epilepsy (Level 2) taking the occurrence of drug-induced side effects and seizure control into account (Kwan et al., 2010):
• Level 1 categorizes the outcome of each medical treatment as either free of seizures or treatment failure, if not seizure free. The ILAE recommends a follow-up time of three times the longest interictal period for therapeutic trials; e.g. if the period in between seizures is 1 month then the follow-up time should be at least 3 months. However, as this is not always practical, seizure freedom is usually defined as an absence of seizures for a 12-month period. Level 1 does not only consider efficacy, but also the occurrence of drug-induced side effects, which will contribute significantly to the overall impact on the patient’s quality of life.
• Level 2 defines pharmacoresistant epilepsy as a failure to achieve freedom from seizures despite adequate trials of two (or more) well-tolerated, correctly chosen and used AED regimens (whether administered as monotherapies or in combination) (Kwan et al., 2010). The number of AED trialled has been restricted to ‘two’ based on the observation that successful seizure control is very unlikely once the patient has not responded to two correctly chosen AED (Kwan and Brodie, 2000; Arts et al., 2004).
It needs to be considered that drug non-responsiveness is not a fixed state but rather a dynamic process, which can be altered by fluctuation of the underlying pathophysiology (Kwan et al., 2010). The seizure frequency can vary over time (Berg et al., 2009; Muñana et al., 2010) and the state of drug non-responsiveness is dependent on the time-point of assessment (Kwan et al., 2010).
Taking into consideration the revised ILAE classification of pharmacoresistant epilepsy, a high proportion of dogs would be classified as pharmacoresistant. Spontaneous and drug-induced epilepsy remission rate in people is around 63% (Kwan and Brodie, 2000), which is markedly higher than that reported in veterinary medicine, which usually ranges between 15 and 24% (Heynold et al., 1997; Berendt et al., 2007; Arrol et al., 2012). However, in a recent study comparing the efficacy and tolerability of phenobarbitone with potassium bromide as a first line treatment, seizures ceased in 85% and 52% of the treated dogs, respectively (Boothe et al., 2012). The study period was only 6 months long and thus it is likely that the percentage of dogs being seizure free would have been less with a longer follow-up period. Lagotto Romagnolo dogs have a documented ‘childhood’ or juvenile epilepsy, which starts when they are around 6 weeks of age (Jokinen et al., 2007). Of the affected dogs, 96% become seizure free at the age of 10 weeks generally without receiving AED treatment. Future, long-term epidemiological studies are necessary to determine prevalence of drug response more accurately in the general epileptic dog population, in addition to the effect of patient age and duration of seizure history on this response.
Risk Factors for Pharmacoresistant Epilepsy
Genetic risk factors
An increased prevalence of epilepsy has been described for many dog breeds (Jaggy et al., 1998; Kathmann et al., 1999; Berendt et al., 2002, 200; Casal et al., 2006; Gullov et al., 2011), which raises the possibility of genetic risk factors also being responsible for drug-responsiveness. The Lagotto Romagnolo, as aforementioned, has a benign childhood or juvenile epilepsy, which goes into remission in many affected dogs with age, as the function of the Lgi2 gene in cerebral synaptic remodelling becomes less important (Jokinen et al., 2007; Seppala et al., 2011). The Lgi2 gene is involved in the immediate post-natal phase of neuronal pruning, but then the Lgi1 gene has a more important role in the regulation of the later part of pruning.
Another gene mutation that has been associated with improved seizure control is located on the ABCB1 gene. The ABCB1 (multidrug resistance (MDR) 1) gene encodes a transmembrane protein, the permeabilityglycoprotein (P-gp). P-gp, an ATP-dependent multidrug transporter, expressed at the blood–brain barrier (BBB), protects the brain from potential central nervous system (CNS) toxins (Schinkel et al., 1994, 1996; Schrickx and Fink-Gremmels, 2008), including AEDs such as PB (Potschka et al., 2002; Basic et al., 2008). Dysfunction of this neuro-protective BBB efflux pump can lead to BBB dysfunction and consequently CNS intoxication (Schinkel et al., 1994, 1996). Collies and other herding breeds can be affected by such a dysfunction of the P-gp transporter caused by a four base-pair deletion (c.296_ 299del) in exon 4 of the ABCB1 gene (Mealey et al., 2001). A homozygous deletion in this region cannot only lead to CNS intoxication by drugs such as ivermectin, but also has been suggested to improve medical seizure control in affected collies (Muñana et al., 2012a).
Border collies are, however, known to have an aggressive seizure phenotype characterized by high seizure frequency, cluster seizures and poor drug responsiveness. In a recent study, 71% of Border collies were classified as pharmacoresistant (Hülsmeyer et al., 2010). PB-resistant epileptic Border collies had a single nucleotide polymorphism (SNP) at intron 1 near the 5-end of the ABCB1 gene, which could influence the promoter elements of this gene (Alves et al., 2011). A polymorphism at the promoter region could influence transcription activity and therefore might increase the level of P-gp expression at the BBB. The same group looked at another breed, the Australian shepherd dog, which is closely related to the Border collie and also has a severe epilepsy phenotype (Weissl et al., 2012). In this dog breed, they were not able to identify a polymorphism related to drug-refractoriness.
Another recent study looked at gene differences between PB responders and non-responders (Kennerly et al., 2009). Five genes were suggested to have an association with PB response, although they did not reach statistical significance after adjustment for multiple comparisons (KCNQ3, SCN2A, GABRA2, EPOX HYD and ABCB4). Three of these genes encode ion channels (KCNQ3, voltage gated potassium ion channel important for post-excitatory membrane re-polarization; SCN2A, sodium ion channel; GABRA2, GABA receptor), all of them are AED targets (Armijo et al., 2005). The other two genes are involved in PB metabolism (EPOX HYD) or transportation (ABCB4) (Kennerly et al., 2009).
Clinical risk factors
Apart from identifying genetic backgrounds, which can be associated with poor response to AEDs, much attention has been spent in the last decade to identify clinical risk factors that predict pharmacoresistant epilepsy. In people, it was commonly believed that seizure freedom was more likely to be achieved when a patient received AEDs immediately after the occurrence of the first seizure. Interestingly, epidemiological studies in developing countries, where AEDs are not freely available, revealed remission rates similar to those in industrial countries (Placencia et al., 1993). Most AEDs despite having good seizure suppressing activities have little influence on the natural course of the disease (epilepsy) itself. New AEDs that were thought to also have an anti-epileptogenic effect have failed in chronic epilepsy models and have not prevented the development of epilepsy (Brandt et al., 2007).
Heynold and colleagues have shown in 1997 that Labrador retrievers, which achieved freedom from seizures, received medication a longer period of time after their first seizure than those dogs that continued to seizure (Heynold et al., 1997). This implies that timing in relation to onset of seizure activity does not influence long-term seizure control in the canine patient. However, there is enough evidence in dogs, rodent models and humans that disease severity, such as high seizure frequency, cluster seizures and status epilepticus can influence drug responsiveness (Heynold et al., 1997; Kwan and Brodie, 2000; Hülsmeyer et al., 2010; Löscher and Brandt, 2010; Weissl et al., 2012). Intact male and female dogs have a higher likelihood of having cluster seizures (Monteiro et al., 2012). Seizure severity will ultimately guide clinical reasoning and the more severely affected patient will receive treatment earlier than dogs with a less severe epilepsy phenotype. It was formerly believed that very
