Abstract

1. INTRODUCTION

The primary goal of antiseizure medication (ASM) treatment in epilepsy is to achieve seizure freedom without intolerable side effects. ¹ Currently, over one‐third of patients with epilepsy have uncontrolled seizures despite treatment with multiple ASMs. ² The International League Against Epilepsy defines drug‐resistant epilepsy (DRE) as the failure to achieve seizure freedom by using two ASMs at the correct dose, either in sequential monotherapy or combination. ³ Few patients with DRE achieve seizure freedom, and the odds of achieving seizure freedom decrease drastically with the failure of each additional ASM prescribed. ² This situation has a relevant impact on patients' lives since DRE is often disabling and associated with more comorbidities, including psychological and social dysfunction, and an increased risk of premature death. ⁴

Furthermore, patients with epilepsy, especially DRE, experience a high burden of polypharmacy, leading to a higher probability of side effects and lower adherence rates. ⁵ Since the ultimate goal of epilepsy treatment is to achieve seizure freedom with the fewest adverse events (AEs), drug overload reduction seems to be a smart strategy to optimize the efficacy, tolerability, and adherence of a newly initiated ASM. In this scenario, there is an urgent need to develop breakthrough ASMs that increase seizure freedom rates with a balanced tolerability profile, allowing a reduction of co‐ASM overload in a rational polytherapy fashion.

Cenobamate (CNB) is a newly marketed ASM with a unique, dual, and complementary mechanism of action, ^{6 , 7 , 8} acting as a positive allosteric modulator of GABA‐A receptors (at a non‐benzodiazepine binding site) and as a sodium channel blocker (SCB) selectively acting on the persistent sodium current. ^{7 , 8} In Europe, it is indicated for the adjunctive treatment of focal‐onset seizures with or without secondary generalization in adults with epilepsy who have not been adequately controlled despite treatment with ≥2 ASMs. ⁶ CNB was authorized by the Food and Drug Administration (FDA) for the treatment of partial‐onset seizures in adult patients, ⁹ and by the European Medicines Agency (EMA), ⁶ based on efficacy and safety results versus placebo in several randomized trials in adults with focal‐onset seizures taking one to three ASMs. ¹⁰ Long‐term data on CNB in the open‐label studies support the results from the double‐blind placebo‐controlled trials. ^{11 , 12 , 13}

Given the high seizure freedom rates reported in the clinical trials ^{10 , 14 , 15} and the potential of CNB to reduce co‐ASM overload, ¹⁶ the main objective of our research was to explore the effectiveness and safety of co‐ASM optimization after initiating CNB in patients with DRE in a real‐world setting. This investigation will enhance the evidence in routine clinical practice.

2. MATERIALS AND METHODS

3. RESULTS

3.1. Patient disposition and baseline characteristics

The electronic medical records of patients receiving CNB as part of the EAP program were screened; 34 subjects were selected for the study as they met the inclusion/exclusion criteria and had received at least one dose of CNB.

The numbers of subjects in the PP population were 32, 29, and 21 at 3, 6, and 12months of follow‐up, respectively. Concerning the ITT population, the numbers of subjects were 34, 33, and 28 at 3, 6, and 12months of follow‐up, respectively. The retention rate of CNB was 94% (32/34 patients), 88% (29/33 patients), and 77% (21/28 patients) at 3, 6, and 12months of follow‐up, respectively (Figure S1). Seven patients discontinued CNB, including two in the first 3months (due to lack of tolerability or seizure worsening), two between months 3 and 6 (due to pneumonia‐associated encephalopathy or seizure worsening), and three between months 6 and 12 (two due to lack of effectiveness and one due to seizure worsening). Although four patients in the study experienced skin rash with SCB, none had a skin reaction due to CNB.

In our cohort, most subjects were female (58.8%), the mean±SD age of disease onset was 13.6±11.4years, the median (IQR) duration of epilepsy was 22 (11) years, and more than half of individuals showed some comorbidity (58.8%). Approximately one‐quarter of subjects (23.5%) had received previous epilepsy surgery. The mean age of treatment initiation with CNB was 36.7±10.5years, and the median (IQR) number of previous ASMs was 12 (7). Finally, the median (IQR) monthly focal seizure frequency at baseline was 15.5 (44) (Table 1).

TABLE 1

Patient demographics and disease characteristics at baseline.

