Clobazam for the Treatment of Intractable Epilepsy in Children

Methods

We retrospectively reviewed medical charts of patients less than 18 years of age treated by the Cook Children’s Comprehensive Epilepsy Program with clobazam for intractable epilepsy since January 2007. All patients were treated with clobazam obtained from a Canadian pharmacy under the supervision of a local physician, as the medication was not yet approved by the Food and Drug Administration (FDA) in the United States. Patients with at least 2 months of follow-up data after initiation of clobazam were included. The study protocol was reviewed and approved by the institutional review board.

Data collected for each patient included baseline demographics of age at drug initiation, age at epilepsy diagnosis, gender, seizure type(s), epilepsy type, and etiology (if known). Treatment characteristics including prior and concomitant antiepileptic drug exposure, along with prior surgical history (including vagal nerve stimulator), were included. Clobazam dosing (mg/kg/d) was recorded at initiation, and at each subsequent patient contact, as were seizure types and response of each seizure type to therapy. Data were collected up until September 2011 or until clobazam discontinuation. Seizure frequency was determined from clinic notes based on parental report at each visit, and routinely recorded as the average number of seizures per unit time (i.e., day/month/year). For the purpose of data analysis, the frequency of seizures was converted to a “per year” standard for each patient. Outcome was reported for each seizure type as <50%, 50% to <90%, or >90% reduction at 6 months, 12 months, and last visit >12 months from clobazam initiation, respectively, with responders defined as ≥50% seizure reduction. Seizure response was recorded for all patients at each time frame, with nonresponse reported for patients subsequent to discontinuation of clobazam. Many patients had multiple seizure types, and in rare cases in which response of individual seizure type was not explicitly reported in clinical notes, the overall reduction in seizure frequency was applied to all. For patients that were initiated on clobazam and expected to have been exposed for 6, 12, and >12 months’ follow-up but failed to return to clinic, data were excluded at the time point under analysis. As this was a retrospective analysis, concomitant antiepileptic drug doses were not required to be stable during follow-up, but changes to concomitant antiepileptic drugs were recorded at each visit as increased, decreased, no change, or discontinued. Changes to vagal nerve stimulator settings were not recorded.

Each seizure type was regarded as an individual clobazam “exposure” for data analysis. Outcome was dichotomized as <50% or ≥50% reduction in seizure frequency and compared against age at diagnosis, duration of epilepsy, age at drug initiation, epilepsy type, and etiology. We compared response of seizures dichotomized as drop (atonic, myoclonic, tonic) and nondrop (all other seizure types), in addition to analysis of response of individual seizure types. A secondary analysis was completed comparing exposures with >90% seizure reduction to all other exposures using identical variables. Statistical analysis was performed using chi-square for discrete variables and ANOVA for continuous variables. Levine’s test was used to assess equality of variances. Data were analyzed using the SPSS 19 statistical package with the level of significance set at P < .05.

To analyze dose response of clobazam, we compared patients at each follow-up time point stratified as low dose (<0.5 mg/kg/d), medium dose (0.5-1 mg/kg/d), or high dose (>1 mg/kg/d) versus seizure reduction. To evaluate the emergence of tolerance on clobazam, we performed Kaplan-Meier survival analysis on all exposures that reported a response to treatment after drug initiation. The survival curve was generated with response defined as time from first reported response after clobazam initiation to event, which included reported decrease to <50% seizure reduction, adjustment in concomitant antiepileptic drugs, increase in clobazam dose, or date of surgery.

Adverse effects occurring during treatment with clobazam were recorded for each patient regardless of whether they were clearly associated with the treatment, as was the date of treatment discontinuation and the reason for medication withdrawal. Adverse effects were compared against clobazam dose and number of concomitant antiepileptic drugs at the time the side effect was reported.

