• Open Access

Epilepsy in Border Collies: Clinical Manifestation, Outcome, and Mode of Inheritance


  • Parts of the study were presented at the “17. Jahrestagung der Fachgruppe – Innere Medizin und Klinische Labordiagnostik – der DVG,” January 31–Feburary 1, 2009, Berlin, Germany.

Corresponding author: Andrea Fischer, Diplomate ACVIM (Neurology), Diplomate ECVN, Section of Neurology, Clinic of Small Animal Medicine, Ludwig-Maximilians-University Munich, Veterinärstr. 13, 80539 Munich, Germany; e-mail: andrea.fischer@med.vetmed.uni-muenchen.de.


Background: There is a lack of data on idiopathic epilepsy (IE) in Border Collies (BCs) in the veterinary literature.

Hypothesis: Genetic epilepsy occurs in BCs and is frequently characterized by a severe clinical course and poor response to medical treatment.

Animals: Forty-nine BCs diagnosed with IE.

Methods: Medical records, seizure data, treatment data, and pedigree information of affected dogs were collected. Cases were classified phenotypically as affected or not affected; mild, moderate, or severe clinical course; active epilepsy (AE) or remission; and drug resistant or not drug resistant.

Results: Clinical manifestations were classified as having a moderate (33%) or severe clinical course (49%), characterized by a high prevalence of cluster seizures and status epilepticus. Survival time was significantly decreased in dogs <2 years of age at seizure onset, and in dogs with a severe clinical course. Drug resistance was apparent in 71% of 24 dogs treated with ≥2 antiepileptic drugs. The epilepsy remission rate was 18%. Median age at onset was significantly higher and initial seizure frequency was significantly lower in dogs with remission compared with dogs with AE. Pedigree analyses indicated a strong genetic founder effect in the appearance of epilepsy, resembling autosomal recessive inheritance.

Conclusion and Clinical Importance: The present study confirms the occurrence of genetically mediated epilepsy with a frequent severe clinical course and drug resistance in BCs. The results provide information about the long-term prognosis of IE in BCs for veterinarians and concerned owners, and may benefit breeders as well.


active epilepsy


antiepileptic drug


Border Collie


confidence interval


cluster seizure


cerebrospinal fluid


discrete seizure


idiopathic epilepsy


International League Against Epilepsy


magnetic resonance imaging


neuronal ceroid lipofuscinosis


status epilepticus

Idiopathic epilepsy (IE) is a common breed-related neurological disorder in contemporary small animal medicine and often may evolve into a life-threatening neurological emergency. Because of a high prevalence of IE in certain breeds within the past few years, familial predisposition and genetic components of disease origin are suspected for many purebred dogs. Consequently, several severely affected breeds have been studied for hereditary and clinical characteristics. Although results of segregation analyses have been mostly consistent with a form of recessive inheritance, in the majority of investigated breeds the exact mode of inheritance has not been determined, and variable clinical manifestations have been described.1–10 Variable genes, mutations, or complex interactions of potentially unknown factors may be involved in the development of the disease among different breeds or even among different families of the same breed.10 Therefore, results of previous studies have only limited application, and each breed must be investigated separately. Until now, identification of a causative gene mutation in canine epilepsy has only been successful in a familial juvenile epilepsy in Lagotto Romagnolo dogs (H Lohi, personal communication) and in the case of autosomal recessive progressive myoclonus epilepsy in miniature wirehaired Dachshunds (Epm2b, tandem repeat expansion).11

Thus, the current study was undertaken against a background of an increasing number of Border Collies (BCs) presenting with severe epileptic seizures that are poorly controlled with antiepileptic therapy. Additionally, although the cases often are serious, no detailed data about the disease's origin or clinical characteristics in this breed have been published previously.

Material and Methods

Study Goal

The aim of the study was to describe the clinical manifestations of IE in BCs with regard to clinical course, response to medical treatment, and outcome. Additionally, heredity was investigated by means of pedigree analyses. The study was designed as a cross-sectional study and took place in Germany over a 24-month time period.

