Exploratory analysis shows GFAP identifies intracerebral hemorrhage, while D-dimers have limited impact on LVO detection.
Background and aims Blood biomarkers may support acute stroke triage. Glial fibrillary acidic protein (GFAP) differentiates intracerebral hemorrhage (ICH) from acute ischemic stroke (AIS), transient ischemic attacks (TIA) and stroke mimics (SM). D-dimers have been proposed to identify large-vessel-occlusion (LVO). We assessed whether a sequential strategy combining GFAP and D-dimers improves stroke subtype classification. Methods Exploratory analyses of a prospective cohort of suspected stroke patients presenting <24 hours from onset. Blood samples were collected upon emergency department arrival. Plasma GFAP levels (pg.mL−1) were measured using the i-STAT Alinity system (results <15 minutes), and D-dimers (ng.mL−1) using a routine laboratory assay (Innovance). Final diagnoses were classified as ICH, LVO-AIS, non-LVO-AIS, TIA, or SM. Results Among 315 participants (mean age 71.6±15.3 years; median NIHSS 8 [3-17]), final diagnoses included 186 AIS (81 LVO), 32 ICH, 88 SM, and 9 TIA (Fig. 1,2). Median GFAP levels were significantly higher in ICH vs. non-ICH (313 [75-4147] vs. 43 [29-80] pg.mL−1; p<0.001), with strong discriminatory performance (AUROC 0.87, 95% CI [0.80-0.93]). An optimal GFAP cut-off of 120 pg.mL−1 achieved 69% sensitivity (95%CI [53-84]) and 87% specificity (95%CI [83-91]). D-dimer levels were higher in LVO compared to other diagnoses (1531 [686-2999] vs 653 [368-1469] ng.mL−1; p<0.001) but showed limited discrimination (AUROC 0.66, 95%CI [0.59-0.73]). Restricting analyses to participants with GFAP <120 pg.mL−1 (n=256), D-dimers did not improve LVO detection (AUROC 0.67, 95% CI [0.59-0.75]). Conclusions GFAP reliably identifies ICH, whereas D-dimers provide limited incremental value for LVO detection. Combined biomarker strategies require further refinement in hyperacute stroke triage. Conflict of interest Julien F. Paul: nothing to disclose; Florence Barrette : nothing to disclose; Sarah-Emilie Godin : nothing to disclose; Ralph Nuhay : nothing to disclose ; David Chiasson-Ricard : nothing to disclose ; Catherine Brassard : nothing to disclose ; Laura C. Gioia : I declare having current funding from the Heart and Stroke foundation of Canada Figure 1 - belongs to Conclusions Figure 2 - belongs to Conclusions
No takes yet. Share an insight, caveat, or question.
Paul et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: