Correspondence J Vet Intern Med 2008;22:243–246 10.1111/j.1939-1676.2008.0070.x

Dear Editor We would like to raise our concerns regarding the recent paper Chetboul et al J Vet Intern Med 2007; 21:742–753. Comparative Adverse Cardiac Effects of Pimobendan and Benazepril Monotherapy in Dogs with Mild Degenerative Mitral Valve Disease: A Prospective, Controlled, Blinded, and Randomized Study. In particular, we are concerned about the statement ‘‘PIMO has adverse cardiac functional and morphologic effects in dogs with asymptomatic MVD’’ as we believe this to be unfounded. We intend to reserve our comments to this part of the study. From the outset, we acknowledge we have an interest in pimobendan, having worked over many years in assessing its clinical efficacy, and have been supported in our efforts by Boehringer Ingelheim Vetmedica Ltd. We are confident that pimobendan is a valuable drug in the treatment of left-sided congestive heart failure (CHF) associated with idiopathic dilated cardiomyopathy and myxomatous mitral valve disease and have published data in support. Pimobendan is now widely used in Europe for the treatment of CHF and has been, in conjunction with ACEi and diuretics, an important contributor to the management of such cases. We are concerned that the blanket statement by Chetboul et al that pimobendan causes pathological changes in canine mitral valves will have an adverse effect on its uptake by practicing veterinarians (we have evidence to suggest that some practitioners are beginning to do so) and thereby compromise animal welfare. We believe the strength of the assertion made by the authors is unreasonable on the basis of the data they provided. Indeed, we feel it unfortunate that such a statement was made in that it is generally accepted in biomedical research that exceptional claims require exceptional proof. With respect to the histopathological changes in the two treatment groups, while differences in the observed changes at day 512 are reported, to then conclude such changes are a consequence of treatment is presumptive. The main fault with this part of the study is the selected population, the numbers in each group, and the time delay from the administration of the treatment to post mortem. Because it is well known that MMVD is a progressive disease, it is not surprising to find in a cohort of 12 dogs that there is a variation in the severity of lesions. Furthermore, we are aware that beagles are commonly affected by MMVD and lesions can be present from as young as 12 months of age, without obvious clinical evidence that they are affected. However, we know little about the manner in which the disease can progress in any one individual dog. It can, therefore, be readily appreciated that the pathological changes noted in the Pimobendan group could be by chance selection alone, especially as the group sizes are small and nearly 2 years elapsed between selection and post mortem. The least the authors could have done was to have commented on the disparity in pathology between the two groups, but to then conclude this was evidence of pimobendan-induced

pathology was unjustified. As there is no sham treated group to monitor the natural change in valve pathology in this population over time, the observed difference between pimobendan and benazapril cannot be interpreted in a valid manner for such a small sample population. Yours sincerely Brendan Corcoran, Geoff Culshaw, Joanna Dukes-McEwan, Anne French, Sarah Smith, Simon Swift The University of Edinburgh Hospital for Small Animals, UK University of Liverpool Small Animal Teaching Hospital, UK The Scarsdale Veterinary Hospital, Derby, UK

Dear Editor, I refer to the article ‘‘Comparative Adverse Cardiac Effects of Pimobendan and Benazepril Monotherapy in Dogs with Mild Degenerative Mitral Valve Disease: A Prospective, Controlled, Blinded, and Randomized Study. Chetboul et al Journal of Veterinary Internal Medicine 2007;21:742–753,’’ with which I have the following concerns:

Hypothesis ‘‘Pimobendan (PIMO) may worsen valve regurgitation in early canine MVD’’ This hypothesis draws on the citation of 3 bibliographic references, all 3 of which have been used in a manner that could mislead the reader. Schneider1 – Chetboul et al cite ‘‘adverse effects (. . .) at dosages close to the therapeutic ones.’’ However, lesions are only reported after administration of pimobendan at doses of 0.75– 2.25 mg/kg/d via intravenous injection for 4 weeks. As the oral bioavailability of pimobendan is about 60%, this corresponds to 2–8 the therapeutic dose. Furthermore, Schneider et al concluded that the secondary effects were the direct consequence of pharmacological overdose. Tissier2 – In these 2 cases of myocardial hypertrophy induced by PIMO, neither of the 2 dogs had underlying cardiac disease and both dogs were chronically overdosed. Amsallen3 – This meta-analysis concludes that there is an overall increased risk ratio with the family of PDE inhibitors, but in the detailed analysis, the risk ratio with pimobendan was not significantly increased.

