Review of NSAIDS: COX selectivity and systemic effects (Proceedings)

Non-steroidal anti-inflammatory drugs (NSAIDs) are widely used to control acute and chronic pain in veterinary patients. The presence and activity of two isoforms of the cyclooxygenase (COX) enzyme, a constitutive COX-1 and an inducible COX-2, have been investigated intensely since the early 1990s. COX-1 is present under basal conditions in many cells including platelets, mucosal cells in the GI tract, endothelial cells, and the renal medullary collecting ducts and interstitium. The COX-2 isoform is present in low concentrations under basal conditions in monocytes, macrophages, smooth muscle cells, fibroblasts, and chondrocytes. Despite widespread use, ongoing research to introduce new drugs, and efforts to improve the safety and efficacy of existing drugs, side effects such as gastrointestinal (GI) irritation, renal and hepatic toxicity, interference with hemostasis, and reproductive problems persist. Remarkably, the true incidence of NSAID-associated side-effects in companion animals is still unknown. Many of the adverse side-effects of the non-selective NSAIDs (drugs which inhibit both COX-1 and COX-2 isoforms) have been attributed to inhibition of the constitutive COX-1. As a result, over the past decade many COX-1 sparing drugs have been developed and introduced into clinical use. NSAIDs produce analgesia and toxic side-effects primarily by inhibiting a key enzyme in the arachidonic acid (AA) pathway, ceasing the production of prostaglandins, most notably prostaglandin E2 (PGE2). The AA pathway is initiated by damage to cell membranes, resulting in the release of prostanoids, which signal inflammation and pain, as well as perform physiologic functions on target tissues. The anti-inflammatory effects of NSAIDs are not solely limited to their function as cyclooxygenase inhibitors. A number of studies have shown that NSAIDs are capable of inhibiting proinflammatory transcription factors, which are independent of cyclooxygenase and prostaglandin activity.

With the testing and marketing of several efficacious NSAIDs over the last decade, veterinarians are more and more comfortable using these drugs. New discoveries about inflammatory mediators and their interactions in the inflammatory cascade, as well as new data on the biochemical mediators associated with osteoarthritis, have led to increased use of NSAIDs. It is now recognized that there are at least three different COX enzymes (COX-1, COX-2, and a COX-1 variant called COX-3) which are active in the metabolism of arachidonic acid, and certain NSAIDs may have selectivity in their actions against these isoenzymes. Furthermore, there is data to support that the ratio of COX and lipoxygenase (LOX) inhibition may be important in the improved gastrointestinal safety seen in dogs, and a dual COX/LOX inhibitor has been introduced to the market. However, these drugs are not a panacea with 100% efficacy and 0.0% adverse effects. In this lecture we will focus on everything from the basic pharmacology of these drugs to some of the tougher, unanswerable questions about potential dosing variations that might be considered due to hepatic or renal insufficiency. We will also touch on multimodal therapeutic ideas with NSAIDs as a component of the therapeutic protocol.

Pharmacology (More than you really need to know):

