Different effects of metformin and insulin on primary and secondary chemoprevention of colorectal adenoma in diabetes type 2: Traditional and Bayesian meta-analysis.

Metformin and insulin differently affect the risk of colon cancer in type 2 diabetic patients, however their effects on colon adenoma is not clear. PubMed, ISI, Scopus and Cochrane databases were searched for studies reporting. The outcomes were total adenoma; advanced adenoma and recurrent adenoma. Traditional and Bayesian meta-analysis were conducted via random-effects models. Odds ratios (OR) with 95 % confidence intervals (CIs) / or credible intervals (CrI) were used to describe the ratio of different events. A random-effects model described by DerSimonian and Laird was performed, when significant between-study heterogeneity existed. Alternatively, an inverse variance fixed effects model was used, when there was no significant heterogeneity across studies. The potential publication bias was assessed with funnel plot, Egger and Begg's regression asymmetry tests. Moreover, "trim and fill" procedure was used to assess the possible effect of publication bias. For metformin intake, 11 studies (51991 patients) were included. The results showed that metformin significantly decreased the risk of advance adenoma (OR= 0.51, p< 0.001). The risk of total adenoma was not associated with metformin use (OR= 0.86, p=0.274), and metformin did not affect the risk of adenoma recurrence (OR= 0.89, p=0.137). Five studies (2678 patients) were included in the analysis to determine the effect of insulin therapy. Insulin significantly increased the risk of colorectal adenoma (OR= 1.43, p=0.002). These findings indicate that metformin has no protective effect on total and recurrent adenoma, whilst it significantly reduces the risk of advanced adenoma, but insulin increases the risk of total adenoma.