Aims. To determine the factors associated with HIV seroprevalence rate
s for victims of drug overdoses. Design. Descriptive epidemiologic sur
vey of a complete 3-year sample of accidental fatal drug overdoses. Se
tting. New York City (population 7 322 564). Participants. All people
over 15 years of age (n = 2159) who died of accidental fatal drug over
doses during 1991-93. Measurements. Using medical examiner data and lo
gistic regression analyses were conducted to examine the association b
etween HIV seroprevalence and gender, race, age and type of drug overd
ose. Findings. There were 646 (29.9%) victims who were HIV positive. W
omen (37.5%) were more likely than men (27.9%) to be HIV positive. Afr
ican-Americans (39.4%) had a higher rate of HIV infection than Latinos
(27.5%), whites (19.2%) or Asians (8.3%). Victims residing in communi
ties with higher levels of poverty had higher rates of HIV infection b
ut poverty did not account for the high rates of HIV infection among A
frican-Americans. The highest rates of HIV infection were found among
victims aged 35-44 years (38.8%) and 45-54 years (33.7%). Dying from a
n opiate overdose was associated with a 2.4 times increase in the like
lihood of being HIV positive. Conclusion. If opiate abuse continues to
rise in the United States, HIV infection will increase in the next fe
w years. Women who abuse drugs participate in risky sexual practices a
nd are more likely than men to develop HIV infection from receptive se
x as well as sharing of needles. Harm reduction programs should addres
s risky sex as well as needle programs. There should be further study
of why African-Americans who died of drug overdoses have the highest r
ates of HIV infection.