Prospective, randomized, placebo-controlled, multicenter, double-blind, parallel-group, registration field trial evaluating izenivetmab, a caninized anti-nerve growth factor monoclonal antibody, administered quarterly to dogs with naturally occurring osteoarthritis

Highlights

• •The objective of this study was to demonstrate the safety and efficacy of izenivetmab administered to dogs for managing OA pain.
• •This study followed Good Clinical Practice guidelines and is intended to support global registration of izenivetmab, a caninized anti-nerve growth factor monoclonal antibody, for management of osteoarthritis pain in dogs.
• •Treatment with izenivetmab every three months resulted in significantly higher frequency of treatment success on the canine brief pain inventory compared to a negative control at all time points tested (Days 7-270).

Abstract

Izenivetmab, a caninized monoclonal antibody that targets nerve growth factor, was evaluated for osteoarthritis-related pain in a prospective, randomized, placebo-controlled, double-blind, parallel-group field trial, conducted at multiple private practices, enrolling 574 individually owned dogs (median age, 9.0 years) with osteoarthritis (e.g., moderate to nearly incapacitating). Dogs were block randomized (1:1) to izenivetmab (n=289; 0.05-0.10mg kg^-1) or placebo (saline, n=285) administered subcutaneously every 3 months for a total duration of 9 months. Day 90 canine brief pain inventory (CBPI) treatment success (TS), the primary endpoint, required a pain interference score (PIS; 0-10) decrease of ≥2 and pain severity score (PSS; 0-10) decrease of ≥1. Izenivetmab showed significant (p<0.05) improvement over placebo for CBPI TS (days 7-270). Of evaluable cases, day 90 CBPI TS (izenivetmab, n=255; placebo, n=257) rates were 36.3% and 21.4%, respectively (p=0.0029); CBPI TS frequency 90 days after each dose was stable for both groups. Izenivetmab showed significant effect on all CBPI elements (PIS, PSS, quality of life) versus placebo. Several adverse events (e.g., polyuria, polydipsia, pain upon injection) occurred at a numerically higher frequency in the izenivetmab group versus placebo. Results demonstrated the safety and effectiveness of izenivetmab (0.05-0.10 mg kg^-1) administered subcutaneously to dogs for osteoarthritis pain every three months for 9 months.

Introduction

Canine osteoarthritis (OA) is a progressive degenerative joint disease characterized by whole-joint structural changes leading to chronic pain, loss of joint function, and impaired mobility (Anderson et al., 2020; Emanuel et al., 2024; Johnston et al., 1997). Management of OA in veterinary patients centers on alleviation of chronic pain with a fundamental objective of improving quality of life (Johnston et al., 2008; Meeson et al., 2019; Rialland et al., 2012; Wright et al., 2022). Anti-nerve growth factor (NGF) monoclonal antibodies (mAbs), bedinvetmab and frunevetmab, have been the most significant therapeutic advances in the management of OA pain in veterinary medicine in over a quarter of a century since non-steroidal anti-inflammatory drugs (NSAIDs) were first registered for this indication in dogs (Corral et al., 2021; Gruen, et al., 2021; Michels et al., 2023). Currently registered anti-NGF mAbs (bedinvetmab for dogs, frunevetmab for cats) require monthly subcutaneous (SC) injections administered by a veterinary professional. Izenivetmab is an anti-NGF mAb with similar target affinity and half-life as bedinvetmab. However, it binds a unique NGF epitope, resulting in an extended pharmacodynamic effect that enables a 3-month dosing interval and a lower dose (0.05 mg kg^-1).

The objective of this study was to demonstrate the safety and efficacy of izenivetmab administered to dogs for managing OA pain. This study was pivotal to the global registration and was therefore compliant with regulatory guidance from both the United States of America Food and Drug Administration (FDA) and the European Medicines Agency (EMA). The study design was similar to the bedinvetmab trial reported by Michels et al. (2023) and the placebo-controlled segment of the bedinvetmab trial reported by Corral et al. (2021) except that it assessed safety and efficacy compared to placebo over 9-months with study treatment administered every three months.

Materials and methods

Study design

The study was a prospective, parallel-group, placebo-controlled, double-blind study with a balanced (1:1) randomization conducted at 37 general practices in Europe. The Veterinary International Conference on Harmonization and Good Clinical Practice Guideline principles ( EMEA, 2000; VICH Topic GL9, 2000) were followed. Prior to start of the study (protocol ID C866C-XC-20-319), protocol review and approval were completed by the Sponsor Ethical Review Board, the respective country regulatory authorities (Hungary no. 5300/3250-2/2020; Ireland CT22006/004; The Netherlands no. BD/2020/821117/p; Portugal no. 86/ECVPT/2020) and, in Germany, the respective local competent authorities, as appropriate. The study design, endpoints, and analysis were formally agreed with the FDA’s Center for Veterinary Medicine (CVM) before the study began. Informed consent was provided in writing by all owners.

Sample size estimates (≥210 evaluable cases per group), aimed to achieve ≥80% power at α=0.05 (two-sided), were derived from power calculations based on variance and effect sizes observed in an unpublished placebo-controlled field study (Sponsor data on file): Proportions of canine brief pain inventory (CBPI) treatment successes were assumed at 0.33 for placebo and 0.49 for izenivetmab . Additional dogs were enrolled to ensure enough evaluable cases per group at the end of the 9-month study.

Study population

Primary care veterinary practices enrolled client-owned dogs from November 2020 – September 2021. Criteria for eligibility were intended to ensure enrollment of dogs without concurrent treatments or uncontrolled disease that could confound assessment of efficacy and safety.

Inclusion criteria

Dogs were at least one year of age with clinical evidence of OA confirmed by radiographs in at least one joint (thoracic or pelvic limb) at the screening visit. Dogs were expected to benefit from continuous treatment for clinical signs of OA over a 3-month period. In compliance with FDA CVM guidance, dogs were naïve to treatment with an anti-NGF mAb (this requirement was intended to prevent bias towards enrollment of dogs whose owners perceived their dog responded well on a previous anti-NGF mAb) and only one dog in a household enrolled. There were no body weight restrictions for enrolment. One or more veterinarian categorical assessment elements (VCA; general musculoskeletal condition, pain on palpation/ manipulation of joint(s), lameness/weight bearing) had a severity of ‘moderate’ or worse at screening and day 0 (severity: “clinically normal”, “mild”, “moderate”, “severe”, “nearly incapacitating”). The owner assessed Pain Interference Score (PIS) and Pain Severity Score (PSS) on the CBPI were both ≥2 (scale, 0-10) at screening and day 0 (Brown et al., 2013). The owner assessed CBPI ‘overall impression of quality of life’ (QoL) but assessments did not apply to eligibility. The CBPI User Guide instructions were adhered to by all users (Brown, 2017), including linguistic validation of CBPI translations from English (Wells et al., 2021). The examining veterinarians confirmed that dogs were otherwise healthy based on physical examination and satisfactory clinical pathology (blood, urine) and any known concomitant diseases were considered well controlled.

Exclusion criteria

Dogs enrolled in a study ≤ one month prior to day 0; intended for breeding (male or female) or pregnant or lactating; cranial cruciate ligament rupture <6 months ago (whether or not surgical joint stabilization was implemented); lameness associated with primary non-OA condition (immunologic, infectious, neoplastic, neurologic), including non-healed fracture; injury with neurologic deficits or intervertebral disc disease history (if the examining veterinarian expected this could confound efficacy assessments); administered prohibited treatment (Table S1). Certain medications were conditionally allowed if specified withdrawal times were obeyed and requirements for frequency of use and minimal use were followed (Table S1).

