Zonisamide Therapy for Refractory Idiopathic Epilepsy in Dogs

Introduction

Seizure disorders represent the most common neurological problem in dogs, with the majority of cases attributable to idiopathic epilepsy. Although most epileptic dogs are managed effectively with “standard” (e.g., phenobarbital, bromide) anticonvulsant drugs, approximately 25% to 30% are refractory to appropriate doses of these drugs.^1–3 Increasing the dosage of phenobarbital or bromide in an effort to improve seizure control can lead to unacceptable side effects (e.g., polyuria/polydipsia, polyphagia, lethargy, weight gain). There are few alternative anticonvulsants to use as additional therapies in refractory canine seizure cases, as most of the available drugs used in people have very short elimination half-lives in dogs.¹² Two drugs often used as additional drugs for refractory epilepsy are felbamate and gabapentin. Evidence of efficacy for both of these drugs is largely anecdotal, especially for gabapentin. Felbamate has an elimination half-life in dogs of approximately 5 to 6 hours; the elimination half-life of gabapentin is approximately 3 to 4 hours in dogs.^124–6 Felbamate probably requires dosing q 8 hours and gabapentin q 6 hours to maintain adequate blood levels. There is a recognized need for safe and effective alternative anticonvulsant drugs that can be used in dogs with refractory seizure disorders.

Zonisamide^a (1,2-benzisoxazole-3-methanesulfonamide) is a sulfonamide-based anticonvulsant drug recently approved for human use in the United States. Potential anticonvulsant mechanisms of action include blockage of T-type calcium and voltage-gated sodium channels in the brain, modulation of dopaminergic metabolism in the central nervous system (CNS), scavenging free radicals, enhancing actions of gamma aminobutyric acid in the CNS, and inhibition of carbonic anhydrase activity.^7–10 This last mechanism is thought to be the least substantial anticonvulsant effect.^7–10 Zonisamide has been shown to be an effective anticonvulsant agent for both focal (i.e., partial) and generalized seizures in humans, and side effects are minimal. The most common side effect reported in people is transient drowsiness. Other reported side effects include ataxia, gastrointestinal upset, tremors, and behavioral disorders (in rare cases). Zonisamide is metabolized primarily by hepatic microsomal enzymes.^7–14

To date, there is no published information documenting the efficacy of zonisamide for canine seizure disorders. However, the drug appears to be well tolerated by dogs, even at very high doses, for extended periods of time. In one study, minimal side effects occurred in beagle dogs administered zonisamide in doses up to 75 mg/kg per day for 1 year.¹⁵ Pharmacokinetic data concerning zonisamide administration in dogs is lacking. In one investigation, an elimination half-life of approximately 15 hours was reported; however, this was based on only two dogs.¹⁶ The suggested therapeutic range of serum zonisamide concentrations in people varies between 7 and 40 μg/mL. The lower end of the therapeutic range is considered to be 10 μg/mL, according to most reports.¹⁰¹¹

The purpose of this investigation was to evaluate the efficacy, tolerance, and pharmacokinetic properties of zonisamide as an additional anticonvulsant in dogs with refractory idiopathic epilepsy.

Materials and Methods

The study was limited to dogs with seizures associated with idiopathic epilepsy. Idiopathic epilepsy was diagnosed via characteristic signalment and clinical features for the disease (i.e., onset of seizures between 1 and 5 years of age, normal neurological examination interictally, normal laboratory tests). Dogs outside of the 1- to 5-year age range were eligible for inclusion in the study if they had normal brain computed tomographic (CT) scans or magnetic resonance (MR) imaging and normal cerebrospinal fluid (CSF) analyses. A dog was considered eligible for zonisamide therapy if seizures were uncontrolled despite adequate serum levels of one or more “standard” (e.g., phenobarbital, bromide) anticonvulsant drugs. A minimum average of two seizures per month and a total of at least six seizures were required for inclusion in the study. Dogs with subtherapeutic serum levels of their anticonvulsants were considered eligible for the study if unacceptable side effects attributable to these drugs were reported by the owner.

