Collagen Hydrolysates as Nutritional Support in Canine Osteoarthritis: A Narrative Review

ABSTRACT

Osteoarthritis (OA) is a common disease in dogs with severe impact on their welfare. The multimodal management of OA includes feeding therapeutic diets and nutraceuticals to slow down OA progression. Collagen hydrolysates (CH) are a nutritional supplement that may exert anabolic effects on osteoarthritic joint cartilage as well as disease‐modifying effects. After oral intake, CH is absorbed, mainly as amino acids, di‐ and tripeptides that are transported amongst others to the joint. In addition to reducing cartilage degradation, CH metabolites may reduce synovial inflammation and subchondral bone sclerosis during OA. Preliminary evidence in dogs suffering from the consequences of OA support the clinical efficacy of CH with reported reductions in lameness. However, effects on biomarker level of cartilage metabolism and inflammation are inconclusive. Additionally, current studies show a lack of standardised dosing regimens and the use of not validated outcomes. Future work should therefore elucidate further on the bioavailability of CH in dogs in order to establish adequate dosing recommendations. Furthermore, high‐quality placebo‐controlled randomised controlled trials are essential to dstudies have evaluated the cetermine the clinical efficacy of CH to reduce lameness, prevent OA progression and thereby improve the level of evidence.

1. Introduction

1.1. Osteoarthritis (OA) in Dogs

OA is a disorder of the synovial joint associated with chronic pain and progressive loss of limb function (Belshaw, Dean, and Asher ²⁰²⁰). Estimates of its prevalence are as low as 2.5% up to > 20%, depending on the investigated population and criteria for diagnosis (Anderson et al. 2018; Enomoto et al. ²⁰²⁴). Whereas OA development can be related to arthropathies such as hip dysplasia (Smith et al. 2012), OA development is also strongly influenced by obesity, breed and advancing age (Anderson et al. 2020; Marshall et al. ²⁰⁰⁹).

OA encompasses an interplay between joint tissues via complex biochemical and cellular signalling. Its first stages can be initiated by various factors, such as metabolism‐triggered inflammation, abnormal load‐bearing or joint trauma (Renberg 2005; Wang et al. ²⁰¹⁵). In reaction to this initial trigger, chondrocytes increase the production of extracellular cartilage matrix (ECM) components in an attempt to repair the damage (Lippiello, Hall, and Mankin ¹⁹⁷⁷; Ryu, Treadwell, and Mankin ¹⁹⁸⁴). When this response fails, and/or the initiating trigger persists, the tightly regulated balance of cartilage ECM metabolism shifts to a catabolic state. Simultaneously, an inflammatory response is triggered that causes influx of immune cells in the synovial membrane and release of inflammatory mediators, such as interleukin (IL)‐1ß and IL‐6 (Scanzello and Goldring 2012; Sanchez‐Lopez et al. ²⁰²²). This maintains the catabolic joint environment and causes synovial hyperplasia and inflammation. In the end, the catabolic environment perpetuates the vicious cycle resulting in loss of joint cartilage and chronic synovitis. This ultimately leads to (chronic) pain and limited joint function in OA patients, hence substantially affecting their quality of life (Anderson et al. 2018; Lee et al. ²⁰¹³; Summers et al. ²⁰¹⁹).

1.2. Treatment of Canine Osteoarthritis and Its Limitations

The majority of canine OA patients are supported through both surgical and nonsurgical means to optimise patient welfare and prevent further deterioration of the joint (Sandersoln et al. 2009). So far, most canine OA patients are prescribed nonsteroidal anti‐inflammatory drugs (NSAIDs) to relieve OA‐related joint pain in addition to life‐style changes such as avoiding strenuous exercise (Anderson et al. 2018; Pye et al. ²⁰²⁴). Additionally, weight management and weight reduction in obese patients can further reduce OA‐related lameness and pain (Marshall et al. 2010). However, the risk of adverse events, such as gastrointestinal ulceration and renal insufficiency, increases with prolonged use of NSAIDs. This underlines the need for alternative management strategies for OA that minimise NSAID use throughout a dog's life (Hunt et al. 2015; Pavlova et al. ²⁰²¹). One such strategy is presented by nutrition and dietary supplements, which is the focus of the present review. Besides having the ability to reduce joint pain in dogs with OA, some ingredients are able to delay OA progression and supporting the pain management of NSAIDs (Johnson, Lee, and Swanson ²⁰²⁰; Vandeweerd et al. ²⁰¹²).

