The incidence of congestive heart failure associated with antidiabetic therapies

BACKGROUND: Increased risk for CHF in persons with type 2 diabetes is well established. Our objectives were to estimate the CHF risk associated with specific therapies for diabetes and to determine the differences in incidence rates of CHF associated with adding various antidiabetic agents. METHODS: Subjects were members of the Kaiser Permanente Northwest (KPNW) diabetes registry as of 1 January 1998, with no prior history of CHF (n = 8063). We identified their therapy as of that date and then defined the start of the subject study period as the date when their drug regimen changed, either by switching to or by adding another antidiabetic drug. We defined the new therapy as the index therapy and the date of initiating the new therapy as the index date. Follow-up on the patients was done until the index therapy was discontinued or changed, or until 31 December 2002, whichever came earlier. We calculated the incidence rate of CHF in patients on various therapeutic regimens adjusting for age, gender, diabetes duration, existing ischemic heart disease, hypertension, renal insufficiency and glycemic control (HbA(1c)). RESULTS: CHF incidence rates were highest in index therapy categories that included insulin and lowest in regimens that included metformin. When insulin was added to an initial therapy, CHF incidence was increased 2.33 times (p < 0.0001) and 2.66 times (p < 0.0001) compared to the addition of sulphonylurea or metformin respectively. CONCLUSIONS: Our findings support the theory that elevated serum insulin levels promote the development of cardiac disease. Consistent with the UKPDS, metformin may offer some protection from incident CHF relative to sulphonylurea or insulin.