Survey of canine use and safety of isoxazoline parasiticides

Abstract A veterinarian and pet owner survey (Project Jake) examined the use and safety of isoxazoline parasiticides given to dogs. Data were received during August 1–31, 2018 from a total of 2,751 survey responses. Forty‐two percent (1,157) reported no flea treatment or adverse events (AE), while 58% (1594) had been treated with some parasiticide for flea control, and of those that received a parasiticide, the majority, or 83% (1,325), received an isooxazoline. When any flea treatment was given, AE were reported for 66.6% of respondents, with no apparent AE noted for 36.1%. Project Jake findings were compared to a retrospective analysis of publicly available Food and Drug Administration (FDA) and European Medicines Agency (EMA) reported AE. The number of total AE reported to FDA and EMA were comparable, although a 7 to 10 times higher occurrence of death and seizures was reported from the EMA or from outside the United States (US). Serious AE responses for death, seizures and neurological effects reported in our survey were higher than the FDA but moderately lower than the EMA reports. These sizable global data sets combined with this pre‐ and post‐parasiticide administration survey indicated that isoxazoline neurotoxicity was not flea‐ and tick‐specific. Post‐marketing serious AE were much higher than in Investigational New Drug (IND) submissions. Although the labels have recently been updated, dogs, cats and their caregivers remain impacted by their use. These aggregate data reports support the need for continued cross‐species studies and critical review of product labelling by regulatory agencies and manufacturers.


| INTRODUC TI ON
The isoxazoline class of compounds are ligand-gated ion channel inhibitor parasiticides that have been developed for at least 20 years (Garcia-Reynaga, Zhao, Sarpong, & Casida, 2013;Nakata et al., 2017;Ozoe, Ozoe, Nakahira, & Mita, 2010;Quan et al., 1999;Shoop et al., 2014;Zhao & Casida, 2014). When initially designed and synthesized as a series of bisbenzamidine isoxazolines by Dupont Pharmaceuticals, their purpose was to act as antithrombotic agents that inhibit the activity of DOI: 10.1002/vms3.285 coagulation factor Xa, although side effects included excessive bleeding and hemolysis (Quan et al., 1999). The earlier reports of these synthesized compounds and their mechanisms of action laid the groundwork for subsequent pre-clinical studies as parasiticides, which were described in a special issue of Veterinary Parasitology in 2014 (Drag, Saik, Harriman, & Larsen, 2014;Shoop et al., 2014). Subsequent to their development as parasiticides they are now widely used throughout the world and are considered by most to be efficacious and safe, even in 8-week old puppies Kuntz & Kammanadiminti, 2017).
However, despite the estimated millions of doses being given, adverse events (AE) that had previously been believed to be rare (Gaens, Rummel, Schmidt, Hamann, & Geyer, 2019;Quan et al., 1999), post-marketing AE have been increasingly reported since the release of these products.
More recently, Gaens et al. (2019) reported suspected transient neurotoxicity in a 7-month old female Danish spaniel (Kooikerhondje) puppy. About 24 hr after administering fluralaner, she exhibited generalized ataxia, myoclonic jerks, tremor of the head and body, muscle twitching and oral dysphagia; she fully recovered after 10 hr without treatment (Gaens et al., 2019). Factors including genetic predisposition that could predispose an individual animal to AE are unknown, although the effects of an MDR1 efflux carrier at the blood-brain barrier could play a role (Saidijam, Dermani, Sohrabi, & Patching, 2017).
The impetus for the present survey inquiry into the potential AE associated with the use of canine flea and tick treatments was prompted by non-AE documented reports and clinical experiential AE reports from veterinarians and pet owners worldwide.
These reports were apparently disregarded or failed to generate concern and responsiveness from the manufacturer(s), which was compounded by the failure to report serious AE on their package labelling, until recent label changes were made from 2018 to 2019.
Consequently, the present team of experts was formed to gain an objective understanding of any toxic effects that may be related to the isoxazolines.
Isoxazolines exhibit killing activity through effects on the nervous system. Isoxazoline efficacy is based on their modality as non-competitive GABA (gamma-aminobutyric acid) receptor antagonists that are putatively much more selective for GABA receptors in fleas or ticks, than for those in mammals, including humans (Gassel, Wolf, Noack, Williams & Ilg, 2014;Shoop et al., 2014). They bind to arthropod gamma-aminobutyric acid (GABA)-gated chloride channels (GABACls) in nerve and muscle cells, which blocks the transmission of neuronal signals. Some of the currently registered isoxazolines claim an adequate safety margin for canine and feline administrations and that the mechanism of action is insect-specific.
However, although some recently marketed isoxazoline package inserts do mention seizures or other neurotoxic effects, these effects, including death, are not insect-specific. The purpose of the current study was to perform an independent survey to document the use and safety of the isoxazoline parasiticides pre-and post-administration.

