The tumor suppressor gene p53 and its downstream effector p21 are thought to play major roles in the development of human malignancy. Polymorphic variants of p53 at codon 72, and p21 at codon 31, have been found to be associated with cancer susceptibility, but few studies have investigated their effect on prostate cancer risk. In this case-control study, we investigated the association of p53 codon 72 and p21 codon 31 polymorphisms with prostate cancer risk in a Taiwanese population. In total, 200 patients with prostate cancer, 247 age-matched male controls, and 181 non-age-matched symptomatic benign prostatic hyperplasia (BPH; American Urological Association symptom score > or = 8 and prostate volume > 20 gm) recruited from two medical centers in southern Taiwan were genotyped. Overall, we found no significant association between p53 polymorphism and risk of prostate cancer. However, for p21 polymorphism, the frequencies of p21 Ser/Ser, Ser/Arg and Arg/Arg were 52 (26.0%), 85 (42.5%), 63 (31.5%) in case patients, 48 (26.5%), 82 (45.3%), 51 (28.2%) in BPH patients, and 76 (30.8%), 119 (48.2%), 52 (21.1%) in controls, respectively. Among the prostate cancer cases and controls, subjects with Arg/Arg genotype were found to have a 1.78-fold increased risk [95% confidence interval (CI), 1.06-3.01] of developing prostate cancer compared with those having the Ser/Ser genotype, after adjusting for other potential covariates. This significant association was slightly stronger [odds ratio (OR), 2.13; 95% CI, 1.16-3.92] in younger men ( 50 mL) compared with those having the Ser/Ser genotype [OR, 2.29; 95% CI, 1.07-4.98]. Our findings suggest that the p21 codon 31 polymorphism may be associated with the development of prostate enlargement and cancer.