Letters

Risk of adverse birth outcomes near landfill sites

Evidence suggests that it is probably safe for fetuses to develop near landfill sites

Risks from landfill sites can be presented in alternative ways

News stories were handled badly

Local registers provide more accurate information

Evidence suggests that it is probably safe for fetuses to develop near landfill sites

EDITORElliott et al report a large geographical study of adverse birth outcomes in populations living near landfill sites.¹ They conclude that there are small excess risks of congenital anomalies and low birth weight in such populations.

Although they advise caution when interpreting their results, the study is nevertheless hailed by the press and by environmental groups as evidence that living near such sites is hazardous to health.² The concerns felt by parents, often with no opportunity to move elsewhere, are fuelled, and perhaps on no grounds. Scientists contribute to the media generated intrigue in subtle ways. For example, the description of the paper in an editorial by McNamee and Dolk as a report "on the risks to fetuses associated with residence" when "a study of the statistical association between reported anomalies and residence" would have been more accurate and less sensationalist.³ Health authorities are left to pick up the pieces when a story breaks, with no advance warning, about a landfill site in their area.

Another interpretation of the results of Elliott et al is that this essentially negative study goes some way to reassuring other scientists who have been studying this question that it is probably safe for fetuses to develop near such sites. Any scientist who has any experience of multivariate epidemiological analysis would have to admit that such small and inconsistent excesses and deficits, even though significant, could well be attributable to inadequately adjusted social deprivation, to name but one of many important confounders. The fact that there was no increase after the landfill sites became operational, and a significant deficit of anomalies around landfill sites in Scotland, must leave even the most determined author feeling uneasy about suggesting a causal association. Epidemiological techniques used in this study are blunt instruments unlikely to detect subtle or unidentified effects.

Perhaps it is time for epidemiologists to admit that their tools are inadequate to answer this question. Other scientists should take this forward by developing better techniques to identify any putative agents in the gas emissions, leachates, and soil that may be of sufficiently high concentrations to affect human health.

Many of these landfill sites are unattractive and foul smelling, attracting seagulls and peppering the landscape with rubbish. More imaginative long term waste-management solutions are long overdue. Consideration should be given to refusing planning permission for new build housing adjacent to landfill sites.⁴

Helene Irvine, consultant in public health medicine (communicable disease and environmental health). 
helene.irvine{at}gghb.scot.nhs.uk

Harry Burns, director of public health. 
Greater Glasgow NHS Board, Glasgow G3 8YU

Risks from landfill sites can be presented in alternative ways

EDITORThe presentation of risks in relative terms does not help individuals, or the population, to understand the risk to themselves. Let us assume that the risks identified by Elliott et al are not due to unmeasured confounding.¹

By taking the absolute risk in the reference area into account, we can translate the relative risk of 1.01 for all anomalies combined to one extra anomaly among 5903 who were exposed to living within 2 km of a landfill site. This can be set against the baseline population risk of one anomaly among each 59 individuals. The relative risk of 1.05 for neural tube defects can similarly be expressed as one extra defect among 35 714 exposed people, against the baseline risk of one per 1786 in the general population. Extending this further, we can calculate the number of the total population among whom one extra anomaly will be produced by the presence of landfill sites (to which 55% of the population are exposed within 2 km)this is termed the population impact number.² For all anomalies it is 10 733 and for spina bifida it is 64 935.

Should the evidence be thought to be robust enough, the individual and the community can use terms beyond the relative risk to help with decision making.

Dick Heller, professor of public health. 
Evidence for Population Health Unit, School of Epidemiology and Health Sciences, University of Manchester, Manchester M13 9PT dick.heller{at}man.ac.uk

News stories were handled badly

EDITOREditor's choice of 18 August refers to the difficulty of transmitting medical stories to the general press.¹ The potential press interest in the article by Elliott et al on birth defects and landfill sites is noted, and a thoughtful and balanced editorial by McNamee and Dolk accompanies the paper. ^{2 3} The relevant paragraph in This week in the BMJ then carries the headline: "People living near landfill sites have an increased risk of adverse birth outcomes." The lay press has, of course, not been slow to pick up this headlinesee, for example, the front page of the Guardian on 18 August.⁴

The whole point of the editorial by McNamee and Dolk is that we do not know whether living near a landfill site is a risk for birth defects or notthe study by Elliott et al is very important but requires further investigation. What hope is there of curtailing the media's obsession with medical scare stories if even a respected medical journal is prepared to highlight a story by using a sensational headline?

Roger A Fisken, consultant physician. 
Friarage Hospital, Northallerton, North Yorkshire DL6 1JG daardon{at}yahoo.com

Local registers provide more accurate information

EDITORElliott et al in their investigation of the risk of adverse outcome associated with residence near landfill sites conclude that one explanation for the finding of a small excess risk of congenital anomalies and low and very low birthweight babies, is a data artefact.¹ Poor ascertainment and low levels of reporting to the national congenital anomaly system at the Office for National Statistics in England and Wales is, as Elliott et al say, well recognised.

Around 45% of births in England and all in Wales are now covered by local congenital anomaly registers, some of which exchange data electronically with the Office for National Statistics. These registers, which actively search for cases and have multisource ascertainment, provide much more accurate data than routine returns of form SD56 to the office. In Oxfordshire we have shown that many serious anomalies, such as heart defects, do not get reported but minor ones, such as clicking hip (on the office's exclusion list), do. In 1999 less than 30% of registerable anomalies occurring around Oxford were notified to the Office for National Statistics. Our local health authority seems unconcerned by this because reporting congenital anomalies is not a statutory requirement and therefore, it seems, not a priority.

The findings of the study by Elliott et al may well cause pregnant women to become more anxious. It is too easy to report that data may be flawed. Elliott et al would have been well advised to restrict their analysis to those areas served by recognised registers. We do pregnant women a disservice by raising concerns from studies using flawed data (garbage in, garbage out). The patchy reporting of anomalies is now being highlighted by the British Isles Network of Congenital Anomaly Registers (BINOCAR), with the aim of improving surveillance of congenital anomalies for the whole of England and Wales. This is crucial not only for investigating potential teratogens but also for monitoring new prenatal screening and diagnostic tests.

Patricia A Boyd, clinical geneticist. 
Paul F Chamberlain, consultant obstetrician. 
Oxford Prenatal Diagnosis Unit, Women's Centre, Radcliffe Hospital, Oxford OX3 9DU

Patricia.boyd{at}orh.nhs.uk