Seventeen studies were included in the review (37,632 cancers): 11 studies looked at both treatments and six studies looked at metformin alone. Seven studies looked at cancer at any site (3,931 cancers) and 10 studies looked at specific cancer sites (33,701). Study sample sizes ranged from 112 to 24,723 cancers.
Compared with control, metformin was associated with statistically significantly reduced risk of all cancers (RR 0.61, 95% CI 0.54 to 0.70; Ι²=84%), colorectal cancer (RR 0.64, 95% CI 0.54 to 0.76; five studies; Ι²=19%) and pancreatic cancer (HR 0.38, 95% CI 0.14 to 0.91; four studies; Ι²=89%). There was no evidence that metformin was associated with risk of breast or prostate cancer. Sulphonylurea was not associated with risk of cancer at any site (RR 0.97, 95% CI 0.82 to 1.14; Ι²=0%).
There was evidence of publication bias with metformin.