The Renal Effects of NSAIDs in Dogs

Introduction

Within the kidney, prostaglandins (PGs) are vasodilators that help maintain renal blood flow (RBF) and glomerular filtration.¹ Upon activation of the renin-angiotensin-aldosterone system (RAAS), increased production of vasodilatory PGs becomes critical within the kidney to offset the vasoconstrictive effects of norepinephrine, angiotensin II (ATII), and vasopressin. Nonsteroidal anti-inflammatory drugs (NSAIDs) have the potential to reduce RBF and glomerular filtration rate (GFR) by inhibiting the cyclooxygenase (COX) production of PGs, especially in the face of RAAS activation.

Healthy canine kidneys express both COX-1 and COX-2, although basal COX-2 expression in dogs is significantly higher than in other species.¹ While COX-1 is most abundant, with expression in renal vasculature, papillary interstitial cells, and collecting ducts, the COX-2 isoform is expressed in the thick ascending limb of the loop of Henle, macula densa, and renal interstitial cells in dogs.¹ When the RAAS is activated, COX-2 becomes more important in the maintenance of RBF and GFR.

NSAIDs that spare COX-1 activity have exhibited less gastrointestinal toxicity, but no NSAID has been proven safe for the kidney. The kidney is the organ with the second highest reports of adverse drug events (ADEs), which usually manifest as functional changes.² However, structural changes, including renal papillary necrosis, can occasionally be observed.

Dogs with chronic kidney disease (CKD) could be expected to be at increased risk for NSAID-related ADEs. Subclinical dehydration and hypertension are common complications of CKD that can result in decreased renal perfusion. As nephrons and renal reserve are lost in CKD, the canine kidney becomes more dependent on COX-2 for production of PGs to maintain fluid balance and RBF.³ Inasmuch as the prevalence of both CKD and osteoarthritis (OA) increases with age, it is expected that many dogs being treated with NSAIDs for OA will have loss of renal reserve and/or early stage CKD.

Renal Hemodynamics

RBF (volume of blood delivered to the kidney/unit of time) and GFR (volume of fluid filtered by the kidneys/unit of time) are two important renal hemodynamic parameters. Although the kidneys account for only 0.5% body weight, they receive approximately 25% of cardiac output.⁴ A majority (90%) of renal blood flow supplies the cortex, with the inner medulla and papilla receiving only 1%. RBF is relatively constant over a broad range of mean arterial blood pressure (80–170 mm Hg) in dogs.⁴

In normal healthy dogs, GFR is largely regulated by tubuloglomerular feedback mechanisms. Those mechanisms involve the juxtaglomerular complex composed of a Na-sensing macula densa in the distal tubule and juxtaglomerular cells located predominantly in the walls of the afferent arterioles. A decrease in GFR slows the flow of filtrate through the loop of Henle, allowing increased time for Na (and chloride) reabsorption. Consequently, less Na reaches the macula densa stimulating vasodilation of the afferent arteriole, which increases glomerular hydrostatic pressure and restores GFR. During that process, renin is released from the juxtaglomerular cells to increase formation of angiotensin I. Subsequently, angiotensin converting enzyme produces ATII from angiotensin I. ATII preferentially constricts the efferent glomerular arteriole, which increases intraglomerular hydrostatic pressure.

The sympathetic nervous system and arachidonic acid metabolites also influence vascular tone in the kidneys. Adrenergic innervation is present along the interlobar, arcuate, and interlobular arteries as well as the afferent arterioles and vasa recta.⁵ Activation of the sympathetic nervous system releases norepinephrine from postganglionic neurons resulting in renal vasoconstriction. This occurs secondary to either hypotension or decreased circulating volume and causes vasoconstriction of both the afferent and efferent arterioles, resulting in a transient (i.e., minutes to hours) decrease in RBF and GFR. PGs and bradykinins counter renal vasoconstriction and tend to enhance RBF and GFR. The most abundant prostanoid in the kidney is PGE₂ with lesser amounts of vasodilatory PGI₂ and PGF_2α.⁶ Prostacyclin synthesis is localized to the cortex, while PGE₂ is found primarily in the medulla.⁶

When hyponatremia and/or hypovolemia occur, renal prostanoid production increases to protect against renal ischemia.⁷ In these situations, synthesis of renal PGs is upregulated by vasoconstrictors such as ATII, catecholamines, and adrenergic input. PGE₂ and PGI₂ are produced in the renal tubules and glomeruli, respectively, to offset vasoconstriction caused by ATII, norepinephrine, and vasopressin.⁷ PGE₂ also acts directly on renal tubules to increase excretion of Na and water and stimulates renin secretion from the macula densa.⁸ Therefore, another potential adverse effect of NSAID administration and decreased renal PG production can be Na and water retention leading to edema.

