(3 November 2007-9 November 2007)))lib.ajaums.ac.ir/booklist/986962.pdfThe Lancet Volume 370, Issue 9598, Pages 1521-1588 (3 November 2007-9 November 2007))) 1. Time to supersize control

The Lancet Volume 370, Issue 9598, Pages 1521-1588 (3 November 2007-9 November 2007)))

1. Time to supersize control efforts for obesity

Page 1521 The Lancet

2. New standards for the health and development of children Page 1522 The Lancet

3. A prescription for psychosocial cancer care Page 1522 The Lancet

4. What will a partly protective malaria vaccine mean to mothers in Africa? Pages 1523-1524 Judith E Epstein

5. Food colourings, preservatives, and hyperactivity Pages 1524-1525 Philippe A Eigenmann and Charles A Haenggeli

6. Counting for health Page 1526 Richard Horton

7. Improving health statistics in Africa Pages 1527-1528 Ngozi Okonjo-Iweala and Philip Osafo-Kwaako

8. Obesity plan lacks foresight Pages 1528-1529 Andrew Jack

9. Lancet/Global Forum essay competition winners Page 1529 Sarah Ramsay

10. Clinical update: cyanotic adult congenital heart disease Pages 1530-1532 Mark S Spence, Michelle S Balaratnam and Michael A Gatzoulis

11. Aboriginal workers key to indigenous health in Australia Pages 1533-1536 Margaret Harris Cheng

12. Carla AbouZahr: making health statistics count Page 1537 Sarah Boseley

13. Bjørn Aage Ibsen Page 1538 Stephen Pincock

14. Cannabis and psychosis Page 1539 John Macleod, George Davey Smith, Matthew Hickman and Matthias Egger

15. Cannabis and psychosis – Authors' reply Pages 1539-1540 Stanley Zammit, Theresa Moore and Glyn Lewis

16. Cannabis and psychosis Page 1540 Daniele Fabio Zullino, Thomas Rathelot and Yasser Khazaal

17. Cannabis and psychosis Page 1540 O Schulte-Herbrüggen and MC Jockers-Scherübl

18. UK classification of cannabis: is a change needed and why? Page 1541 Louisa Degenhardt, Wayne D Hall, Amanda Roxburgh and Richard P Mattick

19. Waist circumference in metabolic syndrome Pages 1541-1542 Satoru Yamada, Youko Tsukamoto and Junichiro Irie

20. New growth standards Page 1542 Colin Binns and MiKyung Lee

21. Department of Error Page 1542

22. Department of Error Page 1542

23. Safety of the RTS,S/AS02D candidate malaria vaccine in infants living in a highly endemic area of Mozambique: a double blind randomised controlled phase I/IIb trial Pages 1543-1551 John J Aponte, Pedro Aide, Montse Renom, Inacio Mandomando, Quique Bassat, Jahit Sacarlal, M Nelia Manaca, Sarah Lafuente, Arnoldo Barbosa, Amanda Leach, Marc Lievens, Johan Vekemans, Betuel Sigauque, Marie-Claude Dubois, Marie-Ange Demoitié, Marla Sillman, Barbara Savarese, John G McNeil, Eusebio Macete, W Ripley Ballou, et al.

24. Cost-effectiveness of drug-eluting stents in patients at high or low risk of major cardiac events in the Basel Stent KostenEffektivitäts Trial (BASKET): an 18-month analysis Pages 1552-1559 Hans Peter Brunner-La Rocca, Christoph Kaiser, Alain Bernheim, Michael J Zellweger, Raban Jeger, Peter T Buser, Stefan Osswald and Matthias Pfisterer

25. Food additives and hyperactive behaviour in 3-year-old and 8/9-year-old children in the community: a randomised, double-blinded, placebo-controlled trial Pages 1560-1567 Donna McCann, Angelina Barrett, Alison Cooper, Debbie Crumpler, Lindy Dalen, Kate Grimshaw, Elizabeth Kitchin, Kris Lok, Lucy Porteous, Emily Prince, Edmund Sonuga-Barke, John O Warner and Jim Stevenson

26. Virchow's node Page 1568 Daniel C Baumgart and Andreas Fischer

27. A scandal of invisibility: making everyone count by counting everyone Pages 1569-1577 Philip W Setel, Sarah B Macfarlane, Simon Szreter, Lene Mikkelsen, Prabhat Jha, Susan Stout and Carla AbouZahr

28. Sudden infant death syndrome Pages 1578-1587 Rachel Y Moon, Rosemary SC Horne and Fern R Hauck

29. Orange and green monkeys jumping around the room Page 1588 Sunku Hemanth Guptha, Tha Han, Qais Arafat and Okubadejo Deyo

Editorial

www.thelancet.com Vol 370 November 3, 2007 1521

Time to supersize control eff orts for obesityThe Health Profi le of England 2007, released on Oct 22, is an unfl attering account of the nation’s health. While the best-ever fi gures for life expectancy and infant mortality deserve praise, compared with the 15 pre-2004 EU member countries, the unhealthy English lifestyle is fuelling preventable mortality, diabetes, and obesity.

The need to reduce obesity (defi ned by WHO as a body-mass index ≥30 kg/m²) stands out as a major public-health priority. Over the past decade, UK rates have increased by 30% in women, 40% in men, and 50% in children. Britons are now among the world’s most obese people: 23% of adults are obese compared with an EU average of 12·6% in other pre-2004 countries.

The problem is highlighted further in the Oct 17 report Tackling Obesities: Future Choices, by the Government’s Foresight Programme. This estimates that more than 50% of the UK population could be obese by 2050. Obesity exacts a high toll on quality of life, and the attendant comorbidity of diabetes, hypertension, cardiovascular disease, musculoskeletal conditions, and some cancers strains NHS resources. The fi nancial cost for obesity-related treatments in 2002 was £1 billion. Estimated indirect costs of £7 billion to business and society are predicted to rise to £45 billion by 2050.

Berating people to eat less and exercise more is futile when energy-dense foods are cheaper and more readily available than healthier alternatives, and when cities have been designed for cars rather than for recreation. More nutritious foods and healthier activities must become so desirable and readily accessible that their uptake is normal, and unhealthy options so inconvenient and unfashionable as to discourage their use.

The Foresight report avoids blaming gluttony and sloth, but instead fi nds that contemporary lifestyles pre-dis pose to passive obesity. While the authors are cor rect to highlight the complexity of factors leading to weight gain and retention, their view is defeatist: behavioural change becomes an insurmountable challenge due to barriers perceived as byproducts of modern lifestyles and beyond an individual’s control. Yet in reality, obesity ought to be a preventable—and reversible—condition, to which an educated population, motivated government, and conscientious food industry can respond. Individuals seldom choose obesity, but are seduced by its harbingers: convenience and economy.

Nonetheless, change is possible as the experience of Finland shows. In the 1970s Finland led the world in cardiac mortality and in the 1980s had an obesity prevalence twice that of the UK. Now 94% of Finns aged 19–65 years take regular exercise, children receive nutritious meals at school, vegetable and fruit con-sumption has increased three-fold, and adult obesity has fallen to 12·8%. Culture change was accomplished by consulting locally to understand and improve the motivation for healthier lifestyles. Interventions included better opportunities for exercise and inter-town competitions to reduce cholesterol levels. Impor-tantly, the food industry was involved in the transition and diversifi ed accordingly.

Although causes of obesity are recognised, prevention and treatment are less well understood. This is a time to consolidate best practice—by promoting breastfeeding, fruit and vegetable consumption, low glycaemic-index diets, and smaller portion sizes—and to explore new interventions, which should be evaluated carefully to inform future practice. The private sector can act by providing healthier food to employees and incorporating physical activity at work, such as making stairs more attractive and convenient. IBM, which increased prod-uctivity and saved US$100 million after introducing an employee fi tness programme 5 years ago, is now providing monetary incentives to staff whose children enrol in obesity education. Most importantly, the UK food industry, the country’s largest manufacturing division with a £70 billion annual turnover, will be key to improving the nation’s health.

The challenge of obesity brings exciting opportunities and large health benefi ts for minor reductions in weight. Just as individuals require initiative, time, income, and multiple interventions to overcome an obesogenic lifestyle, so too will a country’s eff orts require adequate leadership and resources involving many agencies. To facilitate such changes, an apoliti-cal, cross-departmental approach is required, led by an individual who can be responsible for directing and coordinating eff orts, so that henceforth preven-tion and treatment of obesity are interwoven syner-gistically with other initiatives. The obesity epidemic is here and further inactivity will only magnify the consequences. ■ The Lancet

For the Health Profi le of England 2007 see http://www.dh.gov.uk/en/Publicationsandstatistics/Publications/PublicationsStatistics/DH_079716

For the Tackling Obesities: Future Choices report see http://www.foresight.gov.uk/Obesity/Obesity_fi nal/Index.html

See Comment page 1528

Gett

y Im

ages

Great North Run 2007

Editorial

1522 www.thelancet.com Vol 370 November 3, 2007

New standards for the health and development of childrenLast week, WHO published the fi rst internationally agreed classifi cation to document the health and development of children and adolescents. This code extends and builds on the general International Classifi cation of Functioning, Disability and Health (ICF) published in 2001. It is part of the family of classifi cations, of which the International Classifi cation of Diseases (ICD) is the best known. All these codes aim to provide a common language to facilitate comparisons and standardisation in clinical, public-health, and research settings.

The new children’s code is the most complicated one since it rightly puts strong emphasis on diff erent stages of development and on the crucial infl uence of the child’s environment—the immediate family, the living conditions, and the wider societal structure—on health and development. Adverse environmental factors have a much stronger eff ect on children than on adults. At diff erent stages of development, each of these components has a more or less prominent role. The young child’s development is strongly linked to the immediate family, whereas with increasing

independence a teenager’s wellbeing is more dependent on peers and the community. Lack of fundamental elements in the environment, such as nutritious food and clean water, not only leads to diseases but also impairs physical and educational development.

In the explanation to the code, the authors stress that accurate coding requires training and a thorough knowledge of the normal changes in function with development. With all the interlinked factors and the fact that many chronic childhood diseases are complex, applying the code is not an easy task. In addition, the subjective impression of children’s own disability might diff er from the perception of adults. Heather Dickinson and colleagues showed, for example, that the self-reported quality of life of children with cerebral palsy was the same as for other children.

Nevertheless, this new code is a welcome starting point to highlight the unique physical, social, and psychological characteristics of children and to re-emphasise every child’s right to health, and to access to health care and education. ■ The Lancet

For the WHO ICF-CY classifi cation see

http://www.who.int/mediacentre/news/releases/2007/

pr59/en/index.html

For the study in children with cerebral palsy see Lancet 2007;

369: 2171–78

A prescription for psychosocial cancer careAnxiety, depression, and fatigue are just some of the debilitating conditions that can accompany a diagnosis of cancer or a course of cancer treatment. Lack of transportation to health-care appointments, fi nancial problems, and a lack of health insurance are also factors that can adversely aff ect a person’s ability to recover from the disease. However, all-too-often, these psychosocial factors do not receive the attention from cancer-care providers that they deserve, according to a report by the US Institute of Medicine (IOM) released on Oct 23.

To address this gap, the report—Cancer care for the whole patient: meeting psychosocial health needs—presents a new standard of care for patients, which recognises the role of psychological and social factors on people’s ability to function, their quality of life, and disease progression. The report found that, largely thanks to the voluntary sector, a wide range of eff ective psychosocial services are available to patients in the USA. But interventions to identify patients’ psychosocial needs and link them to appropriate services were often lacking. The report makes

ten recommendations for action including the adoption of the new standard of care by health-care providers, the development of performance measures for psychosocial care, and an education drive to empower patients and their family caregivers to expect, and request, cancer care that meets the standard for psychosocial care.

The IOM recommendations provide practical steps towards improving the quality of psychosocial care in the USA. The document deserves to be widely read, not only because it is relevant to the wellbeing of the 10·5 million US people who live with a past or current diagnosis of cancer, but also because it is applicable to other serious illnesses.

Psychosocial factors are problems in their own right but they can also have knock-on eff ects on physical health. Depression, for example, might prevent a patient from complying with their treatment plan. Health-care providers need to realise this link and their responsibility to manage and treat all aspects of their patients’ health—physical and psychosocial. ■ The Lancet

For the IOM report see http://www.iom.edu/CMS/

3809/34252/47228.aspx

Scie

nce

Phot

o Li

brar

y Pa

nos P

ictur

es

The printed journal includes an image merely for illustration

Comment


What will a partly protective malaria vaccine mean to mothers in Africa?

See Articles page 1543An eff ective vaccine would have tremendous benefi t for the estimated 3 billion people living at risk of malaria. In today’s Lancet, John Aponte and colleagues report that the candidate malaria vaccine RTS,S/AS02D is safe, immunogenic, and partly protective against infection in an infant population.1 The primary endpoint was safety. Effi cacy against infection, a secondary endpoint, was assessed during a 3-month follow-up after the last dose of vaccine. Aponte’s report represents the next step in the systematic clinical evaluation of this recombinant protein vaccine with adjuvant, which started more than 15 years ago and has included studies in adults in the USA2–4 and in The Gambia,5 in children aged 1–4 years in Mozambique,6,7 and now in infants in Mozambique. The teams that began working on this vaccine and its forerunners8,9 in the mid-1980s, and brought the vaccine to this point in development, deserve our praise and respect for their dedication, perseverance, and accomplishments.

Previous trials with the RTS,S vaccine in human beings also showed safety and part-protective effi cacy. In Gambian adults, vaccine effi cacy was 71% in the 9 weeks after last vaccination, and 0% in the next 6 weeks.5 In a cohort of children aged 1–4 years in Mozambique, effi cacy for delay of fi rst infections in the 6 months’ follow-up was 45%;6 however, nearly all had at least one infection (83% in the RTS,S group and 93% in the control group). Exciting results were derived from a second cohort of children aged 1–4 years followed up for clinical episodes (febrile clinical malaria).6,7 The rate of acquisition of fi rst clinical episode was, on average, 35% lower in infants vaccinated with RTS,S than in the control group through 18 months of follow-up;6 25% in the RTS,S group and 34% in the control group had at least one episode, a diff erence of 26%. For severe malaria, the rate of acquisition of the disease was 49% less in the RTS,S group. In Aponte and colleagues’ study in infants, effi cacy in delaying time to fi rst infection was 65·9% (95% CI 42·6–79·8%). Effi cacy was defi ned as 1 minus the hazard ratio, and was adjusted for distance to health facility and community. 22 of 93 participants (23·7%) in the RTS,S group and 46 of 92 (50·0%) in the control group had at least one infection, to give a diff erence of 52·6%.

To assess vaccine effi cacy, the investigators mainly used time-to-event analysis based on a proportional hazards regression, an appropriate method for assessing time to outcome with a partly protective vaccine and one that also takes into account varying follow-up for participants. For instance, in the 1–4 year-old children in Mozambique, vaccine effi cacy was 45% for delaying fi rst infection over the fi rst 6 months. However, many are more accustomed to thinking of vaccine effi cacy as being based on the proportion of vaccinated participants versus controls with the outcome during a defi ned follow-up. With the proportional analysis approach for the same data, the vaccine effi cacy against acquiring an infection by age 6 months was 11%. For a vaccine like RTS,S, which delays but does not necessarily prevent infection or clinical malaria, time-to-event analysis is needed. Nevertheless, because we are not certain whether time-to-event or proportional analysis will ultimately predict the long-term eff ect of the vaccine and because of the complexities of assessing a partly protective vaccine, I believe that it would be helpful for the authors to state the results of both types of analyses in future publications.

WH

O/C

hrist

ophe

r Bla

ck

Uganda: child being treated for malaria

Comment


What might be the benefi ts to the population as a whole of delaying infection and disease? Some experts have predicted that the eff ect of the introduction of a partly protective vaccine will be reduction in morbidity and mortality in the fi rst years of life, with negligible eff ect on transmission.10 Data from phase III trials, scheduled to begin in multiple African sites in 2008, will be crucial for determination of the public-health usefulness of the vaccine. Pending licensure, decisions will need to be made about whether or not this vaccine should be used widely and, if so, how best to incorporate it with existing control measures, such as insecticide-treated bednets and early diagnosis and treatment of malaria.

Finally, what do these results mean to the mother who brings her young child to a clinic in sub-Saharan Africa, or to the policymaker who is deciding what malaria controls to use when funds are limited? We can tell them that we are indeed closer to having a vaccine for malaria which might reduce morbidity and mortality. Nevertheless, the malaria-vaccine community must continue eff orts to develop an even more effi -cacious candi date vaccine; ideally, a vaccine that can one day prevent every child from ever having or dying from this dis ease. Dedicated investigators worldwide are pursuing numerous approaches, including prime-boost and adeno virus-vectored vaccines. A radiation-attenu-ated sporozoite vaccine is expected to undergo initial testing in human beings in 2008.11 The next 5–10 years will probably be the most exciting in the long journey to bring a malaria vaccine to the developing world.

Judith E EpsteinUS Military Malaria Vaccine Program, Naval Medical Research Center, Silver Spring, MD 20910, [email protected]

I thank Peter G Smith, London School of Hygiene and Tropical Medicine, London, UK, for helpful discussions and comments. I am in military service and prepared this Comment as part of my duties. The views in this Comment are mine and do not necessarily refl ect the offi cial policy or position of the US Department of the Navy, Department of Defense, or Government.

1 Aponte JJ, Aide P, Renom M, et al. Safety of the RTS,S/AS02D candidate malaria vaccine in infants living in a highly endemic area of Mozambique: a double blind randomised controlled phase I/IIb trial. Lancet 2007; 370: 1543–51.

2 Gordon DM, McGovern TW, Krzych U, et al. Safety, immunogenicity, and effi cacy of a recombinantly produced Plasmodium falciparum circumsporozoite protein-hepatitis B surface antigen subunit vaccine. J Infect Dis 1995; 171: 1576–85.

3 Stoute JA, Slaoui M, Heppner DG, for the RTS,S Malaria Vaccine Evaluation Group. A preliminary evaluation of a recombinant circumsporozoite protein vaccine against Plasmodium falciparum malaria. N Engl J Med 1997; 336: 86–91.

4 Kester KE, McKinney DA, Tornieporth N, et al. Effi cacy of recombinant circumsporozoite protein vaccine regimens against experimental Plasmodium falciparum malaria. J Infect Dis 2001; 183: 640–47.

5 Bojang KA, Milligan PJ, Pinder M, et al. Effi cacy of RTS,S/AS02A malaria vaccine against Plasmodium falciparum infection in semi-immune adult men in The Gambia: a randomised trial. Lancet 2001; 358: 1927–34.

6 Alonso PL, Sacarlal J, Aponte JJ, et al. Effi cacy of the RTS,S/AS02A vaccine against Plasmodium falciparum infection and disease in young African children: randomised controlled trial. Lancet 2004; 364: 1411–20.

7 Alonso PL, Sacarlal J, Aponte JJ, et al. Duration of protection with RTS,S/AS02A malaria vaccine in prevention of Plasmodium falciparum disease in Mozambican children: single-blind extended follow-up of a randomised controlled trial. Lancet 2005; 366: 2012–18.

8 Young JF, Hockmeyer WT, Gross M, et al. Expression of Plasmodium falciparum circumsporozoite proteins in Escherichia coli for potential use in a human malaria vaccine. Science 1985; 228: 958–62.

9 Ballou WR, Hoff man SL, Sherwood JA, et al. Safety and effi cacy of a recombinant DNA Plasmodium falciparum sporozoite vaccine. Lancet 1987; 1: 1277–81.

10 Maire N, Tediosi F, Ross A, et al. Predictions of the epidemiologic impact of introducing a pre-erythrocytic vaccine into the expanded program on immunization in sub-Saharan Africa. Am J Trop Med Hyg 2006; 75: 111–18.

11 Luke TC, Hoff man SL. Rationale and plans for developing a non-replicating,metabolically active, radiation-attenuated Plasmodium falciparum sporozoite vaccine. J Exp Biol 2003; 206: 3803–08.

Food colourings, preservatives, and hyperactivityQuod ali cibus est aliis fuat acre venenum (What is food to one, is bitter poison to others) is attributed to Lucretius (99–55 BC), and the sentiment could well express serious manifestations of hypersensitivity to food. In ancient times, when eating was a major challenge for survival, the perception of adverse reactions to nutrients was probably very diff erent from nowadays. Today, in affl uent parts of the world, adverse reactions to food can be divided into two categories: one being what is perceived as related to food but is only disrupting daily life; the other potentially leading to severe illness, such as IgE-mediated food allergy.

Feingold’s report on hyperkinesis and learning dis-abilities associated with artifi cial food fl avours and colours, published in the 1970s,1 fuelled a large debate and can be attributed to the fi rst category of adverse reactions to food mentioned above. The few pro spective studies published afterwards have been sum marised in a meta-analysis.2 The study in 2004 by Bateman and colleagues3 generated several comments, including one in The Lancet.4 Bateman recruited a sample of 3-year-old children from the general population. A subgroup completed an investigation that included skin-prick tests for allergy and double-blind placebo-controlled

See Articles page 1560

Comment


food challenges, as well as an assessment for hyperactive behaviour. This study revealed a dis crepancy between objective behavioural testing by health professionals, which showed no infl uence of food colouring and additives, and subjective observations of hyperactivity by parents in the group of children getting food colouring and additives. But several questions remained open after careful analysis, because the study emphasised the subjective impression of food-related hyperkinesis but did not provide clear evidence to the professional that foods need to be considered in the assessment of the possible causes of hyperkinesis.

The same group of investigators started a new study, which is published by Donna McCann and colleagues in today’s Lancet.5 In addition to 3-year-old children, a second group of 8–9-year-old children from a general population were studied. Children with hyperactivity were identifi ed and tested with a mixture of food colourings and additives that corresponded to an average daily intake during 1-week crossover challenge periods. Various measures of behaviour by health professionals, parents, or teachers were the primary outcome measures. The study showed statistically signifi cant diff erences in hyperactivity scores when the children were on or off food colourings and additives, as shown by increased hyperactivity scores in both age groups. Interestingly, the authors reported “substantial individual diff erences in the response of children to the additives”. This observation corresponds to the clinical observation of high sensitivity to food colouring and additives in some hyperactive children, and unresponsiveness in others. The researchers are currently characterising these subgroups.

McCann and colleagues did not address intra-individual diff erences over time, but one might speculate that some children have periods of hyperexcitability that correlate with increased sensitivity to food colouring and additives. This idea raises the question of the amount of food containing colourings and additives that can reasonably be eaten by children.

Overall, McCann and colleagues have shown that 3-year-old and 8–9-year-old children might have signs of hyperactivity when on a diet containing usual amounts of colouring and additives. What practical implications does this have for the community? Physicians taking care of children with hyperactivity could advise parents who wish to do so to start an elimination trial of artifi cial colourings and additives. Complete restriction is not

warranted according to McCann’s results. Only some hyperactive children will respond favourably to such a trial, and further research might highlight subgroups of individuals particularly susceptible to colouring and additives. Elimination measures should then focus on these children.

Artifi cial colourings and additives represent only one aspect of hyperactivity in children. Eff orts from health professionals and parents should also focus on adequate drug treatment and psychological measures, and rel-evant guidelines remain valid.6

*Philippe A Eigenmann, Charles A HaenggeliChildren’s University Hospital, 1211 Geneva, Switzerland [email protected]

We declare that we have no confl ict of interest.

1 Feingold BF. Hyperkinesis and learning disabilities linked to artifi cial food fl avors and colors. Am J Nurs 1975; 75: 797–803.

2 Schab DW, Trinh NH. Do artifi cial food colors promote hyperactivity in children with hyperactive syndromes? A meta-analysis of double-blind placebo-controlled trials. J Dev Behav Pediatr 2004; 25: 423–34.

3 Bateman B, Warner JO, Hutchinson E, et al. The eff ects of a double blind, placebo controlled, artifi cial food colourings and benzoate preservative challenge on hyperactivity in a general population sample of preschool children. Arch Dis Child 2004; 89: 506–11.

4 Eigenmann PA, Haenggeli CA. Food colourings and preservatives—allergy and hyperactivity. Lancet 2004; 364: 823–24.

5 McCann D, Barrett A, Cooper A, et al. Food additives and hyperactive behaviour in 3-year-old and 8/9-year-old children in the community: a randomised, double-blinded, placebo-controlled trial. Lancet 2007; 370: 1560–67.

6 Brown RT, Amler RW, Freeman WS, et al. Treatment of attention-defi cit/hyperactivity disorder: overview of the evidence. Pediatrics 2005; 115: e749–57.

Scie

nce

Phot

o Li

brar

y

Comment


Counting for health“The single most critical failure of development over the past 30 years”.1 This judgment comes from the team leading Who Counts?, a Series beginning today in The Lancet on the status of country and global efforts to register every birth and death, and to certify every cause of death. This “scandal of invisibility”2 means that millions of human beings are born and die without leaving any record of their existence. Over three-quarters of them are to be found in sub-Saharan Africa and south-east Asia.

Why does civil registration matter? Tracing the im print of a person’s existence not only confirms their citizenship, but also represents the first step in securing their right to life, freedom, and protection. Registration enables access to services and provides crucial data to policymakers charged with designing and planning health systems. Most fundamentally of all, registration is proof that a state recognises and respects the lives of those it has a responsibility to defend and develop.

Counting human lives and deaths is a pressing priority. Resources invested into global health have been scaled up massively in recent years. Without proper systems in place to monitor and track the impact of these new fi nancial fl ows, donors will do a poor job of holding themselves accountable for their investments.

Sadly, there has been, in the words of the Who Counts? team,3 “widespread neglect” of this issue. Little progress has been achieved during the past four decades. Countries and global institutions have paid only limited attention to vital registration. Today, less than a third of the world’s population is covered by accurate data on births and deaths. Far greater global urgency needs to be injected into this challenge. UN agencies, such as WHO, must do a better job of coordinating their eff orts and supporting countries. The Who Counts? team reaches the verdict that WHO “has made little progress”.3 This conclusion is all the more hard hitting since the team includes senior WHO scientists.

At country level, robust and eff ective national statistics systems are essential if registration is to become a reality. An independent national statistics service requires strong government ministries, a functioning legal system and civil service, devolved

local information networks to collect registration data, and a vocal civil society to press governments to act. The health sector can be an important catalyst in this eff ort.

Globally, there is a gap. No single UN agency cur rently has responsibility for registering births and deaths. This absence has led the Who Counts? team to call for a new international body to improve civil registration eff orts. But they concede that the likelihood of a new organisation being inaugurated is low. In the interim, they urge donors to encourage countries and global partners to do more to promote and support registration systems. Ultimately, this campaign is about how much each of us values the life of every other human being. It is a test of our humanity.

Richard HortonThe Lancet, London NW1 7BY, UK

1 Setel PW, Macfarlane SB, Szreter S, on behalf of the Monitoring of Vital Events (MoVE) writing group. A scandal of invisibility: making everyone count by counting everyone. Lancet 2007; published online Oct 29. DOI:10.1016/S0140-6736(07)61307-5.

2 AbouZahr C, Cleland J, Coullare S, on behalf of the Monitoring of Vital Events (MoVE) writing group. The way forward. Lancet 2007; published online Oct 29. DOI:10.1016/S0140-6736(07)61310-5.

3 Mahapatra P, Shibuya K, Lopez AD, on behalf of the Monitoring of Vital Events (MoVE) writing group. Civil registration systems and vital statistics: successes and missed opportunities. Lancet 2007; published online Oct 29. DOI:10.1016/S0140-6736(07)61308-7.

Pano

s Pict

ures

Published OnlineOctober 29, 2007

DOI:10.1016/S0140-6736(07)61418-4


See Perspectives page 1537

See Series page 1569


Comment


Improving health statistics in AfricaThe Lancet’s Who Counts? Series1–4 raises important issues about the relevance of statistics for eff ective health-care delivery in developing countries. The avail-ability of statistics is crucial in the fi ght against poverty, and is a necessary starting point to quantify outcomes needed to monitor and measure progress towards the Millennium Development Goals. The lack of reliable and good-quality statistics is a major obstacle to assessment of changes in development indicators in many African countries. In particular, health indicators remain of major concern: many African countries with inadequate statistical capacity and measurement systems are also some of the countries worst hit by deadly diseases such as HIV/AIDS and malaria. The lack of health statistics ranges from poor systems for civil registration to poor data on immunisation and child mortality rates.

A World Bank review of 125 middle-income and low-income countries with populations greater than 1 million illustrates this point.5 The countries were assessed for adherence to key international statistical methods and standards of good practices. About 60 countries in the survey—most in sub-Saharan Africa, except for Afghanistan—did not reach the midpoint score. Additionally, the World Bank noted that half the population of African countries had not recently been included in a census.

The absence of statistics in many African countries is both a symptom and a cause of underdevelopment. Im-provement of statistical systems is crucial in devel oping countries for three main reasons. First, without ade-quate capacity for obtaining statistics, assessment of the magnitude of the development problems to be faced is often impossible. Second, if we get the numbers wrong, tackling development problems eff ectively is diffi cult. Scaling up of interventions becomes diffi cult and re sources might be allocated away from more pressing issues. Thus, in this case, we risk solving the wrong problem or solving a problem in the wrong way. Third, without ade-quate statistics, assessment of the eff ectiveness of various programmes after imple mentation becomes diffi cult.

The absence of statistics has immediate implications for policymakers. In many African countries, evidence about the prevalence of child and maternal mortality and about the lack of access to health services is anecdotal. Yet without reliable information, government

offi cials who allocate resources for health budgets in such countries are essentially working in the dark. The low quality of statistics results in adverse outcomes, such as underfunding and poor monitoring of many development programmes.

This Series highlights the importance of improving health statistics with an emphasis on civil registration systems, which provide records of births, deaths, and the causes of death. We agree with the Series authors that many developing countries and donor agencies have not ade quately made a priority of collecting health or social data, compared with other statistical systems for gather ing economic data. Data on gross domestic product and infl ation in most developing countries are probably better estimated than for maternal mortality rates. In our exper ience, the quality of social statistics (compiled by both governmental and international agencies) in countries such as Nigeria is wanting. For example, we question data for the low life-expectancy (at birth) in Nigeria of about 47 years,6 which in our view underestimates the empirical reality. Casual empiricism based on our observations in various parts of Nigeria suggests that life expectancy tends to be a few years higher, particularly given Nigeria’s low and declining HIV/AIDS prevalence rates compared with other African countries.

Strengthening statistical capacity in developing countries will need a concerted eff ort from governments




Pano

s Pict

ures

Birth registration in Agadez, Niger


Comment


there, from donors, and from multilateral institutions. Governments and donors must view reliable data as an important tool in the development process, and must invest both fi nancial and human resources in strengthening their statistical systems. The Who Counts? Series is therefore most welcome, especially if it can spark a broader debate about the importance of statistics in development.

*Ngozi Okonjo-Iweala, Philip Osafo-KwaakoBrookings Institution, Washington, DC 20036, [email protected]

We declare that we have no confl ict of interest.

Obesity plan lacks foresightIf words were eaten and not read, the UK Government would have developed the perfect tool for fi ghting the world’s escalating obesity epidemic. The Tackling obesities—future choices report,1 issued on Oct 17, 2007, by its Foresight programme (part of the Offi ce for Science), is so indigestible that readers will need strong stomachs to get beyond the fi rst page. Try this convoluted example, from the executive summary, and more worthy of a New Age management consultant than a modern governmental department: “The central dynamic of the obesity system is a positive feedback cycle that locks us into a pattern of positive energy balance as individuals and at a societal level.”

The convoluted centrepiece graphic (map 27) of weighted causal linkages, described by a colleague as resembling rhinoceros’ intestines, seems to have no practical value. In a very complex way, the graphic simply makes the banal point that the many factors that contribute to obesity all infl uence each other.

The main 162-page report, and several accompanying publications, are a triumph of vapid political style and fl abby prose over meaningful substance. That does not bode well for a new approach to policy generally—not to mention public health—under Prime Minister Gordon Brown’s fl edgling administration, after the spin-led style of his predecessor, Tony Blair. This state of aff airs is pitiful, because there is no doubt about the urgency of the need to reverse obesity, and hidden away in the document and its supporting references is the collective work of 250 scientists over the past 2 years, which

summarises current understanding and highlights the challenges ahead.

Yet, the launch of the report was a lesson in poor communication, and of packaging for political expedi-ency. At the press conference to release the report, Dawn Primarolo, the Minister for Public Health, became the latest offi cial to spout the triumphalist governmental claim that the UK is “leading the world” in its work on obesity. That claim is both questionable and unnecessary, just as much as others made in recent months on issues as diverse as hospital-acquired meticillin-resistant Staphylococcus aureus and preparations for pandemic infl uenza.

The trumpeting of redoubled governmental com-mitment also sat uneasily with the news, which slipped out discreetly a few days before Tackling obesities was released (and later reported2), that the UK had back-tracked on a previous target unveiled in 2004 to reduce the proportion of obese and overweight children by 2010 to 2000 levels. The deadline for that goal is now 2020.

Health and science correspondents from leading UK media who attended the Foresight press conference left no better enlightened about or equipped for eff orts to tackle obesity than the additional physical exercise off ered to them by walking away with the bulky briefi ng documents.

Health correspondents are not a particularly cynical group, but several were exasperated in their eff orts to fi nd a fresh angle to report. One reason was that the

1 Setel PW, Macfarlane SB, Szretzer S, on behalf of the Monitoring of Vital Events (MoVE) writing group. A scandal of invisibility: making everyone count by counting everyone. Lancet 2007; 370: 1569–77.

2 Mahapatra P, Shibuya K, Lopez AD, et al. Civil registration systems and vital statistics: successes and missed opportunities. Lancet 2007; published online Oct 29, 2007. DOI 10.1016/S0140-6736(07)61308-7.

3 Hill K, Lopez AD, Shibuya, et al. Interim measures for meeting needs for health sector data: births, deaths, and causes of death. Lancet 2007; published online Oct 29, 2007. DOI 10.1016/S0140-6736(07)61309-9.

4 AbouZahr C, Cleland, J, Coullare F, et al. The way forward. Lancet 2007; published online Oct 29, 2007. DOI 10.1016/S0140-6736(07)61310-5.

5 World Bank. Building statistical capacity to monitor development progress. 2002. http://siteresources.worldbank.org/SCBINTRANET/Resources/239410-1113334813340/board-paper-feb4.pdf (accessed Oct 29, 2007).

6 World Bank. World development indicators database. 2007. http://web.worldbank.org/WBSITE/EXTERNAL/DATASTATISTICS/0,,contentMDK:21298138~pagePK:64133150~piPK:64133175~theSitePK:239419,00.html (accessed Oct 29, 2007).

See Editorial page 1521

Crow

n Co

pyrig

ht

Comment


main facts in Tackling obesities were already stale: they had been published a few days earlier in the Sunday newspapers.3 These items included guesstimates that more than half of the UK population would be obese by 2050, by which time annual health-care costs alone, caused by the epidemic, would probably be £6 billion and total costs to society would be £46 billion, unless the trend is reversed.

The Department of Health, hardly the most effi cient communicator, added to the frustration by slapping a strict embargo on the report for the following day, restricting the ability for meaningful reaction by other organi sations. In response to the report, Peter Hollins, chief executive of the British Heart Foundation, com-mented: “This report, whilst welcome, confi rms much of what we already know, and have known since the fi rst reports started warning of this crisis in the mid 1970s.”4 The general concern is that after such a comprehensive review of obesity, there was barely any conclusion other than the claim that there is currently little evidence supporting the launch of new policies—and instead the need for yet more research and pilot projects.

The report’s jargon-ridden, if worthy, core principles for tackling obesity could be applied almost word-for-word to any policy: a system-wide approach, higher priority for prevention, engagement of stakeholders, long-term sustained interventions, ongoing evalu-ation, and focus on continuous improvement. Ano ther

mes sage was that individual eff orts to improve diet and increase physical activity will only have a limited eff ect. This point might contain an element of truth, but would be more convincing if it did not also smack of political correctness and a fear of placing emphasis on personal responsibility.

That the UK Government is paying fresh attention to the enormous challenge of obesity is welcome. But if policymakers are serious, they should start by communicating more clearly, which begins with having something meaningful to say. First, the Government must start to translate prevaricating consultant-speak, hesitant scientists’ views, and hand-wringing into action.

Andrew JackFinancial Times, London SE1 9HL, [email protected]

I am the pharmaceuticals correspondent for the Financial Times.

1 Foresight. Tackling obesities: future choices—project report. Oct 17, 2007. http://www.foresight.gov.uk/Obesity/Obesity_fi nal/Index.html (accessed Oct 25, 2007).

2 Boseley S. Date for halting childhood obesity slips back 10 years. Guardian Oct 17, 2007. http://society.guardian.co.uk/health/story/0,,2192574,00.html (accessed Oct 29, 2007).

3 Campbell D. Obesity crisis to cost £45bn a year. Oct 14, 2007. Observer http://observer.guardian.co.uk/uk_news/story/0,,2190844,00.html (accessed Oct 29, 2007).

4 British Heart Foundation. Foresight report—British Heart Foundation urges Government to wake up to obesity crisis. Oct 17, 2007. www.bhf.org.uk/news_and_campaigning/press_offi ce/latest_news__views/foresight_obesity_report.aspx (accessed Oct 25, 2007).

Lancet/Global Forum essay competition winners“Equitable access: research challenges for health in developing countries” is the theme of this year’s Global Forum for Health Research (Oct 29 to Nov 2, Beijing, China),1 and thus was also the topic chosen for this year’s Lancet/Global Forum essay competition. We challenged young health professionals to impress us with their “passion and pique, argument and analysis, and crisp, compelling prose”.

We were not disappointed. Indeed, as nearly 300 diverse and vibrant entries fl ooded in, we were a little overwhelmed by the prospect of having to select a shortlist of only 40 and fi ve winners. Entries came from 60 countries, from individuals of 64 nationalities. In regional terms—both by residence and nationality—the largest number of entries was from Africa.

Several rounds of reading, debating, and voting were needed before the fi nal shortlisted group was agreed. After more debate and voting, we decided that the fi ve winners should be Seye Abimbola from Nigeria, Denise Nacif Pimenta from Brazil, Laura Sikstrom from Canada, Lee Yung Wong from Malaysia, and Zhang Lingling from China.

The winning essays, published by the Global Forum for Health Research, are now available on The Lancet’s website.

Sarah RamsayThe Lancet, London NW1 7BY, UK 1 Global Forum for Health Research. Forum 11. http://www.globalforum

health.org/Site/004__Annual%20meeting/001__Forum%2011/001__Home.php (accessed Oct 29, 2007).

To read the essays see http://www.thelancet.com/online/focus/essay

Comment


Clinical update: cyanotic adult congenital heart diseaseWorldwide, the annual incidence of adult congenital heart disease (CHD) is 1·5 million,1,2 and advances in medical and surgical techniques now enable 85% of patients with CHD to survive well into adulthood.3 In the UK, about 250 000 adults have CHD and this number is growing.3,4 Cyanosis is common in these patients and is accompanied by adaptive mechanisms and pathological changes to compensate for the associated hypoxia. Patients with cyanotic CHD have a complex multisystem disorder with unique pathophysiology, and need tertiary care. However, many are managed by physicians from a range of disciplines.1,5,6 Longstanding management dogmas and misconceptions about patho physiology are now being challenged in the light of new data.

Secondary erythrocytosis—and not polycythaemia—associated with cyanotic CHD is a physiological and desirable response to chronic hypoxia. In iron-replete patients, the severity of secondary erythrocytosis is inversely related to resting oxygen saturation. Red blood cells increase in number, an adaptive mechanism driven by the chronic cyanosis to increase oxygen-carrying capacity.7 This mechanism has impor tant implications for current practice, especially the use of phlebotomy to decrease haemoglobin con centrations. Phlebotomy has harmful rather than benefi cial eff ects in adults with cyanotic CHD.8

However, routine phlebotomies are still done in appro -priately around the world because of the perceived benefi ts against stroke and on myocardial function.9 Patients might report short-lived relief of hyperviscosity symptoms with phlebotomy. But these symptoms resemble those of iron defi ciency, and we have to question whether routine phlebotomy has a real non-placebo clinical eff ect in these patients.

Stable adults with cyanotic CHD have, by defi nition, abnormal oxygen saturations. In these patients, the extent of cyanosis is usually a manifestation of admixture of oxygenated and deoxygenated blood (as occurs with right-to-left shunting or with uni ventric ular physiology) or inadequate pulmonary blood fl ow. Cyanosis is also dynamic and varies with factors such as exercise. The resultant hypoxia initiates major compensatory mechanisms.9 First, the low con centrations of oxygen delivered to tissues stimulates renal production of erythropoietin, which enhances stimulation of the

production of red blood cells, with a consequential increase in red-blood-cell mass, packed cell volume, and whole-blood viscosity.7,10,11 This process occurs in patients with cyanotic CHD,12 and contrasts with polycythaemia vera, in which polycythaemia in dicates proliferative activity along all cell lines (red and white blood cells and platelets). High haemoglobin con centration (often above 20 g/dL) is therefore a common fi nding in adults with cyanotic CHD.12 Second, a rightwards shift in the oxygen dissociation curve encourages off -loading of oxygen to tissues. Third, cardiac output increases, to achieve optimum oxygen delivery and prevent end-organ damage. Compensatory dilation of the coron ary circulation by a combination of endothelial vaso dilators and mural attenuation is thought to prevent adverse eff ects on myocardial perfusion.13

In the past, there were concerns that this compensatory secondary erythrocytosis could increase plasma viscosity and lead to impairment of the microcirculation, thereby compromising tissue oxygen delivery.7 Symptoms attri b utable to hyperviscosity include headache, dizzi-ness, visual disturbance, paraesthesia, altered mental func tion, tinnitus, fatigue, muscle weakness, and gen-ito urinary and gastrointestinal bleeding.11 In our exper-ience, only rare patients with chronic compensated secondary erythrocytosis and no previous history of phlebotomy show such hyperviscosity symptoms. To add to the complexity, iron defi ciency itself might also cause symptoms similar to hyperviscosity.14 Further-more, symp toms previously attributed to hyperviscosity might actually arise from decreased tissue oxygen delivery.