Materials ‘‘Twelve Beagles with asymptomatic MVD.’’ The authors equate ‘‘dogs with Mild degenerative mitral valve disease’’ (ISACHC Stage Ia) with ‘‘dogs without systolic myocardial disease,’’ while defining the absence of systolic dysfunction ‘‘by a normal fractional shortening

244

Letter to the editor

(%FS).’’ It is widely recognized that FS is a poor contractility index in MVD. Systolic dysfunction exists in humans in stage 1, but is only detectable using invasive methods. In dogs, Borgarelli et al4 demonstrated the existence of early systolic dysfunction in dogs with stage II MVD. The breed of dogs raises several questions: 12 females, selected from 450 dogs, although it is known that there is a significant prevalence of MVD in males. Very young dogs (6.82.1 years).- Symptomatic MVD appears to be uncommon in Beagle. In another recent publication5 by the same team on ruptured chordae tendineae involving 706 dogs, no Beagle was cited. It is, however, a well represented breed in France: number 15 in Central Canine Society’s statistics in 2006. During the 17-month follow-up, none of the 12 dogs advanced to stage Ib, II, or III. Then it appears that in this nonpredisposed breed, MVD remains fixed at a very early stage and, therefore, cannot be considered as a reliable model.

Method The actual posology is not reported. To attain an actual dose of 0.25 mg/kg per dose for a dog weighing 14.2 kg (mean weight of dogs in the study), they should have been given 3.55 mg of pimobendan, while Vetmedin hard capsules are available in 1.25, 2.5 or 5 mg strengths. Were the hard capsules repackaged to avoid overdosing?

References The authors did not mention the ‘‘Vetscope’’ study6 which demonstrates the superiority of PIMO over BNZ in cases of symptomatic MVD (ISACHC stages II-III), published before the revisions.

Ethical aspects Twelve dogs were sacrificed in this study to evaluate adverse effects of drugs in a state where no benefit has been demonstrated for any of the drugs. How would the peer-review team at JVIM regard an article on the toxicity of vitamin D in dogs without rickets? The sponsor of the study is the MA holder for the ‘safe’ product and a direct competitor for the ‘unsafe’ product. Three articles have recently been published on the effects of inotropic drugs and of pimobendan in asymptomatic MVD, and all of them contradict the results presented here: Kanno et al7: on 4 dogs with experimental stage Ib MR, PIMO reduced the regurgitation fraction, diastolic and systolic ventricular diameters, and even plasma noradrenaline. Ouellet et al8: on 26 spontaneous MVD dogs, PIMO reduced the regurgitation fraction after 90 days (p o 0.01). Nakayama et al9: on 9 experimental MR dogs, milrinone did not increase mitral regurgitation. Yours sincerely Dr G. Le Bobinnec, Diplomate ECVIM Cardiology

Results

References

(A) Regurgitations on Day 0 in both groups, and throughout the entire study in the Benazepril (BNZ) group, were reported to be of low velocity (2 m/s) because their duration was short and therefore there was insufficient time for the acceleration to 5–6 m/s that the LV/LA (110 mmHg) pressure gradient normally provokes. But we know that during the early stages of MVD, these small leaks are very irregular in terms of time and space (rapidly changing direction) and are often multiple. So it is very difficult, even with the most experienced operator using high performance material, to obtain a correct evaluation using continuous wave Doppler with 4-chamber view (left) or the CFM method of Regurgitant Jet Area (but both methods have been shown to be valid in the case of massive leaks, namely, in the case of chordae rupture). (B) I believe that a leak at 9 m/s, thus a LV pressure of 330–334 mmHg, maintained for 2 months without parietal hypertrophy (only systolic increase of septal and free wall thickness reported), without aortic or pre-aortic obstruction, without arterial hypertension, without increase of aortic blood flow, without increase in frequency, without arterial dilation, without jet lesions on LA, seems physiologically impossible. (C) A number of essential results for a cardiovascular toxicity study are lacking, such as macroscopic aspect of the heart at necropsy and heart weight.

