Canine Status Epilepticus Treated with Fosphenytoin: A Proof of Principle Study

Abstract

Objectives

There are a limited number of marketed intravenous antiepileptic drugs (AEDs) available for status epilepticus (SE). All were first developed for chronic therapy of epilepsy, not specifically for SE. Epilepsy and canine SE (CSE) occur naturally in dogs with prevalence, presentation, and percentage of refractory cases similar to human epilepsy. The objective of this study was to determine if CSE treated with fosphenytoin (FOS) results in a similar responder rate as for people.

Methods

A randomized clinical trial was performed for dogs with CSE. Dogs who presented during a seizure or had additional seizures after enrolling received an IV benzodiazepine (BZD) followed immediately by IV infusion of 15mg/kg phenytoin equivalent (PE) of fosphenytoin (FOS) or saline placebo (PBO). If seizures continued, additional AEDs were administered per the standard of care for veterinary patients. Total and unbound plasma phenytoin (PHT) concentrations were measured.

Results

Consent was obtained for fifty dogs with CSE. Thirty-one had additional motor seizures and were randomized to the study intervention (22 FOS and 9 PBO). There was a statistically significant difference in the 12 hour responder rate with 63% in the FOS group versus 22% in the placebo group (p=0.043) having no further seizures. The unbound PHT concentrations at 30 and 60 minutes were within the therapeutic concentrations for people (1–2 μg/mL) with the exception of one dog. There was mild vomiting in 36% of the FOS group (7/22) within 20 minutes of FOS administration and none of the placebo group (0/9) (p = 0.064).

Significance

This proof of concept study provides the first evidence that FOS is tolerated and effective in canine SE at PHT concentrations clinically relevant for human SE. Further, naturally occurring CSE can be utilized as a translational platform for future studies of novel SE compounds.

Introduction

Human Status Epilepticus (HSE) is a serious life-threatening neurological emergency consisting of prolonged and/or frequent seizures. HSE is associated with one of the highest mortalities and morbidities of any neurological condition. It has been reported that 152,000 cases with 42,000 deaths (28%) occur each year in the USA.^{1, 2} Treatment guidelines recommend use of benzodiazepines (BZD) and phenytoin (PHT) or its prodrug fosphenytoin (FOS).³ These guidelines are based on studies initiated in the 1990s with drugs developed 30 or more years ago for conditions other than the treatment of HSE. The most comprehensive study of HSE management was a comparison of 4 treatments: lorazepam; phenobarbital; diazepam followed by PHT; and PHT.⁴ PHT had a success rate of 43.6% when given alone, but for diazepam followed by PHT, it was 55.8%. The best treatment, lorazepam, had a success rate of only 64.9%.⁴ There is a need to develop models which provide evidence that new drugs have potentially superior efficacy or safety compared with accepted therapy. Promising new agents discovered using experimental animal models of status epilepticus cannot easily be tested in HSE without more evidence of safety and efficacy. Related to this issue is a mechanism to develop and evaluate drugs that may be specifically effective for HSE and yet may, but not be appropriate for chronic use. Treatment of HSE requires only short term use of drugs in patients who have a decreased level of consciousness. However, using the conventional approach to development of new AEDs, significant adverse effects arising from chronic exposure, will preclude the development of these AEDs for SE. The ideal drugs for HSE are those that have a rapid onset of action, have neuro-protective activity, and are free from significant respiratory and cardiovascular effects.

In contrast to artificially induced seizures in rodent models, epilepsy occurs naturally and spontaneously in dogs. Rodent models also often fail to predict cognitive, behavioral and neurological side effects (irritability, insomnia, balance) and systemic effects (cardiac arrhythmias, hypotension).⁵ The clinical manifestations of seizures in dogs are very similar to those observed in humans, including seizure types,⁶ and electroencephalographic findings.^7,8 Naturally-occurring canine status epilepticus (CSE) is one of the more common emergency conditions treated at veterinary hospitals.⁹ Approximately 59% of dogs with epilepsy have one or more episodes of CSE during their lifetime, and those with CSE had a mean life span of only 8.3 years compared to 11.3 years for those with epilepsy; but no CSE.^10,11 Despite the similarities with epilepsy in humans, use of naturally occurring canine epilepsy for drug development has been underutilized.⁵ A positive proof of concept study with a drug that has been well documented to work in HSE would validate the model and could propel additional studies of novel drugs in CSE. Using dogs as a translational platform has the added advantage of more accurate dose extrapolation to people because body sizes of dogs are closer to humans than are those of rodents (rodent to human is approx.1:4000; dog to human is approx. 1:4).