| Survey
A Survey questionnaire instrument was developed in conjunction with the Project Jake team. It was distributed electronically by mail throughout the United States to veterinarians, veterinary clients, pet caregivers/owners, kennel club groups and on social media sites between August 1 and 31, 2018. The questionnaire addressed the respondent's use of a parasiticide for flea and/or tick control on their dogs pre-and post-administration. We chose to focus just on the dog initially rather than including cats. While a reporting bias can arise for participants of this type of survey, respondents were encouraged to complete initial questions even if they had not seen or were unaware of any AE from the use of isoxazoline or other flea or tick preventive compounds. Pitfalls of electronic online surveys include: absence of interviewer input, may not effectively reach the intended population, and survey bias or fraud (Evans & Mathur, 2018;Nayak & Narayan, 2019;Rice, Winter, Doherty, & Milner, 2017;Vaske, 2011). The authors acknowledge these potential shortcomings of the survey instrument, but believe that the findings described herein reflect valid concerns. In addition to the survey data shown here, the findings were compared to the reported FDA and EMA data. Freedom of Information Act (FOIA) requests were sent to the FDA concerning their isoxazoline AE reports (see Appendix,[1][2][3][4][5][6][7][8]. For the EMA reports, the findings reported by the Committee for Medicinal Products for Veterinary Use (CVMP) at the meeting of 5-7 September 2017 were used. These represented the CVMP Cumulative Reports from Jan 2013 to Sept 2017 (see Appendix, # 9).
An analysis of the occurrence of AE was conducted to compare the number of Project Jake AE, with the FDA and EMA AE pharmacovigilance reports, as listed in the Appendix, #1-9.

| Data analysis and limitations
Analysis of the survey data and of reports made to regulatory agencies consists of event tabulations and summary calculations of percentages of occurrence for AE with a focus in the present summary on serious neurologic effects and deaths reported in canines.
Statistical analysis was made for the severe AE of: death, seizures and convulsions, and ataxia, instability and imbalance by comparing the two-tailed Z-test differences between the Project Jake survey with those reported by the FDA and EMA. The number of reports represents the number of AE received for a particular drug and do not infer the relative toxicity between drugs. AE reports should not be used to calculate incidence rates or estimates of drug risk, since there is no accurate way to determine how many animals were actually given the drug, which is needed as the denominator in calculations of incidence and relative risk. Beyond the manufacturer's advertising and marketing materials, there is no indication of how many of the products have been sold. Furthermore, neither the FDA nor EMA address the total doses or treatments given in their data, and the package inserts do not delineate how many animals were treated and how many had serious AE 7,8;Appendix).
Since an AE may have been related to an underlying disease, use of other drugs at the same time, or other non-drug related causes, these circumstances create uncertain causality between the adverse event and treatment with the drug. The frequency of reporting for a given product may vary over time and may be higher when the drug is newly marketed, or when media publicity occurs, thus underreporting generally occurs with most AE reporting systems.
Regulatory AE reporting systems as well as the Project Jake survey are all voluntarily reported with inherent biases weighted towards adverse effects, since few pet caregivers/owners or veterinarians have cause to report the absence of side effects. However, the Project Jake survey included questions and responses about drug treatment with or without the observation of AE, and thus it provided a population of responders registering treatment without AE.
In fact, 36.1% of the Project Jake survey respondents indicated that they did not see any AEs and another 8.2% were unsure (Table 1).

TA B L E 1 Summary of Project Jake Survey Respondent Numbers and Treatments
Summary: Isoxazoline Survey -August 1-31, 2018

| RE SULTS
The canine products with the largest number of survey responses from the FDA, EMA and Project Jake survey were afoxolaner ( Tables 2 and 8).

| FDA and EMA adverse reaction (AE) reports (Tables 2-4; Appendix)
The authors reviewed FDA and EMA reported AE available via public access of FOIA-requested FDA AE reports (in Appendix; #1-8) and the updated EMA AE reports (Appendix, #9). An analysis of these AE types is presented in Table 2 and the subsequent Tables (3-7), and includes analysis of FDA, EMA and Project Jake survey results with a particular focus on the most severe AE, namely, death and neurological effects (Tables 6-9).
FDA reportable AE events from January 2013 to September 2017 are listed in Tables 2 and 3. These data show 32,374 report- EMA AE reports from January 2013 to January 2019 (Table S1), which reflects about a fourfold increase from September 2017 to January 2019. During this extended reporting period, EMA data reflected notably higher serious AE for the four isoxazolines, ranging from 4.76% to 28.6% for deaths and 9.0 to 55.1% for seizure observations. Note that these EMA data included AE results for spinosad (6,080, Tables 8 and Table S1) with death (30.9%) and seizure (24%) rates markedly higher than serious AE rates for fluralaner, afoxolaner and sarolaner. The AE considered most serious with respect to neurological toxicity are listed at the top of Tables, 3-6, and the remaining reported AE are generally grouped by percentage occurrence and physiologic significance.