Data Collection

Epileptic BCs, living and dead, were identified through requests placed at the clinic's web-page and made to German BC breeding clubs. Cooperating veterinary neurologists were asked to forward information to owners of BCs diagnosed with IE. BCs diagnosed with IE at the study center (Section of Neurology, Munich, Germany) were evaluated retrospectively (2000–2006) and prospectively (2007–2008).


Participating owners completed a detailed questionnaire. The anamnestic assessment included information about the dog's habitat, signalment, medical treatment, diagnostic test results, pedigree information, and seizure data, such as age at seizure onset, seizure type, duration, diurnal rhythm, possible environmental trigger, effect of neutering, and the appearance of pre- and postictal signs. Seizure frequency was assessed by counting a dog's seizure days (a day on which the dog had 1 or more seizures). Copies of medical records were reviewed for all included dogs, and survey information was specified through follow-up telephone interviews with all participating owners and referring veterinarians. Several questions focused on treatment onset, serum concentrations of antiepileptic drugs (AED), adverse-effects of treatment, and occurrence of status epilepticus (SE) or cluster seizures (CSs). Age at death, cause of death, and survival time (duration from seizure onset until death) were recorded for deceased dogs.

Definition of IE

IE was defined as recurrent seizures (≥2 seizure days at least 4 weeks apart) with an onset between 6 months and 5 years of age in dogs with otherwise normal physical, laboratory, and neurological characteristics upon examination. Requested minimal laboratory investigations included a CBC and biochemical profile (ie electrolytes, glucose, creatinine, urea, albumin, total protein, creatine kinase, and liver enzyme activities). Magnetic resonance imaging (MRI) and cerebrospinal fluid (CSF) analysis were requested if age at seizure onset was <6 months or >5 years of age, in concordance with other studies of canine IE and consensus of current clinical experience.6,7,12 Dogs that failed to meet all inclusion criteria were only included if a 1st- or 2nd-degree relative was afflicted by IE or if seizures had occurred for at least 2 years without interictal neurological abnormalities. Testing for neuronal ceroid lipofuscinosis (NCL) was requested if progressive loss of vision or unexplained changes (eg, aggression) were reported. Presence of any initial precipitating event (eg, head trauma leading to loss of consciousness) without subsequent brain imaging, an identified brain lesion, or observational data consisting of <10 h/d (ie, incomplete seizure monitoring) resulted in exclusion from the study.

Seizure Classification

Seizures were classified according to guidelines suggested by Berendt and Gram,13 Licht et al,10 and the Commission on Classification and Terminology of the International League Against Epilepsy (ILAE).14 Seizures that were generalized from the onset were classified as primary generalized seizures (ie, bilateral, symmetrical motor signs). Focal seizures (ie, movements limited to a single part of the body, unusual changes, or autonomic signs) were defined as simple when consciousness was not impaired or complex when consciousness was impaired. Focal onset seizures were defined by focal onset (complex or simple) and secondary generalization. Consciousness was assessed based on the dog's response to the owner during a seizure.

