Authors’ Reply to the Letter by Shoar et al. on “Glycosylated Hemoglobin as a Surrogate for the Prevention of Cardiovascular Events in Cardiovascular Outcome Trials Comparing New Antidiabetic Drugs to Placebo”

Abstract

Dear Editor,According to the suggestion by Shoar et al.[1], we tested the robustness of the correlation between HbA1c reduction and MACE decrease, performing sensitivity analyses [2].The predictive value of HbA1c for coronary events has been consistently shown in the early period following an acute coronary event [3]. There is also some evidence that HbA1c predicts survival in diabetic patients with stable coronary artery disease [4]. In our meta-analysis, 2 trials included patients with a recent acute coronary syndrome and 12 trials included patients either at high risk or with definite coronary artery disease (stable or unstable). Excluding the 2 trials performed after an acute coronary syndrome did not modify the correlation (R = 0.88, p = 0.0004).In the 5 trials testing a DPP-4 inhibitor, the relative MACE reduction was at most 3%, despite a significant (modest) reduction in HbA1c. Excluding these 5 trials did not modify the correlation between HbA1c reduction and MACE decrease (R = 0.84, p = 0.004).Lastly, we excluded the SUSTAIN-6 trial showing the higher MACE decrease (26%). The correlation was not significantly modified (R = 0.88, p < 0.001).Thus, we are confident that HbA1c reduction explains, at least in part, the decrease in MACE observed with the new antidiabetic drugs.