Perampanel monotherapy in pediatric epilepsy: Emphasizing the need for comprehensive safety evaluation

We read with great interest the recent article by Gu et al. titled “Clinical efficacy and safety of perampanel monotherapy as primary anti-seizure medication in the treatment of pediatric epilepsy: A single-center, prospective, observational study”.¹ The study highlighted the high efficacy and safety of perampanel (PER; a noncompetitive antagonist of the (AMPA) - α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid - glutamate receptor) monotherapy in pediatric patients aged 4–18 years with epilepsy, demonstrating seizure freedom rates exceeding 70% at various observation points and a retention rate of 71.58% at 12 months. Notably, the overall adverse event rate was 38.71%, with irritability and dizziness being the most common adverse effects. These findings underscore the potential of PER as an effective therapeutic option for pediatric epilepsy, offering favorable seizure control with a relatively lower maintenance dose for patients who respond well and adhere to long-term treatment.

While the study provides valuable insights into the clinical benefits of PER monotherapy in children, we believe that a more comprehensive evaluation of its safety profile is crucial, especially considering the vulnerable pediatric population. The limited scope of adverse effects reported, primarily mild to moderate, may not fully capture the potential risks associated with PER, particularly in long-term use. To address this gap, we examined existing data from established drug databases and scientific literature to evaluate the toxicological and safety considerations of PER.

Utilizing information from databases such as ChemBL (https://www.ebi.ac.uk/chembl/), PubChem (https://pubchem.ncbi.nlm.nih.gov/), DrugBank (https://go.drugbank.com/), and reports from PreADMET (https://preadmet.webservice.bmdrc.org/), FAFDrugs4 (https://fafdrugs4.rpbs.univ-paris-diderot.fr/), ADMETSAR (http://lmmd.ecust.edu.cn/admetsar2), MolInspiration (https://www.molinspiration.com/cgi-bin/properties), pkCSM (https://biosig.lab.uq.edu.au/pkcsm/prediction), SwissADME (http://www.swissadme.ch/), ADMETlab 2.0 (https://admetmesh.scbdd.com/), and ADMET-AI (https://admet.ai.greenstonebio.com/) web servers, we gathered insights into the structural, physicochemical, and toxicological properties of PER (details in ²). Several concerns regarding potential adverse effects emerged from this analysis (Figure 1) and are discussed below.

First, PER exhibits physicochemical properties that may predispose it to non-specific binding and adverse drug reactions. Specifically, it has a calculated LogP (oil/water partition coefficient) of 3.374 and a topological polar surface area (TPSA) of 58.68 Ų. According to established drug design principles, compounds with LogP values greater than 3 and TPSA less than 75 Ų are associated with a higher likelihood of promiscuous binding and immunotoxicity.³ This suggests that PER may have the potential for non-specific interactions, leading to idiosyncratic hypersensitivity reactions.^{3, 4} This concern is supported by clinical evidence linking perampanel to severe cutaneous adverse reactions (SCAR), including drug reactions with eosinophilia and systemic symptoms (DRESS) and Stevens–Johnson syndrome (SJS), as documented in pharmacovigilance reports and the drug's regulatory labeling.⁵ One case study described a 13-year-old girl who developed a life-threatening DRESS syndrome following an increase in PER dose, presenting with fever, erythroderma, hypotension, acute renal and hepatic dysfunction, and respiratory distress, requiring intensive care and mechanical ventilation. The histopathologic findings confirmed the diagnosis, and her condition improved only after discontinuation of PER and initiation of immunosuppressive therapy.⁵ Such cases emphasize the need for close monitoring of hypersensitivity reactions in patients receiving PER, especially given its potential to trigger severe immune-mediated reactions.⁶

Renal clearance (CLr) plays a key role in drug elimination and pharmacokinetics. Prediction models (ADMETlab 2.0,⁷ pkCSM⁸) estimate a CLr of approximately 2.693–3.67 mL/min/kg for PER, indicating moderate renal excretion. However, empirical data from clinical studies report an absolute CLr of ~12 mL/min in healthy adults,⁹ corresponding to ~0.171 mL/min/kg, which is lower than the in silico predictions. Recent evidence suggests that the clearance of PER is significantly affected by renal function, particularly in moderate to severe renal impairment. Therapeutic drug monitoring (TDM) studies have shown that patients with normal renal function have a concentration-to-dose (CD) ratio of 1740 ± 966 ng·mL·mg·kg, while this ratio increases significantly in patients with severe renal impairment (5327–9113 ng·mL·mg·kg),¹⁰ suggesting reduced drug clearance. In addition, a 14% reduction in clearance was observed in subjects with creatinine clearance (CCr) between 30 and 50 mL/min compared to subjects with normal renal function. These results suggest that although PER is primarily metabolized by hepatic CYP3A4 pathways, renal function modulates systemic exposure, highlighting the need for dose adjustments in patients with impaired renal function. The convergence of in silico predictions and clinical pharmacokinetic assessments underscores the value of integrated approaches in refining dose optimization strategies, particularly for antiepileptic drugs with dual elimination pathways.

The primary route of excretion of PER is hepatic metabolism, which undergoes extensive oxidation and glucuronidation, mainly via CYP3A4/5, CYP1A2, and CYP2B6, resulting in ~70% of the metabolites being excreted in the feces and ~30% in the urine.¹¹ The reported hepatic clearance (Clh) of 0.730 L/h (12.2 mL/min) in adult males and 0.605 L/h (10.1 mL/min) in females emphasizes the predominance of hepatic metabolism over renal excretion, and this profile is consistent with its high lipophilicity (LogP 3.37), which favors hepatic metabolism over direct renal clearance. The prolonged half-life (~105 h) of PER is largely determined by this hepatic metabolism and requires careful dose adjustment in patients with hepatic impairment or concomitant administration of CYP3A4 inducers or inhibitors, which may significantly alter drug plasma levels.

