Response to: "A critical appraisal of principal component analysis of antiseizure medication-induced hostility/aggression and factor analysis of levetiracetam".

We respond to the commentary by Aphale et al. on our study titled "Principal Component Analysis of Antiseizure Medication-induced Hostility/Aggression and Factor Analysis of Levetiracetam" using the Food and Drug Administration Adverse Event Reporting System (FAERS). The commentary raised four methodological points: (1) the interpretation of Component 2 as a "behavioral axis" in the principal component analysis and the classification of borderline personality disorder (BPD), (2) confounding by indication regarding risperidone, (3) the choice of the reporting odds ratio (ROR) over Bayesian methods such as the Empirical Bayes Geometric Mean, and (4) the absence of multivariate analysis in the pediatric group aged less than 12 years. We clarify that BPD was classified among the eight antiseizure medication-associated adverse effects primarily because of its high loading on Component 1 (PC1 = 0.816), the hostility/aggression risk axis, rather than on Component 2 (PC2 = 0.045), which was near-neutral. Component loadings for all 15 preferred terms are provided in Table 1. We acknowledge limitations related to confounding by indication inherent in spontaneous reporting databases, discuss the rationale for selecting the ROR as the primary signal-detection metric, and clarify data-driven constraints that precluded multivariate analysis in the pediatric subgroup. We propose that future studies use larger data sources, such as electronic health records and disease registries, including clinical indications and patient background to enable robust risk factor analyses in this population.