Bioactive peptides derived from collagen hydrolysates (CH), sometimes referred to as ‘specific collagen peptides’ or ‘gelatin hydrolysates’, present a promising source of nutritional support for dog OA patients (Beynen et al. 2010; Dobenecker et al. ²⁰²⁴). It is thought that in order to exert their biological activity, the shorter sequences of the bioactive peptides must be released from the larger protein structure by enzymatic or chemical treatment prior to oral administration (Möller, Scholz‐Ahrens, et al. ²⁰⁰⁸; Rinaldi et al. ²⁰¹⁵; Virgilio ²⁰¹⁹). In humans and horses with OA, orally administered CH exerted positive clinical effects on joint pain on patient reported outcomes and orthopaedic examination (Dobenecker et al. 2018; Honvo et al. ²⁰²⁰). Several studies have investigated the clinical effects of CH in dogs with OA (Table 1) (Beynen et al. 2010; Comblain et al. ²⁰¹⁷; Dobenecker et al. ²⁰²⁴; Eckert et al. ²⁰²¹; Schunck, Louton, and Oesser ²⁰¹⁷). This review summarises and evaluates current evidence on kinetics, potential modes of action and clinical efficacy of CH supplementation in dogs with OA. First, a background on CH and its pharmacokinetics will be provided, followed by current insights into CH's modes of action based on experimental in vitro and in vivo studies. Lastly, directions for future research are discussed based on the identified research gaps.

Table 1.

Overview of currently published clinical evidence on collagen hydrolysates in osteoarthritic dogs.

References	Study type	Patient population	Experimental groups	CH dosage	Treatment duration
Beynen et al. (2010)	Double‐blind, placebo‐controlled trial	Client‐owned dogs with clinical signs of osteoarthritis	CH mixed with soya protein isolate (1:1 ratio); n = 15 Soya protein isolate (placebo); n = 15	10 g/dog	8 weeks
Comblain et al. (2017)	Randomised, double‐blind, placebo‐controlled trial	Client‐owned dogs with clinical and radiographic osteoarthritis	CH, curcuminoids extract and green tea extract enriched diet; n = 21 Control diet; n = 21	Diet containing 15 g/kg CH, 1.5 g/kg curcuminoids extract and 3 g/kg green tea polyphenols	12 weeks
Dobenecker et al. (2024)	Randomised, double‐blind, placebo‐controlled trial	Client‐owned dogs with clinical osteoarthritis	Bioactive collagen peptides; n = 11 Omega‐3 fatty acids; n = 11 Control (cellulose and maltodextrin); n = 9	≥ 200 mg/kg BW	12 weeks
Eckert et al. (2021)	Randomised controlled trial	Client‐owned dogs with osteoarthritis	CH; n = 20 Sulphated glucosamine; n = 21 Therapeutic diet: Hill's J/D (control group); n = 10	20 g/dog	16 weeks
Schunck, Louton, and Oesser (2017)	Prospective study	Client‐owned dogs with osteoarthritis and treatment‐resistant history	Specific collagen peptides; n = 22	5 g/dog	12 weeks

References	Study outcomes	Results
Beynen et al. (2010)	Subjective owner assessments via VAS: Activity, stiffness, joint swelling, lameness, paralysis, pain and body condition	Significant improvement in activity, stiffness and lameness with CH
Comblain et al. (2017)	Force plate gait analysis, biomarkers of cartilage anabolism Veterinary assessed pain at manipulation, lameness, joint mobility Canine brief pain inventory	No statistical difference in ground reaction forces and OA biomarkers Significant reduction of veterinarian assessed pain at manipulation and decrease in owner‐assessed pain severity in enriched diet group
Dobenecker et al. (2024)	Canine brief pain inventory Force plate gait analysis Activity via omnidirectional accelerometers	Improvement in quality‐of‐life item of canine brief pain inventory Significant differences in peak vertical force, vertical impulse and the symmetry index for PVF, with bioactive collagen peptides
Eckert et al. (2021)	Subjective owner assessments: Back pain sensitivity, quality of life assessment, mobility and agility Helsinki chronic pain inventory Subjective veterinary assessment: Degree of lameness, joint pain Bone damage and thigh circumference by X‐ray	Significant reductions in lameness symptoms in collagen hydrolysate and sulphated glucosamine group Improvements in pain at limb manipulation, quality of life score and Helsinki chronic pain index in all three groups
Schunck, Louton, and Oesser (2017)	Subjective veterinary assessment: Lameness, pain at activity, joint swelling, restriction of mobility, joint instability Subjective owner assessment: vitality, discomfort when climbing stairs, problems when standing up, tenderness on palpation	Reduction of lameness at 12 weeks after study start Increased vitality and decreased discomfort in standing at 12 weeks

Abbreviations: CH, collagen hydrolysates; OA, osteoarthritis; VAS, visual analogue scale.

