused on outcomes for sexual abuse. An exception is a prospective study with follow-up at 29 years of age, which reported a signiﬁcant association between physical or sexual abuse or neglect and arrest for prostitution or being paid for sex (13% of cases vs 4% of controls for girls, p=0·001; 15% vs 8% for boys, p=0·17), but no signiﬁcant associations with promiscuity or teenage pregnancy.136 In two prospective studies,91,137 child maltreatment was associated with teenage pregnancy. In one study,136 HIV was twice as common in abused and neglected individuals as in controls, although the diﬀerence did not reach conventional levels of signiﬁcance most likely because of weak statistical power.136 A systematic review and meta-analysis of various types of study, most with retrospective ascertainment of abuse status, similarly reported the strongest associations between child sexual abuse and sex trading in adolescence or adulthood, and showed greater eﬀects for women than for men.112,138–140 Small to moderate eﬀects of child sexual abuse on increased rates of teenage pregnancy have been noted, as well as earlier onset of sexual activity, greater numbers of sexual partners, increased rates of abortion, and increased risks of sexually transmitted disease.4,138,140–145 These eﬀects are stronger with more severe146,147 or repeated145 sexual abuse or exposure to multiple childhood adversities.148,149 Emerging evidence also suggests that exposure to child sexual abuse might be related to later sexual orientation.150 Overall, these ﬁndings suggest associations between exposure to child sexual abuse and subsequent sexual adjustment. Controversy about a possible link between childhood maltreatment and chronic pain in adulthood emphasises the diﬀerences between prospective and retrospective measures of child maltreatment and the advantages of considering both types of study design. A prospective study based on children with maltreatment documented by courts and community-matched controls showed no association with chronic pain reported in adulthood at 29 years of age.151 However, when groups were compared on the basis of retrospective self-reports of child maltreatment, the association with chronic pain was signiﬁcant (p<0·0001).152 Similar evidence of a modest association between child sexual or physical abuse (but not neglect), and pain in adulthood has been reported.151,153–156 These ﬁndings draw attention to the distinction between how people remember and interpret abusive childhood experiences and exposure to child abuse. They establish an association between memories of childhood abuse and chronic pain in adulthood and further suggest that abused individuals with chronic pain are more likely to seek health care than are non-abused individuals with chronic pain.151 However, we cannot conclude that child abuse or neglect causes chronic pain in adulthood. www.thelancet.com Vol 373 January 3, 2009

Series

Despite the evidence for diverse and serious consequences of child maltreatment, a systematic review157 found no studies measuring quality of life during childhood after maltreatment, and only four studies in adults. Further research, based on modiﬁcation of existing methods and development of measures that can be used for younger children, is needed for economic assessments of the burden of child maltreatment and cost-eﬀectiveness of intervention strategies. Studies in North America158,159 and Australia160 have shown increased service use and costs associated with child maltreatment, but research is lacking elsewhere in the world and in other public sectors.

Aggression, crime, and violence In addition to feeling considerable pain and suﬀering themselves, abused and neglected children are at increased risk of becoming aggressive and inﬂicting pain and suﬀering on others, often perpetrating crime and violence. One paper on the cycle of violence161 reported that being physically abused or neglected as a child increased the likelihood of arrest as a juvenile (31% arrested vs 19% of community-matched controls) and as an adult (48% vs 36%). Since that time, similar eﬀects on criminal behaviour have been reported in the USA despite diﬀerences in geographical region, time period, age of adolescent, deﬁnition of maltreatment, and assessment technique.95,137,162–167 These ﬁndings are supported by systematic reviews of retrospective studies, showing that physical and sexual abuse predict delinquency or violence in boys and girls,168 although physical abuse might be most strongly related to youth violence in girls.169 A direct comparison of diﬀerent types of maltreatment found that children who were physically or sexually abused were more likely to carry a weapon in adolescence than were neglected children, because of a perceived need for self protection.170 Evidence that risks of youth violence cumulate when child abuse persists into adolescence suggests a need for interventions to prevent ongoing abuse.169

Future research Child maltreatment is common, and for many it is a chronic condition, with repeated and ongoing maltreatment merging into adverse outcomes throughout childhood and into adulthood. The burden on the children themselves and on society is substantial. At the same time, variation in rates of maltreatment between countries, particularly for infant homicides, and a possible decrease in recent years in sexual and physical abuse in some high-income countries, shows that the present high burden of child maltreatment is not inevitable. International comparative studies are needed, especially in countries outside North America and northern Europe, to help learn lessons from diﬀerent settings about how to prevent child maltreatment and its consequences. The high burden and serious and longlasting consequences of child maltreatment warrant www.thelancet.com Vol 373 January 3, 2009

increased investment in preventive and therapeutic strategies from early childhood. Research into what works at an individual and policy level is a priority.171,172 More research is needed into characteristics of responses by communities, families, and services that help with healthy development rather than exacerbate the child’s problems. This research includes improved understanding of the many ways in which children are victimised at diﬀerent stages of development.27 More attention needs to be given to neglected children. There is mounting evidence that the consequences of childhood neglect can be as damaging—or perhaps even more damaging—to a child than physical or sexual abuse. More attention also needs to be paid to the potentially diﬀerent needs of boys and girls who are maltreated. Although classrooms and neighbourhoods are disrupted more by deviant behaviour of boys than of girls, research shows that maltreatment doubles a girl’s risk of being arrested for a violent crime and increases risk for subsequent alcohol and drug problems, with implications for her children. Conﬂict of interest statement We declare that we have no conﬂict of interest. Acknowledgments We thank the following people who helped provide data, references, or undertook searches for the review: Maria Keller-Hamela, Nobody’s Children Foundation, Warsaw, Poland; Dinesh Seth, WHO Rome Oﬃce, Violence programme; Helen Wadsworth Wilson, City University of New York; and Melissa Harden, UCL-Institute of Child Health, London. We thank Toni Pitcher, Christchurch Health and Development Study, University Otago, New Zealand, for contributing to the web panel on biological mechanisms; and the editorial group for the Series: Rosalyn Proops, Richard Reading, Harriet MacMillan, Danya Glaser, and Pat Hamilton, for commenting on drafts of the review. References 1 Butchart A, Kahane T, Phinney Harvey A, Mian M, Furniss T. Preventing child maltreatment: a guide to taking action and generating evidence. Geneva: WHO and International Society for the Prevention of Child Abuse and Neglect, 2006. 2 Leeb RT, Paulozzzi L, Melanson C, Simon T, Arias I. Child maltreatment surveillance. Uniform deﬁnitions for public health and recommended data elements. Atlanta: Centers for Disease Control and Prevention, 2008. 3 US Department of Health and Human Services, Administration on Children youth and Families. Child Maltreatment 2006. Washington, DC: US Government Printing Oﬃce, 2008. 4 Fergusson DM, Mullen PE. Childhood sexual abuse—an evidence based perspective. Thousand Oaks: Sage, 1999. 5 HM Government. Working together to safeguard children. A guide to interagency working to promote and safeguard the welfare of children. London: The Stationary Oﬃce, 2006. http://www. everychildmatters.gov.uk/_ﬁles/AE53C8F9D7AEB1B23E403514A6C 1B17D.pdf (accessed Oct 16, 2008). 6 Cleaver H, Walker S. Assessing children’s needs and circumstances. London: Jessica Kingsley Publishers, 2004. 7 Department for Children, Schools and Families. Referrals, assessments and children and young people who are the subject of a child protection plan or are on child protection registers: year ending 31 March 2007. London: Department for Children, Schools and Families, 2008. 8 Trocme N, MacMillan H, Fallon B, Marco RD. Nature and severity of physical harm caused by child abuse and neglect: results from the Canadian Incidence Study. Can Med Assoc J 2003; 169: 911–15. 9 Australian Institute of Health and Welfare. Australia’s health 2004. Canberra: AIHW, 2004.

77

Series

10

11

12

13

14

15

16

17

18

19

20

21 22

23

24

25

26

27 28

29

30 31 32

33

78

Woodman J, Pitt M, Wentz R, Taylor B, Hodes D, Gilbert RE. Performance of screening tests for child physical abuse in Accident and Emergency Departments. Health Technol Assess 2008; 12: 1–118. Machado C, Goncalves M, Matos M, Dias AR. Child and partner abuse: self-reported prevalence and attitudes in the north of Portugal. Child Abuse Negl 2007; 31: 657–70. Sebre S, Sprugevica I, Novotni A, et al. Cross-cultural comparisons of child-reported emotional and physical abuse: rates, risk factors and psychosocial symptoms. Child Abuse Negl 2004; 28: 113–27. Browne KD. National prevalence study of child abuse and neglect in Romanian families. Copenhagen: WHO Regional Oﬃce for Europe, 2002. Berrien FB, Aprelkov G, Ivanova T, Zhmurov V, Buzhicheeva V. Child abuse prevalence in Russian urban population: a preliminary report. Child Abuse Negl 1995; 19: 261–64. Finkelhor D, Ormrod R, Turner H, Hamby SL. The victimization of children and youth: a comprehensive, national survey. Child Maltreat 2005; 10: 5–25. Edwards VJ, Holden GW, Felitti VJ, Anda RF. Relationship between multiple forms of childhood maltreatment and adult mental health in community respondents: results from the adverse childhood experiences study. Am J Psychiatry 2003; 160: 1453–60. May-Chahal C, Cawson P. Measuring child maltreatment in the United Kingdom: a study of the prevalence of child abuse and neglect. Child Abuse Negl 2005; 29: 969–84. Janson S, Langberg B, Svensson B. Violence against children in Sweden. A national survey 2006–2007 (in Swedish). Stockholm: Allmanna Barnhuset and Karlstad University, 2007. Nelson EC, Heath AC, Madden PAF, et al. Association between self-reported childhood sexual abuse and adverse psychosocial outcomes: results from a twin study. Arch Gen Psychiatry 2002; 59: 139–45. Andrews G, Corry J, Slade T, Issakidis C, Swanston H. Child sexual abuse. Comparative quantiﬁcation of health risks. Geneva: WHO, 2004. Theodore A, Chang JJ, Runyan D. Measuring the risk of physical neglect in a population-based sample. Child Maltreat 2007; 12: 96–105. Hussey JM, Chang JJ, Kotch JB. Child maltreatment in the United States: prevalence, risk factors, and adolescent health consequences. Pediatrics 2006; 118: 933–42. Carlson BE. Children exposed to intimate partner violence: research ﬁndings and implications for intervention. Trauma Violence Abuse 2000; 1: 321–40. Dong M, Anda RF, Felitti VJ, et al. The interrelatedness of multiple forms of childhood abuse, neglect, and household dysfunction. Child Abuse Negl 2004; 28: 771–84. Fergusson DM, Horwood LJ, Woodward LJ. The stability of child abuse reports: a longitudinal study of the reporting behaviour of young adults. Psychol Med 2000; 30: 529–44. MacMillan HL, Jamieson E, Walsh CA. Reported contact with child protection services among those reporting child physical and sexual abuse: results from a community survey. Child Abuse Negl 2003; 27: 1397–408. Finkelhor D. Childhood victimization. Violence, crime and abuse in the lives of young people. Oxford: Oxford University Press, 2008. Everson MD, Smith JB, Hussey JM, et al. Concordance between adolescent reports of childhood abuse and Child Protective Service determinations in an at-risk sample of young adolescents. Child Maltreat 2008; 13: 14–26. Ghate D, Creighton SJ, Field J. A national study of parents, children and discipline. Swindon: Economic and Social Reserach Council, 2002. Melchert TP, Parker RL. Diﬀerent forms of childhood abuse and memory. Child Abuse Negl 1997; 21: 125–35. Widom CS, Morris S. Accuracy of adult recollections of childhood victimization. Childhood sexual abuse. Psychol Assess 1997; 9: 34–46. Hardt J, Rutter M. Validity of adult retrospective reports of adverse childhood experiences: review of the evidence. J Child Psychol Psychiatry 2004; 45: 260–73. Bifulco A, Brown GW, Lillie A, Jarvis J. Memories of childhood neglect and abuse: corroboration in a series of sisters. J Child Psychol Psychiatry 1997; 38: 365–74.

34

35 36

37

38

39

40

41 42 43 44

45

46 47 48

49

50

51

52

53 54

55

56

57 58

59

Straus MA, Kantor GK. Deﬁnition and measurement of neglectful behavior: some principles and guidelines. Child Abuse Negl 2005; 29: 19–29. Appel A, Holden E. The co-occurrence of spouse and physical child abuse: a review and appraisal. J Fam Psychol 1998; 12: 578–99. Herrenkohl TI, Sousa C, Tajima EA, Herrenkohl RC, Moylan CA. Intersection of child abuse and children’s exposure to domestic violence. Trauma Violence Abuse 2008; 9: 84–99. MacMillan HL, Fleming JE, Trocme N, et al. Prevalence of child physical and sexual abuse in the community. Results from the Ontario Health Supplement. JAMA 1997; 278: 131–35. Lau AS, Leeb RT, English D, et al. What’s in a name? A comparison of methods for classifying predominant type of maltreatment. Child Abuse Negl 2005; 29: 533–51. Fluke JD, Shusterman GR, Hollinshead DM, Yuan YY. Longitudinal analysis of repeated child abuse reporting and victimization: multistate analysis of associated factors. Child Maltreat 2008; 13: 76–88. Hamilton CE, Browne KD. Recurrent maltreatment during childhood: a survey of referrals to police and child protection units in England. Child Maltreat 1999; 4: 275–86. May-Chahal C, Bertotti T, Di Blasio P, et al. Child maltreatment in the family: a European perspective. Eur J Social Work 2006; 9: 3–20. Hindley N, Ramchandani PG, Jones DP. Risk factors for recurrence of maltreatment: a systematic review. Arch Dis Child 2006; 91: 744–52. Bae H-O, Solomon P, Gelles RJ. Abuse type and substantiation status varying by recurrence. Child Youth Serv Rev 2007; 29: 865–69. Drake B, Jonson-Reid M, Sapokaite L. Re-reporting of child maltreatment: does participation in other public sector services moderate the likelihood of a second maltreatment report? Child Abuse Negl 2006; 30: 1201–26. Classen CC, Palesh OG, Aggarwal R. Sexual revictimization: a review of the empirical literature. Trauma Violence Abuse 2005; 6: 103–29. Widom CS, Czaja SJ, Dutton MA. Childhood victimization and lifetime revictimization. Child Abuse Negl 2008; 32: 785–96. Hamilton-Giachritsis CE, Browne KD. A retrospective study of risk to siblings in abusing families. J Fam Psychol 2005; 19: 619–24. Finkelhor D, Ormrod RK, Turner HA. Re-victimization patterns in a national longitudinal sample of children and youth. Child Abuse Negl 2007; 31: 479–502. Clemmons JC, Walsh K, DiLillo D, Messman-Moore TL. Unique and combined contributions of multiple child abuse types and abuse severity to adult trauma symptomatology. Child Maltreat 2007; 12: 172–81. US Department of Health and Human Services. The third national incidence study of child abuse and neglect (NIS-3). Washington, DC: National Clearing House on Child Abuse and Neglect, 2006. Pinheiro PS. World report on violence against children. New York: United Nations Secretary-General’s study on violence against children, 2006. Sullivan PM, Knutson JF. Maltreatment and disabilities: a population-based epidemiological study. Child Abuse Negl 2000; 24: 1257–73. Fisher M, Hodapp R, Dykens E. Child abuse among children with disabilities. Int Rev Res Ment Retard 2008; 35: 251–89. Govindshenoy M, Spencer N. Abuse of the disabled child: a systematic review of population-based studies. Child Care Health Dev 2007; 33: 552–58. Sidebotham P. An ecological approach to child abuse: creative use of scientiﬁc models in research and practice. Child Abuse Rev 2001; 10: 97–112. Sidebotham P, Heron J, Golding J. Child maltreatment in the “Children of the Nineties:” deprivation, class, and social networks in a UK sample. Child Abuse Negl 2002; 26: 1243–59. Berger LM. Income, family characteristics, and physical violence toward children. Child Abuse Negl 2005; 29: 107–33. Roberts I, Li L, Barker M. Trends in intentional injury deaths in children and teenagers (1980–1995). J Public Health Med 1998; 20: 463–66. Ards SD, Chung C, Myers SL Jr. Sample selection bias and racial diﬀerences in child abuse reporting: once again. Child Abuse Negl 2001; 25: 7–12.

