Prophylactic Antiepileptic Drugs in Nontraumatic Intracerebral Hemorrhage: A Network Meta-Analysis.

BACKGROUND: The utility of antiepileptic drugs (AEDs) for seizure prophylaxis following nontraumatic intracerebral hemorrhage (ICH) remains uncertain, with guidelines based on low-quality evidence that often pools heterogeneous medications. We conducted a network meta-analysis (NMA) to evaluate the utility of AEDs and to compare the efficacy and safety of individual agents in this setting.

METHODS: A comprehensive systematic search of PubMed, Embase, Cochrane Library, and Web of Science up to 22 October 2025 was conducted. We included studies involving adults with nontraumatic ICH that compared individual prophylactic AEDs against each other or a control group. We employed a Bayesian random-effects NMA to synthesize the data. Primary outcomes were early seizure incidence and unfavorable functional outcome. Secondary outcomes included adverse events, serious adverse events, NIHSS score at last follow-up, and mortality. Risk of bias was assessed using Cochrane Risk of Bias 2 (RoB 2) and Risk Of Bias In Non-randomized Studies-of Interventions, Version 2 (ROBINS-I V2) tools, and the certainty of evidence was evaluated with the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach.

RESULTS: Out of 2142 identified citations, 17 studies (4 RCTs, 13 observational; N = 6597 patients) met the inclusion criteria. No AED significantly reduced early seizures or improved functional outcomes compared with placebo. However, ranking probabilities (surface under the cumulative ranking curve, SUCRA) suggested valproate was most likely to prevent early seizures (SUCRA = 0.8330) and improve functional outcomes (SUCRA = 0.7938). Valproate also ranked highest for overall safety (SUCRA = 0.8085) and NIHSS improvement (SUCRA = 0.8187), while levetiracetam ranked best for serious adverse events (SUCRA = 0.7935). Eslicarbazepine was associated with increased mortality, whereas valproate showed the most favorable ranking for this outcome (SUCRA = 0.7705). The certainty of evidence across comparisons ranged from low to very low.

CONCLUSIONS: This NMA found no robust evidence to support the routine prophylactic use of AEDs after nontraumatic ICH. Nonetheless, valproate consistently demonstrated the best efficacy and relative safety profiles. This finding warrants investigation in future RCTs, particularly in subgroups of patients at high risk for post-ICH seizures.