A retrospective case-control study was performed to analyze hospital-acquired candidemia in HIV-positive patients. To understand the impact of Highly Active Antiretroviral Therapy (HAART) on the incidence of nosocomial candidemia, two time periods were compared: A (1992-1996) and B (1997-2001). 32 out of 38 (84%) cases of candidemia were hospital-related. A significant reduction in the incidence of all cases of hospital-acquired candidemia has been observed in the post-HAART in respect to pre-HAART period (0.09 episodes vs. 1.1 per 100/py). Multivariate analysis showed that the presence of central venous catheter was the only variable independently associated with the development of nosocomial candidemia. The overall mortality rate was 59%. Univariate analysis indicated three prognostic indicators: presence of concomitant opportunistic infections, isolation of non-albicans Candida species; neutropenia. Multivariate analysis of prognostic indicators showed that isolation of non-albicans Candida species is the only independent variable. Despite the use of HAART, this disease still represents a severe complication of advanced stage of AIDS.