Results

Population Characteristics

One hundred eighteen patients were prescribed clobazam during the period under study. Two patients did not start the medication and 8 had not followed up at the time of data analysis, leaving 108 patients (56 male) in the final cohort. Baseline characteristics of the cohort are presented in Table 1. Patients experienced a variety of seizure types (Table 2) and as each seizure type was considered a separate clobazam exposure, 215 individual exposures were analyzed. An etiology or syndromic diagnosis was available for 86 (80%) patients (Table 3). The most common concomitant antiepileptic drugs at the time of clobazam initiation were valproic acid (n = 37, 12.37%), levetiracetam (n = 28, 9.36%), zonisamide (n = 27, 9%), clonazepam (n = 26, 8.7%), and lamotrigine (n = 26, 8.7%), though a variety of other antiepileptic drugs were represented. Forty-eight (44%) patients were actively treated with vagal nerve stimulator and 17 had previously undergone other epilepsy surgeries: 8 corticectomy, 8 lobectomy, and 1 corpus callosotomy alone.

OPEN IN VIEWER

Table 1.

Baseline Patient Characteristics

Characteristics
Mean age (y) at clobazam initiation (SD, range)	8.66 (4.84, 0.62-17.92)
Mean duration of epilepsy (y) prior to initiation (SD, range)	6.39 (4.18, 0.14-15.85)
Mean seizure types per patient (SD, range)	1.99 (0.89, 0-4)
Mean seizure frequency (SD, range)	1657 (4773 , 10-36500)
Epilepsy type (n, %)
Symptomatic generalized epilepsy	47 (43.5)
Idiopathic generalized epilepsy	10 (9)
Epileptic encephalopathy	9 (8)
Symptomatic localization-related epilepsy	36 (33.3)
Idiopathic localization-related epilepsy	1 (1)
Cryptogenic localization-related epilepsy	4 (3.7)
Undefined	1 (1)
Mean concomitant AEDs (SD, range)	2.93 (1.29, 0-7)
Mean prior AED exposures (SD, range)	4.75 (3.06, 0-16)
Mean duration (mo) of clobazam exposure (SD, range)	16 (11.7, 0.66-55.43)

Abbreviations: AED, antiepileptic drug; SD, standard deviation.

OPEN IN VIEWER

Table 2.

Seizure Types Experienced by Patients (N = 215) on Clobazam

Seizure type	n (%)
Complex partial	45 (21%)
Myoclonic	42 (19.5%)
Generalized tonic-clonic	37 (17%)
Atonic	28 (13%)
Tonic	26 (12%)
Secondarily generalized tonic-clonic	14 (6.5%)
Absence	14 (6.5%)
Infantile spasms	9 (4%)

OPEN IN VIEWER

Table 3.

Etiologies of Epilepsy in Patients Treated With Clobazam

Etiology (n = 86, 80%)	n (%)
Cortical malformation	20 (18.5)
Lennox-Gastaut syndrome (diagnostic criteria)	16 (14.8)
Focal encephalomalacia	11 (10.2)
Doose syndrome	8 (7.4)
Dravet syndrome	8 (7.4)
Chromosomal mutation	8 (7.4)
Tuberous sclerosis	3 (2.8)
Mitochondrial disease	2 (1.9)
Jeavon syndrome	1 (0.9)
Juvenile absence epilepsy	1 (0.9)
Continuous spike wave syndrome	1 (0.9)
Mesial temporal sclerosis	1 (0.9)
Cardiofasciocutaneous syndrome	1 (0.9)
Nonketotic hyperglycinemia	1 (0.9)
Sturge-Weber syndrome	1 (0.9)
Rett syndrome	1 (0.9)
Encephalitis	1 (0.9)
West syndrome	1 (0.9)

The overall attrition for the study was 31% of patients (n = 34) accounting for 69 exposures. Attrition was solely related to the failure of the patients to return for follow-up. Data was available for all 215 exposures within the first 6 months, 124 (n = 189, 65.6%) at 12 months, and 110 (n = 147, 74.8%) at follow-up greater than 1 year (mean = 23.89 months, SD = 8.02). Twenty-seven (25%) patients discontinued clobazam during follow-up with an average exposure of 5.68 months (SD = 4.22, range = 0.66-17.74). The reason for discontinuation was adverse effects (n = 10, 37%), lack of efficacy (n = 9, 33%), lack of efficacy and adverse effects (n = 4, 15%), expense (n = 3, 11%), and conversion to another benzodiazepine (1, 4%).

