AIM: Our study objectives were as follows: assess exposure to psychosocial work demands among working pregnant women and women on preventive withdrawal from work; and measure the association between psychosocial work demands and major depressive symptoms, according to time of withdrawal from work. METHODOLOGY: Karasek's abbreviated scale was used to measure psychosocial work demands (Job strain and "Iso-strain") and CES-D scale (Center for Epidemiological Studies Depression Scale) was used to measure major depressive symptoms (CES-D score>/=23), at 24-26 weeks of pregnancy, among 3043 pregnant women in Montreal (Quebec) who worked at paid jobs at least 15 h/week and at least four consecutive weeks since the beginning of their pregnancy. Multivariate logistic regression models were built. RESULTS: At 24-26 weeks of pregnancy, 31.4% (956/3043) of pregnant women were on preventive withdrawal from work. They were more in "high-strain" (31.1% vs. 21.1%) and "Iso-strain" groups (21.0% vs. 14.2%) than those who continued to work (P/=21 weeks (adjusted OR=2.39; CI 95%: [1.10-5.20]). CONCLUSION: Psychosocial work demands are an important risk factor for the mental health of pregnant workers and require that preventive actions be put forward.