Introduction

Osteoarthritis is a progressive disease of synovial joints in which pathological changes including articular cartilage degeneration, sclerosis and thickening of subchondral bone, fibrosis of periarticular tissue, and osteophyte formation occur.¹ It is very common in dogs, with 6.6% to as many as 20% of dogs reported to suffer from this disease.^1,2

Osteoarthritis in dogs can cause lameness and stiffness, which both suggest the presence of pain.^3,4 These are often accompanied by behavioral changes including reluctance to exercise, unwillingness to jump or climb, changes in sleep patterns, restlessness, anxiety, and aggression.^4,5 These debilitating clinical signs impair quality of life (QOL) and in some cases contribute to the decision to euthanize patients.^4,6,7

Canine osteoarthritis is commonly presumptively diagnosed in first-opinion practice on the basis of the history and physical examination findings (including gait abnormalities, reduced range of joint motion, soft tissue swelling, and muscle atrophy).^3,4 Radiographs of affected joints may be used to exclude other causes of the signs and confirm the presence of typical osseous changes^3,7; however, radiography is not always performed in first-opinion practice because of the associated costs, the dog’s age, and other illnesses.^4,8

As canine osteoarthritis is incurable, it requires lifelong management.^3,9 Treatment aims are first to control the associated pain, which should help normalize behavior and maintain or improve joint mobility and improve QOL, and second to slow disease progression.^6,7 A multifaceted approach is recommended, including pharmacologic and nonpharmacologic analgesia, weight management, modified exercise regimens, physical rehabilitation therapies (eg, hydrotherapy, physiotherapy), and in some cases surgery.^3,6,7,10

NSAIDs are the most common treatment for canine osteoarthritis.^4,7,10 They are highly effective analgesics and relatively safe.¹⁰ Despite this, not all patients can tolerate these drugs,^3,7,11 and even in dogs that do, evidence of ongoing pain is often seen.^12,13 For this reason, additional methods of analgesia are recommended, including other drugs (eg, paracetamol, gabapentin, amantadine, and tramadol), nutraceuticals, acupuncture, and low-level laser therapy.^6,7

Static (permanent) magnets (SMs) are widely available to treat chronic pain in people. Several double-blinded, placebo-controlled studies of people with chronic musculoskeletal pain have shown greater reductions in pain scores in patients treated with SMs compared to those treated with placebos,^14–17 although other studies have shown no difference between the groups.^18–20 A controlled study exposing rats with experimentally induced arthritis to SMs showed increased locomotive activity (suggesting reduced pain) in the treated group.²¹ Collars containing SMs have been developed for dogs and are widely available; however, there are no published efficacy studies.

Studies investigating the efficacy of analgesics for osteoarthritis necessitate tools for the measurement of pain and lameness. Both objective measures of lameness, such as force plate analysis and subjective tools to assess pain and lameness have been developed for use in dogs.^22–27 The Canine Brief Pain Inventory (CBPI) was developed to enable owners to quantitatively score their perceptions of the severity and impact of chronic pain on their dogs with osteoarthritis.²² It is composed of 11 questions in 3 parts: a pain severity score (PSS; questions 1 to 4), a pain interference score (PIS; a measure of interference of pain with function; questions 5 to 10) and a score of overall QOL (question 11; Supplementary Table S1).²² The CBPI was found to be reliable and valid in assessing the severity and impact of chronic pain in canine osteoarthritis.²² It was able to detect improvements associated with treatment, thus supporting its use as an outcome assessment tool.^12,28 Since then, it has been widely used to evaluate the efficacy of canine osteoarthritis treatments.^13,29–34

The aim of this study was to investigate changes in QOL, severity of pain, and interference of pain with function, as determined by the CBPI, in dogs with clinically suspected osteoarthritis when wearing a proprietary static magnetic collar.

Methods

This prospective, blinded crossover study was undertaken in a first-opinion clinic in Malta. Prior ethical approval was granted by an ethics committee. On the basis of a review by Mills et al³⁵ of 116 randomized crossover trials on Medline, which found a median sample size of 15, a sample size of 20 dogs was planned.

