This study expands a previous study of NAT2 polymorphisms and bladder cancer in male subjects occupationally exposed only to benzidine. The combined analysis of 68 cases and 107 controls from a cohort of production workers in China exposed to benzidine included 30 new cases and 67 controls not previously studied. NAT2 enzymatic activity phenotype was characterized by measuring urinary caffeine metabolite ratios. PCR-based methods identified genotypes for NAT2, NAT1 and GSTM1. NAT2 phenotype and genotype data were consistent. A protective association was observed for the slow NAT2 genotype (bladder cancer OR = 0.3; 95% CI = 0.1 = 1.0) after adjustment for cumulative benzidine exposure and lifetime smoking. Individuals carrying NAT1wt/*10 and NAT1*10/*10 showed higher relative risks of bladder cancer (OR = 2.8, 95% CI = 0.8-10.1 and OR = 2.2, 95% CI = 0.6-8.3, respectively). No association was found between GSTM1 null and bladder cancer. A metaanalysis risk estimate of case-control studies of NAT2 acetylation and bladder cancer in Asian populations without occupational arylamine exposures showed an increased risk for slow acetylators. The lower limit of the confidence interval (OR = 1.4; 95% CI = 1.0-2.0) approximated the upper confidence interval for the estimate obtained in our analysis. These results support the earlier finding of a protective association between slow acetylation and bladder cancer in benzidine-exposed workers, in contrast to its established link as a risk factor for bladder cancer in people exposed to 2-naphthylamine and 4-aminobiphenyl. Study findings suggest the existence of key differences in the metabolism of mono- and diarylamines.