Activity Levels in Seven Cats with Osteoarthritis Monitored by GPS Tracker Following Frunevetmab Injection

Abstract Objective  The goal of the current pilot study was to use a commercially available accelerometer (Tractive GPS Tracker for Cats) in a setting evaluating the activity level of cats exhibiting different forms of osteoarthritis in their natural habitat before and after treatment with frunevetmab, using each individual cat as its own control. We hypothesized that treatment with frunevetmab would be associated with increased activity level. Methods  Activity data, measured as minutes of active time per day via accelerometer worn on a collar, were obtained from seven employee-owned cats with radiographically confirmed evidence of osteoarthritis. Cats were randomly assigned to two different protocols in which treatment and no treatment (control) were reversed; that is, undergoing the control period prior to treatment with frunevetmab (“standard” control) versus first assessed without treatment (control period) followed by frunevetmab treatment after minimum 30-day washout period (“inverse” control period). Results  Activity time, on average, was 30 minutes longer per day for frunevetmab versus control ( p  = 0.007). Conclusions  Based on measurements obtained with piloted use of the Tractive, the activity level in osteoarthritic cats appeared greater when treated with frunevetmab compared to self-control without frunevetmab treatment (either pretreatment or following washout period). The small sample size here does warrant caution in interpreting results for a broader population.


Introduction
Osteoarthritis (degenerative joint disease) has a high prevalence among domestic cats, with anywhere from 64 to 90% having radiographic evidence, and 45% exhibiting clinical signs of discomfort and/or decreased mobility. 1,2Historically, feline osteoarthritis has been underdiagnosed and undertreated. 3Growing recognition of this disease has become evident as practitioners seek effective long-term treatment options to improve comfort and quality of life for arthritic cats.While nonsteroidal anti-inflammatory drugs can be effective, they carry risk of kidney damage, precluding use in the many aging cats with concurrent chronic kidney disease. 4Other drugs such as gabapentin, 5 tramadol, [6][7][8] and amantadine 9 have shown varied success in previous studies, however, can cause stress for cats that are not amenable to receiving indefinite daily oral medication.
Nerve growth factor (NGF) has been identified as a therapeutic target for treating osteoarthritis.NGF is pronociceptive, and can trigger peripheral sensitization as well as potentiate inflammatory responses. 10,11Use of anti-NGF monoclonal antibodies has been investigated in companion animals. 12A feline anti-NGF monoclonal antibody (frunevetmab) was demonstrated to exhibit an acceptable safety and efficacy profile in a feline model of chronic pain. 13Frunevetmab is administered as a once-monthly injection and has been approved by the Food and Drug Administration for use in cats with osteoarthritis (Solensia, Zoetis Animal Health, Parsippany, New Jersey, United States). 14 preliminary proof-of-concept study 15 and a multi-site pilot field study 16 have further demonstrated efficacy of frunevetmab in treating chronic pain associated with osteoarthritis in client-owned cats.Outcome assessments incorporated both subjective factors, including clinical metrology assessments completed by owners (client-specific outcome measures and feline musculoskeletal pain index [FMPI]) and physical examination findings, as well as objective factors, including accelerometry data gathered via activity monitor. 15,16ctivity monitors that have been evaluated in the current literature include the Actical (Respironics, Mini Mitter Division, Bend, Oregon, United States), 17,18 PetPace (PetPace LLC, Burlington, Massachusetts, United States), 19,20 and PlusCycle (JARMeC, Kanagawa, Japan). 21The Actical has been validated in the dog 17 and has been shown to correlate well with computer-analyzed video movement data in the cat 18 ; however, the latter study noted that there was variability between subjects, and so the authors advised that future accelerometer-based studies in cats should use each cat as its own control. 18The PetPace, while demonstrated to correlate well with data from the validated Actical, has only been evaluated in dogs. 19,20The PlusCycle demonstrated strong correlations with Actical data in cats, 21 however, is only commercially available in Japan at this time.
The Tractive GPS Tracker for Cats (Tractive Inc, Seattle, Washington, United States) has not been previously evaluated in clinical studies.It offers a simplified user interface whereby a pet owner can easily view and understand their pet's activity data.Compared to the PetPace, the Tractive is offered at a more reasonable retail price ($49.99 vs. $299.99,plus subscription fees, at the time of this writing) and is also lower profile (35 g device that can be affixed to pet's usual collar vs. 60 g standalone collar).Size and comfort of the activity monitor is an important consideration in tolerating wear, particularly for feline patients; indeed, previous work has shown that a more cumbersome harness-type fixation of an activity monitor was associated with decreased activity level in cats over an acclimation period. 18hile clinical metrology assessments such as the FMPI have been refined and shortened to encourage responsiveness, consistent owner evaluation of their cats at home remains a major impediment in assessing pain and activity level. 21Therefore, the Tractive, a relatively small, inexpensive activity monitor with simplified data output, shows promise as a tool to objectively evaluate feline activity levels at home with improved owner responsiveness.
The goal of the current study was to pilot use of the Tractive activity monitor in a clinical study evaluating the activity level of cats when treated with frunevetmab.The study was designed to be a simple, objective, single-outcome, paired analysis, using each individual cat as its own selfcontrol.We hypothesized that treatment with frunevetmab would be associated with increased activity level compared to control.

