Metformin for chemoprevention of metachronous colorectal adenoma or polyps in post-polypectomy patients without diabetes: a multicentre double-blind, placebo-controlled, randomised phase 3 trial

Background The prevalence of, and mortality from, colorectal cancer is increasing worldwide, and new strategies for prevention are needed to reduce the burden of this disease. The oral diabetes medicine metformin might have chemopreventive effects against cancer, including colorectal cancer. However, no clinical trial data exist for the use of metformin for colorectal cancer chemoprevention. Therefore, we devised a 1-year clinical trial to assess the safety and chemopreventive effects of metformin on sporadic colorectal cancer (assessed by adenoma and polyp recurrence) in patients with a high risk of adenoma recurrence. Methods This trial was a multicentre, double-blind, placebo-controlled, randomised phase 3 trial. Non-diabetic adult patients who had previously had single or multiple colorectal adenomas or polyps resected by endoscopy were enrolled into the study from five hospitals in Japan. Eligible patients were randomly assigned (1:1) to receive oral metformin (250 mg daily) or identical placebo tablets by a stratified computer-based randomisation method, with stratification by institute, age, sex, and body-mass index. All patients, endoscopists, doctors, and investigators were masked to drug allocation until the end of the trial. After 1 year of administration of metformin or placebo, colonoscopies were done to assess the co-primary endpoints: the number and prevalence of adenomas or polyps. Our analysis included all participants who underwent random allocation, according to the intention-to-treat principle. This trial is registered with University Hospital Medical Information Network (UMIN), number UMIN000006254. Findings Between Sept 1, 2011, and Dec 30, 2014, 498 patients who had had single or multiple colorectal adenomas resected by endoscopy were enrolled into the study. After exclusions for ineligibility, 151 patients underwent randomisation: 79 were assigned to the metformin group and 72 to the placebo group. 71 patients in the metformin group and 62 in the placebo group underwent 1-year follow-up colonoscopy. The prevalence of total polyps (hyperplastic polyps plus adenomas) and of adenomas in the metformin group was significantly lower than that in the placebo group (total polyps: metformin group 27 [38·0%; 95% CI 26·7–49·3] of 71 patients, placebo group 35 [56·5%; 95% CI 44·1–68·8] of 62; p=0·034, risk ratio [RR] 0·67 [95% CI 0·47–0·97]; adenomas: metformin group 22 [30·6%; 95% CI 19·9–41·2] of 71 patients, placebo group 32 [51·6%; 95% CI 39·2–64·1] of 62; p=0·016, RR 0·60 [95% CI 0·39–0·92]). The median number of polyps was zero (IQR 0–1) in the metformin group and one (0–1) in the placebo group (p=0·041). The median number of adenomas was zero (0–1) in the metformin group and zero (0–1) in the placebo group (p=0·037). 15 (11%) of patients had adverse events, all of which were grade 1. We recorded no serious adverse events during the 1-year trial. Interpretation The administration of low-dose metformin for 1 year to patients without diabetes was safe. Low-dose metformin reduced the prevalence and number of metachronous adenomas or polyps after polypectomy. Metformin has a potential role in the chemoprevention of colorectal cancer. However, further large, long-term trials are needed to provide definitive conclusions. Funding Ministry of Health, Labour and Welfare, Japan.