• Open Access

A Longitudinal Study of Survival in Belgian Shepherds with Genetic Epilepsy


  • The material has not previously been presented

Corresponding authors: Christina H. Gulløv and Mette Berendt, Department of Veterinary Clinical and Animal Sciences, University of Copenhagen, Dyrlægevej 16, Frederiksberg C, DK-1870, Denmark; Department of Large Animal Sciences, University of Copenhagen, Grønnegårdsvej 8, Frederiksberg C DK-1870, Denmark; e-mail: chhg@life.ku.dk; mbe@life.ku.dk



Belgian Shepherds have focal genetic epilepsy. The prevalence of epilepsy has been estimated as 9.5% in the breed and as 33% in the family investigated. Dogs with epilepsy might have an increased risk of premature death.


To investigate survival and selected risk factors for premature death in a Belgian Shepherd family with genetic epilepsy.


One hundred ninety-nine related Belgian Shepherds.


Longitudinal observational study, 2009–2011. Follow-up telephone interviews were all conducted using a structured questionnaire addressing epilepsy, including seizure history and phenomenology, possible remission, possible death, and cause of death.


The life span of epileptic dogs was not significantly shortened by the presence of epilepsy (P = .87). Epilepsy was the predominant cause of death in the population (19/75 = 25%) and epilepsy-related deaths accounted for 70% (19/27) of all deaths in the group of dogs with epilepsy. Two probable sudden unexpected deaths related to epilepsy occurred in dogs with generalized seizures. Cluster seizures occurred in 33% (17/51) but did not significantly influence the life span of epileptic dogs. Dogs with epilepsy had an epilepsy remission proportion of 13.7%.

Conclusion and Clinical Importance

The Belgian Shepherds investigated in the present study display a focal genetic epilepsy with an overall benign course. The life span was not significantly affected by the presence of epilepsy.


antiepileptic drug


Federations Cynologie International

Belgian Shepherds suffer from focal genetic epilepsy. The prevalence of epilepsy has been estimated to 9.5% in the breed.[1-6]

Dogs suffering from epilepsy have an increased risk of premature death compared with the general dog population.[7-9] This also applies for people with epilepsy for whom the risk of death is highest shortly after onset of seizures.[10, 11]

The median survival time after index seizure is 2.07 years in Border Collies and 2.3 years in a population of different purebred and mixed-breed dogs.[7-9] The life expectancy for Irish Wolfhounds is shortened by almost 2 years in epileptic dogs compared with seizure-free relatives. The same study reported that more than 60% of the total number of deaths in the population were related to epilepsy.[12]

Factors such as sex, seizure onset, seizure frequency, and seizure control might influence the life span of dogs with epilepsy. In one study, bitches lived longer with epilepsy compared with males, and daily treatment with antiepileptic drugs did not influence the life span of dogs with epilepsy compared with dogs with epilepsy left untreated.[7] In Border Collies, seizure onset before the age of 2 years has been shown to significantly decrease survival time.[9]

Sudden unexpected death in epilepsy (SUDEP) is the most common cause of death directly related to epilepsy in humans, with an increased frequency of unexpected deaths up to 24 times compared with unexpected deaths in nonepileptics.[13, 14] The mechanisms of SUDEP remain elusive, and it seems unlikely that the identification of a single mechanism will explain all incidents, as SUDEP is more likely a result of several predisposing and triggering factors.[14, 15] Poorly controlled chronic epilepsy with tonic-clonic seizures seems to be the most well-defined risk factor for SUDEP in humans.[14, 15] Suspected SUDEP occurs in epileptic dogs.[11]

A family of 199 Belgian Shepherds with a high prevalence of epilepsy (33%) and seizure phenomenology dominated by focal seizures with and without secondary generalization was described in 2009.[2] The investigation provided evidence for simple Mendelian inheritance of the disease. In 2009, 66 dogs in the family were epilepsy-positive and 116 were epilepsy-negative while 17 had unknown status. The history and the semiology of the epileptic seizures displayed in the family did not differ from what previously has been described for epilepsy as it occurs in Belgian Shepherds in general.[1-3, 5]

The objective of the present study was to further characterize the clinical phenotype of epilepsy in Belgian Shepherds by following the dog family described in 2009 prospectively over time. It was investigated if epilepsy in Belgian Shepherds is associated with an increased risk of premature death. In addition, the possible influence of sex and age of onset on survival after first seizure was investigated. Finally it was an aim to describe the cause of death, survival after first seizure, number of seizures, and possible remission of epilepsy.