Although their sites of action are still unclear, NSAIDs appear to produce analgesia at both central and peripheral levels. They inhibit the peripheral COX-2 enzyme to block the formation of prostaglandins such as PGE2 and PGI2 which function to dilate arterioles and sensitize peripheral nociceptor terminals to the actions of mediators such as histamine and bradykinin to produce localized pain and hypersensitivity. The PGE2 produced by COX-2 plays a pivotal role in sustaining acute pain sensation by increasing the level of cyclic AMP within nociceptors, thus decreasing their threshold of activation. In addition to peripheral pain perception, COX-2 mediated prostaglandins such as PGE2 are involved in spinal nociception and central analgesia. COX-2 is expressed in the brain and spinal cord, and is upregulated in response to traumatic injury and peripheral inflammation. The actions of COX-2 are thought to contribute to neuronal plasticity and central sensitization. The anti-inflammatory effects of NSAIDs are not solely limited to their function as cyclooxygenase inhibitors. A number of studies have shown that NSAIDs are capable of inhibiting proinflammatory transcription factors, which are independent of cyclooxygenase and prostaglandin activity. Aspirin, sodium salicylate, rofecoxib, and ibuprofen have all been shown to mediate some of their anti-inflammatory effects via inhibition of the transcription factor, nuclear factor-kappa B (NF-KB). Briefly, NF-KB is located in the cytoplasm of the cell in an inactive complex with the inhibitory subunit, KB. Upon activation of the cell, the KB subunit is degraded and NF-KB translocates into the nucleus where it interacts with DNA KB binding sites, resulting in transcription of numerous proinflammatory genes. In addition, aspirin and sodium salicylate have also been shown to inhibit another important transcription factor in the immune response, activating protein-1 (AP-1) but again its mechanism of inhibition is unknown. This inhibition of proinflammmatory transcription factors is not a feature of all NSAIDs. NF-KB and AP-1 are key factors in the inflammatory and immune reaction, which make them important targets for modulation of the immune response. Understanding these pathways may help the efficacy of NSAID choice in treatment for certain diseases. For instance in both rheumatoid and osteoarthritis the inflammatory response not only elicits pain, but also is a stimulus for degredation of the cartilage, bone and the other tissues of the joint. Both NF-KB and AP-1 are known to up-regulate some matrix metalloproteinases, which appear to play a significant role in the degradation of cartilage and bone in arthritis. Also, these other mechanisms of NSAID action may be important in understanding the adverse effects associated with NSAID use.

More Practical Pharmacokinetics and Mechanisms of Action

NSAIDs block the rate-limiting step in this pathway at the site of COX, an enzyme which converts arachadonic acid to prostaglandin G2 (PGG2) and prostaglandin H2 (PGH2) through oxidation and reduction reactions. COX-1 is considered to be the constitutive form, as it is important for normal physiological functions. COX-2 is considered to be an inducible form of the enzyme, and is associated with inflammation and pain. COX-3 (a COX-1 variant) has only recently been described in the dog and its clinical or therapeutic significance has yet to be elucidated. Theoretically, an NSAID that selectively inhibits COX-2 without affecting COX-1 will allow analgesia without the common side effects of COX-1 inhibition, which include altered gastrointestinal and thrombocyte function.

An NSAID's ability to inhibit each COX enzyme is often reported as a ratio, which defines enzyme selectivity. There is currently little data to support direct correlation of different in vitro testing techniques to clinical applicability. Along with the development of different COX specific inhibitors, a confusing COX terminology has developed. Most of the confusion surrounds the marketing of drugs which are described as COX-2 preferential, selective, or highly-selective. These classifications are based upon a completely arbitrary system described by in vitro data. The bottom line is all NSAIDs are COX-2 inhibitors. What is different is how the newer compounds do not concomitantly inhibit COX-1. Thus it is probably better to consider these products as COX-1 sparing. A second area of confusion involves the term 'coxib.' Although the term 'coxib' is used rather loosely in the literature, often referring to any COX-1 sparing drug, true coxibs are a specific World Health Organization (WHO) designated subclass of NSAIDs. The WHO states that their 'stem classification system [for drugs] is a working system to help to name new pharmaceutical active substances, rather than a classical pharmacological classification. With the exception of lumiracoxib, all drugs with the suffix '-coxib' consist of a tricyclic ring and a sulfone or sulfonamide group; lumiracoxib consists of two cyclic rings. These chemical features give the class of drugs its high specificity, saturability, and readily reversible binding behavior. Coxibs, like all NSAIDs, inhibit COX-2, but their important feature is their sparing effect on COX-1.

In addition, extensive data exists which indicate that many of the anti-inflammatory effects of NSAIDs are unrelated to the inhibition of arachadonic acid metabolism. NSAIDs have been shown to inhibit a variety of functions of neutrophils including adhesiveness, aggregation, chemotaxis, degranulation and superoxide anion generations. Additionally there is strong evidence to support that NSAIDs act directly in the spinal cord and higher centers. The mechanisms of how the different cyclooxygenase isoenzymes are involved in the generation of pain sensation are not completely understood and new information is appearing almost every month. Pain is a complex experience involving not only the transduction of noxious stimuli from the periphery to the central nervous system (CNS) but also the processing of the stimuli by the higher centers in the CNS.

