P. Soares-da-Silva, Department of Research and Development, BIAL – Portela & Co SA, À Avenida da Siderurgia Nacional, 4745-457 S. Mamede do Coronado, Portugal e-mail: firstname.lastname@example.org
Objectives – To evaluate the efficacy and safety of eslicarbazepine acetate (ESL) as adjunctive therapy in adults with partial-onset seizures.
Material and methods – Double-blind, placebo-controlled, parallel-group, multicenter study consisting of an 8-week baseline period, after which patients were randomized to placebo (n = 87) or once-daily ESL 800 mg (n = 85) or 1200 mg (n = 80). Patients received half dose during 2 weeks preceding a 12-week maintenance period.
Results – Seizure frequency over the maintenance period was significantly (P < 0.05) lower than placebo in both ESL groups. Responder rate was 23% (placebo), 35% (800 mg), and 38% (1200 mg). Median relative reduction in seizure frequency was 17% (placebo), 38% (800 mg), and 42% (1200 mg). The most common adverse events (AEs) (>10%) were dizziness, somnolence, headache, and nausea. The majority of AEs were of mild or moderate severity.
Conclusions – Once-daily treatment with ESL 800 and 1200 mg was effective and generally well tolerated.
Eslicarbazepine acetate (ESL) is a novel voltage-gated sodium channel (VGSC) blocker that is chemically related to carbamazepine (CBZ) and oxcarbazepine (OXC), but with a different metabolism that may result in lower drug interaction potential (1). ESL exhibits anticonvulsant properties through selective interference with rapidly firing neurons and has been designed to retain its anticonvulsant potency even while reducing toxic metabolites such as epoxides, avoiding enantiomeric impurity, and diminishing the unnecessary production of enantiomers or diastereoisomers of metabolites and conjugates (2).
Following oral administration, ESL is rapidly and extensively metabolized to eslicarbazepine (the entity responsible for the pharmacological effect in humans) by a first pass hydrolytic metabolism (3). Plasma levels of ESL usually remain below the limit of quantification. The bioavailability of eslicarbazepine is high and >90% of the metabolites recovered in urine are eslicarbazepine and its glucuronic acid conjugate (3). ESL pharmacokinetics (PK) is not affected by the presence of food (4, 5), age (6), or gender (7), but clearance of ESL metabolites is dependent on renal function (8).
In a Phase II adjunctive therapy study in patients with partial epilepsy, once-daily ESL 800 mg and 1200 mg showed a favorable efficacy and tolerability profile (9). In a placebo-controlled Phase III study recently reported by Elger et al. (10), once-daily doses of ESL 400, 800 and 1200 mg were tested in adult patients with uncontrolled partial-onset seizures despite treatment with one or two antiepileptic drugs (AEDs). ESL was generally well tolerated and seizure frequency over the maintenance period (12 weeks) was significantly lower than placebo in the ESL 800 and 1200 mg groups.
The current Phase III study was designed to further investigate the dose-related efficacy, tolerability, and safety of ESL as adjunctive therapy in adults with refractory partial-onset seizures. In study by Elger et al. (10), ESL starting dose was 400 mg; thereafter, ESL was titrated at weekly 400-mg steps to the full maintenance dose (12 weeks). In the current study, a different titration regimen was investigated: patients were administered half of the maintenance dose for 2 weeks before reaching the full dose.
This double-blind, randomized, placebo-controlled, parallel group, Phase III study was conducted at 39 sites in three countries (Mexico, Portugal and Spain) between 14 December 2004 and 19 January 2007. Individual study participation lasted 26 weeks. Following enrollment, patients entered an 8-week baseline phase. The double-blind treatment phase comprised a 2-week titration period, a 12-week maintenance period and a 4-week tapering-off period. At the start of the double-blind phase, patients were randomized in a 1:1:1 ratio to one of the three following treatment groups: placebo, ESL 800 mg or ESL 1200 mg, once daily.
Randomization codes were prepared by a designated, unblinded statistician, generated using SAS® software (PROC PLAN procedure). Patients who successfully completed the 8-week baseline phase were assigned the lowest randomization number available at their site. During the 2-week titration period, all patients were treated with half their assigned dose before being titrated to their full dose (800 or 1200 mg) at the start of the 12-week maintenance period (Fig. 1). During the first 2 weeks of the tapering-off period, all patients returned to half of the assigned dose, and received placebo for the remainder of the period. Patients who completed the 18-week double-blind phase were eligible to enter open-label long-term treatment with ESL.