Materials and Methods

Study Design and Study Population

An extended dog family (pedigree) of 199 Belgian Shepherds with an accumulation of individuals suffering from genetic epilepsy was followed in a longitudinal study from 2009 to 2011. The family was first described in 2009.[2] In 2009 the owners of the 199 dogs belonging to the pedigree (living and deceased) were contacted by phone and interviewed using a structured standardized questionnaire developed to detect signs of epilepsy, by addressing seizure history, seizure phenomenology, and interictal history. All living dogs were invited to participate in a detailed clinical evaluation at the Department of Small Animal Clinical Sciences, University of Copenhagen, including clinical and neurological examination, full hematological and biochemical profile, thyroid profile, and ECG. By indication, dogs were offered a MRI scan of the brain. Sixty-six dogs were evaluated to be epilepsy-positive. In January 2009, 40 epilepsy-positive dogs were alive while 26 were deceased.

For the present study, a follow-up investigation of the 199 dogs belonging to the pedigree was conducted in spring 2011. The 199 dog owners were contacted by phone and information was collected about status of epilepsy, antiepileptic treatment (AED), remission, time and cause of death, and occurrences of new cases of epilepsy in the population.

In this study, dogs were excluded if the owner was not able to complete an oral interview in Danish. This was done because of the need for a uniform interview,


All owners were interviewed over the phone during a 20–30 minute call. This included questions regarding dog identification, signs of epilepsy, possible death, and cause of death. In addition, owners of dogs identified epilepsy-positive in the 2009 investigation were asked about age of onset, number of seizures, cluster seizures, status epilepticus, and possible remission of epilepsy. This also applied for owners of dogs that were epilepsy-negative in the 2009 investigation, and for which the owner reported development of signs of epilepsy between 2009 and 2011. To ensure standardized interviews, a structured questionnaire was used and all the interviews were conducted by the primary investigator.


We used a definition of remission of epilepsy that has previously been used in dogs.[7] “Spontaneous remission of epilepsy” was defined as no seizures for >3 years without antiepileptic medication. “Epilepsy in remission with treatment” was defined as no seizures for >3 years while being given AEDs. Both terms imply that the dog was seizure free at the time of the investigation in 2011.

Status epilepticus was defined as a single epileptic seizure lasting more than 30 minutes or a series of seizures during which the normal level of brain function is not regained between ictal events in a period longer than 30 minutes.[16]

Cluster seizures were defined as an incidence of seizures within a given period, usually a day or a few days, that exceeds the average incidence for the dog.[17]

Probable sudden unexpected death in epilepsy (SUDEP) was defined as “sudden, unexpected, witnessed or unwitnessed, nontraumatic and nondrowning death in epilepsy, with or without evidence of a seizure and excluding documented status epilepticus”.[18, 19]

Statistical Analysis

A descriptive analysis was initially performed for sex, seizure onset, age at death, cause of death, survival time after index seizure, number of seizures, status epilepticus, cluster seizures, and remission of epilepsy. As this study was a follow-up of a population of dogs that had previously been investigated, only dogs that had remained in the same category as they initially belonged to were included in the analysis. Dogs that had developed signs of epilepsy in the intervening period and dogs for which the status could not be confirmed were thus excluded.

