Comparative Efficacy and Acceptability of Non-pharmacological Therapies and Novel Pharmacotherapies for Autism: A Systematic Review and Network Meta-Analysis.

To assess the efficacy and acceptability of non-pharmacological therapies, novel pharmaceuticals, and their combination for autism and explore the preferred treatment options for autism. A thorough search was conducted across PubMed, Web of Science, The Cochrane Library, Embase, PsycINFO, China National Knowledge Infrastructure, and Medline databases to collect all studies associated with randomized controlled trials on non-pharmacological or new pharmacological interventions for autism, up to October 2024. Two researchers will independently conduct literature screening, data extraction, and quality assessment. The primary outcome will be assessed by the severity of core autistic symptoms using validated scales and standardized mean differences (SMDs). In contrast, the secondary outcome will focus on the acceptability of the treatment. We performed a network meta-analysis within a frequentist framework, utilizing the network package in statistical software Stata 18. 106 RCTs were involved, including 5986 patients, and data from 21 interventions were analyzed. Compared with placebo, the following interventions improved the core symptoms of autism: exercise (SMD = -1.05, 95%CI: -1.55 to -0.55), acupuncture and moxibustion (SMD = -0.87, 95%CI: -1.13 to -0.60), tDCS (SMD = -0.86, 95%CI: -1.58 to -0.13) and bumetanide (SMD = -0.57, 95%CI: -1.12 to -0.01). The cannabinoids group has a lower acceptance rate than most of the other groups. Exercise, acupuncture, non-invasive brain stimulation, and bumetanide can improve the core symptoms of autism and can be used as an alternative or adjunct to autism treatment, but more clinical trials are still needed to support this outcome.