Stress hyperglycemia indexes and early neurological deterioration in spontaneous intracerebral hemorrhage

TL;DR

This study finds that new stress hyperglycemia indexes can predict early neurological deterioration in patients with spontaneous brain hemorrhage.

Contribution

The study introduces and validates new stress hyperglycemia indexes as predictors of neurological decline in intracerebral hemorrhage patients.

Findings

Early neurological deterioration occurred in 12.3% of intracerebral hemorrhage patients.

Glycemic gap, stress hyperglycemia ratio, and glucose-glycated hemoglobin ratio predicted deterioration better than admission glycemia.

Optimized cutoffs for these indexes showed moderate sensitivity and specificity for detecting deterioration.

Abstract

To evaluate the relationship of early neurological deterioration (END) with admission glycemia (aG) and new stress hyperglycemia indexes in spontaneous intracerebral hemorrhage (ICH) patients.

The present retrospective study included 171 ICH patients from two stroke centers. END was defined as an increase ≥ 4 points in National Institutes of Health Stroke Scale and/or a decrease ≥ 2 points in Glasgow Coma Scale within 72 hours from admission. The included stress hyperglycemia indexes were glycemic gap (GGAP), stress hyperglycemia ratio (SHR), and glucose-glycated hemoglobin ratio. GGAP was calculated as aG – 28,7*glycated hemoglobin + 46,7; SHR as aG / (28,7*glycated hemoglobin – 46,7); Glucose-glycated hemoglobin ratio as aG / glycated hemoglobin. We performed univariate and multivariate analyses for END. The receiver operating characteristic curves were built for END-related glycemic measures; area under curves (AUC) were calculated and compared. The optimized threshold values were calculated, and significant glycemic measures were dichotomized. Univariate and multivariate analyses were performed for the dichotomized measures.

END was present in 21 patients (12.3%) and was significantly associated with GGAP, SHR and glucose-glycated hemoglobin ratio, but not with aG. The AUC of the three stress hyperglycemia indexes did not differ significantly. The optimized cutoffs were 35.68 (sensitivity 0.47, specificity 0.81), 1.15 (sensitivity 0.62, specificity 0.68), and 26.67(sensitivity 0.43, specificity 0.80) for GGAP, SHR, and glucose-glycated hemoglobin ratio respectively. END was also associated with all stress hyperglycemia indexes expressed as categorical variables.

GGAP, SHR, and glucose-glycated hemoglobin ratio were predictors of END in ICH patients.

The online version contains supplementary material available at 10.1007/s10072-025-08097-8.

Figures