Background: Treatment options for acute ischemic stroke (AIS) remain limited. We aim to investigate the efficacy of early extracorporeal counterpulsation (ECP) in improving the prognosis of patients with acute cerebral infarction and to establish a short-term prognostic prediction model. Methods: This randomized controlled trial enrolled patients with AIS in a 1:1 ratio to either the ECP plus standard treatment group or the standard treatment-alone group. Outcomes were assessed by single-blinded investigators. The primary outcome was short-term poor prognosis (modified Rankin Scale, mRS score ≥3), while the secondary outcome was short-term poor neurological outcome (National Institutes of Health Stroke Scale score ≥5). Logistic regression and mediation analysis assessed ECP efficacy and potential mechanisms, while interaction tests evaluated outcome stability across subgroups. A machine learning algorithm was employed to establish a short-term prognosis prediction model (90-day mRS), with Shapley value analysis quantifying the contribution of each factor. Additionally, the E -value sensitivity analysis assessed robustness to unmeasured confounders. Results: ECP therapy showed significant negative correlations with both short-term adverse outcomes and adverse neurological outcomes, maintaining stability across all subgroups. Additionally, the mediator analysis indicated a d -dimer-mediated proportion of 6.6%. The optimal model for predicting 90-day adverse outcomes was Light Gradient Boosting Machine (area under the curve = 0.883), with stable results confirmed by fivefold cross-validation. Conclusions: The results of this study indicate that early ECP therapy exerts a positive effect on improving the prognosis of patients with acute cerebral infarction, and this effect may be partially mediated by d -dimer.
Tang et al. (2026) studied this question.