The purpose of this study is to determine if, in men and women with excess abdominal fat and insulin resistance, people with HIV infection respond differently than people without HIV to interventions that typically improve body fat distribution and insulin resistance. The specific interventions are:

1. Diet + exercise program.
2. Rosiglitazone treatment.
3. A combination treatment of diet + exercise program and rosiglitazone.

A constellation of nutritional alterations in HIV-infected patients receiving highly active antiretroviral therapies (HAART), including body fat redistribution with subcutaneous adipose tissue (SAT) wasting and visceral adipose tissue (VAT) accumulation, hyperlipidemia, and insulin resistance (IR) has been described. There is a major concern that these developments will be associated with adverse clinical outcomes related to atherosclerosis, as suggested by several case reports (Henry 1998, Behrens 1998, Gallet 1998, Vittecoq 1998). Although there are well documented associations among body fat distribution, insulin resistance, and adverse health outcomes, especially accelerated atherosclerosis, in non-HIV infected individuals, it is unclear if the relationships are affected by HIV infection, or if they reflect the same outcomes. This information is important, since understanding the interrelationships between body fat distribution and metabolism may guide the development of treatment strategies.

The specific hypotheses to be tested are:

1. HIV infection does not affect the relative reductions in visceral (VAT) and subcutaneous adipose tissue (SAT) resulting from diet + exercise, but decreases the effect of this therapy on insulin resistance.
2. HIV infection decreases the changes in insulin resistance and body composition (increase in SAT and decrease in VAT) expected with rosiglitazone.
3. The combination treatment of diet+exercise and rosiglitazone will reduce VAT to a greater extent than rosiglitazone alone, and will improve insulin resistance to greater extent than diet and exercise alone, however these effects will be blunted in HIV-infected subjects.

Inclusion Criteria:

• HIV-infected or uninfected.
• Body mass index (BMI) at least 25.
• Excess visceral adipose tissue. Excess VAT will be determined in HIV+ and HIV- groups of men by a waist hip ratio > 0.95 and a waist circumference >88.2 cm, and in women by a waist:hip >0.9 and waist circumference >75.3 cm.
• Insulin resistance (fasting serum insulin level >16 μU/ml).

Exclusion Criteria:

• Unable to tolerate magnetic resonance imaging (MRI)
• Clinical evidence of active liver disease or a significantly abnormal liver function test (ALT >2.5x the upper limit of normal).
• Severe hyperlipidemia (fasting plasma triglycerides >500 mg/dL or fasting total cholesterol >300mg/dL)
• Current coronary artery disease including angina
• Peripheral vascular disease
• Uncontrolled hypertension
• Participation in a regular exercise program