Evaluation of Pexion® (Imepitoin) for Treatment of Storm Anxiety in Dogs: a Randomised, Double-Blind, Placebo-Controlled Trial

Received: 20 July 2020 Revised: 7 October 2020 Accepted: 3 December 2020

DOI: 10.1002/vetr.18

ORIGINAL RESEARCH

Evaluation of Pexion® (imepitoin) for treatment of storm anxiety in dogs: A randomised, double-blind, placebo-controlled trial

Irina Perdew1 Carrie Emke1 Brianna Johnson1 Vaidehi Dixit2 Yukun Song2 Emily H. Grifﬁth2 Philip Watson3 Margaret E. Gruen1

1 Department of Clinical Sciences, North Abstract Carolina State University College of Background: While often grouped with other noise aversions, fearful Veterinary Medicine, Raleigh, North Carolina, USA behaviour during storms is considered more complex than noise aversion 2 Department of Statistics, North Carolina alone. The objective here was to assess the effect of imepitoin for the treat- State University College of Sciences, Raleigh, ment of storm anxiety in dogs. North Carolina, USA Methods: In this double-blind, placebo-controlled randomised study, eligible 3 Ingelheim am Rhein, Boehringer-Ingelheim dogs completed a baseline then were randomised to receive either imepitoin Vetmedica GmbH, Ludwigshafen am Rhein, (n = 30; 30 mg/kg BID) or placebo (n = 15) for 28 days. During storms, owners Germany rated their dog’s intensity for 16 behaviours using a Likert scale. Weekly, own-

Correspondence ers rated intensity and frequency of these behaviours. Summary scores were Margaret E. Gruen, Department of Clinical compared to baseline and between groups. Sciences, North Carolina State University Results and Conclusions: Imepitoin was signiﬁcantly superior to placebo in College of Veterinary Medicine, 1060 William Moore Drive, Raleigh NC 27607, USA. storm logs and weekly surveys for weeks 2 and 4, and in the end-of-study sur- Email: [email protected] vey. Mild/moderate adverse events were reported in 26 patients (24 active: two placebo); the most frequent adverse event was ataxia. Owners of dogs in the imepitoin group, compared to placebo, were signiﬁcantly more likely to report that treatment reduced their dogs fear and anxiety during storms (p < 0.001) and other noise events (p < 0.001). Twice daily administration of imepitoin decreased anxiety scores in dogs with storm anxiety. Future work may evaluate optimal dosage regimens.

INTRODUCTION storm’s occurrence. Here, we refer to this condition as ‘storm anxiety’ to address recent shifts in terminology Noise aversion is a common behavioural disorder in and include dogs who do not meet the criteria for a dogs, with affected dogs showing signs of fear and phobic response,1,4,5 but show signs of fear and anx- anxiety in response to noise stimuli. Common signs iety during storms. Meteorological changes including associated with noise aversion include hypersali- shifts in barometric pressure, wind, static electricity vation, panting, hiding, vocalising and attempts to and illumination levels have all been suggested as escape; these signs may be severe enough to result additional triggers for dogs with storm anxiety.6 The in self-harm.1 Theprevalenceofdogswithfearful contribution of each of these factors is difficult to responses to noise has been reported as up to 25– delineate; however, their possible influence means 50%, making it among the most common canine that storm anxiety must be considered separately behavioural disorders.2 When severe, this condition from other noise aversions and is nearly impossible is termed ‘phobia’, denoting a profound, non-graded to model using noise alone. Storm anxiety may be the and extreme response.1,3 Fear of thunder has been only noise-associated condition a dog is affected with, considered a subset of noise aversion; in this defini- or they may have aversions to other noise stimuli. A tion, the sound of the thunder is the fear-inducing recent survey study found a high correlation between stimulus. This has often been referred to as storm fear of thunder and certain other noises; specifically, phobia or thunderstorm phobia. However, storms a majority of dogs with fear of thunder showed fear of encompass more than just the noise of thunder, and fireworks (92.9%) and gunshots (73.8%) with a lower dogs who show signs of fear and anxiety associated percentage showing fear to other noises like vacuums, with storms often show signs well in advance of a sirens, etc. (53.5%). Dogs with fear of fireworks did

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modiﬁcations or adaptations are made. © 2021 The Authors. Veterinary Record published by John Wiley & Sons Ltd on behalf of British Veterinary Association