There is also confl icting evidence about whether iron defi ciency itself increases blood viscosity.15 Iron defi ciency and microcytosis did not actually increase whole-blood viscosity in adults with cyanotic CHD, and hyperviscosity symptoms were relatively common but not related to packed cell volume, haemoglobin, iron indices, or cell size.16 Although symptoms correlated with measured viscosity (after correcting for packed cell volume), exercise capacity was greater in patients with a higher packed cell volume and higher blood viscosity. Furthermore, a normal or increased erythrocyte size (as indicated by mean corpuscular volume) was common in patients with iron defi ciency. Blood rheology in patients

Comment


with cyanotic CHD showed slower blood-passage time in cyanotic patients than in controls, which correlated with red-blood-cell count, haemoglobin concentration, and packed cell volume.17 Consistent with other recent reports, mean corpuscular volume was not diff erent between patients with cyanotic CHD and controls.16,17 Higher mean corpuscular-haemoglobin concentration was associated with faster blood-passage time.17 Although speculative, these studies suggest a larger red-cell volume might enable these patients to maximise haemoglobin capacity per red cell, while keeping blood viscosity low and optimising tissue oxygen delivery.

More than a third of adults with cyanotic CHD are iron-defi cient,12,18 and can be overlooked easily because the common laboratory measures, such as microcytosis and hypochromia, are rarely found.10,18 The cause of iron defi ciency in this group of patients is multifactorial, with the consumption of iron stores from excessive erythropoiesis being a universal and predominant factor. Other causes include recurrent phlebotomies, gastrointestinal blood loss, decreased dietary intake, and menorrhagia. Iron status should be routinely and periodically assessed by measuring serum ferritin and transferrin saturation concentrations, and not by haemoglobin, packed cell volume, and erythrocyte indices alone.7,9,10,18 When ferritin concentrations are increased, assay of soluble transferrin receptor is useful to identify concomitant iron defi ciency.18 Low-dose oral iron or pulses of parenteral iron in patients who are intolerant of or have failed oral iron therapy should be administered until the packed cell volume begins to rise.9,10 Gradual replacement of iron stores is advised to avoid an excessive rebound of iron and resultant erythropoietic response.7 The objective is to provide suffi cient iron to attain steady-state erythropoiesis as appropriate for the patient’s underlying physiology11,16 Correction of iron defi ciency alleviates hyperviscosity symptoms in patients with cyanotic CHD.

Inappropriate phlebotomies in adults with cyan-otic CHD are still done widely, to decrease symptoms attributed to hyperviscosity and the risk of cerebro-vascular events. Cerebrovascular events have been extrapolated inappropriately from studies in poly-cythaemia vera, in which phlebotomy to maintain a packed cell volume of 40–45% remains current practice.19 However, studies in adults with cyanotic CHD have failed

to fi nd an association between red-cell mass and the incidence of stroke.8,9,12,20 The best available evidence derived from a retrospective case-series of 162 patients (3135 patient-years) is to the contrary, which suggests that phlebotomy and iron defi ciency are independent risk factors for cerebrovascular events.8 The decreased oxygen-carrying capacity of iron-defi cient erythrocytes compared with their iron-replete counterparts might contribute to this increased risk. Co-existing abnor-

Assess annually

Anaemia history*Symptoms of hyperviscosity†Measure oxygen saturation‡Laboratory measures:§haemoglobin, packed cell volume,red-cell indices, serum ferritin,transferrin saturation

Serum ferritin ≤15 μg/LTransferrin saturation ≤15%

Patient iron-deficient

Iron supplementationAddress other causes of irondeficiency as identified from history

Reassess symptoms

Repeat laboratory tests

Consider cessation of ironsupplementation when ironreplete (serum ferritin ≥15 μg/Land transferrin saturation ≥15%)

Some patients will requirechronic iron supplementationfor steady-state erythrocytosis

Regularly reassesssymptoms andlaboratory tests

Patient iron-replete

No symptoms ofhyperviscosity

Reassess every6–12 months

Trial of phlebotomywith fluidreplacement

Resolution ofsymptomsPatient remainsiron-replete

Persistent moderate–severe hyperviscositysymptomsPacked cell volume >65%

Patient iron-replete

Symptoms ofhyperviscosity

Assess for other causesof symptoms and treataccordingly: eg,HypovolaemiaGoutBrain abscessHypothyroidismDepression

Serum ferritin ≥15 μg/LTransferrin saturation ≥15%

Figure: Assessment and management of secondary erythrocytosis and iron defi ciency in cyanotic CHD*Including details of haemoptysis, gastrointestinal or other bleeding, phlebotomy, menses, diet, medication, and use of iron supplements. †Headache, faintness or dizziness, visual disturbance, fatigue, muscle aches, joint aches, paraesthesia, easy bruising, epistaxis, gingival bleeding, gout, and poor mental function.16 ‡In seated position, breathing air, after 2 min of complete rest with standard transcutaneous fi nger-pulse oximeter; if diff erential peripheral cyanosis is suspected, saturation should be measured in toe. §If discrepancy between serum ferritin and transferrin saturation, transferrin saturation is more important.

Comment


malities in clotting factors and platelets are also believed to have a role.16 Furthermore, phlebotomy-associated hypovolaemia without volume replacement might lead to an acute fall in systemic blood fl ow, com promising oxygen transport and delivery and resulting in cerebral hypoxia.9

Thus the existing evidence (from case series and translational research) is that prophylactic phlebotomy exacerbates iron defi ciency,10 decreases exercise tol-erance,16 and increases cerebrovascular events.8 Pro-phyl actic phlebotomy for secondary erythrocytosis in adults with cyanotic CHD should therefore be abandoned.

Instead, phlebotomy should be restricted to two in di-cations: temporary relief of moderate-to-severe hyper-viscosity symptoms in the presence of a high packed cell volume (>65%), and preoperatively to improve haemostasis.8,9,11 These indications are based on current expert opinion, although the supportive evidence is scarce and any benefi t might be exaggerated by a placebo eff ect in those treated by phlebotomy. Hypovolaemia and iron defi ciency must have been excluded. Careful clinical and laboratory assessment is vital and might identify other causes for the symptoms. A management algorithm might assist decisionmaking for phlebotomy (fi gure).16 No more than one unit of blood should be removed at any one time, and removal should be combined with equivalent volume replacement with normal saline or Hartmann’s solution.8,9 Air fi lters need to be used for intravenous lines to lower the risk of paradoxical air embolism. Blood pressure should be monitored regularly during the procedure, and further fl uid re place ment might be needed until it stabilises. The patient’s clinical status, rather than haematological laboratory tests, should be closely assessed after removal of each unit of blood to assess whether further phlebotomy is necessary.10,11 Phlebotomy under these circumstances might be done in an outpatient setting.11

Secondary erythrocytosis in patients with cyanotic CHD is a physiologically appropriate response to hyp-oxia. Iron defi ciency is common, produces signifi cant morbidity, and is often exacerbated by inappropriate phlebotomy. Importantly, iron defi ciency produces symp toms that resemble, and are often confused with, those of hyperviscosity. Therefore timely identifi cation

of iron defi ciency, appropriate iron supplementation, and avoidance of inappropriate phlebotomy for patients with cyanotic CHD are all warranted.

*Mark S Spence, Michelle S Balaratnam, Michael A GatzoulisAdult Congenital Heart Centre and Centre for Pulmonary Hypertension, Royal Brompton Hospital, and the National Heart & Lung Institute, Imperial College, London SW3 6NP, UK (MSS, MSB, MAG); and Royal Jubilee Hospital, Victoria Heart Institute Foundation, Victoria, BC, Canada (MSS) [email protected]

We thank Henryk Kafk a for his comments. MAG has received support from the Clinical Research Committee, the Waring Trust, and the British Heart Foundation. MSS and MSB declare that they have no confl ict of interest.

1 Brickner ME, Hillis LD, Lange RA. Congenital heart disease in adults. N Engl J Med 2000; 342: 256–63.

2 Moller JH, Taubert KA, Allen HD, et al. Cardiovascular health and disease in children: current status. Circulation 1994; 89: 923–30.

3 Warnes CA, Liberthson R, Danielson GK, et al. The changing profi le of congenital heart disease in adult life. J Am Coll Cardiol 2001; 37: 1170–75.

4 Wren C, O’Sullivan JJ. Survival with congenital heart disease and need for follow up in adult life. Heart 2001; 85: 438–43.

5 Perloff JK, Warnes C. Challenges posed by adults with repaired CHD. Circulation 2001; 103: 2637–43.

6 Niwa K, Perloff JK, Webb GD. Survey of specialized tertiary care facilities for adults with congenital heart disease. Int J Cardiol 2004; 96: 211–16.

7 Rosove MH, Hocking WG, Child JS, et al. Chronic hypoxaemia and decompensated erythrocytosis in cyanotic congenital heart disease. Lancet 1986; 2: 313–15.

8 Ammash N, Warnes CA. Cerebrovascular events in adult patients with cyanotic congenital heart disease. J Am Coll Cardiol 1996; 28: 768–72.

9 Thorne SA. Management of polycythemia in adults with cyanotic congenital heart disease. Heart 1998; 79: 315–16.

10 Diller GP, Gatzoulis M. Pulmonary vascular disease in adults with congenital heart disease. Circulation 2007; 115: 1051–58.

11 Oechslin E. Eisenmenger’s syndrome. In: Gatzoulis MA, Webb GD, Daubeney PEF, eds. Diagnosis and management of adult congenital heart disease. Philadelphia, PA, USA: Churchill Livingstone, 2003: 363–78.

12 Diller GP, Dimopoulos K, Broberg CS, et al. Presentation, survival prospects, and predictors of death in Eisenmenger syndrome: a combined retrospective and case-control study. Eur Heart J 2006; 27: 1737–42.

13 Dedkov EI, Perloff JK, Tomanek RJ, et al. The coronary microcirculation in cyanotic congenital heart disease. Circulation 2006; 114: 196–200.

14 Fairbanks VF, Beutler E. Iron defi ciency. Beutler E, Lichtman MA, Coller BS, Kipps TJ, Seligsohn U, eds. Williams haematology, 5th edn. New York, USA: McGraw Hill, 1995: 490–511.

15 Milligan DW, MacNamee R, Roberts BE, et al. The infl uence of iron-defi cient indices on whole blood viscosity in polycythaemia. Br J Haematol 1982; 50: 467–71.

16 Broberg CS, Bax BE, Okonko DO, et al. Blood viscosity and its relationship to iron defi ciency, symptoms, and exercise capacity in adults with cyanotic congenital heart disease. J Am Coll Cardiol 2006; 48: 356–65.

17 Katayama Y, Horigome H, Murakami T, et al. Evaluation of blood rheology in patients with cyanotic congenital heart disease using a microchannel array fl ow analyzer. Clin Hemorheol Microcirc 2006; 35: 499–508.

18 Kaemmerer H, Fratz S, Braun SL, et al. Erythrocyte indexes, iron metabolism, and hyperhomocysteinemia in adults with cyanotic congenital cardiac disease. Am J Cardiol 2004; 94: 825–28.

19 Berk PD, Goldberg JD, Donovan PB, et al. Therapeutic recommendations in polycythemia vera based on Polycythemia Vera Study Group protocols. Semin Hematol 1986; 23: 132–43.

20 Perloff JK, Marelli AJ, Miner PD. Risk of stroke in adults with cyanotic congenital heart disease. Circulation 1993; 87: 1954–59.

Special Report


Aboriginal workers key to indigenous health in AustraliaAs a group, indigenous Australians are much less healthy and more likely to die at younger ages than their non-indigenous counterparts. Training more indigenous people as health workers could help to reduce these startling inequalities, say experts. Margaret Harris Cheng reports.

Funeral fund scams are nothing new. But it is usually people in old age who are cheated out of the money they stump up to guarantee a good funeral. When it is the young whose fears of failing to cover funeral expenses are being exploited, you have to ask “why?” But in Australia, where it has been reported that a funeral benefi t company, already in trouble with the authorities, has been targeting indigenous Australians, signing up one in 13 to prepaid funeral schemes, the news raised few eyebrows. It is simply too well known that indigenous Australians die too young and too often.

“On current rates, one-third of young indigenous men aged 15 will be dead before age 60, compared with 8% in the Australian population”, says Alan Lopez in a foreword to a report into the burden of indigenous disease that he and colleagues from the University of Queensland published in September this year. “This four-fold increase in risk of death, comparable to parts of Africa today, is largely due to excess mortality from such causes as ischaemic heart disease, suicide, and Type 2 diabetes. Its reduction must be a priority for indigenous health services” , he writes.

Whereas in the early 1960s infectious diseases were a main cause of morbidity and mortality, the picture is now more a mixed one. According to The burden of disease and injury in Aboriginal and Torres Strait Islanders, cardiovascular disease and mental disorders were the major disease categories in the indigenous Australian population in 2003, accounting for 32% of the disease burden. Chronic respiratory disease, diabetes, and cancers were the next three leading causes, each accounting for 8% of the total indigenous Australian disease burden.

The disease burden caused by cancer in the indigenous population (8%) is much lower than the burden it causes in the total Australian population (19%) because indigenous people are dying earlier, from other causes.

The life-expectancy gap between indigenous Australians and non-indigenous Australians is 17 years. It is a gap that compares especially poorly with similar societies, such as New Zealand, where the gap between indigenous and non-indigenous life expectancies is 7 years. However, that gap is only part of the story.

Challenging pessimism “One of the problems is that we tend to analyse indigenous health in impressionistic ways. It leads to the myth that nothing ever improves. In fact [indigenous] life expectancy has improved by between 10 and 12 years”, says Ian Anderson, professor of indigenous health at the University of Melbourne. In the Northern Territory, where many of Australia’s indigenous groups live in diffi cult, remote country, indigenous life expectancy at birth has risen substantially. It has increased

from around 52 years for men and 54 years for women in the late 1960s to 60 years and 68 years, respectively, according to a study by Tom Wilson and colleagues from the Charles Darwin University. “The widespread pessimism that surrounds indigenous health and mortality is largely unfounded, at least for the Northern Territory. Although much remains to be done to reduce indigenous mortality, the results in this paper demonstrate that improvements are occurring and that sustained and increased eff ort is worthwhile and will succeed”, they concluded in their report published in the Australian and New Zealand Journal of Public Health. Improvements in housing and reductions in overdcrowding and infectious disease rates are thought to have contributed to this gain in life expectancy and the authors believe that these gains can be built on.

David Thomas and colleagues, writing in the Medical Journal of Australia last year agree. “Now is not the time for giving up or changing everything because ‘nothing has worked’, but for investigating further

A visiting indigenous health worker talks to a group of Aboriginal people in the Northern Territory

Corb

is



Special Report

what has worked, and for increased and sustained eff ort to ensure these early promises of the possibility of lasting improvements to Indigenous health are realised in the Northern Territory and beyond.”

Paternalism returnsNot only is the pessimism about Aboriginal health unfounded, say indigenous health workers, it has been used to support the belief that indigenous Australians are a doomed group and that eff orts to improve their health are hopeless or that only extreme measures are productive.

In June, in response to Little Children are Sacred—a report detailing widespread child abuse in Northern Territory indigenous communities and neglect of the issue by authorities—the Australian Prime Minister John Howard announced one such extreme measure: an urgent “intervention”. His plan involved sending extra police to indigenous communities, compulsorily acquiring townships on 5-year leases, tying welfare payments to school attendance, banning alcohol, and instituting compulsory child health checks (later made voluntary).

Although supported by the other major Australian political party, the Labor party—the party expected to win the upcoming Federal election—the intervention has been widely condemned by medical and indigenous groups. “While we support the availability of health checks for all Australians the RACGP [The

Royal Australian College of General Practitioners] would be very concerned at the introduction of a fl y-in, fl y-out model of health assessments for children”, said the college president, Vasantha Preetham. “Aboriginal people need an ongoing and trusted source of high quality health care, not stop-gap measures. Government programmes need to support local doctors and staff to deliver these services with backfi ll from those with less experience dealing with these health needs.”

In late October, Marion Scrymgour, Australia’s fi rst indigenous cabinet minister, the Minister for Family and Community Services, Child Protection, and Young Territorians in the Northern Territory’s Labor government, attacked her Federal Labor counterparts for supporting the intervention. She also said that anyone who dared to criticise the intervention was accused by Mal Brough, the current Federal Minister for Indigenous Aff airs, of being “for the perpetrators” of child abuse. “The new world order for Aboriginal people requires, it seems, a vicious new McCarthyism”, she said.

Speaking at the Charles Perkins Oration at the University of Sydney—an occasion held to commemorate an earlier indigenous leader—Scrymgour revealed that since the intervention, indigenous communities had been deluged with consultants and public servants performing assessments. One community had visits from “164 Commonwealth public servants and consultants related to the inter-

vention over a period of 10 weeks”, said Scrymgour.

Boosting aboriginal workersIt is not surprising that, until recently, indigenous population research has been impressionistic rather than evi-dence based. Indigenous Australians are wary of the term ”research” because much damage has been done to them and their ancestors in the name of research. Communities are pushing to receive and ceremonially bury the remains of ancestors that were taken by “researchers” from museums and pathology departments. School children still learn about the head of Pemulway, the fi rst indigenous leader to oppose the British colonists taking his lands in the Sydney region, which was pickled and sent as a present to the botanist Joseph Banks in 1802.

But, with a gradual increase in the number of indigenous people working in health and research over the past 20 years, long-term population studies are being developed and these are challenging some of the myths about research, researchers, and indigenous health.

Rita Williams, an Aboriginal health worker at the Children’s Hospital, Westmead, Sydney, was so aff ected by the plight of a child from renal disease, and the illnesses experienced by so many others (Aboriginal adults have ten times the rate of renal disease as non-Aboriginal adults) that she pushed to have the issue studied. The end result was the Antecedents of Renal Disease

1788–90

British colonists arrive in the Eora and Dhaurug lands (now Sydney), and claim ownership of Aboriginal lands. The colonists are resisted by groups led by Aboriginal leader Pemulwuy. At least half Eora’s population die from measles and smallpox introduced by the colonists.

1910–70

Legislation results in the forced removal of between one in three and one in ten indigenous children from their Aboriginal or Torres Strait Islander families.

1967

Aboriginal people are recognised as part of the Australian population after 92% of voters elect to include them in the census and give the Commonwealth government the power to legislate on their behalf.

1968

The Commonwealth Offi ce of Aboriginal Aff airs is established and identifi es health as one of four major areas for indigenous development, initiating grants to the States for Aboriginal Health programmes.

1981

The Commonwealth Government begins a A$50 million Aboriginal Public Health Improvement Program focusing on unsatisfactory environmental conditions associated with inadequate water, sewerage, and power systems.

Special Report


in Aboriginal Children (ARDAC), a study that has been following over 2000 Aboriginal and non-Aboriginal children throughout New South Wales for the past 6 years to pinpoint the evolution of renal disease. So far the study has thrown up a few surprises, fi nding no signifi cant diff erences in the incidence of markers of early renal disease between the two groups. It has, however, found that other risk factors for chronic disease, obesity, and hypertension are appearing in the population very early. The ARDAC study will continue for another 6 years, following children who entered the study at age 5 or 6 years through to adulthood, to try and identify what happens to indigenous people through adolescence.

Not only are indigenous Australians living longer, another widely accepted myth—that they are dying out—is not supported by the evidence. According to census data from the Australian Bureau of Statistics, indigenous Australians are one of the youngest and fastest growing groups in the population.

Whereas fertility rates are low in much of the rest of the population, the indigenous population is growing at around 2% per year. And, although the rest of the population is ageing, 40% of indigenous Australians were under age 15 years, according to the 2001 census.

So, indigenous people are living a little longer and are one of the few growing sectors in the Australian population. But these features do not alter the fact that indigenous Australians are, as a

group, much less healthy than non-indigenous Australians. Can anything be done to change that?

Yes, says Anderson. Because more is now known about the causes of ill-health in indigenous communities, something can be done about the eff ects. “In terms of the health gap, 50% of it is due to the fact that when indigenous people get sick they die much more quickly. The other 50% is due to risk factors. We need to provide access to quality care, focus prevention around tobacco control and diet, and invest in a health workforce.”

“We also need Aboriginal doctors and nurses. We did not have any Aboriginal medical graduates until the 1980s. New Zealand had their fi rst indigenous medical graduate in 1890, in the USA it was in the 19th century. We have had a very poor history…we need to create a critical mass.”

Indigenous health workers should not be automatically expected to work in their communities—this is neither reasonable nor practical, said Anderson. But training many more than the Australian system does now, will lead to more indigenous people inside the system able to make it respond more eff ectively to their needs. And one of the simplest but most powerful eff ects of increasing the numbers of indigenous people trained as health workers is that there will be more indigenous people with jobs, many of them the kind of long term, well paid jobs that lead to a permanent escape from poverty. “Some go back

to work in indigenous health, but that is not the only pathway…it is a broader reform strategy, making people available to teach and enabling people to rise out of poverty through employment”, said Anderson.

Addressing social determinantsPoverty and unemployment are widespread in indigenous com-munities, partly because in remote areas there are no jobs but also because staying at school and getting the education needed to get a job in Australia is something too few indigenous people achieve.

Education research in New South Wales, the Australian state with the highest indigenous population, has shown that completing two of the last three high school years (year 10

mid-1980s

Pat O’Shane—Australia’s fi rst Aboriginal law graduate—becomes the country’s fi rst Aboriginal magistrate.

The fi rst Aboriginal medical students begin training and graduating.

1987–90

The Royal Commission into Black Deaths in Custody is held to investigate disproportionate numbers of deaths and suicides among aboriginal people in prisons.

A National Aboriginal Health Strategy is developed.

2003

The National Health and Medical Research Council publishes a strategic framework for improving Aboriginal and Torres Strait Islander health through research to identify research areas that could better inform policy and practice to improve indigenous health.

2004

Prime Minister John Howard abolishes the Aboriginal and Torres Strait Islander Commission. Indigenous specifi c programmes are transferred to mainstream agencies.

2007

The Howard Government responds to Little Children are Sacred by sending police, troops, administrators, and medical workers into indigenous communities in the Northern Territory.

Federal election scheduled for Nov 24.

John Howard has been accused of paternalism

Reut

ers



Special Report

and 11) will increase an Aboriginal person’s chances of employment by 40%. Con tinuing after year 10 and completing year 11 further increases those chances by 13%. Getting a post-secondary qualifi cation will increase employment chances by another 13–23%. But getting young indigenous Australians to stay in school, let alone reach an educational level where they can enter medical or nursing training, is easier said than done.

In 2003, the New South Wales Aboriginal Educational Consultative Group surveyed 413 schools and found disturbing evidence that indigenous

children are much more likely to be suspended from school than their non-indigenous counterparts. For every 1000 Aboriginal boys in the fi rst four high school grades (years 7 to 10) there were 629 short suspensions (suspension lasting from 1–4 days), compared with non-Aboriginal boys who had 188 short suspensions for every 1000.

Not only are there many more sus-pensions among Aboriginal children, the survey also found that excluding children from school starts very early and especially aff ects girls. In the kindergarten years to year 2, and then again for schooling years 3–6, Aboriginal girls were suspended nine to seven times more often than were non-Aboriginal girls. For Aboriginal boys the rate was also higher: they were excluded four to six times more often than were non-Aboriginal boys.

And, possibly worst of all, Aboriginal girls are being suspended from the infant kindergarten six times more often than are non-Aboriginal girls. “It is obvious that suspension will have a signifi cant and detrimental eff ect on any student’s performance. Suspension and the gaps in the students learning will only equip Aboriginal students with long-term low levels of literacy and numeracy skills that will set them on paths similar to that of their parents or other relatives—a path of continual poverty and unemployment”, says the report.

The report also noted that many Aboriginal children do not consider being suspended a punishment. It simply removes the compulsion to go to school. Once free of that they never return. So there is, indeed, plenty that can be done to improve Aboriginal health by focusing on education, community involvement, and improvement of the structure of the health service, particularly staffi ng and the provision of primary care.

New government, new hope?But, according to Anderson, there is also a symbolic infrastructure that

needs improvement. The Howard Government has not signed the UN treaty on indigenous peoples’ rights because its neo-conservative elements balked at the term “self-determination” fearing it would bring on an indigenous secessionist movement. “There is a mounting body of evidence that good health is not achieved if you take away people’s control of their lives. Australia would not sign the indigenous people’s treaty because of the self-determination clause and the idea that it threatens a nation state”, said Anderson. “It is an odd argument. Secession would discharge the colonists of their [fi nancial] obligations and any indigenous states would be poverty stricken!”

Prime Minister John Howard, has also always refused to apologise to indigenous people for the damage wrought to them by the arrival of European colonists. This, say indigenous leaders, has symbolically obstructed reconciliation at all levels of society. Both these issues are in the Labor party platform, which says a Labor government will sign the indigenous peoples’ treaty and formally apologise to indigenous people. If they win offi ce on Nov 24, the strength of those promises will be tested.

Anderson believes those promises will be kept. “The best thing that John Howard could do for reconciliation is to leave offi ce.”

Margaret Harris Cheng

An outback doctor tends to an Aboriginal child

Panel: Interview with an Aboriginal health worker

“I won’t be here tomorrow. I’ve got to drive to Bree for a funeral”, said Angela Trail. “A school mate, 38-years-old, died of a heart attack. Another one, even younger, 32, died last week, up at the hospital. Another massive heart attack.”

Bree is Brewarrina, a town in the middle of the drought-stricken New South Wales outback, all heat, red dust, and empty buildings, but a town that Angela loves because it is where she grew up. Now working in Dubbo, the major town serving western New South Wales, as an Aboriginal health worker, Angela will begin training to be a registered nurse next year. “In 3 years’ time, I can go back to Bree and look after my people”, she says.

Angela took up her job when the last of her four children reached high school and she decided she wanted to go into the workforce. She had no particular qualifi cations but said she did not need any as her job is all about being fl exible, liaising between aboriginal clients and the usually white medical staff , allaying deeply entrenched fears of authority, helping to explain reasons for treatments and investigations, and even simply fi lling out forms for clients with literacy problems. “We do home visits, we go out with the doctor…when they see an Aboriginal health worker they relax, they see it’s to their level.”

Some clients reject treatment or investigations because they are so frightened of institutions or procedures that they don’t understand, said Angela. “One lady, she really needs the treatment but will not take the tablets. I think she’s really frightened. I got on the internet and looked up her disease so I could tell her about it. Then she was better.”

Many of her clients have diabetes, hypertension, hypercholesterolaemia, and heart disease. Mental-health problems are rife too but the Aboriginal Medical Service that Angela works for has no mental-health worker. “Our doctor does what he can and sends them to see a psychiatrist.” There is also a growing drug problem as more of “the heavier stuff ” is fi nding its way from the coast to inland New South Wales. “I wish you could click a fi nger and solve all problems but you will not.”

Corb

is


Perspectives


Profi leCarla AbouZahr: making health statistics countCarla AbouZahr, arriving at WHO nearly 20 years ago, was taken aback to fi nd that the organisation had no defi nitive data about many of the problems it was addressing: “When I joined WHO, I was in the same position as many when they fi rst come. You think WHO knows everything—that the organisation somehow has all the data and it is all reliable and it’s quite a shock to realise how limited the data are.”

Not having the numbers is a serious problem, says AbouZahr, who has pioneered the drive to change all that over the past two decades. She quotes a phrase of Wendy Graham, professor of obstetric epidemiology at the UK’s Aberdeen University: “what you count, is what you do”. Collecting statistical data may seem expensive, but “it is much more expensive to misallocate your resources because you don’t have good data”, AbouZahr says. Unfortunately, the international donor community fails to provide the necessary funds for capacity building in developing countries. “It is not sexy”, she says. “Statistics just don’t light up their eyes.”

AbouZahr has made it her mission to improve the collection of statistical health data and was at the centre of the development of the Health Metrics Network that formally launched in May, 2005, and, having been seconded from WHO, she is now its deputy executive secretary. The Network is a global collaboration between developed and developing nations, multilateral agencies, foundations, and others, focused on improving the quality and quantity of data countries hold.

Statistics have long been AbouZahr’s passion. She studied statistics at the UK’s London School of Economics, and was keen to work with developing countries in some way. “That was a big driving force for me. But it wasn’t a direct route”, she recalls. After her marriage to Omar AbouZahr, with whom she has had three children, she took a job with Eurostat in Luxembourg at a time when the European Union countries were being required to provide statistical data as part of the accession treaty. In 1989, she moved to WHO in Geneva, where she began to work in maternal health and soon discovered a major absence of data.

“We are better at counting births than counting deaths”, she says. The statistics on child mortality are not too bad, although data on cause of death are poor. By contrast, data on adult mortality is almost totally neglected, she says. In poor countries, most adults die at home without a doctor to decide on cause of death and no record is made. That makes it hard to ascertain the scale of maternal mortality, for example. Graham, with whom AbouZahr has worked on maternal mortality since the early 1990s, says AbouZahr has made a diff erence. First, in championing the argument that maternal mortality is greater than has been

assumed because the deaths of pregnant women—who are vulnerable to such diseases as HIV, tuberculosis, and malaria—are not always ascribed to childbirth. And second, she has pushed forward the publication of maternal mortality statistics based on modelling when good data are not available, although it can cause tensions. “She has been on the sharp end of countries’ disquiet”, says Graham.

One of the problems with data collection, says AbouZahr, is that if each disease is treated separately there is a potential confl ict of interest for advocates who know that the greater the burden of suff ering, the more funding is likely to come their way. And there is a danger of double-counting, so that we end up with more deaths than have actually occurred. National data collection is what is needed. The aim of the Health Metrics Network is to try to help countries collect better data and to better use the data they have. It’s an uphill struggle because of the low priority accorded to statistics. But some countries have shown what can be done. Postapartheid South Africa is “a tremendous success story”, says AbouZahr. “They started out with an incredibly biased information system focused on the white ruling classes and grasped the nettle. They introduced a civil registration system covering the whole population.” Other countries that are making systematic eff orts to do better include Tunisia, Mauritius, and, lately, Sierra Leone.

Those with whom she works testify to AbouZahr’s dedication. Kenneth Hill, from the Harvard Center for Population and Development Studies, says she has made “signifi cant contributions to the gathering, evaluation, and interpretation of (particularly) health information in the developing world over the last two decades”. He adds that “Carla has worked tirelessly for HMN to improve national health information systems through a programme of small grants and advocacy programmes, culminating in The Lancet’s WHO Counts series of papers about the scandalously poor condition of civil registration systems in the developing world and how such systems can, and must, be improved.”

Graham points to AbouZahr’s pragmatism and political astuteness, because she takes the view that we will never have perfect data: “She has been able to push forward that need for pragmatism and making the best of existing data.” AbouZahr seems undaunted by the diffi culties she faces. “It is very easy to be disheartened. At the same time, data draw attention to an issue, so you can’t neglect them. You can’t say it is just too diffi cult—let’s not do this”, she says. “Without data it is very diffi cult to get issues onto the agenda.”

Sarah [email protected]

See Comment pages 1526 and 1527


Obituary


Bjørn Aage Ibsen Danish anaesthetist who founded the world’s fi rst intensive care unit. Born in Copenhagen, Denmark, on Aug 30, 1915, he died, aged 91 years, in Copenhagen on Aug 7, 2007.

The patient’s name was Vivi, and when Bjørn Ibsen treated her, she was 12 years old and dying of paralytic poliomyelitis. “Everyone expected her to die”, explains Preben Berthelsen, a Danish anaesthetist who interviewed Ibsen about the case. It was 1952, and Denmark was in the midst of a polio epidemic, so the doctors working under Henry Cai Alexander Lassen at Copenhagen’s Blegdamshospitalet had plenty of experience on which to base their assessment of the girl. But Ibsen, a freelance anaesthetist, thought he could save her. He proposed a radical departure from the standard treatment by suggesting that polio patients could be treated in the same way as he had been managing curarised patients during surgery. “He said, in the OR we intubate and ventilate them and they do alright. We could do exactly the same with the polio patients”, Berthelsen says.

Lassen was sceptical, but the situation was dire enough for him to give Ibsen a chance. A surgeon did a tracheostomy on the girl, but when Ibsen began trying to ventilate her, he found he could not push any air into her lungs. “At that moment, according to the legend, all the other doctors left the room”, Berthelsen says. But Ibsen administered thiopentone, and found he was then able to ventilate without diffi culty using intermittent positive pressure ventilation with a bag and Waters’ to-and-fro system with carbon dioxide absorption connected to a cuff ed tracheotomy tube. She regained

consciousness, and her temperature fell. This demonstration convinced Lassen the new technique was the way forward, so he organised shifts of about 1500 volunteer medical students, nurses, and retired staff to do this ventilation on polio patients. Supervised by anaesthetists and dedicated nursing staff , the work of these volunteer teams using Ibsen’s techniques slashed mortality rates from 87% to less than 15% among patients with bulbar poliomyelitis. The hospital then arranged for the same methods to be applied to other patients with respiratory failure, and concentrated them into three specially designated wards.

Ibsen had earned his medical degree at Copenhagen University, in 1940, and then worked in hospitals on the Jutland peninsula. In 1949, he travelled to the USA where he specialised in anaesthetics at Boston’s Massachusetts General Hospital. On the ship back to Denmark, his wife met Mogens Bjørneboe, deputy to Lassen at Blegdamshospitalet. After returning home, Ibsen worked as a freelance anaesthetist around Denmark and, in 1952, Bjørneboe called on him to help with a tetanus patient, which led to his being on hand weeks later when the polio epidemic erupted. “Ibsen very often attributed his successes to chance and coincidence”, Berthelsen said. But that modesty was balanced by a healthy self-belief, said Jørgen Viby-Mogensen, professor of anaesthesia and intensive care medicine at Copenhagen University Hospital who knew Ibsen for 40 years. “He wanted to be recognised, and was a strong person”, he said. At a time when the role of the anaesthetist was expanding into territory previously dominated by surgeons and others, confi dence was important, Viby-Mogensen said: “It was the same for many anaesthetists at the time. They had to fi ght, they had to believe in themselves.”

In 1953, after the polio outbreak had subsided, Ibsen was off ered a job at Copenhagen’s Kommunehospital, where he was asked to supervise surgical patients in the recovery room and the wards. Later that year, Ibsen turned the surgical recovery ward into a unit where all types of patients could receive critical care. “He took a surgical recovery room, got sole charge of it, and turned it into what he called an intensive therapy unit. He had his fi rst patient there in 1953”, explains Ron Trubuhovich, who has written about Ibsen’s contributions. This was the world’s fi rst dedicated intensive care unit; an achievement in itself, and all the more so because Ibsen was an anaesthetist, says Mogensen. “It was very diffi cult for him to get charge of a unit where patients were under the control of an anaesthetist.”

Although others had raised the idea of treating all critical patients in the same place, Ibsen was the fi rst to make it a reality. “He was in the forefront”, says Berthelsen. Ibsen won the Danish poliomyelitis and anaesthetic medals, and the J E Purkinje medal.

Stephen [email protected]

Correspondence

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


Cannabis and psychosis

Only two relevant studies of cannabis use and psychosis appeared in the 3 years between Theresa Moore and colleagues’ systematic review (July 28, p 319)1 and that of our group.2 One replicates the fi nding that adolescents who report higher cannabis use report more experience of unusual thoughts and perceptions, but cannot resolve issues of residual confounding, reporting bias, and reverse causation. Another reports an adjusted odds ratio for any psychotic symptom associated with any cannabis use of 0·72 (95% CI 0·30–1·74) and an equivalent odds ratio with cannabis dependence of 1·47 (0·55–3·93). Neither of these estimates provides meaningful evidence that cannabis causes schizophrenia.

Moore and colleagues did a meta-analysis; we did not because of the problems of spurious precision when using confounded or biased studies.1 Also problematic is the weakness of statistical tests for small-study bias when only a limited number of studies contribute to the assessment.3 Moore and colleagues interpret a p value of 0·13 as providing “no evidence” for small-study bias, but suggest that similar p values of 0·11 and 0·10 (our calculations from data reported in fi gure 3) provide “weak evidence” for an association between cannabis use and mental health damage.

In their conclusions, Moore and colleagues suggest that there is “now” suffi cient evidence to advise young people against cannabis use, implying that some new and more conclusive data on cannabis and psychosis have recently emerged. This is not the case; there has been an evidence base for some time to advise people not to use cannabis, because most of them smoke it with tobacco. Even when smoked alone, cannabis now seems to be detrimental to respiratory health.4 The evidence that it additionally causes mental health harm seems

to us as equivocal and diffi cult to interpret as it always has been. Moore and colleagues suggest that this issue of causation is “unlikely to be resolved by further longitudinal studies”. We disagree; the application of new methods within observational studies can provide more robust evidence.5 Further systematic reviews, adding little new data, are unlikely to help, particularly if they risk concretising misplaced certainty through emphasis of statistical as opposed to other aspects of evidence provided by observational studies.We declare that we have no confl ict of interest.

*John Macleod, George Davey Smith, Matthew Hickman, Matthias [email protected]

University of Bristol, Bristol BS8 2PR, UK (JM, GDS, MH); and University of Bern, Bern, Switzerland (ME)

1 Moore THN, Zammit S, Lingford-Hughes A, et al. Cannabis use and risk of psychotic or aff ective mental health outcomes: a systematic review. Lancet 2007; 370: 319–28.

2 Macleod J, Oakes R, Copello A, et al. The psychosocial sequelae of use of cannabis and other illicit drugs by young people: systematic review of longitudinal, general population studies. Lancet 2004; 363: 1579–88.

3 Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple graphical test. BMJ 1997; 315: 629–34.

4 Aldington S, Williams M, Nowitz M, et al. Eff ects of cannabis on pulmonary structure, function and symptoms. Thorax 2007; published online July 31. DOI: 10.1136/thx.2006.077081.

5 Macleod J, Hickman M. Understanding pathways to cannabis use and from use to harm. Int J Epidemiol 2006; 35: 680–82.

of observational studies are not inappropriate, but that the statistical combination of results should not be the main component of the review. In keeping with this, the primary focus of our review was not the strength of statistical evidence from the pooled results, but the thoroughness of any attempts made by each individual study to minimise the eff ects of bias, reverse causation, and confounding. It is because of the rigour of these assessments, and not because of the statistical evidence from the meta-analyses, that we concluded that there was weaker evidence for a causal relation between cannabis use and depression than for psychosis.

Macleod and colleagues raise a valid concern that by doing a meta-analysis, people might misinterpret the pooled results as defi nite evidence of a causal eff ect of cannabis on psychosis. That is why we made it very clear in our Summary and Discussion that the eff ects of residual confounding or bias cannot be eliminated as possible explanations for the associations seen, and that uncertainty about causality is inevitable in observational studies such as these.

Future studies can, and will, help to address some of the unknown questions relating to whether cannabis use increases the risk of psychotic disorders. New methods such as Mendelian randomisation2

referred to by Macleod and colleagues are likely to be very useful techniques within epidemiology, but require the presence of genetic variants that are strongly associated with cannabis use. Unfortunately there is no good evidence for any such candidates at present. Therefore we remain of the opinion that further observational studies, such as those included in the review, are unlikely to provide a defi nitive answer given that they will be subject to the same potential problems of confounding and bias.

For any individual, use of cannabis is quite unlikely to lead to psychotic illness, even if the relation is causal.

Authors’ replyWe agree with John Macleod and colleagues that it is not possible to be certain that the relation between cannabis use and psychotic illness is causal given the problems of confounding and reverse causation that are inherent to observational studies.

We carefully considered whether to present a narrative synthesis or a meta-analysis of results from these observational studies. We followed the advice provided by Egger and colleagues1—ie, that meta-analyses

Scie

nce

Phot

o Li

brar

y

Correspondence


However, having taken into account the methodological quality of each study, and the unavoidable limitations in interpreting the statistical evidence from observational studies, we believe that there is enough evidence for people to be made aware of the potential role of cannabis in increasing psychotic illness.We thank Anne Lingford-Hughes, Thomas Barnes, Peter Jones, and Margaret Burke for their contribution to this response. We declare that we have no confl ict of interest.

*Stanley Zammit, Theresa Moore, Glyn Lewiszammits@cardiff .ac.uk

Department of Psychological Medicine, Cardiff University, Cardiff CF14 4XN, UK (SZ, TM, ALH, GL); Department of Psychological Medicine, Imperial College, London, UK (TB); Department of Psychiatry, Cambridge University, Cambridge, UK (PJ); and Department of Social Medicine, University of Bristol, Bristol, UK (MB)

1 Egger M, Davey Smith G, Schneider M. Systematic reviews of observational studies. In: Egger M, Davey Smith G, Altman DG, eds. Systematic reviews in health care: meta-analysis in context, 2nd edn. London: BMJ Publishing Group, 2001: 211–27.

