M1185 Statin use Does Not Appear to Confer a Significant Protection Against the Risk of Colorectal Cancer (CRC): A Meta-Analysis of Eleven Randomized Clinical Trials (RCT)

G A A b st ra ct s data emanating from randomized clinical trials (RCTs) and epidemiological studies. We gathered summary data in a meta-analysis designed to better assess the effect of statins on the risk of developing CRC using epidemiological study data. Methods: Systematic searches were carried querying EMBASE, MEDLINE, CENTRAL, and ISI Web of knowledge. We selected epidemiological studies (non RCTs) published fully or presented in abstract form at major GI meetings. If needed, authors were contacted for additional information. The primary outcome was the occurrence of CRC, or a digestive cancer, depending on available information. Heterogeneity was sought and its presence or absence guided analytical model selection. Results are expressed as adjusted risk ratios (aRR) with their 95% confidence intervals (95%CI). Adapted sensitivity analyses were performed. Analyses were completed using SAS version 9.1 and RevMan 5. Results: 13 case-control (82,1416 patients) and 8 cohort studies (784,818 patients) were included. Statistical heterogeneity was noted in the cohort studies, which is why we adopted random effect models for this analysis; it was not the case for the case-control studies that we analyzed using fixed-effect models. Analysis of the case-control studies suggest a significant beneficial effect attributable to the statins compared to controls (aRR=0.92 [0.90-0.94] ; only a trend to a decrease was noted using cohort (RR=0.89 [0.75-1.05]. When all the studies were analyzed together, a significant but modest decrease was noted in the statin users (RR=0.92 [0.87, 0.98]). Conclusions: This meta-analysis suggests that prolonged statin use may provide a small but significant beneficial effect at decreasing the risk of CRC occurrence. The heterogeneity of the included studies limits any conclusions that can be drawn. The magnitude of the protective effect does not support promulgation of statins in the chemoprevention of CRC, at least in an average risk population.