Definition of Clinical Course

Active epilepsy (AE) was assumed if a dog had had ≥1 epileptic seizure in the last year of the study or in the year preceding death. Further subgrouping of AE into mild (only discrete seizures [DS]), moderate (≥1 episode of CSs), and severe courses (≥1 episode of SE) was conducted according to seizure severity. CSs were defined as >1 seizure within 24 hours. SE was defined as seizure activity lasting ≥5 minutes or CSs without full recovery of consciousness between seizures.15 Remission was defined as seizure absence for ≥1 year (1st grade) or ≥2 years (2nd grade). “Remission with treatment” was defined as remission under current AED therapy and “remission without treatment” as remission under discontinued AED therapy at the time of ascertainment. Definitions of remission and AE were formulated in accordance with ILAE guidelines on epidemiology and prognosis16 and a previous study of dogs,17 but were modified for this study. Drug resistance was classified as ≥1 seizure day per month despite therapy with 2 appropriate AEDs for at least 6 months with adequate serum concentration. Drug resistance was assessed only in dogs for which serum concentrations of AED were monitored. Serum phenobarbital concentrations ≥20 μg/mL, serum potassium bromide concentrations ≥1,500 μg/mL, and serum zonisamide concentrations >10 μg/mL were considered adequate. Serum concentrations of levetiracetam and carbamazepine were not measured because of lack of established therapeutic ranges in veterinary medicine. In addition, levetiracetam has a very high margin of safety and no clear relationship between serum concentration of levetiracetam and efficiency is documented. Nevertheless, for dogs treated with levetiracetam a minimum dosage of 20 mg/kg of body weight q8h was required.18

Pedigree Analyses

Pedigrees of IE-affected dogs were matched for the appearance of common ancestors, and subsequent classification into several subpopulations, each consisting of a group of dogs referable to a common sire, was conducted. Dogs that were ≥7 years old and were observed ≥10 h/d without any observed seizure activity during the dog's lifetime were considered as “not affected.” Dogs <7 years old without any seizure history were considered as “probably not affected.”

Statistical Analysis

Possible predictors of the occurrence of remission or drug resistance and potential influences on age at death and age at onset were analyzed statistically with Fisher's exact test (for 2-tailed results) for categorical variables and with the Mann-Whitney U-test (2-tailed exact significance) for continuous variables. To test association between survival time (calculated from seizure onset) and categorical predictors, Kaplan-Meier curves (evaluated by log rank test) by 2 different methods were conducted. In the 1st method (A), only epilepsy-related deaths were considered events (deaths because of causes other than epilepsy were censored). In the 2nd method (B), all deaths, regardless of cause, were considered events. Dogs that were still alive at the time of analysis were censored in both methods. A Cox regression was applied to test association for continuous predictors. Statistical significance was defined at P≤ .05 and all data analyses were performed by SPSS (version 16.).


Study Population

Data from 90 BCs with a history of seizures were collected. Forty-nine dogs met the given criteria for diagnosis of IE and were enrolled. The remaining 41 dogs were excluded because of mild hydrocephalus (n = 1), canine distemper (n = 1), episodes of unusual behavior not clearly consistent with epilepsy (n = 1), experiencing only 1 seizure (n = 3), severe head trauma in medical history without subsequent brain imaging (n = 1), insufficient diagnostic features (n = 14), notification after data collection was completed (n = 8), or lack of owner response (n = 12). Two dogs that had sustained head trauma were included, because of the absence of brain lesions on MRI (2 mm contiguous slices in transverse, sagittal, and dorsal orientations; pre- and postcontrast T1, T2, fluid attenuated inversion recovery, gradient recall echo); normal CSF analysis; and presence of affected dogs in their kinship. Genetic testing for NCL was negative in 2 dogs that were tested at the owners' request.