PHT can cause hypotension cardiac arrhythmias and serious infusion-site adverse effects.¹² Fosphenytoin is a highly water-soluble prodrug of PHT and circulating phosphatases convert FOS to PHT with a half-life of 8–15 minutes, and was chosen for this study due to less concern with cardiac, blood pressure, and infusion site adverse effects.¹³ For novel drugs, once an effective blood level has been determined in dogs, a dose for humans to attain the same levels can be calculated using standard loading dose equations incorporating parameters such as the volume of distribution with the appropriate caution of species differences in metabolism, protein binding, etc.

The goal of this study was proof of concept of using CSE as a translational platform with a drug, FOS, effective for HSE. Demonstrating similar efficacy with FOS would validate CSE as a translational platform for HSE and allow novel drugs to be moved more quickly from the laboratory into studies of HSE. Also, compounds that may not be pursued for chronic treatment of human epilepsy can be evaluated in CSE as potentially suitable for HSE. As an additional benefit for veterinarians, the results of this study and future studies will also inform the treatment of CSE.

Methods

Design

This was a prospective, double-masked, randomized, placebo-controlled study. Each dog that conformed to the inclusion (established CSE) and exclusion criteria was enrolled. IACUC or equivalent approvals were obtained at all sites.

Inclusion criteria

Dogs with a clinical diagnosis of convulsive status epilepticus defined as continuous convulsion lasting more than 5 minutes, or 2 or more recurrent convulsions without regaining consciousness between seizures.

Study Protocol

Dogs presenting for the emergency treatment of seizures at four veterinary centers (University of Minnesota, University of Pennsylvania, Chicago Veterinary Neurology, and Bush Veterinary Neurology) were considered for entry. Informed consent was obtained directly from the owner of the dog, If the inclusion criteria was met, dogs were admitted to the veterinary hospital for monitoring for 5 hours consisting of: 1) continuous observation for seizures; and 2) hourly checks for alertness, vomiting, diarrhea, or salivation and measures of temperature, pulse and respirations. Pre-treatment CBC, serum chemistry, blood pressure, ECG, and either bile acids or resting ammonia was obtained. If the subject recovered fully, the subject was discharged from the study. If another seizure occurred during the 5 hours monitoring or the dog presented during a seizure, 0.5 mg/kg of diazepam or 0.2mg/kg lorazepam (when diazepam was not available) was given IV. The BZD dose was immediately followed by an IV loading dose of FOS of 15mg/kg phenytoin equivalent (PE) at a rate of 50mg PE/min (1 ml/min), or an equivalent volume of 0.9% NaCl at 1ml/min was given. This dose of FOS in dogs was designed to attain plasma PHT levels similar to that when humans receive 18–20mg/kg PHT.¹⁴ A rescue protocol (diazepam CRI, and/or IV levetiracetam, and/or IV propofol) per the veterinary standard of care¹⁵ was initiated if the dog’s convulsions did not diminish within 10 minutes after the completion of infusion or re-occurred within 12 hours. Post study drug infusion monitoring consisted of blood pressure, 60 second ECG, and modified Glasgow Coma Scale¹⁶ assessment at 10 and 20 minutes. In addition, vital signs were checked hourly for the entire 12 hours.

Primary Endpoints

The primary endpoints were an additional seizure within 2 and 12 hours after completion of infusion. If there were not any seizures within these time periods the case was considered a responder.

Secondary Endpoints

The secondary endpoints were number of seizures within 12 hours, number of bolus injections of diazepam given, number of dogs receiving a constant rate infusion of diazepam and duration of infusion, number of dogs receiving either propofol treatment, hours of propofol, and hours of hospitalization beyond twelve hours.