To identify dogs eligible for inclusion in the study, the clinic’s electronic patient record (EPR) system was searched for dogs’ records containing the term “arthritis” that visited the clinic between January 1, 2019, and December 31, 2019. The EPR was then screened to identify deceased dogs, those under 5 years old on January 13, 2020, those weighing < 10 kg or > 40 kg or with no recorded weight, and those no longer in the country for exclusion. A recorded diagnosis of osteoarthritis was confirmed, and dogs with other orthopedic diseases, neurological diseases, or other painful conditions were excluded. Dogs prescribed analgesics or anti-inflammatories other than NSAIDs, those that had NSAID dose changes in the preceding 28 days, those that had nutraceutical or joint diet changes in the preceding 90 days, and those treated with physiotherapy, hydrotherapy, or acupuncture were excluded.

The list of dogs for potential inclusion was entered into Excel (Microsoft Corp), randomly reordered, and then assigned consecutive study numbers. From this new list, owners were contacted sequentially by telephone, the study was explained and inclusion criteria checked, and, if suitable, they were invited to participate and enrolled until the required sample size was reached.

Clients of enrolled dogs attended the clinic, and full written informed consent was obtained. At baseline, the client completed a questionnaire on patient signalment, all current medications, nutraceuticals, and diet. They were also asked to complete the CBPI pain assessment tool (Supplementary Table S1).³⁶ A veterinary consultation was performed during which the dogs had a full clinical examination and an orthopedic examination including joint palpation, range of motion, pain and crepitus assessments, and gait assessment. A questionnaire was completed by the veterinarian to record the weight, body condition score (using the World Small Animal Veterinary Association Nutritional Assessment Guidelines 9-point scale),³⁷ QOL (using a 5-point Likert scale from poor to excellent), all joints affected by osteoarthritis, and the worst affected joint. Planned changes to medications or management were also recorded. The date osteoarthritis was diagnosed, date of joint radiography (if performed), and presence or absence of osteoarthritis on the radiographs were collected from the EPR system. Cases found not to meet the inclusion criteria above or in which any of the joints radiographed had evidence of diseases other than osteoarthritis were excluded.

The AB/BA crossover study design is shown (Figure 1). The dogs wore the first collar for 4 weeks, followed by a 2-week washout period with no collar and then a 4-week period wearing the other collar. The collars were simple black nylon collars (Figure 2) designed to fit dogs between 10 and 40 kg. The only difference between the magnetic collars and the placebo collars was that the magnetic collars contained a 22 X 11 X 1-mm flat, unipolar permanent magnet made from a neodymium, iron, and boron alloy with a surface field strength of 200 mT, whereas the placebo contained a plastic rectangular spacer. Both the magnet and placebo collars contained a nonmagnetic steel backing plate. The magnetic collars are commercially available (Magnetic Dog Collar; Pulse Magnetics). The order in which the collars were tested was randomized using Excel (Microsoft Corp). Both the owner and veterinarian were blinded as to which collar the dog was wearing. Collars were fitted by one of the authors, and clients were instructed not to remove the collars during the study. Clients were given copies of the CBPI to be completed weekly and a diary to be completed during both 4-week arms detailing any changes to medications (including NSAIDs), nutraceuticals or joint diets, any problems with the collar (including removal and duration of removal), and any deviation from the dogs’ normal exercise regimen. Clients were telephoned by one of the authors at the end of weeks 1 to 3 and 7 to 9 to ensure they were not having any problems with the collar and were completing the CBPIs and diary. Dogs were returned to the clinic at the end of the first arm, beginning of the second arm, and end of the second arm to have collars removed or placed and for clients to return completed paperwork.