Seizure Database

Age at seizure onset was between 1 and 5 years in 36 dogs (74%), ≤1 year in 9 dogs (18%), and >5 years of age in 4 dogs (8%) (Fig 1). Median age at seizure onset was 2.37 years (range, 0.41–8.16; 95% confidence interval [CI], 2.24–3.29), regardless of sex (P= .290) and reproductive status (P= .734). Affected dogs were equally distributed between males (24 males; 12 neutered) and females (25 females; 18 neutered). Of 15 dogs neutered after seizure onset, only 1 owner (7%; 1/15) reported mild improvement in seizure frequency, whereas 13 dogs (86%; 13/15) exhibited no positive trends, and in 1 dog (7%; 1/15) the seizure frequency increased. Four dogs (neutered before seizure onset) experienced their 1st seizure at the time of neutering (2 dogs experienced seizures on the same day and 2 dogs did so within 1 week after neutering). Seizure frequency differed individually and varied from 1 seizure day per year to multiple seizure days per month or per week. No environmental trigger (eg, moon phase, season, feeding habits, and heat cycle) could be identified, and most of the owners declared unspecific stress or anxiety (eg, variation in daily routine) as possible seizure stimuli or reported seizures as totally accidental events. Twenty-eight dogs (57%) had preictal symptoms such as vomiting (n = 1), aggressive behavior (n = 1), salivation (n = 8), seeking the owner's attention (n = 16), or restlessness (n = 20); these were normally observed up to 30 minutes before seizure onset. In 18% (5/28) of dogs, owners recognized preictal signs 1–2 days before seizure onset, consisting of lameness of 1 forelimb (n = 2) and decreased reaction to known commands (n = 3). Diurnal rhythm was dominated by overnight and early morning seizures. In addition, most owners reported that seizures occurred during sleep. Seizures occurring during the daytime occurred most often when the dog was in a resting position. In all 49 IE-affected dogs, the predominant seizure type was generalized. A focal onset was observed in 38 dogs (78%), mostly characterized by staring into space for seconds and lateral head turn or opisthotonos (focal onset seizure). In 4 dogs (8%), seizures were classified as primarily generalized because of absence of any initial motor, behavioral, or autonomic signs. In 7 dogs (14%), seizures remained unclassified, as seizure onset had not been observed. Generalization was dominated by lateral posture, loss of consciousness, tonic-clonic movements of the limbs, salivation, repetitive jaw movements, and urination. Twenty-two dogs (45%) occasionally had isolated focal motor seizures without secondary generalization, usually manifested as sudden uncontrolled head or face twitching mostly associated with impaired consciousness (complex focal seizures). Postictal phase was predominantly distinguished by restlessness (n = 41), thirst (n = 28), hunger (n = 20), lethargy (n = 30), deep sleep (n = 4), aggression (n = 5), vomiting (n = 3), and brief postictal blindness (n = 20); frequently the dogs required long time periods (6 hours to several days) to achieve complete physical and mental recovery. Additionally, several owners reported that their dogs were more sensitive to heat than other BCs in the household, and some dogs occasionally suffered from alternating lameness and movement disorders of unknown origin; however, these observations were not further elucidated in the present study.

Figure 1.

 Distribution of male and female dogs with respect to age at seizure onset.

AED Treatment

Continuous AED therapy was administered to 38 patients (78%); the remaining 11 patients (22%) never received AEDs or received them at irregular time intervals. Five treated dogs (10%) were excluded from subsequent statistical analysis because of lack of adequate treatment information (eg, serum drug concentrations were not monitored). Of the remaining 33 treated dogs (67%), 9 were treated with phenobarbital, 18 were treated with phenobarbital and potassium bromide, and 6 were treated ancillary to phenobarbital and potassium bromide with levetiracetam (n = 1), zonisamide (n = 4), or carbamazepine (n = 1). Twenty-two owners (67%; 22/33) reported adverse effects (eg, sedation, increased appetite, and movement disorders) from AED treatment. These effects generally appeared during the 1st weeks of treatment and were normally mild to moderate. However, 3 owners reported alterations in gait and locomotion (eg, generalized ataxia, paresis) in their dogs after potassium bromide administration, which led to discontinuation of medication after a few days. Decreased working ability (eg, sheep dog trials, tests of agility) attributable to the sedative effect of AEDs was reported in 18 dogs (55%; 18/33). Of the dogs treated with at least 2 AEDs (n = 24), 17 were classified as drug resistant (71%; 17/24). No association was found between medical refractoriness and potential risk factors such as sex (P= .659), reproductive status (P= .191), age at onset (P= .494), or number of seizures before treatment onset (P= .264).

