Determinants of the competing outcomes of intrauterine infection, abruption, or spontaneous preterm birth after preterm premature rupture of membranes

• Published in: The journal of maternal-fetal & neonatal medicine 2016-01, Vol.29 (2), p.258-263
• Main Authors: Hackney, David N., Kuo, Kelly, Petersen, Rebecca J., Lappen, Justin R.
• Format: Article
• Language: English
• Subjects: Abruptio Placentae - epidemiology; Abruptio Placentae - etiology; Chorioamnionitis - epidemiology; Chorioamnionitis - etiology; Competing risk analysis; Female; Fetal Membranes, Premature Rupture - epidemiology; Humans; outcome model; Pregnancy; Premature Birth - epidemiology; Premature Birth - etiology; survival analysis; United States - epidemiology
• ISSN: 1476-7058; 1476-4954
• DOI: 10.3109/14767058.2014.997703

Objective: Patients with PPROM are at risk for a variety of outcomes, including chorioamnionitis (CA), placental abruption (PA), or preterm labor (PTL). Competing risk regression can analyze a cohort's risk of individual outcomes while accounting for ongoing deliveries secondary to competing events. Methods: A secondary analysis of the subjects from MFMU BEAM study of neuroprotection after preterm birth (BEAM) with conservative PPROM management. Deliveries were categorized as: PA, CA, PTL, "elective" or "indicated". The association between outcomes of PA, CA or PTL and clinical predictors of twins, ethnicity, parity, gestational age at rupture, bleeding, contractions, cervical dilation, preterm birth history, weight, and genitourinary infections were evaluated via competing risk regression. Result: 1970 subjects were included. The significance and directionality of predictors varied according to specific outcomes. Patients with twins had an increased PTL hazard (1.85) though reductions in CA- (0.66) or PA-specific (0.56) hazards. Decreased latency in African-Americans was almost entirely due to an increased CA hazard (1.44) without a significant association with PTL. Increasing gestational age at membrane rupture was associated with a decreasing hazard of CA although increasing hazard of PTL. Conclusions: For patients with PPROM, the hazards associated with different clinical predictors vary according to exact outcomes.