2. Collagen Hydrolysates: Production and Absorption

CH is a heterogeneous mixture of collagen‐derived oligopeptides and amino acids (AAs). It is produced by denaturation and subsequent hydrolysis of collagen extracted from raw materials such as skins, bones or fish scales (Hong et al. 2019). The resulting mixture of oligopeptides has a low molecular weight, on average 2–6 kDa, which ensures high digestibility and absorption and has a very low risk for inducing feed‐induced allergic responses (Bai et al. 2024; Oesser et al. ¹⁹⁹⁹; Olivry, Bexley, and Mougeot ²⁰¹⁷).

Raw materials and processing methods determine CH's characteristics. First, the amino acid composition of collagens is species‐specific and the abundance of specific collagen types is different for each tissue (Gelse 2003; Gómez‐Guillén et al. ²⁰¹¹; Lin and Liu ²⁰⁰⁶). In addition, the degree of collagen denaturation and hydrolysis, and therefore the resulting particle size and oligopeptide profile after collagen processing, is influenced by the chosen methods of collagen extraction and processing, as was reviewed in a recent paper (Hong et al. 2019). These include the used proteolytic enzymes, incubation time during hydrolysis (Ao and Li 2012), and processing temperature (Nikolaeva et al. 2021). Differences in collagen processing may therefore affect the bioavailability of specific metabolites and the bioactivity of CH products (Ao and Li 2012; Oesser et al. ¹⁹⁹⁹).

After oral intake, CH are digested and absorbed, releasing CH‐peptides into the systemic circulation (Iwai et al. 2005; Larder, Iskandar, and Kubow ²⁰²¹). Absorption of hydrolysed peptides is thought to occur across the small intestine, as was previously shown in rats and in in vitro digestibility models in dogs and cats (Bai et al. 2024; Kawaguchi, Nanbu, and Kurokawa ²⁰¹²; Osawa et al. ²⁰¹⁸), which may be related to the high abundance of PepT1 transport proteins (Spanier and Rohm 2018). Hydroxyproline (Hyp) and other CH‐derived peptides such as prolyl‐hydroxyproline (Pro‐Hyp) and hydroxyprolyl‐glycine (Hyp‐Gly) reached maximal concentrations in the blood within 1–2 h post‐ingestion in humans, returning to baseline concentrations within 24 h (Iwai et al. 2005; Virgilio et al. ²⁰²⁴). Oral administration of CH with radiolabeled proline (Pro) or prolyl‐hydroxyproline (Pro‐Hyp) in rodents led to accumulation of radioactivity in cartilage, bone and the synovium (Oesser et al. 1999; Kawaguchi, Nanbu, and Kurokawa ²⁰¹²; Watanabe‐Kamiyama et al. ²⁰¹⁰). Even though other CH‐derived oligopeptides may show different accumulation patterns than Pro and Pro‐Hyp after oral ingestion of CH, these data suggest transport and uptake by joint tissues.

Some CH‐peptides that are thought to exert biological activities have been identified in the circulation of humans after oral intake of CH, most notably Pro‐Hyp, alanyl‐hydroxyproline (Ala‐Hyp) and hydroxyprolyl‐glycine (Hyp‐Gly) (Iwai et al. 2005; Iwasaki et al. ²⁰²²; Shigemura et al. ²⁰¹⁸; Virgilio et al. ²⁰²⁴). Yet, it is unknown to what extent the same peptides are absorbed in dogs. Dogs present with differences in digestive physiology compared to humans, such as different jejunal villi morphology and expression of brush border enzymes that allow for intestinal digestion and absorption (Dressman 1986; Hatton et al. ²⁰¹⁵; Hendriks, van Baal, and Bosch ²⁰¹²). Additionally, an impact of commercial processing in pet food on CH bioavailability and efficacy cannot be excluded. Specific studies on CH‐peptide bioavailability and its distribution are therefore warranted in dogs.

3. Collagen Hydrolysates Influence Metabolism of Cartilage and Joint‐Related Tissues

After infiltration into the joint tissues, it is hypothesised that CH‐peptides modulate the joint environment (Oesser et al. 1999; Oesser and Seifert ²⁰⁰³). In vitro studies in primary cells and cell‐lines show that CH increase the synthesis of collagen type II in healthy bovine chondrocytes and glycosaminoglycan production in healthy canine chondrocytes and mouse ATDC5 cell‐lines (Nakatani et al. 2009; Oesser and Seifert ²⁰⁰³; Schunck, Louton, and Oesser ²⁰¹⁷). This is further supported by Pro‐Hyp triggering hyaluronic acid synthesis in a rabbit HIG‐82 synoviocyte cell line (Ohara et al. 2010). These findings suggest that this CH‐derived peptide may affect proteoglycan production. In addition, CH media supplementation reduced catabolic effects of cytokine IL‐1β added in concentrations ranging from 1 (De Luca et al. 2019) to 20 ng/mL (Lee et al. 2021) on aggrecan and type II collagen gene expression in chondrocytes obtained from human OA patients (Bourdon et al. 2021; Lee et al. ²⁰²¹; De Luca et al. ²⁰¹⁹).