www.thelancet.com Vol 373 January 3, 2009

Series

60

61

62

63

64

65 66 67

68

69 70

71

72 73 74

75

76

77

78

79

80 81

82

83

84

Flaherty EG, Sege RD, Griﬃth J, et al. From suspicion of physical child abuse to reporting: primary care clinician decision-making. Pediatrics 2008; 122: 611–19. Falcone RA Jr, Brown RL, Garcia VF. Disparities in child abuse mortality are not explained by injury severity. J Pediatr Surg 2007; 42: 1031–36. Ondersma SJ. Introduction to the second special section on substance abuse and child maltreatment. Child Maltreat 2007; 12: 111–13. Jaﬀee SR, Caspi A, Moﬃtt TE, Polo-Tomas M, Taylor A. Individual, family, and neighborhood factors distinguish resilient from non-resilient maltreated children: a cumulative stressors model. Child Abuse Negl 2007; 31: 231–53. Sellstrom E, Bremberg S. The signiﬁcance of neighbourhood context to child and adolescent health and well-being: a systematic review of multilevel studies. Scand J Public Health 2006; 34: 544–54. Jones L, Finkelhor D. The decline in child sexual abuse cases. Washington, DC: United States Department of Justice, 2001. Gelles RJ, Edfeldt AW. Violence toward children in the United States and Sweden. Child Abuse Negl 1986; 10: 501–10. McKee M, Zwi A, Koupilova I, Sethi D, Leon D. Health policy-making in central and eastern Europe: lessons from the inaction on injuries? Health Policy Plan 2000; 15: 263–69. Walberg P, McKee M, Shkolnikov V, Chenet L, Leon DA. Economic change, crime, and mortality crisis in Russia: regional analysis. BMJ 1998; 317: 312–18. Herczog M, May-Chahal C. Child sexual abuse in Europe: an overview. Strasbourg: Council of Europe, 2002. Ostergren M, Bacchi A, Browne KD. Improving maternal infant and child health in the Russian Federation: a joint DFID/WHO project. Copenhagen: WHO Regional Oﬃce for Europe, 2003. Carter R. Family matters: a study of institutional childcare in Central and Eastern Europe and the Former Soviet Union. London: Everychild, 2005. Hunt K. Abandoned to the state: cruelty and neglect in Russian orphanages. USA: Human Rights Watch, 1998. UNICEF. Child abuse in residential care in institutions in Romania. Bucharest: UNICEF, 2001. UNICEF. A league table of child maltreatment deaths in rich nations. Innocenti Report Card number 5. Florence: UNICEF Innocenti Research Centre, 2003. Brookman F, Nolan J. The dark ﬁgure of infanticide in England and Wales: complexities of diagnosis. J Interpers Violence 2006; 21: 869. Creighton S. Prevalence and incidence of child abuse: international comparisons. London: NSPCC Information Brieﬁngs, 2004. Hunnicutt G, LaFree G. Reassessing the structural covariates of cross-national infant homicide victimization. Homicide Studies 2008; 12: 46–66. Fox JA, Zawitz JA. Homicide trends in the United States. Washington DC: US Department of Justice, 2007. http://www.ojp. usdoj.gov/bjs/homicide/teens.htm (accessed Oct 14, 2008). WHO Regional Oﬃce for Europe. Health for All database (HFA-DB). Copenhagen: WHO Regional Oﬃce for Europe, 2008. http://www.euro.who.int/hfadb (accessed Oct 14, 2008). Behl LE, Conyngham HA, May PF. Trends in child maltreatment literature. Child Abuse Negl 2003; 27: 215–29. Widom CS, Raphael KG, DuMont KA. The case for prospective longitudinal studies in child maltreatment research: commentary on Dube, Williamson, Thompson, Felitti, and Anda (2004). Child Abuse Negl 2004; 28: 715–22. Kendall-Tackett K, Becker-Blease K. The importance of retrospective ﬁndings in child maltreatment research. Child Abuse Negl 2004; 28: 723–27. Jonson-Reid M, Drake B, Kim J, Porterﬁeld S, Han L. A prospective analysis of the relationship between reported child maltreatment and special education eligibility among poor children. Child Maltreat 2004; 9: 382–94. Lansford JE, Dodge KA, Pettit GS, Bates JE, Crozier J, Kaplow J. A 12-year prospective study of the long-term eﬀects of early child physical maltreatment on psychological, behavioral, and academic problems in adolescence. Arch Pediatr Adolesc Med 2002; 156: 824–30.

www.thelancet.com Vol 373 January 3, 2009

85

86

87 88

89

90

91

92

93

94

95

96

97

98 99

100

101

102

103 104

105 106

Perez CM, Widom CS. Childhood victimization and long-term intellectual and academic outcomes. Child Abuse Negl 1994; 18: 617–33. Boden JM, Horwood LJ, Fergusson DM. Exposure to childhood sexual and physical abuse and subsequent educational achievement outcomes. Child Abuse Negl 2007; 31: 1101–14. Leiter J. Child maltreatment and school performance declines: an event-history analysis. Am Educ Res J 1997; 34: 563–89. Kitzmann KM, Gaylord NK, Holt AR, Kenny ED. Child witnesses to domestic violence: a meta-analytic review. J Consult Clin Psychol 2003; 71: 339–52. Widom CS. Childhood victimization: early adversity and subsequent psychopathology. In: Dohrenwend BP, ed. Adversity, stress, and psychopathology. New York: Oxford University Press, 1998: 81–95. Manly JT, Kim JE, Rogosch FA, Cicchetti D. Dimensions of child maltreatment and children’s adjustment: contributions of developmental timing and subtype. Dev Psychopathol 2001; 13: 759–82. Thornberry TP, Ireland TO, Smith CA. The importance of timing: the varying impact of childhood and adolescent maltreatment on multiple problem outcomes. Dev Psychopathol 2001; 13: 957–79. Fergusson DM, Boden JM, Horwood LJ. Exposure to childhood sexual and physical abuse and adjustment in early adulthood. Child Abuse Negl 2008; 32: 607–19. Herrenkohl EC, Herrenkohl RC, Rupert LJ, Egolf BP, Lutz JG. Risk factors for behavioral dysfunction: the relative impact of maltreatment, SES, physical health problems, cognitive ability, and quality of parent-child interaction. Child Abuse Negl 1995; 19: 191–203. Herrenkohl TI, Herrenkohl RC. Examining the overlap and prediction of multiple forms of child maltreatment, stressors, and socioeconomic status: a longitudinal analysis of youth outcomes. J Family Violence 2007; 22: 553–62. Banyard VL, Williams LM, Siegel JA. The long-term mental health consequences of child sexual abuse: an exploratory study of the impact of multiple traumas in a sample of women. J Trauma Stress 2001; 14: 697–715. Yates TM, Dodds MF, Sroufe LA, Egeland B. Exposure to partner violence and child behavior problems: a prospective study controlling for child physical abuse and neglect, child cognitive ability, socioeconomic status, and life stress. Dev Psychopathol 2003; 15: 199–218. Sternberg KJ, Lamb ME, Guterman E, Abbott CB. Eﬀects of early and later family violence on children’s behavior problems and depression: a longitudinal, multi-informant perspective. Child Abuse Negl 2006; 30: 283–306. Kotch JB, Lewis T, Hussey JM, et al. Importance of early neglect for childhood aggression. Pediatrics 2008; 121: 725–31. Appleyard K, Egeland B, van Dulman MH, Sroufe LA. When more is not better: the role of cumulative risk in child behavior outcomes. J Child Psychol Psychiatry 2005; 46: 235–45. Brown J, Cohen P, Johnson JG, Smailes EM. Childhood abuse and neglect: speciﬁcity of eﬀects on adolescent and young adult depression and suicidality. J Am Acad Child Adolesc Psychiatry 1999; 38: 1490–96. Johnson JG, Cohen P, Smailes EM, Skodol AE, Brown J, Oldham JM. Childhood verbal abuse and risk for personality disorders during adolescence and early adulthood. Compr Psychiatry 2001; 42: 16–23. Widom CS, White HR, Czaja SJ, Marmorstein NR. Long-term eﬀects of child abuse and neglect on alcohol use and excessive drinking in middle adulthood. J Stud Alcohol Drugs 2007; 68: 317–26. Noll JG, Trickett PK, Susman EJ, Putnam FW. Sleep disturbances and childhood sexual abuse. J Pediatr Psychol 2006; 31: 469–80. Widom CS, Dumont KA, Czaja SJ. A prospective investigation of major depressive disorder and comorbidity in abused and neglected children grown up. Arch Gen Psychiatry 2007; 64: 49–56. Widom CS. Posttraumatic stress disorder in abused and neglected children grown up. Am J Psychiatry 1999; 156: 1223–29. Brewin CR, Andrews B, Valentine JD. Meta-analysis of risk factors for posttraumatic stress disorder in trauma-exposed adults. J Consult Clin Psychol 2000; 68: 748–66.

79

Series

107 Tolin DF, Foa EB. Sex diﬀerences in trauma and posttraumatic stress disorder: a quantitative review of 25 years of research. Psychol Bull 2006; 132: 959–92. 108 Whiﬀen V, Macintosh H. Mediators of the link between childhood sexual abuse and emotional distress: a critical review. Trauma Violence Abuse 2005; 6: 24–39. 109 Read J, van Os J, Morrison AP, Ross CA. Childhood trauma, psychosis and schizophrenia: a literature review with theoretical and clinical implications. Acta Psychiatr Scand 2005; 112: 330–50. 110 Morgan C, Fisher H. Environment and schizophrenia: environmental factors in schizophrenia: childhood trauma—a critical review. Schizophr Bull 2007; 33: 3–10. 111 Shevlin M, Houston JE, Dorahy MJ, Adamson G. Cumulative traumas and psychosis: an analysis of the national comorbidity survey and the British Psychiatric Morbidity Survey. Schizophr Bull 2008; 34: 193–99. 112 Aﬁﬁ TO, Enns MW, Cox BJ, Asmundson GJG, Stein MB, Sareen J. Population attributable fractions of psychiatric disorders and suicide ideation and attempts associated with adverse childhood experiences. Am J Public Health 2008; 98: 946–52. 113 McHolm AE, MacMillan HL, Jamieson E. The relationship between childhood physical abuse and suicidality among depressed women: results from a community sample. Am J Psychiatry 2003; 160: 933–38. 114 Evans E, Hawton K, Rodham K. Suicidal phenomena and abuse in adolescents: a review of epidemiological studies. Child Abuse Negl 2005; 29: 45–58. 115 Klonsky ED, Moyer A. Childhood sexual abuse and non-suicidal self-injury: meta-analysis. Br J Psychiatry 2008; 192: 166–70. 116 Yates TM, Carlson EA, Egeland B. A prospective study of child maltreatment and self-injurious behavior in a community sample. Dev Psychopathol 2008; 20: 651–71. 117 Widom CS, Ireland T, Glynn PJ. Alcohol abuse in abused and neglected children followed-up: are they at increased risk? J Stud Alcohol 1995; 56: 207–17. 118 Widom CS, Hiller-Sturmhofel S. Alcohol abuse as a risk factor for and consequence of child abuse. Alcohol Res Health 2001; 25: 52–57. 119 Simpson TL, Miller WR. Concomitance between childhood sexual and physical abuse and substance use problems. A review. Clin Psychol Rev 2002; 22: 27–77. 120 Streissguth AP. A long-term perspective of FAS. Alcohol Health Res World 1994; 18: 74–81. 121 Widom CS, Marmostein NR, White HR. Childhood victimization and illicit drug use in middle adulthood. Psychol Addict Behav 2006; 20: 394–403. 122 Widom CS, Weiler BL, Cottler LB. Childhood victimization and drug abuse: a comparison of prospective and retrospective ﬁndings. J Consult Clin Psychol 1999; 67: 867–80. 123 Dube SR, Anda RF, Felitti VJ, Edwards VJ, Williamson DF. Exposure to abuse, neglect, and household dysfunction among adults who witnessed intimate partner violence as children: implications for health and social services. Violence Vict 2002; 17: 3–17. 124 Bair-Merritt MH, Blackstone M, Feudtner C. Physical health outcomes of childhood exposure to intimate partner violence: a systematic review. Pediatrics 2006; 117: e278–90. 125 Glaser D. Child abuse and neglect and the brain—a review. J Child Psychol Psychiatry 2000; 41: 97–116. 126 Lee V, Hoaken PN. Cognition, emotion, and neurobiological development: mediating the relation between maltreatment and aggression. Child Maltreat 2007; 12: 281–98. 127 Johnson JG, Cohen P, Kasen S, Brook JS. Childhood adversities associated with risk for eating disorders or weight problems during adolescence or early adulthood. Am J Psychiatry 2002; 159: 394–400. 128 Noll JG, Zeller MH, Trickett PK, Putnam FW. Obesity risk for female victims of childhood sexual abuse: a prospective study. Pediatrics 2007; 120: 361–67. 129 Thomas C, Hyponnen E, Power C. Obesity and type 2 diabetes risk in mid-adult life: the role of childhood adversity. Pediatrics 2008; 121: e1240–49. 130 Lissau I, Sorensen TI. Parental neglect during childhood and increased risk of obesity in young adulthood. Lancet 1994; 343: 324–27.