Clinical Course

In the study population, 40 dogs (82%) were afflicted by AE and 9 dogs (18%) went into remission (Table 1). Of the 40 dogs with AE, 40% (16/40) had a moderate clinical course and 60% (24/40) had a severe clinical course. Including preceding seizure severity in dogs in remission, 45% (22/49) of all 49 BCs had ≥1 CSs, 4% (2/49) had ≥1 SE and 49% (24/49) had a history of both (CSs and SE). In summary, CSs occurred in 94% (46/49; CSs + CSs and SE) and SE in 53% (26/49; SE + CSs and SE) of all 49 affected BCs. SE occurred in 23% (6/26) as CSs without normal consciousness occurring between seizures, 8% (2/26) had seizures lasting >5 minutes and 69% (18/26) had a history of both. Only 1 dog experienced only DSs and never had any episodes of CSs or SE. This dog subsequently went into remission at the time the study was conducted. Therefore, no dog was classified as having a mild clinical course. Remission occurred independently of preceding seizure severity, as 8 dogs in remission (89%; 8/9) also had a history of CSs, SE, or both. Four dogs fulfilled criteria of 1st-grade remission and 5 dogs were classified as in 2nd-grade remission. “Remission with treatment” was evident in 7 dogs. Of these, 5 dogs were treated with phenobarbital and 2 dogs also were treated with potassium bromide. The remaining 2 dogs received phenobarbital for just a few months at seizure onset, and medication was discontinued thereafter (“remission without treatment”). These 2 BCs in remission without treatment continued to have multiple seizures until seizures stopped spontaneously after 6 months and 2 years, respectively. Brain MRI and CSF analysis was normal in both dogs. Comparison between dogs in AE and dogs in remission identified significant differences in age at seizure onset (P= .029) and age at death (P= .032) (Table 2). Furthermore, initial seizure frequency (occurring within the 1st 6 months of onset) was significantly lower in dogs in remission compared with dogs with continuing AE (P= .032).

Table 1.   Clinical courses and seizure severity of IE in BCs.
Clinical CourseSubgroupsSeverity of
(Preceding)a Seizures
Number of Dogs
DSs, nCSs, nSE, nCSs and SE, nn%b%c
  • a

    For dogs in remission, severity of preceding seizures was assessed.

  • b

    Scale basis: all affected dogs (n = 49).

  • c

    Scale basis: dogs in AE (n = 40) and dogs in remission (n = 9), respectively.

  • DSs, discrete seizures; CSs, cluster seizures; SE, status epilepticus; CSs and SE, history of both.

Active epilepsy     4082%100%
Mild course
Moderate course161633%40%
Severe course1232449%60%
Remission     918%100%
First grade448%44%
 Second grade1211510%56%
Overall prevalencen122224   
of DSs, CSs, and SE%b2%45%4%49%   
Table 2.   Association between clinical variables and the course of epilepsy.
VariableCourse of EpilepsyP-Value
Active EpilepsyRemission
No. of dogs4082b918b 
Therapy (AED)
 With treatment26c657781.000
 Without treatment923222
Seizure severityd
 Status epilepticus24602220.064
 Cluster seizures39987780.083
 Median (95% CI)Median (95% CI) 
  • a

    Scale basis: number of dogs in each group (AE = 40; remission = 9).

  • b

    Scale basis: all affected dogs (n = 49).

  • c

    c Five treated dogs were excluded from statistical analysis owing to lack of adequate treatment information.

  • dIn dogs under remission, severity of preceding seizures was evaluated.

  • CI, confidence interval; AED, antiepileptic drug.

Age at onset2.28 (1.95–2.99)3.71 (2.42–5.77)0.029*
Age at death4.60 (0.95–12.40)10.81 (9.79–11.84)0.032*


At the time of analysis, 26 dogs (53%) were alive and the remaining 23 (47%) were dead. Six dogs (26%; 6/23) died from non-seizure-related causes, and in 17 dogs (74%; 17/23) death was directly attributable to epilepsy. All epilepsy-related deaths, except 2 dogs that died in SE, were the result of euthanasia. The median age at death was 5.17 years (range, 0.95–12.4; 95% CI, 3.58–6.54). In dogs whose death was epilepsy related, median age at death was significantly lower when compared with dogs that died of other causes (P= .002). No effects on age at death of sex (P= .449) or reproductive status (P= .875) were evident.

