Response to letter by Aphale et al.: Long-term outcomes of coronary angiography in patients with and without diabetes
We thank Dr. Aphale and colleagues for their interest in our study investigating the prognostic impact of type 2 diabetes mellitus (T2DM) in a large cohort of 7150 patients undergoing coronary angiography (CA) for different indications [1]. Whereas the present study identified the presence of T2DM as an independent predictor of hospitalization for heart failure (HF) at 36 months (29.0% vs. 18.2%; p = 0.001), even after multivariable adjustments [2], Aphale et al. suggest glycemic variability may serve as even stronger predictor for adverse clinical outcomes than the presence of T2DM itself. From this perspective, glycemic variability – quantified by the standard deviation (SD) – has recently been shown to correlate with myocardial injury in T2DM patients, measured by high-sensitivity troponin levels [3], as well as to predict the risk of mortality in patients with diabetes [4]. Since for the present study the documentation of blood glucose levels comprised minimum, maximum, as well as median glucose levels, we performed additional sub-analyses investigating the prognostic impact of blood glucose ranges during index hospitalization, which may serve as a surrogate of glycemic variability, regarding to the risk of HF-related hospitalization in patients with and without T2DM. Accordingly, serial blood glucose measurements were performed in 768/2233 patients with T2DM with a median range of 50 mg/dL (interquartile range (IQR) 22–102 mg/dL) and in 1584/4917 patients without (median range 46 mg/dL (IQR 21–93 mg/dL). When analyzed by quartiles, the risk of HF-related hospitalization was increased in T2DM patients with higher blood glucose ranges (Q4: 31.2% vs. Q1: 22.9%; hazard ratio (HR) = 1.484; 95% confidence interval (CI) 1.003–2.195; p = 0.048), whereas the risk of HF-related hospitalization was similar among patients in Q2, Q3 and Q4 (Fig. 1, left panel). This association was still observed after multivariable adjustment for the same variables as in the main publication [2] (Q4 vs Q1: adjusted HR = 1.904; 95% CI 1.239–2.925; p = 0.003). By contrast, the ranges of blood glucose during index hospitalization did not predict the risk of HF-related hospitalization in patients without T2DM when analyzed by quartiles (p = 0.747) (Fig. 1, right panel). The observations are in line with previous data suggesting an independent association of glycemic variability with impaired long-term prognosis in T2DM patients and underscore the importance of even mild variations (i.e., similar impaired prognosis in Q2-Q4) with regard to the risk of HF-related hospitalization.