Post-discharge infections and healthcare contact in ugandan women hospitalized for delivery

Abstract

Objectives: Postpartum infection is a common cause of pregnancy related morbidity. Data are lacking from resource-limited settings on post-discharge febrile morbidity and healthcare contact in women with in-hospital postpartum fever. We hypothesized that women febrile postpartum were more likely to report post-discharge fever and infection, seek healthcare, receive antibiotics, and report ill health within 6 weeks postpartum than women normothermic postpartum. Methods: 4,231 largely rural-dwelling women presenting to Mbarara Regional Referral Hospital for delivery or postpartum care were prospectively enrolled. Vital signs were monitored every 8 hours after delivery. Febrile women were evaluated clinically and microbiologically for fever source. All febrile and 1,574 randomly selected normothermic women underwent interview and chart review to collect demographic, health, obstetric, and outcomes data; and were followed by telephone until 6 weeks postpartum. Categorical variables were analyzed using Chi squared and Fisher’s exact tests, and multivariable logistic regression was used to determine whether in hospital postpartum fever was an independent predictor of new postpartum infection and health care contact. Results: Temperature was measured for 4,176 women (99%); 121 (2.9%) developed in-hospital postpartum fever. Febrile women were significantly more likely to report new post-discharge antibiotic prescription (n¼64, 9.9 vs. 3.8%, P¼0.002), readmission (4.5 vs. 1.5%, P¼0.02), infection diagnosis (endometritis, wound, and urinary tract infections, n¼51, 7.2 vs. 3.0%, P¼0.02), wound infection (n¼29, 6.1 vs. 1.5%, P<0.001), and poor health (n¼58, 8.1 vs. 3.4%, P¼0.01) within 6 weeks postpartum than normothermic women. Of 51 new post-discharge infections, 39 (76%) occurred after cesarean delivery (4.6% of cesarean vs. 1.6% of vaginal deliveries, P¼0.001), 36 (71%) within 2 weeks of discharge, and 8 (15.7%) in women within-hospital postpartum fever, of whom 6 (75%) had endometritis inhospital. When controlling for potential confounders, in-hospital postpartum fever was associated with increased odds of new postdischarge infection (aOR 2.5, 95% CI 1.1-5.7, P¼0.03), but not healthcare contact (aOR 1.0, 95% CI 0.4-2.4, P¼0.96). Conclusions: In Uganda, in-hospital postpartum fever was associated with post-discharge hospitalization, infection diagnosis, and poor health. In-hospital postpartum fever should be evaluated and treated more thoroughly to prevent post-discharge febrile morbidity.