Nonlinear reduction in risk for colorectal cancer by oral contraceptive use: a meta-analysis of epidemiological studies.

Abstract

Although the relationship between oral contraceptive (OC) use and colorectal cancer (CRC) risk has been studied extensively, the results of epidemiological studies are controversial. Therefore, we carried out a meta-analysis of epidemiological studies to summarize the available evidence and to quantify the potential dose-response relation. We searched PubMed database for studies of OC use and CRC risk that were published until the end of March 2014. Random- and fixed-effects models were applied to estimate summary relative risks (RRs) and 95 % confidence intervals (CIs). Twelve cohorts and seventeen case-control studies with a total of 15,790 CRC cases were included in the final analysis. The summary RR for the ever versus never category of OC use was 0.82 (95 % CI 0.76-0.88). Similar result was observed when we compared the longest duration of OC use with the shortest duration (RR = 0.86, 95 % CI 0.76-0.96). Furthermore, the results of stratified analysis were comparable to those of overall meta-analysis. In dose-response analysis, significant inverse associations emerged in nonlinear models for the duration of OC use and CRC (P nonlinearity = 0.001). The greatest risk reduction was observed when the duration of OC use was approximately 42 months. There was moderate heterogeneity in the analysis, and no evidence of small-study bias was observed. Based on the findings of this meta-analysis, ever use of OC is associated with lower risk of CRC. Additionally, there is a statistically significant nonlinear inverse association between the duration of OC use and CRC risk.