Randomization and blinding

Dogs were randomized 1:1 at enrollment to placebo (saline) or izenivetmab (Lenivia™) according to a randomized complete block design with one-way treatment structure replicated in multiple clinics. Randomization was done using SAS Release 9.4 (SAS Institute, Cary, NC). Within each clinic, dogs paired based on enrolment order were randomly allocated to saline or izenivetmab (i.e., block size was two). The randomization was provided in an electronic system accessible to the dispenser only. Day 0 was defined for each dog as the day the first dose was administered. Hospital study personnel, with the exception of the dispenser, and owners were blinded to treatment group allocation and block size, and remained blinded at study conclusion. Dispensers prepared the study treatment as allocated (saline or izenivetmab; identical appearance) and did not participate in data collection. Study treatment was administered by the veterinarian or designee.

Treatment administration

Izenivetmab ready-to-use preservative-free formulation was provided in single-use 1 mL vials (vial strengths, 0.5, 1.0, 1.5, 2.0, 3.0 mg mL^-1). Dosage was 0.05–0.1 mg kg^-1 using a dose chart. Dogs treated with placebo received an equivalent volume of saline. Route of administration was by SC injection.

Study schedule

Screening took place at a visit 2 to 21 days prior to day 0. Baseline day 0 data collection occurred on day 0 before randomization. There were nine scheduled hospital visits after day 0, including days 7, 14, 21, 30, 90, 104, 180, 194 and 270. Owners completed a daily diary of any abnormalities in their dog’s health throughout the study.

At all visits, veterinarians conducted physical examination and recorded concomitant medications (newly initiated treatments, or previously documented ones where the posology was adjusted since the last visit) or adverse events (AEs), if any, and owners (without reference to previous assessments) completed the CBPI. Blood for hematology and serum chemistry and urine for urinalysis and measurement of protein creatinine ratio were collected (days 90, 180, 270). Serum aliquots at each visit were prepared to quantify izenivetmab concentrations and to assess for anti-drug antibodies (ADAs).

Escape clause and use of rescue and prohibited therapies

Withdrawal of a dog from the study, by the owner or veterinarian, could occur at any time. ‘Rescue’ medication administered for an OA-related indication, such as lack of efficacy (LOE), was allowed at any time but classified as a ‘prohibited’ treatment for data analyses. A ‘prohibited’ treatment other than rescue medication was one that could confound efficacy assessments and was administered for a non-OA related indication (e.g., an NSAID for a bite injury). Administration of prohibited or conditionally allowed treatments was permitted during the study, if needed, for animal welfare.

Efficacy outcome measures

Dogs were excluded from the effectiveness data set if a protocol deviation affected data collection or integrity or if from practices with less than two evaluable dogs per treatment. Use of rescue medication or withdrawal for LOE resulted in a dog being treated as a failure in the analysis of CBPI TS starting on the day of withdrawal or rescue through the end of the study – affected dogs, including dogs treated with a prohibited treatment that was not a rescue medication, continued study procedures and were required to be withdrawn at the next scheduled treatment visit (i.e., day 90 or 180 visit) or at day 270, as appropriate. However, due to the number of dogs temporarily administered prohibited treatments for reasons unrelated to OA during this 9 month study, and the desire to avoid loss of long-term safety data, this requirement was removed for non-rescue prohibited treatments by amendment (29 June 2021). Instead, efficacy data collected within the required withdrawal time defined in Table S1 was excluded following use of any non-rescue prohibited treatment. For example, efficacy data collected within seven days after treatment with a daily NSAID for a non-OA related reason (e.g., dog bite injury) was excluded from the analyses. See Fig. 1.

Fig. 1.

CONSORT flow diagram describing all cases recruited for the 9-month study. n, number; OA, osteoarthritis. *Animals administered rescue treatment during the study or that were withdrawn because of lack of efficacy were considered treatment failures beginning on the day of rescue or withdrawal, respectively, and were included in the analysis for treatment success. °Animals enrolled but that were not suitable for enrolment, died by vehicular trauma, missing for several months. ^†Data excluded due to protocol deviations (i.e., eligibility criteria not met, incomplete dosing/misdosing on Day 0, visit outside of the permitted window, different owner completing the CBPI assessment), site did not meet enrolment criterion of two or more evaluable cases per treatment group, animal withdrawn for developing an unrelated medical or surgical condition, owner non-compliance/lost to follow-up, or administration of prohibited treatments explains the cases excluded or missing from the efficacy analysis. All animals were included in the safety data analysis.

Statistical analysis of data occurred for days 7 (-0/+3), 14 (± 3), 21 (± 3), 30 (± 4), 90 (± 7), 104 (± 6), 180 (± 7), 194 (± 6) and 270 (± 10). The primary efficacy endpoint was Day 90 CBPI TS. TS, as defined by Brown et al. (2013), required a decrease of ≥ 1 in PSS (0-10) and a decrease of ≥ 2 in PIS (0-10).

Secondary efficacy variables were CBPI TS at remaining time points, all CBPI elements (PSS, PIS, QoL), the proportion of dogs with QoL improvement and a QoL of ‘very good’ or ‘excellent’. Overall VCA improvement (across all 3 VCA components) was a secondary efficacy endpoint and required either improvement in ≥1 component with none worsening, or improvement in ≥2 components; otherwise, a dog was considered not improved if neither condition applied or if withdrawn for LOE (including dogs administered rescue treatment). A post hoc calculation of values was made for number needed to treat (NNT) based on CBPI success/failure rates (Cook and Sacket, 1995).

Safety outcome measures and analysis

Safety evaluations included all dogs. AEs were summarized using Veterinary Dictionary for Drug Related Affairs coding – frequencies of dogs reported with AEs were calculated by organ class and preferred term as recommended by the EMA Committee for Veterinary Medicinal Products (2023). Veterinarians blinded to study treatment recorded suspected causality for each AE (i.e., study treatment, concurrent disease, pre-existing condition, unknown). Shift tables were generated to summarize the number of dogs with a decreased shift (from within or above to below) or an increased shift (from within or below to above) in clinical pathology results compared to screening at each time point based on the respective laboratory reference range.