Vet Rec. 2021;e18. wileyonlinelibrary.com/journal/vetr 1of9 https://doi.org/10.1002/vetr.18 2of9 Veterinary Record not always show fear of thunder (71.8%) suggesting in dogs; these included dogs with fears of noises in that these two stimuli have individual characteristics addition to other behavioural concerns. They found triggering signs of fear in dogs.7 that imepitoin given twice daily, in conjunction with Treatment for dogs with noise aversion involves a a behavioural modification programme was effective combination of environmental and behavioural mod- in the reduction of fear and anxiety scores. This case ification and pharmacologic therapy. Two drugs have series provided evidence to support further investiga- recently been approved for the treatment of noise tion of imepitoin for the treatment of fear and anxi- aversion in dogs,8,9 yet none is labelled for the treat- ety disorders in dogs, including storm anxiety. Given ment of storm anxiety specifically. Indeed, only one its anxiolytic properties and the ability to use imepi- medication combination, clomipramine and alprazo- toin long term, our objective in this study was to assess lam, has been evaluated specifically for the treatment the efficacy of imepitoin specifically for treatment of of storm anxiety (referred to as storm phobia) in client- storm anxiety in dogs. We hypothesised that imepi- owned dogs.6 Existing treatments for storm anxiety toin would decrease the clinical signs of anxiety in often include selective serotonin reuptake inhibitors, dogs with a history of storm anxiety as compared to such as fluoxetine, with an additional episodic treat- placebo. ment administered prior to or during a storm.1,10 This presents a feasibility concern for many owners of dogs with storm anxiety, particularly if a storm is unex- MATERIALS AND METHODS pected. In some parts of the globe, such as the south- ern region of the United States, storms occur with All dogs enrolled in the study were client-owned and some regularity during the ‘storm season’ that extends remained with their owners for the duration of the from early spring through late fall. An effective treat- study. Owners provided written informed consent; all ment that could be used daily during this period, and procedures were carried out with approval from the episodically outside of it (or in other regions), would North Carolina State University Institutional Animal fill a gap in our currently available drugs. Care and Use Committee (Protocol #18-053-O). Imepitoin is a partial gamma aminobutyric acid Subjects were recruited via flyers and advertise- (GABA)-A agonist in the benzodiazepine class that ments on social media and through emails to area has been approved as a treatment for noise aversion veterinarians. Interested owners completed an online in dogs in the United States and European Union. pre-screening questionnaire for preliminary assess- In a randomised, placebo-controlled trial, imepitoin ment of eligibility. To be eligible, dogs had to be was found to be effective in decreasing signs of noise between 1 and 12 years of age, and weigh between aversion (to fireworks) in dogs.8 It is approved at a 2 and 50 kg (4.4–110 lb). Dogs were further required dosage of 30 mg/kg twice daily to be given 2 days to have been living with their owner for at least 1 prior to a noise event and continuing through the end year, have a stable history of storm anxiety (for at of the event. In laboratory-based studies, imepitoin least 1 year) and have a qualifying score (≥ 30) for also showed efficacy in a noise-induced model of fear intensity on the Lincoln Sound Sensitivity Scale.19 To and anxiety in dogs, using thunder track as the noise complete the Lincoln Sound Sensitivity scale, owners stimulus,11 but has not been previously evaluated for rate their dog’s frequency and intensity of signs for 16 clinical cases. As a class of psychoactive drugs, ben- behaviours along a scale from 0 to 3 (frequency) and zodiazepines have also long been used alone or as 1 to 5 (intensity) where lower scores indicate less fre- adjunctive medications for the treatment of anxiety quent or severe behaviour. Dogs were excluded if they conditions, including storm anxiety, in dogs.6,12,13 In had a history of aggressive behaviour, significant sys- a retrospective study on the use of diazepam for anx- temic disease (cardiovascular, hepatic, renal, etc.), or iety conditions in dogs, all owners who administered were taking other psychoactive medications; owners diazepam for storm anxiety reported the treatment were excluded if they had a major life event (move, to be at least somewhat effective for their dogs14;in new baby, etc.) planned during the period of the a prospective study, alprazolam (in conjunction with study. If owners had been using adjunctive treatments clomipramine) was associated with a 52% decrease (pheromones, pressure wraps, etc.) or behavioural in signs of storm anxiety.6 A limitation of the use modification, these were allowed to continue as long of benzodiazepines is the risk of adverse effects – as they had been in use for at least 4 weeks and dogs such as ataxia, behavioural disinhibition and idiosyn- still met entry criteria for their Lincoln Sound Sensitiv- cratic excitatory reactions – and the development of ity Score (LSSS) scores; no new adjunctive treatments tolerance.14 As a low affinity partial agonist, imepitoin or behavioural modifications were allowed during the may decrease the potential for the development of tol- study period. All dogs remained with their owners dur- erance and other adverse effects. Indeed, imepitoin is ing and following completion of the study. approved in several countries (including the European The overview of the study timeline is shown in Fig- Union and United Kingdom) for long-term use in the ure 1. Owners and dogs who were qualified based treatment of canine epilepsy, with dogs remaining on on pre-screening were scheduled for an in-person treatment for months to years.15–17 screening appointment. Following owner consent, all McPeake and Mills18 reviewed their use of imepi- dogs received a physical examination, and routine toin at a range of doses as a long-term treatment for labwork was performed (complete blood count and fear- and anxiety-related problems as a case series serum biochemistry). Once enrolled, dogs began a Veterinary Record 3of9

