Metformin has been proposed in almost all the guidelines to be the initial therapy in diabetes and to be combined with insulin whenever indicated. However, sulfonyl ureas being very popular, very cheap and being in the market even before metformin was introduced, is still started as first drug in diabetes in many clinics across the world. The following study was presented at 49th EASD Annual Meeting providing evidence of higher rates of death when sulfonyl ureas were started alone at the time of diagnosis of diabetes.(eg. glibenclamide)

In a retrospective analysis of UK Clinical Practice Research Datalink (CPRD), patients who were initially treated with a sulfonylurea had a 58% higher risk of all-cause mortality than patients given first-line therapy with metformin (95% CI 1.48 to 1.68, P<0.001). In the assessment of the safety of sulfonylureas as a first-line therapy, 15,687 type 2 diabetes patients treated with the drugs as monotherapy were identified and as a comparison, 76,811 diabetic patients who were treated with Metformin were used each of whom were followed for an average of 2.9 years. There were 13.6 deaths per 1000 person-years with metformin and 44.6 deaths per 1000 person-years with sulfonylureas. This translated into a 58% increased risk for all-cause mortality among diabetics treated with sulfonylureas as monotherapy.The present study showed that approximately 17% of patients sampled were still receiving sulfonyl ureas as monotherapy.

Craig Currie, PhD, an epidemiologist at Cardiff University in Wales, and colleagues also analyzed mortality data when sulfonylureas were used as add-on second-line therapy versus DPP-4 inhibitors as add-on therapy. They compared second-line combination therapy in 33,983 patients treated with metformin and sulfonylureas against combination therapy in 7864 patients treated with metformin and a DPP-4 inhibitor. In total, there were 16.9 deaths per 1000 person-years in the metformin/sulfonylurea patients and 7.3 deaths per 1000 person-years in the metformin/DPP-4 group. This translated into a 35% increased risk for all-cause mortality among those treated with metformin and sulfonylureas compared with the other group.

Currie cautioned that observational studies such as his do have inherent biases -- for instance, patients with worse disease may get higher doses of drugs -- but added that his team has done a significant amount of controlling for potential confounders. "In my view," he concluded, "the safety of sulfonylureas needs urgent evaluation because we are potentially increasing the risk of all-cause mortality."