Pharmacokinetics and immunogenicity assessment

Serum izenivetmab concentrations were determined by a validated LC/MS/MS method. A custom prepared biotin-labeled mouse anti-izenivetmab monoclonal antibody was allowed to bind to izenivetmab in the serum sample and the complex immunoprecipitated with streptavidin magnetic beads. Following dissociation, izenivetmab was digested with a protease mixture, and a surrogate izenivetmab peptide quantified by HLPC with MS/MS detection. The quantitation range was 0.01 µg mL^-1–4.00 ug mL^-1. Accuracy, precision, dilutional linearity (10-fold), selectivity, matrix interference, carryover, short-term stability and long-term frozen stability were assessed during validation and met acceptance criteria. Incurred sample reanalysis performed on study samples met acceptance criteria. ADAs were detected using a validated tiered ligand binding assay (LBA) methods (screening, confirmatory and titer) and were used to classify the immunogenicity status of each dog. The ADA assay utilized a minimum required dilution of 10 (initial sample dilution), had sensitivity within FDA guidance (FDA, 2019) (screening assay 1.97 ng mL^-1, confirmatory assay 2.17 ng mL^-1), and was tolerant to izenivetmab (≥25 µg mL^-1) and target. Titers were performed using 2-fold serial dilutions and reported as the dilution where the assay signal fell below the cut point for the first time multiplied by the minimum required dilution. Two dogs in the placebo group were not evaluable for immunogenicity due to lack of samples or potential mis-dosing. ADAs from dogs with treatment-emergent immunogenicity were further evaluated using a validated competitive LBA method to determine if the ADAs neutralized the ability of izenivetmab to bind to NGF (neutralizing ADA, nAb). Izenivetmab concentrations across all dogs and all time points were integrated with ADA data, including nAb data, to assess clinical relevance of immunogenicity data. Efficacy and safety data were evaluated with these data to assess, by animal, if efficacy (CBPI TS) was impacted by treatment-emergent immunogenicity. Treatment-emergent immunogenicity was assessed for a relationship to AEs.

Non-compartmental pharmacokinetic (PK) parameters of half-life, AUC_0-t(last), AUC_0-∞ and percent AUC_0-∞ extrapolated were only calculated following the first dose of izenivetmab. AUCs were determined using linear trapezoidal summation. The slope of the elimination phase was estimated using at least 3 time points in the log-linear phase of the concentration-time curve.

The sampling schedule precluded estimating C_max and T_max or calculating PK parameters following the second or third dose.

Statistical analysis

Data analysis was conducted using SAS Release 9.4 (SAS Institute, NC, USA). Fixed and random effects in statistical models were pre-specified by the protocol and stayed in the models disregarding their significance and/or numeric values in the analyses. CBPI PSS and PIS were analyzed separately using a general linear mixed model for repeated measures. Day 0 pretreatment scores were utilized as covariates in the model. The model included fixed effects of treatment, day of study and interaction between treatment and day of study. Random model effects included site, interaction between site and treatment, animal within site and treatment, interaction between site, treatment and day of study, and error. TS and any binary data (i.e., yes/no) were analyzed separately for each day of study, using generalized linear mixed models with binomial distribution and logit link. The model included fixed effect of treatment and random effects of site, and interaction between site and treatment. Back-transformed least squares mean were used as estimates of the treatment proportions and corresponding 95% CIs were constructed. Treatment comparisons for TS and any binary data (i.e., yes/no) were evaluated using log-odds ratio. Odds ratios and corresponding 95% CIs were also constructed. Two-sided 5% level of significance was used for all comparisons. Because Day 90 CBPI TS was the single primary efficacy endpoint and other efficacy variables were secondary endpoints providing supportive evidence, no multiplicity adjustment method was applied.

Results

Demographic data

Screening of 689 dogs was completed; of these, eligibility criteria were not met for 115 (Fig. 1). Thus, 574 dogs were randomized (placebo, n=285; izenivetmab, n=289) and received study treatment (Table 1).

Table 1.

Demographics of enrolled dogs with osteoarthritis at day 0. Izenivetmab group: subcutaneous administration of izenivetmab once every three months; placebo group: equivalent volume of saline. Data are presented as n (%) or mean ± standard deviation (range). n, number of animals (all animals enrolled, including animals with protocol deviations excluded from analysis).

Demographic	Groups		Total (n = 574)
	Placebo (n = 285)	Izenivetmab (n = 289)
Breed distribution
Purebred*	161 (56.5)	169 (58.5)	330 (57.5)
Mixed breed	124 (43.5)	120 (41.5)	244 (42.5)
Sex distribution
Male	142 (49.8)	161 (55.7)	303 (52.8)
Female	143 (50.2)	128 (44.3)	271 (47.2)
Neutered/ovariohysterectomized	188 (66.0)	190 (65.7)	378 (65.9)
Age (years)	8.7 ± 3.4 (1.0-18.0)	8.8 ± 3.4 (1.0-19.0)	8.8 ± 3.4 (1.0-19.0)
Weight (kg)	24.5 ± 13.2 (1.0-57.1)	25.2 ± 12.7 (1.0-53.3)	24.8 ± 12.9 (1.0-57.1)
History of ruptured cranial cruciate ligament >6 months prior to enrolment†	19 (6.7)	17 (5.9)	36 (6.3)
Pre-treatment day 0 CBPI PIS (0-10)	5.49 ± 1.84 (2.00-9.33)	5.49 ± 1.82 (2.00-9.83)	5.49 ± 1.83 (2.00-9.83)
Pre-treatment day 0 CBPI PSS (0-10)	5.11 ± 1.61 (2.00-9.25)	4.90 ± 1.54 (2.00-9.00)	5.00 ± 1.58 (2.00-9.25)
VCA general musculoskeletal condition, moderate or worse at study start	233 (81.8)	244 (84.4)	477 (83.1)

CBPI, Canine Brief Pain Inventory; CBPI PIS, pain interference score, CBPI PSS, pain severity score; VCA, veterinary categorical assessment.

^⁎Labrador Retriever was the predominant breed (24.5%; n=81) followed by German Shepherd (11.8%; n=39) and Golden Retriever (7.3%; n=24). No other individual pure breed comprised > 3.6% of the total.

^†Veterinarian responded (yes/no) to ‘Does the dog have a history of a ruptured cranial cruciate ligament?’

Efficacy assessment

Of all enrolled dogs, 171 withdrew before the end of the 9-month study (placebo, n=92; izenivetmab, n=79) (Table 2). There were 106 failures for withdrawal for LOE (placebo, n=64; izenivetmab, n=42); of these, all but two placebo-treated and one izenivetmab-treated dog received rescue treatment. Forty-two dogs were withdrawn for a surgical or medical condition not related to study treatment, 12 were withdrawn for owner non-compliance, including two lost to follow-up. Data were excluded for protocol deviations involving study visits not within the permitted window, assessment completed by different owner than day 0, animal mis-dosed, prohibited treatment administered or eligibility criteria not met. Seventy-five dogs (placebo, n=31; izenivetmab, n=44) were administered non-rescue prohibited medication. As per the FDA CVM guidance, data for seven dogs at three clinics were excluded for sites not having a minimum of two evaluable cases per treatment group.

Table 2.

Overview of dogs with osteoarthritis administered either saline (placebo group) or izenivetmab (izenivetmab group) by quarterly (every three months) subcutaneous injection for nine months that were withdrawn from the study prior to the day 270 visit. Data were excluded from the efficacy analysis for cases with protocol deviations or when efficacy data was collected within the defined withdrawal/wash out timeframe for prohibited treatments^*. All animals were included in the safety analysis.

