Chronic pain control in the dog--when NSAIDs don't work (Proceedings)


The Cartesian model of pain imagines the somatosensory system as a simple, predictable, proportional line of communication directly linking, without altering, pain signals as they are transmitted from the periphery to the brain for conscious perception.

The Cartesian model of pain imagines the somatosensory system as a simple, predictable, proportional line of communication directly linking, without altering, pain signals as they are transmitted from the periphery to the brain for conscious perception. On the contrary, pain pathways are highly labile and modulation in the periphery, the spinal cord, brainstem, and higher centers is the rule. Peripheral nociception is altered by changes in threshold by inflammatory mediators and neuropeptides released from damaged tissue, which stimulate nociceptors, increase their sensitivity, recruit silent nociceptors, and result in up-regulation of receptors. The direct result is peripheral sensitization.

Changes in excitability of neurons and glial cells in the spinal cord result in central sensitization, a process whereby nerve fibers are “trained” to respond more vigorously. This occurs due to bombardment of the dorsal horn with high intensity, high frequency stimuli from the periphery. The result is presynaptic membrane modulation and increased release of excitatory neurotransmitters (e.g. glutamate). Neurotransmitters act on excitatory and inhibitory receptors pre and post synaptically. Excitatory NMDA-receptors, normally silent, are activated by increased glutamate. The consequence is up-regulation of receptors resulting in increased membrane sensitivity, so called spinal facilitation of pain or wind-up.

In addition to subcellular modification neuronal change occurs. Wide dynamic range (WDR) neurons are associated with wind-up. WDR neurons receive information from somatic and visceral structures, noxious and non-noxious stimuli. As a result of high frequency stimulation, previously non-noxious stimuli may cause pain, individual peripheral sensory neuron receptive fields expand, and stimulation of one tissue can be perceived as originating in a distant tissue (i.e. referred pain).

Additionally, tactile, non-noxious impulses reaching the WDR can inhibit transmission of noxious impulse (the gate control theory). Peripheral nerve injury results in loss of normal inhibitory mechanisms due to selective pathologic loss of inhibitory interneurons in the dorsal horn; this so called dis-inhibition can lead to neuropathic pain states. Finally, recent data implicates the role of immune and glial cells in the development of pain disorders particularly in the development of opioid tolerance and opioid-induced hyperalgesia.

Symptom Control vs. mechanism specific pain management

A critical concept in this complex physiology is that pain is not homogenous; there are several types of pain each with distinct neurophysiologic mechanisms: nociceptive (normal or protective pain), inflammatory (non-neuronal tissue damage), neuropathic (damaged peripheral nerves), and functional (abnormal central processing). The last three comprise the bulk of chronic pain. Note not all chronic pain is inflammatory. As our understanding of pain types and mechanisms grows, we move from symptom control to targeted, mechanism-specific pain management, the benefits of which are improved efficacy and reduced adverse effects. The key is diagnosis.

Diagnosis leads us away from empiric analgesia to a rational approach to pain management. For example, NSAIDs of the COX 2 inhibitor type are reasonably employed IF the pain syndrome is inflammatory. However, these products are unlikely to be effective in: functional pain in which inflammation never was or no longer is the primary mechanism; late stage central sensitization in which dis-inhibition is now the primary pain mechanism; or neuropathic pain in which inflammatory pain is only one component of the pain process. More than one pain type can coexist and thus multimodal approaches to management are justified. However, diagnosis coupled with reasoned choices among therapeutic modalities is a prudent and ethical endeavor.

What to do when NSAIDs don't work

The potential reasons why NSAIDs are unable to provide the desired therapeutic effect include: 1) not primarily an inflammatory syndrome, 2) syndrome has gone beyond simple inflammatory, is more complex involving peripheral and central sensitization; 3) More than one pain generator exists of different pain types, and 4) patient cannot tolerate NSAIDs: renal, gastrointestinal or hepatic pathology, concurrent steroid use

Diagnosis of all pain generators, primary and secondary, is the cornerstone of therapy. Secondary pain generators are those arising as a direct consequence of compensatory response to the primary pain generator. The first step to successfully managing any chronic pain patient is a thorough, methodical baseline evaluation including: general health/wellness exam, orthopedic exam, neurologic exam, and a myofascial exam. The goals are to establish diagnosis and functionality of the patient.

Next some means of assessing, documenting, and following pain is necessary to strategically establish and manage outcomes measures (e.g. passive range of motion via goniometry, quality of life scales, gait analysis, muscle development measures). The third step is establishing reasonable goals. The fourth step is establishing a dynamic therapeutic plan in consultation with the family. Chronic pain is not static. Time and diligence are required to establish a new baseline of comfort; and, like any chronic disease, once established, vigilance is required.


