Sudden Infant Death Syndrome and Sleeping Position

Abstract

To the Editor. Several published articles1–3 have assessed the association between sudden infant death syndrome (SIDS) and “sleeping positions” of infants. All these articles point to the prone or lateral position during sleep as a risk factor. The aim of this letter is to present some remarks about the definition of “last sleeping position” used in most of those studies. We consider a paper1 describing a 3-year (1987–1990) case-control study conducted in New Zealand. For the definition of “sleeping position,” the authors considered:The main result of this study was that the infants who were classified as unaccustomed to prone (ie, infants whose usual sleep position was nonprone but were placed prone for the last sleep) had a greater risk of SIDS than those who usually slept in a supine position (adjusted odds ratio [OR]: 19.3, 95% confidence interval [CI]: 8.2–44.8). This result was obtained by comparing the position in which the infant was put to sleep the last time (Fig 1, comparison 1).It is difficult to understand the point of comparing the “last sleep” positions of the cases and controls because we find some ambiguity in the definition of “last sleep” of the controls. Mitchell et al defined “last sleep” position for the healthy infants (controls) as the position in which the infant was put to sleep the last time before the interview with the parents of the infant (nominated sleep). We believe that for the controls, the nominated sleep would have been one particular sleep of many, and, therefore, it could be considered as “usual sleep” and not “last sleep” as defined in the article. Thus, it would be sensible to compare only the “usual sleep” position of cases and controls, as Mitchell et al have done (Fig 1, comparison 2), confirming the danger of usually putting the infant down in a prone position. To establish the etiologic role of last sleep, we believe that an interesting comparison could be the one in which “last sleep” and “previous sleeps” are compared within cases (Fig 1, comparison 3).The reason for the greater risk of SIDS in the case of the prone position has not yet been fully understood. Putting aside the numerous physiopathologic theories for SIDS, the question is: what is the causal model for SIDS? Does the number of times an infant sleeps in the prone position increase the risk of SIDS (similar to a causal model that relates smoking to lung cancer)? Or, although the prone position is a known risk factor, is being unaccustomed to the prone position the bigger risk? This may be likened to a causal model showing that when people unaccustomed to drinking alcohol do drink it in moderation, they increase their risk of a car accident, compared with those who drink alcohol regularly. As is often the case in etiologic epidemiology, hints about causation come before causal models are fully understood. To do this in research about the etiology of SIDS, a better conceptualization of the optimal study is a priority. We hope that what we have discussed above may help in this direction.