80

131 Brewerton TD. Eating disorders, trauma, and comorbidity: focus on PTSD. Eat Disord 2007; 15: 285–304. 132 Felitti VJ, Anda RF, Nordenberg D, et al. Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. The Adverse Childhood Experiences (ACE) Study. Am J Prev Med 1998; 14: 245–58. 133 Draper B, Pfaﬀ JJ, Pirkis J, et al. Long-term eﬀects of childhood abuse on the quality of life and health of older people: results from the depression and early prevention of suicide in general practice project. J Am Geriatr Soc 2008; 56: 262–71. 134 St Amand A, Bard DE, Silovsky JF. Meta-analysis of treatment for child sexual behavior problems: practice elements and outcomes. Child Maltreat 2008; 13: 145–66. 135 Merrick MT, Litrownik AJ, Everson MD, Cox CE. Beyond sexual abuse: the impact of other maltreatment experiences on sexualized behaviors. Child Maltreat 2008; 13: 122–32. 136 Wilson H, Widom CS. An examination of risky sexual behavior and HIV in victims of child abuse and neglect: a 30-year follow-up. Health Psychol 2008; 27: 149–58. 137 Lansford MJ, Berlin D, Bates J, Pettit GS. Early physical abuse and later violent delinquency: a prospective longitudinal study. Child Maltreat 2007; 12: 233–45. 138 Arriola K, Louden T, Doldren M, Fortenberry R. A meta-analysis of the relationship of child sexual abuse to HIV risk behavior among women. Child Abuse Negl 2005; 29: 725–46. 139 Rind B, Tromovitch P, Bauserman R. A meta-analytic examination of assumed properties of child sexual abuse using college samples. Psychol Bull 1998; 124: 22–53. 140 Senn TE, Carey MP, Vanable PA, Coury-Doniger P, Urban M. Characteristics of sexual abuse in childhood and adolescence inﬂuence sexual risk behavior in adulthood. Arch Sex Behav 2007; 36: 637–45. 141 Kalichman SC, Gore-Felton C, Benotsch E, Cage M, Rompa D. Trauma symptoms, sexual behaviors, and substance abuse: correlates of childhood sexual abuse and HIV risks among men who have sex with men. J Child Sex Abus 2004; 13: 1–15. 142 Merrill LL, Guimond JM, Thomsen CJ, Milner JS. Child sexual abuse and number of sexual partners in young women: the role of abuse severity, coping style, and sexual functioning. J Consult Clin Psychol 2003; 71: 987–96. 143 Fergusson DM, Horwood LJ, Lynskey MT. Childhood sexual abuse, adolescent sexual behaviors and sexual revictimization. Child Abuse Negl 1997; 21: 789–803. 144 Paolucci EO, Genuis ML, Violato C. A meta-analysis of the published research on the eﬀects of child sexual abuse. J Psychol 2001; 135: 17–36. 145 Brown J, Cohen P, Chen H, Smailes E, Johnson JG. Sexual trajectories of abused and neglected youths. J Dev Behav Pediatr 2004; 25: 77–82. 146 Fergusson DM, Horwood LJ, Lynskey MT. Childhood sexual abuse and psychiatric disorder in young adulthood: II. Psychiatric outcomes of childhood sexual abuse. J Am Acad Child Adolesc Psychiatry 1996; 35: 1365–74. 147 Mullen PE, Martin JL, Anderson JC, Romans SE, Herbison GP. The long-term impact of the physical, emotional, and sexual abuse of children: a community study. Child Abuse Negl 1996; 20: 7–21. 148 Cohen M, Deamant C, Barkan S, et al. Domestic violence and childhood sexual abuse in HIV-infected women and women at risk for HIV. Am J Public Health 2000; 90: 560–65. 149 Hillis SD, Anda RF, Dube SR, Felitti VJ, Marchbanks PA, Marks JS. The association between adverse childhood experiences and adolescent pregnancy, long-term psychosocial consequences, and fetal death. Pediatrics 2004; 113: 320–27. 150 Tomeo ME, Templer DI, Anderson S, Kotler D. Comparative data of childhood and adolescence molestation in heterosexual and homosexual persons. Arch Sex Behav 2001; 30: 535–41. 151 Davis DA, Luecken LJ, Zautra AJ. Are reports of childhood abuse related to the experience of chronic pain in adulthood? A meta-analytic review of the literature. Clin J Pain 2005; 21: 398–405. 152 Raphael KG, Chandler HK, Ciccone DS. Is childhood abuse a risk factor for chronic pain in adulthood? Curr Pain Headache Rep 2004; 8: 99–110.

www.thelancet.com Vol 373 January 3, 2009

Series

153 Linton SJ. A prospective study of the eﬀects of sexual or physical abuse on back pain. Pain 2002; 96: 347–51. 154 Walsh CA, Jamieson E, MacMillan H, Boyle M. Child abuse and chronic pain in a community survey of women. J Interpers Violence 2007; 22: 1536–54. 155 Raphael KG. Childhood abuse and pain in adulthood: more than a modest relationship? Clin J Pain 2005; 21: 371–73. 156 Brown J, Berenson K, Cohen P. Documented and self-reported child abuse and adult pain in a community sample. Clin J Pain 2005; 21: 374–77. 157 Prosser LA, Corso PS. Measuring health-related quality of life for child maltreatment: a systematic literature review. Health Qual Life Outcomes 2007; 5: 42. 158 Bonomi AE, Anderson ML, Rivara FP, et al. Health care utilization and costs associated with childhood abuse. J Gen Intern Med 2008; 23: 294–99. 159 Chartier MJ, Walker JR, Naimark B. Childhood abuse, adult health, and health care utilization: results from a representative community sample. Am J Epidemiol 2007; 165: 1031–38. 160 Spataro J, Mullen PE, Burgess PM, Wells DL, Moss SA. Impact of child sexual abuse on mental health: prospective study in males and females. Br J Psychiatry 2004; 184: 416–21. 161 Widom CS. The cycle of violence. Science 1989; 244: 160–66. 162 Maxﬁeld MG, Widom CS. The cycle of violence: revisited 6 years later. Archives of Pediatrics & Adolescent Medicine. Arch Pediatr Adolesc Med 1996; 150: 390–95. 163 Smith C, Thornberry TP. The relationship between childhood maltreatment and adolescent involvement in delinquency. Criminology 1995; 33: 451–81.

www.thelancet.com Vol 373 January 3, 2009

164 Stouthamer-Loeber M, Loeber R, Homish DL, Wei E. Maltreatment of boys and the development of disruptive and delinquent behavior. Dev Psychopathol 2001; 13: 941–55. 165 Zingraﬀ MT, Leiter J, Myers KA, Johnsen MC. Child maltreatment and youthful problem behavior. Criminology 1993; 31: 173–202. 166 Herrenkohl RC, Egolf BP, Herrenkohl EC. Preschool antecedents of adolescent assaultive behavior: a longitudinal study. Am J Orthopsychiatry 1997; 67: 422–32. 167 Egeland B, Yates T, Appleyard K, van Dulmen M. The long-term consequences of maltreatment in the early years: a developmental pathyway model to antisocial behavior. Child Serv: Soc Pol Res Prac 2002; 5: 249–60. 168 Hubbard DJ, Pratt TC. A meta-analysis of the predictors of delinquency among girls. J Oﬀender Rehab 2002; 34: 1–13. 169 Maas C, Herrenkohl TI, Sousa C. Review of research on child maltreatment and violence in youth. Trauma Violence Abuse 2008; 9: 56–67. 170 Lewis T, Leeb R, Kotch J, et al. Maltreatment history and weapon carrying among early adolescents. Child Maltreat 2007; 12: 259–68. 171 MacMillan HL, Wathen CN, Barlow J, Fergusson DM, Leventhal JM, Taussig HN. Interventions to prevent child maltreatment and associated impairment. Lancet 2008; published online Dec 3. DOI:10.1016/S0140-6736(08)61708-0. 172 Reading R, Bissell S, Goldhagen J, et al. Promotion of children’s rights and prevention of child maltreatment. Lancet 2008; published online Dec 3. DOI:10.1016/S0140-6736(08)61709-2.

81

Seminar

Obstructive sleep apnoea and its cardiovascular consequences T Douglas Bradley, John S Floras Lancet 2009; 373: 82–93 Published Online October 30, 2008 DOI:10.1016/S01406736(08)61622-0 Sleep Research Laboratory of the Toronto Rehabilitation Institute, Toronto, Canada (Prof T D Bradley MD); Department of Medicine of the Toronto General (University Health Network) and Mount Sinai Hospitals, Toronto, Ontario, Canada, and Centre for Sleep Medicine and Circadian Biology, University of Toronto, Toronto, Ontario, Canada (Prof T D Bradley, Prof J S Floras MD) Correspondence to: Prof T Douglas Bradley, Toronto General Hospital of the University Health Network, Toronto, Ontario, M5G 2C4, Canada [email protected]

Obstructive sleep apnoea (OSA) is a common disorder in which repetitive apnoeas expose the cardiovascular system to cycles of hypoxia, exaggerated negative intrathoracic pressure, and arousals. These noxious stimuli can, in turn, depress myocardial contractility, activate the sympathetic nervous system, raise blood pressure, heart rate, and myocardial wall stress, depress parasympathetic activity, provoke oxidative stress and systemic inﬂammation, activate platelets, and impair vascular endothelial function. Epidemiological studies have shown signiﬁcant independent associations between OSA and hypertension, coronary artery disease, arrhythmias, heart failure, and stroke. In randomised trials, treating OSA with continuous positive airway pressure lowered blood pressure, attenuated signs of early atherosclerosis, and, in patients with heart failure, improved cardiac function. Current data therefore suggest that OSA increases the risk of developing cardiovascular diseases, and that its treatment has the potential to diminish such risk. However, large-scale randomised trials are needed to determine, deﬁnitively, whether treating OSA improves cardiovascular outcomes.

Introduction Obstructive sleep apnoea (OSA) is a disorder in which loss of pharyngeal dilator muscle tone at sleep onset causes recurrent pharyngeal collapse and temporary cessation of breathing (apnoea). An abnormally narrowed or collapsible pharynx puts individuals at greater risk.1 Such apnoeas cause repetitive hypoxia and carbon-dioxide retention, and provoke awakenings (ie, arousals) that restore pharyngeal dilator muscle tone and airﬂow. However, this respite is brief: pharyngeal obstruction recurs once sleep resumes, and recurrent arousals, although protective, disrupt sleep. An OSA disorder is generally deﬁned as ﬁve or more apnoeas–hypopneas per hour of sleep (ie, the apnoea–hypopnoea index [AHI]).2,3 After the initial description of OSA, a proposition was made that cardiovascular stresses evoked by OSA could contribute to the pathogenesis of cardiovascular diseases.4 However, in 1997, Wright and colleagues5 concluded that there was insuﬃcient evidence to support this concept. Over the subsequent decade, data from epidemiological, observational, and interventional studies have transformed our appreciation of how OSA could contribute to the

Search strategy and selection criteria PubMed was searched using the search term “sleep apnoea” with relevant phrases including “cardiovascular disease”, “hypertension”, “cardiac hypertrophy”, “heart failure”, “coronary artery disease”, “ischaemic heart disease”, “stroke”, “cardiac arrhythmias”, and “cerebrovascular disease”. Only papers published in English between 1997 and 2008 were used. Publications referred to in the identiﬁed articles were also reviewed and were cited if they provided important data not addressed in subsequent articles. Emphasis was placed on experimental mechanistic studies, large prospective epidemiological studies, and randomised controlled clinical trials using full polysomnography. Where data from randomised controlled trials are unavailable, non-randomised trials are discussed.

82

burden of cardiovascular diseases, and how its treatment might lessen this burden. The purpose of this Seminar is, ﬁrst, to review our current understanding of mechanisms by which OSA might contribute to the pathogenesis of cardiovascular disease; second, to assess epidemiological evidence concerning a potential link between these two conditions; third, to review cardiovascular eﬀects of treating OSA; and ﬁnally, to identify gaps in our knowledge in need of resolution.

Deﬁnition and diagnosis of OSA A diagnosis of OSA requires the presence of repetitive apnoeas and hypopnoeas during sleep. This presence is most reliably shown by attended overnight polysomnography in a sleep laboratory, in which sleep stages, arterial oxyhaemoglobin saturation, and respiratory movements of the rib cage and abdomen or respiratory eﬀort, or both, are recorded.2 Apnoea is an absence of tidal volume for at least 10 s, and hypopnoea is a decrease in tidal volume of at least 50%, but above zero, for at least 10 s accompanied by at least a 4% decrease in oxygen saturation or terminated by arousal from sleep. Apnoeas are obstructive if accompanied by respiratory eﬀorts against the occluded pharynx, and central if not.2 In the general population, OSA is by far the commonest form of sleep apnoea, whereas central sleep apnoea is rarely recorded in the absence of heart failure.6–10 OSA is frequently associated with a history of habitual snoring, which is a sign of increased pharyngeal airﬂow resistance. An OSA syndrome is deﬁned as an AHI equal to or more than 5 accompanied by either excessive daytime sleepiness or two or more of episodes of choking or gasping during sleep, recurrent awakenings, unrefreshing sleep, daytime fatigue, or impaired concentration or memory.2 The severity of OSA is most often reported as the AHI for two reasons. First, because hypoxia and arousals that disrupt sleep are a direct consequence of apnoeas and hypopnoeas, the AHI provides an index of the frequency of the primary events and their immediate physiological consequences. Second, in general, the higher the AHI, the www.thelancet.com Vol 373 January 3, 2009

Seminar

more severe the clinical manifestations of OSA, such as daytime sleepiness, neurocognitive impairment, and the likelihood of developing cardiovascular complications.3,11–15 The American Academy of Sleep Medicine classiﬁcation of OSA severity considers both the AHI (mild deﬁned as an AHI of 5–15, moderate as 15–30, and severe as >30) and degree of daytime sleepiness (mild: unwanted sleepiness or involuntary sleep episodes occurring during activities that need little attention; moderate: during activities that need some attention, such as during meetings; severe: during activities that need more active attention such as during conversation or driving).2 However, these thresholds are arbitrary.