1. Schneider P, Guttner J, Eckenfels A, et al. Comparative cardiac toxicity of the I.V. administered benzimidazole pyridazinon derivative pimobendane and its enantiomers in female beagle dogs. Exp Toxicol Pathol 1997;49:217–224. 2. Tissier R, Chetboul V, Moraillon R, et al. Increased mitral valve regurgitation and myocardial hypertrophy in two dogs with long-term pimobendan therapy. Cardiovasc toxicol 2005;5:43–51. 3. Amsallen E, Kasparian C, Haddour G, et al. Phosphodiesterase III inhibitors for heart failure. Cochrane Database Syst Rev 2005;25:1. 4. Borgarelli M, Tarducci A, Zanatta R, et al. Decreased systolic function and inadequate hypertrophy in large and small breed dogs with chronic mitral valve insufficiency. J Vet Intern Med 2007; 21:61–67. 5. Serres F, Chetboul V, Tissier R, et al. Chordae tendinae rupture in dogs with degenerative mitral valve disease: prevalence, survival, and prognostic factors (114 cases, 2001–2006). J Vet Intern Med 2007; 21:258–264. 6. Lombard CW, Jo¨ns O, Bussadori C. Clinical efficacy of pimobendan versus benazepril for the treatment of acquired atrioventricular valvular disease in dogs. J Am Anim Hosp Assoc 2006;42:249–261. 7. Kanno N, Kuse H, Kawasaki M, et al. Effects of pimobendan for mitral valve regurgitation in dogs. J Vet Med Sci 2007;69:373–377. 8. Ouellet M, Di fruscia R, Belanger MC. Evaluation of pimobendane in the treatment of early mitral valve disease. J Vet Intern Med 2007;21 ACVIM forum research abstract n 138. 9. Nakayama T, Nishijima Y, Miyamoto M, et al. Effects of 4 classes of cardiovascular drugs on ventricular function in dogs with mitral regurgitation. J Vet Intern Med 2007;21:445–450.

Letter to the editor

Authors’ reply With Regard to the Letter by Corcoran et al We thank Corcoran et al for their comments regarding our paper. We share their concerns regarding animal welfare, and reaffirm that our interest as academics is to provide optimal care for veterinary patients. We agree that pimobendan probably has a valid role in some stages of canine congestive heart failure. Our concerns regarding the safety of pimobendan were initially triggered by two clinical cases in which cardiac lesions had been induced in dogs by pimobendan at dosages close to those recommended.1 Pimobendan-induced cardiac lesions in healthy dogs had already been described in the literature.2–5 Moreover, toxicological studies3 concluded that the dog ‘‘appears to be an especially sensitive species’’ to pimobendan with regard to the development of atrioventricular, endocardial and myocardial lesions as a consequence of exaggerated inotropic activity. Consequently, we concluded that, in the interests of canine welfare, the benefit-to-risk ratio for pimobendan needed to be evaluated in dogs. It is not because a drug is effective that it is safe. In human medicine, safety problems may be not detected during drug development, even with clinical studies involving thousands of patients, and sometimes only become apparent when products are available commercially (eg, Vioxx). Drug safety is defined by the fundamental principle of precaution, and consequently the statement that ‘‘it is generally accepted in biomedical research that exceptional claims require exceptional proof’’ is unacceptable in the field of drug safety. The statement that the ‘‘pathological changes in the pimobendan group could be by chance selection’’ is contradicted by the fact that allocation to treatment group was randomized, and the probability to have randomized the 6 dogs with the most severe valve lesions together into the same group at the start of the study is 1/1000, which is negligible. Moreover, and importantly, pimobendan treatment was associated with specific mitral valve lesions, eg, acute focal hemorrhages and endothelial papillary hyperplasia. These pathological findings could not be due to the spontaneous MVD per se and are in accordance with data previously published2–4 and also results provided by Boehringer Ingelheim using (as here) 6 Beagle dogs per group.5 The life expectancy of dogs with stage 1 MVD is several years,6 and therefore we do not understand Corcoran et al’ objection to the 2-year time period between selection and post-mortem. Corcoran et al’ major criticism is the number of dogs per group. We should point out that our study was a preclinical safety study with euthanasia at the end of the study period, and was not a clinical trial. Our study was in this respect in full compliance with the FDA Safety Guideline #33, which states that ‘‘animal testing should derive the maximum amount of useful scientific information using the minimum number of animals necessary.’’ In accordance with this Guideline too, it would have been unethical and unnecessary to include (and