Site of analgesic action

Although sites of action are still unclear, coxibs and the other COX-1 sparing NSAIDs appear to produce analgesia at both central and peripheral levels. They inhibit the peripheral COX-2 enzyme to block the formation of prostaglandins such as PGE2 and PGI2 which function to dilate arterioles and sensitize peripheral nociceptor terminals to the actions of mediators such as histamine and bradykinin to produce localized pain and hypersensitivity. The PGE2 produced by COX-2 plays a pivotal role in sustaining acute pain sensation by increasing the level of cyclic AMP within nociceptors, thus decreasing their threshold of activation. In addition to peripheral pain perception, COX-2 mediated prostaglandins such as PGE2 are involved in spinal nociception and central analgesia. COX-2 is expressed in the brain and spinal cord, and is upregulated in response to traumatic injury and peripheral inflammation. COX-2 activated PGE2 lowers the threshold for neuronal depolarization, increasing the number of action potentials and repetitive spiking. The actions of COX-2 are thought to contribute to neuronal plasticity and central sensitization. One study found a significant increase in PGE2 levels in the cerebral spinal fluid after noxious stimulus in a rat paw; PGE2 levels returned to normal after administration of celecoxib, but not after administration of a non-selective NSAID.

Indications

NSAIDs can be used to relieve pain in a variety of clinical settings and situations. NSAIDs can be used for cases of acute pain, either traumatic or surgically induced, as well as for chronic pain such as osteoarthritis. There is also increasing evidence which indicate that NSAIDs are an effective component in the treatment of certain types of neoplastic conditions and potentially in certain central nervous system disorders. Choosing and monitoring NSAID usage is a very important consideration to the clinician. There is no right answer to this question but some common sense parameters can help guide us. First it is wise to use products with a history of extensive clinical experience and good safety profiles. Always use only one NSAID at a time and ensure adequate dosing. Review therapy frequently, and change to alternative NSAIDs if there is a poor response to therapy. Observe for potential toxicity as soon as administration is begun with increased vigilance and monitoring required for at-risk patients. If indicated, establish renal and hepatic status of the patient prior to NSAID administration. Remember, efficacy and toxicity is often individualistic and individual monitoring is mandatory.

The mainstay of NSAID use has been in the treatment of osteoarthritis (OA). In these diseases, pro-inflammatory mediators such as interleukin-1 (IL-1), prostaglandins, and nitric oxide (NO) are produced by the synovial membrane and possibly chondrocytes, which induce angiogenesis and inflammation of the synovium, as well as drive the catabolism of cartilage and erosion of subchondral bone. Cartilage destruction ultimately results from an imbalance between synthesis and breakdown of extracellular matrix molecules, notably proteoglycans, which promotes an imbalance between cartilage destruction and effective cartilage repair. COX-2 increases pro-inflammatory prostaglandins (PGE2) and IL-1β , which modulate cartilage proteoglycan degradation. In laboratory studies, however, the effects of COX-2 on cartilage differed depending on whether healthy or damaged cartilage was studied, suggesting that other factors play a more important role in initiating the damage, and that COX-2 serves a role to potentiate such damage. Recent research supports the importance of NO as a co-factor in OA and rheumatoid arthritis, and future management of these diseases will probably involve COX-2 inhibitors as well as NO synthase modulators.