Study Population
Employee-owned cats with suspected osteoarthritis were recruited for enrollment via email listserv and flyers placed in the hospital.Owner informed consent was obtained.Animal ethics approval was not obtained as frunevetmab injection was administered per its labeled use.Cats were eligible for inclusion if they (1) had radiographically documented evidence of osteoarthritis in at least one joint, (2) tolerated wearing of the activity monitor on a neck collar, and (3) were not currently receiving analgesics, anti-inflammatory medication, or supplements to treat osteoarthritis.

Treatment Categories
Cats were randomized via coin-toss to undergo the control period either prior to treatment with frunevetmab ("standard" control, cats A, B, C, D), or after frunevetmab treatment following minimum 30-day washout period ("inverse" control, cats E, F, G; washout periods, 49 days, 35 days, and 68 days, respectively).The inverse control period was utilized in an effort to minimize variation in acclimation period for monitor wear .Each cat was administered a 7 mg subcutaneous injection of frunevetmab (Solensia 7 mg/mL solution in a single-use 1 mL vial, Zoetis Animal Health, Parsippany, New Jersey, United States) 13 on "day 0" of the treatment period.

Activity Data Collection
Owners were instructed to secure a GPS-based accelerometer (Tractive GPS Tracker for Cats, Tractive Inc., Seattle, Washington, United States) to the cat's collar.For the experimental period, activity data were gathered beginning no sooner than "day 1" following treatment.For the standard control period, activity data were gathered immediately; for the inverse control period, activity data were gathered following a minimum 30-day washout period after treatment with frunevetmab.This washout period was justified on the basis of: (1) previous work demonstrating an average plasma half-life of 9 days (range: 7-15 days) following a 2 mg/kg subcutaneous injection of felinized anti-NGF monocloncal antibody, 13 and (2) manufacturer label claim recommending monthly administration of frunevetmab for therapeutic efficacy, 14 citing previous multisite pilot field study. 16he Tractive monitor contains a GPS module as well as an accelerometer.The built-in SIM card connects to cellular data networks via LTE/4G technology, which transmits positional data for the inbuilt GPS module to calculate location via VCOT Open Vol. 7 No. 2/2024 © 2024.The Author(s).
satellite.The "LIVE Tracking Mode" provides location updates to the system every 2 to 3 seconds.The accelerometer outputs data on the user interface as active minutes per day (the number of minutes the cat is was moving, vs. at rest/stationary).Data are transmitted wirelessly to a smartphone-based application where it can be reviewed. 22For this study, data were transmitted directly to the investigators.
A minimum of 7 complete days (24 hours) of activity data were obtained for each cat in each treatment category (control, experimental).The monitor required charging approximately every 48 hours, during which time it would need to be removed from the cat.Due to necessary charging cycles, as well as owner compliance (e.g., forgot to replace monitor on cat's collar for several days following charging period), days following treatment were not consecutive.Days that did not include a full 24 hours of data were not included in analysis.Activity measurements were obtained over a range of 17 days (►Table 1).All cats were indoor-only and maintained in their usual home environment for the entirety of each period in the study.Homes varied from single-level apartment flats to multi-story homes.Owners were instructed not to restrict, confine, or otherwise alter the cat's usual access within the home.

