Prostaglandin Formation

Eicosanoids, such as PGs and leukotrienes, are 20-carbon chain derivatives of cell membranes. Eicosanoids are potent mediators of inflammation and are particularly important in the later stages.^3,8 These compounds are synthesized when oxygen reacts with the polyunsaturated fatty acids of cell membrane phospholipids (see Figures 29-1 and 29-2). The most important of these fatty acids is AA, an omega-6 fatty acid, which is released into the cell from phospholipids of the damaged cell membranes. Release reflects activation of phospholipase A₂ (dependent on calcium and calmodulin) located in the cell membrane.⁹ Once inside the cell, AA serves as a substrate for enzymes that generate intermediate and ultimately the final eicosanoid end product.^8,10 The formation of lipoxygenases and their subsequent inhibition is discussed under “Miscellaneous Antiinflammtories.” COX PG synthase or prostaglandin H (PGH) synthase, located in all cells except mature red blood cells, add oxygen to AA, generating unstable PG endoperoxides (PGG₂). Subsequent peroxidase reactions convert PGG₂ to PGH₂, the precursor of all PGs and thromboxane. The final PG end product depends on the presence of specific isomerase reductase or synthetase enzymes, some of which may be inducible.^8,9

The direct effects of PGs are mediated through cell membrane–spanning G protein receptors located on target tissues and cells. At least nine subtypes have been identified, corresponding to each of the COX metabolites: DP1-2 (PGD₂), EP1-4 (PGE₂), FP A, B (PGF_2α), IP (PGI₂), and TP α, β (TXA₂) (G&G);¹¹ another four have been characterized for PGE₂, opening the potential for selective drug activity.⁹ In addition to their receptor interactions, PGs also act indirectly by enhancing other mediators such as histamine and bradykinin.

The role of PGs in normal physiology as well as in disease might best be understood by considering them as protective in nature, even those associated with inflammation. Their formation is mediated by one of at least two isoforms of cyclooxygenases (see Figures 29-1 and 29-2), located on different genes in humans.^12–14 COX-1, the “housekeeping” isoform,¹⁴ mediates the formation of constitutive PGs produced by many tissues, including gastrointestinal cells, platelets, endothelials cells, and renal cells. PGs generated from COX-1 are constantly present, providing homeostasis through a variety of normal physiologic effects. These include protection of the gastrointestinal mucosa, hemostasis, and the kidney when subjected to hypotensive insults. COX-2 is the product of an “immediate-early” gene that is rapidly inducible and tightly regulated.¹⁴ Regulation of COX expression is complex, with more sites present on the COX-2 compared with the COX-1 gene. Its expression is tightly restricted (but not absent) under basal conditions, but it is dramatically upregulated in the presence of inflammation or other diseases.⁹ Diseases associated with increases include (in humans) rheumatoid arthritis, seizures, ischemia, and (posttranslational) cancer.⁹ Proinflammatory cytokines such as TNFα and the interleukins stimulate the expression of COX-2 in many cell types, such as synovial cells, endothelial cells, chondrocytes, osteoblasts, and monocytes and macrophages.¹⁴ COX-2 catalyzes the formation of inducible PGs, which are needed only intermittently or under specific situations.^12,13,15,16 A third COX isoform has been suggested, but it appears to be a variant of COX-1. Originally identified in high concentrations in canine cerebral cortex (hence the term used by some investigators, “canine COX-3”),¹⁷ it is inhibited by acetaminophen (and other NSAIDs). Its inhibition may influence central analgesic effects of NSAIDs.^1,2 Differential or selective inhibition of COX-2 clearly offers potential some safety benefits by avoiding loss of homeostatic PGs; however, loss of COX-2 activity is also associated with adversities.¹⁴ It is important to note that COX-2 is ofttimes constitutively expressed: in the kidney and brain it mediates a cytoprotective effect in damaged or inflamed gastrointestinal mucosa.

