How much do we know about the safety of NSAID therapy in dogs?

Non-steroidal anti-inflammatory drugs (NSAIDs) are used widely to relieve pain, with or without inflammation, in dogs with acute and chronic musculoskeletal problems, and to relieve pain associated with surgery. In the UK, there are ten  different NSAIDs (in the form of at least 30 different formulations) authorised for use in dogs. All NSAIDs inhibit the enzyme cyclo-oxygenase (COX), which cells use in the synthesis of prostaglandins; some of these have physiological functions and some are synthesised in conditions where they contribute to pain and inflammation. COX exists in at least two isoforms: COX-1 and COX-2. COX-1 is generally considered to be the enzyme associated with normal body functions (e.g. involved in producing prostaglandins in the gastrointestinal tract and the kidneys), while COX-2 is induced in the presence of inflammation.  Consequently, the analgesic and anti-inflammatory actions of NSAIDs are believed to depend mainly on their inhibition of COX-2, and the well-known unwanted gastrointestinal effects of NSAIDs on their inhibition of COX-1. NSAIDs that are more selective for COX-2 have been developed with the aim of minimising gastrointestinal toxicity. However, the true incidence of adverse effects from NSAID use in dogs, is unclear.

The authors of a recent systematic review set out to discover more about the adverse effects of NSAIDs in dogs through a systematic search and assessment of published trials.1 They aimed to assess the quality of evidence on NSAID safety, and also to compare the incidence of adverse effects of NSAIDs with that of placebo. They searched for all prospective trials in English that assessed the safety of acute or chronic administration of NSAIDs in dogs.

They found 64 studies (involving a total of 4,156 dogs), evaluating 14 different NSAIDS. Only 45% of the studies were clinical (i.e. involved dogs more likely to be representative of dogs treated in practice); most of the studies (55%) were in research dogs, which are usually young and free of naturally-occurring disease. Only 22% of studies were randomised controlled and blinded (the highest quality rating for study design).

The strength of evidence was highly variable for the different drugs. For carprofen, firocoxib and meloxicam, the reviewers found a high strength of evidence (defined as relevant studies of high quality, including sufficient animals and giving results relevant to the target population). For deracoxib (not authorised for use in dogs in the UK), ketoprofen and robenacoxib, they found a moderate level of evidence (relevant studies of high-to-moderate quality, including sufficient animals and giving results that could be extrapolated to the target population with some confidence). For etodolac (not authorised for use in dogs in the UK), they found a low level of evidence (studies of moderate-to-low quality with insufficient animals, meaning that the results can be extrapolated to the target population with only low confidence). For flunixin, ketorolac, tepoxalin, licofelone and vedaprofen (not authorised for use in dogs in the UK) and for rofecoxib and tolfenamic acid, there was an “extremely low” level of evidence (few high quality studies involving few animals).

The most common outwardly detectable adverse effects reported in the trials were: vomiting, diarrhoea, anorexia, lethargy and melaena. Also reported, less commonly, were: faecal blood, bleeding, colitis, abdominal pain, aggressiveness or behaviour change, hypersalivation, polydipsia, polyuria, adipsia, constipation, icterus, skin reactions and weight loss. Out of all the studies, two gastrointestinal perforations were recorded: one in a dog that accidentally received a double dose of firocoxib; the other in a dog that received an NSAID while also being treated with a topical corticosteroid. It was not possible from the data provided in the trials to estimate the incidence of any of the adverse effects in dogs.

When NSAIDs were compared with placebo (in a subgroup of 25 randomised blinded trials), there was a statistically significant difference between NSAID and placebo groups (in 5 out of the 25 trials) for the following objective outcome measures: scores for gastrointestinal lesions (which were higher with NSAIDs), and platelet aggregation, food consumption and body weight (which were all reduced compared with placebo).

In their discussion, the review’s authors review the trial results on pathology monitoring. There were no significant changes in liver enzyme activity during or after long-term (at least 28 days) administration of NSAIDs. They comment that in general authors agree that adverse effects on the liver associated with NSAID use are more likely to be idiosyncractic, in contrast to kidney injury and gastrointestinal effects, which can be dose-dependent. In the trials there were no differences in the incidence of renal adverse effects among various NSAIDs or between treated and control groups. Renal function tests failed to detect renal adverse effects in dogs undergoing general anaesthesia and submitted to hypovolaemic and/or hypotensive stress. Outwardly detectable adverse effects that could have been attributable to renal effects were seen in four studies and included polyuria, polydipsia and adpsia although it cannot be certain that these were due to NSAID use.

Briefly, this systematic review is unusual in that it assesses the harms, rather than the benefits, of a treatment. Overall, it shows that most of the published trial evidence on the adverse effects of NSAIDs is in healthy young research dogs, rather than dogs in a clinical setting. It tells us which of the NSAIDs have the best evidence on harms – these are the ones about which there can be greater certainty about their effects. The most common adverse effects of NSAID therapy in dogs appear to be vomiting, diarrhoea, anorexia, lethargy and melaena. However, there is not enough published clinical trial evidence to know the true frequency of any adverse effects of NSAID therapy in dogs, nor how different NSAIDs compare with each other.


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

Originally published by Veterinary Prescriber in September 2013. Read more about adverse effects of NSAIDs in dogs in the context of efficacy in the Veterinary Prescriber review "NSAIDS for use in dogs with musculoskeletal problems"