Effects on the Gastrointestinal Tract:

The primary site of NSAID toxicity is the gastrointestinal tract in companion animals. Effects such as gastro-duodenal erosion, ulceration, perforation, and stricture have been associated with NSAID usage in several species. These effects have been linked to inhibition of the cytoprotective effects of COX-1 induced synthesis of prostaglandins on the mucosal border, local irritation by the drug, and direct inhibition of platelet aggregation by decreasing production of TXA2. When COX-1 sparing drugs are used, however, the incidence of these side-effects decreases (in humans) but is not eliminated, and is still significantly greater than with placebo in several controlled studies. Interestingly, COX-2 appears to be protective on the mucosal border, and important in maintaining mucosal health. Although present at negligible levels in healthy tissue, COX-2 expression is increased on the edges of gastric ulcers where it indirectly accelerates ulcer healing by increasing angiogenesis through the inhibition of cellular kinase activity and upregulation of PGE2 and VEGF in the gastric mucosa. There is nothing known about the physiology and function of the individual COX isoenzymes in the feline. This could be very important information with the availability of NSAIDs that are more or less selective for individual COX isoenzymes. There is also very little information on possible risk factors for GI ulceration in the cat, and therefore, situations in which it might be dangerous to use NSAIDs are not elucidated.

Effects on the Liver:

Hepatotoxicity is a rare but potentially serious adverse consequence of NSAID use. Hepatotoxicosis caused by NSAIDs is generally considered to be idiosyncratic. Administration of carprofen has been documented with an idiosyncratic cytotoxic hepatocellular reaction. Anorexia, vomiting and icterus along with increased hepatic enzyme levels have been seen. The onset of signs was seen by 21 days in the vast majority of dogs. Most dogs recovered with cessation of treatment and supportive care. However other NSAIDs can induce the same problems and liver function must be monitored with use of all NSAIDs.

Effects on the Kidney:

In general, prostanoids regulate glomerular filtration, renin release, and sodium excretion in the kidney. Inhibition of COX-1 had previously been assumed to be the cause of adverse side-effects such as salt retention, decreased GFR, hypertension, and edema associated with NSAID use. Recent studies have however shown that COX-2 is constitutively expressed in the macula densa, thick ascending loop of Henle, and interstitial cells in the dog as well as podocytes and the afferent arteriole in humans. The activation of renin release by COX-2 in response to hypochloremia has been documented in animals, and it has been suggested that the activation of renin is due solely to COX-2, with no involvement of COX-1. Further studies have shown that COX-2-derived PGE2 is released from the macula densa, which causes vasodilatation in the afferent arteriole in the face of vasoconstriction produced by angiontensin II, norepinephrine, and vasopressin. Thus, blockade of this renal vasodilatory COX-2-induced PGE2 may contribute to the decrease in GFR seen with NSAIDs in times of low effective circulating fluid volumes. Salt loading down regulates COX-2 expression in the renal cortex, but upregulates it in the inner medulla and vice versa, suggesting that prostaglandins play varying roles in different regions of the kidney: prostaglandins in the renal cortex protect glomerular circulation in times of decreased blood volume, and prostaglandins in the medulla promote diuresis and natruriesis in times of increased circulating blood volume. Taken together, these findings support the risk of renal decompensation from NSAID usage is highest in hypovolemia or dehydration, pre-existing renal disease, congestive heart failure or hepatic disease. Certainly, the geriatric dogs and cat must be treated with NSAIDs carefully, and when using preoperative NSAIDs in anesthesia, hydration status must be maintained. Renal dysfunction may occur with NSAID administration as a consequence of prostaglandin inhibition. Renal prostaglandin synthesis is very low under normovolemic conditions. When normovolemia is challenged, prostaglandin synthesis is increased and important to maintaining renal perfusion. NSAID use must be considered very carefully in hypovolemic animals. This is especially important to remember with the increasing use of NSAIDs perioperatively for pain management. It must be noted that the COX-1 sparing NSAIDs do not infer greater safety in the kidney.