The median survival time was 2.07 years (range, 0.3–6.58; 95% CI, 1.68–3.28). Survival times, calculated with either method ([A] or [B]), indicated significant differences in those dogs of a young age at seizure onset ([A] P= .003 and [B] P= .044), and with a severe epilepsy course ([A] P= .002 and [B] P= .011) (Figs 2, 3 and 4, 5). No effects on survival time of sex ([A] P= .937 and [B] P= .689), reproductive status ([A] P= .861 and [B] P= .800) or treatment ([A] P= .927 and [B] P= .322) were found. The respective number of AEDs had no positive effects (1 AED–3 AEDs; [A] P= .378 and [B] P= .217). Cox regression analyses for age at onset (continuous data) and body weight indicated no significant effects (P > .05).

Figure 2and3.

 Kaplan-Meier curves (method [A] and [B]) depicting significant reduced survival time in dogs of a young age (<2 years) at seizure onset ([A] P= . 003 and [B] P= .044). Black line, age at onset ≥2 years; broken line, age at onset <2 years; + censored dogs ([A], nonepilepsy-related deaths and dogs that were still alive at time of analysis; [B], only dogs that were still alive at time of analysis).

Figure 4and5.

 Kaplan-Meier curve (method [A] and [B]) depicting significant differences in survival time with respect to clinical course ([A] P= .002 and [B] P= .011). Gray line, remission; black line, moderate course; broken line, severe course, + censored dogs ([A]: nonepilepsy-related deaths and dogs that were still alive at time of analysis; [B]: only dogs that were still alive at time of analysis).

Pedigree Analyses

Pedigrees of 43 affected dogs were analyzed; the remaining 6 dogs had insufficient pedigree data. Twenty-nine affected dogs, from 2 subpopulations, ultimately shared a common founder. An increased manifestation of epilepsy in some subpopulations and the repeated occurrence of the same sires in different subpopulations indicated a strong genetic basis for the condition in this breed. Most affected dogs originated from nonaffected parents (Fig 6).

Figure 6.

 One representative subpopulation of Border Collies including 14 affected dogs (11 affected dogs were referable to a common sire). ○, female; □, male; black, affected with IE; white, dogs aged ≥7 years without any seizure history (not affected); gray, dogs aged <7 years without any seizure history (probably not affected); underlined boxes, affected dogs that were referable to a common sire.


The present study represents the first detailed description of clinical manifestation and inheritance of IE in BCs.

Diagnosing IE still is challenging, because all other possible seizure etiologies must first be excluded. In the present study, MRI examination and CSF analysis were not routinely requested for BCs that experienced their 1st seizure between the age of 6 months and 5 years, the main reason being financial constraints and the lack of interest of some owners because MRI findings may not influence the treatment of epilepsy. The mentioned age interval is indicative for the onset of IE in neurologically normal dogs. Several investigators have demonstrated a low diagnostic yield of MRI and CSF analysis in dogs younger than 5–6 years with epileptic seizures and normal interictal neurological examination.12,19 Consequently, inclusion criteria were defined in accordance with other recent studies of canine IE and the oral consensus statement of the 25th ACVIM Forum 2007 in Seattle.6,7 To further decrease the risk of including single cases of poisoned dogs or dogs with acute CNS diseases, only BCs with ≥2 seizure days and a 4 week interictal period were included.