COX and the Kidney

PGs are produced from arachidonic acid (Figure 1). Arachidonic acid is released from cell membranes into the cytoplasm where it acts as a substrate for COX, lipoxygenase (LOX), and other enzymatic reactions. The precursor PG, PGH₂, is then converted to the prostanoids (PGE₂, PGI₂, PGF_2α, and thromboxane A₂) that exert their biologic effects in close proximity to their site of synthesis.⁶ Although PG production associated with OA contributes to the inflammatory process via a decreased nociceptive threshold, vasodilation, increased vascular permeability, and edema, in the kidney, PGs help maintain RBF and GFR via renal vasodilation.⁹ This renoprotective mechanism can be compromised in dogs treated with NSAIDs (Table 1).

View Figure

• Download as Powerpoint
• Download Figure

TABLE 1 Renal Effects of Prostaglandins, Potential Adverse Renal Effects Associated with NSAIDs, and Recommended Renal Monitoring Parameters

View Fullscreen

The COX-1 and COX-2 isoforms, produced from the parent protein PGH₂ synthase, were discovered in the 1990s. COX-1 is normally present in most healthy tissues. COX-2 can be induced during inflammatory states; however, it is also expressed in, and is necessary for normal function of, gastrointestinal, neural, reproductive, and renal tissues.⁷

Although the renal distribution of expression of the COX-1 isoform is fairly uniform across animal species, interspecies differences in renal COX-2 expression have been recognized. Certain species, such as rats and dogs, have higher basal levels of COX-2 expression in the kidney compared to humans¹. In order to elucidate those differences in COX expression, dogs and monkeys were given the nonspecific COX inhibitor naproxen at 50 mg/kg q 24 hr and 150 mg/kg q 24 hr, respectively, for 2 wk in order to reach a plasma concentration that would maximally inhibit renal COX-1 and 2.¹⁰ Systemic exposures (area under the curve from 0–24 hr) of naproxen were 763 μg/mL/hr and 1918 μg/mL/hr in dogs and monkeys, respectively. Despite similar reductions in renal PG levels, dogs that received naproxen (n = 6) had more significant renal toxicity, manifested by decreases in urine output and Na excretion, than did monkeys, presumably due to a greater degree of COX-2 inhibition. In addition, although GFR decreased in both species, only the dogs also had a decrease in RBF.¹⁰ Immunohistochemistry analysis indicated COX-2 was prominent in the macula densa, thick ascending loop of Henle, and papillary interstitial cells of canine, but not monkey, kidneys.¹⁰ Postmortem examination at 6 wk showed dogs had developed tubular atrophy and interstitial fibrosis in addition to renal papillary necrosis. Sluggish blood flow through the medulla makes this tissue more susceptible to COX-2 induced ischemia.

In a normal canine kidney, the prostanoids are synthesized in both the COX-1 and COX-2 pathways. When hypovolemia occurs in dogs, COX-1 and COX-2 maintain renal blood flow while COX-2 controls tubular function and renin release.⁷ COX-2-derived prostanoids are important for Na excretion and therefore blood pressure regulation under normal healthy conditions as well as in CKD. COX expression in the kidney can be affected by dietary salt intake. Even though dogs have comparatively high basal COX-2 expression, the COX-2 pathway only becomes important in regulation of renal hemodynamics when hypovolemia and/or hyponatremia occur and RAAS is activated.⁷ COX-2 expression in the thick ascending loop of Henle can then triple in dogs, resulting in a 10-fold increase in plasma renin.¹¹

A variety of terms (nonselective, COX-2 selective, COX-2 specific, COX-2 preferential, COX-1 sparing, etc.) have been coined in an attempt to classify NSAIDs according to their ratio of COX activity, but standard use of those terms is lacking. COX-2 inhibitors have exhibited decreased gastrointestinal toxicity compared to nonselective NSAIDs; however, that advantage may be lost in vivo when NSAIDs are administered at recommended dosages.^12,13 Furthermore, renal impairment in dogs can occur in dogs after administration of preferential or nonselective NSAIDs.¹⁰