Modalities for Chronic Pain Management (not all are included due to space and time constraints)

Evidence of efficacy remains an ethical and scientific obligation. In all cases more randomized, controlled trials (RCT) are needed to support use in veterinary patients. The most work has been done in horses.


NSAIDs are the first component of the World Health Organization approach to pain and primarily act by blocking prostaglandin (PG) synthesis peripherally AND in the spinal cord. They are analgesic, anti-inflammatory, and antipyretic. Limiting effects include renal, gastro-intestinal, hepatic, and at least in humans, cardiovascular.

Acetaminophen has analgesic and antipyretic effects, but no known anti-inflammatory effect. It should not be used in cats; it has been used short-term as an alternative to or as an adjunct with NSAID use in dogs. The mechanism of action is not known, but has been linked to COX-3 enzymes in dogs. The reported half-life is 0.96 hours in dogs.

NMDA receptor antagonists include ketamine and amantadine. These agents are thought to decrease central sensitization through blockade of NMDA receptors in the dorsal horn. The only oral product examined in dogs is amantadine; in a study of canine OA, amantadine improved activity up to 21 days. So-called “pain holidays” using short term (2-3 days) CRI of ketamine coupled with other analgesics can be effective in breaking a chronic pain cycle.

Tramadol is a weak mu opioid receptor agonist and inhibitor of serotonin and norepinephrine reuptake. Caution is warranted when used with tricyclic antidepressants, monoamine oxidase inhibitors and selective serotonin reuptake inhibitors. Seizures have been reported rarely in humans. It is hepatically metabolized and cleared through the kidneys; dosing should be altered in those patients with hepatic or renal insufficiency. It has been proven effective in human OA, fibromyalgia and neuropathic pain (NP).  It has not been evaluated as a single agent in veterinary analgesia. The elimination half-life is shorter in dogs and cats than in humans (~1.4hr, 3.4hr, and 5.5hr respectively); thus more frequent dosing may be appropriate.

Opioids are poorly bioavailable by the oral route in dogs and cats. Recent work has shown bioavailability of buprenorphine via the transbuccal route in cats AND dogs. Based on current work, oral administration of butorphanol, morphine, oxycodone, methadone, and codeine cannot be recommended for any pain management protocol in dogs and cats due to poor bioavailability, short and/or weak action. However, peripheral use of opioids in the peri-operative period has implications for reducing the development of chronic pain syndromes.

Gabapentin (GBP) is FDA approved in humans along with pregabalin for the treatment of NP. Controversy exists as to whether gabapentin is only effective centrally or if there is peripheral effect. A current view purports a potent antinociceptive effect in the presence of central sensitization and potency is likely directly related to the degree of sensitization. There is much evidence for use in chronic NP, cancer and inflammatory pain models (primarily rats). The oral bioavailability in dogs is 80% with a serum half-life of 2.9 hours suggesting frequent dosing (> BID). Appropriate dose remains unclear since bioavailability is not linear.

Topical analgesics: Lidocaine patches are effective in human analgesia. Few reports are available regarding use in veterinary species; those that are available report low plasma concentrations. This may not indicate a failure in efficacy since local rather than systemic effect is the intended mechanism. A more accurate assessment might be local drug concentration within the target tissue. Ko et al found excellent local absorption (about 4x plasma concentration) in cats using 5% lidocaine patch.  However, no measure of efficacy was reported. Topical diclofenac liposomal cream (NSAID) has been shown to have beneficial clinical sign modifying and disease modifying effects in horses.



The science behind these products is limited. Supplements are a multi-billion dollar industry in the US. As a direct result of a 1994 law, which deregulated the supplement industry these products are largely unregulated and are not required to demonstrate premarket efficacy OR safety.  The concerns with these products include deceptive-marketing, inaccurate or misleading label claims, and contaminants.

A 2010 report by the General Accountability Office (GAO) found contaminants such as lead and pesticides; many products are manufactured outside of the US and therefore, not inspected by US officials (FDA). As of June 2010 manufacturers are required to comply with Good Manufacturing Practices (GMP). Clinicians must be aware of pharmacokinetic interactions between herbal supplements and pharmaceuticals. Resources:;  

Glucosamine and chondroitin have been shown to produce symptomatic and structure improvement in OA treatment. Cellular effects include: inhibition of IL-1 induced COX2 and PGE2 synthesis, COX independent anti-inflammatory properties (decreased inflammatory cytokine release), inhibition of metalloproteinase synthesis, increased production of proteoglycans, and cartilage repair. Several human RCTs demonstrate improvement in clinical signs and slowed joint space narrowing.

There are mixed canine RCTs. It is not clear that positive results translate to all OTC products particularly in light of significant variability and consistency in manufacturing practices. Data suggests glucosamine and chondroitin are more efficacious in combination. Bioavailability of low molecular weight chondroitin sulfate is >200% due to bioaccumulation and therefore may be preferred.

