| Abstract: | Several studies raise the possibility that smoking during pregnancy is associated with a slightly decreased odds of trisomy 21 at birth. If it is, associations may reflect decreased incidence at conception, increased intrauterine loss (at one or several times in gestation), or both. Women ( n = 13,729) undergoing prenatal diagnosis completed a questionnaire before learning karyotype results. For each women with a trisomy, up to 4 controls with chromosomally normal pregnancies, matched for age and hospital, were selected. Analyses drew on the 89 trisomy 21-control matched m-tuples in which diagnosis was by amniocentesis at 14-26 weeks. We compared the odds of smoking at last menstrual period and in the past in cases and controls. The odds of current smoking versus never smoking were decreased [adjusted odds ratio = 0.8, 95% confidence interval (CI) 0.4-1.6]; and the odds of exsmoking increased (adjusted odds ratio = 1.4, 95% CI 0.9-2.4) in trisomy 21 cases. The association with current smoking was essentially unchanged when the unexposed reference group was defined as exsmokers and women who never smoked (adjusted odds ratio = 0.7, 95% CI 0.4-1.4). These results for current smoking agree well with a summary estimate based on combined studies of births. One interpretation is that at amniocentesis, as has been reported for births, current smoking is associated with a slightly decreased odds of trisomy 21. If associations at amniocentesis and birth are of equal magnitude, the explanation that observations at birth reflect increased loss in the second half of pregnancy with current smoking is unlikely to be correct. However, the present amniocentesis evidence supports this interpretation weakly and is insufficient to rule out the null hypothesis that smoking is unrelated to trisomy 21. © 1993 Wiley-Liss. Inc. |