To compare the survival of dogs with and without epilepsy, a Kaplan-Meir survival analysis was used.[20] The effects of sex, an age of seizure onset <2 years, and occurrence of cluster seizures on the survival after the index seizure were also investigated using Kaplan-Meier survival analyses on the dogs with epilepsy compared with the dogs without epilepsy. The survival curves for different strata were compared using the log-rank test. Furthermore, the effect on survival, of epilepsy as cause of death, was compared with other causes of death in the group of dogs with epilepsy. All analyses were carried out using the survival package in R.12


Total Study Population

The 2009 investigation identified 66 dogs as epilepsy-positive and 116 dogs as epilepsy-negative. For 17 dogs, epilepsy status was unknown as their owners could not be reached.

In 2011, 10 of the epilepsy-positive dogs were lost to follow-up because the owners could not be reached and 2 dogs were excluded as the owners could not be interviewed in Danish. Three dogs were excluded because of inconsistent information given by the owners. This left 51 epilepsy-positive dogs that could be included in the analysis.

Out of the 116 epilepsy-negative dogs, 34 dogs were lost for follow-up in 2011, as the owners could not be reached. Twelve dogs were excluded, as the owner could not be interviewed in Danish. This left 87 epilepsy-negative dogs to be included in the analysis. The 2011 follow-up revealed that 9 dogs originally evaluated as epilepsy-negative in the 2009 investigation had developed classical signs of epilepsy between 2009 and 2011, which indicated an increase in prevalence in the family. These 9 dogs were, however, excluded from the analysis as they were not epilepsy-positive in 2009, and therefore could not be followed throughout the study period from 2009 to 2011.

Finally, a total of 129 dogs were included in the analysis. Of these, 51 dogs that were identified epilepsy-positive in 2009 and 78 that were identified epilepsy-negative in 2009 and remained negative in 2011, were included in the analysis. A flowchart describing the population is shown in Figure 1.

Figure 1.

Flowchart of the longitudinal study.

The group of epilepsy-positives consisted of 43 Groenendael (23 males and 20 females) and 8 Tervueren (4 males and 4 females). The epilepsy-negative group consisted of 54 Groenendael (28 males and 26 females) and 24 Tervueren (10 males and 14 females).

The median survival time for all 129 dogs was 8.8 years and with 50% dead at 11.1 years, CI [10.1, 11.3].

In 2011, 75 dogs were deceased and the estimated median age of death was 8.4 years (range: 0.6–15.4). For 19 dogs (25.3%) death was associated with epilepsy. Fifteen dogs (20%) died because of age-related reasons (arthritis, old age), 14 (18.6%) because of cancer, 8 (10.7%) because of behavioral problems (aggression, anxiety), and 5 (6.7%) were hit by car. Thirteen dogs (17.3%) were euthanized because of other reasons, and for 1 (1.3%) dog the cause of death was undetermined according to the owner.

Dogs without Epilepsy

Forty-eight of the 78 dogs without epilepsy were deceased in 2011. Median age at death was 9.25 years (range: 0.6–15.3),

Dogs with Epilepsy

Twenty-seven of the 51 dogs with epilepsy were deceased in 2011. Median age of death was 8.2 years (range: 2.3–15.4). Between 2009 and the follow-up none of these dogs developed neither signs of extracranial disease that could lead to seizures or seizure-like episodes nor neurological signs (apart from seizures) that could reflect intracranial diseases causing epilepsy. This stressed the validity of the diagnosis of epilepsy and the proposal of epilepsy of idiopathic (genetic) etiology established in 2009.

Age of onset was not provided in 7 of the 51 dogs. For the remaining 44 dogs the median age of onset was estimated to 4 years (range: 0.8–11.2).

In 19 (70%) of the 27 dogs, death was directly associated with epilepsy as 17 were euthanized motivated by the epileptic condition, and 2 died during or immediately after a generalized tonic-clonic seizure that did not differ from any other seizure experienced by the individual dogs in the past. These 2 deaths could be possible SUDEPs. Three dogs were euthanized because of age-related reasons (chronic arthritis and old age), 2 because of cancer, 1 was hit by car, and 2 were euthanized for other reasons.

