SEARCH

SEARCH BY CITATION

Keywords:

  • Levetiracetam;
  • Partial seizures;
  • Epilepsy;
  • Refractory seizures;
  • Taiwanese patients;
  • Adults;
  • Adjunctive therapy

Abstract

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONCLUSION
  7. REFERENCES

Summary: Purpose: To assess the efficacy and safety of adjunctive levetiracetam (LEV) therapy in controlling partial-onset seizures refractory to other antiepileptic drugs (AEDs) in a multicenter study in Taiwanese adults.

Methods: Ninety-four patients aged 16–60 years with refractory partial seizures were randomized to receive LEV (n = 47) or placebo (47) for 14 weeks and composed the intention-to-treat (ITT) population. After the first 2 weeks, LEV patients had their dosage increased from 500 mg twice daily to 1,000 mg twice daily. A 12-week maintenance phase followed, after which patients switched to long-term, open-label LEV therapy or entered a 4-week phase of medication discontinuation.

Results: All patients from the ITT population, except one LEV-treated patient with missing seizure-count data, were included in the primary efficacy analysis. The least square mean of logarithmically transformed weekly partial-seizure frequency was significantly lower in the LEV than in the placebo group (0.813 vs. 1.085; p = 0.001). LEV reduced log-transformed weekly partial-seizure frequency by 23.8% (95% confidence interval, 10.4–35.2%) relative to placebo. Significantly more LEV than placebo patients (43.5% vs. 10.6%) experienced a response of a ≥50% decrease from baseline in weekly frequency of partial seizures [odds ratio, 6.5 (95% CI, 2.2–19.3); p < 0.001]. Adverse events were reported in 34 (72.3%) of 47 LEV-treated patients and 32 (68.1%) of 47 placebo patients. The three most common adverse events in the LEV and placebo groups were somnolence (40.4% and 14.9%), dizziness (14.9% and 8.5%), and headache (10.6% and 8.5%), respectively. Only four patients (three LEV-treated patients and one placebo patient) were withdrawn from the study because of adverse events.

Conclusions: Adjunctive LEV therapy, ≤1,000 mg twice daily, was significantly more effective than placebo and was generally well tolerated in Taiwanese adults with treatment-resistant partial-onset seizures.

Levetiracetam (LEV) is the S-enantiomer of α-ethyl-2-oxo-1-pyrrolidine acetamide (1). In clinical trials, LEV has proved safe and generally well tolerated after short-term or long-term administration to patients with partial-onset or various generalized seizures (1–3). It has demonstrated efficacy as an “add-on” antiepileptic drug (AED) for partial seizures with or without secondary generalization (4,5) and is now registered in >50 countries for such use in adults with epilepsy. Indeed, late-phase clinical trials of LEV were conducted mainly in the United States and Europe in adults with refractory partial epilepsy (4–6). These trials demonstrated the efficacy of add-on LEV, in terms of responder rates and reduced frequency of partial seizures, and highlighted the generally favorable tolerability profile of the drug. One of the trials also showed that the efficacy of LEV monotherapy (3,000 mg/d) was maintained for 12 weeks after discontinuation of concomitant AEDs in responders (6).

Importantly, LEV has a generally favorable pharmacokinetic profile characterized by linear, dose-proportional pharmacokinetics, high absolute oral bioavailability (close to100%), minimal binding to plasma proteins (<10%), lack of oxidative hepatic metabolism and active renal excretion (66% as unchanged drug), and minimal potential for drug–drug interactions.(1,4–6). As such, LEV is unlikely to have vastly different characteristics in different populations and is therefore likely to be “ethnically insensitive” (7). Thus the current study was conducted to corroborate earlier evidence of favorable efficacy and safety for adjunctive LEV therapy in white adults with treatment-resistant partial-onset seizures. It was performed in Taiwanese patients with a LEV dosage of ≤2,000 mg/d and designed to permit future extrapolation of LEV data with white patients to the Taiwanese population.

METHODS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONCLUSION
  7. REFERENCES

Patients

One hundred fourteen subjects were enrolled in the study beginning in November 2000. The last study visit occurred in February 2002. Ninety-four men and women aged 16–60 years with partial-onset seizures, with or without secondary generalization, were randomized to study treatment. Fifteen subjects were withdrawn before randomization because they were found to be ineligible. Five subjects withdrew their informed consent before randomization.

All randomized patients had been diagnosed as having epilepsy for ≥6 months before the study. Partial seizures were treatment resistant in all cases, and, during an 8-week baseline period, all patients had at least four complex or secondarily generalized partial seizures (type IB or IC). For ≥2 weeks before the study, patients had received a stable dosage of one to three AEDs (including benzodiazepines) other than LEV; all patients had been treated with at least two classic AEDs, either simultaneously or consecutively, before the study. In all cases, computed tomography or magnetic resonance imaging in the 2 years before the trial confirmed the absence of neoplasia or progressive cerebral or neurodegenerative disease. Women of childbearing potential were included in the trial only if they were using appropriate contraception. All patients were required to provide written informed consent before inclusion in the trial. For patients aged 16 to 18 years, consent was obtained from both the patient and the legal guardian.

