Puppies and kittens looking over white wall


How it works

You can read or listen to our modules. There is a play button in each heading so you can listen while you read. Or you can listen to the whole module in one go or download it as a podcast. You'll find a play and download button for the whole module at the bottom of this page. At the end of the module there is a quiz, so you can test your knowledge and receive a CPD certificate. You will have to hit the play button on the quiz video and enter your name and email address before you start the quiz.

In acute or chronic pain in cats and dogs, NSAIDs are the first-line treatments unless there is a clear contraindication. However, there are seven different NSAIDs marketed in the UK for use in dogs, four of which are also for cats. This module looks at whether there are any differences between the NSAIDs in terms of efficacy or safety.

By doing this module you will:

  • know what an NSAID is

  • understand the clinical pharmacology of NSAIDs

  • gain a broad understanding of the range of licensed uses of the different NSAIDs

  • understand what is known about the comparative efficacy and safety of NSAIDs

  • be able to make a rational choice.

Non-steroidal anti-inflammatory drug, or NSAID, is a term conventionally used for a heterogeneous group of organic acids (see table) that have analgesic, antipyretic, anti-inflammatory, and platelet inhibitory actions (Myeler 2015). 

NSAIDs by chemical class – those licensed for use in dogs or cats are in bold type


* Phenylbutazone and tiaprofenic acid are licensed for dogs but not marketed in the UK. Piroxicam is licensed for use in humans and is available as an unlicensed special for companion animals; it has been used in veterinary medicine to treat certain tumours that express COX-receptors (BSAVA Formulary 2017)

Paracetamol (an analgesic and antipyretic) is not an NSAID, although its analgesic effects appear to involve inhibition of COX-3 in the central nervous system (see below) (Chandrasekharan et al 2002). Pentosan polysulphate and grapiprant (a newly licensed EP-4 inhibitor, not yet marketed) are not NSAIDs; these are classed as ‘other antirheumatic and anti-inflammatory agents (non-steroid)’ (WHO 2018).

It is widely accepted that the main mechanism of action underlying the therapeutic and adverse effects of NSAIDs is inhibition of the enzyme cyclooxygenase (COX). When cell membranes are damaged, short chain fatty acids (such as arachidonic acid) are released. These are substrates for COX and lipoxygenase enzymes, leading to the production of inflammatory mediators such as prostaglandins, thromboxanes and leukotrienes. Prostaglandins enhance nociceptive stimuli produced by other inflammatory mediators and act in synergy with histamine and bradykinin to increase vessel permeability. NSAIDs have central as well as peripheral analgesic and anti-inflammatory actions (Lees et al. 2004).

In 1991, COX was discovered to exist in the body in at least two isoforms: COX-1 and COX-2. COX-1 is present in most, possibly all, cell types except red blood cells. It is generally considered to be associated with normal body functions (e.g. producing protective prostaglandins in the gastrointestinal tract and kidney). COX-1 is also involved in blood clotting, causing the synthesis in platelets of thromboxane A2, which has potent pro-thrombotic effects. In contrast, COX-2 is an inducible enzyme that is formed at sites of inflammation, resulting in pro-inflammatory prostaglandins. 

The analgesic and anti-inflammatory actions of NSAIDs were originally believed to depend mainly on their inhibition of COX-2, and the unwanted gastrointestinal and renal effects of NSAIDs on their inhibition of COX-1. It followed, in theory, that a drug that inhibits COX-2 at a lower concentration than is needed to inhibit COX-1 might be safer, and so NSAIDs that are more selective for COX-2 have been developed with the aim of minimising toxicity. However, COX selectivity can depend on the dose of the drug and animal species, and different methods for testing and calculating selectivity can give different values for any given NSAID (Lees et al. 2004). Therefore, quoted COX selectivity data may not translate into the expected clinical effects (either beneficial or harmful). Furthermore the terms used to describe inhibition of COX-2 relative to COX-1 (e.g. COX selective, COX preferential) are not standardised. Of the veterinary NSAIDs, carprofen, cimicoxib, firocoxib and robenacoxib are described in the summaries of product characteristics (SPCs) as COX-2-selective, while mavacoxib is described as COX-2 preferential and meloxicam is said to inhibit COX-2 to a greater extent that COX-1. 

In the light of more recent research on COX, the concept of COX-1 as ‘good’ and COX-2 as ‘bad’ is now known to be an oversimplification. For instance, as well as its pro-inflammatory actions, COX-2 also appears to contribute to the production of anti-inflammatory mediators: its activity results in inflammatory effects in the acute phase of injury and anti-inflammatory effects in the resolution phase (Lees et al 2004); however the relevance of this to the effects of NSAIDs is not clear. Also, a third COX isoform – COX-3 has been identified in the CNS of dogs (Chandrasekharan et al 2002). NSAIDs are also being found to have actions other than COX inhibition. For example they may modulate inflammatory gene transcription via actions on the pro-inflammatory transcription factor nuclear factor kappa B (Bryant et al 2003). These actions may vary between different NSAIDs.

