A novel high-protein, low-carbohydrate diet enriched with eggshell membrane and omega-3 fatty acids is associated with improvements in pain, mobility, and biomarkers in dogs with naturally occurring osteoarthritis

Canine osteoarthritis (OA) is a prevalent, chronic disorder causing persistent pain and mobility impairment across a wide age range. Nutritional strategies targeting inflammatory and cartilage degradation pathways are critical components of multimodal OA management. This exploratory study evaluated a novel, complete high-protein diet enriched with eggshell membrane (ESM) and omega-3 fatty acids in laboratory Beagles with naturally occurring OA. Twelve Beagle dogs (mean age 11.0 years) with radiographic and clinical evidence of OA were transitioned to the test diet (37.4% crude protein DM, 1200 mg/kg ESM, 1% EPA DM, 0.66% DHA DM) for a 58-day intervention period. Primary outcomes included technician-assessed pain-related scores and functional impairment scores evaluated using a behavioral assessment tool developed for research colony studies. Secondary outcomes included circulating biomarkers of systemic inflammation (C-reactive protein, CRP) and cartilage degradation (C-terminal telopeptide of type II collagen, CTX-II). Body weight and body composition were monitored using quantitative magnetic resonance (QMR). By Day 58, significant progressive improvements were observed in mean cumulative function scores (from 9.4 at baseline to 0.6; p<0.001) and pain-related scores (from 18.7 to 15.6; p<0.001). Concurrent significant reductions were demonstrated in plasma CRP (approx. 59% reduction; p<0.001) and CTX-II concentrations (approx. 28% reduction; p<0.001). Circulating pro-inflammatory cytokines (IL-2, IL-6, TNF-α) showed no consistent changes. Body weight, body condition scores, and QMR lean/fat mass metrics remained remarkably stable throughout the study. Feeding a complete high-protein diet enriched with ESM and omega-3 fatty acids was associated with improvements in technician-assessed mobility indicators and reductions in biomarkers relevant to systemic inflammation and cartilage turnover. These changes occurred without meaningful changes in body mass or body composition. While these preliminary findings must be interpreted cautiously due to the open-label, single-arm design, they support further evaluation of this multi-component nutritional strategy as a potential adjunctive strategy in canine OA management.

Osteoarthritis (OA) is one of the most common chronic musculoskeletal disorders in dogs and a major cause of persistent pain, reduced mobility, and impaired quality of life (Cachon et al., 2023; Lascelles et al., 2015). Primary-care screening data suggest that OA may be identified in approximately 37% of dogs presenting to general veterinary practices (Wright et al., 2019), and recent evidence indicates that OA is not limited to geriatric populations, with radiographic evidence of OA and associated pain also reported in young adult dogs (Enomoto et al., 2024). Collectively, these findings highlight OA as a substantial welfare burden across the canine lifespan and reinforce the need for safe, sustainable approaches for long-term management.

OA is now widely conceptualized as a whole-joint disease involving cartilage, synovium, subchondral bone, periarticular tissues, with inflammatory and neuroinflammatory mechanisms contributing to chronic pain (Loeser, 2010; Sokolove and Lepus, 2013). In dogs, OA may arise from developmental joint disease, trauma, or instability, resulting in progressive structural and functional deterioration (Cachon et al., 2023). Ongoing synovitis, inflammatory mediator release, altered joint biomechanics, cartilage matrix degradation, and compensatory tissue remodeling contribute to variability in clinical presentation and treatment response (Cachon et al., 2023; Mobasheri and Batt, 2016; Sokolove and Lepus, 2013).

Given its multifactorial nature, current guidance emphasizes multimodal OA management, integrating pharmacologic analgesia and anti-inflammatory therapy with non-pharmacologic strategies such as weight optimization, controlled activity or rehabilitation, and adjunctive nutrition-based approaches (Cachon et al., 2023). Nutritional interventions have gained increasing attention as adjunctive therapies in OA management because they may influence inflammatory pathways, joint metabolism, and body composition. Among diet-linked bioactives, long-chain omega-3 polyunsaturated fatty acids (PUFAs), particularly eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), have the strongest evidence base in canine OA (Barbeau-Gregoire et al., 2022). Controlled studies have demonstrated improvements in owner-reported mobility, weight bearing, and clinician-assessed lameness in dogs fed omega-3-enriched diets compared with controls (Roush et al., 2010a, Roush et al., 2010b). Omega-3 fatty acids may modulate inflammatory pathways by altering membrane lipid composition and reducing signaling through pro-inflammatory mediators, providing a plausible biological rationale for OA symptom improvement (Barrouin-Melo et al., 2016; Roush et al., 2010b).