FIGURE 1 Graphic description of the timeline of the study. Owners visited the veterinary teaching hospital on three occasions (screening, Day 0, and Day 28) while dogs were only required to be seen on one visit (screening) baseline period of at least 2 weeks. During this period, also asked if they believed their dog had received the owners were asked to complete a storm log for each active medication or a placebo. storm their dog was exposed to, if they were present to observe their dog’s behaviour. Storm logs included information about the storm (date, duration and Statistical analysis intensity) and asked owners to rate the intensity of their dog’s behavioural signs on the Lincoln Sound Descriptive statistics were used to compare groups for Sensitivity Scale. To complete the baseline period, age, sex, weight, screening Lincoln Sound Sensitivity owners were required to have completed at least two Scale score and dosage prescribed. Behavioural data storm logs; if an insufficient number of observed were handled separately for weekly survey data and storms had taken place during the first 2 weeks, the storm log data. baseline period was extended until two storms had For weekly survey data, the product of the inten- been observed (up to 6 weeks from screening). Once sity and frequency ratings was calculated by multiply- the baseline period was complete, dogs were ran- ing the intensity score1- 5 by the frequency score (0– domised to treatment or placebo (in a 2:1 ratio) and 3) and taking the sum of these products across the 16 owners returned to the North Carolina State University behaviours (possible range of 0–240). Data were then College of Veterinary Medicine (NCSU-CVM) on Day analysed using a split-plot model. 0 to collect their medication and exchange their com- Data were analysed using both intention-to-treat pleted storm logs for blank storm logs. Owners and and per-protocol analyses to account for dogs with- investigators were masked to the group (treatment or drawn from the study. In the intention-to-treat anal- placebo) with randomisation and medication prepara- ysis, missing values were imputed as follows: If a tion handled by the NCSU-CVM Pharmacy; identical dog withdrew from the study, their last values were packaging and tablet sizes were used for each group. carried forward throughout the data analysis. If there This began a 28-day (±3 days) treatment period where was a missing value in the middle of the study, the dogs received either imepitoin (30 mg/kg BID) or a decision to fill in the missing value was based on a visually identical placebo, and owners recorded their conservative model using immediate neighbouring dog’s behaviour in response to storms on their storm values and group (active or placebo). It is more con- logs. servative to claim a non-existing effect than to neglect In addition to the storm logs, owners were sent a an existing effect for a drug. Therefore, if the dog was link to an online survey to complete once weekly. This in the placebo group, the missing value was set equal survey was similar to the storm logs, except it asked to the neighbouring value corresponding to the better owners to consider all the storms their dog had been performance; conversely, if the dog was in the active exposed to over the past week, and rate both the fre- group, the missing value was set as the neighbouring quency and intensity of behavioural signs. Thus, out- value corresponding to the worse performance. An come measures of interest included the change in alpha-level of 0.05 was set for model significance, and average scores from the storm logs during the treat- all tests used a one-tailed hypothesis (imepitoin would ment period compared to the baseline period, and have a greater improvement than placebo). the change in weekly survey scores completed dur- For storm log data, scores for the Lincoln Sound ing the treatment period compared to the baseline Scale were summed across all 16 behaviours rated by period. Owners were asked to report any adverse owners; while the intensity score ranged from 1 to events that occurred throughout the study. They were 5, if a behaviour was not present, the dog received a asked end-of-study questions about their impressions score of 0 for that behaviour, making the total possible of the treatment their dog received. On this end-of- range 0–80. Storm data were then evaluated using a study survey, owners were asked to rate their agree- linear mixed effects model, using the ‘number of days ment with a series of statements on a 7-point scale since the start of treatment’ as a covariate. In addition, from strongly agree to strongly disagree. Owners were storm log data were averaged over each week of the 4of9 Veterinary Record

TABLE 1 Demographics for the placebo and active group for weight, age, sex ratio, screening Lincoln Sound Scale Score (LSSS) sum and dosage dogs received

Placebo Active

Weight (kg; mean 25.2 (± 8.7) 24.6 (± 10.9) t = 0.19; ± SD) p = 0.85 Age (years; mean 8.1 (± 2.6) 7.1 (±2.6) t = 1.19; ± SD) p = 0.24 Sex (male: female) 3:12 15:15 Fisher’s exact; p = 0.06 Baseline LSSS 44.1 (1.6) 41.3 (1.1) t = 1.44; sum (mean ± p = 0.16 SD) Dosage per dose 29.84 (± 30.4 (± 3.09) t =−0.62; (mg/kg; average 2.46) p = 0.54 ± SD)

