A novel lipodystrophy syndrome characterized by insulin resistance, hypertriglyceridemia, and fat redistribution has recently been described in human immunodeficiency virus (HIV)-infected men and women. Women with the HIV lipodystrophy syndrome exhibit a marked increase in waist-to-hip ratio and truncal adiposity; however, it is unknown whether androgen levels are increased in these patients. In this study, we assessed androgen levels in female patients with clinical lipodystrophy based on evidence of significant fat redistribution in the trunk, extremities, neck and/or face (LIPO: n = 9; age, 35.7+/-1.7 yr; BMI, 24.7+/-0.8 kg/m2) in comparison with age- and BMI-matched nonlipodystrophic HIV-infected females (NONLIPO: n = 14; age, 37.6+/-1.1 yr; BMI, 23.4+/-0.6 kg/m2) and healthy non-HIV-infected control subjects (C: n = 16; age, 35.8+/-0.9 yr; BMI, 23.1+/-0.4 kg/m2). Fasting insulin, lipid levels, virologic parameters, and regional body composition using dual energy x-ray absorptiometry were also assessed. Total testosterone [ LIPO, 33+/-6 ng/dL (1.1+/-0.2 nmol/L); NONLIPO, 17+/-2 ng/dL (0.6+/-0.1 nmol/L); C, 23+/-2 ng/dL (0.8+/-0.1 nmol/L); P < 0.05 LIPO vs. C and LIPO vs. NONLIPO] and free testosterone determined by equilibrium dialysis [LIPO, 4.5+/-0.9 pg/mL (16+/-3 pmol/L); NONLIPO, 1.7+/-0.2 pg/mL (6+/-1 pmol/L); C, 2.4+/-0.2 pg/mL (8+/-1 pmol/L); P < 0.05 LIPO vs. C and LIPO vs. NONLIPO] were increased in the lipodystrophic patients. Sex hormone-binding globulin levels were not significantly different between LIPO and C, but were significantly lower in the LIPO vs. NONLIPO patients (LIPO 84+/-7 vs. NONLIPO 149+/-17 nmol/L, P < 0.05). The LH/FSH ratio was significantly increased in the LIPO group compared with the NONLIPO and C subjects (LIPO, 2.0+/-0.6; NONLIPO, 1.1+/-0.1; C, 0.8+/-0.1; P < 0.05 LIPO vs. NONLIPO and LIPO vs. C). Body fat distribution was significantly different between LIPO and C subjects. Trunk to extremity fat ratio (1.46+/-0.17 vs. 0.75+/-0.05, LIPO vs. C, P < 0.05) was increased and extremity to total fat ratio decreased (0.40+/-0.03 vs. 0.55+/-0.01, LIPO vs. C, P < 0.05). In contrast, fat distribution was not different in the NONLIPO group vs. control subjects. Among the HIV-infected patients, free testosterone correlated with percent truncal fat (trunk fat/trunk mass) (r = 0.43, P = 0.04). These data suggest that hyperandrogenemia is another potentially important feature of the HIV-lipodystrophy syndrome in women. Additional studies are necessary to determine the clinical significance of increased androgen levels and the relationship of hyperandrogenism to fat redistribution and insulin resistance in this population of patients.