| Serious AE findings from the FOIA-requested CVM comprehensive clinical detail report listing (for comparable cumulative date range of January 2013 to September 2017)
These data are presented in Tables 3 and 8. Deaths for the fluralaner, afoxolaner and sarolaner isoxazolines in the FDA AE reports ranged from 2.5% to 3.2%. However, seizures, trembling and ataxia ranged from 2.8% to 7.5% for afoxolaner and fluralaner but were markedly higher for sarolaner at 20.50 and 41.1% (Tables 2, 3 and 8). Additionally, behavioural, cognitive and muscular AE observations for sarolaner were ~ 1.5 to 2 times higher than those noted for afoxolaner and fluralaner (Table 3).
Findings from the EMA AE reports for the same date range of January 2013 to September 2017 are presented in Table 4 Table 3 for the FDA AE data. Of major significance was the fact that deaths in the EMA data were approximately 7 to 10 times higher than those reported in the corresponding time period to the FDA. Serious AE for seizure and tremor events in the EMA data were also notably higher than those reported to the FDA (Table 4). Furthermore, sarolaner reports were markedly higher than for afoxolaner and fluralaner, while the behavioural, cognitive and muscular AE observations in the EMA report data for sarolaner were similar to those noted for afoxolaner and fluralaner.
In comparing the updated cumulative data became available from the EMA for the extended period of January 2013 to January 2019. F I G U R E 1 EMA time from dosing to adverse event The overall number of EMA AE reports (39, 148) for afoxolaner, fluralaner and sarolaner in this update was essentially the same as those of the FDA events (32,374), as shown on Tables 2 and 3. Additionally, death and seizure reports for lotilaner and spinosad became available and are listed on Tables 2 and 8 for comparison. This larger EMA data set, which included an additional 15 months of reports, shows death and seizure events for afoxolaner comparable to the FDA findings with over 32,000 reported events. But, fluralaner and sarolaner EMA reported death and seizure events remained approximately 7 to 10 times higher than the reported FDA events. Remarkably, the death and seizure reports for lotilaner and spinosad were also about 10 times higher than those reported in the FDA data for the isoxazolines (Table 8).
A relationship between the occurrence of death and/or seizures was examined for the EMA cohort data from the January 2013 to September 2017 period versus January 2013 to January 2019 (Table S1). Comparing these two time periods, not only is there a substantial increase in AE reports, but the percentage of seizures and deaths is still remarkable even with a larger data set. The death rate range per isoxazoline type for 2013-2017 was 13%-24% versus 5%-31% for 2013-2019. The seizure rate range per isoxazoline type for 2013-2017 was 19%-60% versus 9%-55% for 2013-2019. The major difference in the two data sets is the type of isoxazoline used with the increase in afoxolaner use, and the introduction of new isoxazolines. These data indicated that not all dogs with seizures died, but the converse result could not be determined. With respect to pet age, serious AE showed death reported in about 60% of dogs older than 5 years with seizures in more than 50% of dogs older than 5 years (Table 9). An analysis of possible concurrent disease or health status was not possible for this data set. An analysis of the reported death and seizure AE associated with product dosage was inconclusive, although the data suggested that these serious AE occurred at most dosages and the Fluralaner dose of 1,000 mg appears to be an outlier, with 310 reports of seizures (30%) and 379 reports of death (38%), some of these could be the same animal, of which cannot be confirmed. (Table S2). The most serious AE as stated in the EMA cumulative reports were observed at 0-24 hr after the first dose, and then again after the second and third doses (Figure 1), which was consistent with the known isoxazoline pharmacokinetics of absorption and elimination following administration of oral chews (Letendre et al., 2014). Additionally, the pharmacokinetics and halflife of isoxazolines were discussed by Drag et al. (2014).