In our study population, age at onset was between 1 and 5 years in 36 dogs (regular onset) and <1 year of age in 9 dogs (early onset). Four dogs with age at seizure onset >5 years were included. These 4 dogs had normal neurological and laboratory examination findings, and seizures continued for ≥2 years in absence of any interictal neurological dysfunction; for 2 dogs relatives also were known to be affected with IE. Brain imaging and CSF analysis were performed in only 2 of the 4 dogs. The owners of the remaining dogs declined further diagnostic investigations because of an acceptable seizure frequency or, for fear of anesthetic complications and of loss of seizure control. The possibility of late onset epilepsy, although rare, was considered as already described in other studies.7,10,20 Yet, even with high definition MRI we may not be able to exclude cryptogenic causes of epilepsy with absolute certainty in these 4 dogs. Neoplasia was unlikely because of the long individual seizure history and normal interictal neurological examination.

The predominant seizure type was focal onset with secondary generalization (focal onset seizures). The majority of dogs with occasional focal motor seizures underwent a complete diagnostic evaluation that failed to identify structural brain disease. In the past, the general consensus has been that partial seizures are more likely to indicate an underlying structural brain lesion,12,21–23 but there recently has been increasing evidence in canine epilepsy studies that focal seizures can be associated with IE.2,6,7,10 Additionally, focal seizures because of IE syndromes have been described in humans.24

A drug resistant fraction of 71% was found among BCs that were treated with at least 2 AEDs (17/24). Approximately 20–30% of treated dogs have been said to respond poorly to treatment with phenobarbital and potassium bromide, but this observation did not correspond with our data.25 Unfortunately, no consistent definition of drug resistance exists, and previous studies in dogs used different criteria to assess medical response, which may explain the variable results.25–30 Although research on epilepsy has advanced substantially in the past decade, the origin of drug resistance or observed heterogeneous response to AEDs among individuals has not been adequately explained yet.31 We did not identify any clinical or environmental predictor for the development of drug resistance in BCs in the present study. Furthermore, the definition of drug resistance is challenging, because it is a multifaceted phenomenon. Several factors must be considered, such as the number of AED failures, inadequate control of seizure frequency, and duration of unresponsiveness to medication.32,33 These factors may apply to humans as well. No single preferred definition for intractable epilepsy exists, even though its definition is of paramount importance with respect to the indication of neurosurgical intervention.32–34 Mainly, inadequate seizure frequency control (which in humans often is referred as seizures' impact on quality of life34) is difficult to define in dogs. Noncompliance of owners in administration of medication or recurrent discontinuation of medication could further complicate the problem and is described in humans to be a common reason for therapeutic failure, as well as the most frequent cause of SE in patients with chronic epilepsy.35 To minimize the risk of incorrectly classifying a case as drug resistant when owner compliance was limited, we required documented serum drug concentrations within therapeutic ranges for each dog. The impact of repeated discontinuation of medication on the occurrence of drug resistance in BCs was not accurately analyzed in our study. However, a lack of standardized definitions impedes comparison of results among epilepsy studies in dogs, and a future aim should be to define universally valid guidelines for conducting epidemiologic studies.

The severity of disease was dominated by moderate (33%) and severe clinical courses (49%), defined by the occurrence of CSs and SE. Only limited data on the prevalence of CSs and SE are available for IE in other breeds.1–10 Compared with a study of IE in 45 English Springer Spaniels (38% of dogs had CSs), the overall prevalence of CSs in BCs was noticeably higher (94%).7 The overall proportion of SE in BCs (53%) correlated with results of a study of 32 dogs with IE in several breeds (59% of dogs had at least 1 episode of SE).36

