Umbilical Cord-Derived Cell Therapy for Autism Spectrum Disorder in Children: A Systematic Review of Interventional Trials

Introduction: Umbilical cord-derived cell therapy is offered to children with autism spectrum disorder (ASD), and narrative reviews of the field report encouraging results. Whether the controlled evidence supports that impression has not been appraised systematically. Methods: This review was conducted and reported in accordance with PRISMA 2020. PubMed and ClinicalTrials.gov were searched on 3 September 2026. Interventional studies administering umbilical cord blood- or cord tissue-derived cells to children with ASD and reporting clinical outcomes or safety were eligible. Companion mechanistic reports were linked to their parent trial, so the unit of analysis was the trial. One reviewer screened all records with language-model-assisted triage; a second independent screening pass and a rule-based recall filter re-screened every exclusion, and flagged records were adjudicated in full text. Randomized trials were appraised with RoB 2 and non-randomized trials with ROBINS-I. Outcome instruments differed, so results were synthesized narratively without pooling. Results: Of 319 reports and 26 registry entries, nine trials were included, comprising 315 children given cord-derived cells; screening agreement was 97.5% and re-screening did not change the included set. Three otherwise eligible trial reports were excluded as retracted. Three included trials were randomized and placebo-controlled (245 children randomized). The largest (n = 180) found no benefit on its primary outcome, adaptive socialization, with signals confined to a post hoc subgroup without intellectual disability; a crossover trial (n = 29) found no significant difference on any endpoint; the smallest (n = 36 randomized, 34 completing) reduced Social Responsiveness Scale-2 total scores by 6.03 points (95% CI -11.93 to -0.14), an imprecise estimate. All six non-randomized or uncontrolled trials reported improvement. No trial was at low overall risk of bias. Serious treatment-related events were absent, but five of twelve recipients of cord tissue mesenchymal cells developed clinically silent class I anti-HLA antibodies, and one autologous cord blood unit yielded coagulase-negative staphylococcus on post-thaw culture. Conclusions: Improvement is reported consistently only where a concurrent control is absent. Short-term safety appears acceptable; efficacy remains unsupported. This review was not registered, no protocol was prepared, and it received no specific funding.