Statistical Analysis
Two data sets with measures of activity were initially evaluated.The first data set contained seven measurements of each patient for each treatment category (control, experimental), numbering those days 1-7.The second data set also contained seven measurements of each patient for each treatment category (control, experimental); however, in this set, measurements were numbered as the actual day that activity was measured over the 17 days of observation.There were missing data, and the model failed to converge when all 17 days were included.The model successfully converged when only actual days 1-8 were included.Mixed-models analysis of variance (treatment category and time fixed, cat random) was performed and revealed no effect of day (p ¼ 0.94 and p ¼ 0.64) or interaction of day Ã treatment category (p ¼ 0.74 and p ¼ 0.55), respectively, in the two analyses.Normality of the errors was confirmed by means of histogram and normal probability plot.
Therefore, we considered the seven measurements obtained over the study period of each patient for each treatment category (control, experimental) as random repetitions, and obtained the mean of these values.A Shapiro-Wilk test was performed and confirmed that the differences in activity were normally distributed (p ¼ 0.38).A paired ttest was then performed to compare frunevetmab versus control.Data were reported as mean and standard deviation.p ¼ 0.05 was considered significant.

Results
Seven cats were eligible for study inclusion.There were four castrated males and three spayed females.The mean age was 10.71 years (range: 7-15 years).The mean weight was 5.48 kg (range: 3.0-9.5 kg).Average dose of frunevetmab administered was 1.5 mg/kg (range: 0.7-2.3mg/kg).Breeds represented included domestic short hair (4), domestic medium hair (1), domestic long hair (1), and Norwegian forest cat (1).Physical examination was performed on all cats prior to initiation of the study and was reactive upon manipulation of the affected joint(s).
Individual cat descriptions were as follows: • Cat A: 13-year-old male neutered Norwegian forest cat, body weight 6.75 kg, two joints affected (bilateral elbows).• Cat B: 8-year-old female spayed domestic long hair, body weight 5.5 kg, two joints affected (bilateral coxofemoral joints).• Cat C: 8-year-old male castrated domestic short hair, body weight 3.36 kg, one joint affected (right elbow).• Cat D: 15-year-old female spayed domestic short hair, body weight 3.18 kg, one joint affected (left elbow).
• Cat E: 7-year-old male castrated domestic short hair, body weight 7.04 kg, four joints affected (bilateral coxofemoral joints, bilateral stifles).• Cat F: 15-year-old female spayed domestic medium hair, body weight 3.0 kg, four joints affected (bilateral coxofemoral joints, bilateral stifles).• Cat G: 7-year-old male castrated domestic short hair, body weight 9.5 kg, multiple joints affected of the spine (lumbar vertebrae).Note: "Day 1" was defined as the first day the cat started wearing the monitor and that data were collected for the particular study period.For the experimental period, treatment was administered on "day 0." The mean minutes of activity per day were obtained for each patient in each treatment category (control, frunevetmab) (►Fig.1; ►Table 2).Mean activity minutes per day were significantly greater for frunevetmab (242.1)compared to control (209.9)(p = 0.007) (►Table 3).

