The decline in Sudden Infant Deaths (SIDS) since the early 1990s reflects the successful application of routine public health activities to an important health problem [1,2]. That this decline occurred despite what is often considered ‘weak’ evidence is worth noting. The ‘causes’ of SIDS were and remain unknown. Allowing for some diagnostic ambiguity , international comparisons of SIDS rates showed large variations with the implication that cultural and environmental factors were likely to be located somewhere in the causal sequence . A number of case control studies identified behaviours that might be important in distinguishing cases from controls. These included not smoking after the birth, prone sleeping position for the child, breastfeeding of the child and not wrapping the child too warmly [5,6]. Based upon this weak evidence, intervention programs were implemented to persuade mothers to act on the above indications with dramatic consequences. SIDS rates in many countries declined rapidly and were often halved or more than halved . Of course it is not fully known which of the behaviours changed and contributed most to the decline. Nor is it known whether other behaviours possibly associated with the few which were initially identified might make a causal contribution to the reduction in SIDS.
The paper by Phillips, Brewer and Wadensweiler  in this issue considers the possibility that maternal (and paternal?) alcohol consumption might make an independent contribution to SIDS deaths. The authors argue that periods of very high alcohol consumption are associated with an increased rate of SIDS deaths and suggest that periods of high alcohol consumption may be associated with reduced vigilance of the child and consequently a greater risk of a SIDS death. They base this argument on their findings that:
- 1Children of alcohol consuming mothers are more likely to die of SIDS (as well as other causes) than do the children of mothers who do not consume alcohol.
- 2SIDS deaths (but not other child deaths) disproportionately occur on Saturdays and Sundays, periods associated with heavier alcohol consumption.
- 3There is a peak in SIDS deaths on New Year's Eve, Marijuana Day and possibly July 4th, periods associated with heavy alcohol (or cannabis) use.
While this research is interesting and potentially important, it remains suggestive. There are at least three reasons why the findings might not be indicative of a causal pathway. Firstly, the research is unable to control for other behaviours associated with alcohol consumption and possibly SIDS (such as particularly heavy tobacco consumption—a known risk factor for SIDS). Second, weekends and New Years Eve are periods when people engage in a range of recreational, social and other activities. Vigilance of a child may consequently be lower or other recreational activities may have a role in SIDS. Thirdly, the social and demographic characteristics of smoking and heavy alcohol consuming parents (e.g. younger age, higher levels of poverty) are different from their non-smoking and non-heavy alcohol consuming peers. Some of these characteristics may make a contribution to SIDS via a range of behaviour differences which are concentrated on weekends and New Year's Eve. For example, the economically disadvantaged have different patterns of social interaction than do their economically better off counterparts. Phillips  add some further possibilities which they address but which cannot be completely dismissed. These include (i) parents sleeping in longer and leaving the child unsupervised, (ii) poorer medical care available on weekends and New Year's Eve, (iii) poorer recording of data on weekends and New Year's Eve, e.g. SIDS deaths tend to be a residual category of death which might be disproportionately selected on busy weekends.
However, even weak evidence may be correct, and it is the issue of whether the findings are likely to be correct that needs to be considered. There are three components of the researchers' argument which can be further considered.
Firstly, is it likely that the level of parental surveillance impacts on the SIDS death rate? Child sleep monitoring equipment has been in use for over 30 years with no evidence that its use reduces the risk of SIDS . The window of time between the child's cessation of breathing and the possibility of an effective response is narrow. Indeed, the American Academy of Pediatrics Review  suggests that apnea (cessation of breathing) is not a risk factor for SIDS. The close monitoring of a child may not improve SIDS outcomes.
Secondly, do women who have recently given birth (SIDS is concentrated in the first six months and, at most, the first year of a child's life) binge drink in significant numbers? There is some evidence that women reduce their alcohol consumption in pregnancy and that only a very small percentage drink alcohol in pregnancy at hazardous levels [9,10]. There is not a great deal of research on postpartum levels of alcohol consumption but data from the ALSPAC study shows an increasing level of maternal alcohol consumption after the birth of a baby. While only 5.1% of mothers consumed an ounce or more per day, some 37.4% of mothers reported at least one episode of binge drinking in the previous month . The evidence does raise some concern about the proportion of young children exposed to high levels of alcohol consumption by their mothers, even though other women may be consuming alcohol at higher levels.
Thirdly, are there findings from other types of studies which would lend support (or contradict) the findings of this paper? The few available case-control studies of maternal alcohol use by women who have experienced a SIDS death have produced inconsistent findings. Scragg and colleagues  found that maternal alcohol consumption in the month prior to the SIDS death was not a risk factor for SIDS. More recently Blair and colleagues  have reported a strong association of co-sleeping and parental use of alcohol or drugs and SIDS deaths. Co-sleeping with the child while affected by alcohol was a risk factor for SIDS. Some 19 of the 77 SIDS deaths had a mother who consumed more than 2 units of alcohol in the past 24 hours prior to the SIDS death, compared to only two of 87 mothers in the randomised controls.
The findings reported by Phillips  raise an important question about one of the possible health consequences of maternal (and perhaps paternal) postnatal alcohol consumption. The authors have directed our attention to the topic of postnatal alcohol consumption and the potential for negative health outcomes as a consequence. There is a need to know more not only about whether SIDS is one such outcome but whether postnatal alcohol consumption is implicated in other possible outcomes as well.