Recognizing the impact of the 1977 San Francisco study of trauma deaths in trauma care, our purpose was to reassess those findings in a contemporary trauma system. Cross-sectional. All trauma deaths occurring in Denver City and County during 1992 were reviewed; data were obtained by cross-referencing four databases: paramedic trip reports, trauma registries, coroner autopsy reports and police reports. There were 289 postinjury fatalities; mean age was 36.8 +/- 1.2 years and mean Injury Severity Score (ISS) was 35.7 +/- 1.2. Predominant injury mechanisms were gunshot wounds in 121 (42%), motorvehicle accidents in 75 (38%) and falls in 23 (8%) cases. Seven (2%) individuals sustained lethal burns. Ninety eight (34%) deaths occurred in the pre-hospital setting. The remaining 191 (66%) patients were transported to the hospital. Of these, 154 (81%) died in the first 48 hours (acute), 11 (6%) within three to seven days (early) and 26 (14%) after seven days (late). Central nervous system injuries were the most frequent cause of death (42%), followed by exsanguination (39%) and organ failure (7%). While acute and early deaths were mostly due to the first two causes, organ failure was the most common cause of late death (61%). In comparison with the previous report, we observed similar injury mechanisms, demographics and causes of death. However, in our experience, there was an improved access to the medical system, greater proportion of late deaths due to brain injury and lack of the classic trimodal distribution.