Elsevier

The Lancet

Volume 365, Issue 9453, 1 January 2005, Pages 29-35
The Lancet

Articles
Repeat sudden unexpected and unexplained infant deaths: natural or unnatural?

https://doi.org/10.1016/S0140-6736(04)17662-9Get rights and content

Summary

Background

There have been suggestions that when two or three unexpected unexplained infant deaths occur within a family they are more likely to be unnatural than natural. We aimed to estimate the probability that a second infant death is natural versus unnatural.

Methods

The Care of Next Infant programme (CONI) supports parents who have previously had an unexpected and apparently unexplained infant death and is currently available in over 90% of health districts in England, Wales, and Northern Ireland. We studied all deaths in 6373 infants who had completed the CONI programme by December, 1999. After a CONI death, we made detailed enquiries into the previous death and the CONI death, including a family interview, a review of autopsies, and case discussion.

Findings

57 (8·9 per 1000) CONI infants died. Nine deaths were inevitable, and 48 were unexpected. 44 families lost one child, and two families lost two children. Of the 46 first CONI deaths, 40 were natural; the other six were probable homicides, five committed by one or both parents (two criminally convicted). The ratio of 40 natural to six unnatural deaths is 6·7 (95% CI 2·8–19·4). Enquiries identified 18 families with two SIDS(sudden infant death syndrome) deaths and two families with probable covert double homicides (ratio 9·0 [2·2 to 80·0]). There were no convictions in 13 incomplete cases. Families with three deaths are reported.

Interpretation

Repeat unexpected infant deaths are most probably natural.

Introduction

Recent court cases in the UK have highlighted the need for epidemiological data on the relative frequency of natural and unnatural deaths in families who have had more than one unexpected infant death. In 1956, Barratt1 introduced the term “cot death” (crib death) to describe unexpected infant deaths without obvious explanation. The term highlights the common feature that many of these infants are discovered dead in their cots. The term sudden unexpected death in infancy provides a broader designation, which includes cot deaths. After thorough post-mortem enquiries, sudden unexpected deaths in infancy divide into various explained deaths and a residual group for which no explanation can be found, designated sudden infant death syndrome (SIDS).

It has long been recognised that some sudden unexpected deaths in infancy are not natural.2, 3, 4, 5 Starting from the assumptions that “SIDS deaths appear to occur in families at random” and that “there is no evidence of a genetic etiology”, DiMaio and DiMaio6 concluded that “Siblings of SIDS have the same risk as the general population”. They therefore implicitly derived the probability of two or three SIDS deaths occurring in one family by squaring or cubing the probability of a single death. This approach led them to conclude that “while a second SIDS with one mother is improbable, it is possible” or “remotely possible” (2nd edition) and “A third case, in our opinion is not possible and is a case of homicide”. Meadow7 and the Confidential Enquiry into Stillbirths and Deaths in Infancy (CESDI) in their report on Sudden Unexpected Death in Infancy (SUDI)8 used a similar argument for deducing the probability of two SIDS deaths. Meadow, apparently following the DiMaios' lead, wrote in the third edition of the ABC of Child Abuse that “two is suspicious and three murder unless proved otherwise is a sensible working rule for anyone encountering these tragedies”.9

The assumptions of independence that underlie these estimates are invalid, because siblings have 25% of genes in common and experience an essentially similar domestic environment, which might, however, be altered by the preceding death. Several studies have shown that siblings of infants with SIDS are at increased risk of SIDS.10, 11, 12 The best estimate was given by the Norwegian population-based study, which reported that the relative risk of recurrence was 5·8 (95% CI 2·1–13·2).13 Such studies of recurrence have been criticised for not considering the possibility of serial infanticide.14 However, a review of the accuracy of the diagnosis of SIDS in Norway during the period of the study suggested that SIDS was originally substantially underdiagnosed.15 Nevertheless, the risk of recurrence is sufficiently low to exclude the possibility of a simple inheritance pattern for SIDS.

In consideration of whether repeated unexpected infant deaths in a family are natural or not, isolated estimates of the probability of more than one SIDS case in a family, however derived, are invalid and easily misinterpreted. Multiple infanticide is also very rare. What is important is the relative likelihood that the deaths are natural versus unnatural, given that two or more deaths have occurred.16

Such information can be obtained only from unselected series of families who have had more than one unexpected infant death. Emery, who was concerned about filicide and also undetected hereditary disease, in a personal series of cases17 reported that filicide was probable in five of 12 cases of two or more unexpected deaths in a family. He and colleagues reported similar results in another ten unselected cases of repeated unexpected infant deaths.18 Stanton,19 in a follow-up study of unselected cases, has recently reported three families who each had had two unexpected infant deaths, two of which raised maltreatment issues. In this paper we report the proportion of natural and unnatural infant deaths (ie, deaths before the age of 1 year), occurring in families enrolled on a support programme for parents who had previously had a sudden unexpected and apparently unexplained infant death.

