Dynamic trajectories of inflammatory biomarkers and post-stroke cognitive impairment: a comprehensive review of neuroimmune mechanisms, longitudinal modeling, and clinical translation

Abstract

Post-stroke cognitive impairment (PSCI) affects 30–50% of acute ischemic stroke survivors within the first year, yet the prognostic relevance of dynamic inflammatory trajectories remains incompletely defined. This narrative review synthesizes clinical and mechanistic evidence linking temporally evolving inflammatory biomarkers to vascular cognitive impairment after acute ischemic stroke. A structured evidence search of PubMed, MEDLINE, Embase, and Google Scholar from database inception to June 2026 prioritized longitudinal inflammatory-biomarker cohorts, trajectory-modeling analyses, machine-learning prediction studies, and mechanistic literature on neuroimmune injury. The available evidence indicates that post-stroke inflammation follows a multiphasic course, including hyperacute activation, acute peaking, subacute resolution, and, in a subset of patients, persistent low-grade inflammation. Delayed resolution or persistent elevation of interleukin-6, C-reactive protein, and the neutrophil-to-lymphocyte ratio appears to predict PSCI more accurately than single admission measurements. Mechanistically, sustained inflammatory signaling may contribute to chronic microglial activation, mCRP/VE-cadherin-mediated blood–brain barrier disruption, hippocampal and Papez-circuit vulnerability, and secondary neurodegeneration. Trajectory-based modeling, including latent class approaches and machine learning, may therefore improve PSCI risk stratification and support personalized follow-up and anti-inflammatory prevention trials. Larger prospective cohorts with standardized sampling and cognitive assessment are required to validate these translational applications.