Patients were excluded if they had status epilepticus in the 3 months before the study, or if they had clusters of seizures that could not be reliably and regularly counted. Other exclusion criteria comprised a history or presence of pseudoseizures; a history of recurrent psychotic or major affective disorder; the presence of clinically significant acute or chronic illness (including cardiac, renal, or hepatic dysfunction); and the use of CNS-influencing medication (other than concomitant AED therapy), unless patients had been stabilized on such medication for ≥1 month before the trial. In addition, pregnant or lactating women were excluded from the study, as were patients who abused alcohol or illicit drugs, patients who participated in another clinical trial in the 3 months before the study, and any patients who had participated in a previous study of LEV. Other exclusion criteria were laboratory test abnormalities, such as elevated liver enzyme levels (>3 times the upper limit of normal), serum creatinine >2.0 mg/dl, and leukopenia (<2,800 white blood cells/μl) or thrombocytopenia (<12,000 platelets/μl); however, an increase in γ-glutamyltransferase level to >3 times the upper limit of normal could be acceptable if the result was due to hepatic enzyme induction caused by concomitant AED therapy.

Study design

The study was conducted in accordance with the new regulation announced by the Department of Health, Taiwan (8), which followed Harmonized Tripartite Guidelines proposed by the International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (7,9). Its results are reported in accordance with the revised CONSORT Statement (10). This was a prospective, double-blind, placebo-controlled, parallel-group study conducted at five centers; two in northern, two in southern, and one in eastern Taiwan. After an 8-week baseline period, patients were randomized to receive add-on LEV therapy or placebo for 14 weeks (Fig. 1). The initial LEV dosage was 500 mg twice daily, which was increased to 1,000 mg twice daily after 2 weeks. The latter dosage was then maintained for 12 weeks, although a “fall-back” option was allowed at the investigators' discretion: that is, the dosage could be reduced to 500 mg twice daily, in patients with poor LEV tolerability, and then increased again to 1,000 mg twice daily before week 10.

image

Figure 1. Study design illustrating time line and components for each segment of the study. bid, twice daily; LEV, levetiracetam; PLA, placebo.

Download figure to PowerPoint

After 14 weeks of treatment, all patients either started a 4-week period of study drug discontinuation (2 weeks at 500 mg twice daily, followed by 2 weeks without the study medication) or switched at the investigators' discretion to long-term, open-label LEV therapy with 1,000 mg twice daily during a 2-week period (Fig. 1). Study blinding was maintained during the discontinuation or switch phase. The study was approved by relevant institutional review boards and local health authorities.

The random treatment-allocation sequence was generated in Belgium by the Clinical Drug Supply Unit of the sponsor on May 22, 2000, by using a randomization program (permuted blocks of size 4) generated by a statistician of the sponsor. Study drug containers were labeled in Belgium by the sponsor in identical fashion for both treatments. Placebo and LEV tablets were identical in appearance. Treatments were assigned sequentially to patients at every site by the investigators. Individual treatment assignments and block size were not revealed to any study participants before the last patient had completed the trial and before unblinding occurred after final determination of statistical analyses during the preanalysis meeting.

Efficacy

Throughout the study, seizures were recorded by the patient or a close relative in a seizure diary issued by the Taiwan Epilepsy Society. The primary efficacy measure was the logarithmically transformed weekly frequency of partial-onset seizures over the 14-week, double-blind evaluation phase (including the uptitration period). This end point, was identical to the primary end point determined for the three regulatory efficacy trials performed with LEV in patients with refractory epilepsy in the United States and Europe. To comply with the assumptions of ANCOVA parametric tests, the log-transformation was chosen to normalize the distribution of the weekly seizure frequency.

Secondary efficacy measures also were assessed over the 14-week treatment period: absolute and percentage reduction from baseline in (untransformed) weekly frequency of partial-onset seizures; weekly frequency of total seizures (partial, primarily generalized, and unclassifiable seizures combined); absolute and percentage reduction from baseline in weekly frequency of total seizures; responder rate (i.e., percentage of patients with a ≥50% decrease from baseline in weekly frequency) for both partial-onset and total seizures; number of seizure-free patients, and number of seizure-free days per 4 weeks; percentage change from baseline in weekly frequency of partial-onset seizures considered in six categories (>25% increase; 25% increase to <25% decrease; 25% decrease to <50% decrease; 50% decrease to <75% decrease; 75% decrease to <100% decrease; 100% decrease).

In addition, weekly frequency of partial-onset seizures, absolute and percentage decrease from baseline in (untransformed) weekly frequency of partial-onset seizures, were calculated as secondary efficacy variables over the last 8 weeks of the maintenance-treatment phase. Investigators and patients also performed a global clinical evaluation of the patient's condition.

Safety

Physical and neurologic examinations, laboratory assessments, and electrocardiography were performed, bodyweight and vital signs were recorded, and detailed records of adverse events were documented, during the baseline and treatment periods and at study completion. In addition, plasma concentrations of LEV and concomitant AEDs were monitored at baseline, and at weeks 6, 10, and 14.