Cyclooxygenases in health and disease

An appreciation of both efficacy and safety of COX inhibition (e.g., NSAIDs) is facilitated by an understanding of the role of COX in healthy and diseased tissues. COX has been described as the most common target of drug therapy with NSAIDs. In general, inhibition of COX-2 is responsible for efficacy, whereas inhibition of COX-1 is responsible for side effects.⁹ However, strict adherence to this simplistic approach will lead to therapeutic failure and increased morbidity with NSAID use. Although COX-1 does indeed appear to be the predominant constitutive enzyme responsible for housekeeping, both COX-1 and COX-2 are constitutively expressed in many tissues. Further, although COX-2 clearly is more active in the promotion of inflammation, COX-1 does appear to have some role.

Gastrointestinal and Other Healing

Both COX-1 and COX-2 are constitutively expressed in the gastrointestinal tract. However, constitutive expression of COX-1 has a predominant role in the protection of the gastrointestinal tract. COX-1 PGs decrease hydrochloric acid secretion, increase mucosal bicarbonate and mucus production, and increase epithelial cell proliferation and mucosal blood flow (Figure 29-5). The latter effect facilitates delivery of oxygen and nutrients to proliferating cells as well as rapid removal of damaging hydrogen ions that make their way into the mucosa. Drugs that express preferential potency for COX-2 PGs generally are associated with fewer gastrointestinal side effects compared to those targeting both COX isoforms, although the relative protection varies with the drugs. Protection appears to be more important in the presence of high drug concentrations (e.g., high doses).²⁰ COX-2 is critically important to the gastrointestinal tract as well, being important for healing of gastrointestinal damage: COX-2 appears within 1 hour of gastrointestinal damage. Newer coxib PGs will inhibit gastrointestinal healing as has been demonstrated in dogs receiving firocoxib (see the discussion of tepoxalin).²¹ Interestingly, in the pancreas, constitutive COX-2 expression dominates, although the clinical relevance of this is not yet known.²⁰ The impact of NSAIDS on bone healing is an emerging concern. COX influences osteoblast and osteoclast; experimental studies have demonstrated impaired bone healing.^20a Their use in controlling pain and limiting ectopic bone growth must be balanced with the increased risk of impaired fracture healing; conventional NSAIDs might be preferred compared to newer NSAIDs. The impact of NSAID on soft tissue healing (or ligaments) remains to be confirmed; whereas short term use may have minimal negative (and potentially a positive) impact, long term used might be approached cautiously. A recent meta-analysis in humans found no advantage in terms of return to function for coxibs, but recommended caution with long term use.^20b A strong association has been reported between NSAIDs and severe necrotizing soft tissue infections.^20c

Figure 29-5 Among the many protective functions of constitutive prostaglandins is protection of the gastrointestinal mucosa from acid and other mediator-induced damage. Protective mechanisms include production of mucus and bicarbonate and inhibition of hydrogen ion secretion; epithelial turnover, which ensures rapid replacement of damaged cells; and increased mucosal blood flow, which ensures provision of oxygen and nutrients to the rapidly dividing epithelial cells as well as removal of hydrogen ions that are able to pass into the cells.

Cardiovascular Disease

The role of PGs in the cardiovascular system is largely beneficial, and the relationship between COX-1 and COX-2 and their respective PG end products exemplifies the complex “ying–yang” balance that characterizes them (Figure 29-6). Platelets contain thromboxane, a synthase, which catalyzes the formation of thromboxane from AA. Thrombosis reflects platelet aggregation and vasoconstriction. The formation of a thrombus is kept in check by the presence of prostacyclin synthase in vascular endothelial cells. This enzyme catalyzes metabolism of AA to prostacyclin (PGI2), a vasodilatory and platelet-inhibiting PG end product. However, whereas thromboxane A2 (TXA2) is associated with COX-2, prostacyclin synthase co-localizes with COX-1. Thus, whereas drugs that target both COX isoforms will potentially allow the balance to be maintained, drugs that preferentially target only one isoform risk disruption of the balance. Such may be the case with COX-2 selective NSAIDs. Their preferential inhibition of COX-2 may allow thrombus formation to go unchecked, increasing the risk of thromboembolic disorders. Indeed, in a meta-analysis of human clinical trials, two coxib drugs (rofecoxib and valdecoxib) were associated with an increased risk of stroke (see “Adverse Events”).^20d