Study day	Group
Study day 0 visit	Placebo (285 dogs)	Izenivetmab (289 dogs)
	Cumulative total number of dogs withdrawn before or on day 90 visit = 52	Cumulative total number of dogs withdrawn before or on Day 90 visit = 42
	39 cases were withdrawn due to worsening of clinical signs of OA or lack of efficacy	25 cases were withdrawn due to worsening of clinical signs of OA or lack of efficacy
	6 cases were withdrawn due to developing an unrelated medical condition: 3 cases were euthanized due to a malignant mass with metastases (1 case), spleen neoplasia with potential lung metastasis (1 case) and presumptive malignant lymphoma (1 case); 3 cases died due to pneumopathy (1 case), bronchopneumonia with congestive heart failure (1 case) and unknown cause (1 case)	7 cases were withdrawn due to developing an unrelated medical condition: 4 cases were euthanized due to multiple organ failure (1 case), malignant salivary gland neoplasia with lung metastases (1 case), bone neoplasia with metastases (1 case), and hemangiosarcoma (1 case); 1 case died due to peritonitis after surgery, 1 case was withdrawn due to developing cystitis and 1 case was withdrawn due to a mammary complex adenoma (1 case)
	2 cases were withdrawn due to administration of prohibited treatment*	4 cases were withdrawn due to administration of prohibited treatment*
	1 case was withdrawn due to owner non-compliance/lost to follow-up	1 case was withdrawn due to owner non-compliance/lost to follow-up
		1 case was withdrawn due to owner no longer wanted to participate in the study (moved out of country)
	2 cases died due to vehicular trauma	3 cases died due to vehicular trauma
	2 cases were withdrawn due to eligibility criteria were not met	1 case was withdrawn due to eligibility criteria were not met
Study day 90 visit	Placebo (233 dogs)	Izenivetmab (247 dogs)
	Cumulative total number of dogs withdrawn before or on day 180 visit = 75	Cumulative total number of dogs withdrawn before or on day 180 visit = 63
	14 cases were withdrawn due to worsening of clinical signs of OA or lack of efficacy	10 cases were withdrawn due to worsening of clinical signs of OA or lack of efficacy
	3 cases were withdrawn due to developing an unrelated medical condition: 1 case was euthanized due to hepatic failure; 2 cases died due to unknown cause	6 cases were withdrawn due to developing an unrelated medical condition: 3 cases were euthanized due to poor general condition (1 case), diabetes and weight loss (1 case) and adenosquamous carcinoma of the larynx with pulmonary, bone, and regional lymph node metastases (1 case); 3 cases died due to respiratory failure (1 case) and unknown cause (2 cases)
	4 cases were withdrawn due to administration of prohibited treatment*	2 cases were withdrawn due to administration of prohibited treatment*
		2 cases were withdrawn due to owner non-compliance/lost to follow-up
	1 case was withdrawn due to dog went missing for several months	1 case died due to vehicular trauma
	1 case was withdrawn due to owner no longer wanted to participate in the study (moved out of country)
Study day 180 visit	Placebo (210 dogs)	Izenivetmab (226 dogs)
	Cumulative total number of dogs withdrawn before or on day 270 visit = 92	Cumulative total number of dogs withdrawn before or on day 270 visit = 79
	11 cases were withdrawn due to worsening of clinical signs of OA or lack of efficacy	7 cases were withdrawn due to worsening of clinical signs of OA or lack of efficacy
	3 cases were withdrawn due to developing an unrelated medical condition: 1 case was euthanized due to a lung tumor and 2 cases died due to respiratory failure (1 case) and unknown cause (1 case)	5 cases were withdrawn due to developing an unrelated medical condition: 4 cases were euthanized due to severe gastrointestinal symptoms (1 case), abdominal neoplasia (1 case), malignant lymphoma (1 case) and poor general condition (1 case); 1 case was withdrawn due to compressive empyema (extradural structure C4-C6)
	3 cases were withdrawn due to owner non-compliance/lost to follow-up	3 cases were withdrawn due to owner non-compliance/lost to follow-up
		1 case died due to vehicular trauma
Study day 270 visit	Placebo (193 dogs)	Izenivetmab (210 dogs)

^⁎Per protocol amendment, from 29 June 2021 onwards, it was no longer required to withdraw animals that received prohibited treatment when the indication for use was considered not osteoarthritis related.

Owner assessment (CBPI)

CBPI TS frequency was significantly greater in the izenivetmab versus placebo group at all visits (Table 3; Fig. 2). TS frequency in both groups was highest up to one month after each dose and decreased thereafter, remaining generally stable at the end of each dose interval (i.e., days 90, 180, 270). The NNT calculated from data at each visit decreased between the end of the first and third dose intervals (i.e., day 90 and day 270): 8.6 (day 7), 5.5 (day 14), 4.7 (day 21), 4.3 (day 30), 6.7 (day 90), 5.6 (day 104), 8.8 (day 180), 4.5 (194), 5.9 (270).

Table 3.

Primary efficacy variable: Canine Brief Pain Inventory (CBPI)-based treatment success of dogs with osteoarthritis assigned to treatment with quarterly (every three months) subcutaneous injections of izenivetmab (izenivetmab group, n=289) or saline (placebo group, n=285) by group and study day. Protocol deviations or administration of prohibited treatments result in exclusion of cases from the efficacy analysis so that the number of animals active in the study is affected and may fluctuate between time points, depending on the nature of the protocol deviations and the collection of efficacy data within or outside the defined withdrawal/wash out timeframe for prohibited treatments. Dogs withdrawn due to lack of efficacy or use of rescue treatment are counted as ‘treatment failures’ from the day of rescue or withdrawal, respectively, through all subsequent time points. Data are presented as proportion of dogs achieving treatment success (95% confidence interval (CI)).

Study Day	Group	Animals (n)	Treatment success Proportion of dogs* (95% CI)	Standard Error	Odds ratio (95% CI)	P
7	Placebo	269	0.115 (0.076-0.168)	0.022	2.328 (1.373-3.950)	0.0026
	Izenivetmab	268	0.231 (0.173-0.303)	0.032
14	Placebo	274	0.168 (0.118-0.235)	0.029	2.659 (1.677-4.217)	0.0001
	Izenivetmab	278	0.350 (0.273-0.436)	0.040
21	Placebo	275	0.223 (0.160-0.301)	0.035	2.672 (1.692-4.220)	0.0001
	Izenivetmab	277	0.434 (0.345-0.526)	0.045
30	Placebo	272	0.261 (0.193-0.342)	0.037	2.734 (1.664-4.492)	0.0002
	Izenivetmab	276	0.491 (0.403-0.580)	0.044
90	Placebo	257	0.214 (0.154-0.290)	0.033	2.084 (1.311-3.314)	0.0029
	Izenivetmab	255	0.363 (0.282-0.452)	0.042
104	Placebo	256	0.285 (0.215-0.368)	0.038	2.151 (1.295-3.575)	0.0043
	Izenivetmab	249	0.462 (0.375-0.552)	0.044
180	Placebo	255	0.236 (0.174-0.310)	0.033	1.739 (1.078-2.805)	0.0247
	Izenivetmab	244	0.349 (0.274-0.432)	0.039
194	Placebo	242	0.254 (0.184-0.340)	0.039	2.672 (1.536-4.650)	0.0010
	Izenivetmab	232	0.476 (0.382-0.572)	0.047
270	Placebo	249	0.205 (0.149-0.276)	0.031	2.322 (1.392-3.874)	0.0020
	Izenivetmab	233	0.375 (0.297-0.460)	0.040

^⁎Back-transformed least squares means proportion.

Fig. 2.

Proportion of dogs with osteoarthritis achieving CBPI TS after quarterly (every 3 months) subcutaneous injections of izenivetmab (izenivetmab group, blue squares; n=289) or saline (placebo group, blue circles; n=285) for 9 months. Data are presented as back-transformed least squares mean ± standard error of the mean. CBPI, canine brief pain inventory; TS, treatment success; n, number of dogs. *Significant difference compared with placebo (p < 0.05).