From the limited pharmacokinetic information available, it was estimated that a dosage of 5 to 10 mg/kg per os (PO) q 12 hours would be necessary to achieve serum concentrations within the therapeutic range (10 to 40 μg/mL).¹⁵¹⁶ Each dog was initially given zonisamide in this dosage range. Dosages were adjusted based upon serum drug concentrations, until the concentration was within the therapeutic range and minimal side effects were apparent. In order to ensure steady-state kinetics were achieved, all dogs were maintained on zonisamide for a minimum of 1 week before serum drug concentrations were measured. A 1-week equilibration period was also allowed between zonisamide adjustments and repeated measurement of serum zonisamide levels. Serum zonisamide concentrations were measured at a commercial laboratory^b via high-performance liquid chromatography (HPLC).¹⁰¹¹¹⁷ The assay was validated for human serum samples, with a lower limit of quantitation of 5 μg/mL. Serum samples were obtained at time 0 (immediately before drug administration) and 3 hours after drug administration in all dogs. The 3-hour estimate for peak serum zonisamide levels was based on the available pharmacokinetic data in dogs.¹⁵¹⁶

Seizure frequency before and after zonisamide therapy was estimated from medical records, logs kept by the owners, and conversations with owners and referring veterinarians. In dogs with cluster seizure activity, individual seizures were counted separately if they were clearly documented by owner logs or medical records. If the number of seizures per cluster was not obvious, each cluster event was counted as a single seizure episode. Seizure frequency was recorded for a minimum of 8 weeks following initiation of zonisamide therapy. For each dog, seizure frequency was recorded for equivalent time periods before and following initiation of zonisamide therapy. The absolute numbers of seizures in the time periods before and during zonisamide therapy were recorded and compared. A positive responder was defined as a dog that experienced at least a 50% reduction in seizures after initiation of zonisamide.

Trough and estimated peak serum zonisamide concentrations were compared for all animals, using a paired Student’s t-test (P<0.05). Serum zonisamide concentrations (trough and peak) were also compared between responders and nonresponders, using an unpaired Student’s t-test (P<0.05). Results of laboratory tests (i.e., complete blood count [CBC], serum biochemical panel) before and during therapy were evaluated if available. Owners were questioned via telephone or at the time of recheck examinations regarding side effects attributable to zonisamide.

Results

Twelve dogs met the criteria for inclusion in the study. Ten of the dogs were patients of Long Island Veterinary Specialists, one was from County Animal Specialty Group, and one was from Veterinary Neurology and Neurosurgery Specialists. The study included five mixed-breed dogs and one each of the following purebred dogs: golden retriever, Akita, giant schnauzer, basset hound, bichon frise, Japanese chin, and Chihuahua. There were seven castrated males, one intact male, and four spayed females. Ages ranged from 9 months to 10 years (mean age, 4.42 years; median age, 4.58 years). The mean and median ages at onset of seizure activity were 2.32 years and 2.46 years, respectively (range, 4 months to 4 years and 9 months). Five dogs had normal MR imaging of the brain, and four of these dogs also had a normal CSF analysis.

Clinical data concerning seizure history, seizure control, anticonvulsant drug changes, and serum zonisamide concentrations are summarized for responders and nonresponders in Tables 1 and 2, respectively. Duration of seizure activity prior to study enrollment ranged from 2 to 82 months (mean, 26 months; median, 20.5 months). All 12 dogs experienced generalized seizures, two of which also displayed focal (i.e., partial) seizures. Nine of the dogs’ owners kept well-maintained seizure logs. Ten dogs were receiving phenobarbital and potassium bromide at the start of the study. One of these dogs was also receiving felbamate, and one was also receiving gabapentin. One dog was treated with phenobarbital and clorazepate. Potassium bromide therapy had previously been attempted in this dog, but it was discontinued when the dog became aggressive after 5 days of therapy. One dog was treated with phenobarbital and felbamate. This last animal was treated briefly with potassium bromide, but the drug was discontinued because of unacceptable side effects (i.e., vomiting, polyuria) prior to attaining serum drug levels. Serum bromide levels in all other dogs were within the therapeutic range (1 to 3 mg/mL). Three dogs receiving phenobarbital had serum drug levels below the therapeutic range (15 to 45 μg/mL). These three dogs had serum bromide levels in the middle (1.4 mg/mL) to high (2.7 mg/mL, 2.9 mg/mL) therapeutic range and had experienced unacceptable side effects (i.e., sedation, ataxia) at higher phenobarbital dosage regimens.