Statistical analyses

A sample size of 47 evaluated patients per group was expected to have a power of 80% to detect 20% superiority for LEV over placebo regarding logarithmically transformed weekly seizure frequency with a one-sided test at 5% (assuming a common standard deviation of 0.43). According to the protocol created for the regulatory submission and the sample-size computation described earlier, all the analyses submitted to the Taiwanese authorities were performed by using one-sided tests at 5% and 90% one-sided confidence intervals. For the specific purpose of this article, all the original analyses were redone by using 5% two-sided tests and 95% two-sided confidence intervals. The two-sided version of all analyses will be the only one presented in this article, but it has to be kept in mind that the power is thus lower than originally planned. An analysis of covariance (ANCOVA) model was used to assess the superiority of LEV over placebo with respect to the primary efficacy variable. The analysis was based on the intention-to-treat (ITT) population and applied to logarithmically transformed seizure data [loge(x+ 1)]. The ANCOVA model included the treatment group as factor, and baseline seizure frequency as covariate. The difference in treatment least square means with two-sided 95% confidence intervals (CIs) was determined and expressed as a percentage decrease versus placebo.

Noninferential statistics were applied to all secondary efficacy variables. Various two-sided tests were performed to assess LEV superiority over placebo regarding absolute and percentage decreases from baseline (Wilcoxon Rank Sum test), responder rate (logistic regression model), treatment response considered in six categories of seizure-frequency change (Cochran-Mantel-Haenszel test), and proportion of seizure-free patients (Fisher's exact test). All demographics and baseline characteristics were described, and inferential tests, comparing LEV with placebo, also were carried out: Fisher's exact test for binary variables, Cochran-Mantel-Haenszel for categorical variables, two-sample t test with hypothesis of equal variance, or Wilcoxon rank sum test with normal approximation for continuous variables. Finally, when presented for adverse event (AE) incidences, the Fisher's exact test comparing LEV with placebo was conducted two-sided.

RESULTS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONCLUSION
  7. REFERENCES

One hundred fourteen patients were screened, and 94 eligible patients were randomized to LEV (n = 47) or placebo (n = 47) (Fig. 2). Of the 94 patients randomized, 90 completed the study (44 LEV-treated patients; 46 in the placebo group). Three (6.4%) LEV-treated patients were withdrawn from the trial because of AEs, as was one (2.1%) placebo patient. Of the 90 patients completing the trial, 43 (97.7%) of 44 LEV-treated patients and 45 (97.8%) of 46 placebo patients switched to open-label LEV therapy. During the 2-week transition from placebo to open-label LEV, one patient was withdrawn from the study because of an adverse event; one LEV-treated patient underwent 4 weeks' downtitration to stop study medication (Fig. 2).

image

Figure 2. Number of patients entering and completing various stages of the study. LEV, levetiracetam; PLA, placebo.

Download figure to PowerPoint

At baseline, the two groups comprising the ITT population had similar demographic characteristics (Table 1) without any statistically significant difference. However, more female subjects were included in the LEV group than in the placebo group (63.8% vs. 46.8%), so that the male/female sex distribution was less even in the LEV group than in the placebo group. In addition, fewer LEV than placebo patients had a history of withdrawal seizures, status epilepticus, and documentation of spikes or spike–waves in the EEG, and fewer had an unknown cause of epilepsy (27.7% vs. 42.6%). Throughout the study, the two groups were generally comparable regarding their use of concomitant AEDs. Two to three concomitant AEDs were taken by 85.1% of LEV-treated patients and 72.3% of placebo patients, whereas the corresponding values for one concomitant AED were 14.9% and 23.4%. Mean rates of compliance with study medication (as assessed by counting of study drug returns) were high: 94.9% and 96.3% of study doses were taken by LEV and placebo patients, respectively.

Table 1. Summary of baseline demographic characteristicsa for the intention-to-treat population of 94 randomized patients
Demographic characteristicLEV (n = 47)PLA (n = 47)p Value (LEV vs. PLA)
  1. AED, antiepileptic drug; LEV, levetiracetam; PLA, placebo.

  2. a Mean (± standard deviation) unless otherwise indicated.

  3. b Recorded on a 10-cm visual analog scale.

  4. c Two-sample t test with hypothesis of equal variances.

  5. d Fisher's exact test.

  6. e Wilcoxon rank sum test with normal approximation.

  7. f Cochran-Mantel-Haenszel (general association) test.