Plasma PHT Concentrations

Blood samples were collected at 0.5, 1.5 and 3 hrs after the end of the FOS infusion. A reversed phase, high performance liquid chromatographic assay was used for measurement of bound and unbound PHT and was adapted for canine plasma using a previously validated method to measure bound and unbound PHT in human plasma.¹⁴ For unbound concentrations, 1 ml of plasma was spun using an Amicon Centrifree ultrafiltration device (Merck Millipore Ltd) at 37°C for one hour at 4000rpm to separate drug bound to proteins from unbound drug. For PHT measurement sample extraction from canine plasma was accomplished with hexane and ethyl acetate followed by evaporation of the solvent under nitrogen in a 37°C water bath. Compounds were eluted using an ion-pair mobile phase containing 25% acetonitrile and 75% phosphate buffer (pH 3.5) and PHT was detected using UV absorbance at 215 nm. The drug concentrations obtained from these subjects were used to confirm attainment of the target concentrations, and examine the relationship between PHT concentration, seizure control, and adverse events.

Power Calculation and Statistical Analysis

The primary endpoints were compared between groups using a Fisher’s exact test. Secondary endpoints were compared between groups using the Wilcoxon rank sum test or a Fisher’s exact test. In addition, the age, sex, clinic location, and number of seizures in the 12 hours prior to admission were compared between the two groups to determine if the groups are otherwise equivalent. Differences were considered statistically significant at a p < 0.05. Sample size and power estimates were based on a literature review of studies of 3 AEDs with human subjects as well as our experience with our canine LEV study.¹⁷ Sample size estimates with 80% power to detect a 40% difference between FOS (50% effective) and placebo (10%) for 12 hours responders using a significance level (alpha) = 0.05 indicated total sample size of 46 enrolled cases receiving study infusion. An a priori, planned interim analysis at 20 enrolled patients was performed to determine exact power calculations with this cohort of dogs with CSE. The initial 20 dogs were randomized to about one half FOS and one half placebo by a statistician using a random number generator. After the interim analysis, the remaining subjects were randomized at a ratio of 5 FOS to 1 PBO due to the potential futility of the placebo group with a 0% 12 hour responder rate.

Results

Interim analysis

This analysis showed that at 12 hours 69.2% (9/13) of the patients in Treatment Group A (later revealed as the FOS group) and 0% (0/7) of the patients in Group B (later revealed as PBO) were responders (p = 0.0031). At 2 hours, Group A (FOS) had 100% (13/13) responders vs. Group B (PBO) with 71% (5/7) responders (p=0.11). With statistical significance already found for 12 hours responders, it was calculated that a total of 32 subjects would be needed to potentially find a statistically significant difference at the 0.05 level for 100% vs. 77% for the 2 hour endpoint. Based on this power calculation, it was decided to try to enroll 12 additional dogs (for a total of 32) at a ratio of 5 Group A (FOS):1 PBO (Group B) in order to possibly achieve the power to detect a difference at 2 hours, and also due to the apparent futility of Group B (PBO) in this population.

Final Study Population

Ultimately, consent was obtained from owners of 50 dogs with SE. One case presented during a seizure 30 cases had additional motor seizures, and were randomized to the study intervention at about an overall 2:1 ratio (22 in the FOS group and 9 in the PBO group). This was just as funding ended and was one short of the goal of 32 total enrolled cases. Breeds with more than one enrolled case included 5 Golden retrievers (3 FOS, 2 PBO), 4 Labrador retrievers (1 FOS, 3 PBO), 3 Bulldogs (2 FOS, 1 PBO), and 3 German shepherd dogs (2 FOS, 1 PBO). There were 16 female dogs (12 FOS, 4 PBO), and 15 male dogs (10 FOS, 5 PBO). Six dogs in the FOS group had secondary epilepsy (1 meningioma, 1 glioma, 2 inflammatory encephalitis, and two cases of hydrocephalus), and two dogs in the placebo group had inflammatory encephalitis. All other dogs in both groups (n=23) met criteria for genetic (idiopathic) epilepsy.