Pathophysiology During non-rapid-eye movement sleep, metabolic rate, sympathetic nervous activity, blood pressure, and heart rate all decrease, whereas cardiac vagal tone increases from wakefulness.16–18 OSA interrupts this cardiovascular quiescence by triggering a cascade of acute haemodynamic, autonomic, chemical, inﬂammatory, and metabolic eﬀects, with chronic after-eﬀects capable of initiating or exacerbating cardiovascular disease (ﬁgure). These cycles of hypoxia and carbon-dioxide retention elicit oscillations in both cardiac parasympathetic and sympathetic nervous activity that aﬀect heart rate. Although bradycardia was originally thought to be a common response to obstructive apnoeas,19 it has subsequently been shown not to be, and heart rate has been shown to either increase, decrease, or remain unchanged during obstructive apnoeas.20 The net response might be a function of both autonomic balance and airﬂow; hypoxia slows heart rate by activating the vagus only in the absence of lung stretch or airﬂow.21,22 When parasympathetic tone predominates, the heart rate can slow, when sympathetic tone predominates, the heart rate can rise, and if vagal and sympathetic eﬀects are equal, the heart rate can remain unchanged. Repetitive apnoea-induced hypoxia and carbon-dioxide retention cause ineﬀectual inspiratory eﬀorts and the generation of negative intrathoracic pressure against the occluded pharynx that, by increasing the diﬀerence between intracardiac and extracardiac pressure, increase left ventricular transmural pressure (ie, afterload, a potent stimulus to left ventricular hypertrophy).23 Negative intrathoracic pressure draws blood into the thorax, augmenting right ventricular preload, while apnoea-induced hypoxia causes pulmonary vasoconstriction, increasing right ventricular afterload.24 These forces distend the right ventricle, causing leftward shift of the interventricular septum during diastole that impedes left ventricular ﬁlling and decreases stroke volume.24 Hypoxia during OSA might also directly impair cardiac contractility and diastolic relaxation.25,26 Such adverse consequences are rapidly relieved by application, via a nasal mask, of continuous positive www.thelancet.com Vol 373 January 3, 2009

Obstructive sleep apnoea

↓ PNA

↑ HR

↓ Intrathoracic pressure

Arousal

↓ PO2↑ PCO2

↑ SNA ↑ Catechols

↓ Myocardial O2 delivery

↑ BP

• Hypertension • Atherosclerosis • Myocardial ischaemia • LV hypertrophy and failure • Cardiac arrhythmias • Cerebrovascular disease

• Oxidative stress • Inﬂammation • Endothelial dysfunction

↑ LV wall tension ↑ Cardiac O2 demand

Figure: Pathophysiological eﬀects of obstructive sleep apnoea on the cardiovascular system PNA=parasympathetic nervous system activity. PO2=partial pressure of oxygen. PCO2=partial pressure of carbon dioxide. SNA=sympathetic nervous system activity. HR=heart rate. BP=blood pressure. LV=left ventricular.

airway pressure (CPAP) to splint the pharynx and maintain its patency.27 These cycles of hypoxia and carbon-dioxide retention elicit oscillations in sympathetically-mediated peripheral vasoconstriction that further augment afterload by raising systemic blood pressure.28 At apnoea termination, asphyxia triggers a brief arousal from sleep that abruptly increases sympathetic activity, and suppresses vagal tone, precipitating surges in blood pressure and heart rate.29 These acute eﬀects can be sustained into wakefulness, causing higher blood pressure and impaired vagallymediated heart-rate variability.28,30–32 Intermittent hypoxia can induce oxygen-free-radical production,33,34 and activate inﬂammatory pathways that impair vascular endothelial function35–37 and increase blood pressure independently of activation of the sympathetic nervous system.38 Non-randomised uncontrolled trials report that CPAP lowered the nuclearfactor-κB-dependent cytokines tumour necrosis factor α (TNFα), and interleukin 8,36,37 and also lowered plasma concentrations of interleukin 6 and C-reactive protein.35 Individuals with OSA have attenuated endotheliumdependent vasodilation,39,40 and decreased circulating markers of nitric oxide that increase after treatment by CPAP.41 OSA can also promote oxidation of lipoproteins, increased expression of adhesion molecules and monocyte adherence to endothelial cells, and vascular smooth-muscle proliferation.33,42–44 These adverse vascular eﬀects, combined with increased sympathetic vasoconstrictor activity and inﬂammation, could predispose to hypertension and atherosclerosis.45,46 In dogs, just 8 hours exposure to severe OSA can induce acute pulmonary oedema,47 and several weeks’ exposure to OSA can induce left ventricular hypertrophy and systolic dysfunction.48 83

Seminar

Platelet activation and aggregability, markers of increased susceptibility to thrombosis, are increased during sleep in patients with OSA and, in a nonrandomised trial, decreased after one night of CPAP.49 Morning ﬁbrinogen concentration is increased, and plasminogen activator inhibitor type-1 activity is decreased in patients with OSA,50 indicating less ﬁbrinolytic potential.51 Fibrinogen has been shown to decrease after one night of CPAP.52 Cerebral bloodﬂow declines signiﬁcantly during obstructive apnoeas due to a decrease in cardiac output.53 In patients with ﬂow-limiting lesions of the cerebral arteries, this can predispose to ischaemic events.54 Compared with control participants, patients with OSA show greater signs of early atherosclerosis, including greater carotid intima–media thickness, decreased arterial compliance, and a higher prevalence of silent brain infarcts.55–57

Epidemiology Prevalence In the USA, the prevalence of OSA (AHI≥5) in the adult, mainly white, population aged 30–60 years, has been estimated at 24% in men and 9% in women, and, at an AHI≥15, at 9% in men and 4% in women, with no major diﬀerences noted between African-Americans and White people.58 The corresponding prevalence of OSA syndrome has been estimated at 4% in men and 2% in women.3 In European populations, the most comprehensive data come from Spain, where 26% of men and 28% of women aged 30–70 years had an AHI≥5, and 14% of men and 7% of women had an AHI≥15.59 In a predominantly oriental male population from Hong Kong, aged 30–60 years, the prevalences of OSA and OSA syndrome at an AHI≥5 were 9% and 4%, respectively, and at an AHI≥15, 5% and 3%, respectively.60 These data suggest that OSA is common in several racial and ethnic groups, but that most individuals with OSA are asymptomatic. This factor could have public-health implications, Cross-sectional (prevalence)

Prospective (incidence)

Unadjusted*

Adjusted†

Unadjusted*

Adjusted†

Hypertension

3,13,75

Yes

Yes

Yes

Yes13

Dysglycaemia

Yes

76–78

Yes

Yes

No78

Coronary artery disease

Yes

14

14

Yes

NA

NA

Heart failure

Yes14

Yes14

NA

NA

Bradyarrhythmias

No79,80

No79,80

NA

NA

Atrial ﬁbrillation

Yes

79

Yes79

NA

NA

Ventricular ectopy

Yes79

Yes79

NA

NA

Cerebrovascular disease

Yes11

Yes11

Yes11,12

No11,12

3,13,75 76–78

13 78

Cardiac arrhythmias

Yes=available data support a signiﬁcant association. No=available data do not support a signiﬁcant association. NA=no data available. *Findings of univariable analyses or multivariate analyses with partial adjustment. †Findings of multivariable analyses in which adjustments have been made for all known confounding factors.

Table 1: Summary of community-based epidemiological studies that used polysomnography to investigate potential links between obstructive sleep apnoea and cardiovascular diseases

84

because relations between OSA and cardiovascular diseases seem not to be related to the presence of symptoms of sleep apnoea.61 Of those with OSA syndrome who could beneﬁt symptomatically from its treatment, 75–80% remain undiagnosed in the USA.62

Risk factors OSA is two to three-times more common in men than in women, and in those aged 65 years or more than in those aged 30–64 years.61 The risk of OSA also increases with increasing body weight: a 10% weight gain increases the risk of developing OSA by six-times.63 Fat accumulation in the neck, as a result of obesity, can impinge on the pharyngeal lumen and predispose its collapse during sleep.64 Nonetheless, OSA occurs in individuals of normal weight in whom other factors can contribute to pharyngeal collapsibility (eg, macroglossia and adenotonsillar hypertrophy).64 Anomalies of craniofacial structure, such as retrognathia, which can retrodisplace the tongue and narrow the pharynx, can be especially important in non-obese oriental populations.65 Nasal obstruction and smoking can also increase the risk for developing OSA, possibly by causing pharyngeal narrowing as a result of inﬂammation.61 Hereditary factors can also increase risk for reasons not fully elucidated.61,66

Association between OSA and cardiovascular diseases Compared with the general population, the prevalence of OSA is higher in populations with cardiovascular conditions, such as hypertension (30–83%),67,68 heart failure (12–53%),8,9,69 ischaemic heart disease (30–58%),70,71 and stroke (43–91%).72–74 However, the coexistence of these conditions with OSA does not prove causality, and potential confounding variables, such as obesity, need to be considered. Nevertheless, epidemiological data (table 1) support the concept that OSA can participate in the initiation or progression of several cardiovascular diseases. There is compelling experimental evidence that OSA can raise blood pressure which, in turn, increases cardiovascular risk.75 Dogs exposed to OSA developed hypertension during both sleep and wakefulness that resolved on reversal of OSA.30 Rats exposed to intermittent hypoxia, mimicking the recurrent hypoxia of OSA, developed hypertension that was prevented by sympathectomy or peripheral chemoreceptor denervation.46,81 Cross-sectional studies have shown increased odds of hypertension in association with OSA, independent of obesity.15,82 A prospective study also showed that for an AHI>15 versus 0, the odds of developing hypertension over 4–8 years was 2·89 (95% CI 1·46–5·64), independent of confounders.13 OSA is also common in patients with drug-resistant hypertension,68 in whom it is often associated with biochemical features of primary aldosteronism83 that could provoke oxidative stress, inﬂammation, and ventricular ﬁbrosis and hypertrophy.84 www.thelancet.com Vol 373 January 3, 2009

Seminar

Dysglycaemia and diabetes also increase the risk of developing cardiovascular disease. Exposure of healthy people to sleep deprivation or intermittent hypoxia increases glucose intolerance, insulin resistance, and activity of the sympathetic nervous system.85,86 In cross-sectional studies, involving as many as 2656 participants, those with OSA had greater resistance to insulin76,87 and higher prevalences of both hyperglycaemia and type 2 diabetes than those without OSA,77,78 after adjusting for confounders including weight. However, prospective data from one of these studies did not show an increased risk of incident type 2 diabetes associated with the presence of OSA during 4 years of follow-up.78 In mice, although exposure to either chronic intermittent hypoxia, or to a high cholesterol diet alone, did not induce atherosclerosis,88 simultaneous exposure to both induced aortic atherosclerosis, probably by potentiating hypercholesterolaemia and worsening lipid peroxidation. In a cross-sectional analysis of the Sleep Heart Health Study, OSA at the highest AHI quartile (AHI>11) was associated with modestly increased odds of coronary artery disease (OR 1·27 [95% CI 0·99–1·62]) versus the lowest quartile.14 Additionally, case–control studies have consistently shown a higher prevalence of OSA in men and women with coronary disease than in those without coronary disease.89,90 In the largest of several prospective observational studies with a mean follow-up of 10·1 years, Marin and co-workers91 noted that, compared with healthy controls matched for age, sex, and weight, those with severe untreated OSA (AHI>30; n=235) had more fatal (1·06 vs 0·3 per 100 patient years; p=0·0012) and non-fatal cardiovascular events (2·13 vs 0·45 per 100 patient years; p<0·0001), whereas event rates in those with severe OSA treated by CPAP (n=372) did not diﬀer signiﬁcantly from the controls. Similar ﬁndings have also been reported elsewhere.92 Obstructive apnoeas, induced by occluding an endotracheal tube during anaesthesia, provoke myocardial ischaemia in the presence, but not in the absence, of coronary artery ligation.93 In human beings, Mueller manoeuvres, which simulate the eﬀects of obstructive apnoeas, cause more pronounced decreases in leftventricular ejection fraction and stroke volume in patients with coronary artery disease or heart failure than in those without these conditions.94,95 Thus, the adverse eﬀects of OSA can be ampliﬁed in the presence of cardiovascular diseases. In patients with coronary disease, cyclic apnoea-induced hypoxia, negative intrathoracic pressure, and hypertension can provoke myocardial oxygen demand–supply mismatch and ischaemic electrocardiographic changes, as well as nocturnal angina that, in an uncontrolled study, were alleviated by CPAP.96 In patients with coronary disease, the presence of OSA is associated with higher mortality (38% vs 9%; p=0·018),97 more major adverse cardiac events (24% vs 5%; p=0·002), and more restenoses after percutaneous coronary interwww.thelancet.com Vol 373 January 3, 2009

ventions (37% vs 15%; p=0·026) than in patients without OSA.98 Milleron and colleagues99 did an observational study involving 54 patients with both coronary disease and OSA (AHI≥15). 25 patients accepted treatment for OSA. During 87 months of follow-up, treated patients had a signiﬁcantly decreased risk for the composite endpoint of cardiovascular death, acute coronary syndrome, hospitalisation for heart failure, or need for coronary revascularisation, than did untreated patients (adjusted hazard ratio [HR] 0·24 [95% CI 0·09–0·62]). In a similar observational study of 168 patients with OSA referred to a sleep clinic, 107 were treated with CPAP. Compared with the untreated group, all-cause mortality over the mean 7-year follow-up was not decreased in CPAP-treated patients, but cardiovascular mortality was (1·9% vs 14·8% deaths; p=0·009).100 Data from randomised trials are now needed to conﬁrm or refute these ﬁndings. Cross-sectional data from 6424 men and women showed a 2·38-times increased likelihood of having heart failure in association with OSA, independent of confounders,14 and prospective data showed that 21 days after acute myocardial infarction, the presence of OSA was associated with impaired recovery of left-ventricular systolic function.101 However, there are no prospective data concerning OSA and the risk of developing either left-ventricular hypertrophy or failure. In the general population, the prevalence of OSA with an AHI exceeding 10–15 is about 7–10%,3 but in patients with heart failure it is higher at 11–53%.8–10,69,102 Because the prevalence of central sleep apnoea with an AHI of 15 or more in patients with heart failure is 15–37%8–10 but less than 1%6,7 in the general population, the overall prevalence of these sleep-related breathing disorders in patients with heart failure is about 50%. Diﬀerentiating between OSA and central sleep apnoea in patients with heart failure is important because their pathophysiologies (central apnoeas are not caused by pharyngeal occlusion, but occur when the partial pressure of carbon dioxide falls below a threshold level needed to stimulate breathing) and treatments diﬀer, as previously reviewed.103,104 The ﬁnding that patients with heart failure have a lower body-mass index for a given AHI than the general population105 suggests that factors unrelated to obesity might contribute more to the pathogenesis of OSA in patients with heart failure. One such factor might be ﬂuid redistribution from oedematous legs to peripharyngeal tissues when moving from the upright to recumbent position, impinging on the pharyngeal lumen and predisposing to collapse during sleep.106,107 In addition to its chronic eﬀects on systemic blood pressure, OSA could, through several mechanisms independent of blood pressure, promote left-ventricular hypertrophy, diastolic and systolic dysfunction, and overt heart failure. These mechanisms include the mechanical and trophic consequences of repetitive increases in left-ventricular wall stress with each obstructed inspira85