Polysulfated glycosaminoglycan (PSGAG). These products are primarily chondroitin sulfate as a semisynthetic preparation from bovine trachea. Reported effects are anti-inflammatory and chondroprotective. There is evidence in horses and humans for reduction in severity of OA clinical signs. SQ routes are equally bioavailable compared to IM.

Hyaluronic acid is a component of cartilage and synovial fluid. OA affected joints are deficient in HA, resulting in decreased anti-inflammatory effects, joint lubrication, and load absorption of synovial fluid. Effects of HA include decreased leukocyte migration, PGE2 release and synthesis, IL-1, chondrocyte apoptosis; increased release of tissue inhibitor of matrix metalloproteinases; and suppression of synthesis of inflammatory cytokines. Molecular weight of HA may be important. Currently, HA in the range of 0.5-2.0 x106 d is recommended. Intra-articular HA is commonly reported in the human literature for treatment of knee and hip OA as viscosupplementation.

There is some data to support use of HA with corticosteroid intra-articularly based on a rabbit model of OA. Unresolved issues for viscosupplementation requiring further study include: determining cost-effectiveness, exploring therapeutic effectiveness and safety, clearly defining the relationship between molecular weight and the clinical effectiveness, and establishing optimal dosing regiments. So far no data exists on the absorption, bioavailability, distribution or efficacy of oral HA products. Only one study is reported documenting efficacy of intravenous HA in the horse.

Omega 3 fatty acids (n3FA). Arachidonic acid is an n6FA. When incorporated into cell membranes it is subsequently metabolized into prostaglandins, leukotrienes and thromboxanes of the 2 and 4 series. Substituting n3 FAs results in metabolism to the less inflammatory eicosanoids of the 3 and 5 series. There are no published RCTs in canine OA. There is one unpublished study in which high and low n6:n3 diets were compared over 21 months in a canine model OA. A 2001 abstract reporting a randomized study of dogs with naturally occurring OA fed a commercial diet with a low n6:n3 ratio found improvement for those on the study diet. Recent data in cats is encouraging.

Microlactin. Special milk protein concentrate from the milk of hyperimmunized cows exerts anti-inflammatory properties by suppression of neutrophil migration. One randomized controlled study in dogs with OA reported improvement in clinical signs and owner assessment.


Bosweilia resin reportedly has anti-inflammatory properties via COX2 inhibition. One open multicenter study in dogs showed some improvement in clinical signs. Human RCTs have shown improvement in pain as well.

Turmeric products have a number of active ingredients. There is some evidence these agents are anti-inflammatory (inhibit COX2 and PGE2). One RCT in canine OA found no improvement in ground reaction force, but found some improvement in subjective outcomes measures.                                                             

Avocado/Soy unsaponifiables. In vitro data using cartilage cultures has shown inhibition of IL-1 and stimulation of collagen synthesis. Clinical studies in humans have shown benefits in OA. In a recent equine model of OA no improvement in clinical signs was found however, there was a disease modifying effect.

Interventional (Joint, spinal and facet injections)

Intra-articular injections have been used for decades in the management of OA. There is mixed support for the use of corticosteroid (CS), HA, PSGAG, etc. A recent study in horses demonstrated a disease modifying effect of intra-articular HA and PSGAG with no improvement in clinical signs. Intra-articular CS is considered in refractory cases of OA.  One study showed potentially beneficial prophylactic effect in a canine model of OA. There is concern over the negative effects of corticosteroids on chondrocyte metabolism; at high concentrations CS can inhibit proteoglycan synthesis and alter structural organization of collagen. However, the level of CS required for adverse effects is greater than the dose required to inhibit synthesis of inflammatory mediators.

Autologous stem cell therapy has been used to treat OA and tendinopathies. There is only one non-controlled clinical trial reported in spontaneous canine OA. Anecdotal clinical results are mixed. The majority of encouraging basic science and clinical work has been done in horses. This is an expensive albeit exciting modality. More clinical studies are required. Basic science is focusing on guided differentiation; this may result in improved clinical outcomes.

Autologous conditioned serum also called platelet rich plasma has been studied in an equine model of OA with good results. Four separate intra-articular injections were administered over a 35-day period. There was significant clinical and histologic improvement; there was an increase in IL-1 receptor antagonist in synovial fluid in treated horses compared to controls.

Use of the capsacin analogue, resinferatoxin, as a salvage procedure in the treatment of osteosarcoma-related pain has been investigated in the past decade with encouraging results. Intrathecal injection results in permanent destruction of C-fibers in the dorsal horn thus elimination of pain sensation in the region. All other sensory perceptions and motor function are preserved. This is a painful procedure and requires deep general anesthesia.

Imaging-guided spinal nerve and facet injections (triamcinolone) are future options in the treatment of chronic cervical and back pain in veterinary species. These techniques are commonplace in human medicine. The translation of these techniques to veterinary medicine is inevitable and has tremendous potential to improve quality of life.