Demographics and characteristics	All patients N=34
Gender
Female, n (%)	20 (58.8%)
Mean age, years (SD)	36.7 (10.5)
Age at epilepsy onset, years
Mean (SD), [range]	13.6 (11.4), [41]
Median (IQR)	10 (17)
Duration of epilepsy, years
Mean (SD), [range]	23.1 (11.5), [49]
Median (IQR)	22 (11)
Monthly seizure frequency
Mean (SD), [range]	57.8 (156), [898]
Median (IQR)	15.5 (44)
Etiology
Structural	13
Genetic	1
Infectious	0
Metabolic	0
Autoimmune	1
Unknown	19
Number of previous ASMs
Mean (range)	12.1 (13)
Median (IQR)	12 (7)
Number of concomitant ASMs
Mean (SD)	2.9 (1)
Median (IQR)	3 (1)
Prior surgery, n (%)	8 (23.5%)
Comorbidities, n (%)	20 (58.8%)
Encephalopathy/Intellectual disability, n (%)	8 (23.5%)

Open in a separate window

Abbreviations: ASMs, antiseizure medications; IQR, interquartile range; SD, standard deviation.

3.2. Safety

In the PP population, 23/32 (71.9%) participants reported at least one ADR at 3months of follow‐up. The most frequent ADRs were somnolence and dizziness. The other reported ADRs were ataxia and diplopia.

At 12months of follow‐up, 11/21 (52.3%) participants in the PP population reported at least one ADR; all ADRs occurred less frequently at 12months than at 3months. The most frequently reported ADR at 12months was somnolence (10/21 participants; 47.6%). The percentage of patients reporting dizziness reduced from 56.3% at 3months of follow‐up to 14.3% at 12months. Similarly, the percentage of patients reporting ataxia improved from 37.5% at 3months to 4.76% at 12months, and the percentage reporting diplopia reduced from 18.8% to 4.76% during this period. See Table 2 for more detail.

TABLE 2

Frequency and percentage of adverse events for subjects treated with cenobamate (PP population).

ADR	Month 3 (n=32)	Month 6 (n=29)	Month 12 (n=21)
Somnolence	22 (68.8)	17 (58.6)	10 (47.6)
Dizziness	18 (56.3)	10 (34.5)	3 (14.3)
Ataxia	12 (37.5)	5 (17.2)	1 (4.76)
Diplopia	6 (18.8)	2 (6.90)	1 (4.76)
Other ADR	3 (9.38)	0 (0)	0 (0)

Open in a separate window

Abbreviation: ADR, adverse drug reaction.

3.3. Co‐ASM management

At the baseline visit, patients treated with CNB received a mean of 2.9±1 co‐ASMs, which decreased significantly to 1.6±0.9 (p<0.0001) after one year of treatment with CNB (PP population; Figure 1A).

Open in a separate window

FIGURE 1

Total reduction in co‐ASMs during follow‐up (PP population). (A) Evolution of the number of co‐ASMs per subject. (B) Mean DDD for subjects treated with CNB by study period. (C) Evolution of percentage of subjects receiving different numbers of co‐ASMs; differences in height between corresponding shaded areas (i.e., the same color) within the bars and areas immediately to the right of the bars indicate discontinuations or incomplete follow‐up between timepoints. Vagus nerve stimulation and a ketogenic diet were removed from the analysis. p‐values are based on a paired t‐test versus baseline; **p<0.0001; ***p<0.0001. Co‐ASMs, antiseizure medications co‐administered with cenobamate; DDD, daily defined dose; PP, per protocol.

The total DDD per patient significantly reduced over time, from a mean of 3.6±1.3 at baseline to 1.4±1.2 (p<0.001) at 1year of treatment with CNB. The evolution of total DDD per patient throughout the study for the PP population is shown in Figure 1B.

After CNB initiation, the number of individuals receiving multiple co‐ASMs decreased significantly. In the PP population, 21/34 (61.8%) patients received three or more co‐ASMs at baseline, which reduced to 3/21 (14.3%) patients after 12months. In contrast, the number of patients treated with one or two co‐ASMs increased from 13/34 (38.2%) at baseline to 17/21 (81%) after 12months. At the end of the study, 1 (4.8%) subject was receiving CNB as the only ASM (Figure 1C).

The most frequently used co‐ASMs at baseline were carbamazepine (CBZ) in 20/34 (58.8%) participants, clobazam (CLB) in 16/34 (47.1%) participants, and brivaracetam (BRV) and lacosamide (LCM) in 9/34 (26.5%) participants each. Reductions in total DDD and in the number of participants taking each different co‐ASM were observed during follow‐up; these occurred for ASMs with different mechanisms of action (Figures 2 and S2; Table S2). The greatest reductions in mean co‐ASM total DDD between baseline and month 12 occurred for SCB (mainly CBZ and LCM) and GABAergics (CLB) (Tables S1 and S2; Figures S2 and S3).

