A meta-analysis was performed to assess the association between the PAI-1 -675 4G/5G polymorphism and susceptibility to diabetes mellitus (DM), diabetic nephropathy (DN), diabetic retinopathy (DR) and diabetic coronary artery disease (CAD). A literature-based search was conducted to identify all relevant studies. The fixed or random effect pooled measure was calculated mainly at the allele level to determine heterogeneity bias among studies. Further stratified analyses and sensitivity analyses were also performed. Publication bias was examined by the modified Begg's and Egger's test. Twenty published articles with twenty-seven outcomes were included in the meta-analysis: 6 studies with a total of 1,333 cases and 3,011 controls were analyzed for the PAI-1 -675 4G/5G polymorphism with diabetes risk, 7 studies with 1,060 cases and 1,139 controls for DN risk, 10 studies with 1,327 cases and 1,557 controls for DR and 4 studies with 610 cases and 1,042 controls for diabetic CAD risk respectively. Using allelic comparison (4G vs. 5G), the PAI-1 -675 4G/5G polymorphism was observed to have no significant association with diabetes (REM OR 1.07, 95% CI 0.96, 1.20), DN (REM OR 1.10, 95% CI 0.98, 1.25), DR (REM OR 1.09, 95% CI 0.97, 1.22) or diabetic CAD risk (REM OR 1.07, 95% CI 0.81, 1.42), and similar results were obtained in the dominant, recessive and co-dominant models. Our meta-analyses suggest that the PAI-1 -675 4G/5G polymorphism might not be a risk factor for DM, DN, DR or diabetic CAD risk in the populations investigated. This conclusion warrants confirmation by further studies.