Multiple outcomes associated with the use of metformin and sulphonylureas in type 2 diabetes: a population-based cohort study in Italy

To compare the effect of metformin and sulphonylureas on the risks of switching to insulin therapy, hospitalisation for macrovascular disease and all-cause mortality. The 70,437 residents of the Italian Region of Lombardy aged 40 to 90 years who started diabetes treatment with metformin or sulphonylureas during 2001-2003 entered the study and were followed until July 2007. We estimated the effects of the first-line agent, early compliance, and persistence with first-line therapy on the risks of switching to insulin, hospitalisation for macrovascular disease and all-cause mortality, by fitting a multistate model and adjusting for age, gender and selected clinical factors. Compared with patients who started on metformin, those who started on sulphonylureas were at a higher risk of switching to insulin (adjusted hazard ratio and 95% CI, 1.55; 1.43, 1.68), hospitalisation (1.15; 1.08, 1.21), and death (1.37; 1.26, 1.49). Compared with patients who stayed on sulphonylureas for 3 months or less, those on sulphonylureas for more than 9 months had an adjusted hazard ratio of 1.24 (1.13, 1.35) for switching to insulin and 1.14 (1.05, 1.23) for hospitalisation. The risks of switching to insulin and hospitalisation were both increased among patients who switched from metformin to another oral hypoglycaemic agent or combined initial monotherapy with another agent. Our study provides evidence that the risks of switching to insulin, hospitalisation because of macrovascular events and death changes according to the first prescribed oral hypoglycaemic agent, as well as to the early compliance and persistence with such agent.