BACKGROUNDS AND AIMS Involuntary weight loss and muscle mass loss among HIV-positive patients are only detectable in late stages, leading poor life quality. The reduction of adductor pollicis muscle thickness (APMT) can be easily and earlier uncovered in those cases. The purpose was to estimate APMT and compare it with immunological and anthropometric parameters of HIV-infected people. METHODS A crosssectional study was carried out in an University Hospital including 103 HIV-infected outpatients by subjective global assessment (SGA). Data were compared to APMT for the whole sample and between gender in univariate analysis. Besides that, simple correlation and multiple linear regression were done to check the APMT relation with gender, age, weight body, body mass index, arm circumference, CD4, CD8 and viral load. RESULTS The APMT average values of the dominant hand (16.2 ± 4.2 mm) and non-dominant hand (14.8 ± 4.3 mm) were lower than in the healthy population. Through stratified analysis by gender, it was found significant difference in weight, arm muscle circumference, arm muscle area, triceps skinfold thickness and arm fat area (p < 0,01 for each). In any age group, men had significantly higher dominant and non-dominant APMT values than women (p < 0.001). Although the fair correlation among cited variables and APMT of both hands, there were no correlation and no difference between the genders in regards to immunological markers (CD4, CD8 and viral load). In a prediction model to APMT values, gender was determinant in multiple linear regression. CONCLUSIONS In a well-nourished HIV sample by SGA with adequate CD4 counts, APMT measures of both hands were lower than in healthy people. In both hands, APMT were positively correlated with weight and male, regardless of other anthropometric data and immunologic factors.