Age (yr)32.8 (±10.5)31.7 (±8.2) 0.564 c
Sex
 Number of males (%)  17 (36.2%)25 (53.2%)0.146 d
 Number of females (%)  30 (63.8%)22 (46.8%) 
Race
 Asian/Pacific Islander [no. of patients (%)] 47 (100%)47 (100%)  
Body weight (kg)63.3 (±14.1)64.7 (±12.6)  0.596 c
Height (cm)160.7 (±8.5)  164.1 (±7.9)    0.051 c
Body mass index (kg/m2)24.4 (±5.0) 24.0 (±4.0)   0.620 c
Epilepsy history
 Duration of illness (yr)18.6 (±8.5) 18.7 (±10.7)  0.968 c
 Age at onset (yr)14.3 (±8.5) 13.1 (±8.7)    0.499 c
 Cause unknown [number (%) of patients]  13 (27.7%)20 (42.6%) 0.194 d
 Withdrawal seizures [number (%) of patients]  4 (8.5%)7 (14.9%)0.523 d
 Status epilepticus [number (%) of patients]   6 (12.8%)10 (21.3%) 0.411 d
 Documentation of spikes or  41 (87.2%)45 (95.7%) 0.267 d
 Spike–waves on EEG [number (%) of patients]  41 (87.2%) 
Baseline seizure frequency per week
 Partial seizures
   Mean (SD) 4.0 (±14.1)4.0 (±5.6)   
   Median (interquartile range)   1.6 (1.2–2.5)2.0 (1.1–3.9)0.378 e
  Total seizures
   Mean (SD) 4.0 (±14.1)4.3 (±7.0)   
   Median (interquartile range)   1.6 (1.2–2.5)2.0 (1.1–3.9)0.378 e
Number of concomitant AEDs taken by patients (overall study) [number (%) of patients]
 One  7 (14.9%)11 (23.4%) 0.314 f
 Two  19 (40.4%)18 (38.3%)  
 Three  21 (44.7%)16 (34.0%)  
 Four or more0 (0.0)2 (4.3%)  
Concomitant AED therapy taken overall study by ≥10% of patients in either treatment group (% of patients)
 Carbamazepine  31 (66.0%)27 (57.4%) 0.525 d
 Lamotrigine  18 (38.3%)20 (42.6%) 0.834 d
 Vigabatrin  15 (31.9%)15 (31.9%) 1.000 d
 Valproic acid (incl. slow-release forms)  11 (23.4%)8 (17.0%)0.608 d
 Topiramate   8 (17.0%)3 (6.4%) 0.198 d
 Phenobarbital   6 (12.8%)10 (21.3%) 0.411 d
 Phenytoin   6 (12.8%)11 (23.4%) 0.284 d
 Clonazepam   5 (10.6%)5 (10.6%)1.000 d
Global clinical evaluationb
 Investigator6.7 (±2.1)7.1 (±1.6)  0.304 c
 Patient6.6 (±1.8)6.9 (±2.0)  0.455 c

Efficacy

Primary efficacy variable

One LEV-treated patient from the ITT population had missing data for weekly frequency of partial seizures and was excluded from the primary efficacy analysis. LEV versus placebo was associated with statistically significantly lower least square mean of logarithmically transformed weekly frequency of partial seizures (0.813 vs. 1.085; p = 0.001). The percentage reduction in log-transformed weekly partial-seizure frequency for LEV relative to placebo was estimated at 23.8% (95% CI, 10.4–35.2%). Median weekly frequency of partial seizures in evaluated patients in the LEV group decreased from 1.6 [interquartile range (IQR), 1.2 – 2.5] during the baseline period to 1.0 (IQR, 0.6–1.7) during the 14-week evaluation period; corresponding values in the placebo group were 2.0 (1.1–3.9) and 1.5 (1.0–3.5) (Fig. 3). Mean weekly frequency of partial seizures in evaluated patients in the LEV group decreased from 4.0 ± 14.1(mean ± standard deviation) at baseline to 1.7 ± 2.0 at week 14; corresponding values in the placebo group were 4.0 ± 5.6 and 2.8 ± 3.7. Over the entire evaluation period, the mean weekly frequency of partial seizures was 1.5 ± 1.5 in the LEV group and 3.2 ± 3.9 in the placebo group.

image

Figure 3. Median weekly frequency of partial seizures in evaluable patients at five time points during the 14-week, double-blind, evaluation phase. LEV, levetiracetam; PLA, placebo.

Download figure to PowerPoint

Secondary efficacy variables

In the LEV group, and over the entire evaluation period, results regarding total seizure frequency were almost identical to those for partial-seizure frequency, thus indicating that nearly all seizures were of the partial-onset type. In the placebo group, however, a slightly greater weekly frequency of total rather than partial seizures was noted during both the baseline and evaluation periods. This difference is due to a single patient in the placebo group, who had atonic phenomena associated with temporal lobe epilepsy, which were coded as atonic seizures, a subcategory of primary generalized seizures. The “total seizure” category did not include any reports of primary generalized tonic–clonic seizures, myoclonic seizures, or absence seizures. When including the atonic phenomena of this patient in the analysis, the least square mean of logarithmically transformed weekly total seizure frequency was still statistically significantly lower in the LEV than in the placebo group (0.819 vs. 1.095; p = 0.001). The percentage reduction in log-transformed weekly total seizure frequency for LEV relative to placebo was 24.1% (95% CI, 10.6–35.6%). Other secondary efficacy results are documented in Table 2. As can be seen, LEV versus placebo was associated with a statistically significantly (p = 0.01) greater median percentage decrease from baseline in weekly frequency over the evaluation phase of both partial and total seizures. Significantly more LEV than placebo patients (43.5% vs. 10.6%) experienced a response of a ≥50% decrease from baseline in weekly frequency of partial seizures [odds ratio, 6.5 (95% CI, 2.2–19.3); p < 0.001]. Moreover, statistical analysis of six categories of ranked response regarding weekly partial seizure frequency revealed a significantly (p = 0.008) more favorable effect for LEV than for placebo: indeed, 28.3% of LEV-treated patients compared with 2.1% of placebo patients had a decrease in weekly partial-seizure frequency of 75–100% (Table 2).