245

consequently euthanize) an additional 6 dogs in order to allow for a sham treated group, all the more because specific cardiac lesions were observed in the pimobendan group. In conclusion, it was indisputably our ethical responsibility to make public these results and to conclude that pimobendan ‘‘has adverse cardiac functional and morphologic effects in dogs with asymptomatic MVD.’’ To have concealed such adverse effects would have been unprofessional and irresponsible.

With Regard to the Letter by Le Bobinnec We understand that he has some questions regarding our protocol design, the understanding of which requires specific expertise in the field of cardiovascular research. We note that Le Bobinnec has not declared any conflict of interest although this information is highly relevant.

Hypothesis (1) Our hypothesis was not so wrong as it was confirmed by our results. (2) If Schneider et al tested 2–8 the therapeutic dose pimobendan (as Le Bobinnec asserts), their findings of cardiac lesions with 4-week treatment support our hypothesis of safety concerns with long-term use of pimobendan at therapeutic doses (as we tested). (3) Concerning the case reports by Tissier et al,1 an unintended effect is by definition an adverse event. We reemphasize that these dogs were in a very poor condition at the time of referral, and they recovered rapidly (disappearance of heart murmur and/or decrease in mitral regurgitation) when pimobendan treatment was stopped. These two cases were judged to be drug-related by the French Pharmacovigilance Office. (4) In order to make their own opinion of this issue, we strongly encourage readers to consult the article by Amsallen et al.7 Their conclusion was that (a) phosphodiesterase inhibitors (PDIs) ‘‘are responsible for an increase in mortality rate compared with placebo in patients suffering from chronic heart failure’’; (b) ‘‘considering mortality from all causes, the deleterious effect of PDIs appears homogeneous whatever the concomitant use (or non-use) of vasodilating agents, the severity of heart failure, the derivative or the molecule of PDI used’’; (c) ‘‘Consequently, the chronic use of PDIs should be avoided in heart failure patients.’’

Animals (1) To confirm normal systolic function, normal tissue Doppler imaging variables were also required. Moreover, the results of Borgarelli et al refer to stage 2 MVD (and not stage 1 with mild regurgitation as here). (2) Canine MVD is indeed more prevalent in males than in females. However, as described in our paper, the dogs used were sourced from a breeding colony for laboratory animals in which there was, for obvious reasons, a clear predominance of females (92%) over males (8%). (3) 6.8 years is not very young, especially for a dog with stage 1a MVD. At our unit, 41/3 of MVD dogs are 5–10-year-

246

Letter to the editor

old.8 Corcoran et al confirm also that Beagles ‘‘are commonly affected by MVD and lesions can be present from as young as 12 months of age.’’ (4) In the study of Serres et al,9 only owners’ (and not laboratory) dogs, including one Beagle, were recruited. This is perfectly in accordance with the expected proportion of Beagles (mainly a hunting breed in France) in the Paris area. (5) The assertion that dogs should have advanced to a higher stage of heart failure over the 17-month test period is not supported by the literature. In Cavalier King Charles Spaniels (which have a faster MVD progression than other breeds), the mean time to heart failure is much longer than 17 months, even in dogs with cardiomegaly (32–43 months).6

important to know the effects of vitamin D in healthy dogs, so that rickets can be safely prevented as well as treated. The toxicity of vitamin D has indeed been studied in healthy dogs.10 (2) Concerning sponsorship, we are independent university-based academics and need to obtain funding for research. Any potential conflict of interest was disclosed in our paper. Our study was randomized and all clinical, imaging, and pathological examinations were conducted blinded, thus ensuring the validity of the results. We hope that these answers will help Le Bobinnec in his understanding of our paper.