Figure 29-6 The role of constitutive prostaglandin products in hemostasis exemplifies the “ying-yang” relationship that often characterizes their action. Platelet activation in response to the damage is accompanied by the release of arachidonic acid, which is catalyzed by thromboxane synthetase, present in platelets, to thromboxane. Thromboxane causes platelet aggregation and local vasoconstriction. Excessive hemostasis is kept in check, however, by the simultaneous release of arachidonic acid from the damaged endothelial cell surface. Prostacyclin synthetase, located in the endothelial cells, results in the formation of prostacyclin, which is vasodilatory and inhibits platelet aggregation. cAMP, Cyclic adenosine monophosphate.

Kidney

In the kidney both COX-1 and -2 are constitutively expressed. Both are formed in the macula densa of humans and animals, but COX-2 may have a more important role than COX-1 (Figure 29-7). In animals, inhibition of COX-2 causes sodium and potassium retention in salt-depleted, but not normal, animals. However, in humans COX-2 appears to influence renal vasculature and podocytes. The role of COX in the kidney requires further elucidation before safety can be assumed for any NSAID; sparing COX-1 and targeting COX-2 can be expected to alter renal function. For example, kidneys do not develop in the embryos of COX-2–null knockout mice.²² The role of COX differs among tissues and species; their impact on adverse reactions are addressed under “Adverse Reactions.”⁹

Figure 29-7 Both constitutive and inducible prostaglandins are important to renal blood flow, with the specific effect of each product varying among species. Renal prostaglandins ensure that intramedullary renal blood flow and urine formation will continue in the presence of decreased renal perfusion. Sites of action are noted by numbers and include shunting of blood from the cortices to the medullary intersitium, stimulation of natriuriesis, and inhibition of antidiuretic hormone. The nephrotoxicity of selected drugs reflects inhibition of renal prostaglandins.

Cycloloxygenase and nonsteroidal antiinflammatory drugs

NSAIDs (Table 29-2) act to block PG synthesis by binding to and inhibiting cyclooxygenase (see Figures 29-1 and 29-2).⁸ This action is both dose and drug dependent. The planar form that characterizes these drugs is thought to facilitate their binding to COX.^6,25 Several investigators have shown that some drugs (e.g., phenylbutazone and flunixin meglumine) also reduce later steps of the formation of PGE₂ in inflammatory exudate at therapeutic doses.²⁶ However, both the major therapeutic and toxic effects of NSAIDs have been correlated extensively with their ability to inhibit PG synthesis, with anti-inflammatory potency related to potency of impaired PG synthesis.⁸

Table 29-2 Dosages of Nonsteroidal Antiinflammatory Drugs in Cats and Dogs

The differential effect of NSAIDs on the isoforms of COX presumably contribute to clinical differences in efficacy and safety, .^17,27 Whereas as a class, NSAIDs appear to inhibit both COX-1 and COX-2, the ratio of the concentration of a NSAID necessary to inhibit the same level of COX-1 vesrus COX-2 activity (e.g., inhibition of 50% or 80% activivity [IC5₀ or IC₈₀]) compares the potency each drug toward each isoform, and as such, provides a basis for screening predicted relative safety and efficacy of each. Thus, a COX-1 to COX-2 ratio greater than 1 (or a COX-2 to COX-1 ratio of less than 1) indicates greater potency for COX-2 which is desirable in that formation of inflammatory PGS should be inhibited preferentially to the housekeeping PGs.^12,13,15,16 Inhibition of platelet activity is commonly used to measure in vitro COX-1 inhibition whereas inhibition of PGs released from macrophages stimulated with endotoxin is used to measure COX-2. A number of drugs associated with in vitro COX-2 selectivity in humans, including rofecoxib (Vioxx) and celecoxib (Celebrex); others include etodolac, meloxicam, and nimesulide; the newest drugs are valdecoxib, etoricoxib, and (soon to be approved) lumiracoxib.⁹ Drugs approved for use in animals include carprofen (the first) followed by etogesic; deracoxib; meloxicam; and, most recently, firocoxib.

