All Johns Hopkins AIDS Service patients starting PI or non-PI containing regimens between January, 1996 and then end of September, 1997, were studied to evaluate the incidence of metabolic complications of antiretroviral therapy. Hypercholesterolemia was defined as an increase of 50 mg/dl over baseline, with a level over 200; hyperglycemia was defined as a glucose over 140 gm/dl, without a history of pre-existing diabetes. During the observation period, 14% of the patients who received PIs developed hypercholesterolemia compared to 3% of those who didn't. The incidence of hypercholesterolemia was 28.9 per 100 person years of follow-up in PI users, compared to 7.8 in those who were on PI-sparing regimens. Factors associated with increased cholesterol were lower liver function tests (AST, ALT) at baseline, increases in CD4 during therapy, greater reductions in HIV plasma viremia. Hypercholesterolemia was more commonly associated with ritonavir use, compared with ritonavir-saquinavir, which was more commonly associated with it compared to nelfinavir or indinavir. 2% of PI-using patients developed hyperglycemia, compared to none of the non-PI patients. The annualized incidence rate of hyperglycemia was 4.1% for PI users. There was no association of hyperglycemia and a specific PI. 34% of patients with hyperglycemia had elevated cholesterol levels, as well. The median time to hypercholesterolemia was: 84 days and for hyperglycemia: 53 days. Three hyperglycemia patients had to be hospitalized, and 6 need insulin. Although hypercholesterolemia is more common, the short term morbidity seems to be less than that associated with hyperglycemia in patients taking PI-containing regimens. Clinicians treating HIV-infected persons on PI-containing regimens need to be alert to these potential complications.