Table 1 Inclusion/Exclusion Criteria.

• 1. Inclusion Criteria

   • Men and women

   • All ethnic groups

   • ≥ 18 years of age

   • IGT with a FPG = 95–125 mg/dl plus at least one additional high risk

   • characteristic (see text). IGT is defined as a two hour plasma glucose

   • concentration = 140–199 mg/dl during a single 75 gram OGTT.

   • BMI ≥ 25 kg/m² (no upper limit) (BMI > 22 kg/m² for Asian Americans)

1. Exclusion Criteria

   • Subjects with diabetes mellitus: FPG ≥ 126 mg/dl or 2-hour plasma glucose ≥ 200 mg/dl during OGTT.

   • Subjects previously treated with a thiazolidinedione (ever) or metformin (within one year prior to randomization)

   • Subjects previously treated with a sulfonylurea, a meglitinide, an alpha glucosidase inhibitor for more than one week within the last year, or within the 3 months prior to randomization

   • Subjects previously treated with insulin (other than during pregnancy) for more than one week within the last year or within the 3 months prior randomization.

   • Medical conditions likely to limit life span and/or increase risk of intervention

- Cardiovascular disease

- Hospitalization for treatment of heart disease or stroke in past 6 months

- New York Heart Association Functional Class > 2

- Left bundle branch block or third degree AV block

- Aortic stenosis

- Systolic blood pressure > 180 mmHg or diastolic blood pressure > 105 mmHg; subjects can be re-screened after treatment of their hypertension

- Renal disease (creatine ≥ 1.6 mg/dl for men or ≥ 1.5 mg/dl for women, or urine protein ≥ 2+)

- Anemia (hematocrit < 33% in men and < 30% in women); if the hematocrit increases above these levels at a later date, they can be included in the study

   • Hepatitis, based on history and/or serum ALT greater than 2.5 times the upper limit of normal

- Other gastrointestinal disease (pancreatitis, inflammatory bowel disease)

- Recent or significant abdominal surgery

- Pulmonary disease with dependence on oxygen or daily use of bronchodilators

- Chronic infection (e.g., HIV, active tuberculosis)

   • Conditions or behaviors likely to affect conduct of the trial

- Unwilling to accept treatment assignment by randomization

- Participation in another intervention research project that might interfere with completion of the study

- Weight loss of > 10% in past 6 months for any reason except postpartum weight loss

- Currently pregnant or within 3 months postpartum

- Currently nursing or within 6 weeks of having completed nursing

- Pregnancy anticipated during the course of the trial

- Unwilling to undergo pregnancy testing or report possible pregnancy promptly

- Unwilling to take precautions to avoid pregnancy if potentially fertile

   • Major psychiatric disorders

   • Excessive alcohol intake, either acute or chronic

   • Medications and medical conditions likely to confound the assessment for diabetes, including:

Thiazide diuretics at a dose greater than 25 mg/day

Non-cardioselective beta-blockers (individuals receiving treatment with a statin or fenofibrate will not be excluded as long as the dose has been stable for 3 months prior to randomization)

Glucocorticoids, systemic

Prescription weight-loss or weight-gain medications

   • Thyroid disease, suboptimally treated as indicated by abnormal serum thyroid-stimulating hormone

   • Other endocrine disorders (e.g. Cushing's syndrome, acromegaly)

   • Fasting plasma triglyceride > 400 mg/dl, despite treatment

   • Individuals with a history of bladder cancer

   • Individuals with hematuria at screening. However, subjects with hematuria may be randomized if the cause of the hematuria is found, treated, and thought unlikely to recur.