Mean PSS and PIS were numerically lower in the izenivetmab versus placebo group at all visits after enrolment; between-group differences were significant on days 7, 14, 21, 30, 104, 194 (p≤0.0005) but not days 90 (p=0.0915), 180 (p=0.2422) or 270 (p=0.0746) (Fig. 3a; Table S2) and days 7, 14, 21, 30, 104, 194 (p≤0.0224) but not days 90 (p=0.0577) or 180 (p=0.0832) (Fig. 3b; Table S3), respectively. There was a numerically higher proportion of izenivetmab-treated dogs with improvement in CBPI QoL at all visits versus placebo (Table S4), except day 90 results were nearly identical. The proportion of animals with ‘excellent’ or ‘very good’ QoL is shown (Fig. 4); differences versus placebo were significant at all visits (p ≤ 0.0135) except days 90 (p=0.0896) and 180 (p=0.1010).

Fig. 3.

Canine brief pain inventory (CBPI) (b) pain interference scores of dogs with osteoarthritis assigned to treatment with monthly subcutaneous injections of izenivetmab (izenivetmab group, squares; n=289) or saline (placebo group, circles; n = 285) for 9 months. Data are presented as least squares mean ± standard error of the mean. The scale of the y-axis was chosen to increase the visibility of error bars. n, number of dogs. *Significant difference compared to placebo (p < 0.05).

Fig. 4.

Proportion of dogs with osteoarthritis assigned to treatment with quarterly (every 3 months) subcutaneous injections of izenivetmab (izenivetmab group, squares; n=289) or saline (placebo group, circles; n=285) for 9 months and classified by the owners as having an ‘excellent’ or ‘very good’ overall impression of quality of life (QoL) on the CBPI assessment. Data are presented as back-transformed least squares mean ± standard error of the mean. CBPI, canine brief pain inventory; n, number of dogs. *Significant difference compared with placebo (p < 0.05).

Veterinarian categorical assessment

The frequency of dogs with overall improvement based on VCA was numerically higher in the izenivetmab compared to placebo group at all visits and the differences were significant (p≤0.0142) except at day 270 (p=0.0738) (Table S5).

Safety assessment

Adverse events and concomitant medications

There were no reports of injection site reactions (e.g., heat, swelling) in either group. Of all izenivetmab doses administered, 1.6 % (12/758) were associated with an adverse response immediately upon administration compared to 1.0 % (7/727) of all placebo doses (i.e., “pain on injection” = immediate pain and/or vocalization). At least one adverse event was reported in a similar proportion of dogs in each group; adverse events that occurred in >2% in the izenivetmab group and at a >1% higher frequency compared to placebo are listed in Table 4. There was a higher number of dogs in the izenivetmab versus placebo group, respectively, with a report of both polyuria and polydipsia (6 versus 1), polyuria only (4 versus 0) and polydipsia only (4 versus 1). Ataxia was reported only in the izenivetmab group. The owner of a 7 year old Samoyed that was last treated with izenivetmab on day 91, discovered the dog was pregnant when it was found to have whelped six puppies on day 131; three were dead upon discovery and the remainder were eventually re-homed without veterinarian exam. The dog was diagnosed with diabetes mellitus on day 170 and euthanized; necropsy revealed pyometra. The veterinarian was first informed of the pregnancy and whelping after the euthanasia. Deaths (including euthanasia) during the study, other than due to vehicular trauma, were attributed by the study veterinarian to concurrent co-morbidities except for one placebo-treated dog where causality was unknown. Other than adverse response upon injection, mentioned above, the only AE attributed by the study veterinarian to study treatment was transient lethargy in one izenivetmab-treated dog on the day of the first dose that did not occur after subsequent doses. All remaining AEs were associated with a wide variety of background comorbidities and appeared in both treatment groups.

Table 4.

Adverse events (AEs) in dogs with osteoarthritis with quarterly (every three months) subcutaneous injections of izenivetmab (izenivetmab group) or saline (placebo group) occurring at least once in >2% of all dogs in the izenivetmab group and at a >1% higher frequency compared to all dogs in the placebo group. AEs are listed by frequency of clinical signs in all dogs in the izenivetmab group using the Veterinary Dictionary for Drug Regulatory Activities (VeDDRA) preferred term classification on a per-animal basis^*. Data are presented as n (%), n, number of dogs (all animals enrolled and by < or ≥ median age of all enrolled dogs (9.0 years)).

			All animals enrolled by < or ≥ 9.0 years
	All animals enrolled		< 9.0 years		≥ 9.0 years
Preferred term clinical sign (VeDDRA)	Placebo (n = 285)	Izenivetmab (n = 289)	Placebo (n = 128)	Izenivetmab (n = 132)	Placebo (n = 157)	Izenivetmab (n = 157)
Any Adverse Event	154 (54.0)	158 (54.7)	63 (49.2)	64 (48.5)	91 (58.0)	94 (59.9)
Lethargy1	22 (7.7)	39 (13.5)	9 (7.0)	15 (11.4)	13 (8.3)	24 (15.3)
Emesis	20 (7.0)	27 (9.3)	6 (4.7)	11 (8.3)	14 (8.9)	16 (10.2)
Anorexia	17 (6.0)	27 (9.3)	6 (4.7)	9 (6.8)	11 (7.0)	18 (11.5)
Abdominal pain	11 (3.9)	15 (5.2)	3 (2.3)	5 (3.8)	8 (5.1)	10 (6.4)
Weight loss	6 (2.1)	12 (4.2)	2 (1.6)	4 (3.0)	4 (2.5)	8 (5.1)
Pain on injection	5 (1.8)	11 (3.8)	2 (1.6)	6 (4.5)	3 (1.9)	5 (3.2)
Arthritis4	5 (1.8)	10 (3.5)	2 (1.6)	6 (4.5)	3 (1.9)	4 (2.5)
Polydipsia2	2 (0.7)	10 (3.5)	1 (0.8)	4 (3.0)	1 (0.6)	6 (3.8)
Polyuria2	1 (0.4)	10 (3.5)	0 (0.0)	2 (1.5)	1 (0.6)	8 (5.1)
Lymphadenopathy	3 (1.1)	8 (2.8)	1 (0.8)	3 (2.3)	2 (1.3)	5 (3.2)
Ataxia3	0 (0.0)	8 (2.8)	0 (0.0)	2 (1.5)	0 (0.0)	6 (3.8)

^⁎Occurrence is calculated on a per-animal basis; no matter how many observations of the same AE were recorded for a dog, it only contributed a single observation to the occurrence calculation. The median age of all enrolled dogs was 9.0 years.

¹Episodes of ‘lethargy’ were reported throughout the study; all reports in izenivetmab-treated dogs were attributed by the veterinarian to concurrent disease (except for transient lethargy attributed to study treatment in one izenivetmab-treated dog on the day of the first dose that did not occur after subsequent doses); reports were frequently transient (e.g., resolving within 1-5 days) and included observations reported by the owner in a daily diary.

²Of these, both polyuria and polydipsia were reported in 6 dogs in the izenivetmab group and 1 dog in the placebo group.