Our study population exhibited an 18% remission rate, although epilepsy remission or seizure freedom has been rarely documented in dogs,2,17,36,37 especially for spontaneous remitted epilepsy (“remission without treatment”).17,37 The 2 BCs with spontaneous remitted epilepsy suffered from recurrent epileptic seizures over a longer time period. Nonepileptic seizure events (eg, syncope, vestibular episodes) were excluded for both dogs based on detailed seizure description. We further excluded symptomatic or reactive causes for the dogs' seizures by MRI, CSF, and laboratory evaluation. Based on these facts, we feel that the diagnosis of IE in these 2 BCs in “remission without treatment” is possible and furthermore is supported by the occurrence of IE-afflicted dogs in their kinship. One might expect that remission is associated with a preceding mild epilepsy course, but in our study population, 89% (8/9) of dogs in remission had a history of CSs, SE, or both. This is in contrast with studies of human patients, which reported SE, among other factors, as a negative predictor for the occurrence of remission in intractable epilepsy.38 We found that increased patient age at seizure onset and low seizure frequency in the beginning of epilepsy may positively impact the occurrence of remission in BCs, as already described in humans.34 However, our analyzed sample size was small (9 dogs in remission) and more epidemiologic studies are needed to further document the nature and occurrence of remission of epilepsy in dogs.

Survival time (calculated with either method) was significantly decreased in dogs of young age at seizure onset (<2 years) and in dogs with severe clinical courses. Evaluation of survival time in pets is controversial, because normally most affected animals are euthanized at the owners' request and thus do not die naturally. Nevertheless, these variables might be useful factors for long-term prognosis in affected BCs, although in this study only 2 dogs died naturally in SE. This is in contrast with sudden unexpected death in humans. The number of AEDs used did not significantly affect survival time. This finding suggests that a given dog's epilepsy was, in general, either easy to control or (more likely for BCs) difficult to control in concordance with human studies.31,39

Pedigree analyses indicated a strong genetic component for disease origin and 29 dogs ultimately were shown to share a common ancestor. Parents were not affected in most cases, suggesting a recessive mode of inheritance. However, we cannot exclude a more complex inheritance pattern resembling recessive inheritance. Also, the study may have been biased, encouraging owners of severely affected dogs to participate whereas owners of mildly affected dogs chose not to participate. An equal sex distribution among affected dogs made a sex-linked chromosomal trait improbable. Previous studies used segregation analyses to determine mode of inheritance of IE in several breeds,3,4,6–8,10,40 but such analyses require knowledge of seizure status of a large number of related dogs and such assignment is susceptible to incorrect phenotypic classification resulting from insufficient data on distantly related dogs. Much data (eg, common founder effect, recessive versus dominant trait) can be gained using family trees as in our study. Use of stringent inclusion criteria and subgrouping into different phenotypes (eg, remission or AE or early, regular, or late onset) may promote an ongoing genome-wide association study.

Frequent occurrence of CSs and SE, a high proportion of drug resistance, nocturnal preponderance of seizures, and frequently reported head deviation at seizure onset suggest some parallels between IE in BCs and seizures of frontal lobe origin in humans, but other characteristic features, such as rapid recovery without postictal disturbance, are missing.41 However, to support these findings and identify the origin of IE in BCs, further investigations with electroencephalogram monitoring and postmortem examination are essential.

Our study had some limitations as a result of data collection by means of questionnaires distributed to owners of affected dogs. Results largely relied on retrospective data evaluation and the subjective perceptions of owners, and the reliability of such results represents the most critical aspect of studies such as this one. Nevertheless, we did attempt to objectify the data through personal interviews with each owner and with the referring veterinarian, establishment of objective criteria for case classification, and collection of complete data by excluding dogs that were observed <10 h/d. Furthermore, tests for NCL only were conducted in 2 dogs (with negative results in both) but this diagnosis appears unlikely for the remaining study population because it is mostly characterized by progressive blindness and behavioral abnormalities, with seizures occurring only rarely.42

In conclusion, results of the present study confirm that IE with a strong genetic component, severe clinical course, and frequent drug resistance frequently occurs in German BCs. Early seizure onset and high initial seizure frequency were associated with a worse prognosis. Nevertheless, remission may occur in a subset of patients.


The 1st author (Velia Huelsmeyer) is supported by a dissertation elite grant (Graduiertenstipendium nach dem Bayerischen Eliteförderungsgesetz—Postgraduate excellence program).