Comparative efficacy and safety of pharmacological prophylactic agents for migraine in children and adolescents: a systematic review and bayesian network meta-analysis

Migraine is a leading cause of recurrent disabling headache in children and adolescents, yet most prophylactic agents have failed to outperform placebo, and head-to-head trials are scarce. We conducted a network meta-analysis (NMA) comparing the efficacy and safety of pharmacological migraine prophylaxis in pediatric patients. We searched PubMed, Cochrane CENTRAL, Scopus, and Embase for RCTs of migraine prophylaxis in patients < 18 years. Outcomes were change in frequency, severity, and duration (SMD), ≥ 50% responder rate (OR), and adverse events (OR), analyzed via Bayesian random-effects NMA (gemtc/JAGS) with node-splitting for inconsistency, SUCRA for ranking, and CINeMA for certainty. Risk of bias was assessed with RoB 2. 52 trials ( 4567 participants, 37 agents) were included. For frequency, riboflavin 200 mg (SMD 1.95), topiramate plus vitamin D (SMD 1.40), and levetiracetam (SMD 0.94) ranked highest, while topiramate, propranolol, and amitriptyline did not separate from placebo. For severity, sodium valproate (SMD 2.04), pregabalin (SMD 1.75), and cinnarizine (SMD 0.87) were superior to placebo. Duration data were sparse and inconsistent. No agent showed a robust responder-rate advantage. Levetiracetam had higher adverse-event odds than placebo (OR 6.54). Risk of bias was high or of some concern in most trials, with moderate-to-high heterogeneity. No agent showed consistent superiority across outcomes; efficacy was outcome-specific rather than universal. Riboflavin, topiramate-vitamin D, and levetiracetam showed the most consistent frequency signal, but CINeMA found no comparison reached high or moderate certainty, most rated very low, reflecting sparse, indirect, high-risk-of-bias evidence. Findings are hypothesis-generating pending larger head-to-head trials. [PROSPERO ID: CRD420261454779]