Discussion
The present pilot study suggests that this group of osteoarthritic cats demonstrated increased activity levels following frunevetmab treatment, compared to self-control, when measured via a GPS-based accelerometry device (Tractive).The self-control study design was selected due to perceived wide variability in activity level between individual cats, as well as differences in home environment (e.g., studio apartment vs. multi-story home).Therefore, we surmised that the most representative means to evaluate a change in activity level with treatment would be to compare the individual to itself, rather than to another control population that may encompass a wider range of variables.Indeed, previous work by Lascelles and others noted the importance of utilizing each cat as its own self-control to minimize such variations  and thereby obtain more clinically meaningful outcome assessments. 18hile multiple schemes have been validated for subjective evaluation of acute or postoperative pain in cats, [23][24][25] and for evaluation of chronic pain in dogs, [26][27][28] currently there is not a validated method to assess chronic pain in cats.Although ongoing refinements have been made to the FMPI, owner responsiveness at this point is poor and often zero. 29Our goal with this study was an attempt to objectively assess activity level in osteoarthritic cats treated with frunevetmab via a means which encourages owner compliance and thereby improves clinical assessment of therapeutic impact.To be fair, this method assumes that "mobility" or "activity" represents "comfort," which may not necessarily correlate perfectly.Additionally, the data available on the user interface do not provide speed/velocity information, and so we could not elucidate details as to how robust or vigorous the activity was (e.g., running/jumping vs. walking)-just that the cat was moving versus stationary.However, considering current limitations in owner-based clinical metrology scales and even subjectivity in clinical examination, it seems reasonable to extrapolate activity data as at least in large part representative of comfort level.This precedent has been established in the veterinary literature for a number of years; Lascelles and others previously showed that accelerometer-based activity data correlate well with computer-detected video movement in cats, serving as an accurate surrogate for mobility and, by extension, effect of pain-relieving therapies. 18 limitation of this study is the use of an as of yet unvalidated activity monitor (Tractive).However, this device represents potential to easily assess feline activity level at home via a modality that may engage owners more consistently than a clinical metrology scale.As it were, the scope of this study was focused on a means of convenient, user-friendly activity monitoring at home that could be adaptable to client use.We did not seek to precisely replicate previous feline accelerometry data obtained in a formal research setting.With Tractive, owners can view their cat's activity clearly on a userfriendly interface, with simple, objective data output.Involving and empowering owners in such a way is instrumental in optimizing clinical treatment plans.Looking toward the future of feline pain management and assessment, we propose that refinement of objective in-home assessments will not only lead to more reliable data acquisition, but will also help solve the issue of poor owner compliance.
A major limitation of this study is the small sample size.Nonetheless, statistical analysis indicated that this was an adequately powered study for the number of data points per patient being used as its own self-control.Due to the short battery life of the activity monitor, and, occasionally, owner compliance with charging, data acquisition typically took multiple weeks to capture 7 full days' worth of activity.The timeline was further prolonged by the 30-day washout period in the inverse control treatment category.This limited the number of cats for which we could obtain complete data during the time period of the study.Certainly, future studies incorporating a larger population of cats are warranted to corroborate these results.
An important future direction from this pilot study with the Tractive is to compare data output to a more extensively studied device such as the Actical, as previously done in cats with the Plus Cycle.Indeed, further studies are warranted on the Actical itself to definitively validate its use in cats.The information gleaned from the Tractive in this study is but a starting point in exploring the means of assessing feline activity level, and, by proxy, pain management, in the home environment.
In conclusion, this population of osteoarthritic cats exhibited an increase in activity level following frunevetmab treatment compared to self-control.This information was obtained via piloted clinical use of the Tractive device, which, while unvalidated, is user-friendly and could encourage owner engagement in future studies.This introductory work corroborates findings in previous literature demonstrating clinical benefit of frunevetmab, 15,16 and suggests an alternative and potentially desirable method of assessing feline activity levels and pain management.

Fig. 1
Fig. 1 Mean activity minutes per day for control and for frunevetmab plotted over day in treatment period (error bars represent standard error of the mean).

Table 1
Days on which activity data were collected

Table 3
Descriptive statistics for paired t-test (p ¼ 0.007)

Table 2
Mean activity minutes per day for each cat in each treatment category