Effect of aspirin and other non-steroidal anti-inflammatory drugs on prostate cancer incidence and mortality: a systematic review and meta-analysis

Background: It has been postulated that non-steroidal anti-inflammatory drugs (NSAIDs) use leads to decreased prostate cancer (PCa) risk. In recent years, NSAIDs' role in PCa development has been extensively studied; however, there is not yet a definitive answer. Moreover, the epidemiological results for NSAIDs' effect on PCa-specific mortality have been inconsistent. Therefore, we performed a meta-analysis to examine the controversy. Methods: We performed a literature database search and included all published studies conducted in the general population exposed to any NSAID, extracting an odds ratio (OR) or a hazard ratio (HR) with 95% confidence intervals (95% CIs) that compared the incidence of PCa or PCa-specific mortality with non-exposure. We derived a pooled OR or HR using random or fixed effects models, as appropriate. Subgroup analyses were also performed. Results: Thirty-nine studies (20 case-control and 19 cohort studies) were included in this analysis. Thirty-one studies were available concerning NSAID use and PCa incidence and eight studies on PCa-specific mortality. Compared to non-use, aspirin use was statistically significantly associated with PCa incidence risk, and the association was slightly stronger for advanced PCa than for total PCa (OR = 0.92, 95% CI = 0.87 to 0.97 for total PCa; OR = 0.81, 95% CI = 0.73 to 0.89 for advanced PCa). Aspirin use seems also to be associated with a modest reduction in PCa-specific mortality (HR = 0.86, 95% CI = 0.78 to 0.96 for total PCa; OR = 0.81, 95% CI = 0.71 to 0.92 for advanced PCa). Generally, the pooled effects for any NSAIDs, NA-NSAIDs and cyclooxygenase-2 inhibitors demonstrated no adverse or beneficial effects on PCa development or PCa-specific mortality, but the results were not consistent. The effect estimates did not vary markedly when stratified by study design and study quality but varied by geographic region. Furthermore, long-term aspirin use (= 4 years) was also significantly associated with reduced PCa incidence (OR = 0.88, 95% CI 0.79 to 0.99). Conclusions: The present meta-analysis provides support for the hypothesis that aspirin use is inversely related to PCa incidence and PCa-specific mortality. The effect estimates, varying by geographic region, deserve further investigation.