Pharmacological and dietary-supplement treatments for autism spectrum disorder: a systematic review and network meta-analysis

Randomized controlled trials (RCTs) of pharmacological and dietary-supplement interventions for ASD with minimum seven days of treatment. Participants diagnosed with ASD using standardized criteria; separate analyses for children/adolescents and adults. Both blinded and open RCTs eligible; placebo-controlled and head-to-head comparisons; first phase of crossover studies used. 125 RCTs in children/adolescents (n=7450) and 18 RCTs in adults (n=1104). Majority were double-blind and placebo-controlled with parallel design; median ages ~8.2 years (children/adolescents) and ~24.6 years (adults); male-to-female ratio ~5:1; most trials used DSM criteria or validated tools (e.g., ADI-R). Risk of bias assessed with Cochrane tool; data pooled with network meta-analysis between age groups.

Some medications could improve core symptom domains in ASD. In children/adolescents, aripiprazole, bumetanide, atomoxetine, and risperidone improved at least one core symptom domain (social-communication difficulties, repetitive behaviors, or overall core symptoms). In adults, fluoxetine, fluvoxamine, oxytocin, and risperidone improved repetitive behaviors. Among dietary supplements, omega-3 fatty acids showed potential small benefits for social-communication difficulties in children/adolescents (low-quality evidence); melatonin reduced caregiver stress but did not improve core symptoms. Overall, most evidence pertains to associated symptoms (irritability, ADHD, sleep) with modest or inconsistent effects on core symptoms, and no ASD core-symptom medication is approved. Benefits should be weighed against adverse effects (e.g., sedation, weight gain, extrapyramidal symptoms for antipsychotics). Evidence quality is generally low to very low with small, short-duration trials and limited adult data; results support cautious use for associated symptoms and highlight need for robust, well-powered trials and targeted treatments.

Small, short-duration trials with heterogeneous measures and populations; risk of bias and reporting bias; limited data in adults; transitivity and network coherence concerns; most evidence relates to associated symptoms rather than core ASD features; safety data incomplete for several interventions.