Dogs were randomised to placebo or active groups in a 1:2 ratio. Group demographics were similar between the groups for age, weight, screening LSSS sum and dosage received; the placebo group had a higher (though not significantly different) proportion of females than males, while the active group was bal- anced with regard to sex (Table 1). Dog were exposed to an average of 5.2 storms over the course of the treatment period, with no difference in average number between dogs in the active and placebo groups (t-test, FIGURE 2 Flow diagram of study participation t = 1.70, p = 0.096). treatment period and compared a split-plot model in Adverse events a weekly model. Responses on the end of study survey were cate- Mild-to-moderate adverse events were reported in gorised as Agree (strongly agree, agree and somewhat 26 patients (24 active:2 placebo); 10 patients with- agree), Neither agree nor disagree, and Disagree drew from the study prior to 25 days (all from the (somewhat disagree, disagree, strongly disagree). active group). Five were withdrawn within one week Response distributions were compared between of starting the medication; one each was withdrawn groups using Fisher’s exact tests. between 1 and 2 and 2 and 3 weeks; and three were withdrawn between weeks 3 and 4. The reasons given for withdrawal included: ataxia and decreased RESULTS appetite2; decreased appetite and elevated liver enzyme concentrations1; impulsive behaviour1; defi- Subjects ant with training and fighting with other dog1; ataxia and subdued behaviour (two from the same family); Over a period of 12 weeks, a total of 307 owners com- ataxia and upset stomach1; pale gums and decreased pleted the pre-screening survey. Dogs were not eligi- appetite1; sedation, ataxia, and hypersalivation.1 The ble for the study if they had a Lincoln Sound Sensitivity owner of the dog with reported pale gums declined an Scale intensity score <30 or had medical, behavioural examination or complete blood count, so this remains or exclusionary medications. Fifty-eight client-owned unconfirmed; her dog was withdrawn from the dogs were screened, with 49 dogs enrolling in the study and recovered without incident. The dog with study; 47 dogs completed the baseline period, and decreased appetite and elevated liver enzyme concen- 45 dogs were fully enrolled (i.e. randomised to treat- trations was seen by their veterinarian and found to ment or placebo). Of the screened dogs, nine were have elevated alanine transferase (1089 U/L; normal excluded due to medical concerns (physical examina- range 10–118 U/L) and alkaline phosphatase (362 U/L; tion findings, labwork or both), one was excluded for normal range 20–150 U/L). This was reported after 25 not observing enough storms after a 5-week baseline days on treatment. The dog was treated for possible period, one owner moved suddenly during the base- bacterial hepatitis and given subcutaneous fluids, an line period, and two owners completed the baseline injection of antibiotic, oral probiotics, oral antiemetics but failed to come for their Day 0 visit (see Figure 2 for and his alanine transferase was followed with repeated flow diagram). This left 45 dogs to be randomised into labwork. In two weeks, his alanine transferase was treatment groups. 142 U/L and alkaline phosphatase was 166 U/L and his Veterinary Record 5of9

TABLE 2 Adverse events experienced by dogs in the imepitoin for all comparisons for imepitoin group; p < 0.003 for and placebo groups all comparisons for placebo group). Significant dif- Imepitoin ferences were found between the placebo and active Event (n = 30) Placebo (n = 15) drug at each week of the study (Table 3). Ataxia 18 (60.0%) 1 (6.7%) The per-protocol analysis excluded dogs as they withdrew from the study. This meant that in the final Impulsive behavior 1 (3.3%) – week analysis, there were a total of 15 dogs in the Hypersalivation 6 (20.0%) – placebo group and 20 dogs in the imepitoin group. Increased appetite 9 (30.0%) 1 (6.7%) For the weekly survey scores, the sum of the prod- Decreased appetite 2 (6.7%) – ucts of frequency and intensity was used, with a split- Somnolence 7 (23.3%) – plot design model fit for each (plots of estimated lines ± Emesis 1 (3.3%) – SD shown in Figure 3). A significant difference was found between the placebo and imepitoin groups dur- Irritability/aggression 2 (6.7%) 1 (6.7%) ing week 2 (p = 0.01) and week 4 (p = 0.02; Table 3). appetite had returned to normal. The two dogs in the placebo group with adverse events were ataxia (mild) Storm log scores and increased appetite with more aggression shown to other dogs in the home over toys; these resolved within For the storm log data, the sums of intensity scores 5 and 7 days, respectively. The overall distribution of were used in the linear models; data from all 45 dogs adverse events in each group is shown in Table 2. were included in the analysis. Significant differences were found between the slopes of the lines for the placebo and active groups, with the imepitoin group Efficacy: intention-to-treat and having a steeper slope (i.e. greater decrease in score) per-protocol analyses than the placebo group (95% CI [−∞, −0.03813]; Figure 4). All 45 enrolled dogs were included in the intention- to-treat analysis. In the intention-to-treat analysis, missing values were imputed based on neighbouring End of study survey values as described. For the weekly survey scores, the sum of the products of frequency and intensity At the end of the study, owners answered ques- were used, with a split-plot design model fit for each tions about their impressions of the treatment their (plots of estimated lines ± SDshowninFigure3). For dog received, and whether they believed their dog the sum of products of frequency and intensity, both had received the active treatment or a placebo. Not placebo and active groups improved significantly from all owners responded to all questions; the number baseline during every week of the study (p < 0.0001 and percentage of respondents for each question are