In the group of epilepsy-related deaths (19) the median age at death was 8.0 years (range: 2.5–11), while it was 9.9 years (range: 2.3–15.4), for the group of deaths (8) not related to epilepsy. In most cases the owners indicated that unsatisfactory seizure control motivated the decision of euthanasia.

The median time of survival after index seizure for all the dogs with epilepsy was 4.3 years. At 5.1 years, 50% of the dogs were dead, CI [3, 8, NA].

One (2%) dog had experienced status epilepticus twice and was euthanized because of the last episode. Seventeen (33%) dogs had experienced cluster seizures. The median survival of these dogs after index seizure was 3.9 years (n = 14). The median survival in the group of dogs without cluster seizures was 4.8 years (n = 28).

Only 18 dogs received daily treatment for epilepsy. Out of these, 14 owners were able to recall the name of the AED used. The dogs were treated with phenobarbital and potassium bromide as monotherapy or combination therapy.

Spontaneous remission of epilepsy was reported in 7 (13.7%) dogs, all untreated with AEDs. In 1 female, remission of seizures occurred after neutering. The frequency of remission in epileptic dogs was 13.7%.

The results are summarized in Table 1.

Table 1. Distribution of dogs with and without epilepsy and data related to epilepsy
  1. EP+, epilepsy-positive dogs; EP−, epilepsy-negative dogs; ALL, all dogs; age at death, age in years; median survival, survival in years, median survival after index seizure, survival in years.

Age at death8.29.38.4
Deaths related to epilepsy70.4% (19) 25.3% (19)
Median survival  8.8
Median survival after index seizure4.3  
Status epilepticus2% (1)  
Cluster seizures33.3% (17)  
Remission13.7% (7)  

Statistical Analysis

The log-rank test showed no significant difference in the life span between dogs with and without epilepsy (P = .865). The life span of all dogs is presented in Figure 2. For dogs with epilepsy, survival after index seizure was not influenced by sex (P = .534), an age of onset <2 years (P = .309), or the presence of cluster seizures (P = .149). The survival after index seizure is displayed in Figure 3.

Figure 2.

The Kaplan-Meier survival curve for the family of Belgian Shepherd. The dashed lines represent the upper and lower 95% confidence limits.

Figure 3.

The estimated Kaplan-Meier curve of survival after index seizure. The dashed lines represent the upper and lower 95% confidence limits.

Among dogs suffering from epilepsy, individuals with epilepsy as cause of death lived significantly shorter compared with the group of dogs that had other causes of death. (P = .03). But when the 2 groups were compared regarding survival after index seizure no significant difference was found (P = .09)


In the present study the median age at death in Belgian Shepherds suffering from epilepsy was 8.2 years, which is not statistically significantly different from the median age at death of 9.25 years in the group of dogs without epilepsy. This study provides evidence that the prognosis for Belgian Shepherds with idiopathic (genetic) epilepsy is better than for a population of dogs including both idiopathic and symptomatic epilepsy. A survival study of epileptic dogs reported a significant difference in life span (7 years) between a group of dogs investigated at a referral hospital, and the life span of dogs in general (10 years).[7, 8]

Approximately 30% of human patients with epilepsy have a known acquired cause of epilepsy, such as head trauma, stroke, tumor, or a congenital lesion, while the remaining 70% include patients with the electroclinically defined “idiopathic” epilepsies, where genetic factors are believed to be central, as well as other less well-defined syndromes.[21, 22] It is well-known that case fatality is highest in patients with symptomatic epilepsies.[10] Human studies on epilepsy have given conflicting results regarding increased mortality in idiopathic epilepsy, but the majority of the studies have shown no or very slightly increased mortality for this group of epilepsies.[11, 23, 24] When looking at the results of the present study, it appear as this might also apply to some canine epilepsies with genetic etiology, as in Belgian Shepherds.[2, 3, 5, 6, 25]

The vast majority of the dogs in this study did not die spontaneously but were euthanized as a result of the owners′ considerations. As opposed to death in humans, death in dogs with epilepsy is mostly induced by euthanasia and strongly influenced by the owners' emotions concerning the dog and personal opinions on what is considered good seizure control and tolerable adverse effects. In this study we did not investigate the influence of adverse effects on survival because of the difficulties of measuring this variable objectively. It is, however, a most important issue to address when working with dogs with epilepsy as a delicate balance exists between seizure control and adverse effects.

