Acupuncture As Adjunctive Therapy for Autism Spectrum Disorder in Children: A Systematic Review-Meta-Analysis and Critical Re-Evaluation of the Field Publisher



Fallahpour B ; Golshan Tafti M ; Bahrami R ; Aarafi H ; Shahbazi A ; Shiri A ; Masoudi A ; Bahrami M ; Golshan Tafti A ; Nematzadeh F ; Neamatzadeh H
Source: Middle East Current Psychiatry Published:2026

Abstract

Background: Autism spectrum disorder (ASD) is a complex neurodevelopmental condition affecting approximately 1 in 100 individuals globally, characterized by persistent deficits in social communication and interaction, and restricted, repetitive behavioral patterns. No pharmacotherapy has demonstrated consistent efficacy for core ASD symptoms. Acupuncture, a central modality of traditional Chinese medicine (TCM), is increasingly used as an adjunctive intervention for pediatric ASD; however, prior meta-analyses have been substantially limited by methodological shortcomings including violation of effect-size independence assumptions, absence of prediction intervals, inadequate publication bias assessment, exclusive reliance on Chinese-language databases, and failure to apply contemporary Grading of Recommendations Assessment, Development and Evaluation (GRADE) standards. Objective: To critically re-evaluate and comprehensively update the evidence base for acupuncture as adjunctive therapy for pediatric ASD through the largest synthesis to date, with robust variance estimation (RVE), prediction intervals, GRADE certainty assessment, and systematic moderator analyses. Methods: Thirteen electronic databases (Cochrane Library, MEDLINE/PubMed, Embase, Allied and Complementary Medicine Database [AMED], Cumulative Index to Nursing and Allied Health Literature [CINAHL], PsycARTICLES, China National Knowledge Infrastructure [CNKI], Wanfang Data, VIP Database, Chinese Biomedical Literature Database [CBM], JAIRO, CiNii, and KoreaMed) were searched from inception through May 20, 2026, without language restrictions. Additional records were obtained through manual reference searching. The primary analysis used RVE with small-sample correction and within-study correlation ρ = 0.7 to properly handle multiple dependent effect sizes per study. Results: Thirty-six unique RCTs contributing 56 effect sizes and involving 3,139 unique participants (5,189 total observations across both outcomes) met inclusion criteria. The pooled RVE estimate demonstrated a large standardized treatment effect: Hedges’ g = − 0.961 (95% confidence interval [CI]: − 1.120 to − 0.801; t = − 12.223, df = 35, P = 3.46 × 10⁻14), with a 95% prediction interval (PI) of − 2.329 to 0.408; I2 = 84.0%; τ2 = 0.381. Because the 95% PI includes zero and I2 exceeds 75%, the pooled point estimate should NOT be used for individual patient decision-making without explicit acknowledgment of this uncertainty. Five moderators were statistically significant after Bonferroni correction: outcome measure (Q = 40.458, P < 0.001), treatment duration (Q = 21.915, P < 0.001), intervention type (Q = 19.302, P < 0.001), publication period (Q = 85.879, P < 0.001), and diagnostic criteria (Q = 38.556, P < 0.001). GRADE assessment rated the primary outcome evidence as MODERATE quality (⊕ ⊕ ⊕ ○○). The adverse event rate was approximately 0.8%; no serious adverse events were recorded across 3,139 participants. Conclusions: Moderate-quality evidence provides a conditional suggestion that acupuncture may reduce autism symptom severity as an adjunct to multimodal treatment in children. Extreme between-study heterogeneity, universal performance bias due to unblinded designs, exclusive reliance on Chinese RCTs, complete absence of sham controls, and substantive publication bias asymmetry significantly limit confidence and prevent definitive clinical recommendations. International, multicenter, sham-controlled RCTs with standardized protocols, biomarker endpoints, and long-term follow-up are urgently required to advance the field. © The Author(s) 2026.