There were no statistically significant differences in pre-treatment CBC, serum chemistry, bile acids, age, weight, sex, neuter status, underlying cause of seizures, number of doses of diazepam prior to study admission, or number of seizures in the 12 hours prior to admission between the FOS and PBO groups (Table 1).

Table 1.

Comparison of fosphenytoin (FOS) treated and placebo treated dogs with status epilepticus by age, weight, seizure frequency, etiology, and sex.

	FOS (22)	Placebo (9)	p-value
Age in years (range)	6.5 (3–13)	4.4 (1.5–12)	0.10
Weight in pounds (range)	28.0 (10–58)	36.8 (29–48)	0.15
Mean # seizures within 12 hrs prior to admission (range)	6.4 (3–10)	5.6 (3–10)	0.59
Genetic (Idiopathic Epilepsy)	16/22 (73%)	7/9 (78%)	0.57
Secondary Epilepsy	6/22 (27%)	2/9 (22%)	0.57
Male	10/22 (45%)	5/9 (55%)	0.70
Female	12/22 (55%)	4/9 (44%)	0.70

Outcomes

There was a statistically significant difference in 12 hour responder endpoint with 63.6% (14/22) in the FOS group vs. 22.2% (2/9) in the placebo group (p=0.043). There was also a significant difference for the 2 hour responders endpoint with 21/22 (95.4%) responders in the FOS group vs 5/9 (55.5%) in the PBO group (p = 0.017) - Table 2. There was not a statistically significant difference in the responder rate between dogs in the lower 50% of the weight range (<32kg – responder rate 66%) vs. dogs in the top 50% of weight (63%). Similarly there was not a significant difference in responder rate for dogs in the lower 50% of age range (< 4.5 yrs – 62% responder rate) versus dogs in the top 50% of age (64%)

Table 2.

Representative primary and secondary endpoints and other important outcome variables comparing fosphenytoin (FOS) versus placebo in canine status epilepticus (CSE).

Endpoint/variable	FOS Group (n=22)	Placebo (n=9)	p-value
% responders at 2 hours	21/22 (95%)	5/9 (55%)	0.017*
% responders at 12 hours	14/22 (64%)	2/9 (22%)	0.043*
# episodes of motor seizures > 5 mins 1st 12 hours	1/22 (4.5%)	0/9 (0%)	0.71
Mean # seizures 12h post Tx (range)	0.61 (0–4)	1.54 (0–4)	0.020*
Vomiting	7/22 (32%)	0/9 (0%)	0.065
In-hospital mortality	3/22 (13%)	1/9 (11%)	0.63
Arrhythmia during or post infusion	2/22 (9%)	0/9 (0%)	0.50

Statistically Significant Differences Indicated with *

Pharmacokinetics

The PHT concentrations in plasma collected at 30, 60 minutes and 120 minutes post-dose (Figure 1) were within the therapeutic concentrations for human subjects (7.5–20 μg/mL total and 1–2 μg/mL unbound) at 30 and 60 minutes with the exception of one dog that had low total and unbound PHT concentrations at 30 min post dose of 7 and 0.7 μg/mL, respectively. There was no significant difference between total or unbound drug concentrations for responders versus non-responders although there was a potential trend toward greater unbound concentrations in responders at 30 minutes post-dose (Figure 2).

Figure 1.

Total and unbound PHT concentrations in plasma collected 30,60 and 120 minutes post fosphenytoin dose.

Figure 2.

Total and unbound PHT concentrations in plasma collected 30 minutes post dose in dogs that were seizure free (responders) or exhibited a seizure(s) (non-responders) within 12 hours of fosphenytoin administration. Mean and 95% CI represented by horizontal lines.