Beyond omega-3 fatty acids, there is increasing interest in ingredients that provide structural support and bioactive components relevant to connective tissue health. Eggshell membrane (ESM) is a naturally derived extracellular-matrix-like material containing structural proteins, collagens, glycosaminoglycans, and other bioactive constituents (Pillai et al., 2023; Shakir et al., 2023; Shi et al., 2021). Because OA is characterized by dysregulated extracellular matrix turnover and chronic low-grade inflammation, ESM has been proposed as a joint-support ingredient with potential relevance to mobility and comfort, consistent with reported anti-inflammatory and regenerative effects in preclinical and translational studies (Casado-Santos et al., 2024). Human studies have reported improvements in joint pain and function following ESM supplementation, with evidence of reductions in biomarkers of cartilage degradation, including CTX-II, in controlled exercise-induced settings (Cánovas et al., 2022; Kiers and Bult, 2021; Ruff et al., 2018). Although direct canine evidence remains limited, studies in dogs have reported improvements in clinical pain, mobility, functional parameters, and/or CTX-II following ESM supplementation (Aguirre et al., 2018; Ruff et al., 2016).

The diet evaluated in the present study was formulated to support joint health through multiple mechanisms, combining ESM with EPA/DHA omega-3 fatty acids and a high-protein, lower-carbohydrate macronutrient profile. Specifically, dogs were fed a complete diet containing 1200 mg/kg ESM as fed (~15 mg/kg body weight) and enriched in EPA (1% DM) and DHA (0.66% DM) over a 58-day feeding period. Additionally, the test diet had a moderate calorie content (337 kcal/100g as fed, NRC 2006 CF), allowing adequate management of body weight. Pain-related scores and functional impairment were assessed using a technician-administered behavioral scoring tool developed for research colony-based assessment of OA-related signs. This instrument was developed and evaluated using research colony-based studies in which the scoring approach demonstrated robust ability to discriminate between osteoarthritic and non-osteoarthritic dogs as well as sensitivity to pharmacologic analgesia using meloxicam (Araujo, 2017), supporting its use in controlled experimental settings. Circulating biomarkers were evaluated to reflect systemic inflammation and cartilage degradation, including C-reactive protein (CRP) and C-terminal telopeptide of type II collagen (CTX-II), which have been widely used in OA research (Sokolove and Lepus, 2013; Ruff et al., 2018). The primary objective of this exploratory study was therefore to evaluate whether feeding of the dietary intervention was associated with changes in pain-related scores, function scores, and biomarker outcomes in laboratory Beagles with naturally occurring OA.

Test diet (omega-3 and eggshell membrane enriched)

The test diet (Virbac Nutrition) was a complete and balanced dry kibble formulation designed to support joint health and overall body composition. The diet was enriched with omega-3 fatty acids, providing 1% EPA and 0.66% DHA (DM), and contained ESM at 1200 mg/kg as fed, corresponding to an approximate target intake of 15 mg/kg body weight per day based on expected daily food consumption. Actual intake, based on measured food consumption, is reported in the Results. The formulation was characterized by a high-protein, moderate-fat macronutrient profile, with relatively low carbohydrate content. The nutritional composition of the diet is outlined in Table 1.

Following baseline assessments and prior to initiation of the treatment phase, dogs were transitioned from a commercial basal kibble (ProPlan ®, Nestlé Purina PetCare; nutritional characteristics presented in Table 1) to the test diet over 5 days by progressively increasing the proportion of the test diet (25%, 50%, 50%, 75%, 100%). Dogs were then exclusively fed the test diet for the duration of the 58-day intervention period.

Test diet feed amounts were determined by calculating each dog’s baseline daily caloric intake from the basal diet, which had previously been established to maintain body condition. The metabolizable energy content of the test diet (337 kcal/100g as fed, NRC 2006 CF) was used to convert these values to the corresponding gram amount required to provide an equivalent caloric intake. Feed rations were offered once daily, following completion of scheduled assessments, for a duration of up to 90 minutes.

Food intake was monitored during both the diet transition and treatment periods by weighing the amount offered and any remaining feed residue. In cases where decreased food consumption was observed, rations were supplemented with a small portion of canned food (Prescription Diet a/d, Hill’s ®) if the addition of warm water did not first increase consumption.

