Obstetric infections and clinical characteristics of maternal sepsis: a hospital-based retrospective cohort study

Abstract

Abstract Sepsis is responsible for 50% of intrahospital maternal deaths worldwide. Incidence is increasing in both low and middle-, and high-income countries. There is little data on incidence and clinical outcomes of obstetric infections including maternal sepsis in the Nordic countries. The aims of this study are to give estimates of the occurrence of obstetric infections and maternal sepsis in a Norwegian hospital cohort, assess the quality of management of maternal sepsis cases, and evaluate the usefulness of diagnostic codes to identify maternal sepsis retrospectively. We conducted a retrospective cohort study of pregnant, labouring, post-abortion, and postpartum women. We assessed the accuracy of the diagnostic code most frequently applied for maternal sepsis, O85. We found 7.8% (95% confidence interval 7.1–8.5) infection amongst pregnant, labouring, and postpartum women. The incidence of maternal sepsis was 0.3% (95% confidence interval 0.2–0.5), and the majority of sepsis cases were recorded in the postpartum period. Two thirds of women were given broad-spectrum antibiotics at the time sepsis was diagnosed, but only 15.4% of women with puerperal sepsis were given antimicrobials in accordance with national guidelines. When used retrospectively, obstetric infection codes are insufficient in identifying both maternal and puerperal sepsis, with only 20.3% positive predictive value for both conditions. In conclusion, obstetric infections contribute significantly to maternal morbidity in Norway’s second largest maternity hospital. This study provides incidences of maternal infections for hospitalised patients in temporal relation to pregnancy, labour, abortion and the postpartum period, knowledge which is valuable for planning of health care services and allocation of resources. In addition, the study highlights areas where improvement is needed in clinical handling of maternal sepsis. There is need for studies on the management quality and use of correct diagnostic codes in this patient category.