Most NSAIDs are well absorbed after oral administration and for many the SPCs recommend giving oral NSAIDs with food. The advice for robenacoxib is to give it with no, or a small amount of, food because the amount absorbed is reduced if given with the total daily food ration (Onsior SPC). Some NSAIDs can be given by injection (usually subcutaneous or intravenous – check the relevant SPC). 

NSAIDs are highly protein bound, which helps distribution to sites of inflammation where NSAIDs act. Plasma half-life varies between drugs and may also vary between individual animals (see table). However, the biological effects of NSAIDs may not be directly related to drug plasma concentrations (Lascelles et al. 1998) and most NSAIDs are given as a once daily dose in cats and dogs (except mavacoxib, which is given monthly because of its very long half life). However, while plasma concentrations are widely reported for the different NSAIDs, tissue concentrations are not reported for all, and so whether there are clinically relevant differences between NSAIDs in terms of their persistence in tissues is not clear.  Typically, NSAIDs are metabolised by the liver to inactive compounds, with only small amounts of the parent drug excreted in the urine (Lees et al. 2004).

Elimination half-lives of NSAIDs

Table of elimination half lives of NSAIDs

Experimental and clinical studies suggest that NSAIDs are most effective in preventing acute pain if given pre-emptively (Veiga et al 2004; Lascelles et al. 1998). In dogs undergoing ovariohysterectomy, carprofen given preoperatively had a greater analgesic effect in the early postoperative period (2 hours after extubation) than when the drug was given postoperatively (Lascelles et al. 1998) 

Trial results also suggest that an NSAID alone will not produce optimal control of pain in all cases (Friton et al 2017). This is illustrated by the results of a randomised controlled trial in dogs undergoing soft tissue surgery that received an NSAID (or placebo) before surgery and for 2 days after surgery (they also received pre-operative butorphanol, a short-acting opioid) (Friton et al 2017). Analgesia was inadequate in 26% of dogs on an NSAID (vs. 42% with placebo). This evidence suggests that multimodal therapy may often be needed for optimal control of peri-operative pain. 

There is no convincing evidence from randomised controlled trials that any NSAID is more effective than any other in managing acute or peri-operative pain (Schmid et al 2010; Gruet et al 2013;  Veterinary Prescriber 2014; BestBetsforVets 2016; Speranza et al 2015). 

An NSAID is usually first-line drug therapy for dogs and cats with chronic pain with an inflammatory cause, as part of a multi-modal approach. As well as being involved in peripheral pain and inflammation, COX enzymes play a role in the process of central sensitisation, whereby pain initiated peripherally (e.g. in the joints) results in central nervous system hyperexcitability that with time becomes self-sustaining and persists despite elimination of the peripheral stimulus (Samad et al. 2001; Innes et al 2010). NSAIDs can inhibit this process (Veiga et al 2004). It is also thought that central sensitisation may worsen joint pathology and that reduction of the central sensitisation could help prevent worsening of the disease (Innes et al 2010). However, therapy with an NSAID alone may become ineffective in more advanced disease. 

The authors of a systematic review aimed to assess the efficacy of long-term NSAID therapy (i.e. lasting at least 28 days) in dogs (Innes et al 2010). They included 15 published trials (7 randomised controlled trials; involving a total of 1,595 dogs) but they were unable to estimate the efficacy of long-term NSAID therapy, because of a lack of data. However, in one short-term trial (lasting 14 days) it was possible to estimate that 5 dogs (95% confidence interval 2–7 dogs) need to be treated for 14 days to see a benefit in one dog (in this case the benefit was a positive response on force-plate evaluation). This calculation is known as the number needed to treat (NNT); it is the inverse of the absolute risk difference. The closer the NNT is to 1, the more effective the treatment. 

No NSAID has been shown to be any more effective than another in relieving chronic pain (Aragon et al 2007; Sanderson et al 2009; Innes et al 2010; Reymond 2012; Payne-Johnson 2015). Unfortunately, most published comparative trials are unhelpful in telling us whether one drug gives better pain relief than another because they aim to show that a drug is only non-inferior to another. (Read more about non-inferiority trials here). 

Study results suggest that there is progressive pain relief with longer-term NSAID therapy (Innes et al 2010; Reymond et al 2012) and so it is worth continuing an NSAID for around 4 weeks, as long as it is well tolerated, before assessing the benefit. Unfortunately some SPCs advise discontinuation if improvement is not seen after a relatively short period (e.g. 10 days)  [Metacam SPC])

A cat or dog that fails to respond to one NSAID may respond to another. The reason for this is not known. One hypothesis is that it may be due to the presence of several isoforms of COX 1 and COX 2 (Warner & Mitchell 2002).