Table 2. Summary of results for primary and secondary efficacy variables in the intention-to-treat population
VariableLEV (n = 47)aPLA (n = 47)aStatistical testp value
  1. LEV, levetiracetam; NA, not applicable; PLA, placebo.

  2. aValues in square brackets indicate numbers of evaluable patients.

  3. bAbsolute decrease from baseline over the evaluation period; median and interquartile range.

  4. cPercentage decrease from baseline over the evaluation period; median and interquartile range.

  5. dProportion of patients with a ≥50% decrease from baseline in weekly seizure frequency.

  6. eMean (± standard deviation).

Partial seizures:
Primary efficacy variable
 Least square mean0.8131.085ANCOVA0.001
 % reduction over placebo23.8% (95% CI: 10.4% to 35.2%) 
 Secondary efficacy variables
 Weekly seizure frequency (absolute decrease from baseline)b0.6 (−0.1 to 1.4)0.3 (−0.2 to 0.7)Wilcoxon Rank Sum0.129
 Weekly seizure frequency (percentage decrease from baseline)c 45.4 (−13.1 to 76.9)15.6 (−5.7 to 41.4)Wilcoxon Rank Sum0.010
 Responder rated20/46 (43.5%)5/47 (10.6%)Logistic regression< 0.001  
 Number (%) of patients free of seizures4 (8.5%); [47]0 (0.0%); [47]Fisher's exact test0.117
 Number of seizure-free days per 4 wke24.2 (±3.3); [46]  21.4 (±6.3); [46]    
 Number (%) of patients in six ranked categories of % change from baseline in weekly seizure frequency:
   >25% increase 7 (15.2%); [46] 8 (17.0%); [47] 
   25% increase to <25% decrease 9 (19.6%); [46]19 (40.4%); [47] 
   25% decrease to <50% decrease10 (21.7%); [46]15 (31.9%); [47] 
   50% decrease to <75% decrease 7 (15.2%); [46]4 (8.5%); [47] 
   75% decrease to <100% decrease 9 (19.6%); [46]1 (2.1%); [47] 
   100% decrease4 (8.7%); [46]0 (0.0%); [47]Cochran-Mantel-Haenszel0.008 (for the six ranks)
Total seizures
 Least square mean0.8191.095ANCOVA0.001
 % reduction over placebo24.1% (95% CI, 10.6%–35.6%) 
 Weekly seizure frequency (absolute decrease from baseline)b0.6 (−0.1–1.4)  0.3 (−0.2–0.7)Wilcoxon Rank Sum0.129
 Weekly seizure frequency (percent decrease from baseline)c45.4 (−13.1–76.9) 15.6 (−5.7–39.1)Wilcoxon Rank Sum0.009
 Number (%) of patients free of seizures4 (8.5%); [47] 0 (0.0); [47]Fisher's exact test0.117
 Number of seizure-free days per 4 wke24.2 (±3.3); [46]   21.3 (±6.4); [46] 

Secondary efficacy results for the last 8 weeks of the maintenance phase were similar to the results for the entire 14-week period. For instance, ANCOVA analysis of log-transformed values for weekly frequency of partial seizures during the last 8 weeks revealed a reduction of 18.1% for LEV versus placebo (95% CI, 0.6–32.6%; p = 0.044). In addition, the median percentage decrease from baseline in weekly partial-seizure frequency over the last 8 weeks of the maintenance phase was 47.8% in the LEV group and 31.5% in the placebo group (p = 0.085).

Safety

Four patients (three LEV and one placebo patient) were withdrawn from the study because of AEs. In two LEV patients, in whom a seizure and dizziness were indicated as the primary reasons for discontinuation of LEV, these AEs occurred 4 and 10 days, respectively, after the patients had actually stopped taking LEV. A mouth ulcer was the primary reasons for discontinuation of LEV in the third LEV patient, and a skin eruption was the principal reason for discontinuation of placebo. Of the 90 patients who completed the 14-week evaluation period, 89 attempted a switch to open-label LEV therapy. One patient converting from placebo to LEV during the “switch” phase was withdrawn from the study because of an AE (dizziness), such that 88 patients successfully switched to open-label LEV (Fig. 2).

During the evaluation period, 34 (72.3%) of 47 LEV-treated patients and 32 (68.1%) of 47 placebo patients reported at least one AE; a total of 170 events was documented (90 in the LEV group and 80 in the placebo group). Overall, 25 (53/2%) LEV-treated patients and 12 (25.5%) placebo patients had at least one AE considered at least possibly related to study medication. Eleven LEV-treated patients and three placebo patients had AEs requiring dosage reduction. In the LEV group, these events were most commonly somnolence (n = 6; 12.8%), dizziness (three; 6.4%), and diplopia (two; 4.3%). During uptitration, two LEV-treated patients had downward dosage adjustments because of AEs [agitation and bad temper (n = 1), drowsiness (n = 1), and one because of noncompliance (n = 1)].