View Full Size

Crossover (AB/BA) study design and protocol. Sixteen dogs were randomly fitted with a placebo (7 dogs) or magnet (9 dogs) collar at the start of the study (arm 1). Following this, the collars were removed for a 2-week washout period and the dogs then crossed over to arm 2, in which they were fitted with the other collar (placebo or magnet) for a further 4 weeks. Owners completed the Canine Brief Pain Inventory (CBPI) pain assessment tool prior to collar placement and each week of the study thereafter. Time points at which the owner completed the CBPI pain assessment tool are denoted by asterisks.

Citation: Journal of the American Veterinary Medical Association 262, 4; 10.2460/javma.23.10.0555

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Photograph showing a dog wearing a magnetic collar.

Citation: Journal of the American Veterinary Medical Association 262, 4; 10.2460/javma.23.10.0555

• Download Figure
• Download figure as PowerPoint slide

Prior to data analysis, client diaries were checked and dogs were excluded that had collars removed for more than half a day cumulatively during either 4-week arm or that had any reported changes in medications (including NSAIDs), joint supplements, joint diets, or exercise regimens.

Results

Seventy-eight eligible dogs were identified from the clinic software. Five of these were excluded due to the presence of neurological disease, other orthopedic disease, treatment with disallowed therapies, or recent NSAID dose changes, leaving 73 dogs with clinically suspected osteoarthritis for randomization. Fifty-four owners were invited to participate, and 35 declined; thus, 19 dogs were enrolled. Three (15.8%) dogs were later withdrawn as additional analgesia was prescribed during the study (n = 2, both at the end of the placebo arm) or the collar was removed due to interference of the collar with domestic electrical devices (1, during the first week of the magnet arm).

Of the 16 dogs that completed the study, 10 (62.5%) were purebreds (2 each of Boxer and English Springer Spaniel and 1 each of Bull Terrier, Cavalier King Charles Spaniel, Dachshund, Galgo Español, Labrador Retriever, and Siberian Husky) and 6 were crossbreeds. There were 8 males (2 entire) and 8 neutered females. The median age was 9.4 years (range, 5.4 to 14.2 years). Their median weight was 21.4 kg (range, 10.3 to 33.0 kg) and median body condition score was 5/9 (range, 4 to 7). Concurrent diseases were present in 5 (31.3%) dogs, including congestive heart failure (n = 2), leishmaniasis (1), skin allergies (1), and urinary incontinence (1). These were all considered nonpainful conditions by the attending veterinary surgeon, and all were on stable treatment throughout the study.

The median time since diagnosis of osteoarthritis was 1.3 years (range, 0.3 to 2.6). Nine (56.3%) dogs had joint radiography, and all 9 had changes consistent with osteoarthritis and no evidence of other orthopedic diseases. At the initial veterinary examination, 12 (75.0%) dogs had more than 1 site of clinically suspected osteoarthritis identified. The worst affected joints/sites were hips (n = 5), spine (4), elbows (4), stifles (2), and carpi (1).

Seven (43.8%) dogs were on a stable regimen of NSAIDs for a median of 82 days (range, 51 to 393) prior to enrollment. These were cimicoxib (Cimalgex; n = 3), firocoxib (Previcox; 2), and robenacoxib (Onsior; 2). Twelve (75%) dogs were on a stable regimen of joint supplements containing glucosamine hydrochloride and chondroitin sulfate, 9 were receiving Stride, 1 dog jGARD, 1 dog Mobilize DS, and 1 dog YuMOVE. One dog was receiving a proprietary mobility diet (J/D prescription diet; Hill’s Pet Nutrition). Six (37.5%) dogs were on both NSAIDs and joint supplements, 5 (31.3%) dogs were on joint supplements alone, and 1 (6.25%) dog was on a joint supplement and a proprietary mobility diet.

The median QOL scores recorded by the owners and veterinarians at baseline were not significantly different (3.5 [range, 2.0 to 5.0] and 3.5 [range, 2.0 to 5.0], respectively; P = .36). The median PSS recorded by the owners at baseline was 2.1 (range, 0.0 to 7.0) and median PIS was 3.4 (range, 0.7 to 5.5).

