Repeat sudden unexpected infant deaths

Abstract

The series of repeat sudden infant deaths presented by Robert Carpenter and colleagues (Jan 1, p 29)1Carpenter RG Waite A Coombs RC et al.Repeat sudden unexpected and unexplained infant deaths: natural or unnatural?.Lancet. 2005; 365: 29-35Summary Full Text Full Text PDF PubMed Scopus (49) Google Scholar is the largest yet reported, and the data are especially valuable in an area where there is a dearth of good evidence. In reaching their conclusions, however, the authors may not have taken sufficient account of two possible sources of bias. First, self-selection at two stages raises the possibility that the sample studied was not a true cross-section of sudden infant deaths. The Care of the Next Infant (CONI) programme, from which the cases were drawn, is voluntary, and the numbers and characteristics of non-participants are unknown. Then, of the 46 participating families who had a repeat unexpected death, 13 (28%) were not subject to full enquiry, seven because they declined. Carpenter and colleagues acknowledge these difficulties but do not adequately allow for them. The second and more serious problem is the scope for bias in decisions as to whether a death was natural or unnatural. Recognition of covert homicide in infancy is notoriously difficult: the history may be fictitious and the autopsy unhelpful, and there are no agreed diagnostic criteria.2Levene S Bacon CJ Sudden unexpected death and covert homicide in infancy.Arch Dis Child. 2004; 89: 443-447Crossref PubMed Scopus (59) Google Scholar Thus, judgments have to be largely subjective, and sometimes it is simply not possible to be certain. Several of the cases attributed to sudden infant death syndrome (SIDS) in this paper have features that, in the context of two unexplained deaths, must give rise to concern: violence in the family (nine cases), parental mental health problems (six), preceding apparent life-threatening events (three), post-mortem findings suggestive of asphyxia (nine), abuse of a previous child (one), open coroner's verdict (three), and professional concern about the safety of other children (four). Instead of dichotomising the cases into natural or unnatural, it would have been more accurate (although less tidy) to have a third category—a grey area in which it was not possible to be certain. The potential fallibility of the categorisations in this study is illustrated by the initial designation of two deaths in one family as natural, with a later change to unnatural when the father admitted killing a third baby. The investigators seem to have taken the benign view that a death should be classified as natural unless there was compelling evidence to the contrary—an approach that is appropriate in the courts but not in scientific debate. Distinction between the legal and scientific arenas is crucial. In law, a person is presumed innocent until proved guilty beyond reasonable doubt. In scientific debate, presumption and benignity have no place: all possibilities compatible with the evidence have to be considered dispassionately. This paper will be influential, but those quoting it should be aware that its data do not support such clear-cut conclusions. Until we have more reliable means of identifying covert homicide, we cannot be certain of the proportion of repeat sudden infant deaths that is unnatural. It may well be rather higher than this analysis suggests. I declare that I have no conflict of interest. Repeat sudden unexpected infant deathsAuthors' reply Full-Text PDF