In order to establish the role of the protease inhibitor nelfinavir in current clinical practice, a prospective 18-month open-label comparison of efficacy and tolerability of nelfinavir was performed among HIV-infected patients who either incorporated nelfinavir in their first-line highly active antiretroviral therapy (HAART) regimen (group A, 57 patients), or who added nelfinavir to a rescue antiretroviral regimen (following at least two attempts with protease inhibitor-based HAART) (group B, 67 patients). All evaluable data were analyzed according to the prior and concurrent antiretroviral therapy, including genotypic resistance assays for patients undergoing salvage therapy. A significantly better virologic outcome (as expressed by a > 2 log(10) drop of plasma viremia versus baseline or attainment of undetectable levels), was shown among patients belonging to group A versus group B, where a number of genotypic mutations possibly elicited by previous anti-HIV treatment strongly impaired a potent and sustained nelfinavir activity. On the whole, the immunologic response (as expressed by the mean CD4(+) lymphocyte count versus baseline), substantially paralleled the virologic one in all analyzed subgroups, but a tendency toward a maintained immunologic competence was also observed in the majority of patients experiencing virologic failure. Nelfinavir introduction was sufficiently safe, because a limited percentage of patients suffered from mild-to-moderate, novel, or continuing adverse events, which proved significantly more frequent in the salvage group but did not affect adherence to HAART.