Duration of anticonvulsant therapy prior to adding zonisamide ranged from 2.5 to 82 months (median, 20 months). Duration of the specific treatment regimen in place prior to beginning zonisamide therapy ranged from 1 to 31 months (median 5.8 months). The mean and median follow-up times after zonisamide administration were 33.5 weeks and 37 weeks, respectively (range, 8 to 71 weeks).

The mean zonisamide dosage administered in this study was 8.9 mg/kg (range, 5 to 11 mg/kg) PO q 12 hours. With the exception of one trough value of 9.7 μg/mL, all trough and peak serum zonisamide concentrations were within the therapeutic range of 10 to 40 μg/mL. The mean and median trough serum zonisamide concentrations were 19.7 μg/mL and 18.4 μg/mL, respectively. The mean and median peak serum zonisamide concentrations were 23.5 μg/mL and 21.2 μg/mL, respectively. Estimated peak zonisamide concentrations were significantly higher than trough levels (P=0.007). There were no significant differences between the serum zonisamide concentrations of responders versus nonresponders for either trough (P=0.36) or peak (P=0.32) values.

Seven of the 12 (58%) dogs experienced at least a 50% reduction in seizures while on the zonisamide and were considered positive responders. Reduction of seizure frequency in the responder group ranged from 54.8% to 100% (mean 81.3%; median 84.5%). Two dogs (case nos. 5, 6) became seizure free after adding zonisamide therapy. A third dog (case no. 2) was seizure free until the owner arbitrarily and abruptly discontinued phenobarbital and lowered the zonisamide dose by 25%; the dog then seized five times in the ensuing 24 hours. After the owner reinstituted the correct drug regimen, the dog was again seizure free throughout the follow-up period (approximately 6 months). Five (42%) of the dogs were nonresponders, with all experiencing an increase in seizure frequency during zonisamide therapy. The median increase in seizures for the nonresponders was 52.6% and ranged from 7.4% to 2500%.

Reduction or elimination of concurrently administered anticonvulsants was possible in nine (75%) dogs (seven responders, two nonresponders) following institution of zonisamide therapy. In the responder group, the dosage of phenobarbital was reduced (in six of seven dogs on the drug) by a mean of 92.2% [Table 1]. Phenobarbital was discontinued in three of these dogs. Potassium bromide was also discontinued in two of the dogs receiving it. Felbamate was discontinued in one responder dog, and clorazepate in another. In the nonresponder group, the dosage of phenobarbital was decreased by 33% in one of the five dogs receiving it [Table 2]. Phenobarbital was reduced in this nonresponder dog because of excessive sedation and uncontrollable urination. In one nonresponder dog, phenobarbital was increased by 13%. Four nonresponder dogs were also receiving potassium bromide. Bromide was discontinued at the start of zonisamide therapy in one dog with multiple cluster seizures, because the dosage required to control the seizures resulted in severe sedation. Another dog also experienced side effects (ataxia, sedation) from the potassium bromide that required transient withdrawal of the drug. The side effects subsequently subsided. Before the potassium bromide was reinstituted at a lower dose (67% reduction), the dog experienced 11 of the 29 seizures that occurred during the zonisamide study period. Overall, this dog experienced only a 7.5% increase in seizure activity while on zonisamide.

Six (50%) dogs experienced side effects attributable to zonisamide. Four were zonisamide responders, and two were nonresponders. Two dogs (case nos. 6, 11) experienced generalized ataxia, which resolved following reduction of the zonisamide dosage. Mild generalized ataxia in a third dog (case no. 2) was noted, but it did not warrant a change in the zonisamide dosage. One dog (case no. 4) developed transient lethargy, which resolved within several days without adjustments in the drug. One dog (case no. 5) developed lameness associated with a polyarthropathy. This dog had a high serum titer for Lyme disease, and signs improved with doxycycline therapy. Transient vomiting occurred in one dog (case no. 9), which resolved over several days with no change in the drug regimen. This dog also developed keratoconjunctivitis sicca (KCS), which responded to topical cyclosporine treatment. The mean trough and peak serum zonisamide concentrations for the six dogs that experienced side effects were 18.22 μg/mL and 20.08 μg/mL, respectively.