Although the gene for COX-1 is about 3 times as large as that for COX-2, the two COX proteins, which are membrane bound, share about 60% homology.⁹ Differences in NSAID selectivity for these two enzymes reflects smaller amino acids at two positions in COX-2 compared with larger amino acids at the same site in COX-1. The result is a larger and more flexible “pocket” into which the newer drugs insert and inhibit the COX-2 isoform.⁹ For example, ”coxib” drugs contain a sulfonamide group that interacts through hydrogen bonding to arginine of COX-2 but not histidine in the same position in COX-1.²⁸ Despite the usefulness of the COX-1:COX-2 ratio in screening drugs for safety or efficacy, its applicability to clinical use is questionable. Further, comparing results among studies is difficult. For example, whereas some studies focus on the IC_50, others determine the IC₈₀, the latter probably being more clinically relevant. Assay methods contribute to variability in ratios among investigators for the same drug in the same species (Table 29-3). In vitro assasys based on cell culture or recombinant enzymes are easier to perform, but it is the ex vivo whole blood assays (as opposed to in vitro cell culture assays) that generally are recognized as the most representative of the clinical patient.¹ However, even whole blood assays do not take into account different distribution patterns to different tissues.^18,22 Extrapolation among species should be avoided.²⁹ For example, wherease the COX-1 to COX-2 ratio of etodolac is much better than that for carprofen in humans,^20,30 the opposite is true in dogs. Indeed, Wilson and coworkers³¹ have demonstrated that, wherease critical residues in the active sites of canine COXs are identical with human homologs, amino acids differ (up to 54) at in active sites and may contribute to differences in activity that preclude extrapolation among species (see Table 29-3). Further complicating extrapolation among and within species is the fact that NSAIDs generally exist and are largely marketed as enantiomers (see Chapter 1), each with a possible different ratios. For example, Ricketts and coworkers³² found the COX-1:COX-2 ratio of carprofen to be 129 for the racemic mixture but 181 for the S isomer and only 4.19 for the R isomer. Little information is available regarding COX selectivity in cats.^33,34 Much of the information that is available is based on manufacturer-sponsored studies and as such might potentially reflect reporting bias. For example, in Table 29-3, drugs that generally performed best in each study tended to be the drug manufactured by the sponsor of the study. As such, COX ratios might be considered a screening procedure^1,2

Table 29-3 COX-1 to COX-2 Ratios for Selected Nonsteroidal Antiinflammatory Drugs in the Dog by Different Investigators

