Efficacy and safety of pharmacologic therapies for irritability in pediatric autism spectrum disorder: A systematic review and network meta-analysis of randomized controlled trials

INTRODUCTION: Irritability is a common and clinically significant feature of pediatric autism spectrum disorder (ASD). We conducted a systematic review and network meta-analysis (NMA) to compare and rank pharmacological interventions for irritability in this population. METHODS: Multiple databases were searched up to September 1, 2025, for randomized controlled trials enrolling pediatric ASD with significant irritability. Risk of bias was assessed using Cochrane RoB 2.0, and confidence in network estimates was evaluated with Confidence in Network Meta-Analysis. Pairwise meta-analyses, subgroup analyses, and frequentist NMA were performed for efficacy and safety outcomes. RESULTS: Forty-three RCTs involving 2711 participants, evaluating 35 drugs across 41 regimens, were included. Traditional pairwise meta-analysis suggested that pharmacotherapy improved irritability more than placebo, with greater effects in younger children (<12 years) and shorter trials (⩽8 weeks). In the NMA, risperidone (Ris; MD = 8.81, 95% CI: 6.01-11.61) and aripiprazole (MD = 5.53, 95% CI: 2.61-8.45) led to larger reductions in ABC-I scores compared with placebo, showing the highest probability among monotherapies (surface under the cumulative ranking curve: 98% and 85%). Among Ris combination regimens, the most favorable effect was observed for the addition of topiramate, with an MD of 7.55 (95% CI: 5.89-9.21) relative to Ris + placebo. In terms of safety, pharmacotherapy was associated with an absolute adverse events (AEs) risk increase of 0.08 versus placebo, with no severe events reported. CONCLUSION: Ris and aripiprazole had the highest ranking probabilities among monotherapies, and topiramate was the most promising Ris adjunct in the NMA. Pharmacotherapy carried a slightly elevated AE risk, with no severe events. However, rankings should be interpreted cautiously because most comparisons were indirect. Large high-quality head-to-head trials are needed to confirm comparative efficacy and long-term safety.