Frontiers in Pharmacology · Nov 9, 2026
Exercise and systemic inflammatory, immune and haemostatic biomarkers across recovery phases in cerebrovascular disease: a systematic review and narrative synthesis
By Edward Jackson, Mudasar Aziz, Anas Al Dabbagh et al.
Background: Stroke provokes a systemic inflammatory response that is coupled to haemostasis and predicts poor outcome. Whether exercise modifies inflammatory, immune and haemostatic biomarkers across the cerebrovascular disease spectrum, and whether any effect depends on recovery phase, training modality or dose, is uncertain. Methods: We systematically reviewed interventional studies of active exercise in adults with stroke at any phase of recovery, or with cerebral small vessel disease, that reported at least one systemic inflammatory or immune outcome (primary outcomes: interleukin-6, tumour necrosis factor-alpha, C-reactive protein and fibrinogen). MEDLINE, PubMed, Embase and the Cochrane Central Register of Controlled Trials were searched from inception to 18th October 2026. Risk of bias was assessed with the Cochrane RoB 2 and ROBINS-I tools, certainty of evidence assessed using GRADE, and results were synthesised narratively in accordance with Synthesis Without Meta-analysis reporting guidance. Results: Twelve studies (nine randomised; 756 participants from 10 countries) were included, and biomarker responses were phase-dependent. In the acute phase, very early exercise limited the day-7 neutrophil rise and preserved lymphocyte counts. In the subacute phase, inflammatory markers declined irrespective of treatment allocation and aerobic training did not modify this trajectory, although early interleukin-6 and fibrinogen were associated with worse six-month outcome. In the chronic phase, aerobic training, or aerobic or resistance training combined, appeared to reduce interleukin-6, myeloperoxidase, oxidative-stress markers and von Willebrand factor and increased ADAMTS13 activity.. Delivered exercise intensity frequently fell below that prescribed. No study reported infection outcomes or immune phenotyping beyond differential leucocyte counts. Conclusion: Biomarker responses to exercise after stroke may depend on recovery phase, modality and dose rather than on exercise participation alone. Small samples, intervention and assay heterogeneity, uncontrolled confounding and incomplete outcome data limit certainty. Adequately powered trials using a core biomarker set, verified exercise dose and functional co-primary endpoints are needed, as are trials that delineate cerebrovascular disease subtypes. Systematic review registration: PROSPERO, identifier CRD420261492285.