Differential Effect of Glycosylated Hemoglobin Value and Antidiabetic Treatment on the Risk of 30-day Readmission Following a Hospitalization for Acute Heart Failure.

INTRODUCTION AND OBJECTIVES In patients with heart failure and type 2 diabetes, low glycosylated hemoglobin has been related with higher risk of mortality but information regarding morbidity is scarce. We sought to evaluate the association between glycosylated hemoglobin and 30-day readmission in patients with type 2 diabetes and acute heart failure. METHODS Glycosylated hemoglobin was measured before discharge in 835 consecutive patients with acute heart failure and type 2 diabetes. Cox regression analysis adapted for competing events was used. RESULTS Mean (standard deviation) age was 72.9 (9.6) years and median glycosylated hemoglobin was 7.2% (6.5%-8.0%). Patients treated with insulin or insulin/sulfonylurea/meglitinides were 41.1% and 63.2% of the cohort, respectively. At 30 days post-discharge, 109 (13.1%) patients were readmitted. A multivariate analysis revealed that the effect of glycosylated hemoglobin on the risk of 30-day readmission was differentially affected by the type of treatment (P for interaction<.01). Glycosylated hemoglobin (per 1% decrease) was inversely associated with higher risk in those receiving insulin (hazard ratio = 1.45; 95% confidence interval, 1.13-1.86; P=.003) or insulin/sulfonylurea/meglitinides (hazard ratio = 1.44; 95% confidence interval, 1.16-1.80; P=.001). Conversely, glycosylated hemoglobin (per 1% increase) had no effect in non-insulin dependent diabetes (hazard ratio = 1.01; 95% confidence interval, 0.87-1.17; P=.897) or even a positive effect in patients not receiving insulin/sulfonylurea/meglitinides (hazard ratio = 1.12; 95% confidence interval, 1.03-1.22; P=.011). CONCLUSIONS In acute heart failure, glycosylated hemoglobin showed to be inversely associated to higher risk of 30-day readmission in insulin-dependent or those treated with insulin/sulfonylurea/meglitinides. A marginal effect was found in the rest. Whether this association reflects a treatment-related effect or a surrogate of more advanced disease should be clarified in further studies.