The additive effect of hemoglobin glycation index and glycemic variability to predict mortality in cardiac intensive care patients with and without diabetes

The hemoglobin glycation index (HGI) and glycemic variability (GV) are important metabolic markers associated with cardiovascular diseases and their prognoses. However, the potential additive effect of these biomarkers on mortality risk remains unclear. This study aimed to investigate the independent and combined effects of HGI and GV on mortality in cardiac ICU patients, both with and without diabetes. We conducted a retrospective cohort analysis of 1636 individuals admitted to the cardiac ICU. The primary outcome was 28-day mortality, with 365-day mortality as a secondary endpoint. Kaplan-Meier survival curves, Cox proportional hazards models, restricted cubic splines, receiver operating characteristic (ROC) curve, and subgroup analyses were used to evaluate the relationship between the HGI and adverse outcomes. During the follow-up period, 163 patients (10.0%) died within 28 days, and 322 patients (19.7%) died within 365 days. ROC analysis showed that the combination of HGI and GV had better predictive accuracy for mortality compared to HGI alone (28-day mortality: 0.658 vs. 0.621, P = 0.025; 365-day mortality: 0.644 vs. 0.562, P < 0.001) and GV alone (28-day mortality: 0.658 vs. 0.622, P = 0.036; 365-day mortality: 0.644 vs. 0.618, P = 0.031). Among non-diabetic patients, low HGI combined with high GV was linked to the highest mortality risk at both 28 days (HR = 4.223, 95% CI: 2.336-7.634) and 365 days (HR = 2.504, 95% CI: 1.550-4.045). However, diabetic patients with low HGI and low GV had the greatest risk of death at 28 days (HR = 3.729, 95% CI: 1.747-7.959) and 365 days (HR = 2.287, 95% CI: 1.324-3.951). The results suggested a potential additive effect of HGI and GV on mortality risk at both 28-day and 365-day in cardiac ICU patients, regardless of diabetes status. This combined biomarker approach may guide personalized glycemic management and improve clinical outcomes.