Le risque de cancer associé aux médicaments du diabète de type 2 (metformine, sulfonylurées et insulines)

Many observational and meta-analysis studies suggested a decreased risk of cancer in patients treated with metformin. However multiple biases have been related to these studies which are known to increase downward the effect of the drug, thus making metformin seems to be protective when in fact it may have less or no effect. Large randomized studies on diabetes treatment did not confirm any protective effect for metformin. The potential effect of sulfonylureas and glinides on cancer risk and mortality has been suggested in several observational studies mainly showing an increased risk for colorectal and pancreatic cancers. The randomized trials ADOPT and RECORD did not find such an exaggerated risk. Two meta-analysis have given divergent results with regard to cancer risk. The first one showing an increased risk for cancer in users of sulfonylureas with or without insulin compared to non users. The second meta-analysis found no risk when users of sulfonylureas without insulin were compared to others patients. A cohort study in the French National Healthcare Insurance Database between 2003 and 2010 suggested a two-fold risk of cancer in patients more exposed to sulfonylureas. It appears difficult to know whether cancer risk excess is due to sulfonylureas itself or to associated insulin treatment, or yet explained by the lower risk seen in control group treated with metformin. According to the results of observational studies, two meta-analysis found a higher risk for any cancer in insulin users. French National Healthcare Insurance Database likewise suggested a two-fold increased risk of cancer in patients highly exposed to insulin. Main criticism pointed out these studies have potential biases as insulin-treated patients have a diabetes with a longer duration and more severe comorbidities (among them cancer). The initial warning on insulin glargine appeared to be erased whereas it remains necessary to have more long term safety studies. In diabetics with high risk or antecedents for cancer, epidemiological, clinical trials and experimental data converge for first promoting combining and reinforcing therapies against insulin resistance.