2 Macleod J, Hickman M. Understanding pathways to cannabis use and from use to harm. Int J Epidemiol 2006; 35: 680–82.

The paper by Theresa Moore and colleagues1 calls for some comment. As Moore and colleagues highlight, only controlled experiments could accurately corroborate a causal relation between cannabis use and psychosis. In the absence of such a possibility, they emphasise three of the criteria previously proposed by Hill:2 temporality (eg, inclusion of only longitudinal surveys), the strength of the association, and the consistency of the association.

We would like to call attention to Hill’s “specifi city” criterion, which has surprisingly been largely neglected until now in the discussion about the specifi c role of cannabis in causing chronic psychoses. For example, in an 18-month population-based longitudinal study, the risk of incident psychotic symptoms among cannabis users was double for those who also smoked tobacco.3 In the same study,

the risk for non-addicted cannabis users was 1·09 and for cannabis-addicted individuals 3·40. The risk of current smokers reporting incident psychotic symptoms was still 1·67 after adjustment for cannabis use. Because no separate data are available for tobacco use only (ie, for non-addicted smokers), the relative risk of 2·14 applies for a mixed group—ie, addicted and non-addicted.

Is there then a causal association between addictive behaviour in general and schizophrenia? Is it therefore reasonable to warn people only about the specifi c eff ects of cannabis on psychosis, or should tobacco smokers also be informed about their possible increased risk?We declare that we have no confl ict of interest.

*Daniele Fabio Zullino, Thomas Rathelot, Yasser [email protected]

Service d’abus de substances, Hôpitaux Universitaires de Genève, Rue verte 2, CH-1205, Geneva, Switzerand

1 Moore THM, Zammit S, Lingford-Hughes A, et al. Cannabis use and risk of psychotic or aff ective mental health outcomes: a systematic review. Lancet 2007; 370: 319–28.

2 Hill AB. The environment and disease: association or causation? Proc R Soc Med 1965; 58: 295–300.

3 Weiser M, Reichenberg A, Grotto I, et al. Higher rates of cigarette smoking in male adolescents before the onset of schizophrenia: a historical-prospective cohort etudy. Am J Psychiatry 2004; 161: 1219–23.

Theresa Moore and colleagues’ systematic review on the infl uence of cannabis use on psychosis and aff ective disorders1 is a milestone in the recent psychiatric literature and clearly provides evidence for a dose-dependent eff ect of cannabis on the development and course of disease.

The paper includes studies on the eff ect of time of fi rst consumption on the course of psychotic disease. These studies either point to a stronger eff ect of cannabis use on psychotic symptoms in those who use it before the age of 16 years, or show no age eff ect of fi rst cannabis use. By contrast, in a prospective

study on cognitive performance in people with schizophrenia and in healthy controls2 (divided into cannabis-users and cannabis-non-users), we reported a positive eff ect on cognitive performance in patients who started cannabis consumption before the development of the fi rst psychotic episode. This eff ect was even more pronounced when regular consumption started before the age of 17 years.

As accurately pointed out in the Comment by Merete Nordentoft and Carsten Hjorthøj,3 there is always the problem of reverse causality. In this case, it could mean that early cannabis consumption might refl ect the early appearance of negative symptoms of schizophrenia that benefi t from use of cannabis. Schizophrenia is associated with an altered endocannabinoid system.4 Thus we do not support the idea of a “well established” harmful eff ect on cognitive function as written by Nordentoft and Hjorthøj. But we agree that conclusions about causal eff ects will only become possible after prospective placebo-controlled studies with psychoactive cannabis components.We declare that w e have no confl ict of interest.

*O Schulte-Herbrüggen, M C Jockers-Scherü[email protected]

Department of Psychiatry and Psychotherapy, Charité-University Medicine Berlin, Campus Benjamin Franklin, Eschenallee 3, 14050 Berlin, Germany

1 Moore THM, Zammit S, Lingford-Hughes A, et al. Cannabis use and risk of psychotic or aff ective mental health outcomes: a systematic review. Lancet 2007; 370: 319–28.

2 Jockers-Scherübl MC, Wolf T, Radzei N, et al. Cannabis induces diff erent cognitive changes in schizophrenic patients and in healthy controls. Prog Neuropsychopharmacol Biol Psychiatry 2007; 31: 1054–63.

3 Nordentoft M, Hjorthøj C. Cannabis use and risk of psychosis in later life. Lancet 2007; 370: 293–94.

4 Laviolette SR, Grace AA. The roles of cannabinoid and dopamine receptor systems in neural emotional learning circuits: implications for schizophrenia and addiction. Cell Mol Life Sci 2006; 63: 1597–613.

Correspondence


UK classifi cation of cannabis: is a change needed and why?

In your July 28 Editorial (p 292),1 you debate the wisdom of the 2004 decision to reclassify cannabis from schedule B to C in the UK, drawing on increasingly strong evidence that cannabis use has a contributory role in psychosis.2

The focus on the association between cannabis use and psychosis is understandable, but it distracts attention from more frequent health problems. Psychosis might aff ect at most 2% of cannabis users, but 16% of adolescent cannabis users develop dependence,3 and in Australia many more users are arrested for personal use than in the UK, with the attendant adverse eff ect of acquiring a criminal record (fi gure).

The implicit assumption seems to be that the reclassifi cation of cannabis, and consequent more severe penalties, will reduce use. Is this plausible? Australian policy suggests not. Australian cannabis use increased steeply through the 1980s and 1990s. During this period,

diff erent States adopted diff erent policies: about half have criminal penalties for possession or use, and the remainder have fi nes. Cannabis use changed at similar rates across States irrespective of these penalties.4 This fi nding strongly suggests that other factors—such as social attitudes and perceived harms—are more important drivers of consumption than penalties for use.5

There is a danger that a debate about the reimposition of criminal penalties will distract the British government from more eff ective communication with the public about the risks of use—which, along with psychosis, should emphasise the more common health risks. Better research is needed on the eff ect (or lack thereof) of increased penalties on use and harm. It would be mistaken for the British community to assume that the public health problems arising from cannabis use can be solved by the stroke of a legislative pen.The National Drug and Alcohol Research Centre is funded by the Australian Government Department of Health and Ageing. The Department did not have any involvement in any aspect of the preparation of this letter. We declare that we have no confl ict of interest.

*Louisa Degenhardt, Wayne D Hall, Amanda Roxburgh, Richard P [email protected]

National Drug and Alcohol Research Centre (NDARC), University of New South Wales, Sydney, NSW 2052, Australia (LD, AR, RPM); and School of Population Health, University of Queensland, Brisbane, QLD, Australia (WDH)

1 The Lancet. Rehashing the evidence on psychosis and cannabis. Lancet 2007; 370: 292.

2 Degenhardt LD, Hall W. Is cannabis a contributory cause of psychosis? Can J Psychiatry 2006; 51: 555–65.

3 Hall W. The mental health risks of adolescent cannabis use. PLoS Med 2006; 3: e39.

4 Australian Institute of Health and Welfare. 2004 National drug strategy household survey—State and Territory supplement. Canberra: Australian Institute of Health and Welfare, 2005.

5 Hall W, Pacula RL. Cannabis use and dependence: public health and the public policy. London: Cambridge University Press, 2003.

Cannabis arrest/caution—AustraliaCannabis arrest/caution—England

6000

Year

Num

ber p

er m

illio

n po

pula

tion 5000

4000

3000

2000

1000

0

1993/94

1994/95

1995/96

1997/98

1998/99

1999/00

2000/01

2001/02

2002/03

2003/04

2004/05

2005/06

Cannabis use disorder hospital episode—AustraliaCannabis use disorder hospital episode—England

Figure: Number of hospital episodes for cannabis poisoning or dependence, and cannabis arrests or cautions per million population, Australia and EnglandSources: Australian Crime Commission; UK Home Offi ce. From 2004 (2003/04 label) the scheduling of cannabis in the UK went from a Class B to a Class C drug; data therefore refer to cautions not arrests. Number per million people aged 15–54 years for Australia, 15–59 years for England. Hospital episodes include separations for cannabis poisoning and cannabis use disorders (not including cannabis-induced psychosis).

Waist circumference in metabolic syndromePaul Zimmet and colleagues (June 23, p 2059)1 defi ne metabolic syndrome in children and adolescents. For consistency with the criteria for adults2 and association with insulin resistance, they included waist measurement in the defi nition.

Although the accumulation of visceral fat is an essential component of metabolic syndrome, and waist circumference measurement can be useful to estimate visceral fat, the reliability of waist circumference measurement has not been deter mined. In fact, the waist circumference that defi nes metabolic syndrome among Japanese people varied so much between studies3–5 that the International Diabetes Federation’s (IDF’s) ethnicity-specifi c cut-off for Japanese people was changed from the original (85 cm for men and 90 cm for women) to an alternative (90 cm for men and 80 cm for women) in just 2 years. If central obesity continues to be defi ned by waist circumference only, ethnicity-specifi c values will be changed by new fi ndings every few

For the International Diabetes Federation’s consensus worldwide defi nition of the metabolic syndrome see http://www.idf.org/webdata/docs/MetS_def_update2006.pdf

Correspondence


New growth standards

The report prepared by the Expert Group on Growth Standards of the Scientifi c Advisory Committee on Nutrition (SACN) and the Royal College of Paediatrics and Child Health (RCPCH),1 which was released in August, 2007, considered the use of the new WHO growth references2 in the UK. The committee rightly recommends that further fi eld trials in the UK are needed .

However, in their report the committee has not put enough emphasis on the fi rst 6 months of life. WHO and most paediatric societies recommend that infants should be exclusively breastfed for 6 months after birth. This provides the best nutritional basis for growth and the possibility of reducing later obesity, and programmes growth for a healthier adult life.

In the fi rst 6 months of life the new WHO growth reference is heavier than the commonly used UK growth reference and the growth reference of the US Centers for Disease Control and Prevention. Formula-fed infants tend to be heavier than breastfed infants, yet the new WHO reference, which is based on breastfed infants, is actually heavier during these early months than our older references that included formula fed-infants. In the Perth Infant Feeding Study, at 4 weeks and 10 weeks of age, exclusively breastfed infants were 350 and 700 g lighter, respectively, than those given formula.3 What is the reason for such a high rate of growth in the WHO study? The reason lies in the sample selection—96% of infants were eliminated and in taking only the top 4%, they were left with very high achievers (at least growth-wise).

All this would be academic except that breastfeeding is so important for infant health. Mothers and health workers usually assess the adequacy of breastfeeding by assessing growth during the fi rst few months of life. With the new WHO growth charts, infants who are exclusively breastfed will often seem to have a degree of growth failure

(see fi gure 1 in the report).4 The report recognises this: “The main diff erences between the charts were observed in the fi rst 2 years and particularly the fi rst 2 months.” Will this aff ect mothers’ breastfeeding behaviour? Will mothers add “top up” feeds of infant formula or even stop breastfeeding altogether to achieve the higher WHO growth rates?

This is a serious issue. Infants should be breastfed exclusively for 6 months. The new WHO chart has the potential to undermine breastfeeding during this period and therefore its introduction should not proceed without more trials.We declare that we have no confl ict of interest.

*Colin Binns, MiKyung [email protected]

School of Public Health, Curtin University, GPO Box U1987, Perth, Western Australia 6845, Australia

1 SACN/RCPCH Expert Group on Growth Standards. Application of the WHO growth standards in the UK, 2007. http://www.sacn.gov.uk/pdfs/report_growth_standards_2007_08_10.pdf (accessed Aug 24, 2007).

2 WHO. WHO child growth standards: methods and development. Geneva: World Health Organization, 2006.

3 Binns C, Graham K. Project report of the Perth Infant Feeding Study Mark II (2002-2004). Perth: Curtin University of Technology, 2005: 95.

4 Binns C, Lee M. Will the new WHO growth references do more harm than good? Lancet 2006; 368: 1868–69.

Department of ErrorOparil S, Yarows SA, Patel S, Fang H, Zhang J, Satlin A. Effi cacy and safety of combined use of aliskiren and valsartan in patients with hypertension: a randomised, double-blind trial. Lancet 2007; 370: 221–29—In this Article (July 21) the units shown for geometric mean plasma renin concentration in fi gure 4A (p 226) and should have been “ng/L”.

McCann D, Barrett A, Cooper A, et al. Food additives and hyperactive behaviour in 3-year-old and 8/9-year-old children in the community: a randomised, double-blinded, placebo-controlled trial. Lancet 2007; 370: 1560–67—In this Article (Nov 3), the second sentence of the third paragraph in the Study design and challenge protocols section should have read: “Active mix B included 30 mg of artifi cial food colourings (7·5 mg sunset yellow, 7·5 mg carmoisine, 7·5 mg quinoline yellow [E104], and 7·5 mg allura red AC [E129]) and 45 mg of sodium benzoate.”

years, and clinicians will not use the criteria.

Our experience shows that waist circumference is valid enough for measurement of prevalence in a group, but that accuracy is not suffi cient for diagnosis in an individual. To demonstrate this, we asked 10 independent, blinded examiners to measure the waist circumference and bodyweight of 20 volunteers (10 men and 10 women, mean age 28·4 years [SD 4·7]). The maximum interexaminer variation in waist circumference measurement was 7·8 cm (for volunteer number 7). However, the accuracy of bodyweight measurement was excellent: the maximum variation was just 0·4 kg (0·1 in body-mass index).

We think that the IDF should foster use of body-mass index as an alternative defi nition of central obesity, especially for adolescents, whose individual variations are larger than those of adults.We declare that we have no confl ict of interest.

*Satoru Yamada, Youko Tsukamoto, Junichiro [email protected]

Department of Internal Medicine, Kitasato Institute Hospital, 5-9-1 Shirokane, Minato-ku, Tokyo 108-8642, Japan (SY, JI); and Department of Nursing, Kitasato Institute Hospital, Tokyo, Japan (YT)

1 Zimmet P, Alberti G, Kaufman F, et al. The metabolic syndrome in children and adolescents. Lancet 2007; 369: 2059–61.

2 Alberti KGMM, Zimmet P, Shaw J. The metabolic syndrome: new world-wide defi nition from the International Diabetes Federation Consensus. Lancet 2005; 366: 1059–62.

3 Ito H, Nakasuga K, Ohshima A, et al. Detection of cardiovascular risk factors by indices of obesity obtained from anthropometry and dual-energy X-ray absorptiometry in Japanese individuals. Int J Obes Relat Metab Disord 2003; 27: 232–37.

4 Shiwaku K, Anuurad E, Enkhmaa B, et al. Predictive values of anthropometric measurements for multiple metabolis disorders in Asian populations. Diabetes Res Clin Pract 2005; 69: 52–62.

5 Hara K, Matsushita Y, Horikoshi M, et al. A proposal for the cutoff point of waist circumference for the diagnosis of metabolic syndrome in the Japanese population. Diabetes Care 2006; 29: 1123–24.

See Articles page 1560

Phot

olib

rary

Articles

www.lancet.com Vol 370 November 3, 2007 1543

Safety of the RTS,S/AS02D candidate malaria vaccine in infants living in a highly endemic area of Mozambique: a double blind randomised controlled phase I/IIb trial John J Aponte, Pedro Aide, Montse Renom, Inacio Mandomando, Quique Bassat, Jahit Sacarlal, M Nelia Manaca, Sarah Lafuente, Arnoldo Barbosa, Amanda Leach, Marc Lievens, Johan Vekemans, Betuel Sigauque, Marie-Claude Dubois, Marie-Ange Demoitié, Marla Sillman, Barbara Savarese, John G McNeil, Eusebio Macete, W Ripley Ballou, Joe Cohen, Pedro L Alonso

SummaryBackground Malaria remains a leading global health problem that requires the improved use of existing interventions and the accelerated development of new control methods. We aimed to assess the safety, immunogenicity, and initial effi cacy of the malaria vaccine RTS,S/AS02D in infants in Africa.

Methods We did a phase I/IIb double-blind randomised trial of 214 infants in Mozambique. Infants were randomly assigned to receive three doses either of RTS,S/AS02D or the hepatitis B vaccine Engerix-B at ages 10 weeks, 14 weeks, and 18 weeks of age, as well as routine immunisation vaccines given at 8, 12, and 16 weeks of age. The primary endpoint was safety of the RTS,S/AS02D during the fi rst 6 months of the study, and analysis was by intention to treat.Secondary endpoints included immunogenicity and analysis of new Plasmodium falciparum infections during a 3-month follow up after the third dose. Time to new infections in the per-protocol cohort were compared between groups using Cox regression models. This study is registered with ClinicalTrials.gov, number NCT00197028.

Findings There were 17 children (15.9%; 95% CI 9.5–24.2) with serious adverse events in each group. In the follow-up which ended on March 6, 2007, there were 31 serious adverse events in the RTS,S/AS02D group and 30 serious adverse events in the Engerix-B group, none of which were reported as related to vaccination. There were four deaths during this same follow-up period; all of them after the active detection of infection period had fi nished at study month 6 (two in RTSS/AS02D group and two in the Engerix-B group). RTS,S/AS02D induced high titres of anti-circumsporozoite antibodies. 68 fi rst or only P falciparum infections were documented: 22 in the RTS,S/AS02D group and 46 in the control group. The adjusted vaccine effi cacy was 65.9% (95% CI 42.6–79.8%, p<0.0001).

Interpretation The RTS,S/AS02D malaria vaccine was safe, well tolerated, and immunogenic in young infants. These fi ndings set the stage for expanded phase III effi cacy studies to confi rm vaccine effi cacy against clinical malaria disease.

IntroductionMalaria remains a leading global health problem, and control of the disease will need the improved use of existing interventions and the accelerated development of new control methods. The malaria vaccine candidate RTS,S (GlaxoSmithKline, Rixensart, Belgium), formulated with the adjuvant system AS02 or AS01, specifi cally targets the pre-erythrocytic stage of Plasmodium falciparum, and has been shown to confer complete or partial protection against experimental infec tion.1,2 Short-term protection against infection was shown in immunised adult men in The Gambia in 1998.3 In 2004, we reported a proof of concept study in African children aged 1 to 4 years, which established that RTS,S reduced the risk of P falciparum infection, uncomplicated malaria, and severe disease, and that this protection lasted for at least 18 months.4, 5

The goal is to register RTS,S for use in infants and children living in P falciparum endemic locations. The plan has two main drivers: the need to protect the youngest age groups and to include RTS,S in the Expanded Program of Immunisation (EPI) scheme. In most areas with stable malaria transmission, children younger than 2 years have a large and disproportionate incidence of severe disease

and death. There is growing recognition that malaria control strategies need to give top priority to protecting infants as soon as possible after birth.6,7 The endemic countries of sub-Saharan Africa often have weak health systems. In this context, the EPI is the best functioning system of regular health contacts with infants, capable of delivering millions of doses across the continent including rural areas.8

The RTS,S clinical development plan involves several African research centres across Africa, doing trials that will generate information on the selection of the fi nal formulation and schedule to be studied in a phase III trial for registration. In this paper, we report the results of the fi rst evaluation of safety, immunogenicity, and effi cacy test of concept of the RTS,S/AS02D vaccine formulation when administered to African infants living in a rural area of intense P falciparum transmission on a vaccination schedule staggered with the existing EPI schedule.

MethodsStudy DesignThe study was done at the Centro de Investigação em Saude de Manhiça (CISM, Manhiça Health Research

Lancet 2007; 370: 1543–51

Published OnlineOctober 17, 2007DOI:10.1016/S0140-6736(07)61542-6


Barcelona Centre for International Health Research (CRESIB), Hospital Clínic/Institut d’Investigacions Biomediques August Pi i Sunyer, Universitat de Barcelona, Barcelona, Spain (J J Aponte MD, M Renom MD, Q Bassat MD, S Lafuente MD, A Barbosa MD, P L Alonso MD); Centro de Investigação em Saude da Manhiça (CISM), Mozambique (J J Aponte, P Aide MD, M Renom, I Mandomando VetMed, Q Bassat, J Sacarlal MD, M N Manaca BSc, A Barbosa, B Sigauque MD, E Macete MD, P L Alonso); National Institute of Health, Ministry of Health, Mozambique (P Aide, I Mandomando, B Sigauque); Faculdade de Medicina, Universidade Eduardo Mondlane, Maputo, Mozambique (J Sacarlal); National Directorate of Health, Ministry of Health, Mozambique (E Macete); GlaxoSmithKline Biologicals, Rixensart, Belgium (A Leach MRCPCH, M Lievens MSc, J Vekemans MD, M-C Dubois MSc, M-A Demoitié MSc, W R Ballou MD, J Cohen PhD); and PATH Malaria Vaccine Initiative, Bethesda, MD, USA (M Sillman MSc, B Savarese RN, J G McNeil MD)

Correspondence to: Pedro L Alonso, Barcelona Center for International Health Research (CRESIB), Hospital Clinic, Universitat de Barcelona, C/ Rosello 132, 4-2, 08036 Barcelona, Spain [email protected]

Articles

1544 www.lancet.com Vol 370 November 3, 2007

Centre) in Ilha Josina and Taninga (50 km north of Manhiça, Mozambique) between June, 2005, and March, 2007. Ilha Josina has been previously described.4 Taninga faces Ilha Jossina directly across the Incomati River. The climate is subtropical with two distinct seasons including a warm and rainy season from November to April. Malaria transmission is perennial with some seasonality and mostly attributable to P falciparum. Anopheles funestus is the main vector. Amodiaquine plus sulphadoxine-pirimethamine was the fi rst line treatment for uncomplicated malaria until September, 2006, when it was changed to artesunate plus sulphadoxine-pyrimethamine. Since 2005, indoor residual spraying, initially with carbamates (lambdacyhalothrin) but changed to dichloro-diphenyl-trichloroethane in December, 2006, has been implemented by the Mozambique Ministry of Health. In accordance with national recommendations, pregnant women at screening were provided with and trained to use an insecticide-treated bednet. Primary health posts and maternity units at the study sites were upgraded including 24-h care and transport to the Manhiça District Hospital.

This phase I/IIb double blind randomised trial assessed the safety, immunogenicity, and effi cacy of RTS,S/AS02D when administered to infants at 10, 14, and 18 weeks of age. The protocol (NCT00197028, BB-IND 10514) was approved by the Mozambican National Bioethics Committee, the Hospital Clinic of Barcelona Ethics Review Committee, and the PATH Human Subjects Protection Committee, and was implemented according to the International Conference of Harmonization Good Clinical Practices guidelines. GlaxoSmithKline (GSK) monitored the study. A local Safety Monitor and a Data and Safety Monitoring Board oversaw the design, conduct, and results of the trial.

ParticipantsInformed consent was obtained from resident pregnant women in their third trimester. An information sheet was explained to groups of pregnant women and individual consent was obtained only after passing an

oral comprehension test of the study information. Signature (or thumbprint if illiterate) were witnessed and countersigned by an impartial member of the community. Consenting women were counselled and screened for HIV (Determine HIV1-2, Abbott Laboratories, Tokyo, Japan, and UNI-GOLD HIV, Trinity Biotech, Bray, Ireland) and hepatitis B (Determine HBsAg, Abbott Laboratories, Abbott Park, IL, USA). HIV-positive women were referred to Manhiça District Hospital for management as per National Guidelines, including reduction of mother-to-child transmission and provision of antiretroviral therapy as indicated. Mothers with hepatitis B were counselled about the risk to their infants, and hepatitis B vaccine was off ered for newborn babies.

Under separate maternal informed consent, infants were screened at between 6 and 12 weeks of age, including a medical history, examination, and blood sampling for baseline haematology, biochemistry, and immunology. Inclusion criteria included normal gestational period and absence of obvious medical abnormalities. Children of mothers who had hepatitis B or HIV, those not receiving oral poliovirus vaccine and Bacille Calmette-Guérin vaccine at least 1 week before study vaccination, and those receiving other pre-study vaccinations were excluded.

Photo identifi cation was provided and included names of the child and mother and census identifi cation number,9 and a unique study number issued at the screening. The fi rst infant was enrolled on Aug 23, 2005, and the last on Sept 12, 2006. Follow-up activities for the double-blind phase were completed on March 6, 2007, when the last recruited child reached their 6 month study visit.

ProceduresFigure 1 represents the trial design and follow-up scheme. Eligible children were enrolled the day of the fi rst vaccination with DTPw/Hib (TETRActHib Aventis Pasteur) and randomly assigned to receive three doses of study vaccines (RTS,S/AS02D [GSK Biologicals, Rixensart, Belgium] or hepatitis B vaccine [Engerix-B, GSK Biologicals, Rixensart, Belgium]) staggered by 2 weeks with DTPw/Hib and oral poliovirus vaccines and

Passive case detection

Screening XEnrolment XRTS,S/AS02D or Engerix-BDTPw/HiB and OPV

Study month 0 3 6 14Child age (weeks) 6–8 8 10 12 14 16 18 20 22 24 26 28 30 32 Blood sampling(safety, immunology) X X X X X XParasite clearance XActive detection of infection X X X X X X X

Visit day –14 0 14 30 44 60 74 90 104 118 132 146 160 180 404

Figure 1: Trial designDTPw=diphtheria-tetanus-whole-cell pertussis. Hib=Haemophilus infl uenzae type b. OPV=oral poliovirus vaccine.

Articles


administered at 10, 14, and 18 weeks of age. Block (1:1 ratio, block size of 2) randomisation was done at GSK Biologicals (SAS version 8). The code was released once databases had been monitored, checked for inconsistencies, and locked.

The infants randomly assigned to malaria vaccine received RTS,S/AS02D (0·5 mL) containing 25 µg of RTS,S and the AS02D (previously described10). RTS,S is a hybrid recombinant protein consisting of the P falciparum

circumsporozoite protein central tandem repeat and carboxy-terminal regions fused to the amino-terminus of the S antigen of hepatitis B virus (HBsAg). The proteins auto-assemble to form a particle that also includes unfused S antigen.

Vaccines were administered intramuscularly in the right (DTPw/Hib) or left (study vaccines) anterolateral thigh. To ensure that the study vaccines were delivered in a double-blind fashion, the preparation and vaccination procedures took place in separate limited-access rooms. Vaccine syringes were masked with opaque tape to prevent the mother from seeing the syringe contents. The vaccination team was not blinded but was not involved in any other study procedures. No other members of the trial team were aware of the which study vaccine any child received.

After each vaccination, infants were observed for an hour. Trained fi eld workers visited the children at home every day for the next 6 days to record local and general adverse events. As described elsewhere,11 unsolicited adverse events were recorded for 30 days after each dose, and serious adverse events were recorded throughout the study by use of the morbidity surveillance system based at the health facility. Haematological renal and hepatic function were assessed at several time points (fi gure 1). Anti-HBsAg antibody titres were measured at baseline and 1 month after the third dose of study vaccines. Anti-circumsporozoite antibody titres were measured at baseline, and at 1 and 3·5 months after the third dose of the study vaccines. Antibodies against diphtheria and tetanus toxins and polyribosyl ribitol phosphate for Haemophilus infl uenzae type b were measured 1 month after dose 3 of DTPw/Hib. Antibodies against Bordetella pertussis, were measured at baseline and 1 month after the third dose of DTPw/Hib.

Malaria infections by P falciparum were discovered by active detection and by passive case detection at health facilities in the study area. Active detection occurred at predefi ned intervals when a blood slide for parasitaemia determination was collected and the axillary temperature was recorded irrespective of symptoms. All children with positive slides were treated with fi rst line treatment and excluded from further assessment of active detection of infection. In all children, parasitaemia was presumptively cleared with amodiaquine (10 mg/kg per day for 3 days) plus sulfadoxine (25 mg/kg) and pyrimethamine (1·25 mg/kg) 2 weeks before the fi nal dose of RTS,S/AS02D or Engerix-B. Parasitaemia was checked at

the time of dose 3 and, if present, treated with second-line treatment based on artemether-lumefantrine. Only children without parasitemia started the active detection of infection, commencing 2 weeks after the third dose and repeated every-other week for 12 weeks.

Passive case detection was done through monitoring of all attendances to health facilities and ascertainment of episodes of clinical malaria, as described in detail elsewhere.4 All children with documented fever (37·5ºC or higher) or history of fever in the preceding 24 h, or pallor, had blood taken for parasite and packed-cell volume determinations. Children meeting admission criteria were referred to the Manhiça District Hospital for hospitalisation. Clinical management was provided according to standard national guidelines. On admission and discharge all relevant information was recorded on standardised forms.

Giemsa-stained blood slides were assessed using standard quality-controlled procedures described elsewhere.4 Biochemical parameters were measured using a dry biochemistry photometer VITROS DT II (Orto Clinical Diagnostics, Johnson & Johnson Company, Rochester, NY, USA). Haematological tests were done with a Sysmex KX-21N cell counter (Sysmex Corporation Kobe, Japan). Packed-cell volume was measured in heparinised microcapillary tubes using a Hawksley haematocrit reader after centrifugation.

Antibodies against the circumsporozoite repeat region were measured by a standard ELISA, using plates absorbed with recombinant R32LR with an assay cutoff of 0·5 EU/mL. Anti-HBsAg antibody levels were measured with a commercial radioimmunoassay (AUSAB, Abbott, IL, USA) with an assay cutoff of 10 IU/mL. Anti-PRP antibodies were measured by ELISA with a cutoff of 0·15 µg/mL. Anti-diphtheria and anti-tetanus antibodies titres were measured by ELISA with an assay cutoff of 0·10 IU/mL. Anti-whole-cell-B pertussis antibody titres were determined by ELISA (Labsystems, Vantaa, Finland) with an assay cutoff of 15 EU/mL.

Statistical analysisThe analysis was based on a prospectively-defi ned report and analysis plan. The primary endpoint was safety of the RTS,S/AS02D during the fi rst 6 months of the study. Secondary endpoints included immunogenicity analyses and estimation of vaccine effi cacy. All children who received at least one dose of DTPw/Hib were included in the intention-to-treat safety analysis.

Anti-circumsporozoite and anti-HBsAg antibody data were summarised by geometric mean titres with 95% CIs. Anti-circumsporozoite seropositivity was defi ned as 0·5 EU/mL or more, whereas seroprotection from hepatitis B was defi ned as 10 IU/mL or more.

The test of concept for vaccine effi cacy was done with a per protocol cohort, which included infants who met all eligibility criteria, completed the vaccination course, and contributed follow-up time during the active detection of

Articles


446 screened mothers

326 eligible mothers

329 livebirths

251 infants screened

214 eligible infants randomised

107 received DTPw/Hib and OPV dose 1 (intention-to-treat population)

1 consent withdrawal1 received DTPw at EPI clinic

1 consent withdrawal1 absent1 received DTPw at EPI clinic

1 consent withdrawal2 absent3 vaccination errors

1 consent withdrawal

1 absent


2 consent withdrawal2 lost to follow-up

2 deaths

102 received DTPw/Hib and OPV dose 2

96 received RTS,S/AS02D dose 2


94 received RTS,S/AS02D dose 3

93 started ADI follow-up (per-protocol population)

89 finished ADI follow-up month 6

87 at end of follow-up 89 at end of follow-up

91 finished ADI follow-up month 6

92 started ADI follow-up (per-protocol population)

94 received Engerix-B dose 3



107 received DTPw/Hib and OPV dose 1 (intention-to-treat population)



2 consent withdrawal1 absent1 received DTPw at EPI clinic3 vaccination errors

2 absent

1 absent



2 deaths

105 received RTS,S/AS02D dose 1 106 received Engerix-B dose 1

97 received Engerix-B dose 2

120 mothers not eligible

6 consent withdrawal30 not fulfilling eligibility criteria

1 migrated

Figure 2: Trial profi leDTPW=diphtheria-tetanus -whole-cell-pertussis. Hib=Haemophilus infl uenzae type b. OPV=oral poliovirus vaccine. EPI=Expanded Program of Immunisation. ADI=active detection of infection.

Articles


infection period. Cases were all fi rst or only infections of asexual P falciparum detected during the follow-up period starting 14 days after dose 3 of RTS,S/AS02D or Engerix-B and ending with the visit at study month 6 (a roughly 3-month follow-up).

Further analyses of vaccine effi cacy against clinical malaria were explored in an intention-to-treat cohort and included fi rst or only episodes from the time of enrolment until the date on which the last enrolled infant completed their active detection of infection follow-up (March 6, 2007) and in the same per protocol cohort and follow-up period as in the effi cacy assessment of new infections. The primary case defi nition for clinical malaria was fever (axillary temperature ≥37·5ºC) with an asexual parasitaemia of P falciparum of 500/μL. This defi nition has a reported sensitivity and specifi city of greater than 90%12. Other exploratory effi cacy analyses used secondary case defi nitions for clinical malaria including fever or history of fever in the previous 24 h plus any asexual P falciparum parasitaemia. Person years at risk was adjusted for absences from the study area and for antimalarial drug usage as previously described.4

Vaccine effi cacy was defi ned as 1 minus the hazard ratio. Vaccine effi cacy was adjusted by distance to health facility, calculated according to previously described methods,4 and community of residence. The adjusted vaccine effi cacy, assessed using Cox regression models, is reported throughout the text unless otherwise stated. The eff ect of anti-circumsporozoite antibody titres on the risk of malaria infection was assessed in the group who received RTS,S/AS02D by comparing the hazard ratio of infants with responses lower than the fi rst tertile against those with responses higher than the second tertile, as well as estimating the hazard ratio per doubling of anti-circumsporozoite antibody titre with Cox regression models. Finally, we compared the geometric mean titre of children who had at least one episode of malaria infection against those without documented malaria infection using a Wilcoxon Rank Sum test.

The sample size was based on an assessment of vaccine safety. A trial with 100 participants in each group had an 80% power to detect a diff erence in the proportion of adverse events of 26% or more if the frequency of an event in the Engerix-B group was 10% or more. A trial of this size also had 90% power to detect an effi cacy against malaria infection of 45% or more, assuming an attack rate of 75% or more in the control group over the surveillance period. Analyses were done with SAS 8 and STATA 9.

This study is registered with ClinicalTrials.gov, number NCT00197028.

Role of the funding sourceGlaxoSmithKline and Centro de Investigação em Saude de

Manhiça both received fi nancial support from the PATH Malaria Vaccine Initiative (MVI), which was involved in all aspects of the study design and interpretation. MVI funded this work through a grant from the Bill & Melinda

Gates Foundation. Core funding for CISM is provided by the Spanish Agency for International Cooperation (AECI). The corresponding author had full access to all the data in the study and had the fi nal decision to submit for publication.

Results214 screened children were enrolled in the trial and received the fi rst dose of DTPw/Hib vaccine (fi gure 2). Baseline characteristics for both groups are presented in table 1.

Figure 3 shows the proportion of all administered doses where a solicited symptom was recorded. The fi gure presents four groups, since the data for DTPw/Hib and oral poliovirus vaccine are segregated by randomisation assignment. The percentage of children with pain was high and similar in all groups. The relative percentage of symptoms was similar after each injection of RTS,S/AS02D and comparable in magnitude (data not shown). No grade-3 solicited symptom was documented in the RTS,S/AS02D group, and the incidence was low in the other groups (data not shown). The number of adverse events after DTPw/Hib vaccine or after study vaccines were similar in both groups (data not shown). In the intention-to-treat cohort, which included all children who had received at least one vaccine dose and who were followed-up until 6 months after the fi rst dose, there were 17 children (15·9%; 95% CI 9·5–24·2) with serious adverse events in each group. In the follow-up which ended on March 6, 2007, there were 31 serious adverse events in the RTS,S/AS02D group and 30 serious adverse events in the Engerix-B group, none of which were reported as related to vaccination. There were four deaths during this same follow-up period; all of them after the active detection of infection period had fi nished at study month 6 (two in RTSS/AS02D group and two in the Engerix-B group). All of the deaths were at home. Verbal autopsies suggested that one death in the RTS,S/AS02D group

Engerix-B (n=92) RTS,S/AS02D (n=93)

Age

Age at fi rst dose (weeks) 8·3 (1·0) 8·3 (1·4)

Sex

Girl 53 (58%) 43 (46%)

Boy 39 (42%) 50 (54%)

Community

Ilha Josina 64 (70%) 63 (68%)

Taninga 28 (30%) 30 (32%)

Distance (km)

0–5 79 (86%) 77 (83%)

5–10 9 (10%) 9 (10%)

10–17 4 (4%) 7 (7%)

Data are mean (SD) or number of children (%).

Table 1: Baseline characteristics

Articles


was due to septic shock, and the remaining three due to gastroenteritis and severe dehydration. The values and proportion of abnormal haematology and biological values after the fi rst dose and third dose were similar in both groups and did not seem suspicious (data not shown).

For the 214 infants enrolled in the trial, data for the EPI antigens responses are available for 151 participants after dose 3 (76 in the RTS,S/AS02D and 75 in the control group). There were no diff erences in seroprotection, seropositivity, and titres between the infants in the two

groups. All but three children reached seroprotective concentrations against all EPI antigens. These three children will be re-immunised with the antigens to which they have failed to respond.

Anti-circumsporozoite and anti-HBsAg antibody levels measured against circumsporozoite and HBsAg are shown in table 2. At screening, 24 (32%) of 76 infants in the RTS,S/AS02D group and 26 (34%) of 77 of infants in the Engerix-B group had low titres of detectable anti-circumsporozoite antibodies. A month after dose 3, 99% (70 of 71) of infants who received RTS,S/AS02D had detectable anti-circumsporozoite antibodies, whereas the corresponding fi gure for infants who received Engerix-B was 4% (3 of 68). 14 weeks after dose 3 (study month 6), the proportion of anti-circumsporozoite positives in the RTS,S/AS02D remained high (98%) but the geometric mean titre had decreased. The anti-circumsporozoite geometric mean titre in the Engerix-B group remained low, even though the prevalence of detectable antibodies increased to 20% (12 of 61).

Figure 4 shows the proportion of children with at least one episode of malaria infection during the active detection of infection follow-up starting 14 days after dose 3 of RTS,S/AS02D or Engerix-B up to study month 6. A total of 68 new infections were documented during this follow-up period, 22 in the RTS,S/AS02D group and 46 in the Engerix-B group. The crude vaccine effi cacy estimate was 62·2% (95% CI 37·1%–77·3%, p=0·0002) over the 3-month follow-up period. Adjusted by distance to the health centre and community of residence, the vaccine effi cacy was 65·9% (42·6%–79·8%, p<0·0001; table 3). The point prevalence of infection at study month 6 was similar between the two groups (5% in the RTS,S/AS02D group vs 8% in the control group, p=0·536), nor were there diff erences between mean parasite densities (2082 parasites per mL (SD 5604) in the RTS,S/AS02D group versus 2579 (6088) in the control group, p=0·85).

Exploratory endpoints contained in the report and analysis included effi cacy estimates for clinical malaria using diff erent cohorts and case defi nitions. Effi cacy based on an intention-to-treat cohort followed from month 0 to month 6, with the primary case defi nition of malaria (fi rst or only episode of fever with more than 500 parasites per mL) detected through both active detection of infection and passive case detection, was 35·5% (95% CI –7·5% to 61·3%, p=0·093). Further effi cacy estimates for clinical malaria in a per protocol cohort starting 14 days after dose 3 of RTS,S/AS02D or Engerix-B up to study visit at 6 months, the same follow-up period used in the primary vaccine effi cacy estimate for infection, are shown in table 3.

The relation of anti-circumsporozoite antibody titres to the risk of malaria was examined in several ways. Firstly, we compared anti-circumsporozoite antibody titres after dose 3 of RTS,S/AS02D or Engerix-B in the group of infants where no malaria infection was documented during the follow-up versus those that had at least one

0102030405060708090

100Pain Swelling Fever Irritability Loss of appetite Drowsiness

DTPw/Hib and OPV(Engerix–B group)

DTPw/Hib and OPV(RTS,S/AS02D group)

Engerix–B RTS,S/AS02D

Perc

enta

ge

Figure 3: Percentage of doses with solicited general symptoms reported during the 7 days after vaccinationDTPw=diphtheria-tetanus-whole-cell pertussis. Hib=Haemophilus infl uenzae type b. OPV=oral poliovirus vaccine.

Engerix-B RTS,S/AS02D

n value ( 95% CI ) n value ( 95% CI)

Anti-circumsporozoite

Baseline 77 0·4 (0·3–0·4) 76 0·4 (0·3–0·5)

30 days after third dose 68 0·3 (0·2–0·3) 71 199·9 (150·9–264·7)

106 days after third dose 61 0·4 (0·3–0·5) 53 58·8 (41·8–82·8)

Anti-HBsAg

Baseline 70 16·6 (11–25) 72 14 (9·6–20·5)

30 days after third dose 64 392·4 (297–518·5) 68 10 081·6 (7394·9–13 744·4)

Table 2: Geometric mean titres for Anti-CS and Anti-HBsAg

Engerix-B (E)RTS,S/AS02D (R)

100

80

60

40

20

00 0·5 1 1·5 2 2·5 3

p value for log-rank test: <0·0001

Cum

ulat

ive

perc

enta

ge

Months from 14 days after dose 3 of RTS,S or Engerix-BNumber at risk(E) 92 81 72 64 55 53 37(R) 93 89 85 82 78 74 45

Figure 4: Kaplan-Meier curves showing the cumulative percentage of participants with at least one episode of malaria infection

Articles


episode. On average anti-circumsporozoite antibody titres were higher in the group without malaria infection (208 vs 132, p=0·026). Secondly, the hazard ratio was 71% lower for infants in the higher tertile of the distribution of antibodies than for infants in the lower tertile (95% CI 8·4%–90·7%, p=0·035). Finally, we examined risk of malaria infection in relation to the increase in antibody titres. Doubling of antibody titres was associated with a reduction in the risk of a new infection of 6·4% (10·8–1·8, p=0·007). A ten times increase in anti-circumsporozoite titres is associated with a 19·8% reduction in the risk of new infec-tions (31·6–5·9).