Pharmacologic Effects

The pharmacologic effects of this class of drugs include analgesia, antipyresis, and control of inflammation. Antithrombosis also occurs, but this effect is variable, being the greatest for aspirin and the least to absent for newer COX-2 selective drugs. The effects are dose and drug dependent and include antithrombosis (when present), which occurs at the lowest concentrations (relatively lower for aspirin compared to all others); followed by antipyresis, analgesia and antiinflammation, with the latter occurring at the highest concentrations, although an exception appears to occur again for aspirin.^10,43d Studies that compared efficacy and safety traditional or newer NSAIDs are generally are based on in vitro or ex vivo methods, with clinical trials limited. Yet, differences should be anticipated even among the newer NSAIDs. The mechanisms by which NSAIDs inhibit (interact with) COX are responsible, in part, for the variable antiinflammatory effects that characterize these drugs. Although several NSAIDs are characterized by a short plasma elimination half-life, the clinical response may last for over 24 hours after a single dose or up to 72 hours after multiple doses. These differential durations reflect, in part, drug-receptor interactions,with irreversible binding to COX most likely contribute to a longer biologic response.⁴⁴ Lees and co-workers have reviewed the relationship of NSAID effect and dose, emphasizing the advantages of PK-PD modeling as a basis of dose design.^44a Inhibition of COX-1 has been described as simple competitive inhibition, whereas that for COX-2 has been described as a slowly reversible, contributing to a longer biologic versus plasma half-life.⁴⁵ Aspirin binds reversibly to the COX activity site on PGH synthase and then irreversibly inactivates the enzyme by acetylating a serine residue.¹⁰ Both thromboxane and prostacycline synthase are impacted. However, platelets cannot produce additional thromboxane synthase; as such, platelet activity depends on new platelet production in the absence of aspirin. In contrast, endothelial cells are able to synthesize more prostacyclin synthetase and are less susceptible to the inhibitory effects of low dose aspirin.¹⁰ In contrast to irreversible binders of COX, ibuprofen binds reversibly with COX and thus competes with AA for binding. In laboratory animals and humans, a relative potency (not to be confused with efficacy) has been established for selected conventional NSAIDs and COX: meclofenamic acid >indomethacin >naproxen >phenylbutazone >aspirin.⁴⁴ However, differential potency for COX-1 versus COX-2 varies among the NSAIDs and their enantiomers. In addition to drug-receptor interaction, prolonged elimination of NSAIDs from inflammatory exudate compared with plasma may also result in differential effects among the drugs and species.^37,44

Along with inhibition of PGs, disruption of cellular signaling is responsible for all three pharmacologic effects (antiypresis, analgesia, anti-inflammation) that characterize all NSAIDs.¹⁰ The antipyretic effect of NSAIDs may or may not be of benefit. Indeed, in human medicine little scientific support exists for the use of antipyretics for relief of discomfort or reduction of morbidity and mortality associated with fever.⁴⁷ Fever may have a beneficial effect on the outcome of bacterial or viral infections and often inversely correlates with the severity of lesions or the time to resolution. Thus the antipyretic effects of NSAID may be of most benefit when increased metabolism induced by fever might prove detrimental, such as with septicemia.⁴⁷ Central analgesic effects have been suggested for some but not all NSAIDs because NSAIDs can provide analgesia at very low intrathecal doses.^15,16,46

The impact of COX on disease has led to a number of new applications for NSAIDs. The antiendotoxic effect of selected NSAIDs has been known (see the discussion of flunixin), but newer mechanisms are being examined. For example, several NSAIDs (carprofen, flunixin meglumine, and phenylbutazone) may specifically inhibit activation of the proinflammatory transcription factor nuclear factor kappa B (NF-κB) and on lipopolysaccharide (LPS) induction of iNOS. Using a mouse macrophages line, carprofen and flunixin, but not phenylbutazone, inhibited iNOS, whereas carprofen and, to a lesser degree, flunixin, inhibited NF-kappaB activation.³⁵

Because NSAIDs are commonly first-choice treatments for arthritis, a focus on the effects of NSAIDs on cartilage is warranted. NSAIDs have been shown to inhibit proteoglycan synthesis in vitro, and for the salicylates this is supported by in vivo studies.⁴⁸ This effect has been attributed to inhibition of uridine diphosphate glucose dehydrogenase, an enzyme important in proteoglycan synthesis.⁴⁸ Hyaluronic acid synthesis, also dependent on this enzyme, does not appear to be as affected. The clinical effects of NSAIDs on cartilage are controversial. Some NSAIDs may, in fact, favorably modify the metabolism of proteoglycans, collagen, and matrix and may decrease the release of proteases or toxic oxygen metabolites.⁴⁸ This may reflect, in part, control of inflammation. Thus, whereas several NSAIDs have documented adverse effects on normal cartilage, ranging from decreased proteoglycan synthesis (e.g., aspirin) to chondrocyte death (phenylbutazone), others (e.g., naproxen, piroxicam, ketoprofen, and possibly carprofen) are recognized for their chondroprotective effects. Dual inhibitors have been noted for their potential efficacy to slow the progression of arthritis.^48b