This clinical study was conducted according to the study protocol approved by local independent ethics committees and authorities, and in adherence to the principles enunciated in the Declaration of Helsinki and the local laws of the countries where the study was performed. Written informed consent was obtained from study participants.
Men and women were eligible for study enrollment if they satisfied all of the following criteria: aged 18 years or older; assessed as being in general good health, other than epilepsy; diagnosed with simple or complex partial seizures (with or without secondary generalization) for a minimum of 12 months prior to screening; experienced at least four partial-onset seizures, as defined by the International Classification of Epileptic Seizures (11) in the two 4-week periods prior to screening as well as during each of the two 4-week periods of the 8-week baseline period; treated with one to two concomitant AEDs in a stable dose regimen for at least 2 months prior to screening.
Patients were excluded from the study if, at the time specified, they had: an uncontrolled, relevant medical disorder; visual field loss caused by vigabatrin use (at least 1 year); simple partial seizures without motor symptoms; primary generalized epilepsy, rapidly progressive neurological disorder, status epilepticus, or cluster seizures (within 3 months); a history of seizures of psychogenic origin (within 2 years); a history of schizophrenia or suicide attempts; a known hypersensitivity to CBZ or OXC, or chemically-related substances. Patients were also excluded if they were: receiving OXC, felbamate, or a benzodiazepine except when used chronically as an AED; pregnant or breastfeeding women.
Efficacy Efficacy data were documented through patient diaries, in which participants (with or without assistance) recorded seizures by type, date and time of occurrence through the 26-week study period. The primary efficacy variable was seizure frequency standardized per 4 weeks, which was evaluated from diary entries. Secondary efficacy variables included the proportion of responders (i.e. patients with a ≥ 50% reduction in standardized seizure frequency), relative reduction in seizure frequency, number of days with seizures, distribution of seizure reduction (the number and proportion of patients with a seizure reduction of <50%, ≥50–75% or >75%), proportion of seizure-free patients (i.e. 100% seizure reduction), and proportion of patients with an exacerbation in seizure frequency ≥25% compared to baseline.
Safety Safety data included adverse events (AEs) that were classified according to the Medical Dictionary for Regulatory Activities (MedDRA; version 9.0) and defined as any undesirable change in the function, structure, or chemistry of the body occurring to a patient during the clinical study, whether or not considered related to study treatment. AEs were recorded from spontaneous reports, indirect inquiry or investigator observation. Safety data were evaluated based on clinical laboratory tests (hematology, coagulation, biochemistry, thyroid function and urinalysis), vital signs and weight, and electrocardiograms (ECGs). All AEs were recorded by the investigator and assessed according to intensity (mild, moderate or severe), causality (not related, unlikely, possible, probable or definite), and seriousness. A treatment-emergent AE (TEAE) was an AE that occurred after the first intake of study treatment during the double-blind phase.
The drug concentrations of eslicarbazepine and concomitant AEDs were analyzed through blood samples collected at the beginning and along the double-blind phase. PK data were obtained from blood trough levels of eslicarbazepine and concomitant AEDs.
All efficacy assessments were based on the intent-to-treat (ITT) population (randomized patients who had been administered at least one dose of investigational product and had at least one post-baseline seizure frequency assessment). Supportive assessments were performed using the per-protocol (PP) population (all ITT patients who completed the 12-week maintenance period without major protocol deviations). The safety population contained all randomized patients who had received at least one dose of investigational product.
For the primary efficacy analysis, the sample size estimation assumed that the change from baseline to the end of the maintenance period (using the natural logarithm [ln] of the seizure frequency per 4 weeks) was –0.05 for placebo and –0.30 for ESL groups. The common standard deviation (SD) was 0.50, with a power of 80% and an alpha-level of 0.025 (which, in turn, was based on a Bonferroni adjustment for two comparisons of the two ESL doses with placebo, each at an alpha level of 0.05). These assumptions yielded a sample size of 71 patients per treatment group. The drop-out rate was estimated at 15%, so 252 total patients were required for enrollment.