DiscussionThe RTS,S/AS02D vaccine had a good safety profi le and was well tolerated. Its reactogenicity did not increase with repeated doses and was indistinguishable to that of the licensed EPI control vaccines. There were no related serious adverse events, nor an imbalance in unsolicited adverse events between both groups. These are encouraging results for the use of the RTS,S candidate and this adjuvant system in infants. The need for an initial assessment of the safety, immunogenicity, and effi cacy of the vaccine, in the absence of potentially confounding interactions with other antigens, led to this staggered design in relation to routinely administered EPI vaccines.

Around 34% of young infants living in this area of high transmission had detectable, albeit low, concentrations of anti-circumsporozoite antibodies at screening, probably transplacentaly transferred. In the control group, anti-circumsporozoite specifi c antibodies remained low over the fi rst 8 months of life in the face of substantial malaria transmission. In infants immunised with RTS,S/AS02D, anti-circumsporozoite antibody responses were similar to those measured in older children (1–4 years of age),4 suggesting that the anti -circumsporozoite responses were not signifi cantly modifi ed either by the presence of maternally-transferred antibodies or by the young age of these infants at the time of immunisation. Anti-hepatitis B antibody responses in the RTS,S/AS02D group were also consistent with those induced by this vaccine in the older children.

The RTS,S/AS02D vaccine or the control were administered at least 14 days apart from the EPI vaccines to keep the risk of interference and confounded safety data to a minimum. Indeed we note that responses to EPI antigens were similar in both groups (data not shown).

The trial was designed mainly to assess safety. However, in view of the high intensity of transmission in this area, the trial also had suffi cient power to assess effi cacy against new infections, a biologically relevant endpoint. Consistent with the idea that pre-erythrocytic vaccines should act through their ability to neutralise sporozoites and limit the number of infected hepatocytes or liver stage merozoites that enter the blood stream, we have previously shown that delay of time to a new infection translates to reductions in both mild and severe malaria disease.4

Vaccine effi cacy for new infections was 65% over a 3-month follow-up after completion of immunisations. This effi cacy estimate is higher than the 45% reduction reported in a previous trial in older children aged 1 to 4 years, who were followed-up in a similar system of active detection in the same study area.4 Nonetheless, caution should be exercised in the interpretation of this higher effi cacy in younger infants. Firstly, the follow-up periods were not identical, being slightly shorter in the infants than in the older children. Secondly, confi dence intervals of the two estimates overlap, and the diff erence could be due to chance. The prevalence of infection at the end of follow-up was lower in the RTS,S group than in the controls (5% vs 8%), but this diff erence is not signifi cant. This low prevalence of infection is in sharp contrast to the high incidence of infection over this same period (around 50%). Prevalence is a function of the incidence and the duration of an infection. In the case of these children, the low prevalence and the lack of power to detect a diff erence is probably the result of the intense follow up, which included regular screening of infection and prompt treatment. Vaccine effi cacy against infection reported here and previously in older children4 was more robust that was seen in malaria-immune adults in The Gambia, and the much higher antibody concen-trations against circumsporozoite in young children than in adults might contribute to these fi ndings.

Engerix–B RTS,S/AS02D Crude vaccine effi cacy Adjusted vaccine effi cacy

Events PYAR Rate Events PYAR Rate Effi cacy (95%CI) p Effi cacy (95%CI) p

Malaria infection

First or only episode of parasitaemia >0 46 17·2 2·7 22 21·8 1·0 62·2% (37·1%–77·3%) 0·0002 65·9% (42·6%–79·8%) <0·0001

Clinical malaria

First or only episode of fever and parasitaemia >500 per µL 22 19·6 1·1 9 22·6 0·4 64·4% (22·6%–83·6%) 0·009 65·8% (25·3%–84·4%) 0·007

First or only episode of fever or history of fever and parasitaemia >0 35 18·2 1·9 17 22·4 0·8 61·0% (30·2%–78·2%) 0·002 63·1% (33·6%–79·6%) 0·0009

PYAR= Person-years at risk. Rate=event/PYAR. Vaccine effi cacy estimates adjusted by distance from health facility and community.

Table 3: Vaccine effi cacy from 14 days after third dose (month 3) of Engerix-B or RTS,S/AS02D until visit at month 6

Articles


Effi cacy against clinical malaria was also suggested by the exploratory analyses, but, again, these positive results should be viewed with some caution. Firstly, the study was not powered to assess effi cacy against a clinical endpoint, but data were collected because they were relevant to our assessment of safety. Secondly, a cohort that is being actively followed up for detection of new infections implies repeated measurements and treatments that are likely to modify the natural course of an infection and its progression to clinical disease, rendering the effi cacy estimates for this endpoint of uncertain validity. The previous proof of concept trial in older children used separate cohorts to assess effi cacy against infection and clinical endpoints in order to avoid the introduction of bias, and it was in this context of separate cohorts that similar effi cacy estimates were seen for reductions in new infections and clinical disease. Nonetheless, this infant study has produced encouraging estimates of effi cacy against disease. Indeed, many of the clinical episodes were detected in the context of the active detection of infection visits. Considering the lack of acquired malaria immunity, infants became sick at low parasite densities,13 possibly explaining the similar levels of vaccine effi cacy documented against infection and disease over the same follow-up periods.

This trial was undertaken in an area of high transmission, but in the context of renewed and intense malaria control activities. All study participants were provided with free insecticide-treated bednets, two rounds of indoor residual spraying were done, with high coverage of the homes of study participants, and artemisinin combination therapies were introduced as fi rst-line treatment for malaria. The future use and deployment of a malaria vaccine should be seen in the context of comprehensive malaria control programmes. This test of concept trial suggests that vaccine-induced protection adds to the protection attributable to other control methods, particularly vector control methods. Hypothetically, synergism between the vaccine and vector control methods could take place through a reduction in the number of infective bites, which results in spacing out challenges, thus allowing for improved immune responses,14 and potentially through reductions in the infective innoculum of injected sporozoites which need to be controlled through vaccine-acquired immunity.

A frequent strategy in the process of antigen selection for inclusion in candidate vaccines has been the use of seroepidemiological data and its relation to infection or disease. Positive associations between certain immune responses induced through natural exposure and a reduced risk of disease have often been assumed deterministic of protection. We believe that immunological determinants of protection will best be established in the context of vaccine-induced effi cacy. This infant study provides evidence of a strong association between vaccine-induced anti-circumsporozoite antibodies and a reduction in the risk of infection. This is of importance

because up until now, immunogenicity was a marker of a response with no clearly proven relation to protection, which in turn could only be established through a clinical trial. This fi nding needs to be corroborated further in other trials, but this observation might be important in the clinical development plan of this vaccine.

The RTS,S/AS02D candidate malaria vaccine was safe, well tolerated, and highly immunogenic in young infants living in a rural area of southern Mozambique. These encouraging data need to be substantiated in phase III trials, but we have shown that a vaccine can reduce the risk of malaria infection in young African infants exposed to intense transmission of P falciparum. These results further strengthen the vision that a vaccine that can partly protect young African children and infants might contribute to the reduction of the intolerable burden of disease and death caused by malaria.ContributorsAll authors participated in the design, implementation, analysis, and interpretation of the study. PA was the principal investigator for the trial and coordinated the malaria vaccine teams in Manhiça and Barcelona. JA, PA, and MR led the implementation team, and together with PA were involved in all phases of the study and led the write up of the manuscript, which all other authors contributed to. AL led the clinical team at GlaxoSmithKline (GSK). JJA and ML led the data analysis. QB, JS, BS, and EM were heavily involved in fi eld and hospital activities, and safety surveillance. MS and SL were the project managers. M-CD is the malaria vaccine project manager at GSK. M-AD coordinated the immunology read-out team. Inacio Mandomando, MNM and AB coordinated all laboratory work at CISM. JV was responsible for safety at GSK. JC and RB had had a key involvement in the design, implementation, and interpretation of this trial and the malaria Vaccine Program at GSK. BS provided key support throught the trial. JGMc is Research Director at MVI.

Confl ict of interest statement MVI supports the development and testing of several malaria vaccines. AL, ML, JV, M-AD, M-CD, WRB, and JC are employees of GSK Biologicals. AL, M-CD, WRB, and JC own shares in GSK. JC and WRB are listed as the inventors of patented malaria vaccines; however, neither individual holds a patent for a malaria vaccine. None of the other authors declare any confl ict of interest.

AcknowledgmentsWe thank the study children and their parents, as well as the Ilha Jossina and Taninga communities and their leaders; the entire staff at CISM, particularly Antonio Timana, Fanuel Maximiano, and Charfudine Sacoor, who was responsible for the demographic surveillance in the area; Sergi Noguera and the administration team for providing invaluable support; the staff at the Barcelona Centre for International Health Research, for providing key support to this trial, particularly Joan Rodés, Antoni Trilla, Gonzalo Vicente, and Joan Vives. We also thank Salut Renom and Marc de Semir, who coordinated our communications team. GSK thank the staff of the Malaria Project team at GSK, in particular, Christine Swysen, Adelheid Davies, Opokua Ofori Anyinam, Conor Cahill, Laurence Vigneron, Nathalie Annez, Sabine Corachan, Manjula Krishna, Joelle Thonnard, Marie Chantal Uwamwezi, Bart Spiessens, Issam Jaimai, Delphine Beauport, and Philippe Dehottay. MVI would like to thank both its past and the current Directors, Melinda Moree and Christian Loucq, for their hard work and dedication; Regina Rabinovich, Director, Infectious Diseases Program, Bill & Melinda Gates Foundation, and Jessica Millman for their signifi cant contributions to the project; Esperança Sevene, the Local Safety Monitor for this trial provided invaluable and continuous support and guidance; Malcolm Molyneaux, Chairman of the Data and Safety Monitoring Board for his advice and support; the District, Provincial, and National Health Authorities, as well as to Joao Fumane, director of the National Institute of Health at the Ministry of Health. Aida Libombo and Pascoal Mocumbi have provided invaluable support over the years to the work at CISM.

Articles


References1 Kester KE, McKinney DA, Tornieporth N, et al. Effi cacy of

recombinant circumsporozoite protein vaccine regimens against experimental Plasmodium falciparum malaria. J Infect Dis 2001; 183: 640–47.

2 Stoute JA, Slaoui M, Heppner DG, et al. A preliminary evaluation of a recombinant circumsporozoite protein vaccine against Plasmodium falciparum malaria. RTS,S Malaria Vaccine Evaluation Group. N Engl J Med 1997; 336: 86–91.

3 Bojang KA, Milligan PJ, Pinder M, et al. Effi cacy of RTS,S/AS02 malaria vaccine against Plasmodium falciparum infection in semi-immune adult men in The Gambia: a randomised trial. Lancet 2001; 358: 1927–34.

4 Alonso PL, Sacarlal J, Aponte JJ, et al. Effi cacy of the RTS,S/AS02A vaccine against Plasmodium falciparum infection and disease in young African children: randomised controlled trial. Lancet 2004; 364: 1411–20.

5 Alonso PL, Sacarlal J, Aponte JJ, et al. Duration of protection with RTS,S/AS02A malaria vaccine in prevention of Plasmodium falciparum disease in Mozambican children: single-blind extended follow-up of a randomised controlled trial. Lancet 2005; 366: 2012–18.

6 Macete E, Aide P, Aponte JJ, et al. Intermittent preventive treatment for malaria control administered at the time of routine vaccinations in Mozambican infants: a randomized, placebo-controlled trial. J Infect Dis 2006; 194: 276–85.

7 Schellenberg D, Menendez C, Kahigwa E, et al. Intermittent treatment for malaria and anaemia control at time of routine vaccinations in Tanzanian infants: a randomised, placebo-controlled trial. Lancet 2001; 357: 1471–77.

8 Hutton G, Tediosi F. The costs of introducing a malaria vaccine through the expanded program on immunization in Tanzania. Am J Trop Med Hyg 2006; 75 (2 Suppl): 119–30.

9 Alonso P, Saute F, Aponte JJ. Manhiça DSS, Mozambique. In: Sankoh OA, Ngom P, Nyarko P, Mwageni E, Kahn K, eds. Population and health in developing countries; vol 1, population, health and survival at INDEPTH sites. Ottawa: International development research center, 2002: 189–95.

10 Macete EV, Sacarlal J, Aponte JJ, et al. Evaluation of two formulations of adjuvanted RTS, S malaria vaccine in children aged 3 to 5 years living in a malaria-endemic region of Mozambique: a Phase I/IIb randomized double-blind bridging trial. Trials 2007; 8: 11.

11 Macete E, Aponte JJ, Guinovart C, et al. Safety and immunogenicity of the RTS,S/AS02A candidate malaria vaccine in children aged 1–4 in Mozambique. Trop Med Int Health 2007; 12: 37–46.

12 Saute F, Aponte J, Almeda J, et al. Malaria in southern Mozambique: malariometric indicators and malaria case defi nition in Manhica district. Trans R Soc Trop Med Hyg 2003; 97: 661–66.

13 Trape JF, Rogier C. Combating malaria morbidity and mortality by reducing transmission. Parasitol Today 1996; 12: 236–40.

14 Diallo DA, Sutherland C, Nebie I, et al. Sustained use of insecticide-treated curtains is not associated with greater circulation of drug-resistant malaria parasites, or with higher risk of treatment failure among children with uncomplicated malaria in Burkina Faso. Am J Trop Med Hyg 2007; 76: 237–44.

Articles


Cost-eff ectiveness of drug-eluting stents in patients at high or low risk of major cardiac events in the Basel Stent KostenEff ektivitäts Trial (BASKET): an 18-month analysis Hans Peter Brunner-La Rocca,* Christoph Kaiser,* Alain Bernheim, Michael J Zellweger, Raban Jeger, Peter T Buser, Stefan Osswald, Matthias Pfi sterer, for the BASKET Investigators†

SummaryBackground Our aim was to determine whether drug-eluting stents are good value for money in long-term, everyday practice.

Methods We did an 18-month cost-eff ectiveness analysis of the Basel Stent KostenEff ektivitäts Trial (BASKET), which randomised 826 patients 2:1 to drug-eluting stents (n=545) or to bare-metal stents (281). We used non-parametric bootstrap techniques to determine incremental cost-eff ectiveness ratios (ICERs) of drug-eluting versus bare-metal stents, to compare low-risk (≥3·0 mm stents in native vessels; n=558, 68%) and high-risk patients (<3·0 mm stents/bypass graft stenting; n=268, 32%), and to do sensitivity analyses by altering costs and event rates in the whole study sample and in predefi ned subgroups. Quality-adjusted life-years (QALYs) were assessed by EQ-5D questionnaire (available in 703/826 patients).

Findings Overall costs were higher for patients with drug-eluting stents than in those with bare-metal stents (€11 808 [SD 400] per patient with drug-eluting stents and €10 450 [592] per patient with bare-metal stents, mean diff erence €1358 [717], p<0·0001), due to higher stent costs. We calculated an ICER of €64 732 to prevent one major adverse cardiac event, and of €40 467 per QALY gained. Stent costs, number of events, and QALYs aff ected ICERs most, but unrealistic alterations would have been required to achieve acceptable cost-eff ectiveness. In low-risk patients, the probability of drug-eluting stents achieving an arbitrary ICER of €10 000 or less to prevent one major adverse cardiac event was 0·016; by contrast, it was 0·874 in high-risk patients.

Interpretation If used in all patients, drug-eluting stents are not good value for money, even if prices were substantially reduced. Drug-eluting stents are cost eff ective in patients needing small vessel or bypass graft stenting, but not in those who require large native vessel stenting.

Introduction Whether drug-eluting stents are good value for money is unclear.1–4 Early studies were criticised for their single vessel interventions in selected patients with high rates of follow-up repeat angioplasties.2,4 By contrast, the BAsel Stent Kosten Eff ektivitäts Trial (BASKET),5 with almost two stents per patient and clinical follow-up only, was questioned because of its short follow-up of 6 months.6 Thus, valid data on cost-eff ectiveness of drug-eluting stents over a reasonable time span in real-world patients are still lacking, but much needed.7 In addition to costs, cost-eff ectiveness is aff ected by the rate of follow-up events, mainly target vessel revascularisations avoided by use of drug-eluting stents.6,8 Thus, duration of follow-up and angiography-driven events aff ect cost-eff ectiveness substantially.9

Additionally, stenting strategies that use drug-eluting stents only have been criticised because of very late stent thrombosis.10–13 This rare event aff ects incremental cost-eff ectiveness assessments and corresponding sensitivity analyses.8 Therefore, we did a long-term cost-eff ectiveness analysis of BASKET at 18 months, as prespecifi ed by protocol, to defi ne the long-term incremental cost-eff ectiveness ratio (ICER) of drug-eluting versus bare-

metal stents in unselected patients, to defi ne such ratios in patients at low risk and those at high risk of major cardiac events, and to do detailed sensitivity analyses including prespecifi ed subgroups to better describe the scope of these fi ndings.

Methods Patients and proceduresThe methods, including patient selection, clinical outcome measures, and cost assessments have been described previously.5 Briefl y, all patients receiving percutaneous coronary interventions (PCI) and stenting between May, 2003, and May, 2004, at our institution were included, irrespective of indication for PCI. Only patients with vessels of 4 mm or greater (n=23), in-stent restenosis (49), or who did not consent (90) were excluded. 826 patients were randomised 2:1 to a drug-eluting stent (sirolimus-eluting Cypher stent, Cordis, Miami Lakes, FL, USA; or paclitaxel-eluting Taxus stent, Boston Scientifi c Corporation, Natick, MA, USA), or bare-metal stents (Vision, Guidant Corporation, Indianapolis, IN, USA) and followed for 18 months for major cardiac events (ie, cardiac death, non-fatal myocardial infarction, symptom-driven target vessel revascularisations) and

Lancet 2007; 370: 1552–59

*Contributed equally

†Listed in reference 5

Department of Cardiology, University Hospital, Basel,

Switzerland (H P Brunner-La Rocca MD,

C Kaiser MD, A Bernheim MD, M J Zellweger MD, R Jeger MD,

Prof P T Buser MD, Prof S Osswald MD,

Prof M Pfi sterer MD)

Correspondence to: Prof Matthias Pfi sterer,

Department of Cardiology, University Hospital, CH-4031

Basel, Switzerlandpfi [email protected]

Articles


related costs including stent costs, costs during the initial hospitalisation, and follow-up costs. Myocardial infarction was diagnosed on the basis of typical cardiac enzyme sequences, symptoms, and ECG changes, in accordance with current guidelines.14 Interventions driven by a lesion in the same epicardial vessel as initially treated was defi ned as target vessel revascularisation. All events were adjudicated by an independent committee blinded to the stent type used. The study was approved by the local ethics committee and patients gave written informed consent.

Quality-of life-adjusted life-years (QALYs) gained were assessed from the self-administered EQ-5D questionnaire at months 6 and 18.15 Since the acute situation at baseline, especially in those with acute coronary syndromes, disproportionally aff ects quality-of-life assessment, an EQ-5D index of 0·9 was used at baseline in all patients. This adjustment was deemed to be appropriate in view of the well-balanced baseline characteristics between patients with bare-metal stents and those with drug-eluting stents. Changing the EQ-5D index at baseline did not aff ect our results (data not shown).

Costs were determined from the perspective of third party payers, on the basis of procedures, stents used, and days spent in hospital, based on the Swiss medical tariff TARMED in 2004, as previously described.5 Specifi c costing was done for each patient. The same stent list prices were used here as in our previous analysis (mean price of sirolimus-eluting stent €2275, paclitaxel-eluting stent €1935, bare-metal stent €1260),5 although the prices of both bare-metal and drug-eluting stents have since decreased. Price reductions were greater for bare-metal stents than for drug-eluting stents, leading to a somewhat larger price diff erence between the two stent types. Considering the short period covered, costs and benefi ts were not discounted. ICERs were determined on the basis of the relation of costs and clinical events or QALYs for drug-eluting stents versus bare-metal stents.16

To investigate cost-eff ectiveness in patients at low risk and in those at high risk of major cardiac events, the two groups of patients showing the strongest interaction with stent type were defi ned as in a previous outcome analysis;17 patients with small vessel or bypass graft stenting were deemed to be high risk (n=268, 32%) and those with large native vessel stenting (≥3·0 mm) deemed to be low risk (n=558, 68%). Subgroups with additional high-risk characteristics were also used as prespecifi ed—including age over 65 years, presence of diabetes, three-vessel disease, long stented segments (cutoff 24 mm), and off -label use of drug-eluting stents5—for subgroup analyses and to test the sensitivity of the results.

Statistical analysisTwo-group comparisons were done with Fisher’s exact test and with the independent samples t test or

Mann-Whitney U test, as appropriate. Distributional assumptions were assessed by use of the one-sample Kolmogorov-Smirnov test. Non-parametric bootstrapping was used to estimate diff erences in average costs and major adverse cardiac events or QALYs for the ICER, and to assess the shape of the joint sampling distribution of the diff erences in average individual costs and eff ects between the two treatment groups.18,19 Findings are presented in cost-eff ectiveness planes using 5000 boot-strap samples drawn from the original dataset. From these data, the probability of a cost-eff ective scenario with drug-eluting stents was calculated in relation to willingness to pay a certain amount to prevent one major adverse cardiac event and to gain one QALY. Arbitrarily, a willingness to pay €10 000 was considered to be cost eff ective to prevent one major adverse cardiac event20 and €40 000 to gain 1 QALY.21 Multiple imputation by both expectation-maximisation and linear regression methods (fi ve imputations) incorporating random variation and including all relevant variables for estimation was used to assess the eff ect of missing data on QALYs.22 The SE of QALY estimates by multiple imputation was 0·05. As per protocol, primary analyses were on the basis of major adverse cardiac events whereas secondary analyses were

Overall(n=826)

Drug-eluting stents (n=545)

Bare-metal stents (n=281)

Sex (male) 650 (79%) 427 (78%) 223 (79%)

Age (years) 64 (11) 64 (11) 64 (11)

Diabetes 154 (19%) 93 (17%) 61 (22%)

Hypertension 550 (67%) 358 (66%) 192 (68%)

Hypercholesterolaemia 628 (76%) 414 (76%) 214 (76%)

Current smoking 238 (29%) 151 (28%) 87 (31%)

Previous myocardial infarction 226 (27%) 151 (28%) 75 (27%)

Previous PCI 133 (16%) 91 (17%) 42 (15%)

Previous CABG 105 (13%) 70 (13%) 35 (12%)

Clinical presentation

STEMI 176 (21%) 115 (21%) 61 (22%)

Unstable 301 (36%) 200 (37%) 101 (36%)

Stable 349 (42%) 230 (42%) 119 (42%)

Glycoprotein IIb/IIIa blockers 212 (26%) 141 (26%) 71 (25%)

Multivessel disease 566 (69%) 371 (68%) 195 (69%)

Bypass graft PCI 47 (6%) 34 (6%) 13 (5%)

Total chronic occlusions 28 (3%) 14 (3%) 14 (5%)

Bifurcations 44 (5%) 27 (5%) 17 (6%)

Number of stented segments 1·5 (0·7) 1·6 (0·7) 1·5 (0·7)

Number of implanted stents 1·9 (1·1) 1·9 (1·1) 1·9 (1·0)

Total stent length (mm) 34 (20) 34 (20) 32 (20)

Stent length per lesion (mm) 28 (15) 28 (15) 27 (16)

≥1 stents <3·0 mm 229 (28%) 160 (29%) 69 (25%)

Off -label use* 548 (66%) 376 (69%)† 172 (61%)

Data are n (%) or mean (SD). CABG=coronary artery bypass grafting. PCI=percutaneous coronary intervention. STEMI=ST-elevation myocardial infarction. *Off -label use according to recent FDA guidelines.23 †p=0·03 vs bare-metal stents.

Table 1: Baseline characteristics

Articles


on the basis of QALYs. The criterion for statistical signifi cance was set at α=0·05.

Sensitivity analyses were done with gradual changes of all factors that could aff ect ICERs—ie, costs (stent costs, hospitalisation costs including intensive care, follow-up angiography, and intervention costs) and eff ectiveness (changes in the diff erence of major adverse cardiac events and in QALYs). Gradual changes in costs were calculated for each patient and these values used for further analysis. The eff ects of all factors on the probability of a cost-eff ectiveness scenario and changes required to reach an arbitrary ICER threshold were calculated with non-parametric bootstrap methods in the whole study sample and in the subgroups.

Analyses were done with SPSS version 14.0.

Role of the funding sourceThe study sponsor had no role in the study design, data collection, data analysis, data interpretation, or writing of the report. All authors had access to the data. HPB-LR, CK, and MP had the fi nal responsibility for the decision to submit for publication.

ResultsBaseline characteristics were similar in patients who had drug-eluting stents and in those with bare-metal stents, except for off -label stent use, which was somewhat more frequent in patients with drug-eluting stents (table 1). Only about 40% of patients had stable coronary disease; complex lesion morphology and multivessel disease were frequent. Overall, 66% (n=548) had off -label indications, comparable with data from recently published registries.24,25

Follow-up was complete in 813 (98%) of 826 patients. Ten patients were alive and well but did not consent to follow-up questioning, whereas three had left the country and could not be located. Follow-up was censored in these patients at the time of last contact, but they were included in this analysis. Exclusion of these patients led to identical results (data not shown).

Rates of major adverse cardiac events or cardiac death and non-fatal myocardial infarction at 18 months were not signifi cantly diff erent between patients with bare-metal stents and those with drug-eluting stents (fi gure 1). The rate of non-myocardial infarction-related target vessel revascularisations, however, was 35% lower after implantation of drug-eluting stents than after implantation of bare-metal stents (fi gure 1). There were 101 major adverse cardiac events in those patients who received drug-eluting stents (45 with sirolimus-eluting stents, 56 with paclitaxel-eluting stents) and 58 in those treated with bare-metal stents. With bootstrap techniques, the mean rate of major adverse cardiac events was 0·185 (SD 0·021) per patient with drug-eluting stents and 0·206 (0·026) per patient with bare-metal stents over the 18 months, corresponding to a mean diff erence of 0·021 (0·033) major adverse coronary events per patient (p=0·50).

0

0

2000

Drug-elutingstents

Bare-metalstents

Drug-elutingstents

Bare-metalstents

Drug-elutingstents

Bare-metalstents

4000

6000

8000

10 000

12 000

14 000

5

10

15

20

25

30

35A

D

B C

MACE

Initial stentCosts for:

Overall

Mea

n co

sts (

€)Pa

tient

s with

eve

nts (

%)

High risk Low-risk

Initial hospitalisation Follow-up hospitalisation Follow-up intervention

Death/MI Non-MITVR

*

* *

*

†

*

*

Bare-metal stentsDrug-eluting stents

MACE Death/MI Non-MITVR

MACE Death/MI Non-MITVR

Figure 1: Event rates and mean overall costs per patient after 18 monthsEvent rates in the whole study sample (A), in high-risk patients (B), and in low-risk patients (C); and mean overall costs, bootstrap technique (D). *p<0·05, †p<0·01 between bare-metal stents and drug-eluting stents. MACE=major adverse cardiac event. MI=non-fatal myocardial infarction. non-MI TVR=symptom-driven target vessel revascularisation.

0·6

0·50·4

0·3

0·2

0·1

QAL

Ys

0·7

0·8

0·9

1

0

1·11·2

1·3

1·4

1·5

p=0·02

Bare-metalstents

Drug-elutingstents

p=0·004 p=0·28

Bare-metalstents

Drug-elutingstents

Bare-metalstents

Drug-elutingstents

Overall High risk Low-risk

Figure 2: QALYs gained with bare-metal and drug-eluting stents in the whole study sample (A), in high-risk patients (B), and in low-risk patients (C)Boxes represent median and IQR, whiskers are 10th and 90th percentile.

Articles


The EQ-5D questionnaire was completed at 6 and 18 months by 703 (85%) patients. Quality of life was higher in patients treated with drug-eluting stents than in those treated with bare-metal stents after 6 months (median 0·99, IQR 0·87–1·00 vs 0·90, 0·79–1·00; p=0·024), but this diff erence was not signifi cant after 18 months (0·99, 0·83–1·00 vs 0·89, 0·79–1·00; p=0·12). Overall, more QALYs were gained in patients with drug-eluting stents than in those treated with bare-metal stents (fi gure 2). With bootstrap techniques, 1·32 (SD 0·02) QALYs were gained per patient in those with drug-eluting stents, compared with 1·29 (0·01) per patient in those with bare-metal stents. The number of QALYs gained in patients with non-fatal myocardial infarction was not signifi cantly lower than in patients without events (median 1·37, IQR 1·22–1·47 vs 1·41, 1·31–1·47; p=0·076); by contrast, signifi cantly fewer QALYs were gained in patients with symptom-driven target vessel revascularisation than in those with no events and in those with non-fatal myocardial infarction (1·33, 1·19–1·39; p<0·0001 vs no events, p=0·032 vs non-fatal myocardial infarction). Multiple imputation for missing data on QALYs resulted, on average, in a mean diff erence between patients with drug-eluting stents and those with bare-metal stents of 0·02 (SD 0·02) QALY gained per patient.

Since 83% of patients had no follow-up events, follow-up costs were fairly low. Total costs were thus driven mainly by initial hospitalisation and stent costs (fi gure 1). The higher costs of drug-eluting stents than of bare-metal stents (p<0·0001 for the diff erence in cost between the two stents) resulted in a signifi cant diff erence in total costs between the two groups. Median costs after 18 months were €9810 (IQR 6020–14 418) per patient in those with drug-eluting stents and €8200 (4270–12 720) per patient in those treated with bare-metal stents (p<0·0001). Using bootstrap techniques, mean costs were €11 808 (SD 400) per patient in those with drug-eluting stents and €10 450 (592) per patient in those with bare-metal stents, resulting in a mean diff erence in total costs of €1358 (SD 717) per patient (p<0·0001).

On the basis of diff erences in mean number of events and overall costs, the ICER was €64 732 to prevent one major adverse cardiac event (similar for sirolimus-eluting stents [€63 691] and paclitaxel-eluting stents [€70 725]). The cost-eff ectiveness plane (fi gure 3) shows that an all-drug-eluting stent strategy had a chance of 71% to be more eff ective and more expensive, of 3% to be more eff ective and cheaper, and of 26% to be less eff ective and more expensive than an all-bare-metal stent strategy. The probability of the use of drug-eluting stents being cost eff ective was only 0·114 at the arbitrary threshold of €10 000 (fi gure 4).

Incremental costs per QALY gained were €40 467 for drug-eluting stents (no diff erence between sirolimus-eluting and paclixatel-eluting stents). The cost-eff ective-

ness planes (data not shown) were slightly more in favour of drug-eluting stents use than of bare-metal stents with respect to QALY (85% more eff ective/more expensive; 5% more eff ective/cheaper, 9% less eff ective/more expensive). The probability of an all-drug-eluting stent strategy being cost eff ective to gain 1 QALY was 0·495 at the arbitrary threshold of €40 000 (fi gure 4). However, multiple imputation for missing data on QALYs reduced cost-eff ectiveness (expectation-maxi-misation method: €56 810, probability of cost-eff ective scenario 0·446; linear regression: €68 703, probability of cost-eff ective scenario 0·377), with no single imputation showing a better cost-eff ectiveness of drug-eluting stents than with the use of the existing data only (range €51 278–153 551).

28·5% 71·1%

0·1% 3·0%

24·7%

0·2% 0·8%

2·2% 27·0%

0·5%

Diffe

renc

e in

mea

n co

sts (

€)Di

ffere

nce

in m

ean

cost

s (€)

Diffe

renc

e in

mea

n co

sts (

€)

70·4%

A

B

C

–7000–0·15 –0·05 0·05 0·15

Difference in mean number of major adverse cardiac events

0·25 0·35

–5000

–3000

–1000

1000

3000

5000

–7000

–5000

–3000

–1000

1000

3000

5000

–7000

–5000

–3000

–1000

1000

3000

5000

74·3%

Figure 3: Incremental cost-eff ectiveness of drug-eluting and bare-metal stents using 5000 samples in bootstrap analysisIncremental cost-eff ectiveness in the whole study sample (A), low-risk patients (B), and high-risk patients (C). Percentages give frequencies of samples in each quadrant (upper right: drug-eluting stents more eff ective/more expensive; upper left: drug-eluting stents less eff ective/more expensive [dominated]; lower right: drug-eluting stents more eff ective/cheaper [dominant]; lower left: drug-eluting stents less eff ective and cheaper).

Articles


In patients aged 65 years or over, in those with one or more treated segment, in those with longer stents per treated lesion, and in those with off -label use, the use of drug-eluting stents resulted in ICERs of about €10 000 to prevent one major adverse cardiac event (table 2). By contrast, drug-eluting stents were either less eff ective and more expensive or clearly not cost eff ective in all other subgroups (table 2). Similar results were found regarding QALYs gained, despite the higher threshold of €40 000 (table 2).

As expected, ICERs were sensitive to stent costs (fi gure 5) and number of stents used per patient. Each 1% reduction in the price of drug-eluting stents would have resulted in a reduction of the ICER to prevent one major adverse cardiac event of €1879, and each 1% reduction in the number of stents in a lowering of the ICER of €809. Thus, to achieve the arbitrary threshold

ICER of €10 000 to prevent one major adverse cardiac event, a reduction in the cost of drug-eluting stents of about 29% would have been required, or a reduction in number of stents from 1·9 to 0·7 stents per patient. By contrast, changes in hospitalisation (€147 increase of ICER per 1% reduction of costs) and intervention costs (€31 increase of ICER per 1% reduction of costs) did not aff ect ICERs signifi cantly. A higher eff ectiveness of drug-eluting stents would also have enhanced ICERs (fi gure 5). Thus, at present costs, a mean reduction in the average number of major adverse cardiac events of 0·075 events per patient—ie, a more than three times higher eff ectiveness—would be required to achieve an ICER of €10 000. Sensitivity analyses of incremental costs per QALY showed similar results. Incremental costs per QALY would have been reduced by €1320 for each 1% reduction in the price of drug-eluting stents and by €420 per 1% reduction in number of stents used; the eff ects of hospitalisation and intervention costs were negligible.

18-month rates of major adverse cardiac events did not diff er between patients with drug-eluting stents and those with bare-metal stents in low-risk groups (fi gure 1; bootstrap mean 0·169 [SD 0·022] vs 0·144 [0·028] events per patient; p=0·42), but was signifi cantly higher in those with drug-eluting stents than in those with bare-metal stents in high-risk groups (fi gure 1; bootstrap mean 0·218 [0·043] vs 0·358 [0·059] events per patient; p=0·002). Similarly, there was no diff erence between the two stent types regarding QALYs in low-risk patients, but signifi cantly more QALYs were gained with drug-eluting stents than with bare-metal stents in high-risk patients (p=0·004; fi gure 2). Using bootstrap techniques, a mean of 1·32 (SD 0·02) QALYs were gained per patient by both those with drug-eluting stents and those with bare-metal stents in the low-risk group; by contrast, 1·31 (0·02) QALYs were gained per patient in those with drug-eluting stents and 1·21 (0·04) in those with bare-metal stents in the high-risk group.

Overall Low risk High risk

MACE QALY MACE QALY MACE QALY

Overall €64 732 €39 641 – €373 785 + +

Age <65 years – €65 014 – €163 243 – €17 742

Age ≥65 years €11 681 €300 – – + +

One or two vessel disease €98 744 €43 770 €269 268 €72 946 €11 333 €5641

Three-vessel disease €442 431 – – – + +

No diabetes €45 173 €30 754 €69 553 €51 690 €10 504 €6733

Diabetes mellitus – – – – + +

One segment €384 091 €70 228 – €146 187 + +

>1 segment €8567 + – – + +

<24 mm per lesion – €112 307 – – + Cost saving, no eff ect on QALY

≥24 mm per lesion €8042 €3471 €109 559 – + +

On-label use – €273 986 – €375 927 + Cost saving, no eff ect on QALY

Off -label use €11 829 €6863 – €224 591 + +

–=Drug-eluting stents less eff ective and more expensive than bare-metal stents. +=Drug-eluting stents more eff ective and less expensive than bare-metal stents. MACE=major adverse cardiac event. QALY=quality-adjusted life-year.

Table 2: ICERs in diff erent subgroups

0

0·2

0·4

0·6

0·8

1·0

Willingness to pay for 1 QALY (€)

0

0·2

0·4

0·6

0·8

1·0

All patientsLow riskHigh risk

0 50 000 100 000 150 000 200 000

Prob

abili

ty o

f cos

t-eff

ecti

vene

ssA B

Willingness to pay to prevent one majoradverse cardiac event (€)

0 50 000 100 000 150 000 200 000

Figure 4: Probability of cost-eff ectiveness in relation to willingness to pay to prevent one major adverse cardiac event (A) and to gain 1 QALY (B) with the use of drug-eluting compared with bare-metal stents

Articles


Initial stent costs were signifi cantly higher in patients with drug-eluting stents than in those with bare-metal stents in both low-risk and high-risk patients (p<0·0001 for both groups, fi gure 1). However, in high-risk patients, follow-up costs were lower in those with drug-eluting stents than in those with bare-metal stents (mean €2062 vs €3747 per patient, p=0·028), whereas they did not diff er signifi cantly in low-risk patients (mean €2210 vs €1369 per patient, p=0·46).

There were substantial diff erences in ICERs between low-risk and high-risk patients (table 2). In low-risk patients, almost 75% of samples were in the less eff ective and more expensive zone of the cost-eff ectiveness plane (fi gure 3) and the probability of drug-eluting stents being cost eff ective was only 0·016 at the ICER threshold of €10 000 to prevent one major adverse cardiac event (fi gure 4). By contrast, in high-risk patients, 71% fell in the more eff ective and cheaper zone (fi gure 3) and the probability of drug-eluting stents being cost eff ective was 0·874 (fi gure 4). A similar trend was seen regarding QALYs: in low-risk patients, 41% of samples were in the less eff ective and more expensive zone with a low probability of a cost-eff ective scenario (probability of 0·111 at the arbitrary threshold of €40 000 to gain 1 QALY), whereas in high-risk patients, 76% of samples were in the more eff ective and cheaper zone, and the probability of drug-eluting stents being cost eff ective was very high (probability of 0·975; fi gure 4).

Sensitivity analyses revealed that changes in hospitalisation and follow-up costs, as well as reduced stent use per patient, had little eff ect on ICERs in these two subgroups (data not shown). Interestingly, there was also little eff ect of reduced prices of drug-eluting stents and changes in event rates (fi gure 5). Thus, whereas even substantial reductions in the price of drug-eluting stents would not have resulted in a cost-eff ective scenario in low-risk patients, this would have been the case in high-risk patients even if the diff erence in stent costs between the two groups had been larger. This was similarly true with respect to QALYs (data not shown).

No subgroup of patients with large native vessel stenting—ie, low-risk patients—with an acceptable ICER could be identifi ed (table 2), irrespective of whether major adverse cardiac events or QALYs were considered, indicating the robustness of this fi nding. In high-risk patients, use of drug-eluting stents was cost eff ective or even cost saving in all subgroups except patients aged under 65 years (table 2), in which the probability of cost-eff ective use of drug-eluting stents (ie, ICER≤€10 000) was only 0·191. These results were robust, with changes in cost and in event diff erences between patients with drug-eluting stents and those with bare-metal stents having little eff ect in sensitivity analyses (data not shown). With respect to QALYs, a similar pattern of fi ndings was seen, but the diff erence between the age-groups was smaller (data not shown).

DiscussionOur data, based on a comprehensive cost-eff ectiveness analysis of a randomised controlled trial in unselected, real world patients, using the latest stent technology with long-term follow-up, show that drug-eluting stents are not cost eff ective if used in all patients, compared with other accepted medical procedures.26 However, in a subset of high-risk patients with small vessel or bypass-graft stenting, which represented about a third of the study sample, drug-eluting stents were truly cost eff ective or even cost saving. By contrast, it seems almost impossible that drug-eluting stents can become cost eff ective in low-risk patients with large native vessel stenting, even in patients with additional high-risk characteristics. Thus, targeted stent use could be the

–20 000

–10 000

0

10 000

20 000

30 00040 000

50 000

60 000

70 000

80 000

ICER

(€)

Actual costsDES –5%DES –10%DES –15%DES –20%DES –25%DES –30%

–20 000

–10 000

0

10 000

20 000

30 00040 000

50 000

60 000

70 000

80 000

ICER

(€)

–20 000

–10 000

0

10 000

20 000

30 00040 000

50 000

60 000

70 000

80 000

ICER

(€)

Actual costsDES –5%DES –10%DES –15%DES –20%DES –25%DES –30%

DES +15%DES +10%DES +5%Actual costsDES –5%DES –10%DES –15%

0·02 0·04 0·06 0·08 0·10 0·12 0·14 0·16

A

B

C

Mean difference in number of major adverse cardiac events (drug-eluting stents vs bare-metal stents)

Figure 5: Sensitivity analysis of ICERs in relation to alterations in the diff erence of number of major adverse cardiac events and drug-eluting stent costs(A) overall, (B) low-risk patients, and (C) high-risk patients.