Most of the information about the adverse effects of NSAIDs comes from clinical trials and pharmacovigilance reports. 

For dogs, data on the adverse effects of 14 different NSAIDs reported across 64 different studies were analysed in a systematic review (Monteiro-Steagall et al 2013). The most common outwardly detectable effects were: vomiting, diarrhoea, anorexia, lethargy and melaena. It was not possible to calculate the rate of individual effects overall, but they ranged from 0% to 38%. The limitations of information from clinical trials is that they do not usually include enough animals to reveal rarer adverse effects, and they usually exclude animals with concurrent diseases (e.g. renal dysfunction) and so may not be a reliable indication of adverse effects seen in clinical practice. The adverse effects of NSAIDs in cats have been less well studied than in dogs. One randomised placebo-controlled trial involving 194 cats with osteoarthritis (including 40 with chronic kidney disease) assessed the safety of 1 months’ treatment with robenacoxib (King et al 2016). It found that vomiting was the most commonly-reported effect (in 20% vs. 21% with placebo), but there was no difference between robenacoxib and placebo in the rate of any adverse effects. 

In an analysis of adverse effects to NSAIDs spontaneously reported to the Veterinary Medicines Directorate, the effects most frequently reported for oral NSAIDs in dogs were vomiting, death, anorexia, lethargy, diarrhoea and hepatopathy; those in cats were vomiting, anorexia, lethargy, death, renal insufficiency and dehydration (Hunt et al. 2015). Overall, the frequency of reported adverse effects to NSAIDs was considered by the authors to be low (when related to the total number of doses sold) and similar for the different NSAIDs. Vomiting, renal insufficiency and death were more common with injectable compared with oral NSAIDs, which is likely to be related to their peri-operative use. In dogs, the analysis found that vomiting, lethargy and death were reported more frequently with ‘coxibs’ than ‘non-coxibs’. The report’s authors judged that this difference is likely to be due to coxibs being newer drugs, for which adverse-effect reporting rates are higher compared to older drugs. One study found that veterinary surgeons perceive coxibs to be safer than non-coxibs (Belshaw et al 2016), and so it is conceivable that they were being used in patients more at risk of adverse effects. The limitations of pharmacovigilance data include the likelihood of under-reporting (especially of known adverse effects), and of serious effects being more likely to be reported than non-serious effects.

All adverse effects to NSAIDs should be reported to the VMD and/or the marketing company. 

An overview of the licensed indications of NSAIDs in the UK*

An overview of the licensed indications of NSAIDs in the UK

Source: VMD product information database. *There may be differences in indication between different brands of the same NSAID. Check the individual SPC.

Choice is partly determined by licensed indication, which varies between NSAIDs (see the table). Choice might also be affected by palatability or ease of use or cost. For the management of chronic pain in dogs, there is a wide choice of oral NSAIDs available as tablets (including chewable and flavoured versions) and oral suspension, for dosing daily or monthly (mavacoxib only). Plasma half-life is an important consideration when treating performance animals to avoid positive drug tests. In such cases it is best to get advice on NSAID choice from the relevant regulator (e.g. the Greyhound Board of Great Britain for racing greyhounds). It is worth stocking, and becoming familiar with, several different NSAIDs so there is an option for switching if needed. For cats, however, there is a limited choice, with only two oral NSAIDs licensed for chronic use.

Published clinical trial evidence does not show any one NSAID to be more effective or safer than any other in the management of acute, surgical or chronic pain. However, individual dogs and cats may respond to one NSAID but not another. It is worth becoming familiar with several different NSAIDs so there is an option for switching if necessary. Licensed indication and compliance are the main factors affecting choice. The choice for cats is limited. Knowledge about the incidence of the adverse effects in dogs and cats is relatively poor. All adverse effects of NSAIDs should be reported. 


If you prefer, you can listen to the whole audio presentation of this module using the following podcast. Don't forget that you can also download the podcast to your iPod, music player, tablet or smartphone using the Download link on the right of the audio player.



How we produced this module

Our modules start with a detailed outline and electronic literature search. We commission a collaborating author, who is a specialist in the module topic, to write a draft module. The collaborating author on this module was Matthew Gurney. The draft is circulated unsigned to a wide range of commentators, include practising first-opinion vets, other topic specialists, the companies that market any mentioned drugs and other organisations and individuals, as appropriate. They can raise points about the interpretation of evidence, ask questions that are important to clinical practice, and present alternative viewpoints. There is a rigorous editing and checking process and the result is a module that is evidence-based, impartial and relevant to clinical practice. The final module is unsigned because it is the result of collaboration. 


Post-operative pain in cats – meloxicam or robenacoxib? Veterinary Prescriber, May 2013. 

