The aim was to investigate the outcomes of individual sulfonylureas in patients with heart failure (HF).
Methods
All patients hospitalized with HF for the first time in 1997–2006, alive 30 days after discharge, and who received anti-diabetic monotherapy with glimepiride (n = 1097), glibenclamide (glyburide) (n = 1031), glipizide (n = 557), gliclazide (n = 251), or tolbutamide (n = 541) were identified from nationwide registers. Risk of all-cause mortality was assessed by multivariable Cox regression models.

Results
Over the median observational time of 744 (Inter Quartile Range 268–1451) days, 2242 patients (64%) died. The analysis demonstrated similar hazard ratio (HR) for mortality for treatment with glimepiride (1.10 [95% confidence interval 0.92–1.33]), glibenclamide (1.12 [0.93–1.34]), glipizide (1.14 [0.93–1.38]), tolbutamide (1.04 [0.85–1.26]), and gliclazide (reference). Grouped according to pancreatic specificity, i.e., with tolbutamide, glipizide, and gliclazide as specific, and glibenclamide, and glimepiride as non-specific agents, no differential prognosis was found between the two groups (HR 1.04 [0.96–1.14], for non-specific, compared to pancreas specific agents). The prognosis was not dependent on prior acute myocardial infarction or ischemic heart disease (p for interactions >0.3).

Conclusions
In current clinical practice, it is unlikely that there are considerable differences in risk of mortality associated with individual sulfonylureas in patients with heart failure.