The use of NSAIDs in the control of acute pain is evolving. Preemptive use (just before and immediately after surgery) has been demonstrated in humans to be more effective than either placebo or control,¹⁷ and similar findings have been reported in the clinical use of COX-1–sparing drugs in animals. Drugs that target COX-2 (NSAIDs) offer a morphine-sparing effect for control of postoperative pain.¹⁷ Their use in multimodal analgesia is increasing.

NSAIDs are currently being investigated for their potential antitumor effects. Knapp and coworkers⁵⁰ found piroxicam to be clinically useful in reducing tumor size and increasing survival time in dogs with transitional cell tumors of the urinary bladder. The result does not appear to reflect direct cytotoxic effects.^49–51 NSAIDs may be beneficial when combined with other anticancer drugs.⁵² The effects of drugs inhibiting COX-2 cancerogenesis may be dose dependent: At least for aspirin, the maximum anticancer effect occurs at a dose less than that associated with control of inflammation but similar to that associated with antiplatelet effects.⁹ Mechanisms other than COX inhibition contribute to NSAID antitumor effects.^51a In addition to the inhibition of cancer growth, the use of COX-1–sparing drugs may improve gastrointestinal tolerance of anticancer drugs.

Risk factors for NSAID-induced gastrointestinal adverse events

Several risk factors for NSAID-induced gastrointestinal adverse events identified in humans are applicable to animals. No chemical characteristic predicts the likelihood of gastrointestinal toxicity by a particular conventional NSAID.^70,72 Drugs that undergo enterohepatic circulation (e.g., naproxen, carprofen, etodolac) may be associated with a greater incidence of gastrointestinal upset. Other risk factors include advanced age (altered disposition coupled with decreased ability to protect damaged mucosa),⁷⁷ concurrent use of glucocorticoids (increasing the risk 4.4-fold in humans) or other NSAIDs (the exception might be low-dose aspirin, as discussed later), or anticoagulants. Comorbidity (renal, cardiovascular, or liver disase) is also a risk factor.⁷³

The combination of steroids and NSAIDs causes worse lesions than either drug alone, as has been demonstrated for flunixin meglumine or dexamethasone.^74,78,79 Further, the FDA adverse drug events site indicates that the risk of gastrointestinal toxicity is greater when drugs are given with glucocorticoids. In a restrospective study of gastrointestinal perforation associated with deracoxib, a risk factor was combination with glucocorticoids (or other NSAIDs) within the past 24 hours.⁸⁰ Other NSAIDs appear to shift use of AA from the COX pathway to the production of cysteinyl LTs and LTB4, eicosonoids that promote leukocyte migration, break down the mucosal barrier, and stimulate gastric acid secretion. Inhibition of COX-2 may preclude angiogenesis critical to the healing ulcer.⁸¹ Persons with previous history of gastric ulcer disease also are predisposed to NSAID-induced gastrointestinal adverse drug events. Although peptic ulcer disease is unusual in animals, prudence dictates that evidence of any gastrointestinal disease associated with mucosal damage (e.g., inflammatory bowel disease) that might require COX-2–mediated healing be considered a potential risk factor.

