|| Checking for direct PDF access through Ovid
We examined highly active antiretroviral therapy (HAART) era and pre-HAART era hospitalization rates among 604 HIV-infected drug users in a prospective study in Bronx, New York. Medical history and risk behaviors were elicited by semiannual interviews. Standardized medical record review abstracted discharge diagnoses for all hospitalizations. Hospitalization rates from January 1997 to December 2000 were compared with rates from January 1992 to December 1996. The rate of hospitalizations per 100 patient-years in the HAART era was 49.3 compared with 44.1 in the pre-HAART era (P = 0.13). Among women, the rate was significantly higher in the HAART era than in the pre-HAART era (68.1 vs. 49.4 hospitalizations per 100 patient-years, respectively; P = 0.01). In the second era, HAART users had lower rates than those who did not use HAART (37.2 vs. 83.4 hospitalizations per 100 patient-years, respectively; P < 0.001) for both HIV-associated and non-HIV-associated illnesses. Multivariate analysis revealed that in the HAART era, female gender (relative risk ratio = 1.72, P = 0.03) and not using HAART (relative risk ratio = 1.82, P = 0.02) independently predicted increased hospitalization risk. In the pre-HAART era, women were at independently higher risk of hospitalization (relative risk ratio = 1.36, P = 0.05). Among HIV-infected drug users, those who use HAART have a decreased risk of hospitalization; those who do not use HAART remain at high risk of continuing morbidity from both HIV-related and non-HIV-related illness and have high hospitalization rates.