Abstract / Summary
Abstract Blood biomarkers after ischemic stroke are commonly evaluated at a single sampling time and against a single outcome, although their concentrations and clinical associations may change during the acute-to-subacute phase. We prospectively studied 218 patients with acute ischemic stroke and measured plasma glial fibrillary acidic protein (GFAP), neurofilament light chain, interleukin-6, ubiquitin C-terminal hydrolase L1, and interferon-α (IFN-α) at admission, 24 hours, and day 7. At 3 months, 116 patients (53.2%) met education-adjusted criteria for post-stroke cognitive impairment (PSCI), and 51 (23.4%) had poor functional outcome (modified Rankin Scale [mRS] 3–6). GFAP showed the most consistent association with PSCI across sampling times (AUC 0.683, 0.680, and 0.683; all FDR q < 0.001). Admission GFAP remained associated with PSCI after clinical adjustment (OR 1.87 per ln-unit increase, 95% CI 1.27–2.75) and with lower 3-month Montreal Cognitive Assessment (MoCA) scores after adjustment for baseline cognition, but was not associated with poor functional outcome (adjusted OR 0.86, 95% CI 0.58–1.28). In contrast, day-7 IFN-α was independently associated with poor functional outcome (adjusted OR 1.97, 95% CI 1.43–2.70), and this association remained essentially unchanged after accounting for pulmonary infection. Paired within-patient comparisons showed greater discrimination of admission GFAP for cognitive than functional outcome, whereas delayed IFN-α showed greater discrimination for functional outcome. These findings indicate that the prognostic information carried by circulating biomarkers after stroke depends on both sampling time and the clinical outcome assessed, supporting the use of serial sampling and outcome-specific analyses in future biomarker studies of post-stroke recovery.