Effects on the Bone:

Bone is a highly dynamic tissue with turnover occurring continuously. In the normal state, NSAIDs do not appear to alter bone homeostasis. However, there is a different dynamic following trauma and fracture of the bone. Bone healing is a complex process: the repair process begins, like any other tissue, with an inflammatory response and is accompanied with synthesis and release of mediators of inflammation. The main mediators are the prostaglandins (PG). Cyclooxygenase (COX) activity is essential for PG production and is the rate limiting enzyme in the synthetic pathway. In other words, without initiation of the inflammatory response, the subsequent resorption and production of bone that is necessary for healing will not occur. Not surprisingly, PGs have been shown to have a major role in bone formation, resorption and repair in numerous studies. Conventional non-specific NSAIDs including indomethacin, naproxen, diclofenac, and aspirin have been shown to slow fracture healing, alter mineralization and inhibit Haversian system remodeling in several (rabbit, rat, mouse) models. With the non-specific NSAIDs, the effects appear to be both time and dose dependent. Whether COX-1 sparing drugs have the same effects on fracture healing, spinal fusions and bone ingrowth remains controversial, but recent data supports the effects of COX-1 sparing drugs also as dose and time dependent and are reversible. Again, these statements are based on rabbit/rat/mouse induced fracture models that show that COX-1 sparing agents do alter bone healing. However the most recent data confirms that following cessation of NSAIDs, fracture healing does return to its normal rate and thus judicious use of postoperative NSAIDs can be recommended. Said another way, any potential adverse effects must be weighed against potential benefits that include but are not limited to improved analgesia and earlier return to function (both mobilization of the limb/patient and weight bearing) and data supports the use of NSAIDs in the immediate post operative period as long as it is not continuous for several weeks.

Effects on the Cardiovascular System:

Non–selective NSAIDs inhibit the platelet COX-1 enzyme, and cause a significant decrease in the amount of thromboxane (TXA2) produced by clotting platelets. Thromboxane is an important activator of platelet aggregation in most dogs, and is released by activated platelets to encourage further platelet recruitment to the site of vessel injury. Thromboxane is also a potent vasoconstrictor. A decrease in thromboxane release can result in prolongations of primary hemostasis. COX-1 sparing (COX-2 selective inhibitor) drugs do not have this effect on thromboxane production or primary hemostasis. The actions of thromboxane are balanced at the vessel level by the presence of prostacyclin (PGI2), which is produced by COX enzymes in the vascular endothelial cells. PGI2 is a strong inhibitor of platelet aggregation, and also results in vasodilation. In the presence of endothelial inflammation (such as that caused by atherosclerotic placques), the expression of COX-2 in the endothelial cells increases, and may produce the majority of prostacyclin in that area. When COX non-selective NSAIDs are administered, the expression of both thromboxane from platelets and PGI2 from endothelial cells is decreased, preserving the balance. In certain circumstances of endothelial inflammation (eg. with atherosclerosis), specific COX-2 inhibitors may decrease the endothelial production of PGI2 (mainly from COX-2), without a concomitant decrease in platelet thromboxane (produced only by COX-1), and consequently may result in the development of a hypercoagulable state. Aspirin is the NSAID that has been most studied in regards to clotting in cats. The hope was that a dose of aspirin could be found that would prevent thrombosis associated with cardiomyopathy and heartworm disease, but not be toxic to the cat. Results have been somewhat conflicting, but overall, it appears that aspirin fails to prevent thrombosis associated with vascular injury in the cat unless administered in extremely high doses. Recent data from humans has elevated concerns about the COX-1 sparing drugs and their ability to increase the occurrence of a thrombotic event by blocking endothelium-derived PGI2, resulting in an unopposed action of TXA2, as an endothelium-contracting and prothrombotic molecule. The significance of this potential adverse effect has not been determined in small animals.

Effects on the Reproductive System and Organ Development:

The role of COX-derived prostaglandins in the female reproductive tract and fetal development has been under investigation. Mice deficient in COX-2 are unable to ovulate and have abnormal implantation and decidualization responses. COX-2 is present in the uterine epithelium in early pregnancy and may be involved in ova implantation, angiogenesis, and the initiation of labor. Additionally, COX-2 receptors have been found in the ductus arteriosus in rats. The secretion of prostanoids, especially PGE2, may keep the ductus patent in utero is indicated by the fact that treatment with COX-1 sparing drugs pre-partum causes premature closure of the vessel. COX-2 is involved in development of the fetal and neonatal kidney in rats and mice. COX-2 deficient mice had significantly retarded development of the renal cortex and glomerulogenesis, and a higher neonatal death rate from renal failure. Although data are limited and the most drug labels do not specifically list pregnancy as a contraindication, prudence suggests that COX-1 sparing drugs not be given to pregnant animals.

