Abstract PD03-03: Metformin and Cancer Risk in Diabetic Patients: A Systematic Review and Meta-Analysis, with Special Emphasis to Breast Cancer

Abstract

Abstract Background: Metformin (M), an insulin-lowering agent, has been associated with decreased cancer risk in epidemiological studies in diabetic patients in comparison with other antidiabetic treatments. Methods: We performed a comprehensive literature search and meta-analysis of epidemiological studies to assess the effect of M on cancer incidence and mortality in diabetic patients, using Pubmed, ISI-Web of Science (Science Citation Index Expanded), Embase, and the Cochrane library until June 2010, with no language or time restrictions. A manual search was also done for references cited in the selected articles, reviews or books. Published independent reports with sufficient information to allow adequate risk estimation of cancer risk/mortality and a corresponding measure of uncertainty after M use compared with other diabetic treatments were reviewed. Association between M and cancer incidence/mortality was computed as a summary relative risk (SRR) with 95% confidence intervals. Random effects models were applied to take into account heterogeneity. Sensitivity analyses were carried out to verify stability of the estimates. Publication bias was investigated using funnel plots and the Macaskill regression test. Results: Eleven studies were selected for relevance in terms of intervention, population studied, independence and reporting of cancer incidence or mortality data, reporting 4042 cancer events and 529 cancer deaths. A significant 31% reduction (overall SRR=0.69, 95%CI, 0.61-0.79) was found in subjects taking M compared with other antidiabetic drugs. The protective effect was significant for pancreatic and hepatocellular cancer, and non-significant trends were noted for colon, and prostate and breast cancer (BC). A non significant protective effect of M on BC incidence was found (SRR=0.75, 95% CI, 0.44-1.29; p for heterogeneity = 0.09, I2=59%). Conclusion: Among diabetics, M use is associated with a significant inverse association with any cancer incidence and a promising trend on BC incidence as compared with other diabetic treatments, including insulin. Further prospective studies of M as a cancer preventive agent are warranted. For breast cancer, Libby et al (Diabetes Care 2009;32:1620-25) found a non-significant trend for a protective effect on M users vs. not users (HR=0.6, 0.32-1.10), whereas Currie et al (Diabetologia 2009;52:1766- 77) noted a weak trend when M was a concomitant treatment among glargine users vs. all other insulin regimens (HR=0.88, 0.48-1.63), or when metformin was given as a monotherapy vs. insulin-based regimens (HR=0.93, 0.69-1.27). No effect was found when M monotherapy was compared with sulphonylureas (HR=1.02, 0.71-1.45). Bodmer et al (Diabetes Care 2010;33:1304-8) found a significant protective effect (OR=0.44, 0.24-0.82) only with a long term use of M, defined as equal to approximately 5+ years. The SRR for M and BC risk is plotted in figure. Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr PD03-03.