Heterogeneous Recovery Trajectories and Prognostic Factors After Ischemic Stroke

Abstract

Background Conventional stroke prognostication focuses on fixed timepoints, typically at 3 months. This study aimed to model recovery trajectories across multiple time points and stroke severity levels, and to identify factors associated with differential recovery patterns. Methods We analyzed data from an 11-year prospective stroke registry at one Medical Center, including ischemic stroke patients with documented modified Rankin Scale scores at 1, 3, 6, and 12 months post-stroke. Patients were stratified into severity groups based on NIH Stroke Scale (NIHSS). Generalized estimating equations were used to model functional trajectories and evaluate the impact of baseline clinical characteristics on recovery over time. Results A total of 6,965 patients were included: 3,421 (49.1%) with mild stroke (NIHSS <5), 2,335 (33.5%) with moderate stroke (NIHSS 5-15), and 1,209 (17.4%) with severe stroke (NIHSS >15). Distinct recovery trajectories were identified across severity groups. The majority of functional improvement occurred within the first 3 months, followed by a slower phase of stabilization or plateau across different stroke severity groups. In the mild stroke group, the respective increases of the proportion of favorable outcome were 15.1%, 5.6%, and 0.3%; in the moderate stroke group, 55.1%, 16.3%, and 6.2%; and in the severe stroke group, 78.8%, 27.1%, and 10.7%. Factors including age, sex, treatment with intravenous thrombolysis and/or endovascular thrombectomy, diabetes mellitus, end-stage renal disease, anemia, leukocytosis, prior cerebrovascular events, and white matter hyperintensities, significantly influenced recovery patterns, with varying significances across different severity strata. Notably, a subset of patients exhibited secondary functional decline after initial recovery, underscoring the dynamic and heterogeneous nature of post-stroke functional outcomes. Conclusions Stroke recovery is dynamic and heterogeneous. Patients with different baseline profiles follow distinct trajectories. This trajectory-based approach enhances prognostic accuracy, supports tailored patient counseling, and informs mechanisms of long-term recovery.