Aragon CL et al. Systematic review of clinical trials of treatments for osteoarthritis in dogs. JAVMA 2007; 230: 514–21.

Belshaw Z et al. The attitudes of owners and veterinary professional in the United Kingdom to the risk of adverse events associated with using non-steroidal anti-inflammatory drugs (NSAIDs) to treat dogs with osteoarthritis. Prev Vet Med. 2016; 131: 121–6.

BestBetsforVets. Robenacoxib for acute musculoskeletal pain control in cats. October 2016. Available: http://bestbetsforvets.org/bet/305.

Bryant CE et al. Evaluation of the ability of carprofen and flunixin meglumine to inhibit activation of nuclear factor kappa B. Am J Vet Res 2003; 64: 211–5.

Chandrasekharan NV et al. COX-3, a cyclooxygenase-1 variant inhibited by acetaminophen and other analgesic/antipyretic drugs: cloning, structure, and expression. Proc Nat Acad Sci of United States America, 2002; 99: 13926–31. 

Friton GT et al. Efficacy and safety of Injectable robenacoxib for the treatment of pain associated with soft tissue surgery in dogs. J Vet Intern Med 2017 31; 832–41. 

Gruet P et al. Robenacoxib versus meloxicam for the management of pain and inflammation associated with soft tissue surgery in dogs: a randomized, non-inferiority clinical trial. BMC Veterinary Research 2013; 9: 92.

Hunt JR et al. An analysis of the relative frequencies of reported adverse events associated with NSAID administration in dogs and cats in the United Kingdom. Veterinary J 2015; 206: 183–90. 

Innes JF et al. Review of the safety and efficacy of long-term NSAID use in the treatment of canine osteoarthritis. Vet Rec 2010; 166: 226–30. 

Jovanovic DV et al. Nitric oxide induced cell death in human osteoarthritic synoviocytes is mediated by tyrosine kinase activation and hydrogen peroxide and/or superoxide formation. J Rheumatol 2002; 29: 2165-75.

King JN et al. Clinical safety of robenacoxib in feline osteoarthritis: results of a randomized, blinded, placebo-controlled clinical trial. J Fel Med Surg 2016; 18: 632–42. 

Lascelles D et al. Guidelines for safe and effective use of NSAIDs in dogs. Vet Therap: Res Appl Vet Med 2005 6: 237–51.

Lascelles BDX et al. Efficacy and kinetics of carprofen, administered preoperatively or postoperatively, for the prevention of pain in dogs undergoing ovariohysterectomy. Vet Surg 1998; 27: 568–82.

Lees P et al. Pharmacodynamics and pharmacokinetics of nonsteroidal anti-inflammatory drugs in species of veterinary interest. J Vet Pharm Ther 2004; 27: 479–90.

Monteiro-Steagall et al. Systematic review of nonsteroidal anti-inflammatory drug-induced adverse effects in dogs. J Vet Intern Med 2013; 27: 1011–9. 

Myeler’s side effects of drugs. Ed Aronson J. 16th edition, 2015. 

Payne-Johnson M et al. Comparative efficacy and safety of mavacoxib and carprofen in the treatment of canine osteoarthritis. Vet Rec 2015; 176: 284–9.

Reymond N et al. Robenacoxib vs. carprofen for the treatment of canine osteoarthritis; a randomized, noninferiority clinical trial. J Vet Pharmacol Ther 2012; 35: 175–83.

Samad TA et al. Interleukin-1beta-mediated induction of Cox-2 in the CNS contributes to inflammatory pain hypersensitivity. Nature 2001; 410: 471–5. 

Sanderson RO et al. Systematic review of the management of canine osteoarthritis. Vet Rec 2009; 164: 418-24.

Schmid VB et al. Analgesic and anti-inflammatory actions of robenacoxib in acute joint inflammation in dog. J Vet Pharmacol Ther 2009; 33: 118–31. 

Sluka KA et al. Reduction in joint swelling and hyperalgesia following post-treatment with a non-NMDA glutamate receptor antagonist. Pain 1994; 59: 95–100.

Speranza C et al. Robenacoxib versus meloxicam for the control of peri-operative pain and inflammation associated with orthopaedic surgery in cats: a randomized clinical trial. BMC Veterinary Research 2015; 11: 79.

Veiga APC et al. Prevention by celecoxib of secondary hyperalgesia induced by formalin in rats. Life Sci 2004; 75: 2807–17. 

Veterinary Prescriber. Post-operative pain in cats: meloxicam or robenacoxib? March 2014.

Warner TD, Mitchell JA. Cyclooxygenase-3 (COX-3): filling in the gaps toward a COX continuum? Proc Nat Acad Sci USA 2002; 99: 13371–3. 

WHO Collaborating Centre for Drug Statistics Methodology. ATCvet Index 2018. [Accessed 1 August 2018]