Abstract Background and aims Post-stroke cognitive impairment (PSCI) encompasses cognitive deficits arising after stroke and affects up to 60% of survivors within the first year. Despite its prevalence, early PSCI identification remains difficult, partly due to the lack of reliable, accessible biomarkers. Identification of biomarkers predictive of PSCI has potential to enhance risk-stratification and guide targeted interventions, ultimately improving patient outcomes. This systematic review aims to evaluate current evidence on blood-based biomarkers associated with PSCI. Methods A systematic search of PubMed, Embase, CINAHL, Scopus, and Web of Science was conducted to identify and evaluate the predictive utility of blood-based biomarkers for PSCI in patients suffering acute stroke within 24 hours of symptoms onset. Two reviewers independently screened and extracted data from studies examining associations between blood-based biomarkers and PSCI. Data on patient cohorts, time from symptoms onset, biomarker type, methodologies and quantification strategies, accuracy, sensitivity, and specificity were compared. Results Seventy-six studies met the inclusion criteria. Elevated inflammatory markers (e.g. C-reactive protein, interleukin-6), neurofilament light chain, and homocysteine were most consistently associated with poorer post-stroke cognitive outcomes. Biomarkers of amyloid and tau pathology were less frequently investigated but showed potential associations. Considerable variability was observed in biomarker sampling time points and cognitive assessment tools. Overall study quality was limited, largely due to small sample sizes and insufficient control for confounding factors. Conclusions Blood-based biomarkers show promise for identifying and prognosticating PSCI, but current evidence is limited by methodological heterogeneity. Large, longitudinal studies with standardized cognitive assessments and repeated biomarker measurements are needed to support clinical translation. Conflict of interest Conor Healy: nothing to disclose; Clodagh Flanagan: nothing to disclose; Anne Hickey: nothing to disclose; David Williams: nothing to disclose; Shona Pfeiffer: nothing to disclose
Healy et al. (2026) studied this question.