Abstract Number: Esoc2026a1481 Infection and Inflammation as Risk Factors for Post-Stroke Cognitive Impairment and Dementia: A Systematic Review and Meta-Analysis

Background and aims Post-stroke cognitive impairment and dementia (PSCID) are common, yet predictors beyond age, stroke severity, and pre-existing pathology remain poorly defined. Acute infection may increase long-term cognitive risk after stroke, but evidence has not been synthesized alongside inflammatory mechanisms. Methods We systematically reviewed observational studies and non-intervention arms of randomized trials reporting acute post-stroke infections and/or inflammatory biomarkers measured ≤ 7 days after stroke, with cognitive outcomes assessed ≥3 months. Six databases were searched from inception to November 2024. Random-effects meta-analyses were performed where feasible. Risk of bias was assessed using modified Joanna Briggs Institute (PROSPERO-CRD42024628145). Results Fifty studies were included (6 infection, 45 inflammatory biomarkers; 1 overlapping; 25 moderate/high risk of bias; n = 78,058 total). In pooled analysis of two large population-based cohorts (n > 60,000), hospital-treated infection was associated with increased dementia risk (HR = 1.83, 95% CI 1.58–2.12), with strongest associations for pneumonia and UTI and evidence of a dose–response relationship by infection frequency (p-trend<0.001). Four additional studies reported associations between infection and global or domain-specific cognitive impairment. Nineteen inflammatory biomarkers were meta-analyzed (33 studies). PSCID was associated with higher acute-phase CRP (SMD = 0.72), IL-6 (0.42), ESR (0.61), fibrinogen (0.50), and MMP-9 (SMD = 1.51; OR = 1.67), although between-study heterogeneity was substantial. One biomarker study stratified by infection status; none adjusted for infection. Conclusions Post-stroke infection is consistently associated with long-term PSCID and may be a modifiable risk factor. Acute systemic inflammation may contribute to PSCID, but studies examining the potential confounding effects of co-existent infection are needed to inform prevention and risk stratification after stroke. Conflict of interest EM is supported by an ARUK-RAD Fellowship (ARUK-RADF2023B-007); TSF is supported by a Sauder Family/Heart & Stroke Professorship of Stroke Research from the University of British Columbia; STP is supported by an NIHR Programme Grant (NIHR204290) and the NIHR Oxford Biomedical Research Centre.