Subjects and husbandry

Twelve Beagle dogs maintained within the Transpharmation Canada Ltd. colony for a minimum of three months prior to study initiation were enrolled. All dogs had historical radiographic evidence of OA, defined by the presence of osteophytes or enthesophytes, subchondral sclerosis or joint effusion in at least two joints. Affected joints in the study population included hips, stifles, and/or shoulders. Evidence of OA-related pain was supported by baseline behavioral assessments, with all dogs exhibiting mean cumulative pain-related scores >17 (maximum score: 26). Dogs were otherwise in good general health and were not receiving concomitant medications expected to confound OA-related outcome assessments. The enrolled dogs included 3 neutered males and 9 spayed females ranging in age from 7.7 to 14.8 years of age (mean age 11.0 years). Baseline demographics are summarized in Table 2.

Dogs were housed in groups of up to four in pens measuring 1.52 x 4.9 m, which included raised resting platforms and a rotation of toys for environmental enrichment. A combination of natural and LED lighting was provided, with artificial lighting maintained on a 12 h:12 h light-dark cycle; however, total daily light exposure varied with fluctuations in natural light. Ambient temperature and humidity were monitored in accordance with institutional standards. Dogs had ad libitum access to water throughout the study.

Study design

This was a prospective, open-label, single-arm, baseline-controlled exploratory feeding study.

During the baseline phase, veterinary examinations, body weight measurements, body condition scoring, muscle condition scoring, quantitative magnetic resonance (QMR) body composition analysis, and blood collection for hematology, biochemistry, and biomarker analysis were conducted to establish health status. Pain-related scores and function scores were assessed using the technician-administered behavioral assessment tool over a 4-day period.

Following test diet wash-in, the 58-day treatment phase was initiated (Day 0 to Day 58). Pain-related scoring and function scoring assessments were conducted over 4-day periods on Days 13-16, 27-30, and 55-58. Veterinary examinations were performed following the first assessment of each period. Body weight measurements, body condition scoring, and muscle condition scoring were performed on Days 0, 17, 31, and 58. Blood samples for hematology, biochemistry, and biomarker analysis were collected on Days 31 and 58, and QMR analysis was repeated on Day 58.

Throughout the study, general health observations and fecal scoring at the group-housing level were performed twice daily, and food intake was measured daily during the diet transition and treatment phases. See Table 3 for the study schedule.

The sample size (n=12) was determined based on the availability of animals meeting predefined inclusion criteria and was considered appropriate for this exploratory pilot study. No formal a priori power calculation was performed. The primary endpoints were pain-related scores and function scores from the behavioral assessment tool. Secondary endpoints included circulating biomarkers, while remaining parameters, including body composition and health monitoring, were considered exploratory.

Technician-assessed pain-related scoring and function scoring

Dogs were approached in their home pen and led on a leash to the observation area by both a handling technician and a trained observing technician, allowing for assessment of behaviors occurring immediately upon interaction. All assessments were conducted by the same observing technician at approximately the same time of day throughout the study to minimize interobserver variability and diurnal variation, respectively; however, formal interobserver variability was not assessed in this study. Because this was an open-label study, the observing technician was not blinded to study phase.

Each dog was evaluated for both functional ability and pain-related responses associated with performing specific behaviors. Behaviors assessed included posture at the start of the session, ease of rising and lying down (if observed), general alertness, initiation of walking, general activity, and willingness to perform a range of movements including walking, trotting, galloping, jumping over a low obstacle, climbing and descending stairs, rearing, and jumping down from an elevated perch.

Function was scored on a scale of 0-4, where 0 = with great ease or very willing, 1 = easily or willingly, 2 = neither or reluctantly, 3 = with difficulty or very reluctantly, and 4 = with great difficulty or does not participate. Pain associated with individual behaviors was scored on a scale of 0–2 where 0 = not due to pain, 1 = unable to determine, and 2 = clearly attributable to pain. Overall assessments of willingness to engage in the session and overall pain were recorded at the conclusion of each observation period. Maximum scores were 48 for function and 26 for pain, with higher scores indicating poorer function and greater pain. For statistical analysis, the mean cumulative pain and function scores across the four test sessions at each time-point was used to evaluate time-point differences.

During assessments, small portions of canned food (ProPlan ®, Nestlé Purina PetCare) were provided as positive reinforcement. Upon completion, dogs were returned to their home pens.

