Objective: Maternal mortality is underreported in the United States in part because traumatic deaths are not included in nationally reported maternal mortality ratios. The overall study goal was to compare women whose deaths had been reported to and investigated by a medical examiner and who had evidence of pregnancy to women without evidence of pregnancy in terms of socio-demographic information, toxicology results, and manner and cause of death. A secondary goal was to compare the pregnancy status and gestational age of women with evidence of pregnancy at the time of death in relation to the manner of death, with particular focus on women who died as a result of violent death.

Methodology: Autopsy charts from 1988–1996 for 651 women aged 15 to 50 from the District of Columbia Office of the Chief Medical Examiner whose autopsies included examination of the uterus were reviewed. Medical examiners' classification of manner and specific causes of death were used as the main outcome measures. Overall, the sample reflected demographic characteristics of women of childbearing age in the District of Columbia, with 82% black, 74.6% unmarried, and 46.5% aged 20 to 34.

Results: Among the 651 autopsy charts evaluated, 30 (4.6%) documented evidence of pregnancy; 43.3% of the women who died due to homicide with evidence of pregnancy were not included in the 21 pregnancy-related deaths officially reported by the District of Columbia State Center for Health Statistics during the study period, and therefore, were also not included in national maternal mortality ratios. Although not statistically significant, 11% more homicides occurred among women with evidence of pregnancy as compared to non-pregnant women. Pregnant women who died a violent death were significantly more likely than non-pregnant women to have died due to gunshot trauma. A significant proportion of pregnant women were < 21 weeks gestation at the time of their death. Additionally, women in this sample with evidence of pregnancy were over 3 times more likely to have been teenagers compared to non-pregnant women.

Conclusion: Medical examiner autopsy records identify violent pregnancy-associated deaths, many of which occur early in pregnancy and are missed by other enhanced case-finding techniques that require a record of a birth or fetal death. These deaths are usually excluded from reported maternal mortality ratios. Few studies have evaluated the prevalence of homicide in women of childbearing age, yet understanding the extent of less commonly associated causes of death during pregnancy such as homicide, may lead to improved identification of preventable problems that contribute to maternal morbidity and mortality. This study, which sheds new light on the identifying and reporting of maternal mortality, and specifically on homicide as a form of violence toward pregnant women, should be of particular interest for all women's health providers, as well as public health professionals, researchers, and advocates who are interested in the design, development, and evaluation of prevention programs, especially those directed toward preventable problems such as domestic violence.