HIV-INFECTION RISK BEHAVIORS AND METHADONE TREATMENT - CLIENT-REPORTED HIV-INFECTION IN A FOLLOW-UP-STUDY OF INJECTING DRUG-USERS IN NEW-ENGLAND

There is wide variation in reported risk factors for HN incidence amon g injecting drug users by community. Available HIV seroprevalence and incidence data indicate that nearly 60% of HIV infection is associated with injecting drug use in Connecticut and 48% in Massachusetts. Usin g 12-month follow-up data on 354 initially HIV-negative New England (M assachusetts and Connecticut) methadone treatment clients, we assessed the association between baseline drug use practices, sexual behavior, partner behaviors, and client-reported HIV infection during follow-up . Variables that predicted client-reported positive HN antibody test r esults were modeled by Cox proportional hazards regression. HN infecti on among those tested was 14.2 per 100 person years (PY) [95% Confiden ce interval (CI) = 9.5 to 21.3] For each injection the relative risk ( RR) was 1.1 (95% CI = 1.1 to 1.2), for males 3.0 (95% CI = 1.2 to 7.3) , for blacks 5.0 (95% CI = 1.6 to 15.5), for Hispanics 3.6 (95% CI = 1 .2 to 10.5). Men who used more than one unclean needle per day and had an HN-infected steady partner had an RR of 28.4 (95% CI = 4.4 to 176. 4). For women, using speedball (RR = 6.1, 95% CI = 1.2 to 38.8) and be ing black (RR = 4.4, 95% CI = 1.0 to 19.8) predicted self-reported HIV infection; having a steady partner who ever injected increased this r isk substantially (RR = 65.3, 95%, CI = 4.0 to 1046.5). These findings for IDUs in Massachusetts and Connecticut indicate that risk factors for HIV infection for men are consistent with expected transmission by unclean needles with an HIV-infected partner, but a preference for us ing speedball predicted HIV infection among women IDUs.