Elevated plasma levels of plasminogen activator inhibitor-1 (PAI-1) increase risk for type 2 diabetes. The PAI-1 4G/5G polymorphism is a major genetic determinant of plasma PAI-1 levels, with 4G/4G homozygotes having elevated PAI-1 levels relative to 5G allele carriers. These observations suggest the hypothesis that the PAI-1 4G/5G polymorphism could be a genetic risk factor for diabetes. We tested this hypothesis among 2169 participants of the Framingham Offspring Study followed for seven examinations over 26 years for 216 cases of type 2 diabetes. PAI-1 4G/4G homozygotes (genotype frequency, 27.4%) were not at significantly (p > 0.05) increased risk of incident diabetes compared with 5G allele carriers and did not have elevated levels of diabetes-related quantitative traits including BMI, fasting plasma glucose, or fasting insulin. In proportional hazards regression models accounting for correlation among siblings, with the 5G/5G genotype as the referent, the hazard ratio for incident diabetes for 4G/5G carriers was 0.93 (95% confidence interval, 0.68 to 1.28) and for 4G/4G carriers was 1.20 (95% confidence interval, 0.83 to 1.92). Results were not altered by further adjustment for sex or levels of BMI, triglycerides, or PAI-1. We conclude that the PAI-1 4G/5G polymorphism is not an important genetic risk factor for type 2 diabetes in this community-based sample. Elevated PAI-1 levels may be associated with an increased risk for diabetes as a marker for underlying endothelial dysfunction rather than by a direct effect of genetically mediated elevated levels.