Vestibular disease in dogs under UK primary veterinary care: Epidemiology and clinical management

Abstract Background Vestibular disease (VD), central or peripheral, can be a dramatic primary‐care presentation. Current literature describes mostly dogs examined in referral centers. Hypothesis/Objectives Describe the prevalence, presentation, clinical management, and outcomes of VD in dogs under primary veterinary care at UK practices participating in VetCompass. Animals Seven hundred and fifty‐nine vestibular cases identified out of 905 544 study dogs. Methods Retrospective cohort study. Potential VD cases clinically examined during 2016 were verified by reviewing clinical records for signalment, presenting clinical signs, treatments, and outcomes. Multivariable logistic regression was used to evaluate factors associated with VD. Results The overall prevalence of VD was 8 per 10 000 dogs (95% CI = 7‐9). Median age at first diagnosis was 12.68 years (interquartile range [IQR], 11.28‐14.64). Compared with crossbreeds, breeds with the highest odds of VD diagnosis included French Bulldogs (odds ratio [OR] = 9.25, 95% CI = 4.81‐17.76, P < .001), Bulldogs (OR = 6.53, 95% CI = 2.66‐16.15, P < .001), King Charles Spaniels (OR = 4.96, 95% CI = 2.52‐9.78, P < .001), Cavalier King Charles Spaniels (OR = 3.56, 95% CI = 2.50‐5.06, P < .001), and Springer Spaniels (OR = 3.37, 95% CI = 2.52‐4.52, P < .001). The most common presenting signs were head tilt (69.8%), nystagmus (68.1%), and ataxia (64.5%). The most frequently used treatments were antiemetics (43.2%), systemic glucocorticoids (33.1%), antimicrobials (25%), and propentofylline (23.25%). There were 3.6% of cases referred. Improvement was recorded in 41.8% cases after a median of 4 days (IQR, 2‐10.25). Conclusions Our study identifies strong breed predispositions for VD. The low referral rates suggest that primary‐care data sources offer more generalizable information for benchmarking to help clinicians review their own clinical activities.


| INTRODUCTION
Vestibular disease (VD) is characterized by a dysfunction of the parts of the nervous system responsible for the maintenance of equilibrium and balance. 1 The diagnosis of VD is based on cardinal neurological examination abnormalities such as ataxia, head tilt, and pathological nystagmus and strabismus. These characteristic clinical signs can be caused by either central or peripheral neurological dysfunction. 2 Neurological signs associated with central involvement include proprioception deficits, altered mentation, cranial nerve (CN) deficits other than CN VII or VIII, and vertical or dysconjugate nystagmus. Recognition of these clinical signs can help the clinician differentiate between central and peripheral variants and has an important role in decision making regarding clinical management and when considering referral for advanced imaging. The most common diagnoses in dogs presenting with peripheral VD are otitis media/interna and idiopathic peripheral VD (IPVD). 2 IPVD is reportedly more common in older dogs and the diagnosis is made by exclusion. 1 For central VD, multiple causes are reported, including anomalous, metabolic, neoplastic, infectious/inflammatory, traumatic, toxic, and vascular etiologies. 3 Regardless of the etiology, VD can be a serious welfare concern for affected dogs. Clinical signs are often dramatic on presentation, 4 with severe disorientation causing distress for both the affected animal and the owner. Although vestibular disorders in dogs are reportedly common, [5][6][7] there is limited information on the prevalence, risk factors, progression, and outcome of VD in the primary-care population of dogs because the current veterinary literature is based mainly on the subset of cases presenting to referral institutes. These referral cases are likely to be more complex cases with more financially committed owners and therefore to generalize poorly to the wider population. 8 The current study aimed to describe the signalment, prevalence, and risk factors in dogs diagnosed with VD in primary-care practice in the United Kingdom and to describe disease progression, clinical management, and outcomes of these cases.

| MATERIALS AND METHODS
The VetCompass Programme collates de-identified electronic patient record (EPR) data from primary-care veterinary practices in the United Kingdom for epidemiological research. 9 VetCompass collects information fields that include species, breed, date of birth, sex, neuter status and bodyweight, and clinical information from free-form text clinical notes and summary diagnosis terms (VeNom codes 10

