To analyse the influence of educational levels on diverse baseline and follow-up characteristics and outcomes of HIV-infected patients, we sequentially evaluated 1352 individuals with known educational levels, who initiated a nelfinavir-based regimen. Higher educational degrees were associated with better baseline clinical (P=0.03) and immunological (P=0.003) conditions, not related to transmission categories, which were also observed during follow-up (P=0.003). However, these differences were only found in antiretroviral-experienced patients (P=0.002), while naive patients had very similar values (P=0.8). Overall, there were different CD4 responses (P=0.06), but not viral load responses (P=0.6), to antiretroviral therapy according to the educational level, but these differences were more marked in the last six months of follow-up (P=0.008). Patients with higher educational degrees had higher rates of adherence to medical appointments both before (P=0.0003) and during the study period (P=0.01), as well as to antiretroviral therapy in univariate (P=0.003) and multivariate analyses (P=0.007). Similarly, baseline CD4 counts proved to be independently associated with education after adjustment for other variables (P=0.0006). The educational groups also differed in diverse socioeconomic parameters and certain beliefs about HIV infection (P<0.0001 for each). We conclude that the patient's educational level influences clinical and immunological outcomes of HIV infection. This impact is probably mediated through differences in the long-term effects of treatment, as a result of adherence to antiretroviral therapy and to medical indications. The evaluation of social aspects such as the patient's education should be incorporated into routine clinical practice to improve the results of treatment.