Articles


preferred strategy today. Lower prices of drug-eluting stents alone are unlikely to result in such stents being cost eff ective in all patients, even if the problems of late stent thrombosis are solved with new generations of drug-eluting stents.

As in other studies,1,4 diff erences in overall costs between patients with bare-metal stents and those with drug-eluting stents were mainly driven by the higher costs of drug-eluting stents. Thus, ICERs were very sensitive to price reductions of drug-eluting stents, but prices would have to be reduced dramatically for such stents to become cost eff ective. In reality, the diff erence in costs between the two types of stent tended to increase; at the time when stents were implanted in BASKET, prices for bare-metal stents dropped more than those for drug-eluting stents, at least in Europe, making the ICERs even more unfavourable for drug-eluting stents. We found no signifi cant diff er-ences in ICERs between the sirolimus-eluting and paclitaxel-eluting stents, despite the somewhat lower prices of paclitaxel-eluting stents. This lower price was off set by a non-signifi cantly higher event rate in patients with paclitaxel-eluting stents than in those with sirolimus-eluting stents, but the study was neither intended nor powered to assess such a diff erence.

ICERs are also very sensitive to changes in event rates,9 which was also true in the present study sample. However, a substantially larger reduction in the rate of major adverse cardiac events with drug-eluting stents than that observed here would be required to achieve an acceptable cost-eff ectiveness level. Even in combination with some reduction in the prices of such stents, cost-eff ectiveness is unlikely to be achieved in unselected patients, which is in agreement with earlier theoretical considerations20 and a recent registry-based analysis.1

The continuous increase of late stent thrombosis after implantation of drug-eluting stents,10–13 at least up to 3 years,27 could diminish the short-term benefi t of drug-eluting stents even more. Such a situation would reduce the longer-term cost-eff ectiveness of drug-eluting stents further, contrasting earlier projections that anticipated the opposite.9 Thus, our ICER regarding major adverse cardiac events in the overall population was substantially worse at 18 months than the one based on results at 6 months, although this was not true with regard to QALYs.5 In fact, the overall ICER regarding QALYs was in a range that could be seen as cost eff ective, despite the controversial discussion about single thresholds for cost-eff ectiveness.21 However, assessments of QALYs are limited, as quality of life was assessed after 6 and 18 months, but not at the point in time of major events where quality of life would have been lowest, which might explain why QALYs were aff ected by symptom-driven target vessel revascularisation, mostly related to refractory angina, but not by non-fatal myocardial infarction. Therefore, assessment of QALYs by EQ-5D seems to overestimate the eff ects of restenosis and to be somewhat overvalued by missing data, as

shown by multiple imputation in this study. Thus, unselected use of drug-eluting stents is hardly cost eff ective, especially since the need for prolonged treatment with clopidogrel in patients with drug-eluting stents, compared with those with bare-metal stents, as recently suggested by the FDA,28 would further increase costs and negatively aff ect ICERs of drug-eluting stents.

However, these fi ndings do not imply that drug-eluting stents are not cost eff ective at all. Our data indicate that drug-eluting stents were truly cost eff ective in high-risk patients, which remains true even if the prices of drug-eluting stents were higher. Further analyses underlines the robustness of these fi ndings, especially in low-risk patients with large native vessel stenting, where the probability of a cost-eff ective scenario with drug-eluting stents was very low, irrespective of the assumptions made in sensitivity analyses and of further subdivisions in subgroups.

Why are these results less favourable for drug-eluting stent use than are previous fi ndings? Previous trials used fewer stents per patient,3,4 thereby reducing the cost diff erences between those treated with drug-eluting stents and those with bare-metal stents. Additionally, event rates were higher because target vessel revascularisations were driven in part by repeat coronary angiography and not by symptoms only, as in the present study. These two factors are most relevant for the ICER. A recent systematic review of cost-eff ectiveness analyses of drug-eluting stents suggested a bias favouring use of such stents in earlier studies,29 using the same stent platforms for bare-metal and drug-eluting stents. Although this was necessary to assess the eff ect of drug-coating on stents it no longer represents the latest stent technology. By contrast, our fi ndings are based on stent technology available in 2003 when the fi rst patient was enrolled.Contributors All authors participated in the conception and conduct of the study and contributed to the fi nal manuscript. MP, as principal investigator, had the original idea for the study, and organised the trial with CK, and did PCIs with CK. HPB-LR was responsible for the statistical analyses and wrote the manuscript with MP and CK. AB, MJZ, RJ, PTB, and SO participated in enrolling and following the patients.

Confl ict of interest statementWe declare that we have no confl ict of interest.

AcknowledgmentsThis study was supported by the Basel Cardiac Research Foundation, Basel, Switzerland and the University Hospital, Basel, Switzerland.

References1 Ong AT, Daemen J, van Hout BA, et al. Cost-eff ectiveness of the

unrestricted use of sirolimus-eluting stents vs. bare metal stents at 1 and 2-year follow-up: results from the RESEARCH Registry. Eur Heart J 2006; 27: 2996–3003.

2 Elezi S, Dibra A, Folkerts U, et al. Cost analysis from two randomized trials of sirolimus-eluting stents versus paclitaxel-eluting stents in high-risk patients with coronary artery disease. J Am Coll Cardiol 2006; 48: 262–67.

3 Cohen DJ, Bakhai A, Shi C, et al. Cost-eff ectiveness of sirolimus-eluting stents for treatment of complex coronary stenoses: results from the Sirolimus-Eluting Balloon Expandable Stent in the Treatment of Patients With De Novo Native Coronary Artery Lesions (SIRIUS) trial. Circulation 2004; 110: 508–14.

Articles


4 Bakhai A, Stone GW, Mahoney E, et al. Cost eff ectiveness of paclitaxel-eluting stents for patients undergoing percutaneous coronary revascularization: results from the TAXUS-IV Trial. J Am Coll Cardiol 2006; 48: 253–61.

5 Kaiser C, Brunner-La Rocca HP, Buser PT et al. Incremental cost-eff ectiveness of drug-eluting stents compared with a third-generation bare-metal stent in a real-world setting: randomised Basel Stent Kosten Eff ektivitats Trial (BASKET). Lancet 2005; 366: 921–29.

6 Ryan J, Cohen DJ. Are drug-eluting stents cost-eff ective? It depends on whom you ask. Circulation 2006; 114: 1736–43.

7 Daemen J, Serruys PW. Drug-eluting stent update 2007: part II: unsettled issues. Circulation 2007; 116: 961–68.

8 Eisenberg MJ. Drug-eluting stents: the price is not right. Circulation 2006; 114: 1745–54.

9 Cooper LM, Linde-Zwirble W. Cost-eff ectiveness of drug-eluting stents: real-world scrutiny of the BASKET trial of real-world usage. Expert Opin Drug Deliv 2006; 3: 305–09.

10 Pfi sterer M, Brunner-La Rocca HP, Buser PT, et al. Late clinical events after clopidogrel discontinuation may limit the benefi t of drug-eluting stents: an observational study of drug-eluting versus bare-metal stents. J Am Coll Cardiol 2006; 48: 2584–91.

11 Lagerqvist B, James SK, Stenestrand U, Lindback J, Nilsson T, Wallentin L. Long-term outcomes with drug-eluting stents versus bare-metal stents in Sweden. N Engl J Med 2007; 356: 1009–19.

12 Kastrati A, Mehilli J, Pache J, et al. Analysis of 14 trials comparing sirolimus-eluting stents with bare-metal stents. N Engl J Med 2007; 356: 1030–39.

13 Stone GW, Moses JW, Ellis SG, et al. Safety and effi cacy of sirolimus- and paclitaxel-eluting coronary stents. N Engl J Med 2007; 356: 998–1008.

14 Anon. Myocardial infarction redefi ned—a consensus document of The Joint European Society of Cardiology/American College of Cardiology Committee for the redefi nition of myocardial infarction. Eur Heart J 2000; 21: 1502–13.

15 Greiner W, Claes C, Busschbach JJ, Graf von der Schulenburg JM. Validating the EQ-5D with time trade off for the German population. Eur J Health Econ 2005; 6: 124–30.

16 Weinstein MC, Stason WB. Foundations of cost-eff ectiveness analysis for health and medical practices. N Engl J Med 1977; 296: 716–21.

17 Brunner La Rocca HP, Kaiser C, Pfi sterer M, BASKET investigators. Targeted stent use in clinical practice based on evidence from the BAsel Stent Cost Eff ectiveness Trial (BASKET). Eur Heart J 2007; 28: 719–25.

18 Efron B. Bootstrap methods: another look at the jackknife. Ann Statist 1979; 7: 1–26.

19 Polsky D, Glick HA, Willke R, Schulman K. Confi dence intervals for cost-eff ectiveness ratios: a comparison of four methods. Health Econ 1997; 6: 243–52.

20 Greenberg D, Bakhai A, Cohen DJ. Can we aff ord to eliminate restenosis? Can we aff ord not to? J Am Coll Cardiol 2004; 43: 513–18.

21 Eichler HG, Kong SX, Gerth WC, Mavros P, Jonsson B. Use of cost-eff ectiveness analysis in health-care resource allocation decision-making: how are cost-eff ectiveness thresholds expected to emerge? Value Health 2004; 7: 518–28.

22 Little RJ, Rubin DB. Statistical analysis with missing data. 2nd edn. New York: John Wiley & Sons, Inc; 2002.

23 Shuchman M. Trading restenosis for thrombosis? New questions about drug-eluting stents. N Engl J Med 2006; 355: 1949–52.

24 Win HK, Caldera AE, Maresh K, et al. Clinical outcomes and stent thrombosis following off -label use of drug-eluting stents. JAMA 2007; 297: 2001–09.

25 Beohar N, Davidson CJ, Kip KE, et al. Outcomes and complications associated with off -label and untested use of drug-eluting stents. JAMA 2007; 297: 1992–2000.

26 Gold MR, Siegel JE, Russel LB, Weinstein MC. Cost-eff ectiveness in health and medicine. New York: Oxford University Press, 1996.

27 Daemen J, Wenaweser P, Tsuchida K, et al. Early and late coronary stent thrombosis of sirolimus-eluting and paclitaxel-eluting stents in routine clinical practice: data from a large two-institutional cohort study. Lancet 2007; 369: 667–78.

28 Anon. Update to FDA statement on coronary drug-eluting stents. http://www.fda.gov/cdrh/news/010407.html (accessed Oct 4, 2007).

29 Ligthart S, Vlemmix F, Dendukuri N, Brophy JM. The cost-eff ectiveness of drug-eluting stents: a systematic review. CMAJ 2007; 176: 199–205.

Articles


Food additives and hyperactive behaviour in 3-year-old and 8/9-year-old children in the community: a randomised, double-blinded, placebo-controlled trial Donna McCann, Angelina Barrett, Alison Cooper, Debbie Crumpler, Lindy Dalen, Kate Grimshaw, Elizabeth Kitchin, Kris Lok, Lucy Porteous, Emily Prince, Edmund Sonuga-Barke, John O Warner, Jim Stevenson

SummaryBackground We undertook a randomised, double-blinded, placebo-controlled, crossover trial to test whether intake of artifi cial food colour and additives (AFCA) aff ected childhood behaviour.

Methods 153 3-year-old and 144 8/9-year-old children were included in the study. The challenge drink contained sodium benzoate and one of two AFCA mixes (A or B) or a placebo mix. The main outcome measure was a global hyperactivity aggregate (GHA), based on aggregated z-scores of observed behaviours and ratings by teachers and parents, plus, for 8/9-year-old children, a computerised test of attention. This clinical trial is registered with Current Controlled Trials (registration number ISRCTN74481308). Analysis was per protocol.

Findings 16 3-year-old children and 14 8/9-year-old children did not complete the study, for reasons unrelated to childhood behaviour. Mix A had a signifi cantly adverse eff ect compared with placebo in GHA for all 3-year-old children (eff ect size 0⋅20 [95% CI 0⋅01–0⋅39], p=0⋅044) but not mix B versus placebo. This result persisted when analysis was restricted to 3-year-old children who consumed more than 85% of juice and had no missing data (0⋅32 [0⋅05–0⋅60], p=0⋅02). 8/9-year-old children showed a signifi cantly adverse eff ect when given mix A (0⋅12 [0⋅02–0⋅23], p=0⋅023) or mix B (0⋅17 [0⋅07–0⋅28], p=0⋅001) when analysis was restricted to those children consuming at least 85% of drinks with no missing data.

Interpretation Artifi cial colours or a sodium benzoate preservative (or both) in the diet result in increased hyperactivity in 3-year-old and 8/9-year-old children in the general population.

IntroductionArtifi cial food colours and other food additives (AFCA) have long been suggested to aff ect behaviour in children.1 Ben Feingold made his initial claims of the detrimental eff ect of AFCA on childhood behaviour more than 30 years ago.2 The main putative eff ect of AFCA is to produce overactive, impulsive, and inattentive behaviour—ie, hyperactivity—which is a pattern of behaviour that shows substantial individual diff erences in the general population. Children who show this behaviour pattern to a large degree are probably diagnosed with attention-defi cit hyperactivity disorder (ADHD). Despite the failure of early studies3 to identify the range of proposed adverse aff ects, a recent meta-analysis4 of double-blinded, placebo-controlled trials has shown a signifi cant eff ect of AFCA on the behaviour of children with ADHD. The possible benefi t in a reduction in the level of hyperactivity of the general population by the removal of AFCA from the diet is less well estab lished. Evidence from our previous study on the Isle of Wight has suggested adverse eff ects on hyperactivity, meas ured by parental ratings for 3-year-old children on a spe cifi c mix of additives.5 These fi ndings needed replication on 3-year-old children, and to establish whether the eff ects could be seen with a wider range of meas ures of hyperactivity. The present community-based, double-blinded, placebo-controlled food challenge was de signed

to extend the age range studied to include 8/9-year-old children to determine whether the eff ects could also be detected in middle childhood.

MethodsParticipantsFigures 1 and 2 present details of recruitment and participation in the study, for 3-year-old and 8/9-year-old children, respectively. The study sample was drawn from a population of children aged between 3 years and 4 years, 2 months, registered in early-years settings (nurseries, day nurseries, preschool groups, playgroups) and from children aged between 8 and 9 years attend ing schools in Southampton, UK. To ensure that the study sample included children from the full range of socioeconomic backgrounds, schools were recruited based on the number of children receiving free school meals (an index of social disadvantage). The distribu tion of the percentage of children receiving free meals in the schools taking part indicated the proportions for the city as a whole. To further check on how representative the sample was, teachers completed a hyperactivity questionnaire6 for all 3-year-old and 8/9-year-old children.

Parents who returned an expression of interest form were contacted by phone and a home visit arranged. On this visit, a research assistant and the study dietitian,

Lancet 2007; 370: 1560–67

Published OnlineSeptember 6, 2007

DOI:10.1016/S0140-6736(07)61306-3


See Department of Error page 1542

School of Psychology (D McCann PhD, A Barrett BSc,

A Cooper MSc, D Crumpler BSc, L Dalen PhD, E Kitchin BSc, K Lok MSc, L Porteous BSc,

E Prince MSc, Prof E Sonuga-Barke PhD,

Prof J Stevenson PhD) and School of Medicine

(K Grimshaw MSc), Department of Child Health, University of

Southampton, Southampton, UK; and Department of

Paediatrics, Imperial College, London, UK

(Prof J O Warner MD)

Correspondence to:Prof Jim Stevenson, School of

Psychology, Faculty of Medicine, Health and Life Sciences,

University of Southampton, Southampton SO17 1BJ, UK

[email protected]

Articles


provided full information about the study and its dietary implications, and written informed consent was obtained. The study dietitian also obtained a report based on 24-h recall by the parent of the child’s pretrial diet, which allowed an assessment of baseline levels of the number of foods containing additives consumed by the child in the previous 24 h. The study was approved by the local research ethics committee (reference no 04/Q1702/61) and written informed consent was obtained from parents. Participating early-years settings received £250 and each school £500 as a contribution towards school funds for the benefi t of all children.

Study design and challenge protocolsThe study design and challenge protocols for both ages were similar. Children were entered into this study with a within-subject crossover between two active mixes (A and B) and a placebo drink.

The two active mixes diff ered both in the quantities of additives and the specifi c additives included. Mix A was similar to the active challenge used in the Isle of Wight study,5 and mix B was selected to indicate the current average daily consumption of food additives by 3-year-old and 8/9-year-old children in the UK.7 Both mixes included sodium benzoate, which had been included in the challenge on the Isle of Wight study and in previous studies.8,9

Mix A for 3-year-old children included 20 mg of artifi cial food colourings (5 mg sunset yellow [E110], 2⋅5 mg carmoisine [E122], 7⋅5 mg tartrazine [E102], and 5 mg ponceau 4R [E124, Forrester Wood, Oldham, UK]) and 45 mg of sodium benzoate [E211, Sigma Aldridge, Gillingham, UK]). Active mix B included 30 mg of artifi cial food colourings (7⋅5 mg sunset yellow, 7⋅5 mg carmoisine, 7⋅5 mg quinoline yellow [E110], and 7⋅5 mg allura red AC [E129]) and 45 mg of sodium benzoate.

Mix A amounts for 8/9-year-old children were multiplied by 1⋅25 to account for the increased amount of food consumed by children at this age. Therefore, mix

A included 24⋅98 mg of artifi cial food colourings (6⋅25 mg sunset yellow, 3⋅12 mg carmoisine, 9⋅36 mg tartrazine, and 6⋅25 mg ponceau 4R) and 45 mg of sodium benzoate. Active mix B included 62⋅4 mg of artifi cial food colourings (15⋅6 mg sunset yellow, 15⋅6 mg

Parents of 898 children sent invitation to participate

556 did notrespond

133 gave negativeresponse

209 gave positiveresponse

3 excluded2 did not meet

age criteria1 not registered

for at least twosessions perweek at earlyyears setting

25 refused toconsent

28 consented,then refused

153 consented

137 completedstudy

16 did not completestudy

Figure 1: Enlistment of 3-year-old participants

Parents of 633 children sent invitation to participate

377 did not respond

96 gave negative response

160 gave positiveresponse

5 excluded 2 food allergy2 allergy to

blackcurrant juice1 on holiday

during challengeperiod

9 refused to consent

2 consented, then refused

144 consented

130 completedstudy

14 did not completestudy

Figure 2: Enlistment of 8/9-year-old participants

3-year-old children in total sample analysis (n=153)

8/9-year-old children in total sample analysis (n=144)

Racial background

White 126 (82%) 130 (90%)

Other 15 (10%) 14 (10%)

Missing data 12 (8%) ..

Marital status

Married/partner 127 (83%) 115 (80%)

Single/separated/divorced/widowed 26 (17%) 29 (20%)

NSSC (father)

Higher occupations 34 (22%) 37 (26%)

Intermediate occupations 26 (17%) 18 (13%)

Lower occupations 51 (33%) 44 (31%)

Never worked/long-term unemployed 4 (3%) 7 (5%)

No father present 26 (17%) 29 (20%)

Missing data 12 (8%) 9 (6%)

NSSC (mother)

Higher occupations 31 (20%) 38 (26%)

Intermediate occupations 18 (12%) 26 (18%)

Lower occupations 66 (43%) 32 (22%)

Never worked/long-term unemployed 26 (17%) 32 (22%)

Missing data 12 (8%) 16 (11%)

Mother’s education

School attendance up to age 16 years (no qualifi cations or certifi cates below “A” level)

53 (35%) 60 (42%)

“A” levels 61 (40%) 42 (29%)

University degree/postgraduate qualifi cation 27 (18%) 27 (19%)

Missing data 12 (8%) 15 (10%)

Data are n (%). NSSC=national statistics social class.20 Higher=managerial and professional. Intermediate=self-employed. Lower=routine work. “A” levels=pre-university, school examinations in the UK.

Table 1: Characteristics of parents of children enlisted in study

Articles


carmoisine, 15⋅6 mg quinoline yellow, and 15⋅6 mg allura red AC) and 45 mg of sodium benzoate.

Doses for mixes A and B for 3-year-old children were roughly the same as the amount of food colouring in two 56-g bags of sweets. For 8/9-year-old children, the dose for mix A was equal to about two bags of sweets a day and for mix B about four bags of sweets a day.

After a week on their typical diet (week 0: baseline diet), the artifi cial colours to be used in the challenges and sodium benzoate were withdrawn from their diet for 6 weeks. Over this period when challenge with active or placebo drinks were given, additive withdrawal continued (week 1: withdrawal period but receiving placebo; weeks 2, 4, and 6: challenge with randomisation to two active periods and one placebo period; weeks 3 and 5: washout continuing on placebo). During this period, 3-year-old children received the challenge and washout-placebo

drinks on a weekly basis and consumed mixed fruit juices (placebo or active) at home (300 mL/day for 3-year-old children, 625 mL/day for 8/9-year-old children), provided in identical sealed bottles. At the beginning of the study, children were assigned by the study administrator by a random-number generator to receive one of six possible sequences of placebo, active mix A, or active mix B challenges across weeks 2, 4, and 6.

A masked testing by two independent panels of 20 young adults showed that the active and placebo juice drinks could not be diff erentiated. When asked if the mix contained additive, 16 (40%), 21 (52%), and 26 (65%) adults responded positively for mix A, mix B, and placebo, respectively. We recorded no signifi cant diff erences between these proportions (Friedman test, χ²=4⋅412, df=2). Therefore, no reliable diff erences were seen between the look and taste of the drinks. The only diff erence in the composition of the placebo and active mixes was the presence of the AFCA in the active mix with some variation in the proportions of the fruit juices to ensure matching colour and taste for the placebo and active drinks. The child’s family and the research team were masked to the challenge allocation. The study administrator assigned the challenge sequence and assisted in the preparation and packaging of juice drinks that were then delivered by the masked research team to homes every week, when questionnaires and other forms were obtained and dispensed. Parents completed a daily diary of juice consumption and compliance with the diet over the study period. Parents also recorded a mistake event when a child consumed a portion of food containing the artifi cial colours or sodium benzoate. Any bottles containing juice not consumed in the previous week were obtained, returned to the study offi ce, and measured to help validate, if possible, parental reports of juice consumption by children.

Global hyperactivity aggregate (GHA)Three measures of behaviour were used to calculate GHA for 3-year-old children, with an additional measure for 8/9-year-old children. First, the abbreviated ADHD rating scale IV (teacher version)6 was used. A total score was obtained for ten of the 18 items (inattentive=5, hyperactive=5) in this questionnaire, which was completed to describe the frequency of the specifi c behaviours displayed over the past week, for every week of the study. Parent behaviour was the second measure, by use of the abbreviated Weiss-Werry-Peters (WWP) hyperactivity scale,10 which has been used in several studies to assess hyperactivity.11,12 Interparent agreement is good for ratings of childhood behaviour (r=0⋅82).13 Parents rated their child’s behaviour during the previous week for seven items previously used (switching activities; interrupting or talking too much; wriggling; fi ddling with objects or own body; restless; always on the go; concentration),4 from which we obtained a total score. For 8/9-year-old children, we used an abbreviated ADHD rating scale IV (parent

Mix A Mix B Placebo

n mean (SD) n mean (SD) n mean (SD)

3-year-old children

Entire sample (n=140) 131 −0·11 (1·03) 134 −0·14 (1·03) 129 −0·32 (1·11)

≥85% consumption (n=130) 104 −0·11 (1·03) 108 −0·15 (1·07) 99 −0·39 (1·07)

Complete case (n=73) 73 −0·14 (1·04) 73 −0·26 (1·05) 73 −0·44 (0·98)

8/9-year-old children

Whole sample (n=136) 132 0·25 (0·97) 133 0·33 (1·10) 127 0·19 (1·03)

≥85% consumption (n=119) 104 0·26 (0·93) 112 0·32 (1·09) 103 0·19 (1·04)

Complete case (n=91) 91 0·27 (0·92) 91 0·35 (1·08) 91 0·19 (1·06)

Table 2: Mean GHA scores for 3-year-old and 8/9-year-old children by challenge type

Entire sample (n=140) Group with ≥85% consumption (n=130)

Complete case group (n=73)

Model 1

Intercept –0·31 (–0·49 to –0·13)* –0·33 (–0·53 to –0·13)† –0·44 (–0·68 to –0·21)†

Challenge type

Mix A vs placebo 0·20 (0·01 to 0·40)‡ 0·24 (0·02 to 0·47)‡ 0·31 (0·04 to 0·58)‡

Mix B vs placebo 0·16 (–0·04 to 0·35) 0·16 (–0·07 to 0·38) 0·19 (–0·08 to 0·46)

Model 2

Intercept –0·54 (–0·89 to –0·18)* –0·51 (–0·92 to –0·11) –0·58 (–1·08 to –0·09)‡

Challenge type

Mix A vs placebo 0·20 (0·01 to 0·39)‡ 0·28 (0·05 to 0·51)‡ 0·32 (0·05 to 0·60)‡

Mix B vs placebo 0·17 (–0·03 to 0·36) 0·19 (–0·04 to 0·41) 0·21 (–0·06 to 0·48)

Week of study

Week 2 vs week 6 0·15 (–0·05 to 0·34) 0·15 (–0·08 to 0·38) 0·19 (–0·08 to 0·46)

Week 4 vs week 6 0·17 (–0·03 to 0·36) 0·23 (0·00 to 0·46)‡ 0·19 (–0·09 to 0·46)

Sex 0·18 (–0·10 to 0·45) 0·22 (–0·07 to 0·51) 0·05 (–0·31 to 0·40)

Baseline GHA score 0·46 (0·26 to 0·66)† 0·54 (0·31 to 0·76)† 0·36 (0·06 to 0·66)‡

Pretrial diet 0·08 (–0·02 to 0·19) 0·07 (–0·04 to 0·18) 0·09 (–0·04 to 0·23)

Maternal education level –0·01 (–0·29 to 0·28) –0·04 (–0·34 to 0·26) –0·03 (–0·41 to 0·35)

Maternal social class 0·15 (–0·44 to 0·13) –0·23 (–0·53 to 0·08) –0·21 (–0·58 to 0·16)

Data are estimate (95% CI). *p<0·01. †p<0·001. ‡p<0·05. Complete case=≥85% consumption and no missing data. Model 1=challenge type alone. Model 2=challenge type with additional factors controlled.

Table 3: General GHA estimates in linear mixed models during challenge period for 3-year-old children

Articles


version)14 to measure parent behaviour, whereby a ten-item questionnaire was completed by parents every week.

A third measure was the classroom observation code,15 which assesses the occurrence of 12 mutually exclusive behaviours during structured didactic teaching and during periods of independent work under teacher supervision. To develop this measure, the behaviours had been selected to indicate components of ADHD that are shown in the classroom. After observers (psychology graduates) were given extensive training, the inter-rater reliability of the classroom observation code, tested before and during the study, exceeded 0⋅87. Children were observed for 24 min every week (three observation sessions of 8 min each) and a total weekly mean score was derived from the total score over every session. The code was slightly modifi ed for 3-year-old children, since preschool children in the UK are not usually given structured or didactic teaching sessions and tend to engage in activities rather than in tasks. Observation took place over a range of activities and the off -task category in the code was scored when the child switched activities.

A fourth measure for 8/9-year-old children was the Conners continuous performance test II (CPTII),16 a test using visual stimuli of 14-min duration and is widely used to assess attention and the response inhibition component of executive control. We used four scores (SE of reaction time, % of commission errors, d́ [discriminability index], and β) to derive a weekly aggregate score. This subset of indicators from the CPTII has been shown to be highly correlated with the ADHD rating scale.17

The GHA was developed to measure individual diff erences in hyperactivity using diff erent sources (teacher, parent ratings, direct observation, and a computerised test) and covering the components of hyperactivity (overactivity, impulsivity, and inattention). Weekly scores for every child were standardised to time 0 at baseline (T0). Weekly standardised (z) aggregate scores were calculated as: (score minus mean score at T0) divided by SD at T0. The GHA was an equally weighted aggregate of the weekly z-scores, and calculated only when at least two (or three for 8/9-year-old children) of these behaviour scores were present for any week (one of which being for the classroom observation code) and averaged across the number of available scores. A high GHA indicates more hyperactivity.

Statistical analysisAlthough the study designs for the two age groups were similar, the diff erence in composition of the GHA, and in the dose of AFCA used, meant that data from the two studies could not be analysed jointly. Therefore, we treated the studies as parallel but independent.

Linear mixed-model methods18,19 in SPSS (version 14.0) were used to analyse data. Several possible covariates were thought to be signifi cantly related to GHA (eg, sex). Two models were tested separately for each age for the eff ects on GHA in challenge weeks. Model 1 used the

challenge type alone as a fi xed eff ect testing for mix A against placebo and mix B against placebo. In model 2, in addition to challenge type, the eff ects of the following factors were adjusted for: week during study, sex, GHA in baseline week, number of additives in pretrial diet, maternal educational level, and social class. A compound symmetry covariance matrix provided the best-fi t model for 3-year-old children and an unstructured covariance matrix for 8/9-year-old children. The study was powered to detect diff erences between the active and placebo periods and, accordingly in each case, the eff ects of mix A and mix B were compared with that of placebo. We anticipated that the additional controls on placebo eff ects would result in an eff ect size smaller than that achieved in the Isle of Wight study.5 A sample of 80 children had

Entire sample (n=153)

Group with ≥85% consumption (n=132)

–0·75

–0·50

–0·25

Estim

ated

mar

gina

l mea

ns

Difference in estimated m

eans

Estim

ated

mar

gina

l mea

ns


eans

Estim

ated

mar

gina

l mea

ns


eans

0

0·25

0·50

0·75

Mix AMix B

Placebo

Mix A vs

placeboMix B vs

placebo

–0·75

–0·50

–0·25

0

0

0·25

0·50

0·75

–0·75

–0·50

–0·25

0·25

0·50

0·75

–0·75

–0·50

–0·25

0

0

0·25

0·50

0·75

–0·75

–0·50

–0·25

0·25

0·50

0·75

0

–0·75

–0·50

–0·25

0·25

0·50

0·75


Figure 3: Estimated marginal means by challenge type and diff erence in estimated means in GHA under model 2 for 3-year-old children Bars=95% CI. Dashed line=zero diff erence between mean GHA under active mix and mean GHA under placebo.

Articles


80% power at α=0·05 to identify an eff ect size of 0·32—ie, the magnitude of the diff erence in GHA mean score changes (SD). This value was lower than that achieved in the previous study (0·51). We were uncertain about the number of children and families who would comply with the demands of a 7-week study, so we set a target of 120 children to reduce the eff ect of attrition on power, which was eventually exceeded in both age groups.

The analyses were replicated for the full sample, a high consumption group (≥85% consumption of drinks in any challenge week), and a complete case group (≥85% consumption in all challenge weeks and no missing GHA). The high consumption and complete case groups were included to determine whether non-compliance and the method of handling missing data aff ected the pattern of results. Analysis was per protocol.

This clinical trial is registered with Current Controlled Trials (registration number ISRCTN74481308).

Role of the funding sourceThe sponsor of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had fi nal responsibility for the decision to submit for publication.

ResultsOf 153 children (mean age 43·5 months [SD 4·5]) enlisted, 79 were boys (43·5 months [4·6]) and 74 were girls (43·4 months [4·3]). Table 1 provides parents’ characteristics for the entire sample. We saw no signifi cant diff erences in

these background characteristics between groups assigned to receive the challenge drinks in diff erent orders during each of the six periods. The proportion of children in each of fi ve quintile ranges on the teachers questionnaire6 was not signifi cantly diff erent for the sample or for the total population (n=898, χ² [4]=1·60).

16 (10%) 3-year-old children failed to complete the study. Age, sex, and marital status of the parents had no eff ect on study completion and children were no more likely to drop out during active challenge weeks than placebo. In only one case was this failure to complete related to problems with the child’s behaviour. Of those children lost to the study, 12 had a mean of 41% consumption in the fi rst challenge week and data were missing for four children. 128 (93%) of the 137 children who completed the study consumed more than two-thirds of all drinks, of which 103 (80%) consumed 85% or more (ie, at least six of seven daily drinks per week). Only one of the remaining nine children drank less than 50% of placebo and active drinks during the study period. The occurrence of dietary infractions or mistakes by 3-year-old children was low (0=33% of children, 1–2=31%, 3–4=18·3%, >4=17%). Rate of infractions did not diff er during active and placebo weeks.

117 (76%) 3-year-old children had complete GHA data over active and placebo weeks, 19 (12%) had two GHA scores, and one had one score. Of children who left the study, 12 provided one score, and four had missing data.

Table 2 shows the mean GHA scores under each of the three challenge types. For the challenge periods in weeks 2, 4, and 6, preliminary analyses had shown no eff ect of the type of challenge in the previous challenge period on the GHA; therefore, the washout periods had eradicated carry-over eff ects. Table 3 shows the results of the linear mixed-model analyses for 3-year-old children. For model 1 (the unadjusted eff ects of challenge type), all three samples had signifi cant adverse eff ects of mix A on GHA compared with placebo. The higher GHA scores for mix B were not signifi cantly greater than for placebo. Under model 2, with the eff ects of other factors controlled, the eff ect of mix A was signifi cant for the entire sample (table 2, p=0·044), by contrast with that of mix B (p=0·093). When the analyses are restricted to those children with at least 85% juice consumption, the adverse eff ect of mix A on behaviour was still signifi cant (p=0·016), but non-signifi cant for mix B (p=0·098). The complete case groups showed the same pattern of results (mix A, p=0·020; mix B, p=0·131). Figure 3 shows estimated marginal mean scores after adjustment for factors in model 2.

Of 144 8/9-year-old children (mean age 106·3 months [SD 5·9]) enlisted to the study, 75 were boys (106·4 months [6·1]) and 69 were girls (106·1 months [5·8]). Table 1 provides parents’ characteristics for the entire sample. We recorded no signifi cant diff erences in these background characteristics between groups of children

Enitre sample (n=136) Group with ≥85% consumption (n=119)


Model 1

Intercept 0·16 (–0·01 to 0·34) 0·09 (–0·09 to 0·27) 0·11 (–0·10 to 0·32)

Challenge type

Mix A vs placebo 0·08 (–0·02 to 0·18) 0·12 (0·02 to 0·23)* 0·14 (0·03 to 0·24)*

Mix B vs placebo 0·12 (0·03 to 0·22)* 0·15 (0·05 to 0·25)† 0·17 (0·06 to 0·28)†

Model 2

Intercept 0·02 (–0·22 to 0·26) 0·14 (–0·08 to 0·37) 0·14 (–0·12 to 0·39)

Challenge type

Mix A vs placebo 0·08 (–0·02 to 0·17) 0·09 (–0·01 to 0·19) 0·12 (0·02 to 0·23)*

Mix B vs placebo 0·12 (0·03 to 0·22)* 0·15 (0·05 to 0·25)† 0·17 (0·07 to 0·28)†

Week of study

Week 2 vs week 6 –0·11 (–0·21 to 0·00)* –0·19 (–0·29 to –0·08)† –0·20 (–0·32 to –0·09) †

Week 4 vs week 6 0·06 (–0·03 to 0·14) 0·04 (–0·06 to 0·13) 0·03 (–0·07 to 0·13)

Sex 0·16 (–0·03 to 0·35) 0·08 (–0·10 to 0·26) 0·11 (–0·09 to 0·31)

Baseline GHA score 0·78 (0·69 to 0·88)‡ 0·79 (0·71 to 0·88)‡ 0·79 (0·70 to 0·89)‡

Pretrial diet 0·04 (–0·02 to 0·10) 0·03 (–0·03 to 0·09) 0·02 (–0·05 to 0·09)

Maternal education level –0·02 (0·20 to 0·16) –0·02 (–0·19 to 0·15) 0·01 (–0·18 to 0·21)

Maternal social class 0·04 (–0·14 to 0·22) –0·03 (–0·20 to 0·14) –0·06 (–0·25 to 0·13)

Data are estimate (95% CI). *p<0·05. †p<0·01. ‡p<0·001. Complete case=≥85% consumption and no missing data. Model 1=challenge type alone. Model 2=challenge type with additional factors controlled.

Table 4: General GHA estimates in linear mixed models during challenge period for 8/9-year-old children

Articles


assigned to receive the challenge drinks in diff erent orders over each of the six periods. The proportion of children in each of fi ve quintile ranges on the teachers questionnaire6 was not signifi cantly diff erent for the sample and for the total population (n=663, χ² [4]=5·05).

14 (10%) 8/9-year-old children failed to complete the study; reasons for failure were unrelated to behavioural problems. Age, sex, and marital status of the parents had no eff ect on study completion and children were no more likely to drop out during active challenge weeks than placebo. Of those children lost to the study, two had a mean of 93% consumption in the fi rst challenge week and data were missing for 12 children. Of the remaining children who completed the study, 98 (75%) consumed

85% or more of the drinks over the challenge weeks (at least six of seven daily drinks per week). Only seven of the remaining 28 children drank less than 50% of placebo and active drinks over the study period. The occurrence of dietary infractions or mistakes by 8/9-year-old children during the study period was low (0=25% of children, 1–2=41%, 3–4=19%, >4=16%). Rate of infractions did not diff er during active and placebo weeks.

Of 125 8/9-year-old children, 114 (87%) had full GHA data during active and placebo weeks, six (4%) had two GHA scores, and fi ve (3%) had one score; eight (6%) had no GHA scores. Table 2 also shows mean GHA scores for 8/9-year-old children for the entire sample, the group with at least 85% consumption, and the complete case sample. For the challenge periods in weeks 2, 4, and 6, preliminary analyses showed no eff ect of the type of challenge in the previous challenge period on the GHA, showing that the washout periods had eradicated carry-over eff ects. For model 1 (the unadjusted eff ects of challenge type) the eff ects of mix A and mix B were signifi cantly greater than that of placebo, with the exception of the entire sample in which the eff ects of mix A versus placebo fail to reach signifi cance (table 4). Under model 2, in which the eff ects of other factors were controlled, the eff ect of mix A for the entire sample was not signifi cant (p=0·123) but mix B did have a signifi cantly adverse eff ect compared with placebo (p=0·012). When the analyses are restricted to those children who consumed at least 85% juice, the adverse eff ect of mix A on behaviour remained non-signifi cant (p=0·066) but was signifi cant for mix B (p=0·003). The complete case groups showed signifi cantly higher GHA scores than placebo for mix A (p=0·023) and mix B (p=0·001). Figure 4 shows the estimated marginal means score after adjustment for factors in model 2.

DiscussionIn this community-based, double-blinded, placebo-controlled food challenge, we tested the eff ects of artifi cial food additives on children’s behaviour and have shown that a mix of additives commonly found in children’s food increases the mean level of hyperactivity in children aged 3 years and 8/9 years. Our complete case data has indicated that the eff ect sizes, in terms of the diff erence between the GHA under active mix and placebo challenges, were very similar for mix B in 3-year-old and 8/9-year-old children. For mix A, the eff ect for 3-year-old children was greater than for 8/9-year-old children. The eff ects for mix B were not signifi cant for 3-year-old children because there was greater variability in the response to the active challenges than placebo in this age group. Thus, we recorded substantial individual diff erences in the response of children to the additives. For both age groups, no signifi cant eff ect of social and demographic factors was seen on the initial level of GHA or in the moderation of the challenge eff ects. The moderating

Entire sample (n=136)

Group with ≥85% consumption (n=119)


eansEstim

ated

mar

gina

l mea

ns

Mix AMix B

Placebo

Mix A vs

placeboMix B vs

placebo

Estim

ated

mar

gina

l mea

nsEs

timat

ed m

argi

nal m

eans

0·2

–0·1

0

0·1

0·3

0·4

0·5

0·2

–0·1

0

0·1

0·3

0·4

0·5

0·2

–0·1

0

0·1

0·3

0·4

0·5

0·2

–0·1

0

0·1

0·3

0·4

0·5

0·2

–0·1

0

0·1

0·3

0·4

0·5

0·2

–0·1

0

0·1

0·3

0·4

0·5


eansDifference in estim

ated means


Figure 4: Estimated marginal means by challenge type and diff erence in estimated means in GHA under model 2 for 8/9-year-old childrenBars=95% CI. Dashed line=zero diff erence between mean GHA under active mix and mean GHA under placebo.

Articles


eff ects of genotype on the child’s behaviour response to AFCA are examined in a separate paper (unpublished data).

The eff ect sizes reported in this study are similar to those calculated in the meta-analysis by Schab and Trinh.4 They estimated the eff ects of AFCA on hyperactivity to be 0·283 (95% CI 0·079–0·488), falling to 0·210 (0·007–0·414) when the smallest and lowest quality trials were excluded. It should be noted that this meta-analysis included studies of hyperactivity in clinical samples, whereas the present study was done on children in the general population with the full range of degrees of hyperactivity. These eff ect sizes recorded by Schab and Trinh are smaller than those reported for stimulant treatment for ADHD in children, for which one meta-analysis21 reported a range of eff ect sizes from 0·78 (0·64–0·91) by teacher report to 0·54 (0·40–0·67) by parent report. We report eff ect sizes that average at about 0·18. Children with ADHD are generally about 2 SD higher on hyperactivity measures than those without the disorder,22 thus an eff ect size of 0·2 is about 10% of the behavioural diff erence between them.