Some of these findings have been disputed. Several CH formulations were tested in a cartilage explant model of human patients with mild and moderate OA, but none of them increased the synthesis of cartilage ECM molecules, while one formulation even increased catabolic mediators nitric oxide (NO) and prostaglandin E₂ (Schadow et al. 2013). Similar results were reported, as some CH formulations caused elevated levels of ECM degradative enzymes MMP‐1, ‐3 and ‐13 (Schadow et al. 2017). Whether these findings relate to specific CH formulations is unclear and require further examination. It must also be considered that a wide range of CH concentrations was used in these models. Whereas the cartilage assembling effects of CH were established in doses ranging from 50 μg/mL (Bourdon et al. 2021) to 1 mg/mL (De Luca et al. 2019), the catabolic effects were generally reported in doses > 1 mg/mL (Schadow et al. 2013, ²⁰¹⁷).

Complimentary to findings in chondrocytes, CH has been found to influence OA‐related immune and bone cell biology in vitro. For instance, CH‐peptide exposure led to reduced secretion of inflammatory cytokine TNF by LPS‐stimulated murine RAW264.7 macrophages by downregulating the inflammatory nuclear factor‐κB (NF‐κB) pathway (Sivaraman and Shanthi 2021). Additionally, Pro‐Hyp reduced COX‐2 and iNOS expression in LPS‐stimulated RAW264.7 macrophages (Zhu et al. 2020) and drove T cell development towards regulatory T cells and T helper‐1 phenotypes by activating the STAT‐1 signalling pathway (Nishikimi et al. 2018). Lastly, osteoclast activity was reduced and osteoblast activity increased after CH exposure due to decreased osteoclast differentiation, which could present itself as reduced subchondral bone sclerosis and cyst formation in vivo (Guillerminet et al. 2010; Kimira et al. ²⁰¹⁴).

It must be noted, however, that the undigested CH‐peptides supplied to most in vitro models may not resemble the oligopeptides that end up in the circulation in vivo (Iwai et al. 2005; Larder, Iskandar, and Kubow ²⁰²¹) and as such the peptides that the joint tissues are exposed to. Additionally, these products may bear low‐grade endotoxin contamination. Endotoxin contamination, such as of lipopolysaccharide (LPS), has previously been reported in pharmaceutical grade gelatine‐based products and is common in both the pharmaceutical and food processing industry (Heinrich et al. 2023; Sandle ²⁰¹⁹; Schneier et al. ²⁰²⁰). While this contamination may not necessarily have biological consequences in vivo, it can act as a strong confounder in vitro (Heinrich et al. 2023). Hence, there is a need for models using digested CH, for example, by using serum obtained after oral CH ingestion, but this has not yet been tested in in vitro models in canine or human OA (Daneault et al. ²⁰¹⁷).

In vivo studies using animal models of post‐traumatic OA (PTOA) seem to correlate with in vitro findings. In animals fed CH after PTOA induction, loss of joint cartilage was reduced whereas the histological cartilage integrity was increased in mice with meniscal‐ligamentous injury (MLI) (Dar et al. 2017), rabbits with anterior cruciate ligament transection (ACLT) (Lee et al. 2021), and in rats that underwent a combination of medial meniscus resection (MMx) and ACLT (Hao et al. 2022). These findings were most likely the result of increased cartilage ECM deposition, as was evidenced by dose‐dependent increases in type II collagen‐ and proteoglycan‐rich tissue deposition and reductions in MMP‐13 immunopositivity compared to untreated controls (Dar et al. 2017; Hao et al. ²⁰²²; Lee et al. ²⁰²¹). Furthermore, chondrocyte viability after PTOA induction was preserved in mice and rats fed CH (Dar et al. 2017; Hao et al. ²⁰²²). These studies support the notion that CH may have a chondroprotective effect and therefore the potential to delay OA progression after joint injury.

