Abstract
BACKGROUND: Previous studies of lifestyle characteristics and risk of miscarriage have mostly been retrospective and failed to account for induced abortions. We examine whether pre-pregnancy body-mass index, alcohol intake and smoking influence the risk of miscarriage after accounting for induced abortions. METHODS: We conducted a prospective cohort study of 9213 women with 26,594 pregnancies participating in the Australian Longitudinal Study on Women's Health. We examined whether body-mass index, smoking and alcohol intake prior to pregnancy was associated with miscarriage. We estimated adjusted relative risks (RR) using generalized estimating equations with an exchangeable correlation matrix. We explored the impact of accounting for induced abortion by first excluding all induced abortions, and secondly including 50% of induced abortions in the comparison group. RESULTS: Of the 26,592 pregnancies which occurred during the follow-up period, 19% ended in a miscarriage. We observed an increased risk of miscarriage according to pre-pregnancy obesity compared to normal weight (adjusted RR 1.13; 95% CI 1.05, 1.21), smoking between 10 and 19 cigarettes per day compared to not smoking (adjusted RR 1.13; 95% CI 1.02, 1.25), but not smoking 20 or more cigarettes per day (adjusted RR 1.07; 95% CI 0.94, 1.21) and risky drinking (≥2 units per day; adjusted RR 1.15; 95% CI 1.03, 1.28) compared to low risk drinking (< 2 units per day). The results for smoking (adjusted RR 1.09 for 10-19 cigarettes per day; 95% CI 0.98, 1.21) was attenuated after including 50% of induced abortions in the comparison group. CONCLUSIONS: We observed a modest increased risk of miscarriage according to obesity and risky alcohol intake prior to pregnancy. There was no evidence of a dose-response relationship with smoking, and the association between smoking and risk of miscarriage was attenuated after accounting for induced abortions.