FIGURE 3 Plots of model estimated lines (± SD) for Lincoln Sound Sensitivity Scores over the weeks of study. Separate plots are shown for each analysis (intention-to-treat and per-protocol analyses for the sum of the product of intensity and frequency) 6of9 Veterinary Record

provided. Responses were collapsed to make three categories for ‘Agree’, ‘Disagree’ and ‘Neither agree nor disagree’. Chi-square analyses of the distribution of responses showed significant differences for two questions related to efficacy, with non-significant differ- 0.0001 0.0001 ences for the remaining questions (Table 4). Within group comparison (with baseline) < < Thirty-nine of the owners (86.7%) correctly identi- fied whether their dog received the active medication or a placebo. For those owners who were incorrect about their dog’s group, 3 of 30 (10%) from the active group believed their dog had received a placebo, while 3 of 15 (20%) from the placebo group believed their dog had received the active medication.

for each group at each DISCUSSION Within group comparisons 0.0001 34.7 0.0001 36.1 Within group comparison (with baseline) Week 4 < < This study found that treatment with imepitoin twice daily was associated with a decrease in clinical signs of storm anxiety compared to placebo. A higher rate of adverse events (particularly ataxia and increased appetite) was found in the imepitoin group, potentially dose related. Both the placebo and active groups showed significant reductions in signs of anxiety, which is characteristic of the caregiver placebo reported in many other studies both in pain management20,21 and behaviour.8,9,22 Context for these findings can be seen in studies of other ther- 0.0001 31.1 0.0001 30.0 Within group comparison (with baseline) Week 3 < < apies for treatment of anxiety disorders in dogs. For example, in the study of fluoxetine for the treatment of separation anxiety, 40–50% of dogs in the placebo group were rated as improved at each week of the study.22 Similarly, both dexmedetomidine orotrans- mucosal gel and imepitoin – when evaluated during fireworks – had respective placebo effects of 37% and 22% (i.e. owners rated the treatment effect as good or excellent while prescribed placebo).8,9 Specifically for thunderstorms, a nutraceutical containing Magnolia officinalis and Phellodendron amurense was evaluated 0.0001 33.7 0.0001 41.5 23 Within group comparison (with baseline) Week 2 < < in a laboratory model of thunder simulation, and 25% of the dogs receiving placebo showed decreased inactivity (the primary measure of success). Despite the significant caregiver placebo effect seen in the present study, the dogs in the imepitoin group had further decreases in reported signs of anxiety for both the storm logs and the weekly surveys over most weeks of the study. While the differences observed var- Week 1 Mean difference from baseline ied across the weeks of the study, the difference was consistently in the same direction, suggesting efficacy for imepitoin for this indication. On the end of study survey, the owners of dogs in the active group were significantly more likely to report that the treatment reduced their dog’s fear and anxiety during storms and during other (non-storm) noise events than owners in Active 32.69 P-A comparisonActive 14.7 33.0 0.045 19.7 0.012 16.5 0.029 16.7 0.027 P-A comparison 12.38the 0.09 placebo group. 27.9 In general, 0.002 owners in 15.4 both 0.055 groups 17.9 0.025 agreed that the medication was easy to administer (87% of the active group and 80% of the placebo Sum of products of frequency and intensity group). In addition, owners were able to correctly identify which group their dog was in 86% of the time. This finding, however, must be considered in light of the predicted side effect of ataxia after initial dosing, evaluated the change in score fromweek baseline, while the between group comparisons evaluated the difference between the mean difference from baseline TABLE 3 Intention-to-treat Placebo 19.7 0.0002 15.8 0.0021 16.4 0.0015 19.7 0.0002 Per-Protocol Placebo 20.8 0.001 13.7 0.023 14.8 0.019 18.4 0.0021 For both the intention-to-treat and per-protocol analyses, the model scores for mean difference from baseline are shown for each week for each group. Veterinary Record 7of9

FIGURE 4 Plot of estimated lines for Lincoln Sound Sensitivity Scale intensity scores during the baseline period and the treatment period. The explanatory variable was ‘number of days from the start of study’ and separate lines were ﬁt for the baseline period and treatment period. The model evaluated the difference in the slopes of the lines during the treatment period for the placebo and active (imepitoin) groups

which could have influenced owner belief about their noted in studies of diazepam14 and imepitoin16 –was dog’s randomisation. reported in two dogs in the imepitoin group and one The number of adverse events observed does affect dog in the placebo group. Both dogs in the imepi- the interpretation of the study. While expected in 35% toin group had additional adverse effects and were of the dogs (based on the data reported previously8), withdrawn from the study, with all signs resolving on the ataxia and effects on appetite experienced by cessation of treatment; the dog in the placebo group some dogs led to their removal from the study, was continued on-study, and signs resolved within or a negative impression left on their owner. The 7 days. While the current study was not performed prevalence of ataxia (60%) is similar to the reported as a safety study, when adverse events warranted, prevalence in a retrospective study on diazepam use labwork was offered, evaluated and followed. In one in dogs,14 but higher than previously reported for case, significant elevations in alanine transferase and imepitoin for noise reactivity. Previous reports of alkaline phosphatase were found which resolved with imepitoin used in the treatment of canine epilepsy antibiotic treatment and withdrawal of medication. have reported mixed results with regard to adverse Significant elevation in hepatic enzyme concentra- events; Gallucci et al. noted transient adverse effects tions has not been found in other studies of imepitoin that resolved within 10 days in dogs treated for 12 in dogs,8,17,24,25 but has been noted in studies of months,15 while Stabile et al. found that 25% of dogs other benzodiazepines (reviewed in Ref. 26). While had persistent adverse effects during treatment with antibiotics were administered by the referring veteri- imepitoin (not significantly different from pheno- narian, and the dog recovered with supportive care, barbital treatment).16 Aggression – which has been a role for imepitoin cannot be ruled out. Imepitoin’s