Seminar

tory eﬀort,108 hypoxia, increased sympathetic drive to the heart, and OSA-related increases in aldosterone secretion.83,84 Compared with patients without OSA, those with OSA, but without overt heart failure, have a higher prevalence of impaired diastolic relaxation, reversible by 3 months of CPAP therapy.109 Several studies reported increased left-ventricular thickness or mass in association with OSA,109–111 but these relations were not signiﬁcant after adjustment for body weight. Only two studies, one in children112 and one in adults with non-ischaemic dilated cardiomyopathy113 noted an independent association between OSA and left-ventricular hypertrophy. In the adult study, left ventricular thickening was more prevalent in patients with OSA than in those without. OSA might also promote the development of intrathoracic aortic aneurysms, presumably by inducing repetitive increases in intrathoracic aortic transmural pressure during obstructed inspiratory eﬀorts.114 Cardiac metabolic gene expression has a circadian rhythm that anticipates diurnal variations in the ratio between workload and substrate availability.115 Disruption of such rhythms by obstructive apnoeas can cause a temporal mismatch between increased myocardial oxygen demand, as a result of increased left-ventricular wall tension, and decreased oxygen supply, as a result of apnoea-induced hypoxia, that might aid contractile dysfunction. Findings of an uncontrolled study involving seven patients with heart failure suggest that such impaired myocardial metabolic eﬃciency might be improved if coexisting OSA is abolished by CPAP.116 In addition to these adverse eﬀects of obstructive apnoeas during sleep, patients with heart failure and OSA have higher daytime sympathetic-nerve traﬃc and systolic blood pressure than those with heart failure alone,117,118 which could promote abnormalities that worsen prognosis, such as cardiac myocyte hypertrophy, β-adrenoceptor desensitisation, and cardiac arrhythmias. In a prospective observational study involving 164 patients with heart failure who were followed up over 7·3 years,10 mortality was signiﬁcantly greater in the 37 patients with untreated OSA (AHI≥15) than in 113 patients without sleep apnoea (8·7 vs 4·2 deaths per 100 patient years; p=0·029) after controlling for confounders. There was also a non-signiﬁcant tendency for lower mortality in those with CPAP-treated OSA compared with those with untreated OSA (p=0·07). In a community-based study, Mehra and co-workers79 noted no signiﬁcant diﬀerence in the prevalence of nocturnal sinus pauses between patients with severe OSA (AHI≥30, n=228) and those without OSA (AHI<5, n=338; 11% vs 9%, p=0·34). Second-degree atrioventricular block was rare, and its prevalence did not diﬀer between the groups. Flemons and colleagues80 reported similar ﬁndings in patients referred to a sleep laboratory for suspicion of OSA. In 66 patients with OSA, Roche and co-workers119 recorded a higher prevalence of sinus 86

pauses (9·1% vs 0%; p<0·01), but not of second-degree heart block, than in 81 participants without OSA. Although there might not be an overall increased prevalence of bradyarrhythmias in individuals with OSA, in non-randomised studies where bradyarrhythmias were detected, they resolved with administration of supplemental oxygen, atropine, or treatment of OSA.19,22,120,121 Increased parasympathetic input to the heart, rather than intrinsic abnormalities of the cardiac conduction system,122 seems to be the main nocturnal bradyarrhythmic mechanism. Thus, if OSA evokes nocturnal bradycardia, its treatment might obviate the need for cardiac pacing. Two studies reported that the prevalence of nocturnal paroxysmal supraventricular tachycardia was not increased in OSA.79,119 Although another study reported a higher prevalence of atrial ﬁbrillation in those with OSA than in those without (4·8% vs 0·9%, respectively; p=0·003),79 this ﬁnding was not replicated in other studies.102,119,123 By contrast, in two of these other studies, there was a strikingly higher prevalence of atrial ﬁbrillation in patients with central sleep apnoea, both in the setting of preserved (27% vs 2% in those with OSA and 3% in those without sleep apnoea; p<0·001)123 or impaired (23% vs 12% in those with OSA and 8% in those without sleep apnoea; p<0·05)102 left-ventricular systolic function. Future investigations of potential relations between sleep-disordered breathing and atrial ﬁbrillation will therefore need careful diﬀerentiation of OSA from central sleep apnoea. In an observational study, Kanagala and co-workers124 reported that patients with untreated OSA had twice the risk for recurrence of atrial ﬁbrillation within 1 year of electrical cardioversion to sinus rhythm than those whose OSA was treated with CPAP (82% vs 42%, respectively; p=0·013). However, in a retrospective analysis, the same group noted that obesity rather than OSA at baseline was an independent predictor of incident atrial ﬁbrillation.125 Mehra and colleagues79 also reported that the prevalence of ventricular premature beats and complex ventricular ectopy was higher in those with, than in those without, OSA (35% vs 21% [p=0·0003] and 25% vs 15% [p=0·002], respectively). By contrast, Flemons and colleagues80 showed no signiﬁcant diﬀerence in the prevalence of these ventricular arrhythmias between patients with and without OSA. Thus, uncertainty remains as to whether OSA induces ventricular arrhythmias. Gami and co-workers126 reported that, in patients with OSA, the relative risk of sudden cardiac death between midnight and 0600 h was 2·57 compared with the general population, whose peak risk for sudden cardiac death was between 0600 h and noon. However, whether these individuals died during sleep is unknown. Thus, although suggestive, these data do not establish a direct causal relation between OSA and an increased risk of sudden death during sleep. Indeed, in an earlier report involving 13 patients with OSA who had sudden death, none died during sleep.127 www.thelancet.com Vol 373 January 3, 2009

Seminar

Treatment

Patients enrolled, n

Patients Treatment completing trial, n period

Treatment outcomes

Becker et al14

Therapeutic vs sham CPAP

60

32

10 mm Hg decrease in systolic and diastolic blood pressure

Robinson et al138

Therapeutic vs sham CPAP

35

32

1 month

No change in blood pressure

Kaneko et al139

Therapeutic CPAP vs no CPAP

24

24

1 month

9% increase in LVEF; 10 mm Hg decrease in systolic blood pressure; decrease in heart rate of four beats per minute

Mansﬁeld et al140

Therapeutic CPAP vs no CPAP

55

40

3 months

5% increase in LVEF; no change in blood pressure; decrease in nocturnal urinary concentration of norepinephrine; improved QOL

Usui et al141†

Therapeutic CPAP vs no CPAP

17

17

1 month

17% decrease in muscle SNA; 15 mm Hg decrease in awake systolic blood pressure

Gilman et al142†

Therapeutic CPAP vs no CPAP

19

19

1 month

Increase in high frequency HRV

Ryan et al143†

Therapeutic CPAP vs no CPAP

18

18

1 month

58% decrease in frequency of VPB during sleep

Egea et al144

Therapeutic vs sham CPAP

61

45

2 months

2·7% increase in LVEF; no change in blood pressure, QOL, or 6-MWT

Therapeutic vs sham CPAP

42

40

3 months

No change in outcomes of HbA1c, euglycaemic clamp, or HOMA%S

Sandberg et al146

Therapeutic CPAP vs no CPAP

63

59

Hsu et al147

Therapeutic CPAP vs no CPAP

30

28

Hypertension* Mean 65 days (SD 50)

Heart failure†

Type 2 diabetes West et al145 Stroke 28 days 8 weeks

Decrease in depressive symptoms; no change in physical or cognitive function No change in physical or cognitive function, sleepiness, QOL, or 24-h blood pressure

CPAP=continuous positive airway pressure. LVEF=left-ventricular ejection fraction. QOL=quality of life. SNA=sympathetic nervous system activity. HRV=heart-rate. VPB=ventricular premature beats. 6-MWT=6-minute walk-test distance. Hb=haemoglobin. HOMA%S=homoeostatic model assessment. *Only trials in which most patients had elevated blood pressure at enrolment are included. †Overlap exists with some patients in these trials because these reports represent components of a larger clinical trial that was extended after the ﬁndings by Kaneko and colleagues139 were reported.

Table 2: Summary of randomised trials of treatment for obstructive sleep apnoea on outcomes contributing to cardiovascular diseases

After a stroke, OSA is frequently noted, adversely aﬀecting patients’ cognitive and physical function.73,74 However, whether OSA predisposes to stroke or vice versa remains unclear. The ﬁnding that there is no association between stroke location, type, or severity and the presence of OSA72,74 favours the former possibility. Cross-sectional data from the Sleep Heart Health Study showed greater odds for stroke in the highest AHI quartile than in the lowest quartile (1·58 [95% CI 1·02–2·46]).14 Subsequently, Arzt and colleagues11 showed, in a community sample between the ages of 30 and 60 years, that, compared with patients without OSA, those with moderate to severe OSA (AHI≥20) had 4·33 (95% CI 1·32–14·24) greater odds for prevalent stroke, independent of other risk factors, and that after 4 years, an AHI of 20 or greater at baseline was associated with an increased risk of incident stroke after adjustment for age and sex (odds ratio 4·48 [95% CI 1·31–15·33]), but not for body-mass index. In an elderly population (70–100 years of age), severe untreated OSA (AHI≥30) was associated with a 2·52-times increased risk of incident stroke over a mean follow-up of 4·5 years compared with patients with an AHI of less than 30 (p=0·04);12 however, because other risk factors were not accounted for, whether severe OSA is an independent risk for stroke in this context remains uncertain. In a longitudinal study of 408 patients with coronary artery disease, Mooe and co-workers71 noted that OSA independently increased the risk for incident cerebrovascular events (HR=2·98 [95% CI 1·43–6·20]; p=0·004). In an observational study, Yaggi and colleagues128 reported an association between OSA and the combined rate of www.thelancet.com Vol 373 January 3, 2009

stroke, transient ischaemic attack, and all-cause mortality, independent of confounders (HR=1·97 [95% CI 1·12–3·48]). However, there was no signiﬁcant association between OSA and either stroke, transient ischaemic attacks, or death alone. Moreover, treated and untreated patients with OSA were included together in this analysis. Because data from other observational studies report that treatment of OSA is associated with a decreased risk of cardiovascular and cerebrovascular events,10,91 the data reported by Yaggi and colleagues are diﬃcult to interpret. Hence, further studies will be needed to establish whether OSA contributes to causation of stroke independently of other risk factors.

Cardiovascular eﬀects of treating OSA In this section we review the ﬁndings of randomised controlled trials of OSA treatment on cardiovascular endpoints. Currently, there is no eﬀective drug therapy for OSA. Oral appliances, electrical stimulation of the pharyngeal dilator muscles, or surgery are less eﬀective than CPAP.129–131 Atrial overdrive pacing, which, in an initial report, caused a modest improvement in OSA in patients with bradyarrhythmias,132 is ineﬀective in those without bradyarrhythmias.133–135 CPAP and mandibular advancement devices136,137 are the only treatments for OSA whose eﬀects on cardiovascular endpoints have been assessed in randomised trials (table 2).

Hypertension Several randomised trials have assessed the eﬀect of CPAP on daytime blood pressure in patients with 87

Seminar

OSA.148–153 However, most participants were normotensive148–150 or had a history of hypertension controlled on medications at the time of enrolment.152 These trials were short (all <3 months) and measured blood pressure by diﬀerent methods at diﬀerent times of day.108 In some instances, the control for therapeutic CPAP was sham CPAP (mask applied with no pressure), which, in two such trials, tended to increase blood pressure.150,153 In two meta-analyses, which included patients with both normal and increased blood pressure, a modest eﬀect-size of CPAP of about 2 mm Hg in mean blood-pressure reduction was reported.154,155 However, a third meta-analysis did not show a signiﬁcant eﬀect of CPAP on blood pressure.156 In a 1-month trial involving 68 patients with OSA with a history of medically-treated hypertension, in whom blood pressure was well controlled in most (mean 24-h blood pressure=130/78 mm Hg), therapeutic CPAP had no eﬀect on blood pressure.152 Robinson and colleagues138 did a 1-month double cross-over trial of 35 patients with OSA who were not sleepy, most of whom had increased 24-h blood pressure at enrolment (mean 24-h blood pressure 142/86 mm Hg). Therapeutic CPAP had no eﬀect on blood pressure.138 By contrast, in a study150 involving 118 patients who were mainly normotensive with hypersomnolence, a 1-month trial of CPAP signiﬁcantly lowered the 24-h mean blood pressure by 2·5 mm Hg. However, in a subgroup with an AHI greater than the median (>33), mean blood pressure decreased by 5 mm Hg, and in those with a history of treated hypertension mean blood pressure decreased by 8 mm Hg. In a 2-month trial151 involving patients with OSA and hypersomnolence, 65% of the 60 enrolled participants were hypertensive on enrolment. CPAP caused a signiﬁcant 10 mm Hg decrease in both 24-h systolic and diastolic blood pressure. Taken together, these data suggest that treatment of OSA by CPAP can lower blood pressure, and is most eﬀective in patients with increased blood pressure and more severe OSA accompanied by hypersomnolence. Because a lowering of blood pressure is more important in hypertensive than in normotensive patients, table 2 lists only the two published randomised trials in which most patients had increased blood pressure at enrolment. Gotsopoulos and colleagues137 compared the eﬀects of a mandibular advancement device, adjusted to alleviate OSA, versus a sham oral appliance on 24-h blood pressure in a randomised trial. After 4 weeks, the device decreased 24-h diastolic blood pressure by 1·8 mm Hg (p=0·001), but systolic blood pressure remained unchanged. In a second randomised trial lasting 3 months, which compared an advancement device versus CPAP versus a placebo tablet, Barnes and colleagues136 documented a 2·2 mm Hg decrease in nocturnal diastolic blood pressure only (p<0·05 compared with placebo and CPAP). Patients in both trials were predominantly normotensive. 88

Dysglycaemia, diabetes mellitus, and atherosclerosis Up to now, only two randomised trials have investigated these conditions. West and co-workers145 studied 42 men with type II diabetes and OSA who were randomly assigned to therapeutic or sham CPAP for 3 months. Therapeutic CPAP had no eﬀect on any measure of glycaemic control or insulin resistance. Drager and colleagues157 tested the eﬀects of 4 months of CPAP versus no CPAP on early signs of atherosclerosis in 24 patients with severe OSA. Compared with the control group, those treated with CPAP had a 9% (p=0·04) decrease in carotid intima–media thickness, and an 11% (p<0·001) decrease in arterial pulse-wave velocity, changes that correlated signiﬁcantly with concordant decreases in daytime plasma C-reactive protein and norepinephrine concentrations.

Heart failure Kaneko and colleagues139 randomly assigned 24 patients with heart failure and OSA (left-ventricular ejaction fraction ≤45%, AHI≥20) to either a control group on optimum medical heart-failure therapy, or to a treatment group that additionally received CPAP. Over 1 month, CPAP increased mean daytime left-ventricular ejaction fraction from 25% to 34% (p<0·001), and lowered morning systolic blood pressure from 126 mm Hg to 116 mm Hg (p=0·02). In an extension of this trial, one of the causes for this decrease in blood pressure was shown to be a CPAP-induced lowering of sympathetic vasoconstrictor nerve discharge.141 CPAP also increased high-frequency heart-rate variability and baroreﬂex sensitivity in such patients, suggesting improved vagal modulation of heart rate.142,158 In another randomised trial that involved 40 patients with heart failure and OSA, Mansﬁeld and colleagues140 tested the eﬀects of 3 months of CPAP treatment. In CPAP-treated patients, leftventricular ejaction fraction increased from 38% to 43% (p=0·04), and nocturnal urinary norepinephrine concentrations decreased (p=0·036). There were signiﬁcant improvements in quality of life and sleepiness, but not in mean blood pressure. Compared with the patients studied by Kaneko and co-workers,139 these individuals had milder systolic dysfunction and OSA, which might explain the lesser improvement in cardiovascular function. Nevertheless, the above studies are consistent in showing that treatment of OSA in patients with heart failure improves left-ventricular systolic function and lowers activity of the sympathetic nervous system.