The primary efficacy variable was assessed in the three treatment groups by comparing the change in seizure frequency over the 12-week maintenance period to baseline; an additional analysis was performed using data from the 2-week titration period. An analysis of covariance (ANCOVA) model was used and an ln-transformation was applied to the standardized seizure frequency (per 4 weeks) to conform to the ANCOVA assumptions as well as to be consistent with the sample size determination. Estimates from the ANCOVA model then were back transformed using the exponential function. Dunnett’s multiple comparison procedure was used to compare ESL treatment least square means (LS Means) to placebo LS Means.
Secondary efficacy variables were analyzed using a Cochran–Mantel–Haenszel (CMH) test, which relied on an analysis of variance (ANOVA) to compare the two ESL treatment groups with placebo. Patients with missing data were considered as non-responders. The median percentage changes in seizure frequency were not tested formally, but an analysis by seizure type was explored. Safety and tolerability data, as well as trough plasma levels of eslicarbazepine and concomitant AEDs, were evaluated descriptively.
Continuous data were summarized using descriptive statistics. Categorical variables were analyzed using frequency (counts) and percentages. The criterion for statistical significance was set at an alpha error of 5% (P < 0.05), two-sided; P-values not related to the primary analysis were of a descriptive nature only.
Of the 330 patients screened, 252 (76.4%) were randomized and treated as part of the study population: 87 to placebo, 85 to ESL 800 mg, and 80 to ESL 1200 mg. There were seven patients who had no post-baseline efficacy data; thus, 245 patients were included in the ITT population. During the 26-week program, 194 (77.0%) patients completed the study (Fig. 2).
In the safety population, 35 patients (13.9%) terminated early during the 12-week maintenance period. The largest proportion of discontinuations occurred in the ESL 1200 mg group (n = 15; 18.8%), while the fewest discontinuations occurred in the ESL 800 mg group (n = 9; 10.6%); by comparison, 11 patients (12.6%) terminated in the placebo group. Most discontinuations were attributed to AEs (n = 14; 5.6%), which occurred in a treatment and dose-dependent manner.
The three treatment groups were similar in terms of demographic and background characteristics (Table 1). The safety population had a median age of 35 years (range: 17–77 years) and approximately half (55.2%) were female. The mean duration of epilepsy was between 22.5 and 23.8 years. All but 15 patients (6.0%) were taking ≥3 AEDs at baseline (protocol violators), and the majority (74%) were taking 2 AEDs. The most common concomitant AEDs among patients were CBZ (49–69%), valproic acid (27–35%), phenytoin (12–21%), and levetiracetam (18–25%).
Table 1. Demographic and baseline disease characteristics (safety population)
Placebo (n = 87)
800 mg (n = 85)
1200 mg (n = 80)
n, total number of patients; SD, standard deviation; AED, antiepileptic drug.
aConcomitant AEDs taken by at least 5% of patients in any treatment group at baseline.
Mean ± SD
37.7 ± 12.07
36.8 ± 10.65
36.0 ± 11.43
Ethnic group, n (%)
Mean ± SD
25.7 ± 4.07
25.5 ± 4.72
26.9 ± 4.94
Duration of epilepsy (years)
Mean ± SD
23.8 ± 13.03
22.5 ± 11.78
23.0 ± 13.01
Seizure frequency/4 weeks
Mean ± SD
11.3 ± 18.48
11.6 ± 22.12
11.3 ± 10.27
Seizure type at baseline, n (%)
Number of concomitant AEDs, n (%)
Type of AEDsa, n (%)
In the ITT population, the standardized seizure frequency was significantly lower during the 12-week maintenance period after once-daily treatment with ESL 800 mg (LS Mean 5.7, P = 0.048) and ESL 1200 mg (LS Mean 5.5, P = 0.021) than with placebo (LS Mean 7.3). The reduction was also statistically significant when the analysis was performed for the titration and maintenance periods combined (Table 2).
Table 2. Analysis of covariance (ANCOVA) for seizure frequency per 4 weeks (ITT population)
ESL 800 mg
ESL 1200 mg
(N = 84)
(N = 84)
(N = 77)
N, total number of patients; n, number of patients with data available; LS Mean, least square mean; CI, confidence interval.