The complete behavioral assessment scoring form, including all items and scoring criteria, is provided as Supplementary File 1.

Biomarker analysis

At each sampling time point, blood was collected into K 2 EDTA tubes and centrifuged under cooled conditions to obtain plasma. Plasma was aliquoted, treated with protein inhibitor cocktail, and stored at approximately -80 °C until shipment on dry ice to the bioanalytical laboratory.

Plasma concentrations of CRP, CTX-II, and pro-inflammatory cytokines (IL-2, IL-6, IL-8, and TNF-α) were quantified using commercially available immunoassays according to the manufacturers’ instructions. Cytokines were measured using a Canine Proinflammatory Panel 3 assay (Meso Scale Diagnostics, LLC; Cat. No. K15035C), CTX-II using a canine CTX-II ELISA kit (Kendall Scientific; Cat. No. CC0853), and CRP using a Canine CRP QuickStep ELISA kit (AssayGenie; Cat. No. AEFI03273). Samples were analyzed in duplicate, and reported concentrations represent the mean of technical replicates.

Body weight and body composition

Body weights were measured using a certified, calibrated scale. Body condition score was assessed using a 9-point WSAVA scale, and muscle condition score was assessed using a 4-point WSAVA scale (WSAVA Nutritional Assessment Guidelines Task Force Members, 2011).

QMR assessments were performed using an EchoMRI-D analyzer (Echo Medical Systems) to measure lean mass, fat mass, free water, and total body water. For each scan, conscious animals were placed in a Plexiglas kennel within the magnet bore. Previously established, system-specific correction factors were applied to raw output values (Zanghi et al., 2013).

Hematology and biochemistry analysis

At each sampling time point, blood was collected into K 2 EDTA and serum separator tubes for hematology and biochemistry analysis, respectively. Serum was isolated by centrifugation, and both sample types were stored under refrigerated conditions until shipment on ice to a commercial diagnostic laboratory (Antech Diagnostics) for standard veterinary analysis.

Veterinary examinations, health observations, and fecal scoring

Veterinary examinations were performed immediately following the first behavioral assessment session at each evaluation period by personnel who were not blinded to study phase. Examinations included an evaluation of body condition, eyes, ears, mouth, nose, throat, heart and heart rate, lungs and respiration rate, abdomen, coat and skin, musculoskeletal system, nervous system, urogenital system, mentation, mucous membrane color, capillary refill time, and body temperature. Qualitative assessment of musculoskeletal findings during the examinations included evaluation of lameness, pain on manipulation, reduced range of motion, stiffness, and palpable abnormalities consistent with OA. Formal goniometry and limb girth measurements were not performed.

General health observations were conducted twice daily throughout the study to monitor for any abnormalities in health or behavior. At each observation time point, fecal consistency was assessed at the group-housing pen level using the Purina Fecal Scoring Chart (Nestlé Purina PetCare).

Statistical analysis

The primary endpoints were pain-related scores and function scores from the behavioral assessment. Secondary endpoints included biomarkers, and remaining parameters including body composition and health monitoring were considered exploratory.

Statistical analysis of pain-related scoring and function scoring data was performed with SAS 9.4 (Cary, NC) using a significance threshold of p<0.05 for all endpoints. A linear mixed model was used to evaluate changes in scores across time points. The models included study day as a fixed effect and a random intercept for animal to account for repeated measurements on the same dog. Residual errors were modeled using a variance components structure, which allowed the residual variance to differ by study day. Estimated marginal means (adjusted means) were estimated for each study day, and comparisons of study day to baseline were adjusted for multiple testing using the Tukey method.

Biomarker data were analyzed using GraphPad Prism 10.6 (San Diego, CA). Normality of data was assessed by the Shapiro-Wilk test. Data were analyzed by ordinary one-way ANOVA with Fisher’s LSD post hoc test (for parametric data) or Kruskal-Wallis followed by uncorrected Dunn’s post hoc test (for non-parametric data). Statistical significance was set to p<0.05. Outliers were removed using ROUT (Q = 5%). Removed data points included n=2 from the CRP analysis on Day 31, n=3 from the IL-6 analysis on Day 31, and n=1 from the IL-8 analysis at all time points (baseline, Day 31, and Day 58). No formal sensitivity analyses including the ROUT-identified outliers were performed. Because multiple biomarker endpoints were evaluated and no adjustment for multiple comparisons was applied, these findings should be interpreted cautiously.