Cardiac arrhythmias In the only randomised trial published, Ryan and co-workers143 reported that the treatment of coexisting OSA in patients with heart failure by CPAP for 1 month caused a 59% decrease in the frequency of ventricular premature beats during sleep (p=0·037). www.thelancet.com Vol 373 January 3, 2009

Seminar

Cerebrovascular disease Sandberg and colleagues146 did a randomised trial of CPAP versus no CPAP in 63 outpatients with OSA 2–4 weeks after a stroke. Compared with controls, CPAP-treated patients had a greater improvement in their depression score, but not in cognitive or physical function over 4 weeks. By contrast with most stroke studies, in which OSA predominates,72–74 there were more central events in this population. Because central sleep apnoea is less likely to respond to CPAP than OSA,159 and because CPAP was titrated only according to oximetry, sleep apnoea in this cohort might not have been treated eﬀectively. Hsu and colleagues147 did a randomised trial of CPAP in 30 outpatients with coexisting OSA within 25 days of a stroke. Patients randomly assigned to CPAP had no improvement in activities of daily living, cognition, depression, sleepiness, quality of life, or blood pressure. However, the eﬀect of CPAP on OSA was not reported and CPAP compliance was poor with overall use of only 1·4 h per night. Thus, the main conclusion to be drawn from this study might be that stroke outpatients have diﬃculty adhering to CPAP. The question of whether eﬀective treatment of OSA with CPAP improves stroke recovery remains unanswered.

Prevention of OSA The most important modiﬁable risk factor for OSA is excess body weight. In patients with OSA, weight loss has been shown to decrease the AHI.160 Unfortunately, because obesity is increasing substantially in industrialised countries, the prevalence of OSA is anticipated to also increase. Thus, the most important public-health measure that could be taken to prevent OSA would be to promote avoidance or reversal of weight gain, especially in children.

Summary, controversies, and future challenges Data from animal models, epidemiological studies, and randomised clinical trials provide strong evidence that OSA can cause hypertension, and that its treatment can lower blood pressure. Indeed, OSA might well be the commonest treatable cause of secondary hypertension. A consistent ﬁnding of trials on CPAP therapy in patients with heart failure and OSA is an improvement in left-ventricular systolic function and a decrease in activity of the sympathetic nervous system, two important clinical markers of adverse outcome in heart failure. OSA can provoke acute nocturnal myocardial ischaemia in patients with coronary artery disease, and in epidemiological studies, untreated OSA increased the risk of developing coronary artery disease. Non-randomised observational studies have consistently described an increased risk of fatal and non-fatal cardiovascular events associated with untreated OSA, independent of other risk factors, and a decrease in this risk in patients treated with CPAP for OSA. The eﬀect www.thelancet.com Vol 373 January 3, 2009

Panel: Cardiovascular consequences of obstructive sleep apnoea (OSA): unresolved issues Pathophysiology • Are processes triggered acutely by OSA epiphenomena or independent pathophysiological mechanisms? • Can animal models simulating human OSA be developed to explore its acute and chronic cardiovascular eﬀects? Epidemiology • Does OSA cause or contribute independently to the development of type II diabetes, coronary artery disease, heart failure, cardiac arrhythmias, and cerebrovascular disease? • If hypertension is an intermediary step between OSA and the development of cardiovascular diseases, how should one control for the potentially confounding eﬀects of hypertension in examining these issues? Treatment • Does treating OSA decrease the risk of developing hypertension and cardiovascular diseases? • In patients with cardiovascular diseases, does treating coexisting OSA decrease cardiovascular morbidity or mortality? • If randomised trials are to be done to resolve these questions, how should the issue of leaving a control group with OSA untreated for lengthened periods be addressed? • Can less cumbersome treatments for OSA be developed as an alternative to CPAP and other forms of positive airway pressure?

of CPAP in patients who have had a stroke and who have OSA remains to be deﬁned. Despite this substantial body of supportive evidence, large, long-term randomised trials are needed to delineate deﬁnitively the role of diagnosing and treating OSA in decreasing the incidence of, and mortality from, cardiovascular diseases (panel). Undertaking such studies and implementing ﬁndings in clinical practice will pose several challenges. Currently, the only clear-cut indication for treating OSA is a complaint of daytime sleepiness.2 Consequently, there is reluctance to deny such patients CPAP for extended periods in a clinical trial. Conversely, if excessive daytime sleepiness is not reported, CPAP treatment of even severe OSA has been shown not to improve neurocognitive function, alertness, or blood pressure.138,161 Compared with the general population, patients with OSA and heart failure have signiﬁcantly less subjective daytime sleepiness.105 Nevertheless, in this population, treatment of OSA by CPAP improves surrogate markers of adverse cardiovascular prognosis.139,140,141 These ﬁndings give rise to complex issues. If a diagnosis of OSA depends on referral to a sleep laboratory, and if referral is contingent on symptoms of an OSA syndrome, 89

Seminar

then contemporary approaches to screening might exclude a large population of individuals at risk of cardiovascular events from the potential beneﬁts of participation in clinical trials, or from OSA management. Simple and cost-eﬀective methods to screen for OSA will therefore need to be developed and tested in unselected community samples. Second, relief of sleepiness might not be the main target of treatment in many patients with OSA. If so, such patients might not sense short-term or long-term symptomatic relief from CPAP therapy and might thus abandon it. Therefore, better methods of encouraging adherence to long-term CPAP are needed. Finally, because of issues with bed-partner acceptance, nasal congestion, and nasofacial discomfort or claustrophobia, not all patients with OSA can or will accept CPAP therapy. Accordingly, there remains an unmet need for eﬀective, well-tolerated alternative therapies for OSA. Conﬂict of interest statement Subsequent to the acceptance for publication of this Seminar, we received research grant support from Respironics Inc to develop a large clinical trial initiated and designed by ourselves. Acknowledgments This paper is supported by a grant from the Canadian Institutes of Health Research (MOP-82731). JSF is supported by a Canada Research Chair in Integrative Cardiovascular Biology and a Career Investigator Award from the Heart and Stroke Foundation of Ontario, Canada. References 1 Remmers JE, deGroot WJ, Sauerland EK, Anch AM. Pathogenesis of upper airway occlusion during sleep. J Appl Physiol 1978; 44: 931–38. 2 Sleep-related breathing disorders in adults: recommendations for syndrome deﬁnition and measurement techniques in clinical research. The Report of an American Academy of Sleep Medicine Task Force. Sleep 1999; 22: 667–89. 3 Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med 1993; 328: 1230–35. 4 Guilleminault C, Cummiskey J, Dement WC. Sleep apnea syndrome: recent advances. Adv Intern Med 1980; 26: 347–72. 5 Wright J, Johns R, Watt I, Melville A, Sheldon T. Health eﬀects of obstructive sleep apnoea and the eﬀectiveness of continuous positive airways pressure: a systematic review of the research evidence. BMJ 1997; 314: 851–60. 6 Bixler EO, Vgontzas AN, Lin HM, et al. Prevalence of sleep-disordered breathing in women: eﬀects of gender. Am J Respir Crit Care Med 2001; 163: 608–13. 7 Bixler EO, Vgontzas AN, Ten Have T, Tyson K, Kales A. Eﬀects of age on sleep apnea in men: I. Prevalence and severity. Am J Respir Crit Care Med 1998; 157: 144–48. 8 Ferrier K, Campbell A, Yee B, et al. Sleep-disordered breathing occurs frequently in stable outpatients with congestive heart failure. Chest 2005; 128: 2116–22. 9 Javaheri S. Sleep disorders in systolic heart failure: a prospective study of 100 male patients. The ﬁnal report. Int J Cardiol 2006; 106: 21–28. 10 Wang H, Parker JD, Newton GE, et al. Inﬂuence of obstructive sleep apnea on mortality in patients with heart failure. J Am Coll Cardiol 2007; 49: 1625–31. 11 Arzt M, Young T, Finn L, Skatrud JB, Bradley TD. Association of sleep-disordered breathing and the occurrence of stroke. Am J Respir Crit Care Med 2005; 172: 1447–51. 12 Munoz R, Duran-Cantolla J, Martinez-Vila E, et al. Severe sleep apnea and risk of ischemic stroke in the elderly. Stroke 2006; 37: 2317–21. 13 Peppard PE, Young T, Palta M, Skatrud J. Prospective study of the association between sleep-disordered breathing and hypertension. N Engl J Med 2000; 342: 1378–84. 14 Shahar E, Whitney CW, Redline S, et al. Sleep-disordered breathing and cardiovascular disease: cross-sectional results of the Sleep Heart Health Study. Am J Respir Crit Care Med 2001; 163: 19–25.

90

15

16 17 18

19

20

21 22

23

24

25

26 27

28

29

30

31

32

33

34

35

36

37

38

Young T, Peppard P, Palta M, et al. Population-based study of sleep-disordered breathing as a risk factor for hypertension. Arch Intern Med 1997; 157: 1746–52. White DP, Weil JV, Zwillich CW. Metabolic rate and breathing during sleep. J Appl Physiol 1985; 59: 384–91. Mancia G. Autonomic modulation of the cardiovascular system during sleep. N Engl J Med 1993; 328: 347–49. Hornyak M, Cejnar M, Elam M, Matousek M, Wallin BG. Sympathetic muscle nerve activity during sleep in man. Brain 1991; 114: 1281–95. Guilleminault C, Connolly SJ, Winkle RA. Cardiac arrhythmia and conduction disturbances during sleep in 400 patients with sleep apnea syndrome. Am J Cardiol 1983; 52: 490–94. Bonsignore MR, Romano S, Marrone O, Chiodi M, Bonsignore G. Diﬀerent heart rate patterns in obstructive apneas during NREM sleep. Sleep 1997; 20: 1167–74. Kato H, Menon AS, Slutsky AS. Mechanisms mediating the heart rate response to hypoxemia. Circulation 1988; 77: 407–14. Zwillich C, Devlin T, White D, Douglas N, Weil J, Martin R. Bradycardia during sleep apnea. Characteristics and mechanism. J Clin Invest 1982; 69: 1286–92. Tkacova R, Rankin F, Fitzgerald FS, Floras JS, Bradley TD. Eﬀects of continuous positive airway pressure on obstructive sleep apnea and left ventricular afterload in patients with heart failure. Circulation 1998; 98: 2269–75. Stoohs R, Guilleminault C. Cardiovascular changes associated with obstructive sleep apnea syndrome. J Appl Physiol 1992; 72: 583–89. Kusuoka H, Weisfeldt ML, Zweier JL, Jacobus WE, Marban E. Mechanism of early contractile failure during hypoxia in intact ferret heart: evidence for modulation of maximal Ca2+-activated force by inorganic phosphate. Circ Res 1986; 59: 270–82. Cargill RI, Kiely DG, Lipworth BJ. Adverse eﬀects of hypoxaemia on diastolic ﬁlling in humans. Clin Sci (Lond) 1995; 89: 165–69. Sullivan CE, Issa FG, Berthon-Jones M, Eves L. Reversal of obstructive sleep apnoea by continuous positive airway pressure applied through the nares. Lancet 1981; 1: 862–65. Somers VK, Dyken ME, Clary MP, Abboud FM. Sympathetic neural mechanisms in obstructive sleep apnea. J Clin Invest 1995; 96: 1897–904. Horner RL, Brooks D, Kozar LF, Tse S, Phillipson EA. Immediate eﬀects of arousal from sleep on cardiac autonomic outﬂow in the absence of breathing in dogs. J Appl Physiol 1995; 79: 151–62. Brooks D, Horner RL, Kozar LF, Render-Teixeira CL, Phillipson EA. Obstructive sleep apnea as a cause of systemic hypertension. Evidence from a canine model. J Clin Invest 1997; 99: 106–09. Arabi Y, Morgan BJ, Goodman B, Puleo DS, Xie A, Skatrud JB. Daytime blood pressure elevation after nocturnal hypoxia. J Appl Physiol 1999; 87: 689–98. Narkiewicz K, Montano N, Cogliati C, van de Borne PJ, Dyken ME, Somers VK. Altered cardiovascular variability in obstructive sleep apnea. Circulation 1998; 98: 1071–77. Dyugovskaya L, Lavie P, Lavie L. Increased adhesion molecules expression and production of reactive oxygen species in leukocytes of sleep apnea patients. Am J Respir Crit Care Med 2002; 165: 934–39. Schulz R, Mahmoudi S, Hattar K, et al. Enhanced release of superoxide from polymorphonuclear neutrophils in obstructive sleep apnea. Impact of continuous positive airway pressure therapy. Am J Respir Crit Care Med 2000; 162: 566–70. Yokoe T, Minoguchi K, Matsuo H, et al. Elevated levels of C-reactive protein and interleukin-6 in patients with obstructive sleep apnea syndrome are decreased by nasal continuous positive airway pressure. Circulation 2003; 107: 1129–34. Ryan S, Taylor CT, McNicholas WT. Selective activation of inﬂammatory pathways by intermittent hypoxia in obstructive sleep apnea syndrome. Circulation 2005; 112: 2660–67. Ryan S, Taylor CT, McNicholas WT. Predictors of elevated nuclear factor-kappaB-dependent genes in obstructive sleep apnea syndrome. Am J Respir Crit Care Med 2006; 174: 824–30. Vongpatanasin W, Thomas GD, Schwartz R, et al. C-reactive protein causes downregulation of vascular angiotensin subtype 2 receptors and systolic hypertension in mice. Circulation 2007; 115: 1020–28.

