C56-44 Characteristics and Processes of Care Among Pregnant vs Non-pregnant Patients With Sepsis

Abstract Rationale Maternal sepsis, a life-threatening condition resulting from infection-related organ dysfunction during pregnancy, childbirth, or post-partum period, is a leading cause of pregnancy-related death in the US, causing 23% of maternal deaths. Delayed recognition and treatment, which may be worse among pregnant/ peripartum patients in comparison to non-pregnant/peripartum peers, may contribute to poor outcomes. We hypothesized that process of care and outcomes differed between pregnant and non-pregnant septic patients. Methods We assessed all encounters among females 18-45yr from 14 hospitals in one Pennsylvania healthcare system from January 2016 to June 2018. Encounters were identified as having sepsis based on presence of implicit or explicit sepsis ICD codes, or meeting Sepsis-2 or Sepsis-3 criteria within 24hr of presentation using electronic health record (EHR) data. Encounters were then stratified as pregnant if there was an obstetric ICD-10 code (O00-O9A) in the discharge diagnoses, versus non-pregnant. We compared population characteristics, processes of care, and outcomes between the groups using Fisher’s exact and chi-square tests. Results There were 440,882 age-eligible female encounters in the 14-facility data set. Of these, 14,662 (3.3%) met sepsis criteria according to ICD or EHR criteria. Among those, 244 (1.7%) were pregnant and 14,418 (98.3%) were non-pregnant while 76.3% of pregnant and 14.4% of non-pregnant patients were admitted to a specialty obstetric hospital. Pregnant patients with sepsis tended to be younger (mean age 30 vs. 32yr, p 0.001), more likely to be identified as African American (28% vs. 19%, p = 0.02), have a lower mean Charlson Comorbidity Index (0.3 vs 1.0, p 0.001) and had shorter ICU stays (median 2 days vs. 3, p 0.001). Hospital mortality was low overall (0.4% vs 1.6%, p = 0.14). There was no detectable difference in ventilator use/dependent days, vasopressor use/dependent days, or organ dysfunction between groups. Pregnant patients were less likely to have a lactate measured within three hours (16.39% vs 29.06%, p = 0.00) but were more likely to have blood cultures drawn before antibiotic initiation (83.61% vs 77.49%, p = 0.02). The groups were equally as likely to receive a 30cc/kg fluid bolus within the first 6 hours. Conclusions Clinical outcomes did not differ between pregnant and non-pregnant patients meeting sepsis criteria in this cohort, which differs from some existing literature. This may be related to the younger age and lower comorbidity burden of pregnant patients, as well as the availability of specialized women’s health centers. However, pregnant patients with sepsis do have delays in lactate measurement. This abstract is funded by: NIH Institutional Research Training Grants (T32)