L-shaped relationship between hemoglobin glycation index and short-term mortality in patients with intracerebral hemorrhage: A retrospective cohort study

Background Intracerebral hemorrhage (ICH) carries a substantial risk of early death and is frequently linked to unfavorable clinical outcomes, yet early prognostication remains challenging. The hemoglobin glycation index (HGI), which quantifies the discrepancy between measured glycosylated hemoglobin A1c(HbA1c) and the level expected based on fasting plasma glucose (FPG), has shown prognostic relevance in various clinical settings. Our objective was to investigate whether the HGI serves as a predictor of short-term mortality among individuals with ICH. Methods We performed a retrospective analysis utilizing data from the critical care database. We identified 1,318 adult ICH patients with available HbA1c and FPG data. HGI was defined as the difference between the observed HbA1c level and the HbA1c value predicted on the basis of admission FPG. Patients were stratified into HGI tertiles. The main outcome was 30-day all-cause mortality (90-day mortality served as a secondary endpoint). The relationship between HGI and mortality was examined using Kaplan–Meier curves, Cox proportional hazards models, and restricted cubic spline analyses. Subgroup analyses were also performed. Results Mortality was significantly greater among individuals in the lowest HGI tertile than among those in the highest tertile (30.18% vs 17.24% at 30-day mortality, p < 0.001; 33.64% vs 22.07% at 90-day mortality, p < 0.001). Multivariable Cox regression showed that higher HGI was an independent predictor of reduced mortality risk. The adjusted hazard ratio(HR) for 30-day mortality was 0.84 (95% CI 0.75–0.93), for 90-day mortality was 0.86(95% CI: 0.79–0.95) in fully adjusted models. Restricted cubic spline (RCS) analysis demonstrated an L-shaped association, with inflection points identified at HGI values of 0.692 for 30-day mortality and 0.472 for 90-day mortality, respectively. Below these thresholds, each one-unit increase in HGI corresponded to a 47.7% reduction in 30-day mortality risk and a 40.4% reduction in 90-day mortality risk, respectively. The association between HGI and mortality was consistent across most subgroups, with a significant interaction by diabetes status (p for interaction < 0.05), indicating the predictive value of HGI was more pronounced in patients with diabetes. Conclusion An L-shaped association exists between HGI and short-term mortality in ICH patients, with low HGI indicating substantially higher risk and holds potential as a novel prognostic indicator for facilitating early risk stratification in patients with acute ICH.