| Project Jake survey AE reports (Tables 1, 5, 6)
The Project Jake survey data results are shown in Tables 1, 5 and 6 and consolidated in Table 8, in comparison to the FDA and EMA data. Table 1 lists the survey findings for isoxazolines and an array of other topical flea and tick parasiticide products administered by survey responders; the number (N) of treated pets shows the percentages of the total number of responders (2,751 However, not all dogs presenting with seizures, ataxia, instability, imbalance, shaking and trembling progressed to death ( Table 6).
The number of AE reported by respondents in the Project Jake survey is shown in Table 7 compared to those AE listed in the manufacturer package inserts for fluralaner and afoxolaner. This tabulation demonstrates that no serious AE are listed in the inserts for either drug (Table 7). Serious neurological AE findings from the FDA, EMA and Project Jake survey are consolidated in Table 8. Seizures: EMA = 16.02%, Project Jake = 13.74% (Tables 6 and 8).

| Statistical analysis
The Project Jake survey percentages of AE were below those reported by the EMA but much higher than those reported by the FDA. However, the p-values for the two-tailed Z-test differences between the Project Jake survey and fluralaner, afoxolaner and sarolaner for both the FDA and EMA reported AE were all < 0.00001.
The statisticians in our group have indicated that an in-depth statistical analysis of the rest of the survey findings would not be applicable to this type of survey.

| D ISCUSS I ON
The Project Jake survey results shown here differed from those of the FDA and EMA AE reports in several ways, namely: (1) Pet caregivers/owners surveyed included those that did not see any AE, as well as those that did report them, providing a more unbiased review of isoxazoline use; (2)   AE; (5) Other concurrent health issues and treatments given were requested; (6) Multiple AE were listed for each report; (7) Recovery information and/or follow-ups were included.
The consolidated FDA, Project Jake and EMA findings (Table 8) showed notable differences between survey populations regarding the percentage of neurological toxicity and serious AE, and fatal effects. Statistical analysis of these serious AE showed highly significant differences between the findings of the Project Jake survey and those reported by the FDA and EMA. While the number of death and seizure AE reported by the EMA was 7 to 10 times higher than those reported to the FDA, the reported responses for the Project Jake survey for death and seizures fell in between those of the FDA and EMA but aligned more closely with the EMA results. Furthermore, the number of reported death and seizure AE for lotilaner and spinosad were considerably higher than suggested with respect to their product labelling for potential neurological effects ( Based upon the FOIA FDA and EMA summary findings and implications, the death and seizure rate differences noted for the EMA were about 3-10 times higher than the FDA (Table 8). Were these FDA and EMA AE reports filed by veterinarians or by pet caregivers/owners or both? Additionally, analysis of the EMA data indicated that more than 60% of canine deaths and 55% of seizures were in dogs > 5 years of age ( In conclusion, isoxazolines exhibit killing activity through effects on the insect nervous system. Prior reports (Gassel et al., 2014) showed that the isoxazoline fluralaner did not demonstrate any inhibitory action on the GABACl channels of rats and suggested that fluralaner is a potent arthropod -specific GABACl channel inhibitor.
However, FDA and EMA AE reports and Project Jake survey evidence consistently demonstrate in three separate, distinct data sets that the neurotoxicity is not arthropod-specific, and that post-marketing serious AE are much higher than in the IND submission studies. Thus, as a class of drugs, these data indicate that isoxazolines can work as intrinsic neurotoxins across species. As pre-approval IND studies were done in only a limited number of animals that did not show neurotoxic serious AE, it is not surprising that higher frequencies of such AE were noted once they were commercialized and given to much larger populations.
The data also suggested notable differences in reporting of AE for the US versus European territories. This may reflect cultural tendencies and/or procedural methodologies for reporting of AE. Nevertheless, the national and international post-market safety survey data summarized here indicate an immediate need for continued cross-species studies and a critical review of product labelling by regulatory agencies and manufacturers. Moreover, this class of drugs had more serious AE than those reported in the package inserts. We believe that the FDA should consider additional changes to their criteria for AE observation, and define what would be needed for future clearances of this class of drugs.

CO N FLI C T O F I NTE R E S T
The authors disclose no financial interest nor conflicts-of-interest in the present study. Prior to the FDA alert informing pet owners and veterinarians of potential for neurological adverse events for isoxazolines on September 21, 2018 product labelling did not contain precautionary or warning labelling of these AE. The excerpted statements are representative of (slightly) modified package labels containing cursory comments related to seizures. The Canadian fluralaner label version of 2019 lists seizure with the qualifier of occurrence rates of less than 1 in 10,000.

AUTH O R CO
Fluralaner 2019 package labelling in Canada states: "The following adverse events have been reported rarely: 1(reported in at least 1 but not more than 10 animals in 10,000 animals exposed) and very rarely: 2(reported in less than 1 in 10,000 animals exposed) and are listed by body system, in decreasing order of frequency: Digestive