Moreover, data from the literature suggest potential cardiotoxicity associated with PER due to its interaction with human ether-à-go-go-related gene (hERG) potassium channels.

Data from established drug databases further identified a potential interaction between PER and hERG potassium channels, utilizing a computational model that assessed the drug against a library comprising over 5000 molecular entities. The blockade of hERG potassium channels is a well-documented mechanism linked to the onset of fatal cardiac arrhythmias, making it a pivotal anti-target in the early stages of drug development due to its high susceptibility to unintended drug interactions.¹² Despite PER's clinical profile being primarily associated with central nervous system-related adverse events—such as dizziness, somnolence, fatigue, irritability, and nausea—without reported cardiotoxicity in conventional therapeutic settings, the potential risk of hERG-related cardiotoxicity remains a significant concern.^{13, 14} This potential risk warrants further investigation, particularly in long-term use or in patients with pre-existing cardiac conditions.

Nevertheless, we recognize that previous clinical studies have not reported significant prolongation of the QT interval in electrocardiography with the use of PER, as shown in the grouped phase III studies.¹⁵ These studies suggest that therapeutic doses (6–12 mg) do not cause QT interval prolongation, with no evidence of clinically relevant arrhythmic risk. In addition, the drug has been shown to improve cardiovagal tone and increase heart rate variability (HRV) in patients with drug-resistant temporal lobe epilepsy, suggesting a cardioprotective effect associated with increased vagal activity.¹⁶ In addition, bradycardia rather than tachyarrhythmias has been reported in cases of pediatric PER overdose, supporting this autonomic modulation.¹⁷ Despite the clinical findings, our in silico predictions remain relevant because computational models, such as those used in this study, assess direct molecular interactions with the hERG channel, which do not necessarily result in QT interval prolongation in vivo due to compensatory mechanisms such as vagal modulation. In silico screening is particularly useful for early assessment of the risk of cardiotoxicity to identify potential off-target interactions before comprehensive clinical data are available. This is critical because many drugs with an affinity for hERG do not lead to QT interval prolongation but may still pose risks under certain physiologic conditions, such as metabolic changes or drug interactions. Although current evidence suggests that QT interval prolongation does not occur, the long-term effects of PER on cardiac electrophysiology have not yet been fully characterized, particularly in populations with pre-existing cardiac disease or polytherapy regimens that could influence hERG-related effects.

Additionally, there is evidence indicating that PER may pose a risk of respiratory system toxicity. Respiratory toxicity is often underdiagnosed due to subtle early signs and can lead to significant morbidity and mortality.¹⁸ While PER has not been associated with respiratory depression at therapeutic doses, cases of respiratory failure have been reported following accidental overdose,¹⁹ highlighting the need for caution and monitoring of respiratory function during treatment.

Furthermore, concerns have been raised about hepatotoxicity, as drug-induced liver damage has been one of the main reasons for the withdrawal of medicines from the market over the last 66 years.²⁰ Clinical observations and two hepatotoxicity models using a dataset of 951 compounds with different effects on the liver in different mammalian species indicate that PER may impair liver function.²¹ Studies have reported liver function abnormalities in children treated with PER,²² emphasizing the importance of regular hepatic monitoring during therapy.

The Ames test, a standard assay for detecting mutagenicity by identifying DNA mutations in bacterial cells, is widely used as a preliminary indicator of carcinogenic potential due to its strong correlation with carcinogenicity.²³ Predictions based on this assay suggested that PER exhibits mutagenic properties, indicating a potential risk for DNA damage in both bacterial and human cells, which warrants further investigation. Similarly, data from the PreADMET server identified PER as both mutagenic and carcinogenic according to Salmonella mutagenicity and mouse carcinogenicity models, respectively. Complementing these findings, a micronucleus assay also classified PER as genotoxic.²⁴ These results collectively point to a concerning genotoxic and carcinogenic profile, emphasizing the need for comprehensive long-term safety evaluations, as no current data address the long-term mutagenic or carcinogenic risks associated with PER.

Despite its demonstrated efficacy in treating epilepsy,^{14, 25} recent studies have highlighted significant rates of PER discontinuation due to inefficacy and adverse effects. Matsuyama et al.⁶ conducted a retrospective study and found that 44.9% of patients discontinued PER, primarily due to non-response, the occurrence of psychiatric adverse effects (PAE), common adverse effects (CAE), and psychiatric comorbidities. Adverse effects occurred in 65% of the patients, with 23.7% experiencing PAE and 49.2% experiencing CAE. These findings suggest that while PER is effective for some patients, a substantial proportion may discontinue treatment due to tolerability issues, underscoring the importance of careful patient selection and monitoring.

Considering these findings, we believe that, while PER shows promise as an effective anti-seizure medication in pediatric patients, a thorough assessment of its safety profile is essential. The potential adverse effects identified from drug databases and scientific literature highlight areas that may not be immediately evident in clinical observations but could have significant implications for long-term patient health.

We commend the authors for their significant contribution to pediatric epilepsy treatment. However, we advocate for additional preclinical and clinical studies to fully elucidate the risk–benefit profile of PER, ensuring its safe and effective use in children. A more comprehensive evaluation of potential adverse effects, including long-term monitoring and larger, multicenter studies, would be valuable in addressing these concerns.

None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.