BACKGROUND AND AIMS: Genome-wide association studies found some variants on chromosome 9p21 associated with type 2 diabetes (T2D). We performed a meta-analysis to estimate strength, accuracy and feature of the association of polymorphisms in 9p21 with T2D. METHODS AND RESULTS: Articles were retrieved screening electronic databases and cross references. Twenty-two publications were identified, for a total of 38,455 T2D patients and 60,516 controls. Twenty-one studies investigated the role of the SNP rs10811661; in some studies three additional SNPs (rs564398, rs10757278, rs1333040) were genotyped. Population attributable risk (PAR) was computed as: risk allele frequency∗(OR-1)/OR, using the per-allele odds ratio (OR). The risk allele (T) of rs10811661 was associated with T2D in most of the studies. In meta-analysis the overall per-allele OR was 1.24 (95% CI: 1.21-1.27; P