A multiple outcomes analysis of effects of smoking and stress on LBW, preterm, and SGA births

Introduction: Smoking and exposure to stress conditions may induce fetal hypoxia; thus it is plausible that stress conditions among pregnant women could modify the effect of smoking on pregnancy outcomes – LBW, preterm, and SGA. When investigating these outcomes simultaneously, appropriate analytic techniques are necessary to account for the correlation among them. Methods: This study applies generalized estimating equations (GEE) to investigate simultaneously the interaction effect of smoking (non-smoker, smoking 3 months before pregnancy only, smoking during last 3 months of pregnancy) and stress (present or absent) on LBW, preterm, and SGA. Type of stress (partner-associated, traumatic, financial, emotional) was investigated separately. Data were from the South Carolina Pregnancy Risk Assessment Monitoring Survey (PRAMS) from 1995 to 1999. Results: The multivariate analysis, simultaneously considering all three outcomes, indicated a possible interaction between smoking and traumatic stress for LBW, but not for preterm or SGA. Smoking during the last three months of pregnancy was positively associated with LBW whether a traumatic stressor was present or absent. However, smoking before pregnancy had a positive association with LBW only in the presence of traumatic stress, OR=1.62; 95% CI=1.05-2.49, compared to OR=0.93; 95% CI=-0.71-1.22 in the absence of stress. No interaction effect was found for the associations with SGA and preterm. Conclusion: GEE analysis of multiple birth outcomes simultaneously exhibited possible effect modification of smoking and traumatic stress related to LBW but not to preterm and SGA. This information could lead to more effective smoking cessation programs for women before and during pregnancy.