³All VeDDRA preferred term reports of ataxia were attributed by the veterinarian to concurrent disease. Mean (median; range) age and body weight for the 8 cases was 11 years (11; 8-15) and 32kg (34; 20-42), respectively. Of these, ataxia did not resolve by study completion or withdrawal in 4 cases, including 1 dog reported with mild “slipping on tile floor” and moderate “ataxia” starting day 30 that withdrew on day 86 for lack of efficacy, 1 dog reported with moderate “ataxia” starting day 135 that was diagnosed with degenerative myelopathy and completed the study (dog was homozygous for high risk factor for degenerative myelopathy), 1 dog reported with “falling on chin” starting day 222, considered likely OA related by the veterinarian, and completed the study, and 1 dog reported by the owner to have mild “momentary imbalance” starting day 226 and completed the study. Resolution of ataxia was reported for the remaining 4 cases, including 1 dog reported “unstable on legs” on day 0 only and “ataxic episode” on days 3 and 24 only and completed the study, 1 dog reported with “unbalanced gait” starting on day 6 (resolved by day 39) was diagnosed with cervical intervertebral disc herniation and was withdrawn from study on day 95, 1 dog reported by the owner with a “staggering episode” on day 107 that resolved before the dog returned to its normal activity within a few minutes (the dog died unexpectedly at home 5 days later due to an unknown cause; no necropsy was done), and 1 dog reported with mild to moderate “ataxia” starting day 235 (resolved by day 239), was diagnosed with suspected vestibular syndrome and completed the study.

⁴Of the 10 dogs treated with izenivetmab that were reported with a clinical sign VeDDRA coded into the preferred term arthritis, only two were withdrawn because they required treatment (one was diagnosed with inflammation/injury of the right tarsal joint and the second one suffered a CCL rupture). Four of these cases were reported with “joint swelling” or “effusion” and another four with “joint stiffness”. Another animal was reported with “degenerative joint disease in both shoulders” as an incidental CT finding. The last animal with a clinical sign coded as arthritis was the CCL case.

Concomitant medications were administered to a similar proportion of dogs in both groups, (Table 5) and were generally well tolerated. There were no clear differences between groups except analgesics categorized under “anti-inflammatory and antirheumatic products, non-steroids” (e.g., carprofen, grapiprant) and “other analgesics and antipyretics” (e.g., bedinvetmab, metamizole sodium) were used at a higher frequency in the placebo group.

Table 5.

Frequency of occurrence of administration of concurrent medications to >2% of dogs both groups (three-monthly subcutaneous injections) during the study. Data are presented as n (%) by ‘drug class’ as categorized by the Anatomical Therapeutic Chemical System for veterinary medicinal products (ATCvet) on a per-animal basis^*. n, number of dogs.

ATCvet drug classification	Placebo (n = 285)	Izenivetmab (n = 289)
Any drug	174 (61.1)	164 (56.7)
Antiinflammatory and antirheumatic products, non-steroids (e.g., NSAIDs, glucosamine)*	78 (27.4)	53 (18.3)
Other analgesics and antipyretics (e.g., bedinvetmab, metamizole)	32 (11.2)	35 (12.1)
Ectoparasiticides for systemic use (e.g., afoxolaner, fluralaner, sarolaner)	24 (8.4)	32 (11.1)
Opioids (e.g., methadone, tramadol, butorphanol, buprenorphine)	24 (8.4)	32 (11.1)
Beta-lactam antibacterials, penicillins (e.g., amoxicillin (± clavulanic acid))	27 (9.5)	29 (10.0)
Anesthetics, general (e.g., ketamine, propofol, isoflurane, remifentanil)	22 (7.7)	29 (10.0)
Hypnotics and sedatives (e.g., (dex)medetomidine, midazolam)	21 (7.4)	28 (9.7)
Immunologicals for Canidae (vaccines)	37 (13.0)	26 (9.0)
Anthelmintics (e.g., praziquantel, fenbendazole)	25 (8.8)	23 (8.0)
I.V. solutions (e.g., lactated ringers)	9 (3.2)	19 (6.6)
Drugs for peptic ulcer and gastro-oesophageal reflux disease (GORD) (e.g., omeprazole, pantoprazole, sucralfate)	10 (3.5)	16 (5.5)
Antidiarrheal microorganisms (e.g., Enterococcus faecium) or Other antidiarrheals (e.g., kaolin)	12 (4.2)	14 (4.8)
Maropitant	11 (3.9)	13 (4.5)
Corticosteroids for systemic use, plain (e.g., dexamethasone, (methyl)prednisolone, triamcinolone)	9 (3.2)	12 (4.2)
Metronidazole	5 (1.8)	11 (3.8)
Other beta-lactam antibacterials (i.e., cefalexin, cefazolin)	6 (2.1)	10 (3.5)
Antiepileptics (e.g., gabapentin, phenobarbital)*	11 (3.9)	9 (3.1)
Antiseptics and disinfectants (e.g., chlorhexidine, povidone-iodine)	8 (2.8)	9 (3.1)
Ectoparasiticides for topical use, incl. insecticides (e.g., permethrin, flumethrin, fipronil)	7 (2.5)	8 (2.8)
Ace inhibitors, plain (e.g., benazepril)	9 (3.2)	7 (2.4)
Quinolone and quinoxaline antibacterials (e.g., enrofloxacin, marbofloxacin)	8 (2.8)	7 (2.4)
Antiinfectives (e.g., chloramphenicol, fusidic acid, ofloxacin)	5 (1.8)	7 (2.4)
Thyroid preparations (e.g., levothyroxine sodium)	13 (4.6)	6 (2.1)
Diet formulations w/ restricted ingredients other than restricted calorie (e.g., k/d Hills)	6 (2.1)	6 (2.1)
Other dermatological preparations (e.g., oclacitinib, lokivetmab, non-medicated shampoo)	6 (2.1)	6 (2.1)

^⁎All dogs shown met conditional use requirements (Table S1) for ‘Antiinflammatory and antirheumatic products, non-steroids' and ‘Antiepileptic’ or were administered the product as a rescue treatment or non-rescue prohibited treatment.

Clinical pathology

Mean and median values for hematology and serum chemistry variables in both groups fell within the respective reference ranges for all visits, with the following exceptions. Mean and median MCHC, at low end of reference range at screening, decreased to ≤4% below reference range at all subsequent visits in both groups. Mean and median MCV, near high end of reference range at screening, increased in both groups to ≤5% above the reference range at day 270. Compared to pre-treatment, clinical pathology results showed shift changes (i.e., increased or decreased versus pre-treatment) throughout the study at low frequency in both groups and were not clinically significant. Exceptions were a higher frequency of decreasing (MCHC) or increasing (MCV, PCV) shifts compared to other variables that occurred in both groups but were not considered clinically significant. Mean results of quantitative urine measures within reference range pre-treatment remained within reference range. Pre-treatment mean UPC was 0.3 and 0.5 in the placebo and izenivetmab groups, respectively; mean day 270 UPC was 0.5 in both groups.

Pharmacokinetic data and immunogenicity

Mean (± SD) izenivetmab serum concentrations at days 7, 14, 21 and 30 in dogs without immunogenicity were 0.301 (0.159), 0.256 (0.127), 0.193 (0.105) and 0.133 (0.0853) µg mL^-1, respectively. Mean concentrations 14 days after the second (day 104) and third doses (day 194) were 0.269 (0.146) and 0.266 (0.133) µg mL^-1, respectively, showing there was no accumulation between doses. Mean concentrations at the end of each dose interval (days 90, 180, 270) could not be determined because >50% of samples were below the assay limit of quantitation. Arithmetic mean izenivetmab serum half-life in dogs without immunogenicity was 13.5 (4.8) days. Mean AUC_0-∞ was 10.0 (6.3) µg day mL^-1.

