OBJECTIVE: To review the effects of monotherapy with alpha-glucosidase inhibitors (AGIs) for patients with type 2 diabetes, with respect to mortality, morbidity, glycemic control, insulin levels, plasma lipids, body weight, and side effects. RESEARCH DESIGN AND METHODS: We systematically searched the Cochrane Central register of Controlled Trials, MEDLINE, EMBASE, Current Contents, LILACS, databases of ongoing trials, and reference lists, and we contacted experts and manufacturers. Inclusion criteria were randomized controlled trials of at least 12 weeks' duration, AGI monotherapy compared with any intervention, and one of the following outcome measures: mortality, morbidity, GHb, blood glucose, lipids, insulin levels, body weight, or side effects. Two independent reviewers assessed all abstracts, extracted all data, and assessed quality. We contacted all authors for data clarification. Continuous data were expressed as weighted mean differences and analyzed with a random-effects model. Possible influences of study characteristics and quality were assessed in sensitivity and meta-regression analyses. RESULTS: Forty-one studies were included in the review (30 acarbose, 7 miglitol, 1 voglibose, and 3 combined), and heterogeneity was limited. We found no evidence for an effect on mortality or morbidity. Compared with placebo, AGIs had a beneficial effect on GHb (acarbose -0.77%; miglitol -0.68%), fasting and postload blood glucose and postload insulin. With acarbose dosages higher than 50 mg t.i.d., the effect on GHb was the same, but the occurrence of side effects increased. Acarbose decreased the BMI by 0.17 kg/m2 (95% CI 0.08-0.26). None of the AGIs had an effect on plasma lipids. Compared with sulfonylurea, AGIs seemed inferior with respect to glycemic control, but they reduced fasting and postload insulin levels. For comparisons with other agents, little data were available. CONCLUSIONS: We found no evidence for an effect on mortality or morbidity. AGIs have clear beneficial effects on glycemic control and postload insulin levels but not on plasma lipids. There is no need for dosages higher than 50 mg acarbose t.i.d.