Aims/hypothesis. Several case-control studies have examined the association between the Gly972Arg variant in the IRS-1 gene and Type 2 diabetes, but most had limited power and results could therefore be conflicting. Methods. We systematically reviewed the literature by means of a meta-analysis and investigated sources of heterogeneity in results of different studies.Results. The summary risk ratio, based on 3408 cases and 5419 control cases from 27 studies, was 1.25 (95% CI 1.05-1.48). The results, however, differed according to the type of study, method of verifying non-diabetic status of the control subjects, and age of the case subjects. Population-based studies reported lower odds ratios than hospital-based studies (OR 0.98, 95% CI 0.74-1.30 vs OR 1.43, 95% CI 1.17-1.74). Also, the diagnostic test to exclude diabetes amongst control subjects interacted with the association between the IRS-1 Gly972Arg variant and Type 2 diabetes (p=0.03). Finally, the odds ratio reduced with increasing age (p=0.03). Conclusion/interpretation. Overall, carriers of the 972Arg variant of the IRS-1 gene are at a 25% increased risk of having Type 2 diabetes compared with non-carriers. The odds ratios are generally higher in hospital-based studies, including relatively young, symptomatic, cases. [Diabetologia (2003) 46:990-995]