This is a study with an approved drug for treating type 2 diabetes, for its effects on treating glucose and lipid abnormalities in patients being treated with first or second-generation antipsychotics, and comparison of effects of this drug with another treatment lifestyle modification. Patients who meet inclusion criteria will be treated with pioglitazone for 12 weeks. They will be evaluated for fasting glucose and lipids, glucose-tolerance tests, and neurocognitive battery and tests of verbal memory at baseline and during treatment with pioglitazone.

The aim of this study is to investigate the effects of pioglitazone added to weight-lifestyle intervention vs. placebo plus lifestyle intervention on reversing or reducing impaired or abnormal triglycerides, HDL and glucose metabolism in schizophrenics treated with first or second-generation antipsychotics.. Another aim is to examine the effects of impaired glucose metabolism on verbal memory and other cognitive function in schizophrenic patients treated with these medications and the relationship to improvements in impaired glucose metabolism to impairments in cognitive function. Clozapine and olanzapine, and some other first or second-generation antipsychotics effective for treating schizophrenia and bipolar disorders, have been reported to be associated with increased incidence of diabetic type metabolic abnormalities, decreases in insulin sensitivity, and abnormal glucose tolerance tests. This can lead to the development of type 2 diabetes and also abnormal lipid metabolic levels which can lead to atherosclerotic changes and increased risk of cardiovascular disease and other diabetes related complications. Drug treatments which could reduce or correct these diabetic metabolic changes would permit many patients to continue to receive the benefits of these antipsychotic medications with reduced drug-induced comorbidity. Previous research using non-psychotic subjects has shown that diabetes and impaired glucose tolerance are associated with cognitive impairments, especially in verbal memory, and provides a rationale for testing whether corrections of impaired glucose metabolism are associated with cognitive improvements in schizophrenic patients.

Inclusion Criteria:

1. Patients will be males or females, 18-70 yrs of age, with a diagnosis of schizophrenia or schizoaffective disorder, and currently being treated with olanzapine or clozapine.

2. Patients will have evidence of:

   1. glucose levels indicating at least impaired fasting glucose: fasting glucose 100 mg/dL or 2 hr glucose tolerance test 140 mg/dL, or current treatment with oral antidiabetic drugs with history of hyperglycemia;
   2. Triglyceride levels > 120 mg/dL and/or HDL levels < 40 mg/dL

Exclusion Criteria:

1. Diabetes mellitus, type 1
2. Recent diabetic ketoacidosis;
3. Patients not currently treated with oral antidiabetic drugs but fasting is glucose 140 mg/dL [WHO criteria] on repeat testing in last three months, or random blood glucose >200 mg/dL plus 2 hr glucose on GTT >200 mg/dL; (these patients may need more immediate treatment with antidiabetic drugs and it is less certain if weight-lifestyle treatment would be effective in treating such high glucose levels);
4. Patients with active liver disease with clinical abnormalities which need current treatment, or liver enzymes (Alt) 3 times upper limit for normal values in chart records in last year, or patients who are recorded as positive for hepatitis C;
5. Congestive heart failure (Class III or IV cardiac status) or history of MI in medical record (because pioglitazone can increase blood volume slightly);
6. Hematocrit greater than 10% below normal (hematocrit may be decreased 2 to 4% due to increased plasma volume);
7. Female patients on current oral contraceptives (because pioglitazone may interfere with effects of some oral contraceptives);
8. Patients taking ketoconazole,
9. Patients who have started on atorvastatin or gemfibrozil in the past 2 months or have had a dose increase in atorvastatin in the last month (since these drugs can also lower triglycerides and raise HDL, recent start of therapy with these drugs could be a confound).
10. Patients are not concomitantly treated with aripiprazole or ziprasidone.

Sufficient RBANS data at baseline and 3 months followup for China sample was not available for analysis. Subjects failed to complete scales or complete scales with sufficient data.