Exploratory variables were summarized descriptively and evaluated qualitatively.

One dog was withdrawn from the study on Day 29 following a cruciate ligament injury that precluded continued participation in performance-based behavioral assessments beginning on Day 27. This study did not use a formal intention-to-treat approach. Available data from completed assessments prior to withdrawal were included in analyses, whereas data from Days 27–58 were not available for this subject. No sensitivity analyses were performed using replacement values for missing data. All other animals were included in the full statistical analysis. A flow diagram of animal disposition and analysis populations is presented in Figure 1.

Diet acceptance and intake of diet components

Overall diet acceptance was considered acceptable. Most dogs consumed their full ration for the majority of the study period, while others exhibited variable intake including short transient reductions, a temporary reduction lasting approximately two weeks in one animal, or persistent partial consumption in three animals. Mean ration consumption across the cohort was 93.1 ± 16.6%, indicating generally adequate acceptance of the test diet.

Based on measured food intake, estimated mean daily intake across the study cohort was 19.8 ± 4.1 mg/kg/day for ESM, 146.8 ± 30.6 mg/kg/day for EPA, and 99.0 ± 20.6 mg/kg/day for DHA. Together these observations reflect variability in individual food consumption but indicates that targeted exposure to key dietary components was achieved across the cohort.

Clinical outcomes: pain-related scores and functional impairment scores

Relative to baseline, mean cumulative pain-related scores decreased by 0.2 (95%CI 0.3 increase to 0.7 decrease; p=0.71) on Days 13-16, by 2.0 (95%CI 0.9 to 3.1; p=0.001) on Days 27-30, and by 3.2 (95%CI 2.0 to 4.5; p<0.001) on Days 55-58 (Figure 2; individual animal trajectories are shown in Figure 3).

Mean cumulative function scores also decreased from baseline by 4.4 (95%CI 1.3 to 7.5; p=0.004) on Days 13-16, by 7.9 (95%CI 5.3 to 10.4; p<0.001) on Days 27-30, and by 8.9 (95%CI 6.3 to 11.4; p<0.001) on Days 55-58 (Figure 4; individual animal trajectories are shown in Figure 5).

These data suggest progressively improved technician-assessed mobility-related performance over the 58-day intervention period, with pain-related scores demonstrating a more gradual pattern of change than function scores.

Biomarker analysis (CRP, CTX-II, and inflammatory markers)

Relative to baseline, plasma concentrations of CRP and CTX-II were significantly decreased on Day 31 (p<0.01 and p<0.05, respectively) and on Day 58 (p<0.001; Figure 6). Relative to baseline, mean plasma CRP concentrations decreased from 87,245.4 ± 39,756.4 ng/mL on Day -8 to 29,961.1 ± 12,154.9 ng/mL on Day 31 (p<0.01) and 35,578.6 ± 27,774.9 ng/mL on Day 58 (p<0.001), corresponding to approximate reductions of 66% and 59%, respectively. Mean CTX-II concentrations also decreased from 1,323.5 ± 227.3 pg/mL at baseline to 1,089.4 ± 258.2 pg/mL on Day 31 (p<0.05) and 957.1 ± 250.8 pg/mL on Day 58 (p<0.001), corresponding to approximate reductions of 18% and 28%, respectively.

Plasma concentrations of the inflammatory cytokines IL-2, IL-6, and TNF-α did not differ significantly from baseline and showed no consistent directional change across time points. IL-2 declined from 6.8 ± 6.4 pg/mL at baseline to 6.2 ± 3.9 pg/mL on Day 31 before increasing above baseline by Day 58 (9.1 ± 7.7 pg/mL), while IL-6 decreased from 6.2 ± 6.1 pg/mL to 3.0 ± 0.7 pg/mL on Day 31 and was 6.7 ± 3.8 pg/mL on Day 58. TNF-α remained low throughout the study, with mean values of 0.18 ± 0.08, 0.16 ± 0.07, and 0.20 ± 0.07 pg/mL at baseline, Day 31, and Day 58, respectively (Figure 7).

In contrast, IL-8 concentrations were significantly increased on Day 58 compared to both baseline and Day 31 (p<0.01). IL-8 concentrations were similar at baseline and Day 31 (33.2 ± 16.2 and 34.8 ± 24.7 pg/mL, respectively) but increased markedly by Day 58 to 117.0 ± 95.0 pg/mL, representing an approximately 252% increase relative to baseline.