This study provides evidence of deleterious eff ects of AFCA on children’s behaviour with data from a whole population sample, using a combination of robust objective measures with strong ecological validity, based partly on observations in the classroom and ratings of behaviour made independently by teachers and by parents in the diff erent context of the home and applying double-blinded challenges with quantities of additives equal to typical dietary intakes. It also replicates the eff ects of mix A previously reported on a large sample (n=277) of 3-year-old children,5 although signifi cant eff ects were only seen with parental ratings in that study.

The specifi c deleterious compounds in the mix cannot be determined for the present study and need to be examined in subsequent studies. The eff ect of artifi cial colours needs to be diff erentiated from the eff ects of preservatives in a 2×2 design. Further investigation would also need to establish whether the age-related diff erence seen in the present study can be replicated—ie, the eff ects of mix A being greater for 3-year-old children than for 8/9-year-old children. We examined the eff ects of additives on changes in behaviour during an extended period in a community-based, double-blinded, placebo-controlled food challenge. A weakness in this approach is the lack of control over when the challenges are ingested in relation to the timing of measures of hyperactivity. This study design also needs extensive resources to obtain multisource and multicontext measures of hyperactivity. We have completed a pilot study showing that changes in hyperactivity in response to food additives can be produced within about 1 h. Therefore, future studies could use more feasible acute double-blinded challenges undertaken in more controlled settings.

The present fi ndings, in combination with the replicated evidence for the AFCA eff ects on the behaviour of 3-year-old children, lend strong support for the case that food additives exacerbate hyperactive behaviours (inattention, impulsivity, and overactivity) in children at least up to middle childhood. Increased hyperactivity is associated with the development of educational diffi culties, especially in relation to reading, and therefore these adverse eff ects could aff ect the child’s ability to benefi t from the experience of schooling.23 These fi ndings show that adverse eff ects are not just seen in children with extreme hyperactivity (ie, ADHD),4 but can also be seen in the general population and across the range of severities of hyperactivity. Our results are consistent with those from previous studies and extend the fi ndings to show signifi cant eff ects in the general population. The eff ects are shown after a rigorous control of placebo eff ects and for children with the full range of levels of hyperactivity.

We have found an adverse eff ect of food additives on the hyperactive behaviour of 3-year-old and 8/9-year-old children. Although the use of artifi cial colouring in food manufacture might seem superfl uous, the same cannot be said for sodium benzoate, which has an important preservative function. The implications of these results for the regulation of food additive use could be substantial.ContributorsJS, JOW, and ES-B participated in the conception and design of the study. The Food Standards Agency assisted with the design of the study. DMC directed the execution of the study. AB, AC, DC, LD, EK, LP, and EP undertook assessments of the children and helped to develop the observational methods employed in the study. KG supervised and KL executed the nutritional aspects of the study in relation to the preparation of suitable challenge drinks and advice on diet for parents. DMC and JS analysed the data and wrote the manuscript with input from all the authors.

Confl ict of interest statementWe declare that we have no confl ict of interest.

Acknowledgments We thank the children, families, and teachers in the participating schools and early years settings in the Southampton area for their help and assistance with the study; and Catherine Varcoe-Baylis and Jenny Scoles and the local steering committee for their assistance, especially Ulrike Munford and the representatives from Southampton City Council Children’s Services and Learning and the Southampton Early Years Development and Childcare Partnership; and the Food Standards Agency for their signifi cant advice and input into the study. This study received funding from the Food Standards Agency (grant T07040).

References1 Overmeyer S, Taylor E, Annotation: principles of treatment for

hyperkinetic disorder: practice approaches for the UK. J Child Psychol Psychiatry 1999; 40: 1147–57.

2 Feingold BF. Hyperkinesis and learning disabilities linked to artifi cial food fl avors and colours. Am J Nurs 1975; 75: 797–803.

3 Editorial. NIH consensus development conference: defi ned diets and childhood hyperactivity. Clin Pediatr 1982; 21: 627–30.

4 Schab DW, Trinh NT. Do artifi cial food colours promote hyperactivity in children with hyperactive syndromes? A meta-analysis of double-blind placebo-controlled trials. J Dev Behav Pediatr 2004; 25: 423–34.

5 Bateman B, Warner JO, Hutchinson E, et al. The eff ects of a double blind, placebo controlled, artifi cial food colourings and benzoate preservative challenge on hyperactivity in a general population sample of preschool children. Arch Dis Child 2004; 89: 506–11.

Articles


6 DuPaul GJ, Power TJ, Anastopoulos AD, Reid R, McGoey K, Ikeda M. Teacher ratings of ADHD symptoms: Factor structure and normative data. Psychol Assess 1997; 9: 436–44.

7 Gregory JR, Collins EDI Davies PSW, Hughes JM, Clarke PC. National Diet and Nutrition Survey: children aged 1·5 to 4·5 years. Vol 1: Report of the Diet and Nutrition Survey. London: HM Stationery Offi ce, 1995.

8 Egger J, Graham PJ, Carter CM, Gumley D, Soothill JF. Controlled trial of oligoantigenic treatment in the hyperkinetic syndrome. Lancet 1985; 325: 540–45.

9 Carter CM, Urbanowicz M, Hemsley R, et al. Eff ects of a few food diet in attention-defi cit disorder. Arch Dis Child 1993; 69: 564–68.

10 Routh D. Hyperactivity. In: Magrab P, ed. Psychological management of pediatric problems. Baltimore: University Park Press, 1978: 3–8.

11 Thompson MJJ, Stevenson J, Sonuga-Barke E, et al. The mental health of preschool children and their mothers in a mixed urban/rural population. Prevalence and ecological factors. Br J Psychiatry 1996; 168: 16–20.

12 Hayward C, Killen J, Kraemer H, et al. Linking self-reported childhood behavioural inhibition to adolescent social phobia. J Am Acad Child Adolesc Psychiatry 1998; 37: 1308–16.

13 Mash EJ, Johnston C. Parental perceptions of child behaviour problems, parenting self-esteem and mother’s reported stress in younger and older hyperactive and normal children. J Consult Clin Psychol 1983; 51: 86–99.

14 DuPaul GJ, Power TJ, Anastopoulos AD, Reid R. AD/HD rating scale IV: checklists, norms and clinical interpretation. New York: Guilford Press, 1998.

15 Abikoff H, Gittleman R. Classroom observation code—a modifi cation of the stony-brook code. Psychopharmacol Bull 1985; 21: 901–09.

16 Conners CK. The Conners continuous performance test. Toronto, ON, Canada: Multi-Health Systems, 1994.

17 Epstein N, Erkanli A, Conners CK, Klaric J, Costello JE, Angold A. Relations between continuous performance test performance measures and ADHD behaviours. J Abnorm Child Psychol 2003; 31: 543–54.

18 Gueorguieva R, Krystal JH. Move over ANOVA: Progress in analysing repeated-measures data and its refl ection in papers published in the Archives of General Psychiatry. Arch Gen Psychiatry 2004; 61: 310–17.

19 Mallinckrodt CH, Watkin JG, Molenburghs G, et al. Choice of the primary analysis in longitudinal clinical trials. Pharm Stat 2004; 3: 161–69.

20 Offi ce for National Statistics. Standard occupational classifi cation. London: Stationery Offi ce, 2000.

21 Schachter HA, King J, Langford S, Moher D. How effi cacious and safe is short-acting methylphenidate for the treatment of attention-defi cit disorder in children and adolescents? A meta-analysis. Can Med Assoc J 2001; 165: 1475–88.

22 Swanson JM, Sergeant J, Taylor E, Sonuga-Barke EJS, Jensen PS, Cantwell DP. Attention-defi cit hyperactivity disorder and hyperkinetic disorder. Lancet 1998; 351: 429–33.

23 McGee R, Prior M, Williams S, Smart D, Sanson A. The long-term signifi cance of teacher-rated hyperactivity and reading ability in childhood: fi ndings from two longitudinal studies. J Child Psychol Psychiatry 2002; 43: 1004–17.


Clinical Picture

Lancet 2007; 370: 1568

Department of Medicine, Division of Gastroenterology

and Hepatology, Charité Medical Centre—Virchow

Hospital, Medical School of the Humboldt University of Berlin,

Berlin, Germany (D C Baumgart MD,

A Fischer MD)

Correspondence to:Dr Daniel C Baumgart,

Department of Medicine, Division of Gastroenterology and

Hepatology, Charité Medical Centre—Virchow Hospital,

Medical School of the Humboldt University of Berlin,

Augustenburger Platz 1, 13353 Berlin, Germany

[email protected]

Virchow’s nodeDaniel C Baumgart, Andreas Fischer

A 70-year-old man presented to his primary care physician with postprandial epigastric pain. He was prescribed antacids, which were ineff ective; his weight then decreased by 11 kg in 3 months. Oesophagogastro duoden-oscopy showed a mass in the antrum; histo pathological analysis of a biopsy sample established that the mass was a signet-ring adenocarcinoma. Even before the biopsy was analysed, examination of the neck supported the diagnosis of cancer: the left supraclavicular lymph node was fi rm and enlarged (fi gure). The left supraclavicular lymph node is near the junction of the thoracic duct and the left subclavian vein, where the lymph from much of the body drains into the systemic circulation. An enlarged node can be the fi rst sign of gastric cancer—and is often called Virchow’s node, because the association between left supraclavicular lymphadenopathy and gastric cancer was fi rst described by the German pathologist Rudolf Virchow (1821–1902). Virchow lectured and practised at the Charité Medical School of the Humboldt University of Berlin, where he discovered several facts fundamental to modern medicine.

Figure: Virchow’s node

Series


Who Counts? 1

A scandal of invisibility: making everyone count by counting everyonePhilip W Setel, Sarah B Macfarlane, Simon Szreter, Lene Mikkelsen, Prabhat Jha, Susan Stout, Carla AbouZahr, on behalf of the Monitoring of Vital Events (MoVE) writing group*

Most people in Africa and Asia are born and die without leaving a trace in any legal record or offi cial statistic. Absence of reliable data for births, deaths, and causes of death are at the root of this scandal of invisibility, which renders most of the world’s poor as unseen, uncountable, and hence uncounted. This situation has arisen because, in some countries, civil registration systems that log crucial statistics have stagnated over the past 30 years. Net of debt relief, offi cial development assistance reached US$80 billion in 2004. Yet because of the weakness in recording vital statistics, we have little authoritative evidence that these funds have their desired eff ects on either mortality or poverty reduction. Sound recording of vital statistics and cause of death data are public goods that enable progress towards Millennium Development Goals and other development objectives that need to be measured, not only modelled. Vital statistics are most eff ectively generated by comprehensive civil registration. Civil registration has a dual function, both statistical and legal; it also helps with economic development. 30 years of stagnation will not be overcome quickly, although new eff orts to develop national statistical capacities off er a unique opportunity to refocus attention on civil registration. Now is the time to make the long-term goal of comprehensive civil registration in developing countries the expectation rather than the exception. The international health community can assist by sharing information and methods to ensure both the quality of vital statistics and cause of death data, and the appropriate use of complementary and interim registration systems and sources of such data. The continued cost of ignorance borne by countries without civil registration far outweighs the aff ordable necessity of action.

Who counts?By posing this question, our goal is to focus worldwide attention on the need to recommit resources to the registration of births and deaths, and to certify the causes of death in the world’s poorest countries. Published fertility, mortality, and cause-specifi c mortality fi gures for rich countries are based on data from functioning civil registration systems and can sensitively monitor long-term and short-term demographic changes, and give up-to-date population counts. Fertility and mortality fi gures for countries without birth and death registration, however, are based on a dwindling supply of data

generated from a variety of suboptimum sources.1 Few data for causes of death are available at all—especially for adults. Information needs in resource-poor settings can no longer be met by continued reliance on enumeration by surveys and econometric modelling. These places need to develop civil registration, and complementary, or interim sources of routinely collecting vital statistics. A holistic approach to the development of health information systems is needed,2–6 the centerpiece of which must be a recommitment to civil registration and to making it a sustainable source of quality information that cannot be obtained otherwise.

Lancet 2007; 370: 1569–77

Published OnlineOctober 29, 2007DOI:10.1016/S0140-6736(07)61307-5

See Comment pages 1526 and 1527


This is the fi rst in a Series of four papers about the importance of collecting data for health development

*Members listed at the end of article

MEASURE Evaluation, Carolina Population Center and Departments of Epidemiology and Anthropology, University of North Carolina at Chapel Hill, Chapel Hill NC USA (P W Setel PhD); Global Health Sciences, University of California, San Francisco, USA (S B Macfarlane MSc); History Faculty and St John’s College, Cambridge, UK (S Szreter PhD); Health Metrics Network, Brisbane, Queensland, Australia (L Mikkelsen PhD); Centre for Global Health Research, St Michael’s Hospital, University of Toronto, Canada (Prof P Jha DPhil); Results Secretariat, World Bank, Washington, DC, USA (S Stout DrPH); and Health Metrics Network, Geneva, Switzerland (C AbouZhar MSc)

Correspondence to:Philip W Setel, MEASURE Evaluation, Carolina Population Center, University of North Carolina at Chapel Hill, 206 West Franklin St, Chapel Hill, NC 27516, USA [email protected]

Key messages

• Civil registration and the resulting vital statistics are essential public goods that benefi t individuals and societies. Legal documents that prove identity and citizenship not only provide access to state services or entitlements, but can also be a defence against exploitation or protracted hardship in times of emergency. If vital statistics of births and deaths are combined with accurate cause of death data, their usefulness for health decisionmaking is greatly increased. The eff ect of having such data for health policies and programmes can be seen across a range of countries with widely diff erent national incomes

• The persistent failure to establish, support, and sustain civil registration systems over the past 30 years, and to ensure that causes of death are accurately known in the world’s poorest countries is a scandal of invisibility, for which aff ordable remedies exist and need to be implemented. The scope of this scandal is enormous. Few countries in greatest need of vital events and information about cause of death have the capacity to obtain it

• Overcoming decades of stagnation will need countries to make a principled long-term commitment to comprehensive civil registration, and to make pragmatic use of complementary or alternative registration systems and sources of data for vital events and cause of death in the short term and medium term

”To make people count, we fi rst need to be able to count people” LEE Jong-Wook, WHO Director-General, 2003–2006, address to WHO staff July 21, 2003

Series


Most people in Africa and Asia, and many in other regions, are born and die without leaving a trace in any legal record or offi cial statistic, and without attaining the UN-proclaimed right to a recorded name and nationality.7 Each year, nearly 50 million newborn children are not registered,8 barely a third of countries outside North America and Europe have the capacity to obtain usable mortality statistics, and half the countries in Africa and Southeast Asia record no cause of death data at all.9 Stagnation in the maintenance of civil registration systems and the resulting failure to develop sound data for vital statistics and cause of death over the past 30 years are at the root of this scandal of invisibility, which renders most of the world’s poor unseen, uncountable, and hence uncounted. Indeed, because of the insuffi ciencies in vital statistics no authoritative evidence is available to show whether or not billions of dollars of aid funds are having their desired eff ect on either mortality or poverty.3

We argue that civil registration, vital statistics, and data for cause of death are essential public goods. They are of crucial importance to the health sector and beyond. The civil registration of births and deaths—which are our prime concern—conveys human rights to individuals and helps with economic development. Vital statistics on births and deaths, together with reliable data for cause of death, provide crucial information for policy, planning, and evaluation in all sectors of development. The health sector has a responsibility to work closely with registration authorities, national statistical offi ces, and other agencies to promote and support the establishment and maintenance of civil registration.

Sources of vital statisticsThe administrative and technical functions of civil registration and vital statistics systems can be confi gured in many ways, and responsibilities for maintaining the system and obtaining the vital statistics vary from country to country.10–13 Figure 1 shows levels and functions in civil registration and the production of vital statistics. Locally, individuals report births and deaths to civil authorities and receive legal documentation—birth and death certifi cates or burial permits. In many countries other parties, such as local leaders, the police, and the health system, also have a role in the reporting of events to local area registration authorities. These local authorities compile and consolidate information received and forward it on. Most commonly, the system is maintained by a national registration offi ce in the Ministry of Interior or Justice, and sometimes, the Ministry of Health. Production of

Figure 1: General structure of civil registration systems

Ministry of Health providestechnical support

Local area registry officeconsolidates and

compiles data

Registrar General(located, eg, in Ministry of

Interior, Home Affairs,Justice, or Health) compiles

and sets standards

Local civil servantregisters eventand produces

legal documents

Health systemreports births, deaths,

and causes of death

Police report deaths(eg, traffic fatalities

or from violence)

Community leadersor lay reporters

report vital events

National Statistics Officecompiles national data set

Individuals reportbirths and deaths

Panel 1: Glossary

In everyday use, the vocabulary of registration is often used in confused and confusing ways. For example, civil registration, vital registration, and vital statistics are often used interchangeably although they are not synonymous. The following terms are used in this paper and refer to the specifi c defi nitions provided here:

Civil registrationAs defi ned by the UN is the continuous, permanent, compulsory, and universal recording of the occurrence and characteristics of vital events (livebirths, deaths, fetal deaths, marriages, and divorces) and other civil status events pertaining to the population as provided by decree, law, or regulation, in accordance with the legal requirements in each country. It establishes and provides legal documentation of such events. These records are also the best source of vital statistics13

Demographic surveillance siteThe continuous registration of all demographic events, including cause of death by verbal autopsy, in a geographically defi ned population; usually established for the purpose of health and development intervention research

EnumerationDistinct from registration; the means by which the presence of individuals in a household or other group is recorded; normally used in reference to a census or survey. Enumeration is anonymous and does not provide any direct benefi t to the individual

Sample registration systemThe longitudinal registration of demographic events, including cause of death by verbal autopsy, in a nationally representative sample of clusters such as exists in China and India

Continues on next page

Series


vital statistics, on the other hand, is usually the responsibility of the national statistical offi ce with technical support provided by the Ministry of Health. The functions and legal framework for civil registration are laid out by the UN Statistics Division in its Principles and Recommendations for

a Vital Statistics System.13 Although civil registration with high and representative

coverage should be the long-term goal, investment in complementary, interim sources of statistics in the short term to medium term is needed—particularly for statistics on mortality levels and causes of death.14 As noted by the UN Statistics Division: “[Although] there is no substitute for the availability of continuous information on vital events as obtained from registration of vital events in civil registration…[a]llowance is made, as appropriate, for the use of other sources of complementary or alternative data.”13 Panel 1 provides defi nitions of technical terms related to civil registration, alternative sources of routine statistics, and data for cause of death.

Health sector demands for dataSeveral UN and WHO reports and publications have summarised the poor state of birth and death registration in poor countries.9,10,15–17 The table shows the estimated proportion and number of births in each WHO region that go unregistered every year. The inequalities in registration rates are large; developing countries account for 99% of the estimated 48 million unregistered births, with South Asia and sub-Saharan Africa together accounting for 79% of all unregistered births. According to WHO (data not shown), these inequalities extend to the capacity to obtain cause-specifi c mortality statistics with only a third of all countries having complete data for cause of death.9 In Africa only the small island states of Mauritius and the Seychelles have complete registration of births, deaths, and cause of death. South Africa is the only other African country in which registration of births and deaths is high, but its cause of death data at the national level, although they have improved in recent years, still have large proportions of deaths attributed to undetermined causes. Of the remaining 43 countries in WHO’s Africa region, only Madagascar and Zimbabwe are able to obtain some data for cause of death. In southeast Asia, Sri Lanka, and Thailand have high registration coverage, but only fi ve other countries obtain even fragmented data for cause-specifi c mortality. In the Americas, by contrast, 24 countries have adequate data for mortality, although not all of these have high registration coverage, and eight countries have poor data for cause of death or none at all.

In countries in which civil registration systems do not exist or do not work effi ciently, the health sector usually calls for statistics on birth and death and lobbies for improved civil registration. The health sector tends to be proactive in this regard, not only because of the importance of vital statistics to health policy and planning, but also because the health system is well placed to inform and support the registration of births and deaths, and to provide

Births (in thousands)

Proportion of unregistered children

Number of unregistered children (in thousands)

South Asia 37 099 63% 23 395

Sub-Saharan Africa 26 879 55% 14 751

Middle east and north Africa

9790 16% 1543

Commonwealth of Independent States and Baltic States

5250 23% 1218

East Asia and Pacifi c 31 616 19% 5901

Latin America and Caribbean

11 567 15% 1787

Industrialised countries 10 827 2% 218

Developing countries 119 973 40% 48 147

Least developed countries 27 819 71% 19 682

World 133 028 36% 48 276

Table: Estimated annual number and proportion of unregistered births by region, 20037

Continued from previous page

Verbal autopsyA structured interview with caregivers or family members of households after a death occurs; used to establish probable cause of death in places where most deaths take place outside of health facilities and direct medical certifi cation is rare

Vital eventAs defi ned by the UN is the occurrence of a livebirth, death, fetal death (defi ned as the death of a fetus before birth or extraction from its mother, irrespective of the duration of pregnancy), marriage, divorce, adoption, legitimation, recognition of parenthood, annulment of marriage, or legal separation13

Vital registrationAll sanctioned modes of registering individuals and reporting on vital events. These modes can include registration activities through complementary systems that are not done as part of the civil formal registration system and do not produce legal birth or death certifi cates

Vital statisticsSummary measures of vital events drawn from all of sources of vital events data. Particularly in developing country settings, where civil registration functions poorly or not at all, the UN acknowledges that many data sources and systems are used to derive estimates of vital statistics

Vital statistics systemAs defi ned by the UN is the total process of (1) obtaining information by civil registration or enumeration on the frequency or occurrence of specifi ed and defi ned vital events, and relevant characteristics of the events themselves; and (2) of compiling, processing, analysing, evaluating, presenting, and disseminating these data in statistical form13

Series


information about cause of death. In such countries the health sector also usually supports demographic surveys and the establishment of demographic surveillance systems or sites. National statistical offi ces in these countries can contribute to collection of data for demographic and health surveys, but are often so overwhelmed with their responsibilities for conducting censuses and other national surveys, and for obtaining important economic statistics, that they have few resources to spare for the establishment and maintenance of registration systems. National registration offi ces might not have resources for a suffi cient network of offi ces, or might fail to perceive the need for civil registration beyond selected urban areas.

The benefi ciaries, benefi ts, and risks of registrationCivil registration and death certifi cation with high and representative coverage is essential for individuals, national and sub-national authorities, and the international community. Other methods of registration and collection of data (eg, sample registration systems or research demographic surveillance sites) for cause of death address some statistical needs, but are clearly less useful than civil registration in the long term, and cannot provide the benefi ts that civil registration provides to individuals.

Benefi ts to individuals Civil registration is the only means of establishing and protecting identities, citizenship, and property rights. Legal documents that prove identity and citizenship not only provide access to state services or entitlements, but can also be a defence against exploitation.18 Additionally, these documents (and those that prove marital status) are crucial to property rights of those who either inherit property or seek to use it as collateral to secure loans.7 In the context of the AIDS epidemic, legal documents and certifi cates are increasingly important with regard to the proper allocation of inheritances. In many countries, an inability to prove a legal relationship to a deceased spouse or father heightens the risk of alienation and deprivation of women and children. Registration can also be an instrument of social protection against forms of systemic abuse and extended hardship—whether from the exploitations of human traffi cking or child labour, or that caused in times of confl ict or other complex emergencies.18 For example, in the aftermath of the Asian tsunami in 2004, the availability of registration records in Thailand and Malaysia helped with the reunifi cation of family members and the identifi cation of those lost, compared with Indonesia in which these tasks were made much more diffi cult because of poor civil registration.19

Benefi ts to societyData for births, deaths, and cause of death are essential for planning of services for populations. Data for mortality, particularly for cause of death, are needed to

set priorities, formulate policies, and monitor and assess such policies nationally and, increasingly, at lower administrative levels. Under sectoral reform programmes in many low-income countries, planning, budgeting, and monitoring responsibilities have been devolved to the regions, provinces, or districts. Yet few data exist at these levels—even from surveys—and the demand for such information is rapidly increasing. As completeness of civil registration and certifi cation rises, more accurate vital statistics can be obtained by, and obtained for, small administrative areas. Additionally, continuous registration enables annual estimates of population size, and should ensure that such estimates are comparable from year to year—a challenge when many surveys use diff erent methods, defi nitions, and variables.

The benefi ts of civil registration, alternative registration system information, and data for cause of death are evident across a range of countries with widely diff erent gross domestic products (panel 2). At one end of the spectrum are developed countries, in which good data for mortality and cause of death have aff ected policy decisions and legislation. Once implemented, these policies can be eff ectively monitored because results can be seen in subsequent data. In middle-income, and even low-income countries, such as Chile and India, investments in civil registration, data for cause of death, and complementary sources of vital statistics have also proven useful. Birth registration and information about perinatal outcomes in Chile and India were available to policymakers and civil society, and had an important eff ect on population policies. In Africa, South Africa is one of the best-documented cases in which the absence of good data for cause of death allowed—for a time—poor national policies to continue, and the improved use of existing data for vital events has led to changes in policy and programme priorities.40–48 In the early 2000s, this country’s available data for vital events pointed unambiguously at a huge increase in adult deaths. Absent information about the causes of those deaths, however, provided an opportunity for a government that was offi cially sceptical of AIDS to persist in casting doubt on the true eff ects of the epidemic in their country. Authorities in Cape Town participated in analysis of existing information about cause of death, particularly about AIDS and homicide, and gained an appreciation of the value of locally generated data for local decisionmaking. In both Uganda and Tanzania civil registration does not function as well as in South Africa, thus other sources of data for vital events have been used to set priorities, formulate essential national health intervention packages, and make decisions about resource allocation in individual districts.49–52

Benefi ts to the international communityInternational commitments to fi ght specifi c diseases are now funded at levels never seen before, and their success is often intended to be assessed on the reduction of deaths due to specifi c causes. As demands to measure the

Series


eff ectiveness of health aid have grown, so have calls for good monitoring of vital events and data for cause of death.4,6,53–57 The Millennium Development Goals (MDGs), in particular, have focused attention on this issue. Donor and development agencies, multilateral development banks, and the UN, and 189 nations committed themselves to achievement of these goals when they signed the Millennium Declaration in 2000.58 This commitment provides new energy and focus to the development enterprise, setting eight major goals to be achieved by 2015.

As these goals were established, development agencies and national governments alike recognised a need not only to increase the overall amount of offi cial development assistance, but also to align it well with a country’s strategies and systems, and to substantially increase eff orts to measure, monitor, and manage results.59 This focus on results has brought realisation about the absence of even basic statistical capacities—especially in low-income countries.

At least six of the MDGs (ie, eradication of extreme poverty and hunger; achievement of universal primary education; attaining gender equality; reductions in child mortality; improvements in maternal health; and reductions in prevalence of HIV/AIDS, malaria, and other diseases) rely on accurate data for fertility, mortality, and causes of deaths. Within the health sector alone, eight health outcome and survival indicators and subindicators need high-quality data for births, deaths, and causes of death. Few countries in Africa have the ability to measure these indicators even nationally, and only Mauritius and the Seychelles have the information needed to measure them adequately subnationally—either by geographic area or by groupings based on socioeconomic status.

Not only health policies, but also general economic and social policies are constrained by the simple absence of information about the distribution and characteristics of populations. The stakes cannot be overstated. In 2004, overall amounts of offi cial development assistance, net of debt relief, reached about US$80 billion.59 Ensuring the eff ectiveness of the specifi c policies and programmes that are supported by this aid is clearly important to the taxpayers in those countries that provide the assistance. It is most important to citizens in developing countries who inevitably pay a deadly price if assistance intended to reach them is either badly managed or ineff ective because decisionmakers are uninformed. How much longer support for eff orts to expand immunisation, and confront AIDS, tuberculosis, and malaria will last is questionable if counting the lives saved, and providing direct evidence of reduction of deaths due to these causes—particularly in the poorest of the poor—remains undone.

Risks associated with registrationThe undoubted benefi ts of civil registration systems need to be weighed against the need for eff ective systems to protect against inappropriate use of

registration information. Unless identities are protected, this powerful instrument can be—and has been—used to do great harm to individuals and vulnerable minorities. Examples include the use of the Dutch population registers by the Nazi regime to locate and exterminate Jewish families,60 and the role of identity cards in the Rwanda genocide.61 Even during peacetime, identity registration has been used by governing authorities to control the movement and liberty of

Panel 2: Vital events data lead to improvements in decisions and policies

Developed countriesInformation about road traffi c fatalities, a leading cause of death and disability for young adults, has long been available from many countries. Steadily increasing trends in road death rates until the early 1970s eventually led to the introduction and enforcement of speed limits, seatbelt laws, and laws on alcohol use and driving. The remarkable reversal of traffi c mortality trends closely followed the introduction of these measures and has been seen in many studies.20

ChileIn 2001, several government departments, led by the Chilean Ministry of Health sought to improve the quality, consistency, coverage, and completeness of information about maternal and newborn health in the country. Interventions were designed to: improve guidelines for registration and reporting; cross-check data for completeness and accuracy; allow direct access to data in cases of death or questionable fi ndings; allow free internet-based access to all information by researchers, the press, and civil society. Between 2001 and 2004, key indicators of data accuracy improved as a result of these measures and, more importantly, key outcome indicators of perinatal health improved.21–24

IndiaThe use of data for birth monitoring is highly evident in India’s national family welfare programmes, and in policies and targets related to national population planning and infant and child mortality.25–29 These varied sources have permitted setting of evidence-based targets and the direct measurement of progress in population, and family planning and welfare. The halving of India’s fertility rate in the second half of the 20th century28,30 would seem to be an unambiguously positive development. Many sources of birth data, however, have allowed some undesirable ramifi cations of population change to come to light. In India, expanded access to, and unethical use of specifi c medical technologies and services (namely fetal sexing by ultrasound) along with economic constraints and the cultural privilege of males has enabled prenatal sex selection of male babies. The scope and scale of the selective abortion of female fetuses in India received enormous international media attention and generated heated debate.31–35 It also spurred the strengthening of the 1994 Prenatal Diagnostic Techniques (Regulation and Prevention of Misuse) Act.36 The mere fact that this issue has come to light so starkly—that debates are even possible about the quality, use, and interpretation of evidence in policies and the laws to back them up—speaks directly for the necessity of better data for births.

TanzaniaBeginning in the 1990s, several districts implemented sentinel demographic surveillance systems that entailed routine monitoring of vital events and data for cause of death derived from a validated set of core verbal autopsy procedures. District councils used this information for annual planning and resource allocation cycles to identify disease burdens to set programme priorities, and for the allocation of resources for cost-eff ective interventions.69 district health boards and management teams reallocated funds in accordance with profi les of local mortality burden to address HIV/AIDS, malaria, and other avoidable causes of childhood and maternal mortality.37,38 In urban areas, they also used these data as a basis for improvement of quality of care in health facilities.39

Series


sections of their populations, for example in communist China, apartheid South Africa, and tsarist and soviet-era Russia.

These historical lessons are warnings that careful thought should be given to the ethical dimensions and data security needs of civil registration. The ownership, administrative custody, and secure use of data are of prime importance and constitute an essential social service.62 If reasonable concern exists about a government’s ability to implement strong measures to protect identity data, some open dialogue can be called for, in which the potential risks of collecting information about religion or ethnicity, for example, are balanced against the undoubted importance of having good information about marginalised subpopulations. The state’s stewardship of the system should also include public and democratic procedures for obtaining permission to use data containing individual identifi ers. At the same time, crucial protections for individuals cannot impede legitimate uses of aggregated anonymous data for the purposes of statistical analysis of social conditions and population health.

Although the UN defi nes civil registration as compulsory, the reality is that many individuals who have historical reasons to fear the negative consequences or abuse of identity registration, or who have strong cultural views that might constrain them from registering infants younger than a certain age, will simply not comply and will risk the consequences. When developing a comprehensive long-term strategy to improve the measurement of vital events, the degree of democracy and the status of ethnic minorities and individual rights in the country should be taken into account. Any legacy of abuse of individual information can aff ect future uptake of registration even after bureaucratic obstacles are removed.63,64 A system that is not trusted by the population it serves is not likely to succeed in the long term. Only by serving as the legal guardians of personal, familial, ethnic, and religious identities can registering authorities engender the necessary trust, and

defend against system abuse. Once the technical and administrative safeguards are in place, they must also explicitly address cultural and behavioural sensitivities while promoting registration. Maintaining these constitutional fi rewalls adds a little to the cost of running the system.

Failure to support civil registrationSince 1968, the international mandate for strengthening vital statistics has been assigned to the statistical community led by the Statistical Commission. Ironically, the Statistical Commission has no authority or resources to demand the improvement of civil registration systems that are not directly under its control, and for which it has no budget. Attention has been focused on the products (ie, vital statistics) of the systems rather than on the systems themselves.65 The health sector, too, has tended to take an instrumental view of civil registration, valuing it solely because of its potential to deliver statistics, and framing the problem as one in need of a technological fi x. Alternative strategies to generate birth and death data have been developed that rely largely on survey enumeration and expensive international technical and fi nancial assistance. As dependence on these sources has grown, national authorities have reduced incentives to further invest in civil registration systems (fi gure 2).66

The International Institute for Vital Registration and Statistics, which had as one of its major objectives the coordination of eff orts in this area, was forced to substantially reduce programmes in the 1990s and ceased to function entirely in 2006 because of lack of fi nancial support. According to one of its former directors, international funding agencies decided that it was cheaper and quicker to obtain macro data through household surveys than to support the long-term development of registration systems (Robert A Israel, personal communication, May 11, 2007).

A little more than a decade later, countries are demanding fi nancial and technical support to build statistical systems capable of monitoring development indicators. Towards this end, national statistical offi ces in low-income and middle-income countries are supported by several donors to develop national strategies for the development of statistics.67

These strategies are comprehensive planning exercises that span all processes for obtaining data that make up a national statistical system. Of these strategies, few have given attention to the registration and certifi cation of births, deaths, and causes of deaths. Countries at the early stages of the strategic planning process have a unique opportunity for their health ministries and statistical offi ces to coordinate their eff orts to develop civil registration.

The aff ordable necessity of actionCivil registration systems, and the statistics they generate are intrinsic to countries’ development, and not one of its

1980–84 1985–89 1990–94

Years

Num

ber o

f cou

ntrie

s

1995–99 2000–04

SurveysCivil register

50

40

30

20

10

0

Figure 2: Collection and reporting of data for mortality by source in 50 low-income countries, 1980–20044

Series


luxury byproducts.7 Although additional studies are needed to understand the complex relations between economic growth and social sector development on the one hand, and quality, high-coverage, and eff ective use of vital statistics on the other, historical and contemporary evidence suggest that they go hand-in-hand. The signifi cance of registration systems for legal property rights has sometimes been the primary motivation for national governments to invest in creating these systems.68

The years of disorganised, underfunded, and poorly coordinated international support; absence of national political commitment and public awareness; disorganised or weak registration infrastructures with unclear roles and responsibilities; inadequate fi nancial and human resources; and geographic or socioeconomic isolation combined to keep complete coverage of civil registration a distant goal for many developing countries.

However daunting, each of these constraints can and should be addressed. First and foremost, the establishment of a civil registration system is an act of political will, a demonstration by national authorities of stewardship, and of reciprocal trust in their government by the registered population. For more than 200 years, systems of civil registration have been established by countries undergoing economic development (eg, the UK and Japan) or perceived as liberating by their populaces on emerging from political strife and confl ict (eg, France, Cuba, or South Africa). The maintenance of civil registration needs both the trust and willing participation of citizens and ongoing political backing and commitment to long-term funding. The absence of national political commitment in low-income countries was noted nearly 30 years ago as a fundamental obstacle,69 which remains true today.

The easiest option is to cite lack of resources as the main reason for not moving forward with registration. Although costs are certainly an important consideration, they are not the crucial barrier to improvement. Gross national income per head correlates somewhat with estimated completeness of registration, but some low-income countries do have good coverage (fi gure 3), showing that much can be achieved, even in low-income countries. Consideration of the costs of little investment for the past thirty years, and what the costs of not investing now will be in the future is important. A small but growing body of research about the costs of obtaining information suggests that civil registration and complementary systems are quite aff ordable.70,71 When the costs are divided by the number of people benefi ting from the information generated, the costs become negligible. Vital statistics data are expensive only if they are obtained but never used.

ConclusionsThe worldwide AIDS pandemic clearly shows that visibility demands accountability, which in turn generates the ability to count. In the 1990s the realities of people living with AIDS in heavily aff ected countries became visible,

and the imperative of action became irresistible. In 2001, the international community put a price-tag on action, and decided that provision of pharmaceutical care at the same standard as developed countries to as many people as could be reached in developing countries was a moral imperative.72,73 Setting targets and measuring progress toward those targets was central to the international response (for example, WHO’s 3 by 5 initiative, which aimed to provide antiretroviral therapy to 3 million people with AIDS by 2005). By contrast, unregistered people remain unseen; no global imperative exists to make them more visible and, far from advancing into this century, the inadequate state of civil registration in developing countries remains mainly as it was three decades ago.

This Series focuses on the need for investment in obtaining civil registration and data for mortality in a large part of the world. After decades of working with the limitations of incomplete data, the fundamental problems of absent and defi cient civil registration have remained unaddressed. Now is a time for action. The continued costs of ignorance borne by countries without fi rmly founded vital statistics far outweigh the aff ordable necessity of action. So, who counts? Everyone. And the international community should set a future date for ensuring that everyone is counted.ContributorsPWS, SBM, SSz, LM, PJ, SSt, and CAZ wrote, edited, and reviewed the manuscript.

MoVE writing groupIn addition to the authors of this article are: Robert N Anderson (National Center for Health Statistics, Centers for Disease Control and Prevention, Hyattsville, MD, USA); Ayaga A Bawah (INDEPTH Network, Accra, Ghana); Ana Pilar Betrán (Department of Making Pregnancy Safer, WHO, Geneva, Switzerland); Fred Binka (INDEPTH Network, Accra, Ghana);

00

20

5000 10 000 15 000

Gross national income per head (international dollars)

Estim

ated

cove

rage

of c

ivil

regi

stra

tion

syst

em (%

)

20 000 25 000 30 000

40

60

80

100

Figure 3: Association between estimated coverage of civil registration and gross national income per head, 1998–2004

Series


Kanitta Bundhamcharoen (International Health Policy Program, Bangkok, Thailand); Rene Castro (Ministry of Health, Chile); John Cleland (London School of Hygiene and Tropical Medicine, London, UK); Francesca Coullare (United Nations Statistics Division, New York, USA); Timothy Evans (Information, Evidence and Research, WHO, Geneva, Switzerland); Ximena Carrasco Figueroa (Department of Health Statistics and Information, Ministry of Health, Chile); Chakkalackal Korah George (Institute of Health Systems, Hyderabad, India); Laragh Gollogly (Department of Knowledge, Management and Sharing, WHO, Geneva, Switzerland); Rogelio Gonzalez (Ministry of Health, Chile, and Center for Perinatal Diagnosis, Hospital Dr Sotero del Rio, Pontifi ca Universidad Catolica de Chile, Santiago, Chile); Danuta Rajs Grzebien (Department of Health Statistics and Information, Ministry of Health, Chile); Kenneth Hill (Harvard Center for Population and Development Studies, Cambridge, MA, USA); Zhengjing Huang (National Center for Chronic and Non-communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention, Beijing, China); Terence H Hull (Australian Demographic and Social Research Institute, Australian National University, Canberra, Australia); Mie Inoue (Department of Measurement and Health Information Systems, WHO, Geneva, Switzerland); Robert Jakob (Department of Measurement and Health Information Systems, WHO, Geneva, Switzerland); Yong Jiang (National Center for Chronic and Non-communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention, Beijing, China); Ruy Laurenti (Department of Epidemiology, School of Public Health, University of Sao Paulo, Brazil); Xiaoyan Li (National Center for Chronic and Non-communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention, Beijing, China); Denise Lievesley (International Statistical Institute, London, UK); Alan D Lopez (School of Population Health, University of Queensland, Brisbane, Australia); Doris Ma Fat (Department of Measurement and Health Information Systems, WHO, Geneva, Switzerland); Prasanta Mahapatra (Institute of Health Systems, Hyderabad, India); Mario Merialdi (Department of Making Pregnancy Safer, WHO, Geneva, Switzerland); Jyh Kae Nien (Center for Perinatal Diagnosis, Hospital Dr Sotero del Rio, Pontifi ca Universidad Catolica de Chile, Santiago, Chile); Francis C Notzon (National Center for Health Statistics, Centers for Disease Control and Prevention, Hyattsville, MD, USA); Chalapati Rao (School of Population Health, University of Queensland, Brisbane, Australia); Keqin Rao (Center for Health Statistics Information, Ministry of Health, Beijing, China); Osman Sankoh (INDEPTH Network, Accra, Ghana); Kenji Shibuya (Department of Measurement and Health Information Systems, WHO, Geneva, Switzerland); Nadia Soleman (Health Action in Crises, WHO, Geneva, Switzerland); Viroj Tangcharoensathien (International Health Policy Program, Bangkok, Thailand); Paul J van der Maas (Erasmus University Medical Center, Rotterdam, The Netherlands); Fan Wu (National Center for Chronic and Non-communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention, Beijing, China); Gonghuan Yang (Chinese Center for Disease Control and Prevention, Beijing, China); Siwei Zhang (National Center for Cancer Registry, Beijing, China); Maigeng Zhou (Chinese Center for Disease Control and Prevention, Beijing, China).