Of animals evaluated, treatment-emergent immunogenicity occurred in 2.83% (8/283) and 3.46% (10/289) of dogs in the placebo and izenivetmab groups, respectively; those in the placebo-treated dogs are considered false positive results. In the 10 izenivetmab-treated dogs with treatment-emergent immunogenicity, the onset of immunogenicity occurred across all three dose intervals (3 dogs during first, 4 dogs during second, and 3 dogs during third dose interval). In 5 of these dogs, ADAs did not persist and were not detected by the end of the study. ADA titers were generally low (≤320). Of the 10 affected izenivetmab-treated dogs, nAbs were detected in eight and were associated with a corresponding decrease in serum izenivetmab concentrations in seven. Immunogenicity was not associated with any adverse effects other than decreased efficacy in two dogs, with unclear impact on efficacy in one, and with no impact in the remaining seven dogs.

Discussion

The efficacy and safety profile of izenivetmab, administered every three months, was established in client-owned dogs with OA based on CBPI owner assessments, veterinary orthopedic examinations (VCA), low number of AEs attributed by the study veterinarians to be related to izenivetmab, clinical pathology and low incidence of treatment-emergent immunogenicity during the 9-month study. The CBPI questionnaire, with use of CBPI TS as the primary endpoint, is a published validated method for assessment of efficacy of therapeutics for OA pain (Brown et al., 2013). The secondary efficacy endpoints provide additional insights although PSS and PIS are predominantly intended to feed into the CBPI TS composite endpoint. As animals are withdrawn from a study for lack of efficacy (LOE), there is a resultant loss of data at later time points. This could have contributed to the observed lack of significance at some time points for the continuous secondary variables. It underlines the importance of using a binary CBPI TS endpoint (yes/no) where withdrawals for LOE count as failures at all subsequent time points, which is particularly relevant for studies with this long in-life duration.

Reliance primarily on a validated method for efficacy assessment helps minimize sources of bias in a clinical OA study. Following the lead of human clinical research, a detailed behavior-based assessment of chronic pain by the owner, although subjective, is routinely relied upon when making clinical decisions and offers the advantages of an extended assessment of a dog in its typical environment by someone who is most knowledgeable about the animal’s behavior. Using a proxy (i.e., the pet owner) to determine the effect of an intervention based on clinical metrology instruments (CMIs) requires the use of validated questionnaires and administering them in an appropriate and consistent manner to decrease variability (Muller et al., 2016). For that reason, a program to develop a valid and reliable, owner-completed questionnaire for use as a chronic pain outcome assessment tool in clinical studies was initiated in the mid-2000s. This resulted in the publicly available CBPI designed to quantify the severity of chronic pain and its impact on routine activities in companion dogs. The CBPI is considered an appropriate, sensitive, and validated pain scoring tool for use in dogs affected by osteoarthritis, especially in the context of placebo-controlled interventional studies (Brown et al., 2007; Brown et al., 2008). The other main CMIs in this area are the Liverpool Osteoarthritis in Dogs (LOAD) (Hercock et al., 2009; Walton et al., 2013), the Helsinki Chronic Pain Index (Hielm-Bjorkman et al., 2009), and the Canine Orthopedic Index (COI) (Brown, 2014a, 2014b, 2014c).

Izenivetmab-treated dogs had a significantly higher frequency of CBPI TS compared to placebo at all assessed time points throughout the 9-month study. All of the secondary endpoints, both owner-assessed as well as veterinarian-based, showed improvement in the izenivetmab group compared to placebo at all time points and the between-group differences were significant at most, but not all, visits.

CBPI TS frequency three months after dosing with either izenivetmab (0.05 mg kg^-1) or placebo on day 0 was somewhat low in both treatment groups in the current study (day 90: izenivetmab, 36.3%; placebo, 21.4%), although the difference (14.9%) was similar to the 16.8% difference reported after 28 sequential days of daily dosing of an NSAID (48.1%) or placebo (31.3%) (Rausch-Derra et al., 2016) but lower than that observed in an unpublished exploratory field study where the difference between the izenivetmab (0.05 mg kg^-1; 51.0%) and placebo (29.0%) groups 84 days after a single dose was 22.0% (Sponsor data on file). The frequency of CBPI TS in the izenivetmab group at the end of each subsequent 3-month dose interval in the current trial remained stable (i.e., day 180, 34.9%; day 270, 37.5%). Despite interventions for OA pain demonstrating significant improvement, a high level of heterogeneity between studies has been reported in veterinary and human patients and has been attributed to differences in assessment tools, study design, placebo effect and patient population (Michels et al., 2023; Sethi et al., 2024). However, regardless of use of the same primary efficacy endpoint (Corral et al., 2021; Michels et al., 2023; Rausch-Derra et al., 2016) and identical study design (Corral et al., 2021; Michels et al., 2023), study to study variability has been reported in canine OA field studies with no clear differences in demographics (i.e., age, breed, weight, sex, neuter status), mean day 0 CBPI PSS or PIS or proportion of dogs categorized with a severity of “moderate” or worse on the VCA of general musculoskeletal condition (Michels et al., 2023). Likewise, Sponsor review of data for the current study and the aforementioned exploratory field study shows no such demographic differences between study populations. The CBPI is expected to minimize any potential effect of recall bias given all CBPI components ask for the owner’s assessment over no more than the past seven days. However, the authors hypothesize more frequent assessments, such as at least monthly, as occurred in other studies that used CBPI TS as the primary endpoint (Corral et al., 2021; Michels et al., 2023), compared to ≥2-month intervals (i.e., between days 30 and 90, days 104 and 180, days 194 and 270) in the current study might have been beneficial. To the best of the authors’ knowledge, there is no precedent of any other canine field trial in this therapeutic area with a total in-life duration of nine months. As the owners knew the animal would be receiving a next injection every three months, they may unintentionally have reported the dog doing worse at the end of each dosing interval based on their expectation, rather than on the actual effect of the treatment. In any case, the possibility of heterogeneity between studies was considered before study start, given the above mentioned precedence in canine OA studies (Corral et al., 2021; Michels et al., 2023; Rausch-Derra et al., 2016); thus, sample size calculations for the current study assumed a difference of 16% between groups.

To date, for the CBPI, the minimal clinically important difference (MCID) has not been defined. Looking at Corral et al., 2021 vs Michels et al., 2023, despite an identical study design and treatment groups (bedinvetmab vs placebo), the observed difference between treatment groups was 26.6% vs 10.8% in these studies. Nevertheless, Librela™ has proven to be a reliable and efficacious treatment for the alleviation of pain associated with OA in the majority of treated dogs. Recent studies using force plate gait analysis (Enomoto et al., 2026) or the Canine Orthopedic Index (COI) (Innes et al., 2025), have demonstrated how, based on the respective primary efficacy variables, monthly bedinvetmab numerically outperformed daily grapiprant or meloxicam, respectively. We therefore stipulate the CBPI, including its well-established definition of treatment success (i.e. reduction in PSS ≥1 and PIS ≥2) can be considered a standardized tool offering a test system to reliably conclude on a differentiation between placebo and actives proven to be clinically relevant in the field.