Despite using different animal models for OA, significant reductions in synovial hyperplasia and inflammation were observed in murine MLI (Dar et al. 2017) and rat MMx+ACLT (Hao et al. 2022) models of PTOA after 12 weeks of CH intake. These observations were supported by decreased release of IL‐1β and IL‐6 in the synovial fluid and lower gene expression of TNF in the synovial membrane (Dar et al. 2017; Hao et al. ²⁰²²; Lee et al. ²⁰²¹). Subchondral bone sclerosis and osteophyte formation were reduced in a rabbit ACLT model after 12 weeks of CH intake (Lee et al. 2021), possibly related to alterations in osteoclast and osteoblast activity that were reported in vitro (Guillerminet et al. 2010; Kimira et al. ²⁰¹⁴).

The exact mechanism behind the bioactive effects is yet unknown. However, it has been hypothesised that CH exert its actions through delivery of specific nutrients and interactions of bioactive peptides with cellular signalling (De paz‐Lugo, Lupiáñez, and Meléndez‐Hevia ²⁰¹⁸; Wu et al. ²⁰¹¹). High concentrations of the nonessential amino acids Hyp, Pro and Gly are supplied by CH (Ao and Li 2012), which are essential for type II collagen synthesis in situation of high demand, such as early‐stage OA (Nelson et al. 1998; de De paz‐Lugo, Lupiáñez, and Meléndez‐Hevia ²⁰¹⁸; Wu et al. ²⁰¹¹). Moreover, bioactive CH peptides influenced important cellular signalling pathways in OA pathophysiology, such as the NF‐κB, Wnt/β‐catenin and MAPK signalling pathways (Hao et al. 2022). Another mode of action of CH could be as prebiotic, since its oral intake counteracted OA‐related changes in the gut microbiome, which correlated with a delayed onset of OA in a mouse PTOA model (Soniwala et al. 2018). Yet, in contrast to humans (Hao et al. 2021), the significance of the gut microbiome in OA pathogenesis in dogs is less established (Cintio et al. 2020).

4. Clinical Studies on the Use of Collagen Hydrolysates in Dogs With OA

4.1. Supplementation of Collagen Hydrolysates Alleviates Clinical Symptoms of OA

To date, few studies have evaluated the clinical efficacy of CH in dogs with OA, but the results so far are promising (Table 1) (Beynen et al. 2010; Comblain et al. ²⁰¹⁷; Dobenecker et al. ²⁰²⁴; Eckert et al. ²⁰²¹; Schunck, Louton, and Oesser ²⁰¹⁷). To the author's knowledge, there are no studies that evaluated the effects of CH on long‐term OA prevention and progression, as such studies typically require long‐term investigations of patients that are at risk of OA development.

Typically, CH has been dosed in dogs with OA according to fixed dosing regimens comparable to those used in people (Beynen et al. 2010; Eckert et al. ²⁰²¹; Kumar et al. ²⁰¹⁵; Schunck, Louton, and Oesser ²⁰¹⁷). A dose of 10 g CH/day for 8 weeks added to the patients' regular OA treatment improved clinical lameness and stiffness by ~30% compared to a placebo, rated by owners on continuous graphical scales (Beynen et al. 2010). Similar improvements in lameness were found in combination with less pressure‐induced joint pain with 5 g CH/day as evaluated by veterinarians (Schunck, Louton, and Oesser ²⁰¹⁷). This study provided CH for 12 weeks, but lacked a control group. More recently, supplementation of 20 g CH/day was compared to treatment with sulphated glucosamine and a therapeutic diet for OA patients (Eckert et al. 2021). Reductions of lameness symptoms and pressure‐induced joint sensitivity, both with unspecified severity levels, were reported in both the CH and glucosamine group by owners and veterinary evaluation. However, chronic pain evaluated with the validated Helsinki chronic pain index (HCPI) did not differ significantly between groups.

The aforementioned studies used fixed dosing regimens, which cause variations in the relative intake of a supplement in a species with a wide range of body weights like the dog. This might result in different bioactive effects in patients of different size. A recent randomised controlled trial therefore investigated the effects of a weight‐based dose of CH (200 mg/kg) on chronic pain evaluated by the canine brief pain inventory (CBPI) score. Two comparison groups were given supplements containing either omega‐3 fatty acids (550 mg/kg) with vitamin E (2 mg/kg) or a placebo (Dobenecker et al. 2024). CH intake increased the CBPI quality of life score from baseline measures compared to the omega‐3 fatty acid group, but not to the placebo. Additionally, owners did not report significant reductions in OA pain and limb function, as evaluated by CBPI (Dobenecker et al. 2024). However, several objective measures performed with pressure plates improved during the study. In this study, the difference Δ between measurements of ground reaction forces (peak vertical impact [PVI], vertical impulse [VI] and stance duration [SD]) at baseline and after 12 weeks of supplementation were evaluated. There was a significant increase of Δ for PVF and VI in the group supplemented with CH, while the Δ DSP increased only numerically. The relative differences in the gait parameters of the affected and the non‐affected limbs revealed that only the dogs in the CH supplemented group had significantly improved during the 12‐week period (Dobenecker et al. 2024). Lastly, the symmetry index for peak vertical force improved in the CH group. Activity of the dogs, measured with omnidirectional accelerometers, was not significantly different between groups (Dobenecker et al. 2024). Altogether, this implies that a supplementation of CH can improve limb load bearing already after 12 weeks, but this may not necessarily translate to improvements visible to the owners.