Dosing and Tolerance

The mean clobazam dose at initiation was 0.88 mg/kg/d (SD = 0.46, range = 0.23-2.17) and the mean dose of clobazam at initial response was 1.05 (SD = 0.71, range = 0.23-4.66). When stratified into low- (n = 44), medium- (n = 63), and high-dose (n = 108) groups, those treated with low-dose clobazam were more likely to be responders at 6 months (30, 68%) versus medium- (35, 56%) or high-dose groups (42, 39%) (χ ² = 11.93, P = .003). Though not significant, there was a maintained trend for patients on low-dose clobazam to more frequently report response compared to medium and high dosing at 12 months and greater than 1 year of follow-up as well. When response to treatment was compared by dose to seizure type, there was no significant difference in response between drop and nondrop seizures at any dosing range.

Of the 215 exposures analyzed, 132 (61%) reported a response to clobazam at some point during their follow-up (mean = 5.37 months, SD = 5.2). Of these, 88 (67%) reported sustained response at subsequent visits (mean duration = 13.47 months, SD = 9.53), whereas 44 (33%) reported <50% reduction in seizures at least once after previously responding to treatment. Kaplan-Meier survival plot of clobazam response is presented in Figure 1. Of the 215 exposures analyzed, 106 (49%) experienced an event indicating a change in response with a mean survival time of 15.51 months. Seizure reduction within the first 6 months was positively correlated with seizure reduction at 12 months (r = 0.456, P < .001) and >12 months (r = 0.417, P < .001). Similarly, seizure reduction at 12 months was highly correlated with seizure reduction at greater than 12 months (r = 0.738, P > .001).

OPEN IN VIEWER

Figure 1.

Kaplan-Meier survival plot for maintained response to clobazam

Discussion

We describe a large cohort of pediatric patients treated with clobazam for epilepsy at a single U.S. center and specifically examine the efficacy of clobazam for individual seizure types. The majority had medically intractable epilepsy representing a variety of epilepsy types and etiologies, providing evidence for the broad-spectrum efficacy of clobazam. Patients were followed for between 6 months and 4.5 years (mean = 16 months) after clobazam initiation, allowing for analysis of sustained response.

Among all patients, 50% responder rates were reported for at least 1 seizure type in 59%, 39%, and 30% of patients at 6, 12, and >12 months’ follow-up. Responder rates were similar to those in prior studies in children reporting response rates from 38% to 80% when followed for durations of 3 months to 5 years. ^{2,6,7,9,11,12} Higher response rates were reported in studies of shorter duration (i.e., 3-6 months), suggesting response may wane as duration of treatment increases. Fifty percent of our cohort was followed for >1 year after clobazam initiation and the favorable responder rate seen in greater than one-third of patients at that duration is encouraging. Many studies have demonstrated that when a response to clobazam is present, the majority of patients experience significant (i.e., >90%) reduction in seizure frequency. ^2,7,11 We found that 60% to 84% of responders reported reduction in seizure frequency >90%. In our cohort of medically intractable patients, such a remarkable reduction in seizure frequency is impressive. Seizure freedom on clobazam therapy has been reported in 12% to 26% of patients with localization-related epilepsy and 9% to 40% of patients with generalized epilepsy, similar to our seizure-free rate of 11%. ^{6 –8,11 –13}

Age at epilepsy onset, epilepsy duration, patient age, gender, and etiology of epilepsy had no impact on outcome. These are important findings, given clobazam is approved in the United States to treat seizures associated with Lennox-Gastaut syndrome in children aged >2 years. Several other studies have demonstrated the efficacy of clobazam for a variety of epilepsy syndromes and etiologies beyond Lennox-Gastaut syndrome. ^{6 –8,11,14 –16} Eighty-five percent of the patients included in this study had a diagnosis other than Lennox-Gastaut syndrome and 9.3% were younger than 2 years of age.

