A retrospective case-control study focused on deaths investigated by the Jefferson County Coroner/Medical Examiner Office (Alabama) between 1986 and 2002. Statute requires that the office investigate all sudden and unexpected deaths in the county. Cases selected for this study were identified by a computer search of the office database for all decedents between age 10 and 70 whose cause and manner of death remained undetermined following an autopsy and toxicological analysis for ethanol and drugs of abuse (n=61). Decomposed remains were included in the study. A control group (n=61) was chosen to closely represent a random sampling of the county's population. The decedents chosen for the control group were either pedestrians or passengers killed in a motor vehicle collision. The charts of all study-group and control-group cases were reviewed for the circumstances surrounding death, a documented history of drug abuse, and any compelling physical signs at autopsy that indicated drug use, such as needle track marks, nasal septum perforation, or polarizing particles in foreign body giant cells within the lungs. In none of the cases did the decedent have sufficient disease to explain death. All toxicology results were noted. In 21 pairs, the case subject had evidence of drug abuse but the control did not, and in 5 cases the reverse was true. The analysis showed that individuals with an undetermined cause of death were 4.2 times more likely to have evidence of drug abuse than were victims of a motor vehicle collision. Further research is required to determine how chronic drug abuse, with attention to specific drugs, may cause sudden death even when drugs cannot be detected in the system at the time of death. 2 tables and 9 references