Open in a separate window

FIGURE 2

Reduction in co‐ASM DDD during follow‐up for each drug. The height of the bars represents DDD and the width represents the number of subjects on each drug (the numerical value is also displayed over each graph). AAZ, acetazolamide; BRV, brivaracetam; CBZ, carbamazepine; CLB, clobazam; CLZ, clonazepam; co‐ASMs, antiseizure medications co‐administered with cenobamate; DDD, daily defined dose; DZP, diazepam; ESL, eslicarbazepine; ETX, ethosuximide; LCM, lacosamide; LEV, levetiracetam; LTG, lamotrigine; PB, phenobarbital; PER, perampanel; PHT, phenytoin; TPM, topiramate; VGB, vigabatrin; VPA, valproic acid; ZNS, zonisamide.

For all ADRs studied, the mean DDD of SCB and GABAergics was reduced to a significantly greater extent relative to baseline in patients with versus without ADRs at month 3; these differences were maximal at month 6 (Figure 3B). At the 6‐month timepoint, the mean percent DDD for SCB and GABAergics dropped to 26.6±22.9% of baseline in patients with any ADR at month 3 versus 58.4±32.9% of baseline in patients without an ADR at month 3; corresponding values at the 12‐month timepoint were 14.2±27.0% versus 43.3%±34.4% of baseline, respectively. This same comparison showed a significant relative reduction of the mean DDD for SCB and GABAergics when analyzing individual ADRs at month 6 for diplopia (17.3±4.3% vs. 42.0±32.0% of baseline; p<0.005), dizziness (20.1±15.9% vs. 50.5±37.7% of baseline; p<0.005), and somnolence (26.4±23.6% vs. 56.3±32.5% of baseline; p<0.01), but not for ataxia (19.1±17.4% vs. 43.6±32.1% of baseline; p=0.057). Additionally, patients who presented with ADRs at 3months of follow‐up showed a higher mean DDD of SCB at baseline when compared with participants with no ADRs (Figure 3A), although these differences were not statistically significant.

Open in a separate window

FIGURE 3

Differential initial dose and management of co‐ASMs between participants with and without ADRs at month 3 of follow‐up. (A) Mean DDD of GABAergics or SCB at baseline for participants with and without ADRs at month 3; (B) Percentage reduction in the mean DDD of GABAergics or SCB from baseline for participants with and without ADRs at month 3. p‐values are based on a paired t‐test comparing “with” versus “without” ADR; *p<0.05. ADR, adverse drug reaction; DDD, daily defined dose; SCB, sodium channel blockers.

3.5. Effectiveness

Responder rates at the ≥50%, ≥75%, ≥90%, and 100% levels are shown in Figure 4 for the PP population. Responder rates at the ≥50% level were 81.2%, 82.8%, and 85.7% at 3, 6, and 12months of follow‐up, respectively. Responder rates at the ≥75% level were 62.5%, 69%, and 71.4% at 3, 6, and 12months, respectively. At the ≥90% level, responder rates at 3, 6, and 12months were 43.8%, 58.6%, and 47.6%, respectively. The percentages of seizure‐free patients (100% responder rate) at 3, 6, and 12months were 21.9%, 44.8%, and 33.3%, respectively. Efficacy results for the ITT population are shown in Figure S4.

Open in a separate window

FIGURE 4

Responder rates during the study, by period. PP population includes all subjects who completed Month 3/6/12 and had not discontinued treatment. PP, per protocol.

At the end of the follow‐up period, 41.2% of participants reported clinically significant satisfaction with the drug in terms of an improvement in seizure frequency (i.e., PGI‐I score 1 [“Very much better”] or 2 [“Much better”]; Table S4).

4. DISCUSSION

Retrospective real‐world studies with newly approved ASMs complement data from clinical trials and inform specialists about the performance of new medications in populations not captured during clinical development. Furthermore, EAPs allow patients with no other options to access drugs and physicians to gain experience before commercialization. ¹⁸ Only a few other studies have analyzed the effectiveness and tolerability of CNB in patients with focal‐onset DRE in a real‐world setting as part of the Spanish EAP. ^{15 , 17 , 19} These studies and the present investigation demonstrate that adjunctive CNB is effective and safe in clinical practice. Specifically, our study revealed that polytherapy optimization while initiating treatment with CNB results in high seizure freedom rates. In real‐world studies, these rates are even higher than those reported by clinical trials. Besides, a high level of satisfaction and a favorable tolerability profile accompanied the improvement in clinical variables.