Confl ict of interest statementWe declare that we have no confl ict of interest. The views expressed in this manuscript are not necessarily those of the Health Metrics Network or USAID.

AcknowledgmentsFunding for the preparation of this manuscript was provided, in part, by the Health Metrics Network and MEASURE Evaluation, Phase 2, a USAID cooperative agreement (GPO-A-00-03-00003-00), implemented by the Carolina Population Center, University of North Carolina at Chapel Hill. The authors thank Luong Nguyen, Dan Williams, and Robert Mswia for research assistance in the preparation of the manuscript.

References1 Walker N, Bryce J, Black RE. Interpreting health statistics for

policymaking: the story behind the headlines. Lancet. 2007; 369: 956–63.

2 Bchir A, Bhutta Z, Binka F, et al. Better health statistics are possible. Lancet 2006; 367: 190–03.

3 Jamison DT, Breman JG, Measham AR, et al, eds. Disease control priorities in developing countries. Part 1. Summary and cross-cutting themes. Washington and New York: The World Bank and Oxford University Press, 2006.

4 Boerma JT, Stansfi eld SK. Health statistics now: are we making the right investments? Lancet 2007; 369: 779–86.

5 WHO. World health statistics 2007. Geneva: World Health Organization, 2007.

6 Evans T, Stansfi eld S. Health information in the new millennium: a gathering storm? Bull World Health Organ 2003; 81: 856.

7 Szreter S. The right of registration: development, identity registration, and social security—a historical perspective. World Dev 2007; 35: 67–86.

8 UNICEF. The ‘rights’ start to life. A statistical analysis of birth registration. New York: United Nations Children’s Fund, 2005.

9 Mathers C, Ma Fat D, Inoue M, Rao C, Lopez AD. Counting the dead and what they died of: an assessment of the global status of cause of death data. Bull World Health Organ 2005; 83: 171–77.

10 Powell NP. Organizational arrangements for a vital registration system. Working paper No 1. Bethesda, MD (USA): International Institute for Vital Registration and Statistics; 1977.

11 UN Department of International Economic and Social Aff airs Statistics Division. Handbook on civil registration and vital statistics systems. Preparation of a legal framework. New York: United Nations, 1998.

12 Rao C, Bradshaw D, Mathers CD. Improving death registration and statistics in developing countries: Lessons from sub-Saharan Africa. S Afr J Demogr 2004; 9: 81–99.

13 UN Department of Economic and Social Aff airs Statistics Division. Principles and recommendations for a vital statistics system. Revision 2. New York: United Nations, 2001.

14 AbouZahr C, Cleland J, Coullare F, et al, on behalf of the Monitoring of Vital Events (MoVE) writing group. The way forward. Lancet 2007; published online Oct 29. DOI: 10.1016/S0140-6736(07)61310-5.

15 Mathers C, Inoue M, Lopez A, Shibuya K. Overview of global mortality data sources. Geneva: World Health Organization, 2003.

16 WHO. World health report 2005. Making every mother and child count. Geneva: World Health Organization, 2005.

17 Lawn JE, Cousens S, Zupan J. 4 million neonatal deaths: when? Where? Why? Lancet 2005; 365: 891–900.

18 UNICEF. Birth registration. Right from the start. Innocenti Digest No 9. Florence, Italy: United Nations Children’s Fund and Innocenti Research Centre, 2002.

19 UNICEF. Birth registration: the record in tsunami-aff ected areas. Bangkok; United Nations Children’s Fund, 2005.

20 Lopez AD. Competing causes of death: a review of recent trends in mortality in industrialized countries with special reference to cancer. Ann N Y Acad Sci 1990; 609: 58–76.

21 Szot J. Mortalidad infantil e indicadores económicos en Chile: 1985–1999. Rev Med Chil 2002; 130: 107–12.

22 Donoso E, Villarroel L. Mortalidad neonatal precoz. Chile 1991–97. Rev Chil Obstet Ginecol 1999; 64: 286.

23 Morgues M, Henríquez MT, Tohá D, Vernal P, Pittaluga E, Vega S, et al. Sobrevida del niño menor de 1500 g en Chile. Revista Chilena de Obstetricia y Ginecologia 2002; 67: 100–05.

24 Gonzalez R, Merialdi M, Lincetto O, et al. Reduction in neonatal mortality in chile between 1990 and 2000. Pediatrics 2006; 117: e949–e54.

25 Ministry of Health and Family Welfare. Annual report of the Ministry of Health and Family Welfare, Government of India, New Delhi, 1996–97. New Delhi: Government of India, 1997.



28 Department of Family Welfare. National Population Policy 2000. New Delhi: Ministry of Health and Family Welfare, Government of India, 2000.

Series


29 Department of Family Welfare. Reproductive and child health programs: schemes for implementation. New Delhi: Ministry of Health and Family Welfare, Government of India, 1997.

30 Registrar General of India. Compendium on India’s fertility and mortality indicators: 1971–1997. Vital Statistics Publication, Sample Registration System. New Delhi: Vital Statistics Division, Offi ce of the Registrar General of India, 1999.

31 Jha P, Kumar R, Vasa P, Dhingra N, Thiruchelvam D, Moineddin R. Low female[corrected]-to-male [corrected] sex ratio of children born in India: national survey of 1·1 million households. Lancet 2006; 367: 211–18.

32 George SM. Sex ratio in India. Lancet 2006; 367: 1725.33 Bhat M. Sex ratio in India. Lancet 2006; 367: 1725–26.34 Bhopal R. Sex ratio in India. Lancet 2006; 367: 1728–29.35 Sen A. Missing women. BMJ 1992; 304: 586–87.36 Supreme Court of India. The Supreme Court Judgment dated

May 4, 2001 in the PNDT Act, 1994. Ind J Medical Ethics 2001; 9: 91. http://www.ijme.in/093do7.html (accessed June 15, 2007).

37 de Savigny D, Setel P, Kasale H, et al. Linking demographic surveillance and health service needs-the AMMP/TEHIP experience in Tanzania. Multilateral Initiative on Malaria; 1999 March 15–19; Durban, South Africa, 1999.

38 Mkusa V, Hemed Y, Masuki G, Setel PW. Setting priorities in health care: use of diverse information perspectives at the district level in Tanzania. 4th International Conference on Priorities in Health Care; 2002 18–22 Sept; Oslo, Norway, 2002.

39 Setel P, Kitange H, Alberti K, Moshiro C. The policy implications of adult morbidity and mortality in Tanzania: from data analysis to health policy—preliminary experiences. Global Forum for Health Research (Forum 2); 1998 June 25–26; Geneva: World Health Organization, 1998.

40 Robins S. Long live Zackie, long live’: AIDS activism, science, and citizenship after apartheid. J S Afr Stud 2004; 30: 651–72.

41 Bradshaw D. What do we know about the burden of cardiovascular disease in South Africa? Cardiovasc J S Afr 2005; 16: 140–41.

42 Bradshaw D, Dorrington RE, Sitas F. The level of mortality in South Africa in 1985—what does it tell us about health? S Afr Med J 1992; 82: 237–40.

43 Bradshaw D, Kielkowski D, Sitas F. New birth and death registration forms—a foundation for the future, a challenge for health workers? S Afr Med J 1998; 88: 971–74.

44 Bradshaw D, Nannan N, Groenewald P, et al. Provincial mortality in South Africa, 2000—priority-setting for now and a benchmark for the future. S Afr Med J 2005; 95: 496–503.

45 Bradshaw D, Schneider M, Dorrington R, Bourne DE, Laubscher R. South African cause-of-death profi le in transition—1996 and future trends. S Afr Med J 2002; 92: 618–23.

46 Bradshaw D, Groenewald P, Bourne DE, et al. Making COD statistics useful for public health at local level in the city of Cape Town. Bull World Health Organ 2006; 84: 211–17.

47 Groenewald P, Bradshaw D, Dorrington R, Bourne D, Laubscher R, Nannan N. Identifying deaths from AIDS in South Africa: an update. AIDS 2005; 19: 744–45.

48 Groenewald P, Nannan N, Bourne D, Laubscher R, Bradshaw D. Identifying deaths from AIDS in South Africa. AIDS 2005; 19: 193–201.

49 Global Forum for Health Research Secretariat. The 10/90 Report on health research 1999. Forum 2. Geneva: World Health Organization, 1999.

50 Kapiriri L, Norheim OF, Heggenhougen K. Using burden of disease information for health planning in developing countries: the experience from Uganda. Soc Sci Med 2003; 56: 2433–41.

51 de Savigny D, Kasale H, Mbuya C, Reid G. Fixing health systems. Ottawa, Canada: International Development Research Centre, 2004.

52 Ministry of Health. The policy implications of Tanzania’s mortality burden: mortality burden profi les from sentinel sites, 1994–2002. AMMP-2 Final Report Volume 4. Dar es Salaam: Adult Morbidity and Mortality Project, Tanzania Ministry of Health, University of Newcastle upon Tyne, and UK Department for International Development, 2004. http://research.ncl.ac.uk/ammp/site_fi les/public_html/fi nrep/index.html (accessed June 16, 2007).

53 Tollman SM, Zwi AB. Health system reform and the role of fi eld sites based upon demographic and health surveillance. Bull World Health Organ 2000; 78: 125–34.

54 Diaz T, Loth G, Whitworth J, Sutherland D. Surveillance methods to monitor the impact of HIV therapy programmes in resource-constrained countries. AIDS 2005; 19 (suppl 2): S31–37.

55 Setel PW, Sankoh O, Mathers C, et al. Sample registration of vital events with verbal autopsy: innovative approaches to measuring and monitoring vital statistics. Bull World Health Organ 2005; 83: 611–17.

56 Baiden F, Hodgson A, Binka FN. Demographic Surveillance Sites and emerging challenges in international health. Bull World Health Organ 2006; 84: 163.

57 Murray CJ. Towards good practice for health statistics: lessons from the Millennium Development Goal health indicators. Lancet 2007; 369: 862–73.

58 UN. The Millennium Development Goals Report 2006. New York: United Nations, 2006.

59 World Bank. Global monitoring report 2006: strengthening mutual accountability—aid, trade, and governance. Washington DC: The World Bank Group, 2006.

60 Moore, B. Victims and survivors: the Nazi persecution of the Jews in the Netherlands 1940–1945. London: Arnold, 1997.

61 Longman, T. Identity cards, ethnic self-perception and genocide in Rwanda, in Caplan J, Torpey J, eds, Documenting individual identity. The development of state practices in the modern world Princeton NJ: Princeton University Press, 2001: 345–57.

62 UN Department of Economic and Social Aff airs Statistics Division. Handbooks on civil registration and vital statistics systems, series F, No 70. Policies and protocols for the release and archiving of individual records. New York: United Nations, 1998.

63 Lewando-Hundt G, Abed Y, Skeik M, Beckerleg S, El Alem A. Addressing birth in Gaza: using qualitative methods to improve vital registration. Soc Sci Med 1999; 48: 833–43.

64 Jewkes R, Wood K. Competing discourses of vital registration and personhood: perspectives from rural South Africa. Soc Sci Med 1998; 46: 1043–56.

65 UN Statistical Commission. ECOSOC resolution 1307 (XLIV), adopting the “1970 world programme for the improvement of vital statistics”. UN Statistical Commission Report, 1968: 1968.

66 Mahapatra P, Shibuya K, Lopez AD, et al, on behalf of the Monitoring of Vital Events (MoVE) writing group. Civil registration systems and vital statistics: successes and missed opportunities. Lancet 2007; published online Oct 29. DOI:10.1016/S0140-6736(07)61308-7.

67 PARIS21 Secretariat. A guide to designing a national strategy for the development of statistics (NSDS). Paris: OECD/PARIS21, 2004.

68 Higgs E. Life, death, and statistics. Civil registration, censuses, and the work of the General Register Offi ce, 1836–1952. Hatfi eld, UK: Local Population Studies, 2004.

69 Powell NP. Major obstacles to achieving satisfactory registration of vital events and the compilation of reliable vital statistics. Technical papers of the International Institute for Vital Registration and Statistics, No 15. Bethesda MD: International Institute for Vital Registration and Statistics, 1981.

70 Rommelmann V, Setel PW, Hemed Y, Mponezya H, Angeles G, Boerma T. Costs and results of information systems for poverty and health indicators in the United Republic of Tanzania. Bull World Health Organ 2005; 83: 569–77.

71 Stansfi eld SK, Walsh J, Prata N, Evans T. Information to improve decision making for health. In: Jamison DT, Breman JG, Measham AR, et al, eds. Disease control priorities in developing countries. Part 3. Strengthening health systems. Washington and New York: The World Bank and Oxford University Press, 2006: 1017–30.

72 Schwartlander B, Stover J, Walker N, et al. AIDS. Resource needs for HIV/AIDS. Science 2001; 292: 2434–36.

73 Lange JM, Perriens J, Kuritzkes D, Zewdie D. What policymakers should know about drug resistance and adherence in the context of scaling-up treatment of HIV infection. AIDS 2004; 18 (suppl 3): S69–74.

Seminar


Sudden infant death syndromeRachel Y Moon, Rosemary S C Horne, Fern R Hauck

Despite declines in prevalence during the past two decades, sudden infant death syndrome (SIDS) continues to be the leading cause of death for infants aged between 1 month and 1 year in developed countries. Behavioural risk factors identifi ed in epidemiological studies include prone and side positions for infant sleep, smoke exposure, soft bedding and sleep surfaces, and overheating. Evidence also suggests that pacifi er use at sleep time and room sharing without bed sharing are associated with decreased risk of SIDS. Although the cause of SIDS is unknown, immature cardiorespiratory autonomic control and failure of arousal responsiveness from sleep are important factors. Gene polymorphisms relating to serotonin transport and autonomic nervous system development might make aff ected infants more vulnerable to SIDS. Campaigns for risk reduction have helped to reduce SIDS incidence by 50–90%. However, to reduce the incidence even further, greater strides must be made in reducing prenatal smoke exposure and implementing other recommended infant care practices. Continued research is needed to identify the pathophysiological basis of SIDS.

IntroductionSudden infant death syndrome, or SIDS, is defi ned as “the sudden death of an infant under one year of age, which remains unexplained after a thorough case investigation, including performance of a complete autopsy, examination of the death scene, and review of the clinical history”.1 Despite declines in SIDS rates following risk reduction campaigns, SIDS continues to be the leading cause of death for infants aged between 1 month and 1 year in developed countries. Currently, Japan and the Netherlands have the lowest reported SIDS rates, at 0·09 and 0·1 per 1000 livebirths, respectively,2,3 whereas New Zealand has the highest reported SIDS rate in developed countries, at 0·8 per 1000 livebirths.4 The USA and the UK have intermediate SIDS rates of 0·57 and 0·41 per 1000 livebirths, respectively.5,6

In this Seminar, we focus on newer epidemiological and pathophysiological fi ndings, risk reduction recommendations, and controversies related to some of these recommendations.

EpidemiologyIn the 1980s and 1990s, after epidemiological studies showed a decreased incidence of SIDS in infants who slept supine, many countries implemented public-health campaigns to encourage families to place infants on their back for sleep. In most of these countries, the rate

of placing infants prone for sleep has decreased 50–90%, and the rate of SIDS has similarly decreased 50–90%. As prone sleeping has become a less common risk factor, new epidemiological risk factors have emerged. We will discuss the epidemiology of SIDS both before and after the decrease in prone sleeping.

Demographic factorsSIDS occurs less frequently in the fi rst month of life, peaks between 2 and 4 months of age, and decreases thereafter. Around 90% of SIDS deaths happen in the fi rst 6 months of life. Boys are more likely to die than girls, at a ratio of 60:40.

Despite the overall decline in SIDS worldwide, there are still racial and ethnic disparities. In the USA, infants who are African American, Native American, or Alaska Native have SIDS rates that are two to three times the national average, irrespective of socioeconomic status.7,8 Maoris in New Zealand9 and Aboriginal Australians10–12 are also at higher risk for SIDS. Maoris are six times more likely to die of SIDS than non-Maori New Zealanders.13 Although biological diff erences (such as racial diff erences in tobacco metabolism14) might partially explain the disparity, behavioural factors also have an eff ect on the disparity. For instance, African Americans are twice as likely to place infants prone for sleep15 and twice as likely to bedshare than white people.16 Although the supine sleep position is the norm for infants who are Native American, Alaska Native, Aboriginal Australian, and New Zealand Maori, there are also high rates of smoke exposure17 and bedsharing in these groups, which place these infants at higher risk of SIDS.11,18

Maternal smoking and smoke exposureMaternal smoking during pregnancy is a major risk factor in almost every epidemiological study of SIDS.19–21 Postnatal exposure to tobacco smoke has emerged as a separate risk factor in a few studies,21,22 although this variable is diffi cult to separate from maternal smoking prenatally. Prenatal smoke exposure results in decreased lung volume and compliance,23 and decreased heart rate variability to stress.24 Additionally, nicotine has neuroteratogenic eff ects,

Lancet 2007; 370: 1578–87

Goldberg Center for Community Pediatric Health,

Children’s National Medical Center and George Washington

University School of Medicine and Health Sciences,

Washington, DC, USA (R Y Moon MD); Ritchie Centre

for Baby Health Research, Monash Institute for Medical Research, Monash University,

Melbourne, Australia (R S C Horne PhD); and

Departments of Family Medicine and Public Health

Sciences, University of Virginia School of Medicine

Charlottesville, Virginia, USA (F R Hauck MD)

Correspondence to:Rachel Y Moon,

Children’s National Medical Center, Michigan Avenue NW,

Washington, DC 20010, [email protected]

Search strategy and selection criteria

We used the PubMed database to search for publications, using the search terms “SIDS,” “crib death,” “infant death,” and “sudden infant death syndrome.” Additionally, we reviewed listings of articles received through various mailing lists to identify publications that were not yet in PubMed. Citations were selected from articles published in English. We mostly selected publications from the past 5 years but did not exclude commonly referenced and highly regarded older publications. Reference lists in key textbook chapters and review articles were also checked for relevant publications and references.

Seminar


resulting in alterations in autonomic pathways, including decreased arousal to hypoxia and other stimuli.25–30 All of these eff ects could directly aff ect SIDS risk. If in-utero smoke exposure was eliminated, a third of SIDS deaths could possibly be prevented.31

Sleep practices and environmentSince the introduction of international Back to Sleep campaigns and the subsequent large decrease in the proportion of infants placed prone to sleep, the contribution of the side sleep position to SIDS risk has increased, such that the risk of side and prone placement are similar.9,32–35 The side sleep position is unstable, and many infants placed this way will roll to prone,36,37 placing them at high risk for SIDS.34,35,38 Infants usually placed supine are sometimes placed prone by secondary caregivers (grandparents, babysitters, child care providers, or other relatives).

Soft bedding and soft surfaces, including pillows, quilts, comforters, sheepskins, and porous mattresses, have been shown to be important risk factors.39–45 In particular, a strong interaction has been found between prone sleep position and soft bedding surface.45 Soft bedding is also a likely factor contributing to infant deaths occurring on adult beds.46–48

Warmer room temperatures and multiple layers of clothing or blankets on the infant have also been associated with an increased SIDS risk.43,49,50 This increased risk of overheating is particularly evident when infants are sleeping in the prone position and heat loss from the face is reduced;43 the risk is less clear when infants are supine.

Bed sharing between an infant and adult facilitates breastfeeding and enhances parent-infant interactions.51,52 However, epidemiological studies have shown that this practice can be hazardous, especially when there are multiple bedsharers,45 when the infant is younger than 11 weeks,53,54 and when bed sharing occurs for the entire night.55,56 The risk for infant death might also be increased when the person sharing the bed has consumed alcohol or is overtired.53,56 Bed sharing is particularly hazardous with mothers who smoke,55,56 but sharing with non-smoking mothers is also a risk factor among infants younger than 11 weeks.53,54 Bed sharing on a couch or sofa is particularly dangerous and should always be avoided.45,54,57,58 There is growing evidence that room sharing without bed sharing is associated with a reduced risk of SIDS.53–55 The USA, UK, Canada, Australia, and several other countries currently recommend that infants sleep in a crib or bassinet next to the parents’ bed.

Published case-control studies have shown a signifi cantly reduced risk of SIDS when a pacifi er is used at sleep time,45,53,58–63 and two meta-analyses have reported a strong protective eff ect.64,65 Several mechanisms have been postulated to explain this protective eff ect. Franco and colleagues66 found a lower arousal threshold in infants who frequently used a pacifi er than those who

did not during sleep, perhaps allowing increased responsiveness to a life-threatening challenge, such as obstructive apnoea, cardiac arrhythmia, or external conditions leading to hypoxia and asphyxia. Other theories include an improved ability to breathe through the mouth if the nasal airway becomes obstructed61,67,68 and decreased likelihood of oropharyngeal obstruction by bringing the tongue forward.66,67 Use of a pacifi er might also reduce SIDS risk by aff ecting sleep position.59,61,68 Even when pacifi ers become dislodged from the mouth after an infant falls asleep, which generally occurs soon after sleep onset,66,69 the protective eff ect still persists. Displacement of the pacifi er might contribute to increased sleep disruption and greater arousability of the infant.70

Infants in child care settings (ie, cared for by a non-parental caregiver, including babysitters and child care providers) make up around 20% of SIDS deaths.71–73 Why these infants are at higher risk is unclear, since there is less exposure to environmental risk factors, such as smoke exposure, soft bedding, prone sleep position, and bed sharing.71,73 Furthermore, a large proportion of these deaths take place during the fi rst week of child care.72 Stress and sleep disruption during the transition to child care could be contributing factors.71,73

Infant medical conditionsInfants who are born prematurely or who have low birthweight have up to four times the risk of SIDS than those infants born at term, and this risk increases with decreasing gestational age or birthweight.74–76 Apnoea of prematurity and other complications of prematurity do not explain this increased risk.74 Infants born preterm who are placed prone are at equally high risk for SIDS as infants born at term.77 However, infants born preterm are often placed prone in the intensive care setting to enhance respiratory mechanics.78,79 The results of one study showed that before neonatal unit discharge, infants born prematurely slept longer, had fewer arousals from sleep, and more central apnoeas when sleeping prone.80 Thus, premature infants should be placed supine as soon as is medically safe, and well before hospital discharge, to ensure and assure that the infant and parents are accustomed to supine placement.

Although apnoea was for many years thought to be the predecessor of SIDS, results of studies such as the Collaborative Home Infant Monitoring Evaluation have shown that it does not precede or predict SIDS.81 Apnoea monitors might be useful in selected patients who have had apparent life-threatening events, but there is no evidence that they are useful for reducing the occurrence of SIDS.82

Evidence of a recent infection has been a common fi nding of SIDS autopsies. Upper respiratory tract infections within 4 weeks of death have been reported in 53% of SIDS cases, compared with 38% of controls.83 The same study also reported that fewer SIDS cases (33%)

Seminar


than controls (68%) had received age-adequate pertussis immunisation.83 There is no evidence that immunisations increase the risk for SIDS.84,85

BreastfeedingEpidemiological studies have produced inconsistent fi ndings on the protective eff ects of breastfeeding for SIDS. Results of some studies have shown a protective eff ect,86–88 whereas others, after adjustment for socioeconomic status, did not show such an eff ect.89–92 Studies in Germany have found breastfeeding initiation to be associated with a reduced risk of SIDS.93,94 Chen and Rogan95 reported that postneonatal death in general, but not SIDS, is decreased in breastfed infants. Physiological studies have shown that breastfed infants arouse more readily at 2–3 months of age than do formula-fed infants,96 and post-mortem studies have shown that formula-fed infants who died of SIDS have signifi cantly lower concentrations of docosahexaenoic acid in the frontal lobes of their brains than do breastfed infants.97 This long chain fatty acid is concentrated in neural tissue and has been reported as a predictor of developmental indices at 1 year of age in preterm infants.98 It has also been associated with more mature infant sleep patterns.99

Recurrence of SIDS in siblingsSiblings of SIDS victims are at increased risk for SIDS. However, the exact extent of the increase in risk is as yet unclear. Metabolic or genetic disorders, such as fatty acid oxidation disorders or prolonged QT syndrome, might go unrecognised and subsequent deaths attributed to SIDS. Although homicide should be considered as a possibility, a sudden infant death in a subsequent sibling is six times more likely to be SIDS than homicide.100 If a family has more than one infant death or if more than one infant dies while in the care of the same caregiver, a careful and complete autopsy, death scene investigation, and review of medical history are especially imperative so that causes other than SIDS can be ruled out.101

PathophysiologySIDS is regarded to be multifactorial in origin. The so-called triple risk hypothesis is a useful model to organise current knowledge. It proposes that SIDS results when three factors coincide: a vulnerable infant; a critical developmental period in homoeostatic control; and an exogenous stressor, such as being placed for sleep in the prone position.102 The triple risk hypothesis posits that infants will die of SIDS only if they possess all three factors, and that the vulnerability lies latent until they enter the crucial developmental period and are exposed to an exogenous stressor. The fi nal pathway to SIDS is widely believed to involve immature cardiorespiratory autonomic control, together with a failure of arousal responsiveness from sleep.

Autonomic control and arousalMore than 20 years ago, autonomic dysfunction or impairment was suggested as a possible cause of SIDS.103 Infants who had been studied and subsequently died from SIDS, compared with surviving infants, had higher baseline heart rates,104 lower heart rate variability,105 prolonged QT indexes,106 and low parasympathetic tone, high sympathovagal balance, or both.107,108

Arousal from sleep is an important survival response to a life-threatening event, such as hypotension or prolonged apnoea.109 Arousal from sleep can occur spontaneously in response to internal physiological changes, and can also be induced by external environmental factors. Arousal involves both autonomic and behavioural components. During arousal from sleep, heart rate, arterial pressure, and ventilation are increased,110 and importantly, a behavioural response is evoked, allowing body movements aimed at avoiding the life-threatening stimulus.109 The cardiorespiratory responses at arousal are similar to fi ght-or-fl ight reactions, which also increase arterial pressure, heart rate, and ventilation.110 These responses are relayed and integrated in specifi c regions of the hypothalamus and the brainstem. The underlying neuronal activity that elicits cortical activation also involves specifi c neuro-transmitter-modulated discharge patterns of thalamo-cortical neurons. The defence-alerting responses range from mild cardiorespiratory activation (eg, a sigh), through awakening, and fi nally to redistribution of blood fl ow and gasping during asphyxia. The carotid bodies are believed to play a major part in the mediation of the arousal response to arterial hypoxaemia, as surgical denervation of the carotid body chemoreceptors produces a delayed or incomplete hypoxic arousal.111 Similarly, the arterial baroreceptors are vital for arousal responses to hypotension and hypertension.112,113

Abnormalities of arousal are now believed to be involved in the fi nal SIDS pathway. Prospective studies of infants who subsequently died from SIDS have identifi ed fewer spontaneous arousals from sleep and immature sleep patterns, compared with those infants who survived.114–116 The most compelling evidence was provided by Kato and colleagues117 in a study of more than 20 000 infants, which found both reduced numbers of spontaneously occurring arousals and a pattern of arousals suggestive of an impaired arousal pathway in infants who subsequently died from SIDS.

Multiple studies of sleep and arousal responses in healthy infants exposed to major risk factors for SIDS have shown that these risk factors alter sleep patterns and physiology and impair infant arousal responses to both internal and external stimuli. Prone sleeping has been shown to increase the total amount of time infants spend asleep80,118,119 and, in particular, the time spent in quiet sleep, a state of reduced arousability.80,118–121 Additionally, infants sleeping in this position have decreased spontaneous arousability,80,118,122 diminished

Seminar


induced arousability,120,121,123–125 and fewer full cortical arousals,126 than those sleeping in the supine position. Prone sleeping has also been associated with altered autonomic control, manifest by raised heart rates,118,121,127–133 decreased heart rate variability,120,123,124,128,133,134 and increased sympathetic tone.120,128,135 Similarly, infants exposed in utero to maternal smoking have been shown to have both decreased spontaneous arousal from sleep29,136–138 and diminished stimulus-induced arousal responses.25,28,137,139 The results of other studies also suggest that autonomic control is altered in these infants.24,27,140 These fi ndings provide further support for the theory that autonomic control and arousal responsiveness are implicated in SIDS causation.

Autopsy fi ndings Although there are no pathognomonic autopsy fi ndings for SIDS, there are several common fi ndings. Petechial haemorrhages of the thymus gland, visceral pleura, and epicardium occur in 68–95% of SIDS deaths and are generally more extensive than in non-SIDS deaths.100 Additionally, pulmonary congestion and oedema indicative of terminal left ventricular failure are more common in SIDS cases.100 Oronasal secretions, which are typically frothy, mucoid, and pink or bloody, are also more common in SIDS cases.141

Other pathological fi ndingsSustained and intermittent hypoxia promotes the formation of vascular endothelial growth factor, which is an endothelial cell-specifi c mitogen that increases peripheral oxygen delivery by stimulating angiogenesis. Even small changes in tissue oxygenation can substantially alter expression of vascular endothelial growth factor.142 In one study, concentrations of this growth factor in cerebrospinal fl uid were signifi cantly higher in 51 infants who died from SIDS than in 33 control infants who died from known causes, suggesting that hypoxia frequently precedes death from SIDS.143

Infants who died from SIDS have also been found to have structural and neurotransmitter alterations in the brainstem in areas associated with autonomic control, control of respiration, sleep, and arousal. Kinney and colleagues144 have reported that the serotonin or 5-hydroxytryptamine (5-HT) system is abnormal in at least 50% of SIDS cases. Serotonin is a neurotransmitter that aff ects various autonomic functions, including cardiorespiratory and circadian regulation. Recently, the same group reported that infants who died from SIDS had a signifi cantly higher number and density of 5-HT neurons, a signifi cantly lower density of 5-HT1A receptor binding sites in regions of the medulla involved in homoeostatic function, and a lower ratio of 5-HTT binding density to 5-HT neuron count in the medulla than infants who died of known cases.145 These fi ndings suggest that the defi ciencies in the medullary 5-HT system are quite extensive and involve abnormal 5-HT

neuron fi ring, synthesis, release, and clearance. Importantly, this study also identifi ed decreased 5-HT1A receptor binding in male infants who died of SIDS compared with female infants, which could explain the increased incidence of SIDS in male infants.145

Recent studies by Lavezzi and colleagues146 show that the cerebellum, a structure not normally associated with respiratory and cardiovascular control, might have an important role in compensatory responses to hypoxic insults. The investigators reported that, compared with 10% of controls, 62% of sudden perinatal and infant deaths showed alterations, such as neuronal immaturity, altered apoptotic programmes, negative expression of somatostatin and the EN2 gene, intense c-fos expression positivity, and astrogliosis in the cortex and dentate nucleus. There was also a strong correlation between these fi ndings and a history of maternal smoking, further supporting the strong link between maternal smoking and SIDS.

Other neuropathological defects and abnormalities found in the brains of infants who died of SIDS include increased numbers of neuronal dendritic spines, which are thought to be indicative of delayed neuronal maturation.141 Waters and colleagues147 have reported increased neuronal apoptosis in the hippocampus and brainstem, which might lead to functional loss in regions sensitive to hypoxia and regions associated with sensation in the face and position of the head, suggestive of an hypoxic insult before death. These fi ndings could explain why infants whose faces have been covered by bedding, as is commonly noted in SIDS deaths, do not simply turn or lift their heads to avoid asphyxia.

Infl ammatory changes have been reported in several systems including the respiratory tract, digestive tract, nervous system, and blood. SIDS could be due to a rapid, uncontrolled release of infl ammatory mediators in response to an infectious agents or their toxins.148 In support of this idea, one study reported that 10 of 20 (50%) infants with SIDS had levels of interleukin-6 (IL-6) in their cerebrospinal fl uid equivalent to those found in infants who died of infectious diseases.149 The common bacterium Staphylococcus aureus might have a role in this infection, since despite being isolated in 56% of healthy infants, 86% of SIDS infants had these bacteria in the respiratory tract.150 Some of the risk factors for SIDS, such as smoke exposure and lower socioeconomic status, parallel the risk for serious infection, particularly respiratory tract infections in infants.

In summary, the pathological fi ndings from research protocols support the theory that SIDS is associated with abnormalities or immaturity of the autonomic nervous system, immune system, and arousal pathways.

Genetic factorsCurrently, there are no specifi c genotypic diff erences in infants who died of SIDS that can be linked to specifi c clinically defi ned phenotypes. However, several dif-

Seminar


ferences in gene polymorphisms between infants with SIDS and control infants have been identifi ed. These include polymorphisms in sodium channel (SCN5A) and potassium channel genes that are associated with prolonged QT syndrome;151,152 ion channel defects are thought to be associated with up to 5–10% of SIDS cases.153,154 Polymorphisms have also been identifi ed in a serotonin transporter (5-HTT) gene that increases serotonin transporter activity, in turn reducing serotonin concentrations at nerve endings.155,156 Other poly-morphisms have been identifi ed in genes that aff ect autonomic nervous system development (PHOX2a, RET, ECE1, TLX3, EN1).157 Additionally, deletions in genes associated with infection have been reported. In a case-control study, infants who died from SIDS who had a mild upper respiratory tract infection before death were more likely to have deletion of the complement C4A or C4B gene than control infants.17 In SIDS infants, polymorphisms in the promoter region of the anti-infl ammatory cytokine interleukin 10 (IL10) could lead to decreases in antibody production or alternatively to increases in infl ammatory cytokine production.158 All of these fi ndings are consistent with post mortem and physiological fi ndings of abnormalities in the control of cardiorespiratory and immune function and arousal.

The triple risk model suggests that these and possibly other gene polymorphisms might make certain infants more vulnerable to SIDS; this vulnerability then manifests when there is an environmental challenge, such as prone positioning or tobacco exposure, at a critical age (fi gure). The interaction between genetic and environmental factors is thus an emerging area of research that needs to be further explored.

DiagnosisBy defi nition, SIDS is a diagnosis of exclusion. Protocols for standardised autopsies and death scene investigations in sudden unexpected infant deaths have been published.159–161 However, there is wide variability in both

the content and frequency with which these protocols are implemented across jurisdictions, both within countries and across diff erent countries. For example, autopsy rates in the Netherlands61 and Japan70 are lower than in most other developed countries. Diff erences could be a result of political, cultural, religious, and economic factors.

Although autopsies are crucial in identifying the cause of death in many kinds of sudden infant death, such as infections, congenital anomalies, and trauma, deaths from suff ocation are often diffi cult to diagnose from the autopsy alone. Thus, examination of the circumstances present immediately before death, including detailed description of the infant’s sleep environment, has become increasingly emphasised in recent years. Greater attention has also been paid to establishing the critical elements of these investigations and how to interpret the data collected. Analyses of trends in causes of sudden unexpected deaths in infancy indicate that a diagnostic shift might be taking place. Malloy and colleagues162 showed that during the period 1999–2001, the US postneonatal SIDS rate declined slightly, but the total postneonatal mortality rate did not change, and that the decrease in SIDS rate was largely accounted for by an increase in deaths due to suff ocation and cause unknown or unspecifi ed. Similar fi ndings were reported by Shapiro-Mendoza and colleagues163 for all infants from birth to 1 year of age, for the years 1989–2001. In a South Australian investigation, the number of SIDS deaths declined from 1989–93 to 1994–98, and the number of suff ocations in bed and cause undetermined deaths increased.164 A survey of medical examiners and coroners representing six US jurisdictions reported that “…almost all of the pathologists refl ected on their experiences with the changing SIDS diagnostic criteria and protocols from when they fi rst began practicing in the 1980s and discussed how much more complicated, confusing, and time-consuming SIDS cases have become. Most of them also noted that they used to label many more infant death cases as SIDS than they do now.”159

Part of the uncertainty in diagnosing SIDS seems to be the growing trend of confusing risk factors for the syndrome with causes of sudden unexpected infant deaths. For example, some pathologists might code the cause of death as positional asphyxia when an infant is found unresponsive in the face-down position, although this is a well-established risk factor for SIDS. Other pathologists might not code a death as SIDS if the infant had been sleeping with another person (bed sharing) at the presumed time of death.159

Several classifi cation schemes for sudden unexpected infant death have been developed as a way to standardise the assignment of cause of death on the basis of autopsy, scene investigation, and clinical information.165–67 Reaching consensus internationally on a classifi cation scheme is essential to accurately monitor trends in sudden infant death and to make appropriate use of data obtained through the autopsy and death scene investigation.

Environmental risk factors

Smoking

Genetic risk factors

Prone or side sleeping

Soft bedding

Prematurity

5-HTT polymorphism

Cardiac ion channel polymorphism

ANS polymorphism

Complement orinterleukin polymorphism

Impaired autonomicregulation and arousal

SIDS

Figure: Gene-environment interactionsAdapted from Hunt.153

Seminar


Risk reduction interventions and their eff ect Campaigns to reduce the risk of SIDS were initiated in the Netherlands in 1987; in the UK, New Zealand, and Australia in 1991, in the Scandinavian countries 1990–92, and in the USA in 1994.168 These campaigns largely focused on reducing prone sleeping. Initially, recommendations stressed side or back placement, but after new research identifi ed an increased risk with the side position com-pared with supine, subsequent recommendations include supine placement only.84 Some campaigns have also inc-luded messages about other behaviours and practices to reduce risk, such as reducing smoking during pregnancy and increasing breastfeeding, but there have been no signif-icant changes seen in these behaviours, and thus reduced SIDS rates have been attributed mainly to avoid ance of prone sleeping.169–171 Although there have been barriers to changing parental practices about infant sleep position, the prevalence of supine sleeping now approaches 100% in many countries or in subgroups within countries or regions.172 Individual behaviours, such as maternal smoking during pregnancy, are more diffi cult to change.

Research continues to seek factors associated with increased and decreased risk of SIDS to expand the eff ectiveness of risk reduction campaigns and to overcome the levelling in SIDS rates in many countries.173 Such eff orts are particularly salient for population groups disproportionately aff ected by SIDS, such as African Americans in the USA,174 Aboriginal Australians,10 and Maoris in New Zealand.9 In its most recent SIDS policy statement, the American Academy of Pediatrics reiterated some previous recommendations, revised others, and added new recommendations to potentially further reduce SIDS risk (panel),84 such as off ering a pacifi er to infants at bedtime and discouraging bed sharing while encouraging room sharing during infant sleep periods. These recommendations have met with resistance, mainly from breastfeeding advocates who worry that these measures will reduce breastfeeding frequency and duration and prevent families from enjoying the experience and benefi ts of bed sharing.175 The controversy is likely to continue for some time, until new research specifi cally targeted to these issues is done. Monitoring is needed of the adoption patterns of these recommendations among parents and health personnel and the eff ect they have on the SIDS rate and on other important outcomes, such as breastfeeding. We should also not lose sight of the more established and proven risk reduction measures, such as supine sleeping for all infants including those born preterm and avoidance of smoke exposure in utero, and to continually reinforce these measures to new generations of health providers, parents, and others who care for young children.

Management and supportThe loss of an infant is devastating for everyone concerned. However, in addition to the loss of their infant, families whose infant has died of SIDS could face police investigation, a long wait for autopsy results, and

continued uncertainty, leading to prolonged emotional distress, all of which complicate the grieving process. The physician can play an active part by advocating for an autopsy in all cases of sudden unexpected death, discussing the results of the autopsy with the family, and providing emotional support. Surviving siblings and other family members also need age-appropriate support. If appropriate, the family should be referred for genetic counselling, metabolic testing, or both. Additionally, the family should be directed to local counselling and support groups, which are available in most communities.

Future directions Despite declines in prevalence of SIDS, work still needs to be done on many fronts. Further refi nement in elucidation of the risk and protective factors, with appropriately targeted and implemented interventions leading to increased adoption by families, could bring the number of SIDS deaths to lower and lower levels. However, the disorder is unlikely to be completely eliminated or reduced to the lowest possible rates until the specifi c causative mechanisms are understood more fully. Continued research to identify the pathophysiology and genetics of SIDS must therefore be supported and expanded to include larger sample sizes of both aff ected and control infants, and infants from the highest risk groups.176 Additionally, systematic investigations of stillbirths and sudden unexplained deaths in those over 1 year of age are equally important and might provide additional insight into the pathogenesis of SIDS. New investigators, both in the basic and epidemiological sciences, need to be funded, mentored, and encouraged to enter the fi eld of SIDS research.

Panel: American Academy of Pediatrics SIDS risk reduction recommendations, 200584

• Back to Sleep: infants should be placed for sleep in a supine position for every sleep. Side sleeping is not as safe as supine sleeping and is not advised

• Use a fi rm sleep surface• Keep soft objects and loose bedding out of the crib• Do not smoke during pregnancy• A separate but proximate sleeping environment is recommended: the infant’s crib or

bassinet should be placed in the parents’ bedroom, which, when placed close to their bed, will allow for convenient breastfeeding and contact

• Consider off ering a pacifi er at nap time and bedtime: the pacifi er should be used when placing the infant down for sleep and not be reinserted once the infant falls asleep. If the infant refuses the pacifi er, he or she should not be forced to take it. For breastfed infants, delay pacifi er introduction until 1 month of age to ensure that breastfeeding is fi rmly established

• Avoid overheating• Avoid commercial devices marketed to reduce the risk of SIDS• Do not use home monitors as a strategy to reduce the risk of SIDS• Avoid development of positional plagiocephaly: encourage “tummy time” when the

infant is awake and observed. This will also enhance motor development. Avoid having the infant spend excessive time in car-seat carriers and “bouncers,” in which pressure is applied to the occiput. Upright “cuddle time” should be encouraged

• Continue the Back to Sleep campaign

Seminar


Surveillance of trends in rates of SIDS and other sudden unexpected infant deaths is crucial. Comparisons across jurisdictions within countries and across countries would be made more meaningful with standardised classifi cation guidelines. The international community needs to work together to either accept previously recommended approaches or develop a new one to achieve standardisation. Ideally, these changes would be incorporated into a new revision of the International Classifi cation of Diseases and thus refl ected on death certifi cates, the major way we have to establish trends in the causes of infant death internationally.

