NSAIDs: Comparative Toxicity and Drug Interactions
World Small Animal Veterinary Association World Congress Proceedings, 2013
Lauren A. Trepanier, DVM, PhD, DACVIM, DACVCP
University of Wisconsin-Madison, Madison, WI, USA

NSAID Mechanisms of Action

Cyclooxygenase 1 (COX-1) is constitutively expressed in many tissues and generates protective prostaglandins in the stomach, intestine, and kidney. COX-1 also generates thromboxane (TXA2), which mediates platelet aggregation. Cyclooxygenase 2 (COX-2) is induced by inflammation and generates proinflammatory prostaglandins. However, COX-2 also generates protective renal prostaglandins and is important for the healing of gastric ulcers once they occur. NSAIDs are antiinflammatory, antipyretic, and analgesic via COX-2 inhibition and have antiplatelet activity via COX-1 inhibition. For clinical decision-making, NSAIDs can be divided into three major groups: classical nonselective agents, newer COX-2 preferential drugs, and even newer COX-2 selective coxibs.

Nonselective NSAIDs

Both COX-1 and COX-2

Low doses selectively inhibit platelet function (0.5–1 mg/kg q 12 h)
May be useful for prothrombotic states (hypertrophic cardiomyopathy, protein-losing nephropathy)
GI upset and bleeding at higher dosages

Both COX-1 and COX-2

Short-term analgesia
Has been associated with bleeding and azotemia
There are safer agents

Both COX-1 and COX-2

Effective to slow growth of transitional cell, nasal, and colon carcinomas in dogs and cats
Avoid if coagulopathy present

COX-2 preferential agents

Carprofen (Rimadyl)
COX-2 > COX-1

Good pain control with minimal risk of bleeding
Risk of renal decompensation remains
Rare risk of idiosyncratic hepatic necrosis

Meloxicam (Metacam)
COX-2 > COX-1

Good pain control with minimal risk of bleeding
Risk of renal decompensation remains
Acute renal failure with chronic dosing at label dose in cats

Etodolac (EtoGesic)
COX-2 > COX-1

Excessive bleeding in dogs during experimental surgery (EtoGesic label)
Fall in serum T4 in dogs with orthopedic disease, sometimes into the hypothyroid range (Ness 2003)

COX-2 selective agents

Deracoxib (Deramaxx)
Firocoxib (Previcox)
COX-2 >>> COX-1

Apparent low risk of new GI ulceration
Can impair healing of preexisting ulcers
Minimal risk of bleeding
Risk of renal decompensation remains

Other NSAIDs

Tepoxalin (Zubrin)
Both COX-1 and COX-2

Safer than coxibs for dogs with preexisting ulcers? Also inhibits 5-lipoxygenase
Unfortunately, manufacture has been stopped

Weak, reversible COX inhibition

No significant antiinflammatory effects
Antipyretic or analgesic in dogs intolerant of NSAIDs (acetaminophen 10-15 mg/kg q 8 h)

GI Toxicity of Available NSAIDs

Vomiting and diarrhea are common early side effects, due to direct gastric and duodenal irritation. NSAIDs directly alter phospholipids in the mucus gel layer overlying the gastric mucosa and can damage this hydrophobic barrier (this is the rationale for enteric-coated aspirin). Gastric ulceration, which is more serious, results from inhibition of PGE2 generation. PGE2 is important for maintenance of healthy gastric mucosa (epithelial turnover, mucus and bicarbonate secretion).

It is difficult to directly compare the relative GI toxicity of veterinary NSAIDs; most published studies using the gold standard (endoscopy) enrol relatively few dogs, and they are usually young and healthy. This may not translate to GI risk in geriatric or critically ill patients. We do know that there is a lower risk of GI upset and ulceration with COX-2 preferential compared to nonselective agents. For example, carprofen and etodolac lead to less GI ulceration than aspirin in dogs (Reimer 1999), and carprofen is associated with fewer and milder gastric lesions compared to ketoprofen in dogs (Forsyth 1998).

Coxibs have a better safety profile than other NSAIDs in humans. This may be true in dogs as well, but veterinary coxibs have not been compared directly to COX-2 preferential NSAIDs. GI ulceration and perforation have been reported with deracoxib in dogs. COX-2 is important for healing of gastric ulcers experimentally, and coxibs may impair healing of preexisting ulcers. Therefore, coxibs should never be directly substituted for other NSAIDs in animals with GI upset, without a washout period to allow any subclinical ulcers to heal. Ulcer risk from NSAIDs is also potentiated by glucocorticoids, which inhibit prostaglandin synthase, and decrease peroxidase-mediated scavenging of free-radical precursors. The use of multiple concurrent NSAIDs, or the combination of NSAIDs with glucocorticoids, is therefore contraindicated. While the optimal period for NSAID washout has not been established, the goal is to wait at least 4–5 NSAID elimination half-lives and allow time for gastric mucosal turnover. I typically wait one week for an NSAID to glucocorticoid transition, and add omeprazole during this period to promote healing of any subclinical ulcerations.

Monitoring for GI Bleeding from NSAIDs

Clinical monitoring for GI bleeding should include owner vigilance for vomiting, diarrhea, lethargy, and inappetence. Darkened stools are a late finding; substantial chronic GI bleeding can be present without overt melena. Routine blood work is also important in at-risk older patients. CBC may show polychromasia with a drop in PCV and TP acutely, or microcytosis with thrombocytosis (both signs of iron deficiency) chronically. On biochemical panel, low albumin, low globulin, and increased BUN (with no increase in creatinine) also suggest GI bleeding. Fecal occult blood can detect GI bleeding before overt melena, but available tests lack high sensitivity and specificity in dogs and cats and are not recommended.

