Answers are grounded in this paper and cite their sources.
Retrospective cohort study reveals a reverse J-shaped association between hemoglobin glycation index and mortality in critically ill atrial fibrillation patients, suggesting low values signal high...
The hemoglobin glycation index (HGI), a marker of individual glycation propensity, has demonstrated prognostic value in cardiometabolic diseases. Its association with outcomes in critically ill patients with atrial fibrillation (AF) remains poorly characterized, particularly over the longer term. In this retrospective cohort study, we analyzed data from 2458 ICU patients with AF in the MIMIC-IV database. HGI was calculated as the difference between measured glycated hemoglobin (HbA1c) and the value predicted from fasting plasma glucose. The primary endpoints were 30-day and 365-day all-cause mortality. We employed Kaplan–Meier analysis, Cox proportional hazards models, restricted mean survival time (RMST), and restricted cubic splines. Subgroup analyses were exploratory. Prespecified sensitivity analyses addressed bias via exclusion of continuous enteral/parenteral nutrition or dextrose fluids, restriction of HbA1c to within 7 days, exclusion of recent transfusion or extreme hemoglobin, and a lipid complete-case analysis. The analysis revealed a reverse J-shaped relationship between HGI and all-cause mortality. Patients in the lowest HGI quartile (Q1) faced the highest risk of death across all time points. RMST was significantly shorter in Q1 versus Q2. Adjusted spline curves confirmed this nonlinear association, with inflection points at HGI values of −0.54 (30-day) and −0.65 (365-day). In the fully adjusted model, the Q2 versus Q1 hazard ratio (HR) was 0.704 (95% CI: 0.530–0.933) for 30-day and 0.714 (0.552–0.924) for 365-day mortality. Directionally consistent estimates were observed across all sensitivity analyses, for example, excluding enteral/parenteral nutrition [0.641 (0.464–0.885)], restricting HbA1c to 7 days [0.611 (0.408–0.913)], excluding transfusion/extreme hemoglobin [0.713 (0.527–0.964)], and lipid complete-case [0.647 (0.455–0.920)]. Subgroup analyses did not identify significant effect modification after FDR correction; however, these analyses were underpowered. In this MIMIC-IV cohort of critically ill AF patients, HGI exhibits a reverse J-shaped association with mortality, indicating that low HGI, and to a lesser extent high HGI, may identify higher-risk phenotypes. These findings require external validation and should not be interpreted causally. Integrating HGI into initial metabolic profiling could enhance risk stratification and inform personalized monitoring strategies in the ICU, pending prospective confirmation.
No takes yet. Share an insight, caveat, or question.
Chen et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: