In developing countries, the variations in the clinical spectrum of human immunodeficiency virus (HIV)-related oral lesions over time, and the possible effects of antiretroviral therapy, have not been described. In this study we evaluate the clinical spectrum of oral lesions in a series of HIV-infected patients when first examined at the acquired immunodeficiency syndrome (AIDS) clinic of a tertiary care institution in Mexico City, Mexico, and the changes observed over 12 years. All HIV-infected adult patients had an oral examination performed by specialists in oral pathology and medicine who used established clinical diagnostic criteria for oral lesions. Four periods were defined according to the evolving pattern of antiretroviral use: the first 2 were before the introduction of highly active antiretroviral therapy (HAART) and the last 2 were during more established use of HAART. For the statistical analysis the chi-square test for contingency tables and the chi-square test for trend were utilized. For dimensional variables, except age, the Kruskal-Wallis or Mann-Whitney rank sum tests were used when applicable and trend was tested with the Spearman correlation coefficient. Age was tested through analysis of variance (ANOVA) and linear regression analysis. Alpha value was set at p = 0.05 for each test. In the 12-year study, 1,000 HIV-infected patients were included (87.9% male). At the baseline examination, oral lesions strongly associated with HIV were present in 47.1% of HIV-infected patients. Oral candidosis (31.6%), hairy leukoplakia (22.6%), erythematous candidosis (21.0%), and pseudomembranous candidosis (15.8%) were the most frequent lesions. Oral Kaposi sarcoma (2.3%), HIV-associated periodontal disease (1.7%), and oral non-Hodgkin lymphoma (0.1%) were less frequent. HIV-related oral lesions decreased systematically-by half during the course of the 4 study periods (p < 0.001). Except for Kaposi sarcoma, all oral lesions strongly associated with HIV showed a trend to decrease significantly during the study period. No apparent variation in the occurrence of salivary gland disease or human papillomavirus-associated oral lesions was found. A significant trend to a lower prevalence was observed in the group of patients who were already taking antiretroviral therapy, non-HAART and HAART (p < 0.001 and p = 0.004, respectively). Only a discrete reduction, barely significant, was noted among untreated patients (p = 0.060). By Period IV (1999-2001), those who received HAART showed the lowest prevalence of oral lesions strongly associated with HIV (p < 0.001). Patients with oral lesions strongly associated with HIV had significantly lower median CD4+ counts and higher viral loads than those without oral lesions strongly associated with HIV (p < 0.001 and p = 0.005, respectively). When CD4+ counts were correlated with prevalence of oral candidosis, a consistently negative association was found; this association prevailed even after the study group was partitioned according to period. In this selected cohort of 1,000 patients with HIV infection, the clinical spectrum of HIV-related oral lesions has changed over the 12-year study, with a decreased prevalence of most oral lesions. Our findings probably represent improvements in medical care of HIV-infected persons, earlier detection of HIV-infected patients at the AIDS clinic, the increasing use of prophylactic drugs to prevent secondary AIDS-related opportunistic infections, and, perhaps most important, the availability of potent antiretroviral therapy in recent years, since the introduction of HAART.