Dose-dependent toxicity has been demonstrated during the approval process essentially for all NSAIDs as is indicated by product package inserts (Table 29-5). The single most common cause of NSAID-induced toxicity found by a review of NSAID-induced adverse drug event reports by the FDA is overdosing, in particular failure to adhere to the recommended dosing regimen. Overdosing was associated with an increased risk of NSAID-induced gastrointestinal perforation in dogs.⁸⁰ Drugs for which low-concentration preparations are not available increases the risk of toxicity because of inaccurate dosing. Use of compounded preparations (for which oral bioavailability is not known) should be implemented with extreme cautions; increasing bioavailability will result in a proportional increase in drug concentrations. The risk for overdosage is greater for selected drugs for which higher doses may result in zero-order elimination (e.g., deracoxib, phenylbutazone). For such drugs half-life is no longer germane, and the risk of drug accumulation dramatically increases.

Table 29-5 Safety Data for Selected NSAIDs in Dogs and Cat Based on Target Safety Studies and Clinical Trial Field Studies

The relationship between dose and risk of toxicity remains complex, even for NSAIDs, as is exemplified by aspirin. At low doses (concentrations), aspirin blocks COX activity and thus shunts AA into the lipoxygenase pathway. Although gastrotoxic LTs are produced (cysteinyl LTs and LTB4), aspirin also simultaneously triggers lipoxin (aspirin-triggered lipoxin [ATL]) formed by lipoxygenase-15. Lipoxin is antiinflammatory and thus inhibits the gastrotoxic leukocyte recruitment and activation caused by cysteinyl LTs, as well as mediates the antiadhesive platelet effects of aspirin. As such, low-dose aspirin tends to be both cardioprotective as well as safe to the gastrointestinal tract. However, at higher doses production of LTs is sufficient that the protective effects of lipoxin are masked and gastrotoxicity emerges. Interestingly, COX-1–sparing (COX-2 preferential drugs) inhibit the formation of lipoxin, contributing to the toxicity of the newer drugs when combined with low-dose aspirin.⁸¹ Theoretically, whereas COX-inhibiting drugs should not be combined with low-dose aspirin, dual-acting NSAIDs, which do not target lipoxygenase-15 (see the discussion of dual-acting NSAIDs), should not inhibit lipoxin and thus should maintain the low-dose aspirin gastroprotective and cardioprotective effects.⁸¹

In general, conventional NSAIDs are associated with a greater risk of gastrointestinal adverse drug events in humans, but exceptions may occur. For example, in humans (but not dogs) ibuprofen, which is essentially equivalent in its COX-1 versus -2 selectivity, is ranked as low risk, along with the COX-1–sparing drugs celecoxib, rofecoxib, meloxicam, and etodolac.⁷³ Comparison in dogs of gastrointestinal safety among newer NSAIDs, including both COX-1–sparing drugs and dual-acting NSAIDs, might be based on toxicity data generated during the approval process and available on package inserts. This latter source includes both targeted safety studies (the dose is multiplied by 3, 5, or 10) and field studies (clinical trials at recommended doses under conditions of anticipated use) (see Table 29-5).

Comparison of clinical safety among NSAIDs ideally should reflect nonmanufacturer-sponsored, well-designed clinical trials involving a sufficient number of animals such that the power is sufficient to detect a significant difference. Placebo controls are imperative; positive responses have been reported in 40% or more of placebo-treated animals.⁸² However, such studies in veterinary medicine are few and far between. Data that do exist thus far support the safety of the newer drugs compared with conventional NSAIDs. For example, using a placebo-controlled, parallel, randomly assigned design in dogs (n = 6/group), Reimer and coworkers⁸³ found that buffered aspirin caused endoscopic gastric lesions within 5 days of starting therapy compared with no or minimal lesions in animals treated with carprofen or etodolac (sample size may have limited the ability to discern a difference between the two groups); all drugs were administered at mid-recommended doses. Endoscopic lesions followed the same relative pattern among drugs throughout the 28-day study period, indicating that aspirin is more commonly associated with lesions in the canine gastrointestinal tract compared with the COX-1–sparing drugs. A separate study found no differences in gastrointestinal lesions in animals receiving ketoprofen (nonselective), carprofen, meloxicam, or placebo. However, the power of this latter study to detect a significant difference was not reported.

