Glycated Albumin and Cardiovascular Mortality in CKD Stage V Patients with Diabetes Mellitus: A Five-Year Follow-Up Study

Glycemic assessment in patients with chronic kidney disease (CKD) Stage V on hemodialysis (HD) is limited by the inaccuracy of hemoglobin A1c (HbA1c), mainly due to anemia, shortened erythrocyte lifespan, and erythropoiesis-stimulating agent (ESA) therapy. Glycated albumin (GA), independent of erythrocyte turnover, may better reflect glycemic exposure. We evaluated the diagnostic performance and clinical utility of GA as a biomarker of poor glycemic control and cardiovascular risk in patients with diabetes mellitus (DM). A cross-sectional analysis and five-year prospective follow-up were conducted in three subgroups: HD+DM+ (n = 40), HD- DM+ (n = 15), and HD+DM– (n = 22). Glycemic markers (mean plasma glucose over 3 months (PG3m), GA, and HbA1c) were compared between groups. GA levels were significantly higher in HD+DM+ patients (p < 0.001) and showed the strongest correlation with PG3m. GA independently predicted poor glycemic control (OR 3.23; 95% CI 1.54–6.77; p = 0.002) and demonstrated high diagnostic accuracy (AUC 0.873; optimal cut-off 10%). During the five-year follow-up, CV mortality was 35%, and 85% of deceased patients had GA > 10%. Although Cox regression did not reach statistical significance, GA showed a consistent trend toward higher CV mortality risk after adjustment for age and HD duration (HR 2.56; 95% CI 0.57–11.44; p = 0.21), whereas HbA1c was not prognostic (HR 1.39; 95% CI 0.49–3.91; p = 0.28). GA appears to be a clinically useful marker of glycemic control in diabetic patients with CKD Stage V receiving maintenance hemodialysis. In this cohort, GA showed high diagnostic accuracy and a more balanced sensitivity/specificity profile compared with HbA1c, with a consistent trend toward association with cardiovascular mortality. In this regard, the wider use of this marker in clinical practice is worth considering. Nevertheless, larger prospective studies are warranted to validate these observations.