Incorporation of nutraceuticals in diets for pets with OA is a common practice in veterinary medicine, and as such Comblain et al. (2017) compared a pet food enriched with 17.1 g CH/kg dry matter (DM), 1.71 g curcuminoids/kg DM and 3.4 g green tea polyphenols/kg DM to a similar pet food without these ingredients. These diets were fed for 12 weeks for adult maintenance, leading to an approximate intake of 400 mg/kg^0.75 CH. Dogs fed the therapeutic diet showed reduced pain at manipulation by veterinarians, assessed on a 5‐point scale. In contrast to the aforementioned studies, no differences were found in lameness evaluated by veterinarians, load‐bearing via ground force plate analysis, and the other CBPI domains. It is possible that other nutrients reduced CH's efficacy by affecting CH absorption (Walrand et al. 2008), but this requires further investigation.

4.2. Effects of CH Supplementation on Biomarkers for OA

In addition to clinical measures, concentrations of biomarkers have been measured to gain insights in CH's metabolic effects on joint tissue metabolism in OA patients (Comblain et al. 2017; Eckert et al. ²⁰²¹; Schunck, Louton, and Oesser ²⁰¹⁷). Biomarkers are naturally occurring molecules that may act as indicators for physiological and pathological processes (Califf 2018). In the case of OA, they are mainly related to joint tissue metabolism or inflammation, such as ECM‐derivatives or matrix degrading enzymes, and are primarily measured in synovial fluid or peripheral blood (Rousseau, Chapurlat, and Garnero ²⁰²¹).

Evaluation of biomarkers in only three studies and with a limited number of patients studied, that is, between 3 (Schunck, Louton, and Oesser ²⁰¹⁷) and 23 (Comblain et al. 2017; Eckert et al. ²⁰²¹), indicated conflicting effects of CH on cartilage metabolism in dogs with OA (Table 1). Serum levels of joint‐catabolism marker MMP‐3 decreased during CH treatment (Eckert et al. 2021; Schunck, Louton, and Oesser ²⁰¹⁷), whereas concentrations of MMP‐inhibitor TIMP‐1 and collagen synthesis marker procollagen type II increased (Schunck, Louton, and Oesser ²⁰¹⁷). On the other hand, the same study reported increased serum concentrations of cartilage oligomeric matrix protein (COMP) and hyaluronic acid that have both been associated with knee OA progression in people (Bastick et al. 2015; Hoch et al. ²⁰¹¹). However, it must be noted that this study measured concentrations of biomarkers in only three patients (Schunck, Louton, and Oesser ²⁰¹⁷). TIMP‐1 and collagen synthesis marker Coll2‐1 concentrations were unaltered in the other studies (Comblain et al. 2017; Eckert et al. ²⁰²¹). It is tempting to hypothesise that CH improves the net collagen content in the ECM, evidenced by increased procollagen type II levels and reduced degradation via MMP‐3 in OA patients (Eckert et al. 2021; Schunck, Louton, and Oesser ²⁰¹⁷). Yet, these findings need to be confirmed in larger cohorts of canine OA patients treated with CH.

4.3. Limitations of Clinical Research and Future Study Directions on CH in Canine OA

It is clear from the studies above that CH supplementation can improve clinical complaints of OA in dogs. Differences were seen in the magnitude of effects as measured by different outcome parameters (Beynen et al. 2010; Dobenecker et al. ²⁰²⁴; Comblain et al. ²⁰¹⁷; Eckert et al. ²⁰²¹). This might be dose‐related, as dosages varied between fixed dosages of 5–20 g/dog (Beynen et al. 2010; Eckert et al. ²⁰²¹; Schunck, Louton, and Oesser ²⁰¹⁷) to weight‐based dosages of 200 mg/kg (Dobenecker et al. 2024) or incorporation in therapeutic diets (Comblain et al. 2017) (Table 1). In addition, there is a limited number of studies available, of which only three were randomised‐controlled trials (Beynen et al. ²⁰¹⁰; Dobenecker et al. ²⁰²⁴; Comblain et al. ²⁰¹⁷), with high heterogeneity in study design between them. Hence, more studies are needed to further substantiate these findings.

