Abstract

PIP: A review of the literature from 1957-79 on the pregnancy outcomes associated with smoking yielded estimates of various attributable risk statistics for these outcomes. Rates were extracted for 28 pregnancy outcome variables in smoking and nonsmoking mothers found in cohort studies of pregnancy. Data for similar outcomes were combined to produce average rates of each outcome variable; the relative risk for each type of outcome was then obtained by dividing the rate in smokers by the rate in nonsmokers. The amount of damage to be attributed to smoking was determined by estimating the attributable risk (AR) for the 25 pregnancy outcomes found to occur more often in smokers. The percentage of women reported as smoking during pregnancy has ranged from 14% in American Orientals to 61% in US Whites. Of the total 260,000 women in these studies, approximately 42.1% smoked while pregnant. ARs for various outcomes, expressed as rates per 1000 maternal smokers, vary from 1.2 (congenital abnormalities) to 56.7 (perinatal plus deaths), about 1/2 of all rates lying below 10. These are the measure of the number of events per 1000 smoking mothers that might in theory have been prevented if they did not smoke. An average of estimates of nonfatal events in mothers and fatal events in offspring suggests about 20 nonfatal bleeding episodes per 1000 smoking mothers and 24.1 offspring deaths from conception to the end of the first year of life. On average, between 5% and 25% of unfavorable outcomes in all pregnancies may be attributable to smoking. Although low relative risk values suggest clinical insignificance, the estimates are probably too low because of varying definitions of smokers and nonsmokers. Dose-related outcomes and other possible causes of pregnancy outcome risks must also be considered. A causal relation implies that the portion of pregnancy outcomes attributable to smoking would be prevented if pregnant women did not smoke.