Effects on Neoplastic Cells:

A variety of NSAIDs have shown promising effects controlling the growth of neoplastic cells and may have a therapeutic role in the prevention, treatment, and palliation of certain cancers. COX-2 is expressed on mesenchymal cells such as fibroblasts and macrophages and may potentiate tumor growth by stimulating the production of various growth factors (VEGF, bFGF, HGF), angiogenesis, and tumor invasion. Over-expression of COX-2 on epithelial cells has been shown to induce resistance to apoptosis, and cause phenotypic changes resulting in increased adhesion to extracellular matrix proteins. In human and rat colonic tissue, cancerous tissue had greater concentrations of PGE2 compared to normal tissue. Increased expression of COX-2, but not COX-1, has been documented in colorectal carcinomas, and it is reported that approximately 85% of adenocarcinomas have a two-to-50 fold increase in COX-2 expression at the mRNA and protein levels. A study of naturally occurring colorectal tumors and polyps in dogs demonstrated that COX-2 was upregulated in pathologic tissue, and absent in normal colonic and small intestinal tissue. Similarly, in a study of canine transitional cell carcinoma of the bladder, COX-2 was not present in normal bladder tissue but was highly upregulated in all primary and metastatic tumors, as well as the surrounding blood vessels in several dogs. Elevated levels of COX-2 expression have also been documented in lung, gastric, breast, head and neck, and pancreatic tumors. A recent retrospective study revealed a marked increase in median survival time when canine prostatic carcinomas were treated with carprofen or meloxicam, compared to a placebo.

Contraindications:

One must adapt therapy to suit the patient's requirements. In patients with chronic disease, begin with the recommended dose, and if efficacious, attempt to reduce dose at regular intervals (ex. weekly) until you have the lowest dose providing the maximum benefit. Avoid NSAIDs in patients with known contraindications to their use. Contraindications for NSAID use can be fairly obvious to quite subtle. The following recommendations are general guidelines. Patients with the following conditions should be given NSAIDs judiciously and under close supervision: renal or hepatic insufficiency or dysfunction, any clinical syndrome that creates a decrease in the circulating blood volume (e.g. shock, dehydration, hypotension, or ascites), any type of confirmed or suspected coagulopathy (this may be less important with COX-1 sparing drugs), active GI disease, patients who have sustained traumatic injury with known or suspected significant active hemorrhage or blood loss, pregnant patients or females attempting to become pregnant, and patients with significant pulmonary disease (this may be less important with COX-1 sparing drugs).

TAKE HOME MESSAGE OF ADVERSE EVERNTS:

The most common problems associated with NSAID administration to dogs and cats involve the gastrointestinal (GI) tract. Signs may range from vomiting and diarrhea, including hematemesis and melena, to a silent ulcer which results in perforation. Realistically, the true overall incidence of GI toxicity in dogs or cats treated with NSAIDS is unknown. Concurrent administration of other medications (especially other NSAIDs or corticosteroids), previous GI bleeding, or the presence of other systemic diseases may contribute to adverse reactions. The effect aging has on an individual patient's ability to metabolize NSAIDs is likely to be quite variable. Hepatotoxicosis caused by NSAIDs is generally considered to be idiosyncratic. Most dogs recover with cessation of treatment and supportive care. Renal dysfunction may occur with NSAID administration as a consequence of prostaglandin inhibition. Renal prostaglandin synthesis is very low under normovolemic conditions. When normovolemia is challenged, prostaglandin synthesis is increased and important to maintaining renal perfusion. NSAID use must be considered very carefully in hypovolemic animals. This is especially important to remember with the increasing use of NSAIDs perioperatively for pain management.

References available upon request