Adjunctive traditional Chinese medicine for ischemic stroke: a systematic review with structured narrative synthesis...

Ischemic stroke is a leading cause of death and long-term disability, and traditional Chinese medicine (TCM) is widely used alongside conventional care in East Asia. We reviewed randomized controlled trials (RCTs) of adjunctive TCM in ischemic stroke, analyzing the evidence by clinical phase and by individual formulation rather than as a single treatment class. We searched seven bibliographic databases and one clinical trial registry to 8 April 2026, for trials of adults with imaging-confirmed ischemic stroke given a TCM formulation in addition to standard care, compared with placebo plus standard care or standard care alone. Nine RCTs enrolling 12,348 randomized participants were eligible. They evaluated eight formulations across three clinical phases, with different outcome instruments and follow-up of 14 days to 24 months, so we did not pool them; each trial is reported with its published effect measure and synthesized under Synthesis Without Meta-analysis guidance. In acute treatment, Tongxinluo improved a favorable functional outcome at 90 days (modified Rankin Scale [mRS] 0 to 1: 640/973 vs. 575/973; odds ratio [OR] 1.33, 95% confidence interval [CI] 1.11 to 1.60), and Panax notoginseng saponins improved functional independence at 3 months (mRS 0 to 2: 1,328/1,487 vs. 1,218/1,479; OR 1.95, 95% CI 1.56 to 2.44); these thresholds are related but not interchangeable. Angong Niuhuang Pills and NeuroAiD (MLC601) showed no clear difference, and a small trial of Xueshuan Xinmai tablets reported no between-group estimate. In early recovery, Qizhitongluo improved lower-limb motor function at 12 weeks (mean difference 1.81 points, 95% CI 0.88 to 2.74). In secondary prevention, Naoxintong reduced 2-year recurrence [hazard ratio (HR) 0.665, 95% CI 0.492 to 0.899] and Dengzhan Shengmai reduced 1-year stroke incidence (HR 0.70, 95% CI 0.50 to 0.98). GRADE certainty, rated separately by formulation, clinical phase, and outcome, ranged from moderate to very low. Safety reporting was incomplete, so comparative safety cannot be established. The review was not prospectively registered and no dated protocol preceded screening. The findings are formulation-specific, unreplicated, and do not support routine clinical use or any conclusion about TCM as a treatment class.