Several previous studies have suggested that maternal smoking is associated with a decreased incidence of trisomy 21. By using the Swedish health registries, 1,321 infants with Down's syndrome (DS) were selected among 1,117,021 infants born in 1983-1993 with known smoking exposure in early pregnancy. No association between maternal smoking and all cases of DS was found [age-adjusted odds ratio (OR) for maternal smoking: 0.98; 95% confidence interval (CI): 0.86-1.11], but heterogeneity over strata existed. A slightly decreased OR (0.91; 95% CI: 0.72-1.15) for any maternal smoking was indicated among primiparas, but among multiparas, no effect of smoking on the incidence of DS could be detected (OR: 1.01; 95% CI: 0.87-1.17). The difference between the OR for smoking > or = 10 cigarettes per day among primiparas (OR: 0.59; 95% CI: 0.38-0.90) and multiparas (OR: 1.06; 95% CI: 0.86-1.31) was statistically significant. If not due to statistical fluctuation, the findings indicate that no direct effect of smoking on DS risk exists but the association observed in primiparas is due to covarying factors.