Provided for non-commercial research and education use ... · van der Meer, Frank Wakshlag, Joseph...

The Canadian Journal of Veterinary Research is published by Canadian Veterinary Medical Association. The attached copy is

furnished to readers for personal, non-commercial research and education use. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party Web sites are prohibited. Those who require further

information regarding reprints or archiving and manuscript policies are encouraged to contact [email protected].

Provided for non-commercial research and education use. Not for reproduction, distribution or commercial use.

VETERINARY RESEARCHRECHERCHE VÉTÉRINAIREREVUE CANADIENNE DE

JANUARY/JANVIER 2020, VOL. 84, NO. 1

CANADIAN JOURNAL OFFOR PERSONAL USE ONLY

Published quarterly by the Canadian Veterinary Medical AssociationPublication trimestrielle de l’Association canadienne des médecins vétérinaires

C a n a d i a n J o u r n a l o fVe t e r i n a r y R e s e a r c h

Formerly CANADIAN JOURNAL OF COMPARATIVE MEDICINE

The journal publishes the results of original research in veterinary and comparative medicine. Manuscripts must be as concise as pos-sible, and the research described must represent a significant con-tribution to knowledge in veterinary medicine. Full-length papers, short communications, and review papers are welcome. All manu-scripts will be reviewed for scientific content and editorial accuracy.

Manuscripts must conform to the Instructions for Authors found on the CVMA website. Consultation of a recent issue of the journal to ensure that the manuscript conforms to current style is also recom-mended. Please use the website (www.canadianveterinarians.net) to submit articles. A publication charge of $95 per published page for CVMA Members and $140 per published page for Non-members, plus a cost of $25 per table, $25 per illustration, and $155 per color illustration is payable by contributors. Instructions for ordering reprints are sent with the galley proofs. Color illustrations will only be published if the author(s) agree to pay the extra cost.

R e v u e c a n a d i e n n e d er e c h e r c h e v é t é r i n a i r e

Remplace la REVUE CANADIENNE DE MÉDECINE COMPARÉE

La revue publie les résultats de recherches originales en médecine vétérinaire et comparative. