BACKGROUND: HAART has been associated with metabolic abnormalities (hyperlipidemia, insulin resistance, alterations in cortisol metabolism) and fat redistribution. SETTING: A prospective study of 26 Caucasian men (median age 43.5 years) with HIV-1 viral loads < 500 copies/ml for 12 months while on highly active antiretroviral therapy (HAART) who interrupted treatment for a median of 7.0 weeks (range 4.9-10.3 weeks). Seventeen (65.4%) patients reported at least one fat redistribution symptom at baseline. METHOD: Serum lipids, glucose and insulin levels during an oral glucose tolerance test, 24-h urinary free cortisol and 17-hydroxycorticosteroids, and anthropometric parameters were measured before HAART cessation and prior to its reinstitution. RESULTS: When baseline values were compared with those obtained after HAART interruption (means +/- SD), there was a significant decrease in total cholesterol (194+/-47.3 versus 159+/-29.3 mg/dl; P < 0.0001), low density lipoprotein (LDL) cholesterol (114+/-32.6 versus 96+/-24.7 mg/dl; P = 0.0013), triglycerides (261+/-244.3 versus 185+/-165.4 mg/dl; P = 0.008), and 24-hour urinary 17-hydroxycorticosteroids (15+/-7.9 versus 5+/-2.5 mg/24 h, P < 0.0001) and a significant increase in 24-hour urinary free cortisol (45+/-34.1 versus 62+/-32.2 microg/24 h; P = 0.016). There were no significant changes in glucose or insulin levels or in anthropometric measurements. CONCLUSIONS: A relatively brief interruption of HAART resulted in significant improvements in total cholesterol, LDL cholesterol, and triglyceride levels. No changes were observed in insulin resistance profiles or anthropometric measurements, perhaps because of the brief duration of HAART interruption. These results suggest that hyperlipidemia and alterations in corticosteroid metabolism in the setting of HAART are a direct drug effect that reverses with drug withdrawal. However, glucose metabolism and fat redistribution do not change over the short term.