Effect of collars on QOL

For all dogs, the median QOL score was 4.0 (range, 2.0 to 5.0) at the start of the placebo arm and 4.0 (range, 2.0 to 5.0) at the end of the placebo arm (Figure 3). The estimated median change in QOL score during the placebo arm was 0.0 (range, –2.0 to 1.0) and was not significantly different from 0 (P = .61).

View Full Size

Dot plots of mean quality of life (A), pain severity (B), and pain interference (C) scores at the start and end of the placebo arms (gray circles) and magnet arms (gray triangles) of the crossover study described in Figure 1. Black diamonds denote the median of each data set.

Citation: Journal of the American Veterinary Medical Association 262, 4; 10.2460/javma.23.10.0555

• Download Figure
• Download figure as PowerPoint slide

The median QOL score was 3.0 (range, 2.0 to 5.0) at the start of the magnet arm and 4.0 (range, 3.0 to 5.0) at the end of the magnet arm (Figure 3). The estimated median change in QOL during the magnet arm was 0.0 (range, –2.0 to 1.0) and was not significantly different from 0 (P = .27).

The median difference between the change during the magnet arm and change during the placebo arm was 0.0 (range, –2.0 to 3.0) and was not significantly different from 0 (P = .82). The median QOL in the placebo and magnet arms at each time point over the 4-week periods are shown (Figure 4).

View Full Size

Line graphs of the weekly median scores for quality of life (A), pain severity (B), and pain interference (C) for all 16 dogs in both the placebo and magnet arms of the crossover study described in Figure 1.

Citation: Journal of the American Veterinary Medical Association 262, 4; 10.2460/javma.23.10.0555

• Download Figure
• Download figure as PowerPoint slide

Discussion

This blinded, randomized crossover study has shown for the first time that collars containing an SM had a positive effect, reducing pain severity and interference of pain with function scores in dogs with clinically suspected osteoarthritis. Significantly lower PSS and PIS were observed by owners after 4 weeks of dogs wearing the magnetic collars. This was not observed when dogs wore placebo collars. Additionally, a significantly greater change in both PSS and PIS was found during the magnet arm compared to the placebo arm.

Although this was the first time an effect of SMs on reducing pain scores has been found in dogs, they have been shown to reduce pain in several double-blinded, placebo-controlled studies of people with chronic musculoskeletal pain.^14–17,38 Analgesic effects of SMs have also been reported in randomized, double-blinded, placebo-controlled studies of other painful conditions in people including diabetic neuropathy, temporomandibular disorders, and chronic pelvic pain.^39–41 There has been less work in animals; however, analgesic effects of SMs have been reported experimentally in mice with induced visceral and orofacial pain.^42–44 Additionally, exposure to SMs increased rats’ locomotor activity in a chronic arthritis model when compared to nonexposed rats, a finding assumed to be due to reduced pain.²¹

The mechanisms by which SMs work are poorly understood. Proposals include neurological, vascular, and anti-inflammatory effects. Suggested neurological analgesic mechanisms include changes to nociceptors or primary sensory neurons, alterations in C-fiber conduction velocity, alterations in opioid receptors, and changes in enkephalin or endorphin release.^16,41–44 Alterations in blood flow or viscosity have also been proposed^21,45; however, Hinman et al¹⁵ suggested that the rapid onset of the observed analgesic effect makes these vascular mechanisms less likely to be the means by which magnets work. There is also some evidence to suggest SMs reduce concentrations of serum inflammatory proteins and other inflammatory markers.^21,46

Vallbona et al¹⁶ suggested that SMs should be placed over the site of pain for efficacy; however, this was not the case in the present study, nor in a study of people with hip and knee osteoarthritis treated successfully with magnetic bracelets.¹⁴ In a study of post-polio patients, some people reported benefits in areas not directly underneath the applied magnets.⁴⁷