In conclusion, the multifactorial, complex, and changing knowledge base and approach to SIDS requires a multidisciplinary and collaborative eff ort that engages health professionals and policymakers, researchers, medical examiners and coroners, grief counsellors and agencies that provide support to families, and above all, families and communities, especially those at highest risk. Continued research, surveillance, risk reduction campaigns, and standardisation of autopsy and scene investigation protocols and classifi cation of deaths are all essential pieces to illuminating the SIDS puzzle and reaching our shared goal of eliminating it as a cause of infant death.Confl ict of interest statementRYM has received research grants to undertake studies of SIDS epidemiology and risk reduction interventions from the National Institutes of Health and the Gerber Foundation. RSCH’s salary is currently funded by a National Health and Medical Research Council of Australia Senior Research Fellowship. Funding for research has been awarded by the National Health and Medical Research Council of Australia, the Sudden Infant Death Research Foundation of South Australia and the National SIDS Council of Australia. FRH is currently funded by the National Institute of Child Health and Human Development and the University of Virginia Children’s Hospital. Both RYM and FRH have received speaker’s fees from various SIDS organisations.

References1 Willinger M, James LS, Catz C. Defi ning the sudden infant death

syndrome (SIDS): deliberations of an expert panel convened by the National Institute of Child Health and Human Development. Pediatr Pathol 1991; 11: 677–84.

2 Maternal and Child Health Statistics of Japan: Boshi Eisei Kenkyuu Kai, 2006.

3 Central Bureau of Statistics, Netherlands, 2006. http://www.cbs.nl/en-GB (accessed July 5, 2007).

4 New Zealand Health Information Service, 2003. http://www.nzhis.govt.nz (accessed July 5, 2007).

5 Hoyert DL, Mathews TJ, Menacker F, Strobino DM, Guyer B. Annual summary of vital statistics: 2004. Pediatrics 2006; 117: 168–83.

6 Report: Unexplained Deaths in Infancy, 2005. Health Stat Q 2006: 82–86.

7 Hoyert DL, Arias E, Smith BL, Murphy SL, Kochanek KD. Deaths: fi nal data for 1999. Natl Vital Stat Rep 2001; 49: 1–113.

8 Mathews TJ, Menacker F, MacDorman MF. Infant mortality statistics from the 2002 period: linked birth/infant death data set. Natl Vital Stat Rep 2004; 53: 1–29.

9 Mitchell EA, Tuohy PG, Brunt JM, et al. Risk factors for sudden infant death syndrome following the prevention campaign in New Zealand: a prospective study. Pediatrics 1997; 100: 835–40.

10 Freemantle CJ, Read AW, de Klerk NH, McAullay D, Anderson IP, Stanley FJ. Sudden infant death syndrome and unascertainable deaths: trends and disparities among Aboriginal and non-Aboriginal infants born in Western Australia from 1980 to 2001 inclusive. J Paediatr Child Health 2006; 42: 445–51.

11 Panaretto KS, Whitehall JF, McBride G, Patole S, Whitehall JS. Sudden infant death syndrome in Indigenous and non-Indigenous infants in north Queensland, 1990–1998. J Paediatr Child Health 2002; 38: 135–39.

12 Beal SM. Sudden infant death syndrome in south Australia 1968–97. Part 1: Changes over time. J Paediatr Child Health 2000; 36: 540–47.

13 Blakely T, Ajwani S, Robson B, Tobias M, Bonne M. Decades of disparity: widening ethnic mortality gaps from 1980 to 1999. N Z Med J 2004; 117: U995.

14 Perez-Stable EJ, Herrera B, Jacob P, 3rd, Benowitz NL. Nicotine metabolism and intake in black and white smokers. JAMA 1998; 280: 152–56.

15 National Infant Sleep Position study. http://dccwww.bumc.bu.edu/ChimeNisp/Main_Nisp.asp (accessed July 5, 2007).

16 Willinger M, Ko CW, Hoff man HJ, Kessler RC, Corwin MJ. Trends in infant bed sharing in the United States, 1993–2000: the National Infant Sleep Position study. Arch Pediatr Adolesc Med 2003; 157: 43–49.

17 Blackwell CC, Moscovis SM, Gordon AE, et al. Ethnicity, infection and sudden infant death syndrome. FEMS Immunol Med Microbiol 2004; 42: 53–65.

18 Tuohy PG, Smale P, Clements M. Ethnic diff erences in parent/infant co-sleeping practices in New Zealand. N Z Med J 1998; 111: 364–66.

19 Haglund B. Cigarette smoking and sudden infant death syndrome: some salient points in the debate. Acta Paediatr Suppl 1993; 82 (suppl 389): 37–39.

20 MacDorman MF, Cnattingius S, Hoff man HJ, Kramer MS, Haglund B. Sudden infant death syndrome and smoking in the United States and Sweden. Am J Epidemiol 1997; 146: 249–57.

21 Schoendorf KC, Kiely JL. Relationship of sudden infant death syndrome to maternal smoking during and after pregnancy. Pediatrics 1992; 90: 905–08.

22 Mitchell EA, Milerad J. Smoking and sudden infant death syndrome. WFO/TFI Reports 1999. http://www.who.int/tobacco/media/en/mitchell.pdf (accessed July 9, 2007).

23 DiFranza JR, Aligne CA, Weitzman M. Prenatal and postnatal environmental tobacco smoke exposure and children’s health. Pediatrics 2004; 113 (4 suppl): 1007–15.

24 Sovik S, Lossius K, Walloe L. Heart rate response to transient chemoreceptor stimulation in term infants is modifi ed by exposure to maternal smoking. Pediatr Res 2001; 49: 558–65.

25 Lewis KW, Bosque EM. Defi cient hypoxia awakening response in infants of smoking mothers: possible relationship to sudden infant death syndrome. J Pediatr 1995; 127: 691–99.

26 Froen JF, Akre H, Stray-Pedersen B, Saugsted OD. Adverse eff ects of nicotine and interleukin–1B on autoresucitation after apnea in piglets: implications for sudden infant death syndrome. Pediatrics 2000; 105: E52.

27 Browne CA, Colditz PB, Dunster KR. Infant autonomic function is altered by maternal smoking during pregnancy. Early Hum Develop 2000; 59: 209–18.

28 Chang AB, Wilson SJ, Masters IB, et al. Altered arousal response in infants exposed to cigarette smoke. Arch Dis Child 2003; 88: 30–33.

29 Franco P, Groswasser J, Hassid S, Lanquart JP, Scaillet S, Kahn A. Prenatal exposure to cigarette smoking is associated with a decrease in arousal in infants. J Pediatr 1999; 135: 34–38.

30 Horne RS, Franco P, Adamson TM, Groswasser J, Kahn A. Infl uences of maternal cigarette smoking on infant arousability. Early Hum Dev 2004; 79: 49–58.

31 Mitchell EA, Milerad J. Smoking and the sudden infant death syndrome. Rev Environ Health 2006; 21: 81–103.

32 Fleming PJ, Blair PS, Bacon C, et al. Environment of infants during sleep and risk of the sudden infant death syndrome: results of 1993–5 case-control study for confi dential inquiry into stillbirths and deaths in infancy. Confi dential Enquiry into Stillbirths and Deaths Regional Coordinators and Researchers. BMJ 1996; 313: 191–95.

33 Hauck FR, Moore CM, Herman SM, et al. The contribution of prone sleeping position to the racial disparity in sudden infant death syndrome: the Chicago Infant Mortality Study. Pediatrics 2002; 110: 772–80.

34 Li DK, Petitti DB, Willinger M, et al. Infant sleeping position and the risk of sudden infant death syndrome in California, 1997–2000. Am J Epidemiol 2003; 157: 446–55.

35 Oyen N, Markestad T, Skjaerven R, et al. Combined eff ects of sleeping position and prenatal risk factors in sudden infant death syndrome: the Nordic epidemiological SIDS study. Pediatrics 1997; 100: 613–21.

Seminar


36 Waters KA, Gonzalez A, Jean C, Morielle A, Brouillette RT. Face-straight-down and face-near-straight-down positions in healthy, prone-sleeping infants. J Pediatr 1996; 128: 616–25.

37 Willinger M, Hoff man HJ, Wu K-T, et al. Factors associated with the transition to nonprone sleep positions of infants in the United States: the National Infant Sleep Position Study. JAMA 1998; 280: 329–35.

38 Mitchell EA, Thach BT, Thompson JMD, Williams S. Changing infants’ sleep position increases risk of sudden infant death syndrome. Arch Pediatr Adolesc Med 1999; 153: 1136–41.

39 Brooke H, Gibson A, Tappin D, Brown H. Case-control study of sudden infant death syndrome in Scotland, 1992–5. BMJ 1997; 314: 1516–20.

40 Kemp JS, Kowalski RM, Burch PM, Graham MA, Thach BT. Unintentional suff ocation by rebreathing: a death scene and physiologic investigation of a possible cause of sudden infant death. J Pediatr 1993; 122: 874–80.

41 Mitchell EA, Thompson JMD, Ford RPK, Taylor BJ. Sheepskin bedding and the sudden infant death syndrome. New Zealand Cot Death Study Group. J Pediatr 1998; 133: 701–04.

42 Ponsonby A-L, Dwyer T, Couper D, Cochrane J. Association between use of a quilt and sudden infant death syndrome: case-control study. BMJ 1998; 316: 195–96.

43 Ponsonby A-L, Dwyer T, Gibbons LE, Cochrane JA, Wang Y-G. Factors potentiating the risk of sudden infant death syndrome associated with the prone position. N Engl J Med 1993; 329: 377–82.

44 Scheers NJ, Dayton CM, Kemp JS. Sudden infant death with external airways covered: case-comparison study of 206 deaths in the United States. Arch Pediatr Adolesc Med 1998; 152: 540–47.

45 Hauck FR, Herman SM, Donovan M, et al. Sleep environment and the risk of sudden infant death syndrome in an urban population: the Chicago Infant Mortality Study. Pediatrics 2003; 111: 1207–14.

46 Flick L, White DK, Vemulapalli C, Stulac BB, Kemp JS. Sleep position and the use of soft bedding during bed sharing among African American infants at increased risk for sudden infant death syndrome. J Pediatr 2001; 138: 338–43.

47 Kemp JS, Livne M, White DK, Arfken CL. Softness and potential to cause rebreathing: diff erences in bedding used by infants at high and low risk for sudden infant death syndrome. J Pediatr 1998; 132: 234–39.

48 Scheers NJ, Rutherford GW, Kemp JS. Where should infants sleep? A comparison of risk for suff ocation of infants sleeping in cribs, adult beds, and other sleeping locations. Pediatrics 2003; 112: 883–89.

49 Fleming P, Gilbert R, Azaz Y, et al. Interaction between bedding and sleeping position in the sudden infant death syndrome: a population based case-control study. BMJ 1990; 301: 85–89.

50 Ponsonby A-L, Dwyer T, Gibbons LE, Cochrane JA, Jones ME, McCall MJ. Thermal environment and sudden infant death syndrome: case-control study. BMJ 1992; 304: 277–82.

51 McKenna JJ, Mosko S, Richard CA. Bedsharing promotes breastfeeding. Pediatrics 1997; 100: 214–19.

52 Mosko S, Richard C, McKenna J. Infant arousals during mother-infant bed sharing: implications for infant sleep and sudden infant death syndrome research. Pediatrics 1997; 100: 841–49.

53 Carpenter RG, Irgens LM, Blair PS, et al. Sudden unexplained infant death in 20 regions in Europe: case control study. Lancet 2004; 363: 185–91.

54 Tappin D, Ecob R, Brooke H. Bedsharing and sudden infant death syndrome in Scotland: a case control study. J Pediatr 2005; 147: 32–37.

55 Blair PS, Fleming PJ, Smith IJ, et al. Babies sleeping with parents: case-control study of factors infl uencing the risk of the sudden infant death syndrome. CESDI SUDI research group. BMJ 1999; 319: 1457–62.

56 Scragg R, Mitchell EA, Taylor BJ, et al. Bed sharing, smoking, and alcohol in the sudden infant death syndrome. New Zealand Cot Death Study Group. BMJ 1993; 307: 1312–18.

57 Carpenter RG, Waite A, Coombs RC, et al. Repeat sudden unexpected and unexplained infant deaths: natural or unnatural? Lancet 2005; 365: 29–35.

58 McGarvey C, McDonnell M, Chong A, O’Regan M, Matthews T. Factors relating to the infant’s last sleep environment in sudden infant death syndrome in the Republic of Ireland. Arch Dis Child 2003; 88: 1058–64.

59 Arnestad M, Andersen M, Rognum TO. Is the use of dummy or carry-cot of importance for sudden infant death? Eur J Pediatr 1997; 156: 968–70.

60 Fleming PJ, Blair PS, Pollard K, et al. Pacifi er use and sudden infant death syndrome: results from the CESDI/SUDI case control study. CESDI SUDI Research Team. Arch Dis Child 1999; 81: 112–16.

61 L’Hoir MP, Engleberts AC, van Well GTJ, et al. Dummy use, thumb sucking, mouth breathing and cot death. Eur J Pediatr 1999; 158: 896–901.

62 Mitchell EA, Taylor BJ, Ford RPK, et al. Dummies and the sudden infant death syndrome. Arch Dis Child 1993; 68: 501–04.

63 Tappin D, Brooke H, Ecob R, Gibson A. Used infant mattresses and sudden infant death syndrome in Scotland: case-control study. BMJ 2002; 325: 1007–12.

64 Hauck FR, Omojokun OO, Siadaty MS. Do pacifi ers reduce the risk of sudden infant death syndrome? A meta-analysis. Pediatrics 2005; 116: e716–23.

65 Mitchell EA, Blair PS, L’Hoir MP. Should pacifi ers be recommended to prevent SIDS? Pediatrics 2006; 117: 1755–58.

66 Franco P, Scaillet S, Wermenbol V, Valente F, Groswasser J, Kahn A. The infl uence of a pacifi er on infants’ arousals from sleep. J Pediatr 2000; 136: 775–79.

67 Cozzi F, Albani R, E. C. A common pathophysiology for sudden cot death and sleep apnoea: “the facuum-glossoptosis syndrome”. Med Hypotheses 1979; 5: 329–38.

68 Engleberts A, L’Hoir MP. Pacifi er use and SIDS. Arch Dis Child 2000; 82: 267.

69 Weiss P, Kerbl R. The relatively short duration that a child retains a pacifi er in the mouth during sleep: implications for sudden infant death syndrome. Eur J Pediatr 2001; 160: 60–70.

70 Kahn A, Sawaguchi T, Sawaguchi A, et al. Sudden infant deaths: from epidemiology to physiology. Forensic Sci Int 2002; 130 (suppl): S8–20.

71 de Jonge GA, Lanting CI, Brand R, Ruys JH, Semmekrot BA, van Wouwe JP. Sudden infant death syndrome in child care settings in the Netherlands. Arch Dis Child 2004; 89: 427–30.

72 Moon RY, Patel KM, Shaefer SJ. Sudden infant death syndrome in child care settings. Pediatrics 2000; 106: 295–300.

73 Moon RY, Sprague BM, Patel KM. Stable prevalence but changing risk factors for sudden infant death syndrome in child care settings in 2001. Pediatrics 2005; 116: 972–77.

74 Hoff man HJ, Hillman LS. Epidemiology of the sudden infant death syndrome: maternal, neonatal, and postneonatal risk factors. Clin Perinatol 1992; 19: 717–37.

75 Malloy MH, Hoff man HJ. Prematurity, sudden infant death syndrome, and age of death. Pediatrics 1995; 96: 464–71.

76 Blair P, Ward Platt MP, Smith IJ, Fleming PJ. Sudden Infant Death Syndrome and sleeping position in pre-term and low birthweight infants: an opportunity for targeted intervention. Arch Dis Child 2005.

77 Blair PS, Ward Platt M, Smith IJ, Fleming PJ. Sudden infant death syndrome and sleeping position in pre-term and low birthweight infants: an opportunity for targeted intervention. Arch Dis Child 2006; 91: 101–06.

78 Baird TM, Paton JB, Fisher DE. Improved oxygenation with prone positioning in neonates: stability of increased transcutaneous pO2. J Perinatol 1991; 11: 315–17.

79 Martin RJ, DiFiore JM, Korenke CB, Randal H, Miller MJ, Brooks LJ. Vulnerability of respiratory control in healthy preterm infants placed supine. J Pediatr 1995; 127: 609–14.

80 Bhat RY, Hannam S, Pressler R, Raff erty GF, Peacock JL, Greenough A. Eff ect of prone and supine position on sleep, apneas, and arousal in preterm infants. Pediatrics 2006; 118: 101–07.

81 Ramanathan R, Corwin MJ, Hunt CE, et al. Cardiorespiratory events recorded on home monitors: comparison of healthy infants with those at increased risk for SIDS. JAMA 2001; 285: 2199–207.

82 American Academy of Pediatrics Committee on Fetus and Newborn, Apnea, sudden infant death syndrome, and home monitoring. Pediatrics 2003; 111: 914–17.

83 Heininger U, Kleemann WJ, Cherry JD. A controlled study of the relationship between Bordetella pertussis infections and sudden unexpected deaths among German infants. Pediatrics 2004; 114: e9–15.

84 Kattwinkel J, Hauck FR, Keenan ME, Malloy MH, Moon RY. Task Force on Sudden Infant Death Syndrome, American Academy of Pediatrics. The changing concept of sudden infant death syndrome: diagnostic coding shifts, controversies regarding the sleeping environment, and new variables to consider in reducing risk. Pediatrics 2005; 116: 1245–55.

Seminar


85 Vennemann MM, Butterfass-Bahloul T, Jorch G, et al. Sudden infant death syndrome: No increased risk after immunisation. Vaccine 2007; 25: 336–40.

86 Hoff man HJ, Damus K, Hillman L, Krongrad E. Risk factors for SIDS. Results of the National Institute of Child Health and Human Development SIDS Cooperative Epidemiological Study. Ann N Y Acad Sci 1988; 533: 13–30.

87 Ford RP, Taylor BJ, Mitchell EA, et al. Breastfeeding and the risk of sudden infant death syndrome. Int J Epidemiol 1993; 22: 885–90.

88 Alm B, Wennergren G, Norvenius SG, et al. Breast feeding and the sudden infant death syndrome in Scandinavia, 1992–95. Arch Dis Child 2002; 86: 400–02.

89 Kraus JF, Greenland S, Bulterys M. Risk factors for sudden infant death syndrome in the US Collaborative Perinatal Project. Int J Epidemiol 1989; 18: 113–20.

90 Ponsonby AL, Dwyer T, Kasl SV, Cochrane JA. The Tasmanian SIDS Case-Control Study: univariable and multivariable risk factor analysis. Paediatr Perinat Epidemiol 1995; 9: 256–72.

91 Gilbert RE, Wigfi eld RE, Fleming PJ, Berry PJ, Rudd PT. Bottle feeding and the sudden infant death syndrome. BMJ 1995; 310: 88–90.

92 Fleming P, Bacon CJ, Blair P, Berry PJ. Sudden unexpected deaths in infancy. The CESDI SUDI studies 1993–1996. London: Stationery Offi ce, 1999.

93 Vennemann MM, Bajanowski T, Butterfass-Bahloul T, et al. Do risk factors diff er between explained sudden unexpected death in infancy (SUDI) and SIDS? Arch Dis Child 2007; 92: 133–36.

94 Vennemann MM, Findeisen M, Butterfass-Bahloul T, et al. Modifi able risk factors for SIDS in Germany: results of GeSID. Acta Paediatr 2005; 94: 655–60.

95 Chen A, Rogan WJ. Breastfeeding and the risk of postneonatal death in the United States. Pediatrics 2004; 113: e435–39.

96 Horne RS, Parslow PM, Ferens D, Watts AM, Adamson TM. Comparison of evoked arousability in breast and formula fed infants. Arch Dis Child 2004; 89: 22–25.

97 Byard RW, Makrides M, Need M, Neumann MA, Gibson RA. Sudden infant death syndrome: eff ect of breast and formula feeding on frontal cortex and brainstem lipid composition. J Paediatr Child Health 1995; 31: 14–16.

98 Bjerve KS, Brekke OL, Bonaa K, Johnsen H, Vatten L, Vik T. Clinical and epidemiological studies with alpha-linolenic and long chain omega-3 fatty acids. In: Sinclair AJ, Gibson RA, eds. Essential fatty acids and eicosanoids: invited papers from the Third International Congress. Champaign, IL: American Oil Chemists Society, 1992.

99 Cheruku SR, Montgomery-Downs HE, Farkas SL, Thoman EB, Lammi-Keefe CJ. Higher maternal plasma docosahexaenoic acid during pregnancy is associated with more mature neonatal sleep-state patterning. Am J Clin Nutr 2002; 76: 608–13.

100 Hunt CE, Hauck FR. Sudden infant death syndrome. CMAJ 2006; 174: 1861–69.

101 American Academy of Pediatrics: Distinguishing sudden infant death syndrome from child abuse fatalities. Pediatrics 2001; 107: 437–41.

102 Filiano JJ, Kinney HC. A perspective on neuropathologic fi ndings in victims of the sudden infant death syndrome: The triple-risk model. Biol Neonate 1994; 65: 194–97.

103 Kahn A, Riazi J, Blum D. Oculocardiac refl ex in near miss for sudden infant death syndrome infants. Pediatrics 1983; 71: 49–52.

104 Kelly DH, Golub H, Carley D, Shannon DC. Pneumograms in infants who subsequently died of sudden infant death syndrome. J Pediatr 1986; 109: 249–54.

105 Schechtman VL, Harper RM, Kluge KA, Wilson AJ, Hoff man HJ, Southall DP. Heart rate variation in normal infants and victims of the sudden infant death syndrome. Early Hum Dev 1989; 19: 167–81.

106 Schwartz PJ, Stramba-Badiale M, Segantini A, et al. Prolongation of the QT interval and the sudden infant death syndrome. N Engl J Med 1998; 338: 1709–14.

107 Kluge KA, Harper RM, Schechtman VL, Wilson AJ, Hoff man HJ, Southall DP. Spectral analysis assessment of respiratory sinus arrhythmia in normal infants and infants who subsequently died of sudden infant death syndrome. Pediatr Res 1988; 24: 677–82.

108 Franco P, Verheulpen D, Valente F, et al. Autonomic responses to sighs in healthy infants and in victims of sudden infant death. Sleep Med 2003; 4: 569–77.

109 Phillipson EA, Sullivan CC. Arousal: the forgotten response to respiratory stimuli. Am Rev Respir Dis 1978; 118: 807–09.

110 Horner RL. Autonomic consequences of arousal from sleep: mechanisms and implications. Sleep 1996; 19 (10 suppl): S193–95.

111 Fewell JE, Kondo CS, Dascalu V, Filyk SC. Infl uence of carotid-denervation on the arousal and cardiopulmonary responses to alveolar hypercapnia in lambs. J Dev Physiol 1989; 12: 193–99.

112 Horne RS, Berger PJ, Bowes G, Walker AM. Eff ect of sinoaortic denervation on arousal responses to hypotension in newborn lambs. Am J Physiol 1989; 256: H434–40.

113 Horne RS, De Preu ND, Berger PJ, Walker AM. Arousal responses to hypertension in lambs: eff ect of sinoaortic denervation. Am J Physiol 1991; 260: H1283–89.

114 Kahn A, Groswasser J, Rebuff at E, et al. Sleep and cardiorespiratory characteristics of infant victims of sudden death: a prospective case-control study. Sleep 1992; 15: 287–92.

115 Schechtman VL, Harper RM, Wilson AJ, Southall DP. Sleep state organization in normal infants and victims of the sudden infant death syndrome. Pediatrics 1992; 89: 865–70.

116 Harper RM. State-related physiological changes and risk for the sudden infant death syndrome. Aust Paediatr J 1986; 22 (suppl 1): 55–58.

117 Kato I, Franco P, Groswasser J, et al. Incomplete arousal processes in infants who were victims of sudden death. Am J Respir Crit Care Med 2003; 168: 1298–303.

118 Kahn A, Groswasser J, Sottiaux M, Rebuff at E, Franco P, Dramaix M. Prone or supine body position and sleep characteristics in infants. Pediatr 1993; 91: 1112–15.

119 Myers MM, Fifer WP, Schaeff er L, et al. Eff ects of sleeping position and time after feeding on the organization of sleep/wake states in prematurely born infants. Sleep 1998; 21: 343–49.

120 Franco P, Groswasser J, Sottiaux M, Broadfi eld E, Kahn A. Decreased cardiac resonses to auditory stimulation during prone sleep. Pediatrics 1996; 97: 174–78.

121 Horne RS, Ferens D, Watts AM, et al. The prone sleeping position impairs arousability in term infants. J Pediatr 2001; 138: 811–16.

122 Ariagno RL, van Liempt S, Mirmiran M. Fewer spontaneous arousals during prone sleep in preterm infants at 1 and 3 months corrected age. J Perinatol 2006; 26: 306–12.

123 Galland BC, Reeves G, Taylor BJ, Bolton DP. Sleep position, autonomic function, and arousal. Arch Dis Child Fetal Neonatal Ed 1998; 78: F189–94.

124 Galland BC, Hayman RM, Taylor BJ, Bolton DP, Sayers RM, Williams SM. Factors aff ecting heart rate variability and heart rate responses to tilting in infants aged 1 and 3 months. Pediatr Res 2000; 48: 360–68.

125 Horne RS, Bandopadhayay P, Vitkovic J, Cranage SM, Adamson TM. Eff ects of age and sleeping position on arousal from sleep in preterm infants. Sleep 2002; 25: 746–50.

126 Kato I, Scaillet S, Groswasser J, et al. Spontaneous arousability in prone and supine position in healthy infants. Sleep 2006; 29: 785–90.

127 Skadberg BT, Markestad T. Behaviour and physiological responses during prone and supine sleep in early infancy. Arch Dis Child 1997; 76: 320–24.

128 Sahni R, Schulze KF, Kashyap S, Ohira-Kist K, Fifer WP, Myers MM. Postural diff erences in cardiac dynamics during quiet and active sleep in low birthweight infants. Acta Paediatr 1999; 88: 1396–401.

129 Chong A, Murphy N, Matthews T. Eff ect of prone sleeping on circulatory control in infants. Arch Dis Child 2000; 82: 253–56.

130 Galland BC, Bolton DP, Taylor BJ, Sayers RM, Williams SM. Ventilatory sensitivity to mild asphyxia: prone versus supine sleep position. Arch Dis Child 2000; 83: 423–28.

131 Tuladhar R, Harding R, Cranage SM, Adamson TM, Horne RS. Eff ects of sleep position, sleep state and age on heart rate responses following provoked arousal in term infants. Early Hum Dev 2003; 71: 157–69.

132 Tuladhar R, Harding R, Adamson TM, Horne RS. Heart rate responses to non-arousing trigeminal stimulation in infants: eff ects of sleep position, sleep state and postnatal age. Early Hum Dev 2005; 81: 673–81.

133 Galland BC, Taylor BJ, Bolton DP, Sayers RM. Heart rate variability and cardiac refl exes in small for gestational age infants. J Appl Physiol 2006; 100: 933–39.

Seminar


134 Ariagno RL, Mirmiran M, Adams MM, Saporito AG, Dubin AM, Baldwin RB. Eff ect of position on sleep, heart rate variability, and QT interval in preterm infants at 1 and 3 months’ corrected age. Pediatrics 2003; 111: 622–25.

135 Hunt CE. The cardiorespiratory control hypothesis for sudden infant death syndrome. Clin Perinatol 1992; 19: 757–71.

136 Tirosh E, Libon D, Bader D. The eff ect of maternal smoking during pregnancy on sleep respiratory and arousal patterns in neonates. J Perinatol 1996; 16: 435–38.

137 Horne RS, Ferens D, Watts AM, et al. Eff ects of maternal tobacco smoking, sleeping position, and sleep state on arousal in healthy term infants. Arch Dis Child Fetal Neonatal Ed 2002; 87: F100–05.

138 Sawnani H, Jackson T, Murphy T, Beckerman R, Simakajornboon N. The eff ect of maternal smoking on respiratory and arousal patterns in preterm infants during sleep. Am J Respir Crit Care Med 2004; 169: 733–38.

139 Parslow PM, Cranage SM, Adamson TM, Harding R, Horne RS. Arousal and ventilatory responses to hypoxia in sleeping infants: eff ects of maternal smoking. Respir Physiol Neurobiol 2004; 140: 77–87.

140 Franco P, Chabanski S, Szliwowski H, Dramaix M, Kahn A. Infl uence of maternal smoking on autonomic nervous system in healthy infants. Pediatr Res 2000; 47: 215–20.

141 Byard RW, Krous HF. Sudden infant death syndrome: overview and update. Pediatr Dev Pathol 2003; 6: 112–27.

142 Xu F, Severinghaus JW. Rat brain VEGF expression in alveolar hypoxia: possible role in high-altitude cerebral edema. J Appl Physiol 1998; 85: 53–57.

143 Jones KL, Krous HF, Nadeau J, Blackbourne B, Zielke HR, Gozal D. Vascular endothelial growth factor in the cerebrospinal fl uid of infants who died of sudden infant death syndrome: evidence for antecedent hypoxia. Pediatrics 2003; 111: 358–63.

144 Kinney HC, Filiano JJ, White WF. Medullary serotonergic network defi ciency in the sudden infant death syndrome: review of a 15-year study of a single dataset. J Neuropathol Exper Neurol 2001; 60: 228–47.

145 Paterson DS, Trachtenberg FL, Thompson EG, et al. Multiple serotonergic brainstem abnormalities in sudden infant death syndrome. JAMA 2006; 296: 2124–32.

146 Lavezzi AM, Ottaviani G, Mauri M, Matturri L. Alterations of biological features of the cerebellum in sudden perinatal and infant death. Curr Mol Med 2006; 6: 429–35.

147 Waters KA, Meehan B, Huang JQ, Gravel RA, Michaud J, Cote A. Neuronal apoptosis in sudden infant death syndrome. Pediatr Res 1999; 45: 166–72.

148 Blackwell CC, Moscovis SM, Gordon AE, et al. Cytokine responses and sudden infant death syndrome: genetic, developmental, and environmental risk factors. J Leukoc Biol 2005; 78: 1242–54.

149 Vege A, Rognum TO, Scott H, Aasen AO, Saugstad OD. SIDS cases have increased levels of interleukin-6 in cerebrospinal fl uid. Acta Paediatr 1995; 84: 193–96.

150 Blackwell CC, MacKenzie DA, James VS, et al. Toxigenic bacteria and sudden infant death syndrome (SIDS): nasopharyngeal fl ora during the fi rst year of life. FEMS Immunol Med Microbiol 1999; 25: 51–58.

151 Ackerman MJ, Siu BL, Sturner WQ, et al. Postmortem molecular analysis of SCN5A defects in sudden infant death syndrome. JAMA 2001; 286: 2264–69.

152 Plant LD, Bowers PN, Liu Q, et al. A common cardiac sodium channel variant associated with sudden infant death in African Americans, SCN5A S1103Y. J Clin Invest 2006; 116: 430–35.

153 Hunt CE. Gene-environment interactions: implications for sudden unexpected deaths in infancy. Arch Dis Child 2005; 90: 48–53.

154 Tester DJ, Ackerman MJ. Sudden infant death syndrome: how signifi cant are the cardiac channelopathies? Cardiovasc Res 2005; 67: 388–96.

155 Narita N, Narita M, Takashima S, Nakayama M, Nagai T, Okado N. Serotonin transporter gene variation is a risk factor for sudden infant death syndrome in the Japanese population. Pediatrics 2001; 107: 690–92.

156 Weese-Mayer DE, Berr-Kravis EM, Maher BS, Silvestri JM, Curran ME, Marazita ML. Sudden infant death syndrome: association with a promoter polymorphism of the serotonin transporter gene. Am J Med Genet 2003; 117A: 268–74.

157 Weese-Mayer DE, Berry-Kravis EM, Zhou L, et al. Sudden infant death syndrome: case-control frequency diff erences at genes pertinent to early autonomic nervous system embryologic development. Pediatr Res 2004; 56: 391–95.

158 Korachi M, Pravica V, Barson AJ, Hutchinson IV, Drucker DB. Interleukin 10 genotype as a risk factor for sudden infant death syndrome: determination of IL–10 genotype from wax-embedded postmortem samples. FEMS Immunol Med Microbiol 2004; 42: 125–29.

159 Maternal and Child Health Bureau, Health Resources and Services Administration (HRSA), Variations in SIDS diagnosis procedures. Rockville, MD: Health Resources and Services Administration, 2003.

160 Krous HF. Instruction and reference manual for the International Standardized Autopsy Protocol for Sudden Unexpected Infant Death. J SIDS Infant Mortality 1996; 1: 203–46.

161 Centers for Disease Control and Prevention. Guidelines for death scene investigation of sudden, unexplained infant deaths: recommendations of the interagency panel on sudden infant death syndrome. MMWR 1996; 45 (RR–10).

162 Malloy MH. Trends in postneonatal aspiration deaths and reclassifi cation of sudden infant death syndrome: impact of the “Back to Sleep” program. Pediatrics 2002; 109: 661–65.

163 Shapiro–Mendoza CK, Tomashek KM, Anderson RN, Wingo J. Recent national trends in sudden, unexpected infant deaths: more evidence supporting a change in classifi cation or reporting. Am J Epidemiol 2006; 163: 762–69.

164 Mitchell E, Krous HF, Donald T, Byard RW. Changing trends in the diagnosis of sudden infant death. Am J Forensic Med Pathol 2000; 21: 311–14.

165 Krous HF, Beckwith JB, Byard RW, et al. Sudden infant death syndrome and unclassifi ed sudden infant deaths: a defi nitional and diagnostic approach. Pediatrics 2004; 114: 234–38.

166 Beckwith JB. Defi ning the sudden infant death syndrome. Arch Pediatr Adolesc Med 2003; 157: 286–90.

167 Rognum TO. [Do we need a new defi nition of crib death?] (in Norwegian). Tidsskr Nor Laegeforen 1995; 115: 85–88.

168 Hauck FR. Changing epidemiology. In: Byard RW, Krous HF, eds. Sudden infant death syndrome: problems, progress and possibilities. London: Arnold Publishers, 2001.

169 Dwyer T, Ponsonby A-L. SIDS epidemiology and incidence. Pediatr Ann 1995; 24: 350–56.

170 Markestad T, Skadberg B, Hordvik E, Morild I, Irgens L. Sleeping position and sudden infant death syndrome (SIDS): eff ect of an intervention programme to avoid prone sleeping. Acta Paediatr 1995; 84: 375–78.

171 Mitchell EA, Brunt JM, Everard C. Reduction in mortality from sudden infant death syndrome in New Zealand: 1986–92. Arch Dis Child 1994; 70: 291–94.

172 Wennergren G, Alm B, Oyen N, et al. The decline in the incidence of SIDS in Scandinavia and its relation to risk-intervention campaigns. Nordic Epidemiological SIDS Study. Acta Paediatr 1997; 86: 963–68.

173 Creery D, Mikrogianakis A. Sudden infant death syndrome. BMJ Clinical Evidence 2005; 13: 434–443.

174 Mathews TJ, MacDorman MF. Infant mortality statistics from the 2003 period linked birth/infant death data set. Natl Vital Stat Rep 2006; 54: 1–29.

175 Fleming P, Blair P, McKenna J. New knowledge, new insights, and new recommendations. Arch Dis Child 2006; 91: 799–801.

176 Weese-Mayer DE. Sudden infant death syndrome: is serotonin the key factor? JAMA 2006; 296: 2143–44.

Case Report


Orange and green monkeys jumping around the roomSunku Hemanth Guptha, Tha Han, Qais Arafat, Okubadejo Deyo

In July, 2006, a 62-year-old solicitor, with a 14-year history of Parkinson’s disease, began to have worse tremor, stiff -ness, and fatigue, and reduced motor control. His illness had previously been well controlled by co-careldopa and selegiline. Ropirinole was introduced, and his symp toms decreased as the dose was increased. However, after 2 months, when the dose was 12 mg daily, he started to experience visual hallucinations of orange and green monkeys jumping around his room, and men in green uniforms marching around his bed. He also had vivid dreams in which he fought gory battles with soldiers, or was tortured by creatures resembling aliens. His partner reported that he was shouting and agitated in his sleep: once, he violently attacked his pillow. He developed per-secutory delusions, with a constant urge to move his belong ings and furniture around the house, to protect them from strangers; when his partner tried to persuade him not to do so, he became verbally aggressive. He had never used illicit drugs, or used alcohol harmfully; he had no previous or family history of psychiatric illness. 3 weeks after the onset of hallucinations, the patient became dis oriented in space and time, so we admitted him to hospital.

The patient had an asymmetrical resting tremor, rigidity, bradykinesia, and shuffl ing gait—but no other abnor-malities were found on examination. Blood tests, including a full blood count, ESR, standard biochemistry screen, thyroid function tests, and concentrations of glucose, C-reactive protein, vitamin B12, and folic acid gave normal results—as did chest radiography, urinalysis, and an electrocardiogram. The patient was thought to have a iatrogenic psychosis. Selegiline was discontinued immediately; 7 days later, ropirinole was discontinued. 2 weeks

into his admission, the patient realised that his hallucinations did not represent reality, and his aggression subsided. He tended to be lucid and independent in the mornings, but disoriented and in need of continuous supervision, even to use the toilet, in the evenings. In the lucid periods, he scored 10 out of 10 on the abbreviated mental test score. His motor symptoms fl uctuated similarly, with recurrent falls in the evenings. There was no evidence of multisystem atrophy. Serial testing of visual fi elds and acuity did not reveal any defi cits. Risperidone, at a dose of 500 μg twice daily, did not diminish the hallucinations. The patient therefore underwent MRI of the brain, which showed a butterfl y glioblastoma multi forme in the occipital lobes and right temporal lobe, extending up to the right side of the thalamus (fi gure). The tumour was so extensive that surgical treatment was not considered appropriate. Treatment with dexamethasone was associated with a reduction in frequency of the hallucinations. The patient developed signs of raised intra cranial pressure. In November, 2006, he was dis-charged to a nursing home, where he died within a week.

About a quarter of patients with Parkinson’s disease have visual hallucinations, which are transient, often show human beings and animals, and may be accompanied by sleep disturbances.1 Risk factors for visual hallucinations in Parkinson’s disease include a long duration of illness, visual impairment, cognitive impairment, and the use of dopaminergic drugs.1–3 Visual hallucinations are also a typical symptom of Lewy-body dementia, which can cause parkinsonism and is pathologically on a continuum with Parkinson’s disease. Other causes of complex visual hallucinations in elderly people include Charles Bonnet syndrome; disruption of the visual pathway, when hallu-cinations are localised to the aff ected area; recreational drug use; alcohol intoxication; schizophrenia, notably very late-onset schizophrenia; and, rarely, epilepsy, migraine, peduncular hallucinosis, and posterior circulation stroke.4 Our patient’s visual hallucinations resembled those of ped-un cular hallucinosis—in which the lesion is not necessarily in the cerebral peduncles, but typically in the rostral brainstem (particularly the midbrain) or the thalamus.4 The normal fi ndings on repeated neurological assess ment emphasise that the standard neurological examin ation does not adequately assess all parts of the brain. References1 Fénelon G, Mahieux F, Huon R, Ziégler M. Hallucinations in

Parkinson’s disease: prevalence, phenomenology and risk factors. Brain 2000; 123: 733–45.

2 Sanchez-Ramos JR, Ortoll R, Paulson GW. Visual hallucinations associated with Parkinson disease. Arch Neurol 1996; 53: 1265–68.

3 Matsui H, Udaka F, Tamura A, et al. Impaired visual acuity as a risk factor for visual hallucinations in Parkinson’s disease. J Geriatr Psychiatry Neurol 2006; 19: 36–40.

4 Manford M, Andermann F. Complex visual hallucinations. Clinical and neurobiological insights. Brain 1998; 121: 1819–40.

Lancet 2007; 370: 1588

Department of Medicine for Older People, Edith Cavell

Hospital (S H Guptha MRCP, T Han MB, O Deyo FRCP), and

Department of Radiology, Peterborough and Stamford

NHS Trust (Q Arafat FRCR), Peterborough, UK

Correspondence to:Dr Sunku Hemanth Guptha,

Edith Cavell Hospital, Bretton Gate, Peterborough PE3 9GZ, UK

[email protected]

Figure: MRI of the brain, showing the tumour

Documents

(3 November 2007-9 November 2007)))lib.ajaums.ac.ir/booklist/986962.pdfThe Lancet Volume 370, Issue 9598, Pages 1521-1588 (3 November 2007-9 November 2007))) 1. Time to supersize control