Currently available epidemiologic evidences concerning the chemopreventive effect of aspirin on ovarian cancer are inconsistent. Therefore, we aimed to further explore the association by synthesizing evidence from population-based studies.

We searched PubMed, EMBASE, Web of Science, and Scopus using key words and controlled vocabularies. Title/abstract screening, full-text review, data extraction, and quality assessment were performed independently by reviewers, and a random-effects model was utilized for meta-analysis. Subgroup analysis was conducted based on study locale, and sensitivity analysis was performed by synthesizing studies that adjusted for certain covariates or studies with good quality. Dose-response relation was assessed by a two-stage linear dose-response model. Statistical heterogeneity was evaluated by the I-squared value and a chi-squared test for the Cochrane Q statistic.

We identified 8 cohort studies and 15 case-control studies. In overall meta-analysis of risk ratios (RRs) of binary exposure, the synthesized RR was 0.89 (95%CI, 0.83-0.96), and no substantial statistical heterogeneity was observed (I(2)=22.5%, PCochrane=0.168). After stratification by study design, the synthesized RR was 0.85 (95%CI, 0.77-0.94) and 0.95 (95%CI, 0.85-1.05) for case-control and cohort studies, respectively. In sensitivity analysis, the synthesized estimate of long-term use was not statistically significant, whereas the effect measure (RRmeta=0.60, 95%CI, 0.39-0.93) was significant by synthesizing RRs of the highest frequency of use from 2 cohort studies. The dose-response analysis showed an inverse significant association between aspirin use and the risk (RRper 1 time/wk=0.94, 95%CI, 0.89-1.00; n=2). Egger's tests showed that publication bias existed for overall meta-analysis, meta-analysis for case-control studies, and studies conducted in the United States.

In summary, our study suggests that aspirin can reduce the risk of ovarian cancer. In addition, we observed a possible dose-response relation between frequency of use and ovarian cancer risk, but further studies are needed to examine this association.