RESUMEN
OBJECTIVE: To investigate risk factors for maternal mortality among HIV-infected women in Tanzania. DESIGN: Prospective cohort study. SETTING: HIV care and treatment clinics in Dar es Salaam, Tanzania. POPULATION: HIV-infected pregnant women. METHODS: Data were collected for all patients enrolled in an HIV/AIDS care and treatment program. Between November 2004 and September 2011, there were 18 917 women pregnant at least once during the follow-up. Thirteen percent of these women had more than one pregnancy, with 21 645 pregnancies occurring. Logistic regression was used to explore the predictors of maternal death among these women. MAIN OUTCOME MEASURES: Maternal mortality. RESULTS: During the study period, 363 maternal deaths occurred, giving a maternal mortality ratio of 1729 [95% confidence interval (CI) 1553-1905] per 100 000 live births. Being wasted [odds ratio (OR) 3.38, 95% CI 2.58-4.45] or anemic (OR 2.26, 95% CI 1.70-3.00) was associated with a higher risk of maternal mortality. Women who were initiated on antiretroviral therapy before their pregnancy had a 55% decreased risk of maternal mortality (95% CI 0.29-0.70) compared with women who were not. The risk of maternal mortality decreased with the length of time on antiretroviral therapy during pregnancy, by 8% for each additional month (OR 0.92, 95% CI 0.88-0.96). CONCLUSIONS: Maternal mortality was high among HIV-infected women. Initiating women on antiretroviral therapy as early as possible and providing nutritional interventions during pregnancy should be considered as means to reduce the maternal mortality among these women.