Section snippets

Methods

The Foundation for the Study of Infant Deaths facilitates a programme of home support for parents with young babies when there is a history of sudden unexpected and apparently unexplained infant death in a sibling.20 The Care of Next Infant (CONI) scheme, which started in 1988, was available in 75% of health districts in England, Wales, and Northern Ireland by 1994 and is currently available in 91%. It is coordinated nationally from Sheffield and run locally in collaboration with the hospital

Results

From 1988 to December, 1999, 6373 babies from 5229 families had completed the CONI programme. 1144 (22%) were siblings to an earlier CONI baby. 57 (8·9 per 1000) CONI babies died under the age of 1 year. Nine deaths were not unexpected. 48 presented suddenly and unexpectedly; seven of these (1·10 per 1000 [95% CI 0·42–2·26]) were classified as probable homicides and 41 (6·43 per 1000 [4·62–8·73]) as natural sudden unexpected deaths in infancy. Thus, the proportion of sudden unexpected deaths in

Discussion

When considering these and many previous findings, it is essential to distinguish between cases in which homicide was or might have been legally proven and cases in which homicide appears to be the most probable explanation.12, 17, 18, 19 Emery's work on this issue was primarily concerned with the identification and protection of subsequent infants of families in which homicide appeared probable.17, 18, 23, 24 Similarly, in this study, one object of our confidential enquiries was to identify

References (32)

  • WG Guntheroth et al.

    Risk of sudden infant death syndrome in subsequent siblings

    J Pediatr

    (1990)
  • JL Emery

    Families in which two or more cot deaths have occurred

    Lancet

    (1986)
  • AM Barrett

    Sudden death in infancy

  • C Templeman

    Two hundred and fifty eight cases of suffocation of infants

    Edinburgh Med J

    (1892)
  • MA Valdes-Depena

    Sudden and unexpected death in infancy: a review of the world literature 1954–66

    Pediatrics

    (1967)
  • WA Kukull et al.

    Sudden infant death and infanticide

    Am J Epidemiol

    (1977)
  • WG Guntheroth

    Crib death: the sudden infant death syndrome

    (1995)
  • VJ DiMaio et al.

    Forensic pathology

    (1989)
  • R Meadow

    Unnatural sudden infant death

    Arch Dis Child

    (1999)
  • P Fleming et al.

    Sudden unexpected deaths in infancy: the CESDI SUDI Studies 1993–1996

    (2000)
  • R Meadow

    Fatal abuse and smothering

  • DR Peterson et al.

    The sudden infant death syndrome: repetition in families

    J Pediatr

    (1980)
  • SM Beal et al.

    Recurrence of incidence of sudden infant death syndrome

    Arch Dis Child

    (1988)
  • N Oyen et al.

    Population-based recurrence risk of sudden infant death syndrome compared with other infant and fetal deaths

    Am J Epidemiol

    (1996)
  • R Firstman et al.

    SIDS and infanticide

  • Vege et al.

    Diagnosis of sudden infant death in the Nordic countries since 1970, revised

  • Cited by (54)

    • Sudden Infant Death Syndrome and Apparent Life-Threatening Events

      2012, Kendig and Chernick's Disorders of the Respiratory Tract in Children
    • Sudden death in infants during sleep

      2011, Handbook of Clinical Neurology
      Citation Excerpt :

      There have been suggestions that when two or three unexpected unexplained infant deaths occur within a family they are more likely to be unnatural than natural (Meadow, 1999). In one study, cases of repeat sudden unexpected and unexplained infant deaths were reviewed (Carpenter et al., 2005). Second deaths were not rare (6.2 per 1000 living births) and the majority of deaths, 80–90%, were natural.

    • Underascertainment of child abuse fatalities in France: Retrospective analysis of judicial data to assess underreporting of infant homicides in mortality statistics

      2010, Child Abuse and Neglect
      Citation Excerpt :

      When the present study was set up, it was widely agreed that a certain percentage of SIDS were in fact homicides. However, studies such as Carpenter's (Carpenter et al., 2005) have since questioned this position. And yet Carpenter's study was seriously criticized, in particular for its lack of analysis of undetermined cases and the classification of natural and unnatural deaths (Bacon, Braithwaite, & Hey, 2007; Gornall, 2006).

    View all citing articles on Scopus
    View full text