Quantitative Synthesis the Strength of Evidence and Extent of Potential Biases in the Claimed Associations Between Metformin Use on the Risk and Prognosis of Cancer

Jianhong Zhu,Jianfang Li,Chen Su,Zhichao He,Yin Lin,Xiaoxia Yu,Kaifeng Qiu,Wenxia Zhao,Zepeng Chen,Junyan Wu

DOI: https://doi.org/10.2139/ssrn.3314452

2019-01-01

Abstract:Background: The claimed associations between metformin and cancer outcomes have not been documented. We aimed to summarize the evidence of meta-analysis of metfoemin on cancer outcomes, describe the magnitude, direction, and evaluate the potential biases in the claimed associations. Methods: We searched PubMed, Embase, Cochrane Library to identify meta-analyses that examined associations between metformin and cancer outcomes. For each meta-analysis, we estimated the summary effect size by use of random-effects model, the 95% CI, and the 95% prediction interval. We measured the heterogeneity by the τ2 statistic and the I2 statistic, publication bias by Egger test, quality of included meta-analyses by AMSTAR score, and evidence of outcomes by GRADE methodology (grading of recommendations assessment, development, and evaluation). Findings: The umbrella review identified 21 meta-analyses with 31 unique cancer outcomes. Metformin use was associated with significant lower incidence of cancer at the site of colorectal liver, lung, pancreatic, gastric, head and neck, stomach, oesophagus. Metformin prolonged overall survival in patients with colorectal cancer, breast cancer, endometrial cancer, liver cancer, lung cancer, pancreatic cancer, prostate cancer, head and neck. Finally, there seems to be beneficial associations between metformin use and cancer-specific survival of colorectal cancer and bladder cancer. The quality of evidence was low or very low according to the GRADE methodology. Interpretation: Metformin use may improve certain cancer outcomes, but robust randomized controlled trials are needed to understand whether the observed associations are causal. Funding: This work was supported by Grant [2013]163 from Key Laboratory of Malignant Tumor Molecular Mechanism and Translational Medicine of Guangzhou Bureau of Science and Information Technology; Grant KLB09001 from the Key Laboratory of Malignant Tumor Gene Regulation and Target Therapy of Guangdong Higher Education Institutes. Declaration of Interest: We declare that we have no conflicts of interest.