This study supports the hypothesis that users of metformin are at lower risk of cancer compared with people with type 2 diabetes on other treatments. Fewer than 8% of a cohort of metformin users were diagnosed with cancer during a maximum of 10 years of follow-up, compared with 11% of a comparator cohort of nonusers. The median time to cancer was 3.6 years among metformin users, compared with 2.5 years among comparators, and they also had reduced overall and cancer-related mortality.
This was an observational study; therefore, we could not control for all differences between study groups. Metformin users could have been at lower baseline risk of cancer than the comparators. Indeed, they were younger than their comparators (but mean BMI and A1C were higher). Metformin users did seem to be a different group clinically from nonusers, with a much lower rate of mortality (some of which could be explained by lower risk of cardiovascular mortality [16
]). Although this limitation is inherent in the observational nature of the study, we adjusted results for known potential confounders and there were sizable changes to the risk estimates. There may still have been residual confounding or unknown confounders, but it is unlikely that this could account for the entire 37% reduced risk of cancer observed.
Adjusting for use of other diabetic drugs was necessary because there was a higher proportion of metformin users who were treated with sulfonylureas compared with the comparators but a lower proportion treated with insulin. This probably reflects the heterogeneity of the pool of potential comparators. Patients not treated with metformin will encompass those who do not yet require oral therapy as well as those who have progressed to insulin after treatment with sulfonylureas only. However, we found no statistically significant independent effects of sulfonylureas and insulin on risk of cancer in the Cox regression analysis. In contrast, men and older people were at increased risk of cancer, as might be expected. The results were similar for specific cancer types.
In a dose-response analysis, metformin users appeared to have a higher risk of cancer during the first 2 years of follow-up. This may be because patients who begin treatment with metformin are more likely to have cancer diagnosed because they have increasing contact with health care professionals. In later years of follow-up, high maximum doses of metformin were associated with the greatest reduction in risk of cancer. Metformin dose usually increases with increasing duration of use; therefore, dose variables can be confounded by duration. This could produce a survival bias, with higher doses spuriously associated with reduced cancer because patients have survived to receive a higher dose. This is the reason for stratification by length of follow-up (although residual confounding may still be present).
Within the known limitations of observational data, we are confident in our study design and data sources. The data sources used were independent of each other, and they provided objective measures of exposure and outcome. The diabetic population of Tayside, Scotland, U.K., is well defined, and the MEMO database used to identify metformin users has been widely used for drug safety research (11
). The likelihood of misclassification of metformin exposure due to data error is low because we ensured that all patients had multiple metformin prescriptions. We were otherwise unable to judge whether patients actually took the metformin as prescribed, although we know that the drug was collected from the pharmacies (11
). We are confident that we eliminated survival bias in our choice of comparators. The national cancer registry (SMR6) was used to identify cancer diagnoses. Specificity is likely to be higher than sensitivity in this register, but if any cancer diagnoses were missed, this would not occur differentially with respect to metformin status.
This study has produced sufficient epidemiological evidence that metformin reduces the risk of cancer to make further investigation a high priority. A plausible biological mechanism hinges on the discovery that the upstream LKB1 regulator of AMPK is a tumor suppressor and that activation of AMPK by LKB1 plays an important role in inhibiting cell growth when cellular energy levels are low (17
). Metformin activates AMPK by inhibiting mitochondrial respiration and increasing 5′-AMP, which enhances activation of AMPK by LKB1 (1
). The blood glucose–lowering properties of metformin are mediated through AMPK restoring cellular energy levels by phosphorylating regulatory proteins that lead to stimulation of glucose uptake into muscle tissues as well as inhibition of gluconeogenesis in the liver. The anticancer properties of metformin are likely to be mediated by AMPK's ability to preserve cellular energy levels by phosphorylating proteins such as p27KIP and TSC2 that lead to inhibition of cell growth and proliferation signaling networks (18
Prior to the discovery that the LKB1 tumor suppressor activated AMPK, there was little interest in the role of AMPK in cancer. However, the ability of AMPK to gauge and control cellular energy places it in an ideal position to ensure that cell growth and proliferation is coupled to the availability of a sufficient supply of nutrients and energy. Recent laboratory evidence showing that three distinct drugs activate AMPK-delayed tumorigenesis in tumor-prone mice suggests that activators of AMPK could have therapeutic benefit for the treatment of cancer in humans (4
). The protective effects of metformin on cancer development could potentially be rapid and may occur at quite a late stage of cancer development. Treatment of animal cells with metformin significantly activates the AMPK pathway within 30 min (20
). Metformin also inhibits growth of cancer cells (5
) or mouse embryonic stem cells (4
) within 1–2 days. We believe that there is now a strong case to conduct a randomized trial to establish whether metformin is protective in a population at high risk for cancer.