Altered expression of central muscarinic and nicotinic acetylcholine receptors in hippocampal and cortical regions may contribute to the cognitive impairment exhibited in patients with schizophrenia. Increasing cholinergic activity through the use of a cholinesterase inhibitor (ChEI) therefore represents a possible strategy for cognitive augmentation in schizophrenia. We examined the efficacy and safety of the ChEI donepezil as cotreatment for mild to moderate cognitive impairment in schizophrenia or schizoaffective disorder in a prospective, 12-week, placebo-controlled, double-blind, parallel-group study. In total, 250 patients (18-55 years) with schizophrenia or schizoaffective disorder who were clinically stabilized on risperidone, olanzapine, quetiapine, ziprasidone, or aripiprazole, alone or in combination, were enrolled at 38 outpatient psychiatric clinics in the United States. Patients were randomized to donepezil 5 mg q.d. for 6 weeks then 10 mg q.d. for 6 weeks, or placebo administered as oral tablets. The primary outcome measure was the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) neurocognitive battery composite score. In the intent-to-treat sample (donepezil, n ¼ 121; placebo, n ¼ 124), both treatments showed improvement in the composite score from baseline to week 12. At week 12, cognitive improvement with donepezil was similar to that with placebo (last-observation-carried-forward effect size, 0.277 vs 0.411; p ¼ 0.1182) and statistically significantly inferior for the observed-cases analysis (0.257 vs 0.450; p ¼ 0.044). There was statistically significant improvement in the Positive and Negative Syndrome Assessment Scale negative symptoms score for placebo compared with donepezil, while total and positive symptom scores were similar between both treatments. Statistically significant improvements in positive symptoms score and Clinical Global Impression-Improvement for donepezil compared with placebo were noted at Week 6. Treatment-emergent adverse events (AEs) were observed for 54.5% of donepezil-and 61.3% of placebo-treated patients; most AEs were rated as mild to moderate in severity. Donepezil was safe and well-tolerated but was not effective compared with placebo as a cotreatment for the improvement of cognitive impairment in this patient population. A significant and surprisingly large placebo/practice effect was observed among placebotreated patients, and is a serious consideration in future clinical trial study designs for potential cognitive enhancing compounds in schizophrenia.