TABLE 4 Results from the end of study survey completed by owners

Neither agree Fisher’s Agree nor disagree Disagree Exact (P) p-value

Q1. The treatment was easy for me to Placebo (n = 15) 12 (80.0%) 0 (0.0%) 3 (20.0%) 0.27 0.67 administer to my dog Active (n = 30) 26 (86.7%) 0 (0%) 4 (13.3%) Q2. The treatment reduced my dog’s Placebo (n = 15) 4 (26.7%) 0 (0%) 11 (73.3%) 0.0001 0.0005 fear/anxiety during storms Active (n = 28) 22 (78.6%) 2 (7.1%) 4 (14.3%) Q3. The treatment reduced my dog’s Placebo (n = 15) 2 (13.3%) 2 (13.3%) 11 (73.3%) 2.7e-5 0.0001 fear/anxiety during other noise events Active (n = 24) 18 (75.0%) 3 (12.5%) 3 (12.5%) Q4. After the first few doses, the treatment Placebo (n = 15) 8 (53.3%) 4 (26.7%) 3 (20.0%) 0.027 0.22 effect on my dog was predictable Active (n = 28) 22 (78.6%) 3 (10.7%) 3 (10.7%) Q5. Side effects (if any) were tolerable Placebo (n = 4) 3 (75%) 0 (0%) 1 (25%) 0.23 0.65 Active (n = 24) 11 (45.8%) 1 (4.2%) 12 (50.0%) Q6. Duration of the treatment effect was Placebo (n = 11) 7 (63.6%) 2 (18.2%) 2 (18.2%) 0.12 0.87 acceptable Active (n = 25) 18 (72.0%) 4 (16.0%) 3 (12.0%) Q7. I would choose to continue giving this Placebo (n = 15) 3 (20.0%) 3 (20.0%) 9 (60.0%) 0.01 0.086 treatment to my dog Active (n = 29) 15 (51.7%) 2 (6.9%) 12 (41.4%) For each question, the number and percentage of responses from each group is shown for the categories: Agree, Neither agree nor disagree, and Disagree. Fisher’s Exact tests were used to test for differences in response distributions between groups. 8of9 Veterinary Record advantage over other benzodiazepines is its specific added significantly to the results. The FDA approvals approval for long-term use in dogs with epilepsy, low- for dexmedetomidine oromucosal gel and imepitoin ered risk of tolerance development and lower abuse were both based on studies performed using stan- potential. dardised ratings of behaviours made by owners.8,27 The immediate negative effect of ataxia was One previous study of storm anxiety in dogs did reflected in the withdrawal of five dogs within the use client-owned dogs in a home-setting6; this study first week of starting treatment. This higher than evaluated the effect of clomipramine (given BID) and expected rate of adverse events may be due to the alprazolam (given 1 h prior to storms) and found dosage and dosing schedule of the medication used a 52% improvement in overall signs over 120 days in this study. Imepitoin was used at its approved of treatment. However, this study was an open- dosage and dosing schedule for noise aversion; how- label trial and included a structured desensitisation ever, when imepitoin is used for other, longer-term and counter-conditioning protocol. It is unclear how indications, dogs are typically started at a lower dose much improvement would have been seen with the (approximately 10 mg/kg) and slowly increased to behavioural modification protocol alone. No specific an effective dose. In the case series described by behavioural modification protocol was used in the McPeake and Mills, dogs were initially prescribed a current study. An advantage of the current study’s dosage of 10.6 mg/kg (range 8.6–13.6 mg/kg) BID, design was the twice daily dosing; an additional with dosages titrated for effect; at week 11, dogs were layer of study complexity is added when owners are receiving and average dosage of 17.7 mg/kg (range required to be present prior to a storm in order to 3.9–30.5 mg/kg) BID.18 These authors concluded that treat their dog, with a variable amount of time elapsed ‘a dose of 20 mg/kg BID appears to be an appropriate between treatment and the onset of the storm. Indeed, starting dose for anxiolytic effects in most patients, it is our clinical impression that one source of stress with a fast onset of action with improvements being for owners of dogs with storm anxiety is the inabil- seen within the first week of commencing treatment’. ity to predict when a storm will occur and be avail- It is possible that the 30-mg/kg dosage may be used to able to treat their dogs effectively. As the treatment achieve rapid effect prior to a noise event, but that for was administered daily, this also allowed us to inquire longer term use, such as a storm season, the use of a about the effects of the treatment on other noise trig- lower dose with upward titration for individual dogs is gers in dogs with concomitant noise reactivity; an preferable. important application given the high co-morbidity of There were limitations to the study that merit dis- storm anxiety and fear of other noises.4 However, with- cussion. One limitation is inherent in the study of out baseline data regarding other noise triggers, these storm anxiety: the inability to predicts storms or con- results remain the impression of the owner rather than trol the number of storms a given patient is exposed to being directly compared. during the treatment period. We were able to extend the baseline period to ensure that every patient had at least two storm logs to form the basis of their baseline CONCLUSIONS anxiety scores; however, once the treatment period was started, the number, duration and intensity of the In conclusion, this study showed that twice daily storms was unable to be controlled. Other studies of administration of imepitoin, compared to placebo, storm anxiety have avoided this limitation by relying decreased anxiety scores in dogs with storm anxiety. on laboratory models13,23; while important, these do We used a conservative data imputation method, not provide as much ecologic validity as field studies. favouring the inability to reject the null hypothesis As noted in these studies, the lack of other signs asso- (that there would be no difference between groups) ciated with storms (changes in barometric pressure, rather than incorrectly rejecting the null hypothesis etc.) means that the response seen is limited to the (claiming a difference where none existed). Over 4 effect of the noise of the thunder. This difficulty is weeks of study, dogs receiving imepitoin had sig- likely a reason for the relative absence of studies spe- nificantly decreased anxiety scores from baseline cific to storm anxiety in dogs. In this study, the decision for weeks two through four, with significantly lower was made to standardise the length of time rather than scores than placebo for weeks two and four. Own- the number of storms dogs were exposed to during the ers reported that the medication was easy to give, treatment period. This standardisation was selected as and that it was decreased their dogs’ fear and anx- duration of treatment could affect response (positively iety during storms and during other noise events. or negatively) and because dogs were randomised over A higher than expected rate of adverse events was a relatively short window of time (May through July). observed, particularly ataxia and changes to appetite. Future studies could evaluate an alternate design Given the efficacy of the treatment, and its ability to where the number of storms, rather than length of be given daily during a period of frequent storms, treatment, is standardised. Finally, this study relied on future work should be pursued to determine whether owner ratings of behaviours, rather than any physio- a lower dose would provide similar efficacy with fewer logic or video-based behavioural coding. Behavioural adverse events. Based on previous work highlighting coding would have provided a further measure of effi- the efficacy of lower dosages and the advantage of cacy for the treatment groups, but would have been individually tailored dosing,18 a lower dose is likely to logistically difficult to standardise and may not have be beneficial. Veterinary Record 9of9