The observed reduction in PSS in this study was estimated to be 38% of baseline, which lies between the 18% reduction in pain score in people with hip and knee osteoarthritis treated with magnetic bracelets for 4 weeks¹⁴ and the 62% improvement in perceived pain (measured on a visual analog scale) in people with chronic knee pain after 2 weeks of magnet therapy.¹⁵ The observed reduction in PIS of approximately 53% from baseline was very similar to the 52% reduction in physical function score (measured using the Western Ontario and McMaster Universities Osteoarthritis Index) seen in the Hinman et al¹⁵ study. The magnitude of improvements is also similar to those reported with NSAID use in dogs. The reductions in PSS and PIS were similar to those reported by Brown et al^12,28 when carprofen was administered to dogs with untreated osteoarthritis (PSS change, –1.2; PIS change, –1.1) and were greater than the PSS changes (but similar to the PIS changes) reported by Vijarnsorn et al³⁴ when firocoxib was administered for 4 weeks for chronic hind limb lameness (mean pretreatment PSS, 1.89; mean post-treatment PSS, 1.88). Comparison with other studies of osteoarthritis interventions using the CBPI score is difficult because of the methods used to report results in those studies. In people, the magnitude of the effects of magnetic bracelets on pain caused by hip and knee osteoarthritis was comparable to that seen in other studies using oral NSAIDs or topical NSAID creams.¹⁴ It is important to note that about 40% of the dogs in this study were already being administered NSAIDs and 75% were on nutraceuticals; despite this, an additional beneficial effect was seen. Similarly, an additive effect has been observed in people.¹⁴

Although this study was not designed to determine the speed of onset of any analgesic effect, median PSS and PIS started to decrease by 7 days (Figure 4). A rapid onset of analgesia has been reported in people, with efficacy within 45 minutes in 2 blinded studies^15,16 and within minutes in 2 case reports.⁴⁸ The effect on median PSS and PIS continued to increase during the 4 weeks of the magnet arm. Studies in people have similarly seen a trend of decreasing pain scores over several weeks.^14,39,41,49 Further studies are needed to investigate the speed of onset of analgesia and whether the trend of decreasing PSS and PIS would continue beyond 4 weeks.

The effect of magnetic collars on the PSS and PIS was short-lived after collar removal, with scores not significantly different to baseline after the 2-week washout period. This suggests that the magnetic collar must be worn to maintain its effect. The rapidity with which the analgesic effect wanes after discontinuation of SM therapy has not been widely reported in the human literature. In a study of post-polio pain, some people reported pain relief lasting days and even weeks after removal of the magnets.⁴⁷ Similarly, one of the patients in the case studies by Holcomb et al⁴⁸ remained pain free despite removing the magnets for increasing periods of time; however, it was not clear whether this was due to improvement in the underlying condition or a persistent effect of the magnet therapy.

Despite the significant changes in PSS and PIS by the end of the magnet arm, significant changes in QOL were not detected in either arm. This may be because QOL is challenging to score in animals, as it is highly individual and involves proxy scoring on behalf of the animal.⁵⁰ Multiple factors aside from chronic pain also contribute to QOL, including age, comorbidities, the environment, and personality.⁵⁰ In this study, we used the simple, unstructured QOL question from the CBPI in which a subjective overall score out of 5 for the preceding 7 days was allocated by the owner, and we didn’t attempt to define or explain QOL for them. By contrast, the PSS and PIS were each based on several questions, scored out of 10 using observations more easily recognized by clients. In addition, the median QOL score of 3.5 at the start of the study was reasonably high, giving less scope for improvement. It could be argued that the lack of a significant improvement in QOL scores suggests the intervention is of limited value; however, one could also argue that the PSS and PIS are indicators of QOL, and their improvement suggests improved QOL despite the lack of change in QOL score. Interestingly, other studies using the CBPI tool have either not reported the QOL results or reported them only in combination with the PSS and PIS as a total CBPI score.^{13,29–31,33,34}

The cases included in this study were dogs between 10 and 40 kg that had been diagnosed with osteoarthritis in first-opinion practice. The signalment of these cases was similar to those found in other studies of dogs with osteoarthritis.^4,12,31–33 About 40% of dogs were on NSAIDs and 75% were on joint supplements, and therefore, as would be expected, the median PSS and PIS at baseline were lower than most of those reported in untreated dogs with osteoarthritis^12,28–31 but higher than most of those on NSAIDs.^28,34 Our findings can be generalized to dogs in first-opinion practice on standard therapies for osteoarthritis.