During the evaluation period, five patients (two LEV and three placebo patients) reported a total of seven serious AEs (SAEs). All SAEs were considered serious because they led to hospitalization of the patient. Three SAEs were severe in intensity (common cold, fever, and frequent generalized tonic–clonic seizures), all reported by one placebo patient, and all three were considered unrelated or unlikely to be related to study medication. The other SAEs were complex partial seizure status and burn injuries during a seizure in two patients taking placebo and secondarily generalized partial seizures and multiple fractures in two patients taking LEV. An additional patient reported three serious AEs: convulsion, fall, and subcutaneous hematoma, 3–4 days after stopping LEV. These events were severe in intensity and considered possibly related to study medication.

AEs noted with an incidence of ≥5% during the evaluation period in either study group are listed in Table 3. As can be seen, somnolence, dizziness (excluding vertigo), and headache were the three most frequent events in the LEV group. They were also the only AEs that occurred with an incidence of ≥5% in LEV patients when only those AEs that were considered by the investigator to be at least possibly related to study medication are seen. Of these causally related AEs, somnolence was noted in 40.4% of LEV-treated patients and 14.9% of placebo patients; the difference in incidence of somnolence between LEV and placebo was statistically significant (two-sided p < 0.05; Fisher's exact test). Corresponding incidences were 10.6% and 4.3% for dizziness, and 6.4% and 2.1% for headache, and these were not statistically significantly different. Overall, more than three fourths of AEs in the LEV group were mild in intensity. No findings of major clinical significance were noted regarding the effects of study medication on laboratory test results, vital signs, electrocardiographic recordings, physical and neurologic examinations, and psychiatric and mental status.

Table 3. Number (%) of patients with adverse events observed with an incidence of ≥5% during the evaluation period (data for the intention-to-treat population)
Adverse eventAll adverse eventsAdverse events considered at least possibly related to study medication by the investigator
LEV (n = 47)PLA (n = 47)LEV (n = 47)PLA (n = 47)
  1. ECG, electrocardiogram; LEV, levetiracetam; PLA, placebo.

  2. ap < 0.05, Fisher's exact test.

Somnolence19a (40.4%)  7 (14.9%)19a (40.4%)   7 (14.9%)
Dizziness (excluding vertigo)7 (14.9%)4 (8.5%)   5 (10.6%)  2 (4.3%)
Headache5 (10.6%)4 (8.5%) 3 (6.4%) 1 (2.1%)
ECG abnormalities4 (8.5%) 2 (4.3%)  
Diplopia3 (6.4%) 0        
Pruritus3 (6.4%) 0        
Nasopharyngitis2 (4.3%)  6 (12.8%) 
Dissociation (psychiatric)0       3 (6.4%)  
Memory impairment0       3 (6.4%)  

DISCUSSION

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONCLUSION
  7. REFERENCES

The principal goal of this study was to determine the efficacy and safety of LEV, ≤2,000 mg/d, in controlling seizures in Taiwanese adults with refractory partial-onset seizures. The dosage of 2,000 mg/d was selected based on data from previous, similarly designed United States and European studies in white patients (4–6), which confirmed the favorable efficacy and tolerability of LEV 1,000–3,000 mg/d, and on the premise that Taiwanese patients in the current trial would weigh ∼25% less than white patients in the previous studies. However, mean body weight for the ITT population in the current study (∼63–65 kg) was only ∼10% lower than that for study populations in earlier United States and European trials.

Furthermore, baseline median weekly partial-seizure frequency was 1.6 and 2.0, in LEV and placebo treatment groups, respectively, in the present trial, but ∼1.9 and 2.3 in LEV and placebo treatment groups, respectively, in earlier studies (4–6). Another difference was in concomitant AED therapy: up to three concomitant AEDs were permitted in the current trial and were taken by ∼40% of patients, whereas up to two concomitant AEDs were allowed in the earlier studies. Although the rate of concurrent carbamazepine (CBZ) use in the present and earlier trials was similar (∼62–68% of patients), Taiwanese patients in the current trial versus white patients in the earlier studies tended to have greater use of novel AEDs such as lamotrigine and vigabatrin (≈40% vs. 7–8% of patients).

The principal efficacy result of the current study in Taiwanese patients, a 23.8% reduction in log-transformed weekly partial-seizure frequency for LEV relative to placebo, was similar to results from earlier studies with LEV, 1,000–3,000 mg/d, in white patients (1,4–6). In particular, this principal result agreed with a 17.7% decrease (98% CI, 4.1–29.4%; p = 0.003) reported for LEV, 2,000 mg/d, over placebo in an earlier European study (5).

Secondary efficacy end points for which preplanned or post hoc exploratory analyses were performed showed a consistent trend favoring LEV over placebo, with the only notable exception being absolute decrease in weekly partial-seizure frequency over the last 8 weeks of the maintenance period, which showed no between-group difference.