Physical medicine

This is an under-recognized and valuable modality in comprehensive pain management. This category includes weight control, conditioning, and professionally directed rehabilitation. Rehabilitation techniques include: traction, massage, joint mobilization, cryo and thermotherapy, and therapeutic exercises, among others. The value of “hands on work” cannot by underestimated in the diagnosis, treatment, and reassessment of pain syndromes. There is a tremendous body of supportive literature in human medicine.



Acupuncture works by modulating activity of the peripheral, central, and autonomic nervous systems. There is much anecdotal evidence for acupuncture's clinical effects. A review in 2006 concluded that there is no compelling evidence to recommend or reject acupuncture in domestic animals; this review further noted that some encouraging data exists warranting further investigation. A recent study showed acupuncture resulted in shorter time for return to ambulation and return of deep pain in dogs with thoracolumbar intervertebral disc disease; a second study showed acupuncture induced analgesia in a thermal threshold pain model.

LASER (Light Amplification by Stimulated Emission of Radiation). Caution: This modality is being heavily marketed. There is a growing body of basic science and clinical evidence for the use of low-level laser therapy (LLLT) in the near infrared spectrum.

Studies show LASER has positive effects on pain control by altering afferent nerve conduction, neurochemistry of the central and peripheral nervous system, pain threshold, and inducing endorphin release. LASER treatment has a positive impact on wounds and bone healing by increasing myofibroblast activity; collagen synthesis; cartilage synthesis; tenocyte and myocyte activity; release of growth factors from macrophages; ATP levels by altering oxygen utilization in mitochondria; blood flow through angiogenesis and vasodilation; and axonal sprouting both centrally and peripherally.

Additionally, it reduces microorganism population by increasing lymphocyte numbers and activity. Although mixed results have been found in clinical trials and meta-analyses, this may be attributed in part to a lack of uniformity in study design and thus, difficulty in making direct comparisons. Animal models and clinical trials in humans have revealed positive results in the management of tendinopathy, ligament injury, osteoarthritis, fractures, spinal cord injury, and refractive wounds.

The currently recommended dose for analgesia is in the range of 8-10J/sec (Watts/sec/cm2) and is thought to be cumulative. Multiple exposures at doses in the 16J/cm2 range result in inhibitory and potentially damaging effects. Importantly, the difference between Class 3b (<500mW) and 4 LASERS (>500mW) is power. This only affects TIME to deliver the dose, not DEPTH of penetration. Though more efficient in terms of treatment time, Class 4 LASERs carry a higher risk of thermal damage and must be used with care.

TENS (transcutaneous electrical neuromuscular stimulation) applies electrical current at frequencies demonstrated to facilitate pain relief. The mechanism is thought it involve the “Gate Control Theory” and endorphin release.  Acute pain is treated at a pulse rate between 60-200Hz and a pulse width of less than 60msec. Effects are reported to be fast acting, but relatively short lived. Chronic pain has been treated with lower frequencies (2-4 Hz) and a pulse width of 200msec with effect lasting hours to days. Handheld units are used by many human patients at home.

PEMF (pulse electromagnetic field therapy). Data only supports the use of PULSED rather than static magnetic field. A 2009 systematic review of RCTs found improved clinical and functional scores in the treatment of human knee OA using PEMF. The theoretical mechanism of action is based on the observation that resting membrane potentials are altered in diseased tissue; magnetic fields, by virtue of their ability to impact ion exchange across semipermeable membranes, positively influence cartilage synthesis, cell proliferation, glycosaminoglycan synthesis, oxygen utilization, nerve conduction and angiogenesis. Therapeutic effect is reported to last 6-8 hours.

One product is FDA approved for adjunctive palliative treatment of post-operative pain and edema in superficial soft tissue. However, this is an FDA 510-K premarket notification (PMN) only; there is no requirement for new supportive data.      

Extra Corporeal Shock Wave Therapy (ECSWT) has been examined in an equine model of OA. This modality requires general anesthesia. Lameness scores improved however there were no disease modifying effects.



Must be considered, bearing in mind that surgery may not be possible/appropriate for valid reasons.


Veterinary Orthotics and Prosthetics (VOP) is an emerging field. The biomechanical concepts and technology developed in humans over the last century are being adapted for use in veterinary species. Decreased function of a limb and lack of a limb segment whether traumatic or congenital can result in significant secondary pain due to changes in posture and gait. Chronic conditions such as muscle tension and triggerpoints, ligament damage, OA, and joint collapse are commonplace and, until recently, accepted as inevitable in these patients. In the coming years it is likely that VOP will become integral in the practice of modern veterinary pain medicine as it is now in human medicine.

Related Videos
© 2023 MJH Life Sciences

All rights reserved.