The combination of didanosine (ddI) and lamivudine (3TC) is attractive considering its low cost, potency, tolerability, and convenience (once daily administration), but it is not recommended as first-line therapy for HIV infection. A prospective, multicenter, open, comparative trial was conducted in HIV-infected, antiretroviral-naive persons in Spain who begun a QD regimen with efavirenz (EFV), 3TC, plus ddI, the latter with or without food. A total of 103 patients were recruited in the study. Median baseline CD4 count was 229 cells/microl and plasma HIV-RNA was 4.9 logs copies/ml. In an intent-to-treat analysis, 78 (75.8%) had undetectable viremia at week 48 of therapy, without significant differences when comparing patients on and without fasting ddI (75% vs. 76.6%, respectively). The mean CD4 increase was of 199 cells/microl, with no significant differences between groups. Overall, 29 adverse events were recorded in 23 patients, the majority associated with neuropsychiatric symptoms of EFV. Treatment was discontinued in 10 (9.7%) patients due to adverse events. Virological failure was recognized in only six patients, four taking ddI with and two without food (p = 0.3). Drug resistance mutations were recognised in four of them. Plasma ddI concentrations did not differ significantly between groups. Mitochondrial DNA within peripheral blood mononuclear cells tended to increase in most subjects over 48 weeks of therapy regardless of treatment group. A QD regimen with ddI, 3TC, and EFV shows potency and tolerance similar to that reported using other currently recommended regimens in drug-naive HIV-infected patients. Its efficacy does not seem to be compromised when ddI is administered with food. Therefore, this regimen merits further investigation in larger comparative trials.