Analysis of the baseline characteristics reveals that most subjects had not undergone previous epilepsy surgery. Our cohort includes DRE patients with ≥12 previous ASMs. This population, the so‐called absolute drug‐resistant (failure of ≥6 ASMs), had a probability of achieving seizure freedom close to zero. ^{2 , 20 , 21} However, after treatment with CNB, a quarter of patients in the ITT population achieved seizure freedom at month 12.

Despite an individualized approach to epilepsy therapy and the availability of many ASM combinations, physicians may find themselves in a loop of transitional polytherapy with DRE patients. ¹⁶ In this population, polytherapy is associated with a high number of adverse effects, which are stronger predictors of quality of life than seizures. Besides, treatment complexity has been associated with lower adherence and pharmacological interactions, possibly impacting effectiveness and safety. ^{22 , 23}

A “start low, go slow” titration approach has improved tolerability when starting a new ASM. When adding a new ASM, the adjustment of co‐ASMs to achieve seizure control with the lowest drug load is also key to minimizing tolerability issues. ²⁴ In addition, if tolerability issues arise when adding a new ASM, lowering the doses of co‐ASMs that failed to achieve seizure freedom is an option rather than discontinuing the new ASM. In patients with DRE, such a strategy allows continuation of the new ASM titration and reduction of adverse effects without negatively affecting seizure control, ²⁵ allowing seizure freedom to be reached in the optimal scenario. The titration schedule of CNB (every 2weeks) and its demonstrated early efficacy ^{26 , 27} allow physicians to increase the CNB dose slowly while reducing co‐ASMs.

A panel of epileptologists developed ASM dose‐reduction recommendations for initiating CNB in adults. ^{28 , 29} These include proactive dose adjustments before the report of side effects for clobazam, SCBs, phenytoin, and phenobarbital. Reactive dose adjustments in response to side effects are recommended for all other ASMs at standard doses. A study conducted by Steinhoff et al. ²⁶ has demonstrated the feasibility of reducing co‐ASM DDD in patients with focal‐onset seizures who initiated CNB. Similarly, in a post hoc analysis by Rosenfeld et al., ¹⁶ patients who continued CNB tended to have more significant reductions in co‐ASM doses than those who discontinued the drug.

Our study highlights the therapeutic relevance of managing co‐ASMs while escalating CNB doses to improve the effectiveness‐tolerability balance. At the beginning of the study, 61.8% of patients in the PP population received three or four co‐ASMs. In comparison, after 12months, this proportion was only 14.3%, significantly increasing the rate of subjects treated with only one or two co‐ASMs after 1year (more than three‐quarters of the PP population). Improvement in tolerability was followed by a high reduction in seizure frequency. A third of patients in the PP population achieved seizure freedom at month 12, and patients showed high retention rates after reducing or discontinuing ≥1 co‐ASM. It is also worth noting that co‐ASM doses were reduced for many patients very early in the CNB titration.

Following the principles of rational therapy, the highest reduction in drug load after starting CNB in our cohort occurred with SCB and GABAergics (mainly carbamazepine, lacosamide, and clobazam). This reduction of co‐ASMs was most evident in those patients who presented with ADRs, which might have allowed us to achieve better tolerability results than in previously reported studies.

Furthermore, as both an enzyme inducer and inhibitor of different CYPs, CNB affects the metabolism of other ASMs, such as clobazam and its active metabolite N‐desmethylclobazam (N‐CLB). ³⁰ Potential interactions between CNB and clobazam have been described, ^{6 , 31} and their combination results in a large and clinically significant increase in the serum concentration of N‐CLB. ³² In our study, some patients who received a single ASM and were seizure‐free experienced seizure recurrence at month 6. However, seizure freedom was achieved again after re‐introducing very low clobazam doses (2.5mg/day in most cases). This potentially beneficial synergy between CNB and low clobazam doses between months 6 and 12 in five of our patients has been observed in another series published by Osborn et al. ³³ As CNB has CYP3A4‐induction effects, decreasing levels of clobazam with higher doses of CNB are to be expected. However, we proactively reduced clobazam to reduce pharmacodynamic interactions that, in our experience, are a major contributor to adverse effects. Consequently, after the introduction of CNB, the management of polypharmacy in this study achieved very good effectiveness results despite very low numbers of co‐ASMs.

