OBJECTIVETo explore the association of the androgenic receptor (AR) CAG repeats with the risks of benign prostatic hyperplasia (BPH) and prostate cancer (PCa).

METHODSWe searched the major databases at home and abroad for the literature addressing the correlation of the AR gene CAG repeats with BPH and PCa. Based on the results of heterogeneity tests, we used the M-H fixed effect model and random effect model to pool the odds ratio (OR) effect size. We evaluated publication bias by Begg and Egger bias analysis, investigated the association of CAG repeats with the risks of BPH and PCa by systematic review, and stratified their relationship according to the races of the patients.

RESULTSBased on the selection criteria, 4 of the 29 identified studies were included, with 485 cases of BPH, 767 cases of PCa, and 709 controls. There was no heterogeneity between the BPH and control groups, and no correlation between short CAG repeats and BPH after pooling the odds ratio (OR) effect size. Heterogeneity was found among the BPH, PCa and control groups. Random effects model suggested an association of short CAG repeats with the risk of PCa (OR(PCa/control) = 1.45, OR(PCa/BPH) = 1.86, OR(PCa/(BPH + control)) = 1.66), while subgroup analysis with racial stratification indicated inter-ethnic differences between the two. Begg and Egger bias analysis showed no significant publication bias.

CONCLUSIONShorter CAG repeats are positively correlated with the risk of PCa but not with that of BPH.