Is metformin associated with a lower prevalence of polyps, adenomas and colorectal carcinoma in patients with diabetes mellitus?

Abstract

Abstract Purpose. Recent studies suggested a protective role of metformin in the development of colorectal cancer (CRC) and its precursors. We aimed to investigate if metformin was associated with a lower prevalence and number of colorectal polyps in diabetic patients and, secondarily, adenomas, high-risk adenomas and CRC.Methods. Retrospective study on adult patients with diabetes mellitus followed in our hospital with a total colonoscopy between 2015–2019, treated with either metformin for > 5 years or other antidiabetic agent (control group). We assessed the number, size and histology of proliferative lesions detected on colonoscopy.Results. We included 401 patients aged 69 ± 9 years, 57% males, divided in two groups: treated with metformin (n = 260) and without (n = 140). The number of polyps detected was significantly lower in patients under metformin (p = 0.014). There was a nonsignificant trend towards lower polyp detection rates in the metformin compared to the control group both in unadjusted analysis (50% vs 60%, p = 0.058) and adjusted analysis (odds ratio [OR] 0.68, 95% confidence interval [CI] 0.43–1.09, p = 0.111). In the latter we identified male gender (OR 2.24, 95%CI 1.44–3.49, p < 0.001), age (OR 1.35 for every 10 years, 95%CI 1.07–1.71, p = 0.012), glycated hemoglobin value (OR 1.20 for every 1% increase, 95%CI 1.06–1.37, p = 0.005) and hypertension (OR 1.76, 95%CI 1.01–3.08, p = 0.046) as factors associated with a higher prevalence of polyps. We saw no statistically significant differences regarding adenoma (p = 0.231), high-risk adenoma (p = 0.810) and CRC (p = 0.705) diagnoses between groups.Conclusion. In our study, metformin was associated with less colorectal polyps in diabetic patients compared to other treatment modalities. The polyp detection rates, although lower in the metformin group, had no statistically significant differences between groups both in unadjusted and adjusted analysis.