Taken together, the biomarker profile supports a preliminary signal for reduced CRP and CTX-II during dietary intervention; however, the isolated IL-8 increase highlights the complexity of interpreting circulating inflammatory biomarkers and therefore, caution should be exercised when evaluating the overall biomarker response.

Body weight and body composition

Mean body weight at study conclusion (10.77 kg) was similar to baseline (10.84 kg), representing a minimal change over the 58-day intervention. Individual body weight changes were modest and varied in direction. All dogs entered the study with an ideal or near-ideal body condition score of 5 or 6 (9-point scale). Body condition scores remained largely stable, and muscle condition scores showed only minor fluctuations in a small number of animals.

QMR analysis was performed at baseline and study conclusion to characterize body composition. No meaningful changes were observed in corrected fat mass, lean mass, or total body water over the 58-day intervention period (Figure 8). Individual values remained generally stable across animals, and overall changes were small relative to baseline variability, without evidence of a consistent directional shift.

These findings indicate that major shifts in body mass or body composition were not evident over the study period, suggesting that the observed clinical and biomarker changes were unlikely to be explained primarily by gross weight loss.

Hematology and biochemistry

No consistent treatment-related patterns were identified in hematology or serum biochemistry parameters. Values were generally stable over the study period, with group-level means remaining within or close to reference intervals for most analytes at all scheduled time points.

Occasional values outside reference ranges were observed in a limited number of animals; however, these were typically isolated, mild, and not associated with clinical abnormalities. Many out-of-range values were present at baseline or reflected pre-existing or historical findings and did not demonstrate consistent worsening or emergence during treatment. No abnormalities required medical intervention.

When restricted to newly observed values outside reference ranges during the intervention phase, abnormalities were limited in number and did not suggest a consistent treatment-related pattern. The most frequent newly observed finding was platelet count elevation, occurring in 4 dogs on Day 31 and 2 dogs on Day 58, with 2 of these cases persisting across both time points. All other newly observed abnormalities were isolated to single animals at individual time points and did not demonstrate consistent trends or clinical relevance. Hematology and biochemistry analysis data are provided as Supplementary File 2.

Veterinary examinations, health observations, and fecal scoring

Veterinary examinations identified musculoskeletal findings consistent with OA, including reduced range of motion, joint pain, and stiffness; however, no clear trends were observed across examination time points. No clinically significant abnormalities were identified in other examination parameters.

Following introduction of the test diet, four isolated episodes of vomiting were recorded over the treatment period. These events were sporadic, non-recurring, and did not require treatment. Fecal scores remained within the normal range for the majority of observations, with three instances of loose stool occurring in two individual dogs.

OA is a prevalent, progressive musculoskeletal disorder in dogs and a major contributor to chronic pain and impaired mobility (Cachon et al., 2023; Lascelles et al., 2015). In the present study, feeding of a complete high protein diet enriched with ESM and omega-3 fatty acids was associated with improvements in technician-assessed function scores, more gradual reductions in pain-related scores, and lower circulating CRP and CTX-II concentrations over 58 days in laboratory Beagles with naturally occurring OA.

A notable feature of the clinical response was the magnitude and timing of functional improvement relative to improvement in pain-related scores. Function scores improved at all post-baseline time points, whereas pain-related scores changed more gradually. The early improvement observed by Days 13–16 also suggests that repeated testing, acclimation to study procedures, handling familiarity, or observer-related factors may have contributed to the observed functional response. In the current model, the functional measure is based on willingness to engage in normal activities and therefore is a composite measure that is independent of pain per se. This pattern is consistent with the concept that nutritional interventions may exert effects gradually through modulation of inflammatory structural pathways rather than through immediate analgesic effects (Calder, 2010; Patel et al., 2026; Wang et al., 2025), although this interpretation remains speculative in the absence of a control group. The magnitude and direction of improvement observed in this study are consistent with findings from controlled studies in dogs with OA, in which reductions in pain and improvements in function have been demonstrated using a range of outcome measures assessing OA-related clinical signs (Lascelles et al., 2015). Similar patterns have been reported in canine studies evaluating omega-3-enriched diets, where improvements in mobility and functional parameters were observed over time, while changes in pain-related indices were less pronounced or developed more gradually (Moreau et al., 2012; Roush et al., 2010a). Although clinically meaningful response thresholds have not been established for the technician-administered behavioral assessment used in this study, improvements were observed in a high proportion of dogs. By the final assessment period, both pain-related scores and function scores had improved relative to baseline in all evaluated animals (11/11). The proportion of dogs demonstrating improvement increased over time for pain-related scores and was consistently high for function scores across assessment periods. This pattern suggests that the observed changes were not driven solely by a small number of extreme responders and supports the consistency of the response across the study population.