Essential study information regarding patient selection criteria, patient allocation and randomisation as well as OA severity and location was not reported in three studies (Beynen et al. 2010; Eckert et al. ²⁰²¹; Schunck, Louton, and Oesser ²⁰¹⁷), even though these factors explicitly affect intervention effectiveness and biomarker concentrations (Bannuru et al. 2019; Rousseau, Chapurlat, and Garnero ²⁰²¹). Based on findings in the preclinical studies described earlier, dogs with early‐onset OA may benefit most from CH supplementation; it could counteract early stage collagen losses by providing amino acids that are essential in this stage to increase, but this hypothesis still needs to be confirmed (Nelson et al. 1998; de De paz‐Lugo, Lupiáñez, and Meléndez‐Hevia ²⁰¹⁸; Oesser and Seifert ²⁰⁰³). Since OA is a complex heterogeneous disorder that is affected by many phenotypical characteristics of the patient (Deveza et al. 2017), wide inclusion criteria, for example, recruiting patients with advanced or end‐stage OA, may lead to uneven treatment responses and therefore would require large patient numbers to account for these factors. Further stratification can be achieved by grouping according to the affected joint, radiological and clinical OA severity, OA origin, or even by biomarker concentrations when sufficient sample sizes can be achieved. Stratifying inclusion to these criteria will lead to homogenisation of treatment groups that are optimised to detect treatment effects of CH (McAlindon et al. ²⁰¹⁵) and even assist in identifying which canine patients will benefit most from CH treatment.

There is a need to investigate the bioavailability and circulating small CH peptides after oral intake of CH in dogs. Information on the bioavailability of bioactive peptides after oral CH intake in dogs is essential to establish an effective dosing strategy for clinical studies. Yet, the bioavailability and release of the specific oligopeptides can be affected by the collagen source and manufacturing processes (Nikolaeva et al. 2021). Differences in the resulting particle sizes have previously been reported to affect the bioavailability and therefore the biological efficacy of CH‐related peptides (Oesser et al. 1999; Ohara et al. ²⁰⁰⁷). The clinical relevance of these changes have not been reported yet, but this uncertainty can be mitigated by complete bioavailability studies and identification of circulating oligopeptides after CH intake of specific CH products in dogs. Until the bioactive CH‐related peptides have been identified, minimum study information provided for a product should ideally be the product's raw material origin, purity of the product, peptide size and concentration, included collagen types, amino acid composition and mineral content of the product and dosing rationale to allow for evaluation.

Whereas the study of Dobenecker et al. exclusively used previously validated outcomes in the form of the CBPI and force plate gait analysis (Dobenecker et al. 2024), the majority of studies evaluated unvalidated subjective outcomes of pain and lameness. To illustrate, some studies used pressure‐induced joint pain, which is prone to inter‐rater variability, or stiffness evaluated by continuous graphical scales that were unique to each study (Beynen et al. 2010; Comblain et al. ²⁰¹⁷; Eckert et al. ²⁰²¹). As a result, many similar outcomes showed differences in parameter definitions and scale‐width. To illustrate, the outcome lameness was defined differently; whereas one study defined the highest grade of lameness as: ‘Reluctant to rise and will not walk more than five paces’ (Comblain et al. 2017), another specified lameness of a similar degree as: ‘High degree of lameness with relieving posture of the affected limb’ (Schunck, Louton, and Oesser ²⁰¹⁷). Multiple validated clinical metrology instruments (CMIs) are available for clinical monitoring of the canine OA patient. These include the Liverpool Osteoarthritis in Dogs (LOAD) (Walton et al. 2013), HCPI (Hielm‐Björkman, Rita, and Tulamo ²⁰⁰⁹), and the CBPI (Brown et al. 2008) for evaluation of OA lameness and chronic pain by dog owners, and the Canine OsteoArthritis Staging Tool (COAST) for OA evaluation by veterinarians (Cachon et al. 2018). Using such outcome as major or primary outcome will minimise the risk of collecting unrepeatable and unreliable data on treatment effectiveness, but also enables translation of study findings into future treatment recommendations (Belshaw, Asher, and Dean ²⁰¹⁶; Vandeweerd et al. ²⁰¹²).