ACKNOWLEDGEMENTS 11. Engel O, Masic A, Landsberg G, Brooks M, Mills DS, Rundfeldt The authors would like to thank Dr. Barbara Sherman C. Imepitoin shows benzodiazepine-like effects in models of for her role in the design of the study and reviewing anxiety. Front Pharmacol. 2018;9. 12. Ibanez M, Anzola B. Use of fluoxetine, diazepam, and behavior results. We also thank all the owners who participated modification as therapy for treatment of anxiety-related disor- in this study with their dogs. ders in dogs. J Vet Behav. 2009;4(6):223-9. 13. Araujo JA, de Rivera C, Landsberg GM, Adams PE, Milgram NW. FUNDING AND COMPETING Development and validation of a novel laboratory model of INTERESTS sound-induced fear and anxiety in Beagle dogs. J Vet Behav. 2013;8(4):204-12. This study was funded by Boehringer Ingelheim 14. Herron ME, Shofer FS, Reisner IR. Retrospective evaluation of with input into the study design and review of the the effects of diazepam in dogs with anxiety-related behavior manuscript. No input was provided for the data col- problems. J Am Vet Med Assoc. 2008;233(9):1420-4. lection, analysis, or interpretation of the data. PW is an 15. Gallucci A, Gagliardo T, Menchetti M, Bianchi E, Bucci D, employee of Boehringer Ingelheim, the manufacturer Gandini G. Long-term efficacy of imepitoin in the treatment of naive dogs affected by idiopathic epilepsy. Vet Rec.144-, of Pexion. MEG has served as a consultant and spon- 2017;181(6). sored speaking for Boehringer Ingelheim. 16. Stabile F, van Dijk J, Barnett CR, De Risio L. Epileptic seizure frequency and semiology in dogs with idiopathic epilepsy after AUTHORS CONTRIBUTION initiation of imepitoin or phenobarbital monotherapy. Vet J. MEG, CE, BN and PW designed the study; IR, CE, BN, 2019;249:53-7. 17. Rundfeldt C, Tipold A, Loscher W. Efficacy, safety, and tolera- MEG were responsible for data collection and data bility of imepitoin in dogs with newly diagnosed epilepsy in a quality; EHG, YS and VD analysed the data; IR and randomized controlled clinical study with long-term follow up. MEG drafted the manuscript; all authors participated BMC Vet Res. 2015;11. in review and editing of the manuscript. 18. McPeake KJ, Mills DS. The use of imepitoin (Pexion (TM)) on fear and anxiety related problems in dogs – a case series. BMC Veterinary Research. 2017;13. ORCID 19. Mills DS BDM, Zulch H. Stress and pheromonatherapy in small Vaidehi Dixit https://orcid.org/0000-0002-7157- animal clinical behaviour. Chichester, UK: Wiley-Blackwell; 3354 2013. Emily H. Griffith https://orcid.org/0000-0002-9745- 20. Conzemius MG, Evans RB. Caregiver placebo effect for dogs 7722 with lameness from osteoarthritis. J Am Vet Med Assoc. 2012;241(10):1314-9. Margaret E. Gruen https://orcid.org/0000-0002- 21. Gruen ME, Dorman DC, Lascelles BDX. Caregiver placebo 6036-8849 effect in analgesic clinical trials for cats with naturally occur- ring degenerative joint disease-associated pain. Vet Rec. -, 2017;180(19): 473. REFERENCES 22. Landsberg GM, Metese P, Sherman BL, Neilson JC, Zimmer- 1. Ballantyne KC. Separation, confinement, or noises: What is man A, Clarke TP. Effectiveness of fluoxetine chewable tablets scaring that dog? Vet Clin North Am Small Anim Pract. in the treatment of canine separation anxiety. J Vet Behav. 