Succes/failure can be expressed as ‘number needed to treat’ (NNT) to compare across studies and get a sense of the relative efficacy of therapeutics. NNT calculated for the current study [6.7 at day 90 (13 weeks)] are within the range of that reported for NSAIDs for OA pain in humans [range, 4.7-8.4 (following treatment for 12 weeks] (Cadet et al., 2021) and similar to that reported for bedinvetmab for OA pain in dogs [4.3 at day 84 (12 weeks)] (Michels et al., 2023). Such data from similar studies involving any other licensed canine OA pain product are not available for comparison. For izenivetmab, the NNT is somewhat lower at the end of the study on Day 270 (5.9) vs Day 180 (8.8) and Day 90 (6.7), which demonstrates sustained efficacy over time.

Although osteoarthritis is frequently associated with aging, the majority of OA in dogs is secondary to developmental orthopedic diseases such as hip dysplasia, elbow dysplasia, osteochondritis dissecans, and patellar luxation, which can produce clinically significant OA in dogs as young as one year of age (Anderson et al., 2020; Enomoto et al., 2024). The inclusion of young dogs in this study, albeit a small proportion (1.7% of enrolled dogs were between 1 and 1.5 years old) reflects the clinical reality of canine OA and enhances the translatability of these results.

This study provided positive izenivetmab safety data. Considering the length of the study, the frequency and range of AEs were within that expected in an older population of dogs with OA. An exception is that polyuria (PU) and/or polydipsia (PD) were each reported in 3.5% of dogs in the izenivetmab group compared to less than 1% of the placebo group; a detailed review of affected cases revealed no apparent underlying cause, other than the administration of izenivetmab, including no relationship to the emergence of renal disease or progression of pre-existing renal disease. Although reports of ataxia occurred only in the izenivetmab group, the clinical signs and duration varied widely across the eight cases, as described in Table 4 footnote, suggesting there is not a common mechanism related to treatment. Review of cases with the remaining AEs in Table 4, including lethargy (the most frequently reported AE in the izenivetmab group in Table 4), reveals no apparent safety concerns associated with izenivetmab treatment compared to placebo. No findings consistent with unexpected joint pathology were identified. Over half of dogs in each treatment group received at least one concurrent medication, largely for co-morbidities. Clinical pathology summary statistics showed no clinically relevant changes. Post marketing pharmacovigilance surveillance in a wider-ranging population of dogs will provide further izenivetmab safety and efficacy information.

The pharmacokinetic data show the half-life of izenivetmab is similar to that of bedinvetmab (Krautmann et al., 2021; Michels et al., 2023). Despite the similar half-life and similar binding affinity to NGF (Sponsor data on file), this study demonstrated a 90-day duration of efficacy. This extended pharmacodynamic effect is attributed to izenivetmab’s unique binding epitope on NGF. In this study, C_max and T_max were not reported due to sparse sampling. However, in a previous laboratory study in healthy beagle dogs, maximal serum drug concentrations occurred at approximately 3 days following SC administration (Sponsor data on file). Therefore, the concentrations observed at 7 days in this study were likely near the true C_max.

Incidence of treatment-emergent immunogenicity in izenivetmab-treated dogs was low and only slightly higher than that in the placebo group (3.46% vs. 2.83%, respectively). All ADA assay methods developed and validated per FDA guidance (FDA, 2019) have a target false positive rate of 1% in the confirmatory ADA assay. The low incidence in the placebo group is consistent with this false positive rate, especially given that multiple samples over nine months were analyzed per dog. The low incidence of immunogenicity observed with izenivetmab is presumably due to caninization of the mAb to minimize foreign amino acid sequences (analogous to human mAbs) as well as effective stabilization of the mAb by the formulation (Beck et al., 2010; Gaughan, 2015). The low immunogenicity incidence was consistent with that observed for bedinvetmab and lokivetmab (Michels et al., 2016a and 2016b; Corral et al., 2021; Krautmann et al., 2021; Michels et al., 2023). There was no apparent trend observed as to when ADAs to izenivetmab were first detected. However, in half of the izenivetmab-treated dogs with treatment-emergent immunogenicity, ADAs were only transiently detected and were not detected at the end of the study. This may reflect development of immune tolerance in some treated animals. Detection of ADA in a sample was often associated with a corresponding reduction in izenivetmab concentration. While this may be expected to result in some reduction of efficacy, correlation to CBPI TS was difficult to establish. This is likely due to either the weak and transient nature of the ADA response or due to the variability of the CBPI TS measurement. Immunogenicity to izenivetmab was not associated with any other adverse effects.

Conclusion

Izenivetmab is the first approved anti-NGF monoclonal antibody with a demonstrated 90-day duration of efficacy for managing OA pain in dogs. The results of pet owner and veterinary assessments, as well as clinical pathology assessments, reported in this study demonstrate the safety and efficacy of izenivetmab administered once every three months to dogs suffering from pain associated with OA with efficacy maintained for up to 9 months following repeat quarterly administration. Compared with bedinvetmab, this molecule is designed to reduce the practical burden associated with achieving treatment adherence and patient comfort. Treatment choice should always be based on a holistic risk-benefit assessment for an individual patient.

Ethics declaration

This study was conducted in accordance with the following guidelines for animal welfare and/or reporting: All applicable national regulations were adhered to. This study was approved by the Zoetis Ethical Review Board.

Statement for studies in animals

Prior to start of the study (protocol ID C866C-XC-20-319), protocol review and approval were completed by the Sponsor Ethical Review Board, the respective country regulatory authorities (Hungary no. 5300/3250-2/2020; Ireland CT22006/004; The Netherlands no. BD/2020/821117/p; Portugal no. 86/ECVPT/2020) and, in Germany, the respective local competent authorities, as appropriate. The study design, endpoints, and analysis were formally agreed with the FDA’s Center for Veterinary Medicine (CVM) before the study began. Informed consent was provided in writing by all owners.

Ethical approval statement

Prior to start of the study, protocol review and approval were completed by the Sponsor Ethical Review Board and the respective country regulatory authorities.

Generative AI statement

The authors declare that no Gen AI was used in the creation of this manuscript.

Funding statement

All research presented herein was fully funded by Zoetis, the Sponsor.

CRediT authorship contribution statement

Hilde Moyaert: Writing – review & editing, Visualization, Validation, Supervision, Project administration, Data curation, Conceptualization. Gina M. Michels: Writing – review & editing, Writing – original draft, Visualization, Conceptualization. Helena Berlamont: Writing – review & editing, Visualization, Validation, Data curation. Karen Akkermans: Writing – review & editing, Validation, Data curation. Bernard D. Hummel: Writing – review & editing, Validation, Formal analysis, Data curation, Conceptualization. Jezaniah Kira S. Tena: Writing – review & editing, Visualization, Formal analysis, Conceptualization. Maria J Corral: Writing – review & editing, Validation, Project administration, Data curation, Conceptualization. Ignacio Gisbert: Writing – review & editing, Conceptualization. Dawn M. Cleaver: Writing – review & editing, Conceptualization. Michael R Stegemann: Writing – review & editing, Supervision, Project administration, Conceptualization.

Declaration of competing interest

All authors were full-time employees of Zoetis at the time of study conduct.

Data availability statement

All raw data is proprietary to the Sponsor and available for review by regulatory agencies. The authors declare that all results presented herein accurately reflect the raw data.