Is Glycosylated Hemoglobin a Reliable Marker of Diabetes Management in Advanced Heart Failure Patients on Durable Mechanical Circulatory Support?

Abstract

Background Diabetes mellitus (DM) is an independent predictor of mortality in patients with heart failure (HF). Glycosylated hemoglobin (HbA1c) testing is the standard of care for monitoring glycemic control in DM patients. We sought to investigate the reliability of HbA1c testing in Stage D HF patients who require left ventricular assist device (LVAD) support. Methods We retrospectively reviewed patients who underwent LVAD implantation at an academic medical center between 1/1/2012 and 12/31/2018. We reviewed pre-VAD clinical data to help identify patients who carried a pre-LVAD diagnosis of DM, based on the following American Diabetic Association (ADA) criteria: 1) pre-LVAD glycosylated hemoglobin (HbA1c) ≥6.5% or 2) pre-LVAD HbA1c Results Amongst 108 patients with LVAD implants, 52 (48%) carried a pre-LVAD diagnosis of DM based on the above outlined criteria. Baseline characteristics of DM vs non-DM patients were comparable for age (56 ±11.7 years), gender (male 70%), bridge-to-transplantation LVAD (54%), HeartWare LVAD (77%), and days on LVAD support (325 ±318.9 days), but different for ischemic HF etiology (39% vs 17%, p=0.016), BMI (31.5 ±5.9 vs 27.9 ±6.8 kg/m2, p=0.007), and length of stay following LVAD implant (39 ±29.7 vs 25 ±17.4 days, p=0.005). Thirty-eight out of 52 DM patients had more than 90 days of follow-up while on LVAD support, accounting for a cumulative set of 72 HbA1c/APG readings. Prior to LVAD implant, mean comparisons between HbA1c and eHbA1c were similar (7.2 ±1.4% vs 7.1 ±1.5%, p=0.517). During LVAD support, means of HbA1c and eHbA1c were different (6.4 ±1.6% vs 7.2 ±1.9%, p Conclusion Discrepancies noted between HbA1c and eHbA1c raises concern for the reliability of HbA1c testing for DM management in Stage D HF patients on LVAD support.