| RESULTS
The overall study population comprised 905 544 dogs under primary veterinary care at 886 veterinary clinics during 2016. Of 2600 dogs identified as candidate VD cases, manual checking identified 759 confirmed VD cases, giving an overall prevalence of 8 per 10 000 dogs (95% CI = 7-9 per 10 000 dogs). The prevalence in dogs that were 9 years or older (636/177 754) was 36 per 10 000 dogs (95% CI = 33-39) and the prevalence for each breed is shown in Table 1. Out of the 759 VD cases, only 39 (5.13%) were pre-existing cases that were diagnosed with VD before 2016, therefore the incidence of VD was similar to the prevalence: 8 per 10 000 dogs (95% CI = 7-9).   Out of the 759 cases diagnosed with VD, 139 (18.31%) were hospitalized for management and the treatments prescribed are summarized in Table 5.
There were 27/759 (3.55%) cases referred. A specific cause of VD in the clinical notes or in a referral letter was recorded in 33/759 (4.34%) of the cases. Out of these, 14 were recorded with IPVD, 11 with brain disease, 7 with either otitis media or interna, and 1 case with metronidazole toxicity. Of the 14 cases diagnosed with IPVD, 3/14 (21.42%) were reported to have facial nerve palsy.
Of the 11 cases with a brain disease confirmed, 6 were diagnosed with meningoencephalitis of unknown origin (MUO), 2 with intracranial neoplasia, 1 with Chiari-like malformation, 1 with rostral cerebellar ischemic stroke (suspected to be caused by thromboembolic disease), and 1 with an unspecified cerebellar lesion. in the United Kingdom reported "brain disorder" as the most common cause of mortality. 26 It is interesting to note that both French Bulldogs and Bulldogs did not show significantly increased odds for being diagnosed with VD in the univariable logistic regression analysis (Table 2). We believe that, in the univariable model, the age was a confounding factor. Both breeds had a lower median age at presentation (Table 3), which could suggest that diseases associated with older age (such as IPVD or neoplastic diseases) were less likely to be the cause for VD in these breeds.
King Charles Spaniels and CKCS were also shown to have increased odds of VD compared with crossbred dogs, 4.96 and 3.56 times, respectively. A previous study on the disorders causing presentation to primary care practice in CKCS found a prevalence of neurological causes to be 4.3%. 29 CKCS are known to be predisposed to several neurological syndromes that can cause vestibular signs, including occipital hypoplasia/syringomyelia, granulomatous meningoencephalomyelitis (GME), and VD. 30 Moreover, CKCS have a tendency toward cerebrovascular disease, particularly rostral cerebellar artery infarction. [31][32][33] In immature dogs, the first presenting sign of syringomyelia could be scoliosis, which can appear similar to a head tilt of vestibular origin. 30 VD in CKCS can be also idiopathic-CKCS is one of the most commonly seen breeds with idiopathic facial nerve paralysis 34 and facial and vestibular neuropathy of unknown origin. 35 Aside from central and IPVD, CKCS are also predisposed to primary secretory otitis media, known to commonly be associated with peripheral VD. 3,36 Springer spaniels were shown to have 3.37 times the odds of diagnosis with VD compared to crossbreds. Springer Spaniels are predisposed to cerebellar infarcts that occur most commonly within the region supplied by the rostral cerebellar artery 33 40 and this could potentially contribute to an increased predisposition toward IPVD or other causes of VD.
In our study, the median age at first diagnosis was 12.68 years.
This suggests that IPVD might be implicated in most cases, as the mean age of onset for IPVD in published literature is between 9.4 and 12.5 years. 5,18,21 Other causes of VD cannot be excluded, however, as it has been documented that the incidence of brain tumors, for example, is higher in dogs older than 5 years. 41 Regardless of the primary cause, VD should be considered a high-risk disease in older dogs, especially in the predisposed breeds described above.
The most common clinical signs on presentation were ataxia, nystagmus, and head tilt, which are clinical signs that are already known as the hallmark signs of VD. 42 Vertical nystagmus is generally present only in central VD 1,3 and was documented in 7.3% of the cases that presented with nystagmus. This should be interpreted with caution however, as vertical nystagmus can be difficult to differentiate from a rotary one (which can be seen in either peripheral or central VD). The presence of pathological nystagmus tends to be short-lived, especially in cases of peripheral VD, because it can often be rapidly compensated by voluntary visual fixation. 3 The high prevalence of nystagmus among the VD cases in our study therefore suggests an acute presentation.
Vomiting is a common clinical sign of VD in dogs 43 and in our study 25.7% of cases presented with vomiting, similar to previously published literature. 18 Although vomiting occurs with VD regardless of the cause, it is more common in dogs with acute peripheral VD and is a result of the connection between the vestibular nuclei axons and the vomiting center in the brainstem reticular formation. 3,21 Vestibular disorders in people are known to cause unpleasant autonomic signs such as nausea and vomiting that can lead to anxiety and panic attacks. 44 Even though vomiting is usually transitory in dogs with VD, 21 it should therefore be considered to have welfare implications.
Collapse was recorded in 32.1% of the cases, but it is difficult to discern whether this was due to central disease and loss of proprioception or due to marked imbalance and disorientation. We therefore suspect the presence of collapse is an indication of the severity of the disease or the acute onset.
Differentiating between central and peripheral VD is an important goal for the clinician when assessing a dog with signs of VD. Central VD generally requires more expense to diagnose and treat, and can be associated with a poorer prognosis. 45 In our study, signs of central involvement were identified in 14.1% of cases. However, the true proportion with central VD could be higher as many cases did not have a full neurological examination described in the EPRs. A large number of cases (11.5%) were euthanized on presentation and where VD contributed to the decision. This could be due to veterinary recommendations or at the request of the owner. VD cases can show a dramatic presentation causing a lot of distress to the owners.
Also, considering that the median age at diagnosis was quite old, other concurrent diseases could be contributing to the decision of euthanasia. However, it is possible that owners could now be reassured by their veterinary surgeons about a good probability of improvement reported in both the previous literature as well as the current study and might opt to trial treatment which is generally affordable and to monitor for signs of recovery for at least a few days after diagnosis in a greater proportion of cases.
The limitations of the study mainly arise from the retrospective design. The data collected from the EPRs were not recorded primarily for research purposes and were therefore limited by incomplete information and potential lack of accuracy. We suspect the prevalence of 8 per 10 000 dogs in our study is underreporting the prevalence of VD in primary-care clinical practice. By using the search terms "nyst" and "vest," we retrieved 2600 potential cases and we manually evaluated each against the case definition, but potential true cases could have been missed due to misspelling or lack of complete entry of clini-

CONFLICT OF INTEREST DECLARATION
Authors declare no conflict of interest.

OFF-LABEL ANTIMICROBIAL DECLARATION
Authors declare no off-label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION
Royal Veterinary College Ethics and Welfare Committee (reference number URN 2015 1369).

HUMAN ETHICS APPROVAL DECLARATION
Authors declare human ethics approval was not needed for this study.