www.thelancet.com Vol 373 January 3, 2009

Seminar

39

40

41

42

43 44

45

46

47

48

49

50

51

52

53 54

55

56

57

58

59

60

61

Kato M, Roberts-Thomson P, Phillips BG, et al. Impairment of endothelium-dependent vasodilation of resistance vessels in patients with obstructive sleep apnea. Circulation 2000; 102: 2607–10. Imadojemu VA, Gleeson K, Quraishi SA, Kunselman AR, Sinoway LI, Leuenberger UA. Impaired vasodilator responses in obstructive sleep apnea are improved with continuous positive airway pressure therapy. Am J Respir Crit Care Med 2002; 165: 950–53. Ip MS, Lam B, Chan LY, et al. Circulating nitric oxide is suppressed in obstructive sleep apnea and is reversed by nasal continuous positive airway pressure. Am J Respir Crit Care Med 2000; 162: 2166–71. Bedwell S, Dean RT, Jessup W. The action of deﬁned oxygen-centred free radicals on human low-density lipoprotein. Biochem J 1989; 262: 707–12. Faller DV. Endothelial cell responses to hypoxic stress. Clin Exp Pharmacol Physiol 1999; 26: 74–84. El-Solh AA, Mador MJ, Sikka P, Dhillon RS, Amsterdam D, Grant BJ. Adhesion molecules in patients with coronary artery disease and moderate-to-severe obstructive sleep apnea. Chest 2002; 121: 1541–47. Carlson JT, Rangemark C, Hedner JA. Attenuated endothelium-dependent vascular relaxation in patients with sleep apnoea. J Hypertens 1996; 14: 577–84. Fletcher EC, Lesske J, Culman J, Miller CC, Unger T. Sympathetic denervation blocks blood pressure elevation in episodic hypoxia. Hypertension 1992; 20: 612–19. Fletcher EC, Proctor M, Yu J, et al. Pulmonary edema develops after recurrent obstructive apneas. Am J Respir Crit Care Med 1999; 160: 1688–96. Parker JD, Brooks D, Kozar LF, et al. Acute and chronic eﬀects of airway obstruction on canine left ventricular performance. Am J Respir Crit Care Med 1999; 160: 1888–96. Bokinsky G, Miller M, Ault K, Husband P, Mitchell J. Spontaneous platelet activation and aggregation during obstructive sleep apnea and its response to therapy with nasal continuous positive airway pressure. A preliminary investigation. Chest 1995; 108: 625–30. Steiner S, Jax T, Evers S, Hennersdorf M, Schwalen A, Strauer BE. Altered blood rheology in obstructive sleep apnea as a mediator of cardiovascular risk. Cardiology 2005; 104: 92–96. Rangemark C, Hedner JA, Carlson JT, Gleerup G, Winther K. Platelet function and ﬁbrinolytic activity in hypertensive and normotensive sleep apnea patients. Sleep 1995; 18: 188–94. Chin K, Ohi M, Kita H, et al. Eﬀects of NCPAP therapy on ﬁbrinogen levels in obstructive sleep apnea syndrome. Am J Respir Crit Care Med 1996; 153: 1972–76. Balfors EM, Franklin KA. Impairment of cerebral perfusion during obstructive sleep apneas. Am J Respir Crit Care Med 1994; 150: 1587–91. Klingelhofer J, Hajak G, Sander D, Schulz-Varszegi M, Ruther E, Conrad B. Assessment of intracranial hemodynamics in sleep apnea syndrome. Stroke 1992; 23: 1427–33. Drager LF, Bortolotto LA, Lorenzi MC, Figueiredo AC, Krieger EM, Lorenzi-Filho G. Early signs of atherosclerosis in obstructive sleep apnea. Am J Respir Crit Care Med 2005; 172: 613–18. Minoguchi K, Yokoe T, Tazaki T, et al. Increased carotid intima-media thickness and serum inﬂammatory markers in obstructive sleep apnea. Am J Respir Crit Care Med 2005; 172: 625–30. Minoguchi K, Yokoe T, Tazaki T, et al. Silent brain infarction and platelet activation in obstructive sleep apnea. Am J Respir Crit Care Med 2007; 175: 612–17. Young T, Shahar E, Nieto FJ, et al. Predictors of sleep-disordered breathing in community-dwelling adults: the Sleep Heart Health Study. Arch Intern Med 2002; 162: 893–900. Duran J, Esnaola S, Rubio R, Iztueta A. Obstructive sleep apnea-hypopnea and related clinical features in a population-based sample of subjects aged 30 to 70 yr. Am J Respir Crit Care Med 2001; 163: 685–89. Ip MS, Lam B, Lauder IJ, et al. A community study of sleep-disordered breathing in middle-aged Chinese men in Hong Kong. Chest 2001; 119: 62–69. Young T, Skatrud J, Peppard PE. Risk factors for obstructive sleep apnea in adults. JAMA 2004; 291: 2013–16.

www.thelancet.com Vol 373 January 3, 2009

62

63

64 65

66

67

68

69 70

71

72 73

74

75

76

77

78

79

80

81

82

83

84 85 86

Young T, Evans L, Finn L, Palta M. Estimation of the clinically diagnosed proportion of sleep apnea syndrome in middle-aged men and women. Sleep 1997; 20: 705–06. Peppard PE, Young T, Palta M, Dempsey J, Skatrud J. Longitudinal study of moderate weight change and sleep-disordered breathing. JAMA 2000; 284: 3015–21. Ryan CM, Bradley TD. Pathogenesis of obstructive sleep apnea. J Appl Physiol 2005; 99: 2440–50. Li KK, Kushida C, Powell NB, Riley RW, Guilleminault C. Obstructive sleep apnea syndrome: a comparison between Far-East Asian and white men. Laryngoscope 2000; 110: 1689–93. Buxbaum SG, Elston RC, Tishler PV, Redline S. Genetics of the apnea hypopnea index in Caucasians and African Americans: I. Segregation analysis. Genet Epidemiol 2002; 22: 243–53. Fletcher EC, DeBehnke RD, Lovoi MS, Gorin AB. Undiagnosed sleep apnea in patients with essential hypertension. Ann Intern Med 1985; 103: 190–95. Logan AG, Perlikowski SM, Mente A, et al. High prevalence of unrecognized sleep apnoea in drug-resistant hypertension. J Hypertens 2001; 19: 2271–77. Chan J, Sanderson J, Chan W, et al. Prevalence of sleep-disordered breathing in diastolic heart failure. Chest 1997; 111: 1488–93. Peker Y, Kraiczi H, Hedner J, Loth S, Johansson A, Bende M. An independent association between obstructive sleep apnoea and coronary artery disease. Eur Respir J 1999; 14: 179–84. Mooe T, Franklin KA, Holmstrom K, Rabben T, Wiklund U. Sleep-disordered breathing and coronary artery disease: long-term prognosis. Am J Respir Crit Care Med 2001; 164: 1910–13. Bassetti C, Aldrich MS. Sleep apnea in acute cerebrovascular diseases: ﬁnal report on 128 patients. Sleep 1999; 22: 217–23. Good DC, Henkle JQ, Gelber D, Welsh J, Verhulst S. Sleep-disordered breathing and poor functional outcome after stroke. Stroke 1996; 27: 252–59. Kaneko Y, Hajek VE, Zivanovic V, Raboud J, Bradley TD. Relationship of sleep apnea to functional capacity and length of hospitalization following stroke. Sleep 2003; 26: 293–97. Nieto FJ, Young TB, Lind BK, et al. Association of sleep-disordered breathing, sleep apnea, and hypertension in a large community-based study. Sleep Heart Health Study. JAMA 2000; 283: 1829–36. Punjabi NM, Sorkin JD, Katzel LI, Goldberg AP, Schwartz AR, Smith PL. Sleep-disordered breathing and insulin resistance in middle-aged and overweight men. Am J Respir Crit Care Med 2002; 165: 677–82. Punjabi NM, Shahar E, Redline S, Gottlieb DJ, Givelber R, Resnick HE. Sleep-disordered breathing, glucose intolerance, and insulin resistance: the Sleep Heart Health Study. Am J Epidemiol 2004; 160: 521–30. Reichmuth KJ, Austin D, Skatrud JB, Young T. Association of sleep apnea and type II diabetes: a population-based study. Am J Respir Crit Care Med 2005; 172: 1590–95. Mehra R, Benjamin EJ, Shahar E, et al. Association of nocturnal arrhythmias with sleep-disordered breathing: The Sleep Heart Health Study. Am J Respir Crit Care Med 2006; 173: 910–16. Flemons WW, Remmers JE, Gillis AM. Sleep apnea and cardiac arrhythmias. Is there a relationship? Am Rev Respir Dis 1993; 148: 618–21. Fletcher EC, Lesske J, Behm R, Miller CC 3rd, Stauss H, Unger T. Carotid chemoreceptors, systemic blood pressure, and chronic episodic hypoxia mimicking sleep apnea. J Appl Physiol 1992; 72: 1978–84. Stamler J, Stamler R, Neaton JD. Blood pressure, systolic and diastolic, and cardiovascular risks. US population data. Arch Intern Med 1993; 153: 598–615. Pratt-Ubunama MN, Nishizaka MK, Boedefeld RL, Coﬁeld SS, Harding SM, Calhoun DA. Plasma aldosterone is related to severity of obstructive sleep apnea in subjects with resistant hypertension. Chest 2007; 131: 453–59. Marney AM, Brown NJ. Aldosterone and end-organ damage. Clin Sci (Lond) 2007; 113: 267–78. Oltmanns KM, Gehring H, Rudolf S, et al. Hypoxia causes glucose intolerance in humans. Am J Respir Crit Care Med 2004; 169: 1231–37. Spiegel K, Leproult R, Van Cauter E. Impact of sleep debt on metabolic and endocrine function. Lancet 1999; 354: 1435–39.

91

Seminar

87

88

89

90

91

92

93

94

95

96 97

98

99

100

101

102

103 104 105

106

107

108

109

110

92

Ip MS, Lam B, Ng MM, Lam WK, Tsang KW, Lam KS. Obstructive sleep apnea is independently associated with insulin resistance. Am J Respir Crit Care Med 2002; 165: 670–76. Savransky V, Nanayakkara A, Li J, et al. Chronic intermittent hypoxia induces atherosclerosis. Am J Respir Crit Care Med 2007; 175: 1290–97. Mooe T, Rabben T, Wiklund U, Franklin KA, Eriksson P. Sleep-disordered breathing in women: occurrence and association with coronary artery disease. Am J Med 1996; 101: 251–56. Mooe T, Rabben T, Wiklund U, Franklin KA, Eriksson P. Sleep-disordered breathing in men with coronary artery disease. Chest 1996; 109: 659–63. Marin JM, Carrizo SJ, Vicente E, Agusti AG. Long-term cardiovascular outcomes in men with obstructive sleep apnoea-hypopnoea with or without treatment with continuous positive airway pressure: an observational study. Lancet 2005; 365: 1046–53. Buchner NJ, Sanner BM, Borgel J, Rump LC. Continuous positive airway pressure treatment of mild to moderate obstructive sleep apnea reduces cardiovascular risk. Am J Respir Crit Care Med 2007; 176: 1274–80. Scharf SM, Graver LM, Balaban K. Cardiovascular eﬀects of periodic occlusions of the upper airways in dogs. Am Rev Respir Dis 1992; 146: 321–29. Bradley TD, Hall MJ, Ando S, Floras JS. Hemodynamic eﬀects of simulated obstructive apneas in humans with and without heart failure. Chest 2001; 119: 1827–35. Scharf SM, Bianco JA, Tow DE, Brown R. The eﬀects of large negative intrathoracic pressure on left ventricular function in patients with coronary artery disease. Circulation 1981; 63: 871–75. Franklin KA, Nilsson JB, Sahlin C, Näslund U. Sleep apnoea and nocturnal angina. Lancet 1995; 345: 1085–87. Peker Y, Hedner J, Kraiczi H, Loth S. Respiratory disturbance index: an independent predictor of mortality in coronary artery disease. Am J Respir Crit Care Med 2000; 162: 81–86. Yumino D, Tsurumi Y, Takagi A, Suzuki K, Kasanuki H. Impact of obstructive sleep apnea on clinical and angiographic outcomes following percutaneous coronary intervention in patients with acute coronary syndrome. Am J Cardiol 2007; 99: 26–30. Milleron O, Pilliere R, Foucher A, et al. Beneﬁts of obstructive sleep apnoea treatment in coronary artery disease: a long-term follow-up study. Eur Heart J 2004; 25: 728–34. Doherty LS, Kiely JL, Swan V, McNicholas WT. Long-term eﬀects of nasal continuous positive airway pressure therapy on cardiovascular outcomes in sleep apnea syndrome. Chest 2005; 127: 2076–84. Nakashima H, Katayama T, Takagi C, et al. Obstructive sleep apnoea inhibits the recovery of left ventricular function in patients with acute myocardial infarction. Eur Heart J 2006; 27: 2317–22. Sin DD, Fitzgerald F, Parker JD, Newton G, Floras JS, Bradley TD. Risk factors for central and obstructive sleep apnea in 450 men and women with congestive heart failure. Am J Respir Crit Care Med 1999; 160: 1101–06. Arzt M, Bradley TD. Treatment of sleep apnea in heart failure. Am J Respir Crit Care Med 2006; 173: 1300–08. Yumino D, Bradley TD. Central sleep apnea and Cheyne-Stokes respiration. Proc Am Thor Soc 2008; 5: 226–36. Arzt M, Young T, Finn L, et al. Sleepiness and sleep in patients with both systolic heart failure and obstructive sleep apnea. Arch Intern Med 2006; 166: 1716–22. Chiu KL, Ryan CM, Shiota S, et al. Fluid shift by lower body positive pressure increases pharyngeal resistance in healthy subjects. Am J Respir Crit Care Med 2006; 174: 1378–83. Shiota S, Ryan CM, Chiu KL, et al. Alterations in upper airway cross-sectional area in response to lower body positive pressure in healthy subjects. Thorax 2007; 62: 868–72. Floras JS, Bradley TD. Treating obstructive sleep apnea: is there more to the story than 2 millimeters of mercury? Hypertension 2007; 50: 289–91. Arias MA, Garcia-Rio F, Alonso-Fernandez A, Mediano O, Martinez I, Villamor J. Obstructive sleep apnea syndrome aﬀects left ventricular diastolic function: eﬀects of nasal continuous positive airway pressure in men. Circulation 2005; 112: 375–83. Niroumand M, Kuperstein R, Sasson Z, Hanly PJ. Impact of obstructive sleep apnea on left ventricular mass and diastolic function. Am J Respir Crit Care Med 2001; 163: 1632–36.

