The human immunodeficiency virus (HIV) infection rate was examined in a selected cohort of healthy clients of an inner-city alcohol treatment center from 1990 through 1993. These subjects were also participating in a research protocol (n = 258) designed to assess immunity and HIV risk behaviors in inner city alcohol-dependent persons. Healthy alcohol-abusing heterosexual clients (165) had HIV testing conducted in an inner-city ambulatory alcohol treatment center between September 1990 and December 1993. Respondents were 93.9% African-American and 3.6% Hispanic; 72.1% were male. Anonymous HIV-1 antibody testing was conducted retrospectively for an additional 80 subjects who participated in the research protocol during the same interval, but for whom HIV-1 antibody testing was not conducted clinically at the time. HIV infection rate among the clinic-tested subjects (n = 165) was 4.4% for individuals who were exclusively alcohol-dependent, 1.4% for non-injecting drug use (IDU) mixed substance abusers, and 46.8% for clients with a history of IDU. Rates did not differ among cohorts tested in different years. Among non-injecting drug users tested in the clinic, all infected respondents (n = 3) were women (p = 0.03). Among those tested anonymously (n = 80), however, infection rate for exclusively alcohol-dependent persons was 16.7%, non-IDU mixed abusers 11.1%, and injecting drug users 48.3%, with seropositive males as well as females in each group. HIV infection rates for the pooled samples (n = 245) were 8.7% for exclusively alcohol-dependent persons, 5.1% for mixed abusers, and 54.5% for injecting drug users. Among non-injecting drug users, exclusively alcohol-dependent women had a significantly higher (p < 0.01) infection rate (20.0%) than the remaining females and males. Infection rates among exclusively alcohol-dependent males, male and female polysubstance non-IDU abusers, and injecting drug users were comparable with that seen in an earlier screening in the same clinic in 1989, with apparently little diffusion of infection from the IDU population to other substance abusers. An exception seemed to be exclusively alcohol-dependent females, who show substantially elevated rates. Age, housing, and other social differences may help segregated substance-abusing populations in the relatively small Newark metropolitan area, although not protecting exclusively alcohol-dependent females.