Yet, all subjective outcomes bear a risk of placebo effects and biases, and inclusion of objective measures has been recommended as unbiased method to evaluate effects of interventions in OA patients (Conzemius and Evans 2012; Walton et al. ²⁰¹³). Force plate gait analysis was used in two CH studies (Comblain et al. 2017; Dobenecker et al. ²⁰²⁴) and is generally considered a sensitive, or even ‘gold standard’, method to monitor limb lameness (Brønniche Møller Nielsen et al. 2020; Carr and Dycus ²⁰¹⁶). Accelerometer‐based activity monitoring could also objectively monitor patient activity (Brown, Boston, and Farrar ²⁰¹⁰) but is limited in the sense that activity is primarily regulated by dog owners and that dogs may adapt to chronic OA pain by reducing their activity (Dobenecker et al. 2024; Walton et al. ²⁰¹³). Furthermore, developments in deep‐learning artificial intelligence technologies are continuously improving the sensitivity of these methods to detect dog pain behaviours, and their use in future studies with CH may provide for repeatable and reproducible objective evaluation in large cohort studies (Chambers et al. 2021).

The effects of CH on canine joint cartilage health have not been evaluated in any of the aforementioned clinical studies and it is debatable whether the structural improvements caused by CH can be visualised within the limited timespan of these studies. Only MRI can directly visualise cartilage in a clinical modality, but its use is unfeasible for regular evaluation due to limited availability, a need for general anaesthesia and significant costs (D'Anjou et al. ²⁰⁰⁸). However, MRI evaluations of OA‐affected joints in humans have revealed improvements in cartilage structure after 24 weeks of taking CH (Gonçalves 2017, McAlindon et al. ²⁰¹¹). Earlier evaluations of cartilage structure have not been performed to the author's knowledge, and the onset of these structural changes in cartilage in relation to the duration of CH intake is yet to be clarified. Ultrasound has been studied in humans to evaluate synovial proliferation and synovial fluid volumes visually, but this use has not yet been reported in dogs with OA (Möller, Bong, et al. ²⁰⁰⁸) and may be significantly challenged by the relatively smaller joints and considerably thinner articular cartilage layer compared to man (Frisbie, Cross, and McIlwraith ²⁰⁰⁶). Structural joint changes also correlate poorly with clinical signs of OA (Felson and Neogi 2018), and it is therefore feasible to focus on clinical signs of OA as a leading outcome in future work.

Within this context, OA biomarkers could provide information of joint metabolism and inflammation, and some of these have been validated for use in dogs (de Bakker et al. ²⁰¹⁷). For instance, concentrations of IL‐6, IL‐8 and monocyte chemoattractant protein (MCP)‐1 correlate well with clinical progression of OA symptoms and synovitis as well as the response to interventions (Garner et al. 2011; Mabey ²⁰¹⁵). They can be complemented with markers of ECM metabolism, but to the authors' knowledge many of these, including COMP, TIMP‐1, Pro‐Collagen II and MMP‐3 used in aforementioned studies on CH (Comblain et al. 2017; Eckert et al. ²⁰²¹; Schunck, Louton, and Oesser ²⁰¹⁷), have not been validated for clinical use in dogs.

Lastly, further research on the biological mechanisms of CH in dog‐specific in vitro models may give additional insight on the patient population that may benefit most from its supplementation. Ideally, these studies simulate changes in CH‐peptides that occur during gastro‐intestinal passage, digestion and absorption, for instance by using CH‐enriched serum methods (Daneault et al. 2017; Wauquier et al. ²⁰¹⁹).

5. Conclusion

This review evaluated currently available evidence on the clinical efficacy of CH in canine OA. Based on in vitro and experimental in vivo studies, CH likely acts in a chondroprotective manner by impeding cartilage degradation, improving collagen synthesis and reducing synovial inflammation. Clinical studies supported this by reporting improvements in clinically apparent lameness in dogs with OA. However, to improve the validity of treatment recommendations in canine OA cases with CH, the number of well‐organised, high‐quality clinical studies must be increased. Future clinical studies evaluating objective gait analysis combined with validated subjective outcomes such as the LOAD and CBPI in a well‐described large patient population will give further insights in CH's clinical efficacy. Evaluation of biomarkers could give additional insights in CH's metabolic effects and specifying OA patient populations that benefit from its supplementation. Future studies should also explore the bioavailability of their CH formulation in dogs to establish dosing regimens or specify the rationale of their chosen dosing strategy. This will add high‐quality evidence on CH and its actions during OA and is necessary for informed recommendation of its use in dogs with OA.

Author Contributions

Niels R. Blees: conceptualisation, original manuscript draft. Ronald Jan Corbee: conceptualisation. Marianna A. Tryfonidou: conceptualisation. Michelle Teunissen: conceptualisation. All authors contributed to the manuscript review, editing and manuscript revisions.

Ethics Statement

The authors confirm that the ethical policies of the journal, as noted on the journal's author guidelines page, have been adhered to. No ethical approval was required as this is a review article with no original research data.

Conflicts of Interest

N.B. has received research funding from Darling Ingredients. J.P. is an employee of Rousselot. The other authors declare no conflicts of interest.

Data Availability Statement

The authors have nothing to report.