111 Noda A, Okada T, Yasuma F, Nakashima N, Yokota M. Cardiac hypertrophy in obstructive sleep apnea syndrome. Chest 1995; 107: 1538–44. 112 Amin RS, Kimball TR, Bean JA, et al. Left ventricular hypertrophy and abnormal ventricular geometry in children and adolescents with obstructive sleep apnea. Am J Respir Crit Care Med 2002; 165: 1395–99. 113 Usui K, Parker JD, Newton GE, Floras JS, Ryan CM, Bradley TD. Left ventricular structural adaptations to obstructive sleep apnea in dilated cardiomyopathy. Am J Respir Crit Care Med 2006; 173: 1170–75. 114 Sampol G, Romero O, Salas A, et al. Obstructive sleep apnea and thoracic aorta dissection. Am J Respir Crit Care Med 2003; 168: 1528–31. 115 Young ME. The circadian clock within the heart: potential inﬂuence on myocardial gene expression, metabolism, and function. Am J Physiol Heart Circ Physiol 2006; 290: H1–16. 116 Yoshinaga K, Burwash IG, Leech JA, et al. The eﬀects of continuous positive airway pressure on myocardial energetics in patients with heart failure and obstructive sleep apnea. J Am Coll Cardiol 2007; 49: 450–58. 117 Sin DD, Fitzgerald F, Parker JD, et al. Relationship of systolic BP to obstructive sleep apnea in patients with heart failure. Chest 2003; 123: 1536–43. 118 Spaak J, Egri ZJ, Kubo T, et al. Muscle sympathetic nerve activity during wakefulness in heart failure patients with and without sleep apnea. Hypertension 2005; 46: 1327–32. 119 Roche F, Xuong AN, Court-Fortune I, et al. Relationship among the severity of sleep apnea syndrome, cardiac arrhythmias, and autonomic imbalance. Pacing Clin Electrophysiol 2003; 26: 669–77. 120 Tilkian AG, Guilleminault C, Schroeder JS, Lehrman KL, Simmons FB, Dement WC. Sleep-induced apnea syndrome. Prevalence of cardiac arrhythmias and their reversal after tracheostomy. Am J Med 1977; 63: 348–58. 121 Simantirakis EN, Schiza SI, Marketou ME, et al. Severe bradyarrhythmias in patients with sleep apnoea: the eﬀect of continuous positive airway pressure treatment: a long-term evaluation using an insertable loop recorder. Eur Heart J 2004; 25: 1070–76. 122 Grimm W, Hoﬀmann J, Menz V, et al. Electrophysiologic evaluation of sinus node function and atrioventricular conduction in patients with prolonged ventricular asystole during obstructive sleep apnea. Am J Cardiol 1996; 77: 1310–14. 123 Leung RS, Huber MA, Rogge T, Maimon N, Chiu KL, Bradley TD. Association between atrial ﬁbrillation and central sleep apnea. Sleep 2005; 28: 1543–46. 124 Kanagala R, Murali NS, Friedman PA, et al. Obstructive sleep apnea and the recurrence of atrial ﬁbrillation. Circulation 2003; 107: 2589–94. 125 Gami AS, Hodge DO, Herges RM, et al. Obstructive sleep apnea, obesity, and the risk of incident atrial ﬁbrillation. J Am Coll Cardiol 2007; 49: 565–71. 126 Gami AS, Howard DE, Olson EJ, Somers VK. Day-night pattern of sudden death in obstructive sleep apnea. N Engl J Med 2005; 352: 1206–14. 127 Gonzalez-Rothi RJ, Foresman GE, Block AJ. Do patients with sleep apnea die in their sleep? Chest 1988; 94: 531–38. 128 Yaggi HK, Concato J, Kernan WN, Lichtman JH, Brass LM, Mohsenin V. Obstructive sleep apnea as a risk factor for stroke and death. N Engl J Med 2005; 353: 2034–41. 129 Ferguson KA, Ono T, Lowe AA, Keenan SP, Fleetham JA. A randomized crossover study of an oral appliance vs nasal-continuous positive airway pressure in the treatment of mild-moderate obstructive sleep apnea. Chest 1996; 109: 1269–75. 130 Oliven A, O’Hearn DJ, Boudewyns A, et al. Upper airway response to electrical stimulation of the genioglossus in obstructive sleep apnea. J Appl Physiol 2003; 95: 2023–29. 131 Sher AE. Upper airway surgery for obstructive sleep apnea. Sleep Med Rev 2002; 6: 195–212. 132 Garrigue S, Bordier P, Jais P, et al. Beneﬁt of atrial pacing in sleep apnea syndrome. N Engl J Med 2002; 346: 404–12. 133 Krahn AD, Yee R, Erickson MK, et al. Physiologic pacing in patients with obstructive sleep apnea: a prospective, randomized crossover trial. J Am Coll Cardiol 2006; 47: 379–83.

www.thelancet.com Vol 373 January 3, 2009

Seminar

134 Luthje L, Unterberg-Buchwald C, Dajani D, Vollmann D, Hasenfuss G, Andreas S. Atrial overdrive pacing in patients with sleep apnea with implanted pacemaker. Am J Respir Crit Care Med 2005; 172: 118–22. 135 Pepin JL, Defaye P, Garrigue S, Poezevara Y, Levy P. Overdrive atrial pacing does not improve obstructive sleep apnoea syndrome. Eur Respir J 2005; 25: 343–47. 136 Barnes M, McEvoy RD, Banks S, et al. Eﬃcacy of positive airway pressure and oral appliance in mild to moderate obstructive sleep apnea. Am J Respir Crit Care Med 2004; 170: 656–64. 137 Gotsopoulos H, Kelly JJ, Cistulli PA. Oral appliance therapy reduces blood pressure in obstructive sleep apnea: a randomized, controlled trial. Sleep 2004; 27: 934–41. 138 Robinson GV, Smith DM, Langford BA, Davies RJ, Stradling JR. Continuous positive airway pressure does not reduce blood pressure in nonsleepy hypertensive OSA patients. Eur Respir J 2006; 27: 1229–35. 139 Kaneko Y, Floras JS, Usui K, et al. Cardiovascular eﬀects of continuous positive airway pressure in patients with heart failure and obstructive sleep apnea. N Engl J Med 2003; 348: 1233–41. 140 Mansﬁeld DR, Gollogly NC, Kaye DM, Richardson M, Bergin P, Naughton MT. Controlled trial of continuous positive airway pressure in obstructive sleep apnea and heart failure. Am J Respir Crit Care Med 2004; 169: 361–66. 141 Usui K, Bradley TD, Spaak J, et al. Inhibition of awake sympathetic nerve activity of heart failure patients with obstructive sleep apnea by nocturnal continuous positive airway pressure. J Am Coll Cardiol 2005; 45: 2008–11. 142 Gilman MP, Floras JS, Usui K, Kaneko Y, Leung RS, Bradley TD. Continuous positive airway pressure increases heart rate variability in heart failure patients with obstructive sleep apnoea. Clin Sci (Lond) 2008; 114: 243–49. 143 Ryan CM, Usui K, Floras JS, Bradley TD. Eﬀect of continuous positive airway pressure on ventricular ectopy in heart failure patients with obstructive sleep apnoea. Thorax 2005; 60: 781–85. 144 Egea CJ, Aizpuru F, Pinto JA, et al. Cardiac function after CPAP therapy in patients with chronic heart failure and sleep apnea: a multicenter study. Sleep Med 2007; published online September 27. DOI:10.1016/j.sleep.2007.06.018. 145 West SD, Nicoll DJ, Wallace TM, Matthews DR, Stradling JR. Eﬀect of CPAP on insulin resistance and HbA1c in men with obstructive sleep apnoea and type 2 diabetes. Thorax 2007; 62: 969–74. 146 Sandberg O, Franklin KA, Bucht G, Eriksson S, Gustafson Y. Nasal continuous positive airway pressure in stroke patients with sleep apnoea: a randomized treatment study. Eur Respir J 2001; 18: 630–34. 147 Hsu CY, Vennelle M, Li HY, Engleman HM, Dennis MS, Douglas NJ. Sleep-disordered breathing after stroke: a randomised controlled trial of continuous positive airway pressure. J Neurol Neurosurg Psychiatry 2006; 77: 1143–49.

www.thelancet.com Vol 373 January 3, 2009

148 Dimsdale JE, Loredo JS, Profant J. Eﬀect of continuous positive airway pressure on blood pressure: a placebo trial. Hypertension 2000; 35: 144–47. 149 Faccenda JF, Mackay TW, Boon NA, Douglas NJ. Randomized placebo-controlled trial of continuous positive airway pressure on blood pressure in the sleep apnea-hypopnea syndrome. Am J Respir Crit Care Med 2001; 163: 344–48. 150 Pepperell JC, Ramdassingh-Dow S, Crosthwaite N, et al. Ambulatory blood pressure after therapeutic and subtherapeutic nasal continuous positive airway pressure for obstructive sleep apnoea: a randomised parallel trial. Lancet 2002; 359: 204–10. 151 Becker HF, Jerrentrup A, Ploch T, et al. Eﬀect of nasal continuous positive airway pressure treatment on blood pressure in patients with obstructive sleep apnea. Circulation 2003; 107: 68–73. 152 Campos-Rodriguez F, Grilo-Reina A, Perez-Ronchel J, et al. Eﬀect of continuous positive airway pressure on ambulatory BP in patients with sleep apnea and hypertension: a placebo-controlled trial. Chest 2006; 129: 1459–67. 153 Norman D, Loredo JS, Nelesen RA, et al. Eﬀects of continuous positive airway pressure versus supplemental oxygen on 24-hour ambulatory blood pressure. Hypertension 2006; 47: 840–45. 154 Bazzano LA, Khan Z, Reynolds K, He J. Eﬀect of nocturnal nasal continuous positive airway pressure on blood pressure in obstructive sleep apnea. Hypertension 2007; 50: 417–23. 155 Haentjens P, Van Meerhaeghe A, Moscariello A, et al. The impact of continuous positive airway pressure on blood pressure in patients with obstructive sleep apnea syndrome: evidence from a meta-analysis of placebo-controlled randomized trials. Arch Intern Med 2007; 167: 757–64. 156 Alajmi M, Mulgrew AT, Fox J, et al. Impact of continuous positive airway pressure therapy on blood pressure in patients with obstructive sleep apnea hypopnea: a meta-analysis of randomized controlled trials. Lung 2007; 185: 67–72. 157 Drager LF, Bortolotto LA, Figueiredo AC, Krieger EM, Lorenzi GF. Eﬀects of continuous positive airway pressure on early signs of atherosclerosis in obstructive sleep apnea. Am J Respir Crit Care Med 2007; 176: 706–12. 158 Ruttanaumpawan P, Gilman MP, Usui K, Floras JS, Bradley TD. Sustained eﬀect of continuous positive airway pressure on baroreﬂex sensitivity in congestive heart failure patients with obstructive sleep apnea. J Hypertens 2008; 26: 1163–68. 159 Bradley TD, Logan AG, Kimoﬀ RJ, et al. Continuous positive airway pressure for central sleep apnea and heart failure. N Engl J Med 2005; 353: 2025–33. 160 Smith PL, Gold AR, Meyers DA, Haponik EF, Bleecker ER. Weight loss in mildly to moderately obese patients with obstructive sleep apnea. Ann Intern Med 1985; 103: 850–55. 161 Barbe F, Mayoralas LR, Duran J, et al. Treatment with continuous positive airway pressure is not eﬀective in patients with sleep apnea but no daytime sleepiness. a randomized, controlled trial. Ann Intern Med 2001; 134: 1015–23.

93

Case Report

Unexplained seizures in an infant Marinella Astuto, Carmelo Minardi, Giuliana Rizzo, Antonino Gullo Lancet 2009; 373: 94 Department of Anaesthesia and Intensive Care, Catania University Hospital, Catania, Italy (M Astuto MD, C Minardi MD, G Rizzo MD, Prof A Gullo MD) Correspondence to: Dr Marinella Astuto, Department of Anaesthesia and Intensive Care, Catania University Hospital, Catania, Italy [email protected]

In June, 2004, a girl was delivered by caesarean section, at 38 weeks, after the placenta became detached. When she was 2 months old, the girl was brought to hospital with generalised tonic-clonic seizures, tremors in arms and legs, hypotonia, and inconsolable crying. Blood tests, EEG, and MRI of the head and spine showed nothing abnormal. The girl’s mother, a housewife (the father was a farmer) was prescribed amitriptyline for postnatal depression and tension headache; we knew of no other family history of neurological illness. Concluding that the girl had epilepsy, doctors prescribed phenobarbital; the hypotonia was treated with physiotherapy. However, in early 2005, the girl had two further episodes of seizures. Each time, she had serial seizures for 20 s, then cried inconsolably for 5–7 min, and slept for 8–12 h. On hospitalisation, she had similar episodes almost daily. Finding the presentation inexplicable and unusual, doctors consulted international experts, who suggested that the girl might have a rare genetic condition, or intermittent hemiplegia. Flunarizine was prescribed, but stopped after 3 days, because of extrapyramidal sideeﬀects. Subsequently, the girl was admitted to hospital around once a month, with seizures, which resolved on treatment. After the girl was admitted to our paediatric ward in September, 2005, doctors witnessed a seizure lasting 20 s, and prescribed diazepam, then phenobarbital. Electroencephalography (EEG) showed focal seizure activity. After the seizure, doctors monitored the girl’s EEG and electrocardiogram (ECG). Around 30 min after the seizure, the ECG showed ventricular tachycardia, which developed into ventricular ﬁbrillation. The girl was given cardiopulmonary resuscitation, and transferred to the intensive-care department, where she was deﬁbrillated, and given epinephrine and lidocaine, as

Panel: Some psychiatric disorders that typically present as physical illness • Factitious disorder: the patient, apparently unaware of his motives, seeks medical attention by the intentional production or feigning of symptoms • Munchausen’s syndrome: a type of factitious disorder, characterised by pathological lying and peregrination • Factitious disorder by proxy (=Munchausen’s syndrome by proxy): a caregiver, with unmet psychological needs, seeks medical attention by fabricating or creating an illness in a child • Malingering: the patient knowingly fabricates a medical illness for known gain. • Somatisation: the patient is preoccupied by physical symptoms, and believes them to be physically caused, despite medical opinions to the contrary • Hypochondriasis: the patient believes himself to have a serious disease, despite medical opinions to the contrary • Dissociative (conversion) disorder: grief or anxiety manifests as a neurological or possession syndrome, of no neurological cause

94

well as 0∙9% saline. Sinus rhythm returned. The girl was now haemodynamically stable; however, her Glasgowcoma-scale score was 8. Her pupils were reactive but dilated, and she had generalised hypotonia. We sedated the patient with midazolam, intubated her, and ventilated her mechanically. The next morning, the girl was breathing independently, moving all her limbs well, and speaking. After extubation, we told the mother that her child was getting better. The mother did not seem relieved; she almost insisted that the girl’s condition was unimproved. Finding this incongruous, we requested toxicological analysis of blood and urine samples from the infant. Concentrations of anticonvulsant drugs were in the therapeutic range. However, the serum concentration of amitriptyline was 350 μg/L; the serum concentration of nortriptyline (a metabolite of amitriptyline), and the urine concentrations of amitriptyline and nortryptyline, were higher than our laboratory could quantify. We contacted the police. The mother admitted that, since the child was 1 month old, she had been administering amitriptyline drops to her. The child was removed from the family home, to live with her maternal grandmother. Subsequently, the girl has been seen every 3 months by a paediatrician and a neurologist. When last seen, in July, 2008, she was well, and well looked after. Poisoning is one of the most common medical emergencies of childhood; and tricyclic antidepressants are frequent causes of poisoning in adults and children.1 However, deliberate poisoning is quite rare. We were unable to interview the mother at length, so know little about her motivations for poisoning the child. However, our patient may have been a victim of factitious disorder by proxy, also known as Munchausen’s syndrome by proxy (panel).2,3 If our surmise is correct, the mother did not primarily intend to kill the child. Nonetheless, she could have done. In children, amitriptyline overdose commonly causes seizures and tachycardia, as well as lethargy, hyperglycaemia, and leucocytosis;4 in adults, a serum concentration of 1000 μg/L is associated with a high risk of death.5 References 1 Bronstein AC, Spyker DA, Cantilena LR Jr, Green J, Rumack BH, Heard SE. 2006 annual report of the American Association of Poison Control Centers’ National Poison Data System (NPDS). Clin Toxicol 2007; 45: 815–917. 2 Turner J, Reid S. Munchausen’s syndrome. Lancet 2002; 359: 346–49. 3 Schreier H. Munchausen by proxy deﬁned. Pediatrics 2002; 110: 985–88. 4 Caksen H, Akbayram S, Odabaş D, et al. Acute amitriptyline intoxication: an analysis of 44 children. Hum Exp Toxicol 2006; 25: 107–10. 5 Ash SR, Levy H, Akmal M, et al. Treatment of severe tricyclic antidepressant overdose with extracorporeal sorbent detoxiﬁcation. Adv Ren Replace Ther 2002; 9: 31–41.

www.thelancet.com Vol 373 January 3, 2009