Abstract DP216: Glucose-to-Lymphocyte Ratio as a Predictor of In-Hospital Mortality in Spontaneous Intracerebral Hemorrhage

Introduction: Intracerebral hemorrhage (ICH) accounts for approximately 10% of strokes in the United States and is the most fatal subtype, with early mortality rates of 30–40%. The glucose-to-lymphocyte ratio (GLR) has emerged as a novel prognostic marker reflecting systemic metabolic and inflammatory responses in critical illness. Patients with ICH often present with systemic inflammation, immune cell dysfunction, and stress-induced hyperglycemia, all linked to poor outcomes. We aimed to evaluate the association between GLR, in-hospital mortality, and hematoma expansion (HE) in patients with spontaneous ICH. Hypothesis: Higher GLR levels are associated with increased in-hospital mortality and HE in patients with spontaneous ICH. Methods: We conducted a retrospective cohort study of patients admitted to Beth Israel Deaconess Medical Center with spontaneous ICH from 2008 to 2024. GLR was analyzed as a continuous variable and categorized in quartiles. We examined associations between GLR and (a) in-hospital mortality and (b) hematoma expansion dichotomized as significant (>33% relative increase) vs not. We also fitted univariate and multivariable logistic regression models, adjusting for confounders (age, baseline hematoma volume, intraventricular extension, Glasgow Coma Scale, infratentorial location and HE), and using the first GLR quartile as reference. Results: We included 1273 patients, 584 (46%) were men and mean age was 70.9 (SD 14.2). 201 (15.6%) hemorrhages were infratentorial and 501 (39%) had intraventricular extension. Median GLR was 105.2 (IQR 70.3-175.4), median hematoma volume 14.7 mL (IQR 4.7–42.9), and HE occurred in 20%. In-hospital mortality was 23.4%. In unadjusted analyses, GLR was significantly associated with in-hospital mortality (OR 5.18, 95% CI (3.39-7.94), p<0.0001) with a dose-response association (p-trend <0.0001). Adjusted analyses attenuated the strength of the association which remained statistically significant (OR 1.82, 95% CI (1.06-3.14), p=0.03) consistent with a dose-response relationship (p-trend 0.006, figure 1). We observed no association between GLR and HE (p=0.559). Conclusions: Our findings suggest that GLR may serve as an independent predictor of mortality in patients with spontaneous ICH, potentially reflecting the combined impact of inflammation and metabolic stress.