A monocenter observational study was conducted to determine the clinical and virological effects of cidofovir added to highly active anti-retroviral therapy (HAART) in AIDS-associated progressive multifocal leukoencephalopathy (PML). Exposure to other anti-viral drugs or late initiation of cidofovir were exclusion criteria. Of the 53 consecutive patients with virologically proven PML admitted at the NeuroAIDS Unit of Bicêtre Hospital between May 1996 and July 2000 and having received HAART with or without cidofovir, 46 met the inclusion criteria. Cidofovir was initiated in most cases on compassionate grounds. The 22 patients treated with HAART only (HAART group) were compared to the 24 patients treated with HAART and cidofovir (CDV group). Survival, neurological outcome assessed by the expanded disability status scale (EDSS), and monitoring of the JC virus (JCV) load in CSF were investigated prospectively. At baseline (date of initiation or intensification of HAART), both groups were similar regarding CD4 cell count, plasma HIV load, CSF JCV load, EDSS, and demographic features. Both groups had similar response to HAART in terms of plasma HIV load and CD4 cell count. At month 6, CSF-JCV load was below the detection level in 8 out of 24 (33%) patients from the CDV group and 7 out of 18 (39%) patients from the HAART group (P = 0.71). One-year cumulative probability of being alive was 62% in the CDV group and 53% in the HAART group (P = 0.72). However, an additional benefit with respect to survival was observed in patients who were given cidofovir after adjustment to the following baseline variables (CSF-JCV load, CD4 cell count, and EDSS). Despite the addition of cidofovir to HAART, no significant benefit had been observed in neurological outcome, particularly in patients with an early worsening.