Adverse Effects

There were no significant differences for any adverse effects or other medical data variables (Table 2). Mild, drug controllable vomiting was more common in the FOS group of 32% (7/22) vs 0% (0/9) in the PBO group, but this did not quite reach statistical significance in this number of patients (p = 0.065) The initial vomiting was within 20 minutes of FOS infusion completion in all 7 dogs, with only one dog having a second episode at 40 minutes. Unbound PHT levels were similar at 30 mins in the dogs that vomited (1.90 μg/mL) compared to the dogs that did not vomit (1.70 μg/mL) (p = 0.36). Dogs with vomiting were treated with maropitant or ondonsetron and no vomiting occurred beyond 40 minutes after completion of the study infusion. No aspiration pneumonia was documented to have occurred in any of the dogs. Due to the known risk of vomiting from FOS in dogs^15,18 it was recommended in the study protocol to keep the head lowered for 15 minutes post study drug infusion, to try to prevent aspiration. Occasional premature ventricular contractions occurred in two dogs in the FOS group (9%) just after the FOS infusion, but treatment was not indicated or initiated in either dog. Hypotension was not detected in any dogs, as systolic blood pressures were greater than 80mm in all dogs during and after the study drug infusion.

The overall in-hospital mortality was 12.9% (4/31) which is within the general expected range for dogs with SE.^9–11 There was a 13.6% (3/22) mortality rate in the FOS group vs 11.1% (1/11) in the PBO group which was not statistically significantly different (Table 2). The one death in the PBO group was an owner elected euthanasia 16 hours after study infusion in a dog that had stable vital signs but was a non-responder with continuing seizures. In the FOS group 2/3 deaths were also owner-elected euthanasia’s in stable non-responders. One FOS case died just prior to 3 hours post FOS infusion. This dog did vomit one time, 1 minute post FOS infusion completion, and was treated with 0.2mg/kg IV ondansetron at that time, and there was not any additional vomiting. The dog had blood pressures, ECGs, and respiratory rates within normal limits for the planned 10 and 20 minute post infusion monitoring, and appeared stable until just prior to 3 hours post FOS infusion. At that time the dog was found to be very weak, tachycardic with pulse rate of 144, and a capillary refill time of greater than 3 seconds. It was taken to the veterinary ICU and shortly after went into respiratory arrest, and CPR was not performed due to a do not resuscitate status. The 30 and 60 minute post infusion free PHT levels were 1.60 ad 0.92 mg/dl respectively in this case, which was within the study target range. A necropsy (animal post mortem) was requested but was not approved by the owners.

Discussion

The 64% responder rate for the 12 hour endpoint in the BZD/FOS treated dogs is similar to the 55.8% reported for people in the Trieman study. ⁴ This study, therefore, met its aims by demonstrating that response to FOS of subjects with CSE is similar to that reported for HSE. The results of this study show that FOS is an effective drug for the treatment of CSE in dogs, with relatively few adverse events except some vomiting shortly after administration. At both 2 and 12 hours there was a better responder rate for FOS vs. PBO, and statistical significance can potentially be found with a limited number of CSE cases in a randomized controlled clinical trial.

Study limitations include lack of EEG for non-convulsive SE, not all cases were managed by the same clinician, and consequently there was some variation in treatments received (other than FOS), and MRI was not performed for all patients and therefore a final diagnosis was not precisely determined in all cases. Even with the lack of EEG, the results of CSE studies can be directly translatable to convulsive HSE as evidenced by the very similar response rate to FOS for SE in both species. Additionally it is not yet clear at what hour responder rate is the best translational model endpoint for CSE, though the 12 hour endpoint for CSE does appear to be a reasonable clinical endpoint, as some of the dogs who were responders at 6 hours, had additional seizures between 6 and 12 hours.

Our results are significant because: 1) the study offers proof of concept for future testing of the efficacy and safety of non-FDA approved, investigational agents discovered in experimental models, which may be significantly superior in efficacy or neuro-protection to present treatments for humans; and 2) a new platform to speed the translation of experimental agents to human use has been developed. In this study we established a national network of four veterinary emergency care departments to evaluate FOS in CSE.. We have just expanded this network to seven veterinary centers and have future plans to test novel drugs in CSE, that have shown promise in rodent SE models, to try to initiate a process of efficient results for potential translation into HSE studies. If a drug has promising results in the canine platform, this could speed its development into human clinical trials, and if it is ineffective in canine SE, money could be saved in not pursuing development for HSE.