Notably, there were no serious ADRs in our study, and ADRs could be adequately addressed or resolved. Although four of the patients in the study had suffered a skin rash with SCB before initiating treatment with CNB, none experienced skin reactions during CNB treatment. At 3months of follow‐up, the rate of ADRs was comparable to rates described in previous studies. ³³ Most importantly, the ADR rate was significantly reduced at 6 and especially at 12months of follow‐up. The low rate of ADRs occurred in parallel with the simplification of co‐ASMs, especially of GABAergics and SCB. As stated previously, co‐ASMs were significantly reduced in all patients. However, the reduction of co‐administered GABAergics and SCB occurred fastest and to the greatest extent when ADRs were detected. Therefore, reactive co‐ASM reduction might have allowed us to reduce ADRs. It is also important to note that GABAergic and SCB DDD showed a tendency to be highest in patients who presented with ADRs 3months after initiation of CNB. Thus, the appearance of ADRs should be more closely followed in patients with higher dosages of these ASM groups in order to optimize ADR management.

CNB may also be distinguished from other ASMs by the high rates of seizure freedom not seen in previous placebo‐controlled trials, which has the potential to improve quality of life significantly. ^{15 , 34} Other retrospective studies, including in patients receiving CNB during the EAP, ^{14 , 15} reported lower seizure freedom rates than those observed in our study. We believe these results might have been possible due to the high doses of CNB that were reached in our study, which might have been better tolerated thanks to the previously described reduction of co‐ASMs. Along with the marked seizure freedom rates observed, the reported 35.7% rate of patients in the ITT population with a ≥90% reduction in seizure frequency at 12months is also remarkable. Also, the retention rate of CNB after 12months of therapy was high since 77% of the subjects were still on CNB treatment at the end of the study.

The high effectiveness results shown in our study were obtained thanks to several factors. First, CNB has a well‐characterized safety profile and exhibits predictable side effects, which are mainly mild to moderate and, in most cases, dose dependent. ³² Second, there was a low rate of tolerability issues, which might have been due to the management and reduction of co‐ASMs. Third, the long half‐life of CNB allows for a single daily dose. Fourth, the biweekly titration schedule and the low starting dose (12.5mg) recommended in the Summary of Product Characteristics (SmPC) ⁶ favor tolerability and allow co‐ASM reduction, ³⁵ alongside the early efficacy of CNB.

This research also served to accumulate experience in the management of CNB, so there is increasing knowledge regarding how to titrate and deal with the tolerability issues related to the addition of CNB in highly polymedicated patients with DRE. Finally, it is also noteworthy that the high PGI‐I scores were encouraging, as many participants were satisfied with the results achieved with CNB.

This study has several limitations, including its retrospective, single‐center design. Also, results are mainly focused on the per‐protocol population, although efficacy data for the ITT population are also provided. Additionally, seizure counts in observational studies are less accurate than in clinical trials. The limited sample and the lack of adherence data suggest that the results should be interpreted cautiously. Nonetheless, our results align with those previously reported in clinical trials and other real‐world studies, providing additional evidence of the effectiveness and tolerability of adjunctive CNB in polymedicated patients with highly resistant focal epilepsy. In addition, this study may inform other physicians about the importance of treatment optimization while introducing CNB to avoid drug overload, and the subsequent impact on tolerability, adherence, and treatment satisfaction. More real‐world studies are needed to consolidate our findings.

In conclusion, the results of this study demonstrate that high seizure freedom rates, together with a significant dose reduction in the number and dose of co‐ASMs, can be achieved with the use of adjunctive CNB in patients with absolute drug‐resistant focal‐onset seizures.

AUTHOR CONTRIBUTIONS

JRU led clinical research, data collection, review of the analysis of results, and approved the final version of the manuscript. JMSC and RHR collected data, reviewed and revised manuscript drafts, and approved the final version.

CONFLICT OF INTEREST STATEMENT

Juan Jesús Rodríguez‐Uranga has been a paid consultant for EISAI Farmaceutica, GW Pharma, Jazz Pharmaceutical Iberia, UCB Pharma, Bial, Novartis, Livanova, Angelini Pharma, Arvelle Therapeutics, GlaxoSmithKline, Pfizer, and Esteve. Juan M. Sánchez Caro has served as a consultant for Angelini Pharma and Nutricia. Roshan Hariramani Ramchandani has carried out consulting work for Eisai. 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.

Supporting information

ACKNOWLEDGMENTS

Notes

Rodríguez‐Uranga JJ, Sánchez‐Caro JM, Hariramani Ramchandani R. Treatment simplification to optimize cenobamate effectiveness and tolerability: A real‐world retrospective study in Spain. Epilepsia Open. 2024;9:1345–1356. 10.1002/epi4.12959 [Europe PMC free article] [Abstract] [CrossRef] [Google Scholar]

REFERENCES