COX versus LOX

In addition to the expression of COX-2, increased levels of 5-LOX occurred in a study of canine coxofemoral OA.¹⁴ Dual inhibitors, NSAIDs that can inhibit both the COX and LOX pathways, may therefore provide additional beneficial effects in decreasing pain and inflammation. The 5-LOX pathway may have the most clinical significance in chronic inflammatory disease because an end product, leukotriene (LT)B₄, attracts leukocytes via chemotaxis.¹⁵

In addition to potentiating inflammation in OA, LTs can also cause renal impairment. Renal upregulation of LOX secondary to kidney injury increases production of those proinflammatory lipids. LTD₄ causes vasoconstriction of smooth muscle in the glomerular mesangium.¹⁶ Glomerular macrophages generate LTB₄, which is chemotactic for leukocytes. Activated leukocytes produce histamines, reactive O₂ species, and cytokines, further increasing glomerular injury. Glomerular function improved and proteinuria decreased by 50% when an indirect LOX inhibitor was administered to rats with experimental glomerulonephritis.¹⁷ Similar data is not available in dogs.

Tepoxalin is the only veterinary approved dual inhibitor NSAID. A dual inhibitor in phase III trials for approval in humans, ML-3000 (licofelone), decreased interleukin-1β and collagenase 1 synthesis, reducing experimental evidence of OA in a group of mongrel dogs treated for 8 wk. PGE₂ and LTB₄ production was also significantly decreased.¹⁸ It should be noted that there are no studies in veterinary medicine comparing the efficacy and ADE observed with COX inhibitors with that of dual inhibitors. Table 2 contains a list of NSAIDs approved for use in dogs.

TABLE 2 NDAIDs Approved for Use in Dogs

View Fullscreen

CKD and OA

CKD affects 0.5–1.5% of the canine population and is defined as structural or functional changes of the kidneys, usually present for at least 3 mo.^19,20 Both OA and CKD are more common in older dogs, so it's reasonable to assume that some subset of dogs with OA will also have subclinical [International Renal Interest Society stage I/early stage II] CKD. Even when CKD is a known diagnosis, the use of NSAIDs may be a clinical dilemma due to poor quality of life from OA. Control of pain associated with OA may require long-term treatment with NSAIDs. Although NSAIDs are often used for chronic management of OA, few long-term safety studies exist. A recent review of the safety and efficacy of long-term NSAID use in the treatment of canine OA identified 15 trials that evaluated treatment of 28 days or more in duration, with the longest study being 120 days.²¹ The evidence reviewed suggested an increased beneficial clinical effect with long-term use.²¹ More research to assess the effects of long-term NSAID administration in dogs would provide beneficial information on ADEs and could direct monitoring guidelines.

Renal Effects and Toxicity

The kidney is the organ with the second most number of ADEs from NSAIDs.² Most of those ADEs occur secondary to interference with renal hemodynamics and electrolyte balance due to decreased prostanoid synthesis. Within the kidney, decreased prostanoid synthesis commonly manifests as decreases in RBF and/or GFR and in severe cases, acute tubular injury that may lead to acute kidney injury (AKI). Anesthesia, even for elective procedures, may be associated with hypotension and/or hypovolemia, which can enhance the potential ADEs of NSAIDs on RBF. If NSAIDs are administered preoperatively for postoperative pain management, IV fluid support and blood pressure monitoring are recommended during anesthesia and recovery. AKI from NSAIDs is more likely to occur in dogs that already have decreased renal function.²² Maintenance of RBF becomes increasingly PG-dependent in dogs with CKD; therefore, decreased PG production secondary to the use of NSAIDs increases the risk of renal vasoconstriction.²³

Concurrent medication administration may also change renal hemodynamics. Dogs with CKD and concurrent hypertension and/or proteinuria are frequently treated with angiotensin-converting inhibitors (ACEis). ACEis not only inhibit the generation of ATII but also decrease the degradation of kinins like bradykinin.²⁴ Kinins exert their vasodilatory effects via PGs. Therefore, if NSAIDs are administered in combination with ACEis, the kinin/PG vasodilatory arm of the ACEi may be compromised. The effects of combined ACEi and NSAID treatment in dogs with CKD is largely unknown, although in one study, no changes in GFR or RBF were observed when tepoxalin and an ACEi were administered to healthy beagles for 28 days.²⁴