2018;48(3):367-86. 2008;3(1):12-9. 2. Blackwell EJ, Bradshaw JWS, Casey RA. Fear responses to 23. DePorter TL, Landsberg GM, Araujo JA, Ethier JL, Bledsoe noises in domestic dogs: Prevalence, risk factors and co- DL. Harmonease chewable tablets reduces noise-induced fear occurrence with other fear related behaviour. Appl Anim Behav and anxiety in a laboratory canine thunderstorm simula- Sci. 2013;145(1-2):15-25. tion: A blinded and placebo-controlled study. J Vet Behav. 3. Overall KL, Dunham AE, Juarbe-Diaz SV. Phenotypic determi- 2012;7(4):225-32. nation of noise reactivity in 3 breeds of working dogs: A cau- 24. Rundfeldt C, Gasparic A, Wlaz P. Imepitoin as novel treat- tionary tale of age, breed, behavioral assessment, and genetics. ment option for canine idiopathic epilepsy: pharmacokinetics, J Vet Behav. 2016;16:113-25. distribution, and metabolism in dogs. J Vet Pharmacol Ther. 4. Lopes Fagundes AL, Hewison L, McPeake KJ, Zulch H, Mills 2014;37(5):421-34. DS. Noise sensitivities in dogs: an exploration of signs in dogs 25. Tipold A, Keefe TJ, Loscher W, Rundfeldt C, de Vries F. Clinical with and without musculoskeletal pain using qualitative con- efficacy and safety of imepitoin in comparison with phenobar- tent analysis. Front Vet Sci. 2018;5:17. bital for the control of idiopathic epilepsy in dogs. J Vet Phar- 5. McPeake K, Affenzeller N, Mills D. Noise sensitivities in dogs: a macol Ther. 2015;38(2):160-8. new licensed treatment option. Vet Rec. 2017;180(14):353-5. 26. Skolnick P. Anxioselective anxiolytics: on a quest for the Holy 6. Crowell-Davis SL, Seibert LM, Sung WL, Parthasarathy V, Curtis Grail. Trends Pharmacol Sci. 2012;33(11):611-20. TM. Use of clomipramine, alprazolam, and behavior modifica- 27. Korpivaara M, Laapas K, Huhtinen M, Schoning B, Overall K. tion for treatment of storm phobia in dogs. J Am Vet Med Assoc. Dexmedetomidine oromucosal gel for noise-associated acute 2003;222(6):744-8. anxiety and fear in dogs-a randomised, double-blind, placebo- 7. Tiira K, Sulkama S, Lohi H. Prevalence, comorbidity, and behav- controlled clinical study. Vet Rec. 2017;180(14):356. ioral variation in canine anxiety. J Vet Behav. 2016;16:36- 44. 8. Engel O, Muller HW, Klee R, Francke B, Mills DS. Effective- ness of imepitoin for the control of anxiety and fear associated with noise phobia in dogs. J Vet Intern Med. 2019;33(6):2675- How to cite this article: Perdew I, Emke C, 84. Johnson B, et al. Evaluation of Pexion® 9. Korpivaara M, Laapas K, Huhtinen M, Schoning B, Overall K. (imepitoin) for treatment of storm anxiety in Dexmedetomidine oromucosal gel for noise-associated acute dogs: A randomised, double-blind, anxiety and fear in dogs-a randomised, double-blind, placebo- placebo-controlled trial. 2021;e18. controlled clinical study. Vet Rec. 2017;180(14):356. 10. Moesta A. Animal behavior case of the month. JAVMA-J Am Vet https://doi.org/10.1002/vetr.18 Med Assoc. 2014;244(5):538-40.