The observed reductions in CRP and CTX-II provide biologically relevant support for changes occurring during the dietary intervention. CRP was included as a nonspecific marker of systemic inflammation rather than as an OA-specific monitoring tool. Although CRP can be influenced by inflammatory conditions or other physiological processes unrelated to OA, systemic inflammation is implicated in OA pathophysiology, providing a rationale for its inclusion. Similarly, CTX-II is commonly used as an indicator of type II collagen turnover and cartilage degradation in OA research, although reductions in CTX-II alone do not establish structural disease modification. This pattern is consistent with evidence that nutraceutical and dietary interventions may modulate inflammatory and catabolic pathways implicated in OA and contribute to improvements in clinical signs (Barbeau-Gregoire et al., 2022). Because this was an uncontrolled exploratory study, these biomarker changes cannot be attributed definitively to the diet; however, the direction of change is consistent with a potentially favorable association with improved inflammatory and cartilage-related processes.

In contrast, no consistent changes were observed in circulating concentrations of IL-2, IL-6, or TNF-α, while IL-8 concentrations were elevated at study conclusion. This isolated cytokine finding should be interpreted cautiously, as such measurements are subject to biological and analytical variability, and may not consistently align with broader clinical outcomes, particularly in small exploratory studies (Barrouin-Melo et al., 2016). Additionally, some studies have shown that the evaluation of these biomarkers as indicators of OA in serum was less robust than in synovial fluid (Muller et al., 2019). Variability in inflammatory biomarker responses has been reported in OA models (but not in dogs) and may reflect transient physiological or methodological factors rather than a direct treatment effect (Sokolove and Lepus, 2013). The increase in IL-8 despite reductions in CRP and CTX-II highlights the complexity of interpreting circulating inflammatory biomarkers in OA. Although IL-8 may reflect inflammatory activity, circulating cytokine concentrations can be influenced by factors unrelated to joint disease, including transient systemic immune activation, stress, assay variability, and non-OA inflammatory processes. Therefore, the isolated IL-8 increase should not be interpreted as evidence of worsening OA or as evidence against the observed reductions in CRP and CTX-II. The clinical significance of this finding remains uncertain, and the discordant biomarker responses reinforce the exploratory nature of these results.

The test diet, characterized by a high-protein profile, incorporated both omega-3 fatty acids and ESM, each of which has mechanistic relevance to OA management. Omega-3 fatty acids, particularly EPA and DHA, modulate inflammatory pathways in part by competing with omega-6 fatty acids for eicosanoid synthesis, resulting in production of less pro-inflammatory mediators (Roush et al., 2010a). A systematic review by Barbeau-Gregoire et al. (2022) further supports omega-3-enriched diets and omega-3 supplements as having consistent evidence of clinical benefit, particularly for analgesic outcomes, in canine OA. Notably, the same analysis reported low or inconsistent evidence for several other tested nutraceuticals, including collagen-derived products, glucosamine/chondroitin combinations, and β-glucans.

Eggshell membrane provides a matrix of collagens, glycosaminoglycans, and structural proteins, and has been investigated for joint health and cartilage-related outcomes (Shi et al., 2021). Human studies have demonstrated improvements in joint pain following ESM supplementation, with some evidence also indicating reductions in cartilage degradation markers such as CTX-II (Cánovas et al., 2022; Ruff et al., 2018). While direct canine evidence remains relatively limited, a controlled study in dogs has demonstrated improvements in clinical pain and function alongside reductions in cartilage degradation markers following ESM supplementation (Ruff et al., 2016), with an additional study reporting improvements in mobility, lameness, and owner-assessed pain in dogs with OA (Aguirre et al., 2018). Together, these findings support the biological and clinical relevance of ESM as part of a multi-component nutritional strategy; however, because the present intervention was a whole-diet strategy rather than a component-isolation study, the specific contribution of ESM, omega-3 fatty acids, or the overall macronutrient profile cannot be determined. Accordingly, the present findings should not be interpreted as evidence for the independent efficacy of ESM. Rather, they support further evaluation of the complete dietary strategy used in the current study in which ESM was one component.