The most commonly used measures of efficacy across studies of AEDs, such as the percentage of patients attaining a decrease of ≥50% from baseline in partial-seizure frequency (i.e., responder rate) and median percentage reduction in seizure frequency from baseline showed results that were clearly in favor of LEV. The median seizure frequency with LEV remained stable throughout the treatment period after a drastically initial decrease in seizure frequency. Rapid onset of action of LEV has been reported (11,12). The tendency toward convergence at the last visit (Fig. 3) could be a chance finding of the small sample size in this study and could also be a consequence of spontaneous fluctuation in seizure frequency (13). The possibility of developing tolerance was found to be unlikely in an analysis of LEV as long-term add-on epilepsy therapy (14). On the contrary, sustained short-term efficacy of LEV has been reported (12), and the long-term efficacy for over 1 year and longer has also been confirmed (14).

In the present study, the 50% responder rate was >4 times greater in LEV than in placebo patients (43.5% vs. 10.6%), and the likelihood of such a response was 6.5 times greater in the LEV than in the placebo group (p < 0.001). This finding compares favorably with responder rates documented in earlier placebo-controlled studies of LEV: ∼30–40% with LEV, 1,000–3,000 mg/d, and ∼10% with placebo (1,4–6).

Moreover, in the current trial, 8.5% of LEV-treated patients became seizure free from the first day of treatment, compared with no patients in the placebo group (two-sided p = 0.117). These findings also are consistent with those of earlier studies, in which ∼3–6% of white patients treated with LEV, 1,000–3,000 mg/d, became seizure free (4–6).

Overall, the favorable findings for LEV in Taiwanese patients in the current trial are also likely to manifest in other Asian populations. LEV does not undergo oxidative hepatic metabolism and has no known potential for drug–drug interactions, in particular with other AEDs, such as phenytoin, CBZ, and valproic acid (15). The pharmacokinetic characteristics of LEV make it highly likely that its metabolism is insensitive to intrinsic ethnic factors such as genetic polymorphism of drug-metabolizing enzymes (e.g. of the cytochrome P450 isoenzyme CYP2C19, which is known to be present in substantial proportions of Asian subpopulations, notably in Chinese, Koreans, and Japanese) (16–18). The only major enzymatic pathway assumed to be relevant for the limited metabolism of LEV is through an amidase believed to be widely distributed throughout the human body (15). It is currently unknown whether this enzyme is subject to practically relevant genetic polymorphism. Extrinsic ethnic factors (e.g., climate, sunlight, pollution, culture, medical practice) may somewhat differ between various Asian subpopulations (e.g., Chinese, Japanese, Korean, Malay, Indian), but could not reasonably be expected to contribute substantially to differential efficacy or tolerability of AEDs.

Regarding tolerability in the present trial, slightly more than one half of LEV-treated patients and one fourth of placebo patients had at least one AE considered at least possibly related to study medication. However, only four patients were withdrawn from the study during the evaluation period because of AEs (three LEV-treated patients and one placebo patient), and only two of these withdrawals (one LEV, one placebo) occurred while the patients were taking study medication. Importantly, the overall tolerability profile of LEV was similar to that documented in the three earlier placebo-controlled studies in white patients (4–6). Somnolence, dizziness, and headache were the three most frequent events in the current study, compared with somnolence, asthenia, infection, and dizziness in the earlier trials. The incidence of somnolence in LEV-treated patients was greater in the current study than in previous trials (40.4% vs. 12.0%), as was the incidence in placebo patients (14.9% vs. 9%). The overall higher incidence of somnolence in both treatment groups may have been due to the greater mean number of concomitant AEDs taken by the Taiwanese patients.

No patients in the current trial discontinued LEV because of somnolence, and only six LEV-treated patients had their dosage reduced because of this AE. In addition, in >80% of somnolence reports in LEV-treated patients, the event was mild in intensity; and none was severe. Duration of side effects was not assessed in the present study. However, previous studies do suggest that somnolence occurs predominantly during the initial phase of treatment with LEV (2). As in the current study a different dictionary for coding was used; it is possible that the AEs currently coded as somnolence included symptoms coded in earlier United States and European trials as “asthenia,” based on the COSTART dictionary (Coding Symbols for Thesaurus of Adverse Reaction Terms) item used mainly to describe tiredness and fatigue and reported in three earlier studies in 7% of LEV-treated patients and 3% of placebo patients (4–6).

Because the side-effect profile of LEV was already known when this study started, a higher awareness of the principal side effect could in itself have increased the sensitivity to notice and report such an effect. Nonetheless, the possibility remains that Taiwanese patients may be somewhat more sensitive to LEV-induced somnolence than are white patients. However, because of the generally mild intensity of somnolence reported in the Taiwanese patients, this difference with the white populations studied previously may not be of clinical relevance.

LEV as add-on therapy has a favorable tolerability profile in patients with refractory partial-onset seizures, as it has been shown by the results of several placebo-controlled studies (1,4–6) and from some studies of large-scale meta-analysis of eight newer AEDs introduced over the past 10 years (3,19–21). LEV was associated with one of the lowest relative probabilities of treatment discontinuation, combined with one of the highest relative probabilities of a ≥50% response (3,21). In another review of placebo-controlled clinical trials, LEV was found to have the lowest Summary Complaint Score of seven different newer AEDs (20).

