Abstract
Background. A randomized clinical has shown the effectiveness of intramuscular electrical stimulation for the treatment of poststroke shoulder pain. Objective. Identify predictors of treatment success and assess the impact of the strongest predictor on outcomes. Method. This is a secondary analysis of a multisite randomized clinical trial of intramuscular electrical stimulation for poststroke shoulder pain. The study included 61 chronic stroke survivors with shoulder pain randomized to a 6-week course of intramuscular electrical stimulation (n = 32) versus a hemisling (n = 29). The primary outcome measure was Brief Pain Inventory Question 12. Treatment success was defined as ≥ 2-point reduction in this measure at end of treatment and at 3, 6, and 12 months posttreatment. Forward stepwise regression was used to identify factors predictive of treatment success among participants assigned to the electrical stimulation group. The factor most predictive of treatment success was used as an explanatory variable, and the clinical trials data were reanalyzed. Results. Time from stroke onset was most predictive of treatment success. Subjects were divided according to the median value of stroke onset: early (<77 weeks) versus late (> 77 weeks). Electrical stimulation was effective in reducing poststroke shoulder pain for the early group (94% vs 7%, P < .001) but not for the late group (31% vs 33%). Repeated-measure analysis of variance revealed significant treatment (P < .001), time from stroke onset (P = .032), and treatment by time from stroke onset interaction (P < .001) effects. Conclusions. Stroke survivors who are treated early after stroke onset may experience greater benefit from intramuscular electrical stimulation for poststroke shoulder pain. However, the relative importance of time from stroke onset versus duration of pain is not known.