Gastric acid is necessary for the development of ulcers from NSAIDs. In humans, pump blockers such as omeprazole protect against the development of NSAID ulcers. This has not yet been established in dogs but is a reasonable approach in at-risk animals until more is known. Misoprostol, the synthetic PGE2 analog, has proven efficacy in preventing aspirin-induced ulcers in dogs. The main misoprostol drawback is diarrhea and cramping.

Renal Decompensation

Any NSAID can adversely affect renal function. Prostaglandins increase renal arterial blood flow in response to a drop in renal perfusion, and stimulate renin release. Both COX-1 and COX-2 generate protective renal prostaglandins, and COX-2 selective agents can drop glomerular filtration to the same extent as classical NSAIDs. The risk of renal decompensation from NSAIDs is greatest with preexisting renal disease, hypovolemia or dehydration, congestive heart failure, sodium restriction, or cirrhosis.

Most studies of renal function with NSAIDs have been done in healthy animals undergoing elective procedures. Carprofen at the label dose showed no adverse effect on glomerular filtration rate (GFR) in one study; however, in another study, carprofen and ketoprofen both led to a decrease in GFR in dogs undergoing castration (Forsyth 2000). In addition, ketoprofen has been associated with transient azotemia, even in healthy dogs being spayed (Lobetti 2000).

Meloxicam has no adverse effects on GFR in healthy euvolemic cats (Goodman 2009), and, unlike other NSAIDs, clearance of meloxicam in cats is not slower than in dogs. However, meloxicam has been associated with acute renal failure and death in client-owned cats given the label dose (0.3 mg/kg SC) chronically. Lower daily dosages (0.01–0.03 mg/kg daily) were well tolerated in 46 geriatric cats with osteoarthritis, but renal function was not consistently monitored (Gunew 2008). Older cats with IRIS stage 1–2 renal disease, treated with meloxicam at a median dose of 0.02 mg/kg/day, did not show clinical evidence of renal decompensation when followed for at least 6 months (Gowan 2011). However, this chronic use of meloxicam is specifically contraindicated in cats in the U.S.

NSAID Use with Preexisting Azotemia

NSAIDs should be avoided, whenever possible, in any patient with preexisting renal azotemia. Consider alternative analgesics to include fentanyl CRI or patch, buprenorphine, acetaminophen, tramadol and/or gabapentin. If NSAIDs are necessary in azotemic animals, use conservative doses and add fluid support.

Platelet Dysfunction and Bleeding

Classical NSAIDs (aspirin, ketoprofen, piroxicam) inhibit platelet function most readily, via inhibition of COX-1 mediated TXA2 generation. Prolonged bleeding times were reported with ketoprofen in dogs undergoing elective orthopedic surgery (Grisneaux 1999); one ketoprofen-treated dog developed a hematoma at the surgical site. COX-2 preferential agents have less platelet inhibitory effects; while carprofen leads to mild subclinical decreases in platelet aggregation, neither carprofen nor meloxicam prolong buccal mucosal bleeding times in healthy dogs. However, etodolac has been associated with excessive bleeding in dogs during experimental surgery (Etodolac label). COX-2 selective coxibs do not affect buccal mucosal bleeding time in dogs (Deramaxx label; Steagall 2007), and coxibs are probably the safest NSAIDs in dogs with preexisting coagulopathies (e.g., von Willebrand's disease).

Coxibs and Hypercoagulable Patients

Conversely, coxibs may be a poor choice in hypercoagulable patients (protein-losing nephropathy, immune-mediated hemolytic anemia, vasculitis, Cushing's). Unopposed activity of COX-1 may lead to platelet overreactivity and impaired small vessel dilatation, resulting in thrombosis in at-risk patients. I would avoid coxibs in these patients until more is known.

Drug Interactions with NSAIDs


NSAID interaction

Possible outcome


Protein-binding displacement of NSAID or benzodiazepine

Enhanced sedation from benzodiazepine?


P450 induction, enhanced bioactivation of acetaminophen

Increased hepatotoxicity of acetaminophen


Volume contraction; NSAIDs inhibit compensatory renal prostaglandin production

Enhanced nephrotoxicity


NSAIDs inhibit compensatory renal prostaglandin production

Enhanced nephrotoxicity


Some NSAIDs can impair potassium excretion

Possible hyperkalemia

Meadowsweet Willow

Salicylate-containing herbs

Enhanced bleeding risk with aspirin?

Gingko, garlic, ginger, ginseng

Inhibit platelet aggregation

Enhanced bleeding risk with aspirin (reported for gingko)

Minimizing NSAID Complications

 Screening CBC and biochem panel in older patients

 Maintain hydration

 Absolutely no concurrent glucocorticoids

 Consider at least one-week washout between NSAIDs (can bridge with tramadol and omeprazole)

 Client education and careful clinical monitoring for GI upset, darkened stools, inappetence, or lethargy in any treated patient

 Periodic monitoring of CBC and biochem panel in older patients, or those with underlying risk factors for GI bleeding, such as hepatic disease or early renal insufficiency

 Be aware of potential drug-NSAID or herb-NSAID interactions


Speaker Information
(click the speaker's name to view other papers and abstracts submitted by this speaker)

Lauren A. Trepanier, DVM, PhD, DACVIM, DACVCP
University of Wisconsin-Madison
Madison, WI, USA

MAIN : Pharmacology : NSAIDs Comparative Toxicity
Powered By VIN