Abstract Background and aims Longitudinal data on functional recovery after stroke with upper-extremity impairments is limited. This study aimed to investigate functional recovery during the first year after stroke in individuals with early upper-extremity impairment and to identify predictors of global functional recovery. Methods Data were obtained from the Stroke Arm Longitudinal Study in Gothenburg. Participants with first-ever stroke and impaired upper-extremity motor function were assessed using the Stroke Impact Scale (SIS) at 4 weeks, 3 months, 6 months, and 12 months. Model-based clustering was applied to classify participants into global good versus poor functional recovery phenotypes. Predictors of global recovery were examined using a gradient boosting model. Results Ninety-five participants (median age 68 years; 59% men) were included. Significant improvement over 12 months was observed in SIS domains related to strength, activities of daily living, mobility, hand function, and participation, whereas memory, emotion, and communication remained stable at high levels. Two global recovery groups were identified: good (65%) and poor (35%) recovery. Participants with good recovery demonstrated consistently higher SIS domain scores and greater perceived recovery. The gradient boosting model showed good discrimination for global functional recovery (AUC = 0.93). Upper-extremity motor function at 4 weeks was the strongest predictor (relative influence = 66%), followed by age (15%) and admission NIHSS score (11%). Conclusions Functional recovery after stroke is heterogeneous among individuals with early upper-extremity impairment. Early motor function, age, and stroke severity are key predictors of global functional recovery in this group. Conflict of interest Adam Viktorisson: nothing to disclose. Hanna C Persson: nothing to disclose. Margit Alt Murphy: nothing to disclose.
Viktorisson et al. (Fri,) studied this question.