Pharmacologic prevention and treatment

Future pharmaceutical manipulations designed to deliver NSAIDs by alternative routes may decrease risk of gastrointestinal adverse drug effects, although their application to veterinary patients should not necessarily be assumed. For example, formation of nitroso derivatives of conventional NSAIDs may improve the safety margin of these drugs as a result of in vivo release of NO that provides gastroprotection while improving antiinflammatory and analgesic potency.² However, NO also has been associated with the pathogenesis of osteoarthritis, thus exemplifying the continuously complicated nature of designing safe NSAIDs (see later discussion).²² In contrast to current drugs that only slowly dissociate with COX-1, newer NSAIDs appear to be weak and rapidly reversible binders of COX-1, thus enhancing safety.^17,45

Topical (including transdermal) NSAID administration is appealing because it avoids direct contact between the drug and target tissue. Indeed, a meta-analysis of clinical trials comparing topical NSAIDs to placebos and oral NSAIDs in humans found this route to be effective, with no difference in response compared with oral. Topical administration was safe, although not necessarily safer than oral. Of the NSAIDs reviewed, ketoprofen was described as the best.⁸⁴ Thus far, topical NSAID administration has not proved to be a vital means of avoiding gastrointestinal adverse drug events in small animals, primarily because of failed drug delivery. Because the amount of drug delivered cannot be predicted, administration of NSAID in novel drug delivery systems offered by compounding pharmacists is not recommended unless the amount of drug delivered by that system has been scientifically demonstrated, as would occur for an approved drug. Enteric coatings, combination drugs (e.g., with gastroprotectants), and other approaches may be of some benefit for some drugs (e.g., aspirin) and, if currently available, are discussed with the individual drugs.

Both the prevention and treatment for gastrointestinal toxicity focus on, in order of priority, control of gastric acid secretion, replacement of the missing PGs, and (for treatment) protection of the damaged mucosa.^64,85 The two major categories included antisecretory drugs and the PG analog misoprostol; cytoprotectants, such as sucralfate, also play a role in treatment.

Among the antisecretory drugs, proton pump inhibitors (PPIs) generally have proved more effective than H₂ receptor blockers for prevention and treatment of NSAID-induced gastrointestinal adverse drug events. An exception has been demonstrated in humans for famotidine, but only when it is administered at higher than recommended doses (40 mg twice rather than once daily).⁷³ Boulay and coworkers⁸⁶ demonstrated that cimetidine did not protect the gastric mucosa from developing lesions in dogs receiving nonbuffered aspirin (35 mg/kg every 8 hours). PPIs appeared to be more effective, as well as better tolerated, compared with misoprostol. In one human study, close to 30% of persons reciving misoprostol could not tolerate the full dose.⁷⁶ Omeprazole generally is the PPI of choice, although lansoprazole may also be a reasonable choice. Although lansoprazole was not found to be superior to misoprostol for prevention of gastrointestinal adverse drug events in humans, improved compliance and better tolerance made it the preferred PPI drug in humans. However, the use of PPI will not prevent ulcers in all patients, as has been demonstrated in humans. Symptoms of GI side effects still occur in 20% of human patients receiving omeprazole.⁷³ Although routine use of PPI in nonrisk animals receiving NSAIDs is not indicated, strong consideration should be given to their use in at-risk animals. As a class, the PPIs are inhibitors of drug-metabolizing enzymes, and care should be taken to review drug interactions before their use. Among them, esomeprazole may be the least likely to impact metabolizing enzymes.^86a However, PPI also have been shown to induce selected CYP45O enzymes (P45OIA1 and A2) but this impact appears to be most clinically relevant to (human) poor metabolizers that are deficient in enzymes.^86b

Stay updated, free articles. Join our Telegram channel

Join