Potent diuretics like furosemide may enhance ADEs in dogs treated with NSAIDs. Intrarenal PGs play a major role in mediating the hemodynamic effects of furosemide in conscious dogs.²⁵ The renal effects of ibuprofen and carprofen have been investigated in euvolemic and volume-depleted healthy dogs.²⁶ Ibuprofen (a nonspecific NSAID) and carprofen (a COX-2 preferential NSAID) caused similar decreases in GFR in dogs that had also received furosemide, indicating that both nonspecific and preferential NSAIDs are capable of hemodynamic renal impairment in the face of volume depletion.²⁶ A follow up study was performed comparing the renal effects of carprofen and etodolac in euvolemic and volume-depleted healthy dogs.²⁷ Dogs that received either NSAID in combination with furosemide experienced an increase in creatinine and decrease in GFR that was reversible when treatment was discontinued. Renal plasma flow (RPF), the volume of plasma reaching the kidneys/unit time, was preserved. A decrease in GFR without a decrease in RPF suggested preglomerular vasoconstriction and a postglomerular reduction in vascular resistance.²⁷

Renal effects of furosemide and NSAIDs have been evaluated in rodents and humans as well. The diuretic effect of furosemide was neutralized by rofecoxib in rats, and renal cortical COX-2 increased significantly in rats treated with rofecoxib compared with untreated controls.²⁸ In another study, COX-1 expression was decreased in rats treated with both diclofenac and the combination of diclofenac and furosemide.²⁹ COX-2 expression was increased in rats treated with either diclofenac or furosemide as well as with a combination of diclofenac and furosemide.²⁹ The clinical concern of using furosemide in combination with an NSAID is an additive risk of fluid and electrolyte imbalances superimposed on decreased production of PGs, which are necessary to counter renal vasoconstriction. Alterations in Na and potassium as well as hypotension secondary to hypovolemia from a diuretic effect can result in decreased renal perfusion. As an example, the use of furosemide in combination with indomethacin in neonatal infants with patent ductus arteriosus increased the incidence of AKI.³⁰

AKI from NSAIDs is more likely to occur in dogs that already have decreased renal function.²² Many older dogs with OA also have other conditions that could predispose them to NSAID ADEs, such as liver disease, cardiac disease, or neoplasms in addition to CKD. Decreased NSAID elimination could occur with liver disease, increasing the possibility of ADEs. Cardiac and liver disease could result in either decreased effective circulating volume or activation of the RAAS. Because NSAIDs are highly protein-bound, their half-lives could be decreased in hypoalbuminemic states and liver or kidney disease. Other concurrent conditions, such as decreased metabolic rate and altered volumes of distribution, are risk factors for NSAID toxicity in elderly humans and may have a role in dogs as well.⁹

NSAIDs are commonly thought to be only indirectly nephrotoxic. Reversible hemodynamic changes are the most common renal effects of NSAIDs, but structural changes to the kidney can also occur. AKI, interstitial nephritis, and renal papillary necrosis are all renal effects of NSAIDs that have been reported in dogs.¹⁰ NSAIDs most commonly affect the proximal tubules, although the collecting ducts may also be susceptible to NSAID-induced nephrotoxicity. The mechanism is unclear, but long-term NSAID exposure may cause toxicity to the collecting ducts through either increased osmolality of the tubular fluid or further decreases to the already scant medullary blood flow.³¹ At excessively high NSAID doses, drug accumulation may also have a direct toxic effect in the kidney, as in renal papillary necrosis.

Clinical Safety Studies

ADEs of veterinary NSAIDs in the literature are commonly associated with high doses and/or prolonged administration. When deracoxib was administered to dogs [10 dogs/group, 2 mg/kg q 24 hr (i.e., labeled dosage) and 4 mg/kg q 24 hr for 6 mo], no adverse clinical effects were noted; however, GFR was not measured. When administered to dogs at 6 mg/kg q 24 hr (three times the label dose) for 6 mo, 2 dogs developed hyposthenuria. Increases in blood urea nitrogen and dose-dependent renal tubular degeneration occurred with doses of 6 (n = 2), 8 (n = 2), and 10 (n = 4) mg/kg q 24 hr. Renal papillary necrosis developed at 6 mo in 1 dog receiving 8 mg/kg q 24 hr and in 3 dogs receiving 10 mg/kg q 24 hr.³²