Complete blood counts and serum biochemical profiles were performed in eight dogs before and during zonisamide therapy. Laboratory tests were performed, on average, 24 weeks (range, 7 to 42 weeks) after initiation of zonisamide. There were no abnormalities noted in any of the CBCs. Marked elevations of serum alanine aminotransferase (ALT) and alkaline phosphatase (ALP) were evident in one dog in which a hepatic mass was visualized on ultrasound examination (case no. 4). The ALT activity in this dog was 342 U/L pre-zonisamide and 550 U/L during zonisamide therapy (reference range, 12 to 118 U/L). The ALP activity was 3290 U/L pre-zonisamide and 1740 U/L while on zonisamide (reference range, 5 to 131 U/L). For the remaining seven dogs, all serum ALT measurements taken during zonisamide therapy were within the reference range. Two dogs had elevated ALT measurements on their pre-zonisamide profiles. Serum ALP activity during zonisamide therapy was elevated in five of the remaining seven dogs. The ALP levels during zonisamide therapy ranged from 137 to 1740 U/L (mean 622 U/L). The pre-zonisamide serum ALP activity was abnormal in three of the five dogs, and it was actually higher than the serum ALP activity measured during zonisamide therapy.

Two of the nonresponder dogs (case nos. 9, 11) were euthanized for uncontrollable seizure activity approximately 9 and 11 weeks after entering the study, respectively. One of the responder dogs (case no. 4) was euthanized 44 weeks after entering the study, following the diagnosis of a hepatic mass on ultrasonography. Necropsies were not performed in these three dogs.

Discussion

Results of this study demonstrated that oral zonisamide was well tolerated and that a q 12-hour dosing interval maintained adequate serum drug concentrations. Although case numbers are low in this study, it appears that zonisamide is often effective in reducing seizure frequency in dogs with refractory idiopathic epilepsy, when it is used as an additional therapy. In human anticonvulsant trials, the number of patients that experience at least a 50% reduction in seizures (responder rate) is a commonly used measure of efficacy.^710131418 Seven of the 12 refractory epileptics in this investigation were similarly classified as zonisamide responders. The dosages required to achieve serum zonisamide concentrations within the therapeutic range in the study reported here were well within dosages previously shown to be safe in dogs.¹⁵

Five dogs were classified as nonresponders. Three appeared to derive no benefit from the addition of zonisamide to their anticonvulsant drug regimen. Two of these dogs were euthanized because of uncontrollable seizure activity. The third dog experienced some improvement in seizure control after zonisamide was discontinued and felbamate therapy was added. The other two nonresponders (case nos. 8, 12) continued to receive zonisamide and may actually have benefited from the therapy. In one dog that developed side effects from the potassium bromide, the addition of zonisamide allowed the potassium bromide dosage to be reduced, and the side effects subsequently subsided. The owner of this dog was pleased with the effects of the zonisamide.

The last nonresponder dog was receiving a very large dose of potassium bromide prior to the study in an attempt to control multiple episodes of cluster seizures. Because of the high dosage of potassium bromide, the dog was very sedate and could not ambulate without assistance. However, the dog experienced only one cluster seizure during the intense bromide therapy. The number of seizures during this cluster episode was not clear from the owner’s log, so it was counted as a single seizure. Potassium bromide was discontinued coincident with starting zonisamide therapy, in an effort to improve the dog’s quality of life. Although this dog had the greatest increase in seizure frequency (2500%) of the nonresponders, the pre-zonisamide comparison period incorporated only a single cluster episode, not multiple cluster episodes. The owner did not feel this dog’s seizure frequency was any worse than before the administration of the high dosages of potassium bromide, and was pleased with the dog’s improved quality of life while on zonisamide.

