We would like to express our concern over incorrect statements made in a recent review article (Systematic review of nonsteroidal anti-inflammatory adverse effects, J Vet Intern Med 2013;27: 1011–1019, B.P. Monteiro-Steagall, P.V.M. Steagall, and B.D.X. Lascelles) with respect to one of the articles referenced – 55. Nephrotoxicity in dogs associated with methoxyflurane anesthesia and flunixin meglumine analgesia, Can Vet J. 1990;31:766-771, Mathews KA, Doherty T, Dyson DH et al.
We would like to commend the authors on preparing this review as such reviews are an important contribution to our knowledge of a subject. However, it is also important to maintain the accuracy and intent of the contents of articles presented as, frequently, it is this information that is perpetuated rather than the reference to the original document.
Our concern is that the above study, referring to reference 55, states that flunixin meglumine (FM) was administered with a corticosteroid or another NSAID and implied the adverse effects of FM were associated with this co-administration. This statement is incorrect as FM was not administered with a corticosteroid or another NSAID, as this was known by the authors to be absolutely contraindicated. Another concern is that our study was also cited related to FM being “administered with a well-known nephrotoxic agent”, and implying methoxyflurane (MOF). In 1986 when the study was performed, MOF was not known as a “well-known nephrotoxic agent” to the practicing veterinary population with only one or two articles in the human literature commenting on potential nephrotoxicity. MOF was in common use in 1986 when the study was performed and FM was recently approved as an analgesic in Canada. Our study included a group of dogs in the surgical exercise program where third year veterinary students are given the opportunity to experience current anesthesia and analgesia practices as a component of their anesthesia training. The first group of 5 dogs received FM at the termination of anesthesia. The following day all of these dogs demonstrated signs of acute kidney injury (AKI) and increased serum creatinine. This was of great concern but numbers too small to publish, even though 100% of this group was affected. Based on the 100% affected, our epidemiologist indicated that another 5 dogs with the potential for 100% being affected would give a power of 84% and adequate to confirm toxicity. This was carried out in the next surgical exercise laboratory where 100% of dogs were again affected. While the power analysis was not mentioned in the article, the facts presented indicated 100% of 10 dogs, which implies this would be a consistent problem in veterinary practice. With respect to design, we respected the 3 R's and were not permitted to use placebo, nor would we, so a commonly used analgesic was used as a comparator.
This study was influential in ending MOF followed by FM at the moment of discovery yet this study was rated extremely low (because of the rating system), yet met all the criteria for appropriate study design and necessary information for practitioners to know, and it was published in a refereed veterinary journal available to Canadian practitioners. This information was presented at the American Veterinary Anesthesiologists conference in Illinois in 1987 (noted in the original article). These facts contradict its inclusion in the category of extremely low strength of evidence that can be extrapolated to a target population with low confidence even though all studies were ‘clumped’ to give the final rank. In fact, this study prevented dogs from receiving MOF and FM with the consequence of subsequent AKI.
As future information gathering and citations may be obtained from the review article, rather than the original study, we request that this erratum be published so as to not perpetuate the incorrectly stated interpretation.