ANCOVA model: treatment as factor and ln-transformed baseline seizure frequency as covariate. Model was based on ln-transformed seizure frequencies, and estimates were back transformed using the exponential function. Dunnett’s multiple comparison procedure was used to assess ESL treatment means to placebo mean.
95% CI for mean
LS Mean difference to placebo
2-week titration + 12-week maintenance
95% CI for mean
LS Mean difference to placebo
Responder rate was 22.6% (placebo), 34.5% (800 mg; P = 0.106), and 37.7% (1200 mg; P = 0.020). The overall difference between the two ESL doses and placebo for the degree of seizure reduction was statistically significant for the ESL 1200 mg (P = 0.008) group, but not for the ESL 800 mg (P = 0.059) group. The median relative reduction in standardized seizure frequency was higher in the ESL 800 mg (37.9%) and ESL 1200 mg (41.9%) groups than in the placebo (17.0%) group (Fig. 3); when the 2-week titration and 12-week maintenance periods were combined, the reduction in seizure frequency was highest in the ESL 800 mg group (37.2%) compared to the ESL 1200 mg (34.1%) and placebo (11.9%) groups.
The proportion of seizure-free subjects (100% seizure reduction) was greater in the ESL 800 mg (4.8%) and ESL 1200 mg (3.9%) groups than in the placebo (1.2%) group, although the difference was not statistically significant compared to placebo in either group. This finding is consistent with the proportion of placebo patients who reported an exacerbation in seizure frequency ≥25% (22.6%), which was nearly double that reported in the ESL groups (16.7% in the ESL 800 mg group vs 13.0% in the ESL 1200 mg group).
The median number of days with seizures decreased from 5.8 at baseline to 4.0 during the maintenance period in the ESL 800 mg group, and from 5.6 to 3.0 in the ESL 1200 mg group; the reduction in the placebo group was from 5.7 to 4.9. Results were similar in the PP populations for all efficacy analyses.
In the safety population, the overall incidence of TEAEs generally increased with increasing ESL dose (Table 3), both for all TEAEs (39.1% of placebo, 52.9% of ESL 800 mg, and 61.3% of ESL 1200 mg) as well as for those considered at least possibly related to treatment (26.4% of placebo, 41.2% ESL of 800 mg, and 51.3% of ESL 1200 mg). The dose-dependent increase in TEAE frequency was associated with specific events within only two system organ classes: gastrointestinal (GI) disorders (e.g. nausea, vomiting) and nervous system disorders (e.g. abnormal coordination, dizziness). No other TEAEs exhibited this relationship. The most common TEAEs (≥10% in all groups) were dizziness, somnolence, headache and nausea (Table 3). The majority of TEAEs attributed to ESL were mild or moderate in intensity (81.9%) compared to 85.3% in the placebo group. Most reported TEAEs were in the system organ class of nervous system disorders (n = 94; 37.3%), of which 80 cases (31.8%) were considered at least possibly related to study medication. Rash occurred in only 3 (1.8%) out of the 165 patients treated with ESL.
Table 3. Summary of treatment-emergent adverse events (safety population)
Treatment-emergent adverse events
Placebo (N = 87)
ESL 800 mg (N = 85)
ESL 1200 mg (N = 80)
N, total number of patients; TEAE, treatment-emergent adverse event.
aTreatment-emergent adverse events affecting >3% of patients and leading to discontinuation in >3% of patients in any treatment group.
Incidence: n (%) of subjects
TEAEs leading to discontinuationa
TEAE onset in relation to start treatment (Day)
The incidence of TEAEs was higher during the first 41 days of treatment (Table 3). From day 42 onwards there was a small difference among the three treatment groups.
The incidence of serious TEAEs was minimal across the treatment groups (0% of placebo, 0% of ESL 800 mg, and 1.3% of ESL 1200 mg). The single case in the ESL 1200 mg group (cerebellar syndrome) led to study discontinuation.
Overall, 22 patients discontinued the study as a direct result of a TEAE: nine patients (11.3%) in the ESL 1200 mg group, seven patients (8.2%) in the ESL 800 mg group and six patients (6.9%) in the placebo group. TEAEs leading to discontinuation in >3% of patients included abnormal coordination, dizziness and nausea. No deaths occurred during the study.
There were no apparent trends noticeable in any treatment group for laboratory parameters, vital signs and weight and ECG assessments. Hyponatremia occurred in only one (0.6%) out of the 165 patients treated with ESL in the current study.
