Abstract WP224: Hematoma location and symptom onset to sampling time are predictors of GFAP release into plasma in acute intracerebral hemorrhage

Background: The astroglial protein GFAP (glial fibrillary acidic protein) is a promising biomarker candidate for prehospital differentiation between intracerebral hemorrhage (ICH) and ischemic stroke. GFAP is rapidly released into plasma from disrupted astrocytes in expanding ICH, but increases more slowly following ischemic cell necrosis. Besides hematoma volume, factors associated with GFAP release in acute ICH have not been explored. It has been noted in previous studies that a certain proportion of ICH patients did not show elevated GFAP levels, leading to a reduced sensitivity of the test. The aim of this study was to analyze GFAP release kinetics according to ICH location and the time span between symptom onset and blood sampling. Methods: Consecutive ICH patients admitted within 6 hours of symptom onset were derived from a prospective biomarker study in acute stroke. All blood samples were collected in the prehospital phase. Plasma GFAP measurements were performed on the i-STAT Alinity® (Abbott) point of care device. We determined both ICH volume and hematoma location from the first available brain scan after admission in a standardized way. Results: We analyzed 74 ICH patients (mean age 73.5 ± 12.4 years). The median [IQR] time span from symptom onset to blood sampling was 144.0 minutes [113.3] and the mean ICH volume was 35.7 ± 38.8 mL. A strong correlation between hematoma volume and GFAP plasma levels was observed (Spearman rho 0.697; p<0.001). After adjustment for hematoma volume, ICH located in the striatum showed lowest GFAP release rates (44.2 ± 99.8 ng/mL per mL hematoma volume), whereas for cerebellar and brainstem this value was significantly higher (262.5 ± 177.0 and 192.5 ± 212.1). Our findings well match the histopathological distribution of GFAP in the brain. In order to validate this finding, we analyzed datasets of two previous GFAP studies which showed an identical pattern of GFAP release depending on location. We found a positive correlation between time span from symptom onset to blood sampling and GFAP release rate (Spearman rho 0.378; p<0.001). Conclusion: Besides hematoma volume, ICH location and time between symptom onset and blood sampling are predictors of GFAP release in acute ICH. Thus, particularly small striatal bleedings in an early time window may be difficult to detect. This shortcoming has to be considered when biomarker-based identification strategies for acute ICH patients are applied in the prehospital phase.