The way in which the extensive use of highly active antiretroviral therapy (HAART) has influenced the incidence of visceral leishmaniasis (VL) among human immunodeficiency type 1 (HIV-1)-infected patients is not yet understood. The present study assessed whether the incidence of symptomatic VL in HIV-infected patients has decreased since the introduction of HAART. Likewise, the role of other potential risk factors for VL was also analyzed. Therefore, 479 HIV-1-infected patients receiving antiretroviral treatment, according to the available drugs at each moment, were prospectively followed from April 1989 to June 2000 in two university hospitals in southern Spain. A bone marrow aspiration was performed when patients showed symptoms suggestive of kala-azar. A diagnosis of VL was made when Leishmania amastigotes were seen in Giemsa-stained samples or promastigotes were cultured in specific media. The median follow-up time was 1,380 [8 to 4,536] days. Twenty-one patients were diagnosed with symptomatic VL. The density of incidence of VL has decreased 64.8% as of January 1997, when HAART began to be used extensively in our area. The use of HAART was the main independent factor associated with VL; this therapy was a protective factor (adjusted hazard ratio [HR], 0.05; 95% confidence interval [CI], 0.02 to 0.15). CDC clinical category C at entry in the cohort (HR, 4.08; 95% CI, 1.46 to 11.35) and CD4(+) cell counts below 300 cells/mm(3) during the follow-up (HR, 3.96; 95% CI, 1.56 to 10.01) were also independently associated with kala-azar. A VL diagnosis prior to follow-up and low compliance with antiretroviral therapy were not independently associated with symptomatic VL, although statistical significance was almost reached (P = 0.1 and P = 0.08, respectively). In summary, the use of HAART has led to a fall in the incidence of symptomatic VL in HIV-infected patients. The main risk factor associated with kala-azar emergence in patients infected with HIV is deep immunosuppression.