Methods

1. Tissier R, Chetboul V, Moraillon R, et al. Increased mitral valve regurgitation and myocardial hypertrophy in two dogs with long-term pimobendan therapy. Cardiovasc Toxicol 2005;5:43–51. 2. Schneider P, Guttner J, Eckenfels A, et al. Comparative cardiac toxicity of the IV administered benzimidazole pyridazinon derivative pimobendan and its enantiomers in female Beagle dogs. Exp Toxicol Pathol 1997;49:217–224. 3. Schneider P, Bauer M, Eckenfelds A, et al. Acute, subacute and chronic toxicity studies of pimobendan in laboratory animals. Pharmacometrics 1992;42:561–578. 4. Schneider P. Hemodynamically induced heart lesions in the dog after the administration of cardio-active substances. Exp Pathol 1990;40:155–168. 5. FDA freedom of information summary, Vetmedin, Boehringer Ingelheim Vetmedica (http://www.fda.gov/OHRMS/ DOCKETS/98fr/2007-141-273-fois001). 6. Kvart C, Ha¨ggstro¨m J, Pedersen HD, et al. Efficacy of enalapril for prevention of congestive heart failure in dogs with myxomatous valve disease and asymptomatic mitral regurgitation. J Vet Intern Med 2002;16:80–88. 7. Amsallem E, Kasparian C, Haddour G, et al. Phosphodiesterase III inhibitors for heart failure. Cochrane Database Syst Rev 2005;25:1. 8. Gouni V, Serres FJ, Pouchelon JL, et al. Quantification of mitral valve regurgitation in dogs with degenerative mitral valve disease by use of the proximal isovelocity surface area method. J Am Vet Med Assoc 2007;231:399–406. 9. Serres F, Chetboul V, Tissier R, et al. Chordae tendineae rupture in dogs with degenerative mitral valve disease: Prevalence, survival, and prognostic factors (114 cases, 2001–2006). J Vet Intern Med 2007;21:258–264. 10. Spangler WL, Gribble DH, Lee TC. Vitamin D intoxication and the pathogenesis of vitamin D nephropathy in the dog. Am J Vet Res 1979;40:73–83.

The hard capsules were not repackaged (2.5 and 5 mg capsules, see footnote list). The exact dosages ranged from 0.22 to 0.30 mg/kg PO q12h, in strict compliance with the recommendations of the drug company (with such a formulation, a 14-kg dog should be given 0.357 mg/kg q12h).

Results (A) In our study, none of the dogs showed multiple mitral regurgitations. Moreover, based on current evidence in canine cardiology and our expertise,8 Le Bobinnec’s comment about validation is totally erroneous. (B) Evidence-base medicine is based on research results and fortunately not on personal perception. (C) Samplings of appropriate specimens for histopathology were of course identified by macroscopic appearance of the heart. In our view, heart weight is a very rough indicator of alterations of cardiac valves.

References and postscripts Our paper was submitted in June 2006, and the Vetscope study was published in July-August 2006. Nakayama et al used an experimental model and did not test pimobendan. The study by Kanno et al has several major weaknesses (eg, surgically induced MVD, no blinding, no validation data for imaging variables, and no pathological data provided whereas ethically these dogs should have been euthanized at the end of the study period). The Ouellet et al study has been published only as an abstract and therefore we cannot reliably assess the results.

Ethical aspects (1) As defined in FDA/EMEA guidelines, safety studies are generally performed in laboratory and not diseased animals. The argument regarding vitamin D seems neither appropriate nor correct because unlike rickets and vitamin D, heart failure in dogs is not caused by a deficiency of pimobendan. In addition, it is certainly

References

Vale´rie Chetboul, Herve´ P Lefebvre, Carolina Carlos Sampedrano, Vassiliki Gouni, Vittorio Saponaro, Francois Serres, Didier Concordet, Audrey P. Nicolle, and Jean-Louis Pouchelon Unite´ de Cardiologie d’Alfort UMR-INSERM U841 (National Institute of Health and Medical Research), Ecole Nationale Ve´te´rinaire d’Alfort, 7 Avenue du Ge´ne´ral de Gaulle, 94 704 Maisons-Alfort cedex, France UMR 181 Physiopathologie et Toxicologie Expe´rimentales INRAENVT, Ecole Nationale Ve´te´rinaire de Toulouse, 23 chemin des Capelles, 31076 Toulouse cedex 03, France