Can the fall in Avon's sudden infant death rate be explained by changes in sleeping position?

Abstract
OBJECTIVE--To examine the impact of changing practice with regard to infant sleeping position on mortality from the sudden infant death syndrome. DESIGN--A population based study of all infants dying suddenly and unexpectedly during February 1990 to July 1991, and two groups of controls; one comprising every 125th baby born to Avon residents and the other comprising pairs of infants matched to each index case for age, neighbourhood, and date of study. Information about sleeping position was collected at home visits soon after the index baby's death or, for the population based controls, on several occasions in the first six months of life. The design was comparable to that of an earlier study of the same population. SETTING--County of Avon. SUBJECTS--35 infants who died suddenly and unexpectedly (32 of the sudden infant death syndrome), 70 matched controls, and 152 population based controls. RESULTS--The prevalence of prone sleeping in the matched controls was much lower than that found in an earlier study in Avon (28% (18/64) 1990-1 v 58% (76/131) 1987-9; p less than 0.001) and was comparable with the prevalence in population based controls (29%). This would be expected to lead to a reduction in the incidence of the sudden infant death syndrome to 2.0/1000 live births (95% confidence interval 1.8/1000 to 2.5/1000). The actual mortality fell from 3.5/1000 in 1987-9 to 1.7/1000. CONCLUSION--The fall in mortality can be almost entirely accounted for by the reduction in prone sleeping, suggesting a causal relation exists between them. Side and supine positions confer protection but the side position is unstable and the infant may roll prone. We therefore recommend supine as the safest sleeping position for babies.