The high-protein profile of the diet has relevance to OA management through its effects on maintenance of lean body mass and mobility. Adequate protein intake is important for preserving muscle mass, particularly in aging or less active dogs, and loss of lean muscle has been associated with declines in mobility and physical function in older dogs (Bellows et al., 2015; Hutchinson et al., 2012). Nutritional strategies incorporating higher protein content, often in combination with reduced carbohydrate intake, have been shown to support fat loss while preserving lean mass during weight management in dogs (Diez et al., 2002; Vanelli et al., 2025). Maintenance of lean body mass is an important consideration in OA, as muscle function contributes to joint stability, load distribution, and overall mobility. Together, these factors suggest that the macronutrient composition of the diet may have contributed to the observed improvements in functional outcomes.

Body weight and body composition remained generally stable throughout the study. Given prior evidence that weight loss alone can improve lameness and mobility in dogs with OA (Marshall et al., 2010; Mlacnik et al., 2006), the absence of meaningful body mass change suggests that the observed improvements in pain-related scores and function scores were unlikely attributable to reductions in body weight.

In this context, the diet was formulated with a specific nutritional profile, highly adapted to the prevention of overweight and obesity, as excess body weight is known to be detrimental to OA related clinical signs. Indeed, with a high protein (37.4% DM, 101 g/1000 kcal), low starch (20.6% DM), and moderate energy content (337 kcal/100g; calculated NRC 2006), it allows for control of body weight and body composition, even in neutered dogs or patients with decreased activity, provided it is appropriately rationed (National Research Council, 2006; Kawauchi et al., 2017; Vendramini et al., 2020).

The diet was generally well tolerated over the study period. Vomiting was infrequent and transient, and fecal consistency remained largely within normal limits. Such transient gastrointestinal signs are commonly observed following dietary transitions during feeding trials, and no clinically meaningful hematology or biochemistry abnormalities were identified. These findings support the feasibility of sustained dietary administration, which is an important consideration for long-term OA management strategies in which nutritional approaches may play a supportive role (Cachon et al., 2023).

Several limitations of the present study should be acknowledged. First, the study employed an open-label, single-arm, baseline-controlled design without a parallel control diet, which limits the ability to attribute observed effects solely to the dietary intervention. A randomized, controlled design including a control diet and component-specific comparator groups would be required to distinguish the effects of the complete diet from those of individual ingredients. In addition, regression to the mean, acclimation to repeated behavioral testing, handling effects, natural fluctuation in OA signs, and observer expectation bias cannot be excluded. Because the primary endpoint relied on technician-assessed behavioral scoring by unblinded personnel rather than objective gait or mobility assessments (e.g., force-plate analysis or activity monitoring), the magnitude of functional improvement should be interpreted cautiously. Second, the sample size was small and based on animal availability rather than a formal a priori power calculation. In addition, the use of a homogeneous research colony Beagle population may reduce biological and environmental variability but limits generalizability to client-owned dogs with more diverse breeds, environments, OA phenotypes, owner management practices, and comorbidities. Third, the primary behavioral outcome was assessed using a technician-administered behavioral assessment instrument developed for use in a research colony setting that has not been formally validated as a clinical metrology instrument. Although prior internal evaluation work with the behavioral scoring approach supports discrimination between OA and non-OA populations and responsiveness to meloxicam, formal interobserver reliability, repeatability, and clinically meaningful response thresholds have not been established. Fourth, the 58-day intervention provides only short-term information and does not establish durability of response. Longer follow-up would be required to determine whether clinical or biomarker changes persist, plateau, or diminish with continued feeding. Finally, one dog was withdrawn mid-study due to a cruciate ligament injury, further reducing the already limited sample size for later assessments. Despite these limitations, the directional consistency of improvements in the technician-assessed functional outcomes, alongside improvements in pain-related scores and reductions in CRP and CTX-II, support further evaluation of this dietary approach. These findings should be considered hypothesis-generating and require confirmation in randomized, controlled, blinded designs including larger and more heterogeneous populations.

In this exploratory, open-label, baseline-controlled study, feeding a complete high protein diet enriched with ESM and omega-3 fatty acids was associated with improved technician-assessed function scores, reduced pain-related scores, and lower circulating CRP and CTX-II concentrations in research colony dogs with naturally occurring OA. These hypothesis-generating findings support further evaluation of targeted nutritional strategies within multimodal OA management and warrant confirmation in randomized, controlled studies in broader canine populations.