Predicting adverse outcomes following trauma in pregnancy.

Objective: To identify variables predicting adverse maternal or fetal outcome following trauma and to establish a composite morbidity model to predict poor obstetrical outcomes.

Study Design: A retrospective study of pregnant women following major and minor trauma from a single institution from 1985-2007 was performed. Abstracted data included maternal demographics, Injury Severity Score (ISS), laboratory and radiology studies,fetal monitoring and delivery information. Linear algorithm and logistic regression analysis estimated predictors of adverse obstetrical outcomes. Accuracy of composite morbidity models was tested using receiver operating characteristic.

Results: A total of 292 pregnant trauma patients were analyzed. Forty-eight (13%) women had an ISS > or = 10. Adverse pregnancy outcomes occurred in 71 patients (24%). Predictors of poor outcomes included trauma in the third trimester, length of stay (LOS) >2 days, abdominal trauma, ISS >2 and a positive Kleihauer-Betke (KB) test. Composite morbidity models resulted in sensitivity and specificity ranging from 54.3% to 70.4% and 59.5% to 87.5%, respectively.

Conclusion: Predicting adverse perinatal outcomes following major trauma remains challenging. Composite morbidity models using a combination of third trimester trauma, LOS > 2 days, abdominal trauma, ISS >2 or a positive KB test improves identification of those at risk for adverse perinatal outcomes.