Peripartum infections and associated maternal mortality in rural Malawi.

OBJECTIVE To assess associations between maternal mortality and severe morbidity and human immunodeficiency virus (HIV) infection, uptake of antiretroviral therapy, obstetric infections, and nonobstetric infections in a rural Malawian district, where the estimated HIV prevalence is 21%. METHODS We studied the incidence and outcomes of maternal peripartum infections between September 2007 and September 2009 at the district hospital. We used a facility-based prospective cohort study design, including all cases of severe maternal peripartum infection up to 42 days postpartum, and recorded maternal and pregnancy-related characteristics. We assessed the association between mortality and covariates (including nonobstetric infection, HIV prevalence, and uptake of antiretroviral therapy) using univariable and multivariable logistic regression models. RESULTS In total, 140 infections occurred: 79 (56%) obstetric and 53 (38%) nonobstetric (eight unknown). Half of the women were HIV-positive. Multivariable analysis showed that nonobstetric infection was the most important explanatory variable for mortality (adjusted odds ratio [OR] 4.23, 95% confidence interval [CI] 1.53-11.73). HIV-positive women not on antiretroviral therapy were at higher risk of mortality (adjusted OR 3.02, 95% CI 1.06-8.60) but there was no significant mortality increase among those on treatment (adjusted OR 0.51, 95% CI 0.10-2.71). The most common infections were puerperal sepsis (obstetric, case fatality rate 7%) and pneumonia (nonobstetric, case fatality rate 41%). CONCLUSION Untreated HIV infection and nonobstetric infections are independently associated with maternal mortality. Prompt treatment of HIV and nonobstetric infections in pregnant women must be prioritized to reduce maternal mortality. LEVEL OF EVIDENCE II.