ObjectiveTo determine he role of homicide and other injuries in maternal deaths in North Carolina over the three-year period from 1992 through 1994.MethodsMaternal deaths were identified from death certificates that indicated a maternal death and through an enhanced surveillance system that matches death certificates with live-birth and fetal-death certificates. Deaths were classified as direct, indirect, medically unrelated, or injuryrelated. Patterns of prenatal care were ascertained from the matching live-birth or fetal-death certificates. Maternal death rates for whites and nonwhites wee calculated.ResultsThe most common cause of maternal death was injury, accounting for 62 of the 167 deaths (37%). Homicide was the most common cause of injury-related death (35.5%). The relative risk of maternal death for nonwhites compared with whites was 1.8 (95% confidence interval [CI]1.6, 2.1]. Similarly, their relative risk for injury-related maternal death was 1.7 (95% CI 1.4, 2.2).ConclusionIt is essential to include an analysis of injuryrelated deaths in maternal mortality reporting. As the most common cause of maternal deaths, injury is not limited to densely populated, metropolitan areas. Counseling regarding injury prevention, domestic violence, and depression should be a part of both prenatal and postpartum care.