The magnets within devices differ in strength, size, composition, and polarity. The magnets used in this study were 200 mT, a similar strength to those showing an analgesic effect in the study by Harlow et al¹⁴ and stronger than those used successfully by Vallbona et al¹⁶ and Colbert et al.³⁸ Magnet size and composition also affect the strength of the magnetic field.¹⁵ Stronger magnets appear to have better effects and are thought to have deeper tissue penetration.¹⁵ Magnet polarity and the type of pain have also been suggested to influence efficacy.^15,45 The differences in magnet characteristics (particularly magnet strength) have been suggested as a possible reason why beneficial effects have been seen in some but not all human trials.^15,45 This means that our results cannot be generalized to all magnetic devices designed for dogs, and further investigations into the effects of these other devices on pain scores are therefore required.

Not all individual dogs in this study had improvements in PSS or PIS. This may be partly because of variable disease control at the start of the study. Other patient characteristics or the distance between the collar and the affected joint(s) may also play a role. In people with post-polio syndrome, 34% did not report a greater reduction in pain score with magnet therapy than with the placebo¹⁶ and Holcolm et al⁴⁸ reported that about 20% of 2,000 patients with lower back pain experienced no benefits with magnetic devices. These findings are important and highlight the importance of ongoing monitoring of dogs treated with magnet therapy to ensure adequate analgesia is achieved.

No adverse effects were reported in this study. Similarly, most studies of people have not reported any side effects^16,40,41; however, a few subjects in 1 study¹⁵ reported mild tingling in the area close to the magnet and another reported rare dizziness and increased pain or stiffness, but these were also reported in the placebo group.¹⁴ Further, larger studies in dogs would be needed to investigate adverse effects. In this study, only 1 owner reported interference with home electrical devices when the dog was wearing the magnetic collar.

This study had several limitations including the small number of dogs enrolled and the heterogeneous population; however, the placebo-controlled, randomized crossover design allowed each dog to act as its own control reducing bias associated with imbalance in known and unknown confounding variables³⁵ and the population reflected that of a first-opinion practice. Not all dogs had a radiographic diagnosis of osteoarthritis, and some dogs may have had other musculoskeletal disorders. Many owners declined to participate, and those that did may have been more open to believe an effect would occur, although the inclusion of a placebo arm and blinding should have mitigated this. Maintaining blinding during studies using magnets is notoriously difficult,^14,39,41 and owners may have guessed which collar was which, thereby influencing the results; however, to our knowledge, none of the owners of included dogs had ascertained this. The outcome measures used were subjective and owner-reported; however, the CBPI score was designed for this purpose and client opinions of their dogs’ pain is clinically relevant. It is known that pain due to osteoarthritis is not constant and flare-ups can occur, which could have affected the results; however, both the placebo and magnetic arms should have been equally affected.

This study was the first to attempt to investigate the analgesic effect of SMs in dogs. The results showed that, in dogs with clinically suspected osteoarthritis on standard treatments, owner-reported pain severity and the interference of pain on function scores were reduced when wearing a collar containing magnets for 4 weeks but not the placebo collar. This suggests that SMs could play a role in these patients as an adjunct to traditional medical therapies such as NSAIDs and nutraceuticals. The findings warrant further study in a larger group of dogs over a longer period to confirm the reduction in pain scores and determine its duration.

Supplementary Materials

Supplementary materials are posted online at the journal website: avmajournals.avma.org

Disclosures

The authors have nothing to disclose. No AI-assisted technologies were used in the generation of this manuscript.

Funding

The authors have nothing to disclose.