Side effects are uncommon and mild in humans receiving zonisamide, and they primarily include sedation, gastrointestinal upset, ataxia, and tremors.^7–14 Rarely, behavioral disorders such as mania have been attributed to zonisamide therapy in people, yet zonisamide has also been effective in treating mania in humans.^19–21 Side effects attributable to zonisamide in this study were generally mild and included ataxia, transient vomiting, and lethargy. Side effects were not severe enough in any dog to warrant termination of the drug. Considering that all dogs were concomitantly receiving other anticonvulsant medications, it could not be determined whether the side effects were specifically related to the zonisamide or arose from a combination of drugs. The KCS (n=1) and polyarthropathy (n=1) that occurred in this study were of particular interest, because these disorders have also been associated with the use of other sulfonamide medications in dogs.^22–25 The high serum titer to Borrelia and the positive response obtained with doxycycline therapy in the dog with polyarthropathy suggests that zonisamide therapy may not have caused the dog’s joint disease. However, it is possible that zonisamide therapy exacerbated a previously subclinical arthropathy. Clinicians should be cognizant of potential side effects from sulfonamide-based zonisamide similar to those associated with other sulfonamides.

The only consistent laboratory abnormalities associated with zonisamide therapy in humans and dogs are elevations in liver enzyme activities, particularly that of serum ALP.⁷⁸¹⁵ Results of liver enzyme assays in the study reported here were similar to these prior observations. Since all dogs were concurrently receiving phenobarbital, it was not possible to ascertain the exact effect zonisamide had on these enzyme elevations. The only dog that had an elevated ALT level during the zonisamide treatment period also had an elevated ALT prior to zonisamide and a hepatic mass on abdominal ultrasonography. The rest of the dogs only had elevations of serum ALP, and most of the values were lower than the pre-zonisamide levels.

There are several noteworthy limitations to the study reported here. In addition to the small number of cases studied, there are recognizable inaccuracies in counting seizures both retrospectively and prospectively. Fortunately, most of the owners of the dogs in this study kept accurate logs of their pets’ seizure activity. Despite this factor, retrospective documentation of seizure frequency is inherently less accurate than prospective documentation. Because the dogs were refractory epileptics, it was not considered ethical to incorporate a pre-zonisamide time period for the purpose of counting seizure activity prospectively. Whenever there was a discrepancy in pre-zonisamide seizure numbers, the investigators chose to utilize the most conservative number.

The limited pharmacokinetic data obtained in this study appeared to justify a q 12-hour dosing regimen. Despite this useful information, a thorough pharmacokinetic analysis of zonisamide in dogs was not feasible in this clinical-based study. Such an analysis would be useful in verifying important pharmacokinetic properties of zonisamide in dogs, such as peak serum concentration (C_max) and elimination half-life. The elimination half-life of zonisamide in people already receiving drugs that stimulate hepatic microsomal enzymes (e.g., phenobarbital) is approximately half that of patients not receiving such medications.^111226–28 All dogs in this study were receiving phenobarbital, and repeat zonisamide serum concentrations were not available for dogs in which a substantial reduction in phenobarbital was possible. Therefore, it could not be ascertained from this study whether a similar decrease in zonisamide elimination half-life occurs in dogs receiving phenobarbital. The HPLC assay utilized for this investigation was validated for human, rather than canine, serum samples. For future pharmacokinetic studies, it would be preferable to develop an HPLC assay specifically validated for dogs.

Conclusion

In this study, zonisamide appeared to be effective in reducing seizure frequency in a majority of refractory canine epileptics. Side effects were generally mild and manageable. A q 12-hour dosing interval was effective in maintaining serum drug concentrations within the therapeutic range. In addition, a reduction in dosage or the elimination of concurrently administered anticonvulsant drugs was possible in 75% of the dogs. Because of the high cost of zonisamide in comparison to other anticonvulsant drugs, it is unlikely to become a popular first-line or sole therapy for canine idiopathic epilepsy. Based on the results of this study, further investigation is warranted on the efficacy of zonisamide as an additional anticonvulsant drug for dogs with refractory idiopathic epilepsy.