Trauma and/or accidental injury complicates 6-7% of all pregnancies. The management protocols for trauma in pregnancy are based largely on case reports and small series. The purposes of this study were to: describe the demographics of pregnant trauma patients at a tertiary care center and a large community hospital; identify variables predictive of fetal outcome including an examination of Kleihauer-Betke and nonstress testing; and recommend an evaluation and management protocol after trauma based on empirical data rather than anecdotal reports. Data from pregnancies complicated by trauma from July 1987 through October 1993 were retrospectively reviewed. Statistical analysis included Chi-square and Kruskall-Wallis testing. There were 476 medical records available for review. Of the trauma cases, 54.6% were motor vehicle accidents, 22.3% were domestic abuse and assaults, 21.8% were associated with falls, and 1.3% were secondary to burns, puncture wounds, or animal bites. Mean maternal age was 24 years, 49.9% were Caucasian, and 43% were primigravid. Mean ges-tational age at occurrence of trauma was 25.9 weeks and mean gestational age of delivery was 37.9 weeks. Domestic abuse occurred most frequently before 18 weeks, falls between 20-30 weeks' gestation, and motor vehicle accidents occurred with equal frequency throughout gestation. Uterine contractions occurred in 39.8% of patients and as often as every 1 to 5 min in 18% of patients. Preterm labor occurred in 11.4%, preterm delivery in 25%, and abruptions in 1.58% of the trauma population. Fetal heart rate monitoring was abnormal in 3% of cases. Twenty-seven perinatal deaths were noted and in 14 pregnancies the deaths were related to trauma. Eight of these perinatal deaths were associated with motor vehicle accidents, four with domestic violence, and two with falls. The only preventable perinatal deaths were a twin pregnancy transferred with nonreassuring fetal heart tones. Early warning symptoms of vaginal bleeding, uterine contractions, and/or abdominal and/or uterine tenderness were not predictive of either preterm delivery or adverse pregnancy outcome, sensitivity 52%, specificity 48%. Abnormal monitoring and positive Kleihauer-Betke tests were also not predictive of adverse pregnancy outcome. However, there were no adverse outcomes directly related to trauma when monitoring was normal and early warning symptoms were absent (negative predictive value 100%). Two hundred eighty-nine Kleihauer-Betke tests were performed and only one affected management. Repetitive monitoring over several days did not uncover any patients whose heart rate tracings evolved from normal to abnormal monitoring. Given our findings that prolonged monitoring was not helpful in management of pregnant trauma patients, we support the recommendation that initial external fetal monitoring be performed for 4 hr, and, if reassuring, the patient may be sent home with precautions. We also recommend an Rh-immunoglobulin work-up for all Rh-neg-ative pregnant trauma patients, but do not recommend Kleihauer-Betke testing for Rh-positive women. Given the frequency with which trauma affects pregnancy and the difficulty encountered with identifying variables predictive of pregnancy outcome, there may be great benefits of incorporating trauma prevention into routine prenatal care.