Cluster seizures did not significantly affect survival after first seizure. However, as a relatively high number of the dogs investigated in this study experienced cluster seizures, it is important to stress the need for adequate seizure control and instructions regarding treatment of acute seizures, to avoid that owners lose courage.[26]

In the group of dogs with epilepsy, the deaths motivated by epilepsy occurred almost 2 years earlier than the deaths because of other reasons. This result was statistically significant and it is comparable with the results of a study on Border Collies with epilepsy, where the median age of death was 5.17 years. In the group of dogs whose death was epilepsy-related, the median age was significantly lower than in the group of dogs that died of other causes.[9] This was also the case in a study on Irish Wolfhounds, in which the life expectancy for the Irish Wolfhound was shown to be shortened by almost 2 years in epileptic dogs compared with seizure-free relatives.[12]

Epilepsy accounted for 25.3% of deaths in the Belgian Shepherd family as opposed to a study on the Irish Wolfhound in which more than 60% of the total number of deaths were related to epilepsy.[12] This information supports a suspicion of a more benign course of epilepsy in Belgian Shepherds.

The median survival time of 4.25 years after index seizure in the present study shows that Belgian Shepherds live longer with epilepsy compared with the 2.07 years found in Border Collies and 2.3 years in a mixed population.[11, 12] The survival time after index seizure was not significantly decreased if seizure onset was <2 years of age which is in contrast to what has been reported in Border Collies.[9] A possible explanation for the increased survival time in Belgian Shepherds could be the idiopathic focal nature of the epilepsy displayed combined with the remission pattern observed.

Probable sudden unexpected death occurred in 2 epileptic dogs, while no unexpected deaths were observed in the group of epilepsy-negative dogs. Both dogs died during or immediately after a generalized tonic-clonic seizure. The most important preventive tool against SUDEP in humans is adequate seizure control as many SUDEPs are related to poorly controlled chronic epilepsy with generalized tonic-clonic seizures.[14, 15] The figures of this study are very low, but they do stress the importance of monitoring and continuously adjusting AED treatment in dogs, to reach the best seizure control possible.

One important outcome variable of epilepsy studies is epilepsy remission, both spontaneous and AED induced. The present study found a remission of epilepsy of 13.7%. In a study on Border Collies, the remission rate was 18% and defined as 1 or 2 years with seizure absence and with or without treatment.[9] A study on Labrador Retrievers reported a remission rate of 24% with remission defined as 2 years with absence of seizures, and a study on a mixed population of epileptic dogs found a remission rate of 15%.[11] As several studies have now documented, remission of epilepsy is not rare in dogs, and this information should definitely be included when discussing prognosis with the owners.


Belgian Shepherds experience genetic epilepsy of a focal nature where the life span of affected individuals is not significantly influenced by the presence of epilepsy. With a remission rate of 13.7% and a very low frequency of status epilepticus, the epilepsy displayed in this breed appears to have a relatively mild course. However, with a very high prevalence, the disease has a significant impact on the breed

We are presently participating in a multicenter investigation of the genetic foundation of epilepsy in Belgian Shepherds. This will hopefully lead to an identification of the genetic background of epilepsy in this breed.


Conflict of Interest: Authors disclose no conflict of interest.


  1. 1

    Terry Therneau and original Splus->R port by Thomas Lumley (2011). survival: Survival analysis, including penalized likelihood. R package version 2.36–5. http://CRAN.R-project.org/package=survival .

  2. 2

    R Development Core Team (2011). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org/.