The results of this Phase III study demonstrated that once-daily doses of ESL 800 mg and 1200 mg were generally well tolerated and effective as adjunctive therapy for partial epileptic seizures in patients who were refractory to treatment with 1–2 concomitant AEDs.
During the 12-week maintenance period, the seizure frequency was significantly lower, the median number of days with seizures was significantly reduced, and more patients became seizure free in the ESL 800 and 1200 mg groups than in the placebo group. Though the study was not powered to show significant differences in the responder rates, the proportion of responders was significantly higher in the ESL 1200 mg group.
Although both CBZ and ESL are VGSC blockers and are chemically related, they do not share any metabolites. Furthermore, there is evidence of improved efficacy when different VGSC blockers are combined (12). Therefore, it was considered beneficial to evaluate the effect of ESL when given concomitantly with CBZ, which is one of the most widely used AEDs. In this study, ESL significantly reduced the seizure frequency in a population in which approximately 50% of patients were co-treated with CBZ. In the study by Elger et al. (10), the percentage of patients co-administered with CBZ was 56–62%.
Overall, an analysis of TEAEs revealed no special concerns with respect to ESL safety, or any relevant differences in AE incidence among the three treatment groups. A dose-dependent association with TEAEs was evident, but the occurrence of new TEAEs was limited to the first 41 days of treatment in the study. Thus, patients are more likely to develop an AE when they first initiate treatment with ESL; the AEs are unlikely to be serious, and the risk of an event subsides after 6 weeks as patients adjust to the dosage level.
While the overall incidence of TEAEs did increase with increasing dose of ESL, this was due to events that are common to all AEDs, namely dizziness, somnolence, headache and nausea. Furthermore, problems associated with other AEDs, such as the occurrence of rash or hyponatremia, did not emerge as a concern during ESL treatment. There were neither no changes in laboratory parameters (such as in plasma sodium levels or lipid profile) that indicated a safety concern, nor were there any clinically relevant differences in vital signs, body weight or ECGs.
Finally, the majority of currently available AEDs is dosed several times a day and involve complex and long-term titration schemes, which can hinder compliance. Once-daily dosing with ESL offers a low risk of experiencing neurological, GI, psychiatric or skin disorders, while minimizing compliance concerns. Thus, the added efficacy apparent in this study is expected to have a wider application in the difficult-to-treat population of patients with refractory partial epilepsy who are taking concomitant AEDs.
Given the convenience of once-daily dosing and a short and simple titration regimen, the effectiveness with which ESL reduces partial seizures and the good tolerability of the drug, ESL offers a valid alternative to current AEDs for the treatment of partial epileptic seizures in patients refractory to treatment with 1–2 concomitant AEDs. Although both ESL doses were well tolerated, the incidence of at least possibly related TEAEs was higher in the 1200 mg group. Therefore, a once-daily dose of 800 mg ESL appears to offer the optimal initial treatment regimen, with the option of titrating to 1200 mg (based on individual response and tolerability), if increased efficacy is necessary.
This study was sponsored by BIAL (Portela & Ca SA).
Clinical Investigator’s list
Mexico: Marco Antonio Diaz Torres, Leticia Hoyos Gómez, Fiacro Jiménez Ponce, Manuel Ojeda Chavarria, José-Luis Oropeza De-Alba, Sandra Quiñones, Ricardo Ramos, Ricardo Rangel-Guerra, Silvia Cristina Rivera Nava, Jerónimo Rodríguez, and Rosalía Vásquez.
Portugal: Luis Bigotte de Almeida, Rui Chorão, Georgina de Sousa, Manuel Dílio Alves, José Lopes Lima, Esmeralda Lourenço, Isabel Luzeiro, António Martins, Fernando Martins, Henrique Pereira, José Pinto Marques, Ana Paula Santos, and José Vieira Branco.
Spain: Montserrat Asensio, Francisco Astriazarán, Vicente Bertol, Oscar Fernández, Juan Carlos García-Moncó, Antonio Gil-Nagel, Rafael Gonzélez Maldonado, José Angel Mauri, Alberto Molins, Maria Dolores Morales, Ricardo Navarro, José Maria Serratosa, Javier Tejada, Juan Jesús Vilchez, and Francisco Villalobos.