A total of 21 published articles were retrieved via both computerized searches and review of references. Estimates of OR/RR for highest versus non/lowest tea consumption levels were pooled on the basis of random effect model or fixed effect model as appropriate. Stratified analyses on tea type, population and study design were also conducted.

No statistical significance was detected between tea consumption and prostate cancer risk in meta-analysis of all included studies (odds ratio (OR) = 0.86, 95% CI (0.69-1.04)). Furthermore, stratified analyses on population (Asian, OR = 0.81, 95% CI (0.55-1.08); non-Asian, OR = 0.89, 95% CI (0.72-1.07)) and tea type (green tea, OR = 0.79, 95% CI (0.43-1.14); black tea, OR = 0.88, 95% CI (0.73-1.02)) also yielded non-significant association. Only the case–control study subgroup demonstrated a borderline protective effect for tea consumption against prostate cancer (OR = 0.77, 95% CI (0.55-0.98)).