In a placebo-controlled study, ketoprofen was administered at 1 mg/kg to five clinically healthy beagles for 30 days (the labeled dose was 1 mg/kg daily for up to 5 days).^33,34 No significant differences were observed in either RBF or GFR between pre- and post-NSAID treatment; however, one dog in the ketoprofen group was below the reference range for RBF at 20 and 30 days and developed mild to moderate renal proteinuria and urine sediment abnormalities. Renal tubular epithelial cells (2–3/high-power field) were present on urine sediment exam. Two dogs in the ketoprofen group also had increased urinary N-acetyl-β-D-glucosaminidase (NAG) and/or gamma-glutamyl transpeptidase (GGT) excretion. One of those dogs showed increased urine NAG and GGT excretion between days 6 and 18, while the other dog only had increased urine GGT excretion at day 30. At necropsy, those same two dogs had mild lymphoid cell infiltration in the renal medulla.³³ In another ketoprofen study, the effects of a low dose (0.25 mg/kg per os given once daily for 30 days) on urinary enzyme excretion was assessed. No increase in either urinary NAG or GGT occurred suggesting renal tubular cell injury did not occur at that dose; however, histopathology was not performed to corroborate the laboratory findings.³⁵

One study in dogs with both OA and International Renal Interest Society stage 2 or 3 CKD administered tepoxalin for up to 7 mo found no change in serum biochemical analysis, urinalysis, urine protein/creatinine ratio, urine GGT/creatinine ratio, iohexol plasma clearance, and indirect blood pressure measurement in dogs completing the study.³⁶ ADEs resulting in discontinuation of tepoxalin and/or withdrawal from the study included increased serum creatinine concentration (one dog in week 1), collapse (one dog in week 1), increased liver enzyme activities (one dog in week 4), vomiting and diarrhea (one dog in week 8), hematochezia (one dog in week 24), and gastrointestinal ulceration and perforation (one dog in week 26).³⁶ Some of the dogs that experienced ADEs had pre-existing medical conditions and/or were receiving other medications in addition to tepoxalin during the study period.

Conclusion

Using the lowest effective dose of a veterinary-approved NSAID to control pain and improve mobility is recommended for all older dogs necessitating NSAID treatment, but especially for dogs with concurrent health problems such as CKD. Alternate forms of OA management that have fewer potential ADEs on the kidney should be employed first in dogs with CKD. Opioids, milk protein, chondroitin sulfate, glycosaminoglycans, gabapentin, amantadine, acupuncture, omega-3 fatty acid supplementation, maintaining an ideal body condition, and routine, moderate exercise may improve the quality of life in dogs with OA without adversely affecting kidney function. In a recent study, dogs with OA that were fed a diet supplemented with omega-3 fatty acids were able to tolerate more rapid reductions in NSAID dose without adversely affecting quality of life as compared to arthritic dogs fed a control diet.³⁷

In some dogs with CKD, OA may so adversely affect the quality of the dog's life that NSAIDs are necessary. In those cases, it is imperative to avoid additional risk factors like hypotension, dehydration, anesthesia, furosemide, and other drugs with potential adverse renal effects (e.g., aminoglycosides). In addition, baseline evaluation of blood pressure, hematocrit, and renal and hepatic parameters are recommended prior to prescribing an NSAID in all dogs. Repeat evaluation of laboratory parameters 2 wk after initiating treatment with periodic monitoring during treatment is recommended.³⁸ Evaluation of urine sediment, enzymuria, and GFR is recommended to detect changes in renal function and/or tubular damage prior to changes in serum creatinine concentration. Re-evaluation of standard clinicopathological parameters after 1 mo of NSAID administration should also be considered because some ADEs (e.g., hepatocellular injury) can be clinically silent.³⁶ If all parameters are stable, the patient should be evaluated q 3 mo to evaluate any clinical changes and to monitor CKD progression.

A systematic review of NSAID-induced ADEs in dogs was recently published.³⁹ ADEs from NSAIDs are more likely to occur in the first 14–30 days of administration but have been reported from 3 to 182 days.^2,36 Because it is impossible to predict which animals will experience an ADE, the owner of every animal receiving NSAIDs should be educated regarding NSAID ADEs such as vomiting, diarrhea, inappetence, and dark stools. The administration of nonveterinary-approved NSAIDs is not recommended in dogs due to increased elimination times and an extremely narrow margin of safety. By practicing vigilance, the quality of life for dogs with severe OA can be improved without overlooking the warning signs that could lead to more serious problems.