Although clobazam is approved for the treatment of seizure types associated with Lennox-Gastaut syndrome, the phase III trial that led to this approval was designed to measure efficacy against drop seizures specifically. ^2,3 In children with intractable localization-related seizures, Marcondes da Silviera et al ⁸ found 38% demonstrated at least 75% reduction in seizure frequency when treated with adjunctive clobazam therapy. Other studies have included children with both localization-related and generalized seizures, but have not specifically analyzed response based on seizure type. ^{10 –12} We found no seizure type responded more favorably to clobazam. Some seizure types (i.e. infantile spasms) were present in a small number of patients and may not have been adequate to detect a difference in treatment response. However, when we dichotomized the population into drop versus nondrop seizure types, the data still failed to demonstrate a difference in response to treatment. Although this is contrary to data by Ng et al, ² which showed an increase in nondrop seizures for low- (0.25 mg/kg/d) and medium-dose (0.5 mg/kg/d) clobazam and a reduction of 40% in the high-dose group (1 mg/kg/d), our cohort was exposed to a mean clobazam dose of 0.88 mg/kg/d at initiation, which may account for the improved efficacy seen in nondrop seizures. Although drop seizures did not respond more favorably during the time period analyzed, there was improved response in patients with a diagnosis of generalized epilepsy at all follow-up time points, which reached significance at 1 year. It is possible that a difference in drop versus nondrop response could occur over a longer duration, though prior studies demonstrated such a difference in less than 3 months of treatment. ² It should not be unexpected that a variety of seizure types would respond to clobazam, as other benzodiazepines exhibit a similar broad efficacy.

Patients in our study were exposed to a range of clobazam dosing. Similar to prior studies, we examined the response based on a low, medium, and high dosing and found that patients on low dosing were responders earlier in treatment and tended to demonstrate higher rates of response at subsequent visits. Twenty-eight patients demonstrated achievement of seizure reduction at doses greater than 1 mg/kg/d (maximum dose = 4.66). This suggests that response to clobazam will be evident in most patients at a lower dose, but there may be benefit in some patients to titrating dosing higher if the medication is tolerated.

Tolerance to treatment is a significant concern for most antiepileptic drugs, but may be especially prominent among benzodiazepines used as long-term therapy. There may be a lower likelihood of developing tolerance to clobazam versus other benzodiazepines. ⁴ Several studies have reported rates of tolerance ranging from 5% to 10% at 6 months to 40% at longer than 1 year. ^8,10 Singh et al ¹⁷ specifically examined clobazam tolerance in patients with initial favorable response and found a 50% rate at a mean follow-up of 17 months, similar to our 50% mean survival of 15 months. Studies ^13,18 have found that patients responding to clobazam for greater than 1 year are more likely to maintain response long-term. Thus, an early, low-dose response may be a favorable predictor of sustained response to long-term therapy.

Adverse effects are reported to occur in 20% to 88% of patients treated with clobazam and are most often central nervous system effects of somnolence or irritability, similar to our findings. ^{2,7 –11,13,16} We did not find a relationship between clobazam dosing or number of concomitant antiepileptic drugs to the presence of any adverse effect. Overall, adverse effects were mild, as only 14 (13%) led to discontinuation of the medication.

This study is limited by retrospective design. As a result, follow-up data were not available for all patients throughout the time period under study. Certainly, the overall responder rates may have been different if all patients’ data were included at each follow-up time point. Likewise, we relied on chart review to determine frequency reduction for each seizure type, which is an imprecise measurement. However, given the frequency of seizures experienced by the patients included in this study, a significant response of >50% seizure reduction is likely to be accurate. Finally, patients were not required to keep concomitant antiepileptic drugs stable during treatment; thus, response to therapy cannot be solely attributed to clobazam. The rates of concomitant antiepileptic drug changes were not significantly different between responders and nonresponders at any time point, and changes occurred in less than one-third of patients.

Clobazam is an effective antiepileptic drug for the treatment of intractable childhood epilepsy and provides significant seizure reduction for a broad spectrum of seizure types and epilepsy etiologies. The medication is well tolerated in children, with sedation and agitation being the most commonly encountered adverse effects. The response to clobazam is often present early in treatment at doses of less than 0.5 mg/kg/d and is maintained for an average duration of 15 months, though in some cases response may be achieved at doses beyond 1 mg/kg/d. Although tolerance appears to develop in some cases, an early and low dose response is a favorable indicator of sustained response. When a response to therapy is present, patients often experience significant seizure reduction (i.e., >90%). Given the intractability of the cohort under study, the response rate experienced is a considerable achievement and supports the consideration of clobazam as adjunctive therapy for a variety of childhood seizures types and epilepsy syndromes. Additional prospective studies should be undertaken to support these conclusions.