CONCLUSION

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONCLUSION
  7. REFERENCES

This double-blind, placebo-controlled study demonstrates that LEV, ≤2,000 mg/d, is an effective and safe adjunctive therapy for Taiwanese patients with treatment-resistant partial-onset seizures. Results from the current study in Taiwanese patients confirm favorable findings with LEV, 1,000–3,000 mg/d, in previous placebo-controlled studies in white patients, and they suggest that the effects of LEV are likely to be ethnically insensitive.

REFERENCES

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONCLUSION
  7. REFERENCES
  • 1
    Shorvon SD, van Rijckevorsel K. A new antiepileptic drug. J Neurol Neurosurg Psychiatry 2002;72: 426428.
  • 2
    French J, Edrich P, Cramer JA. A systematic review of the safety profile of levetiracetam: a new antiepileptic drug. Epilepsy Res 2001;47: 7790.
  • 3
    Marson AG, Hutton JL, Leach JP, et al. Levetiracetam, oxcarbazepine, remacemide and zonisamide for drug resistant localization-related epilepsy: a systematic review. Epilepsy Res 2001;46: 259270.
  • 4
    Cereghino JJ, Biton V, Abou-Khalil B, et al. and the United States Levetiracetam Study Group. Levetiracetam for partial seizures: results of a double-blind, randomized clinical trial. Neurology 2000;55: 236242.
  • 5
    Shorvon SD, Löwenthal A, Janz D, et al. and the European Levetiracetam Study Group. Multicenter double-blind, randomized, placebo-controlled trial of levetiracetam as add-on therapy in patients with refractory partial seizures. Epilepsia 2000;41: 11791186.
  • 6
    Ben-Menachem E, Falter U, and the European Levetiracetam Study Group. Efficacy and tolerability of levetiracetam 3000 mg/d in patients with refractory partial seizures: a multicenter, double-blind, responder-selected study evaluating monotherapy. Epilepsia 2000;41: 12761283.
  • 7
    International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. ICH Harmonised Tripartite Guideline: ethnic factors in the acceptability of foreign clinical data E5. Recommended for Adoption at Step 4 of the ICH Process on 5 February 1998 by the ICH Steering Committee (including Post Step 4 corrections agreed by the Steering Committee on 11 March 1998). Accessible at http://www.ich.org.
  • 8
    Announcement of Bridging Study, Department of Health, Taiwan. Accessible at http://www.cde.org.tw/BS.web.cht/main.htm
  • 9
    International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. ICH Harmonised Tripartite Guideline: Guideline for Good Clinical Practice E6. Recommended for Adoption at Step 4 of the ICH Process on 1 May 1996 by the ICH Steering Committee (including Post Step 4 corrections agreed by the Steering Committee on 10 June 1996). Accessible at http://www.ich.org.
  • 10
    Altman DG, Schulz KF, Moher D, et al. for the CONSORT Group. The Revised CONSORT Statement for Reporting Randomized Trials: explanation and elaboration. Ann Intern Med 2001;134: 663694.
  • 11
    French J, Arrigo C. Rapid onset of action of levetiracetam in refractory epilepsy patients. Epilepsia 2005;46: 324326.
  • 12
    French J, di Nicola S, Arrigo C. Fast and sustained efficacy of levetiracetam during titration and the first 3 months of treatment in refractory epilepsy. Epilepsia 2005 (in press).
  • 13
    Somerville R. Aggravation of partial seizures by antiepileptic drugs: is there evidence from clinical trials? Neurology 2002;59: 7983.
  • 14
    Ben-Menachem E, Edrich P, van Vleymen B, et al. Evidence for sustained efficacy of levetiracetam as add-on epilepsy therapy. Epilepsy Res 2003;53: 5764.
  • 15
    Nicolas J-M, Collart P, Gerin B, et al. In vitro evaluation of potential drug interactions with levetiracetam, a new antiepileptic agent. Drug Metab Dispos 1999;27: 250254.
  • 16
    Weinshilboum R. Inheritance and drug response. N Engl J Med 2003;348: 529537.
  • 17
    Goldstein JA. Clinical relevance of genetic polymorphisms in the human CYP2C subfamily. Br J Clin Pharmacol 2001;52: 349355.
  • 18
    Spear BB. Pharmacogenetics and antiepileptic drugs. Epilepsia 2001;42(suppl 5):3134.
  • 19
    Marson AG, Kadir ZA, Hutton JL, et al. The new antiepileptic drugs: a systematic review of their efficacy and tolerability. Epilepsia 1997;38: 233234.
  • 20
    Cramer JA, Ben Menachem E, French J. Review of treatment options for refractory epilepsy: new medications and vagal nerve stimulation. Epilepsy Res 2001;47: 1725.
  • 21
    Eng CO, Arrigo C, van Rijckevorsel K, et al. Meta-analysis and indirect comparison of levetiracetam with second-generation antiepileptic drugs in partial epilepsy. Clin Neuropharmacol 2005;28: 7278.