RESUMO
OBJECTIVE: To assess the direct, indirect, and total effects of violence during pregnancy on perinatal outcomes, and to evaluate the effect of violence as a moderator of the mediated relationship of depression with perinatal outcomes. METHODS: Data was collected from the prenatal study and follow-ups of the BRISA cohort, São Luís, Maranhão, Brazil. The perinatal outcomes investigated were: birth weight (BW), intrauterine growth restriction (IUGR) and gestational age (GA). Violence against women was evaluated using the World Health Organization Violence against Women instrument (Violence during pregnancy - regardless of the type of violence; Physical violence during pregnancy; Psychological violence during pregnancy). Depressive symptoms during pregnancy were evaluated as a mediating variable. Moderated mediation analysis was performed to estimate the effects of violence and depression on perinatal outcomes. RESULTS: Three types of violence analyzed by depression had an indirect effect in BW and GA. None of the types of violence showed an association with IUGR. All types of violence analyzed showed a moderated mediation effect with BW and GA. Only among women who experienced violence were birth weight and gestational age lower the higher the values of depressive symptoms. CONCLUSION: Violence and depression are only associated with lower BW and GA when they occur simultaneously.