In a case-control study in a veterans hospital in Taiwan, we compared 237 histology-confirmed prostate carcinoma cases with 481 controls, frequency matched by age, for their consumption of vegetarian food, namely soybean products, rice, wheat protein and other vegetables. The multivariable logistic regression analysis showed a significant association with such food (odds ratio (OR)=0.67, 95% confidence interval (CI)=0.47, 0.94). This beneficial effect presented for men with body mass index (BMI) < or =25 kg m(-2) (OR=0.50, 95% CI=0.32, 0.76) but not for men with greater BMI. The OR of prostate carcinoma for men with BMI < or =25 kg m(-2) was 1.74 (95% CI=1.21, 2.51), compared with men with higher BMI (>25 kg m(-2)). Other significant risk factors associated with the disease included higher income (OR=2.40, 95% CI=1.07, 5.42), physical activity (OR=1.75, 95% CI=1.08, 2.83), being married (OR=2.49, 95% CI=1.40, 4.43) and coffee consumption (OR=1.88, 95% CI=1.07, 3.30). Stratified analysis also showed that the consumption of fish/shellfish had an adverse association for men with higher BMI. This study suggests that the intake of the low fat local vegetarian food has a protective effect against prostate carcinoma for thin men in this study population.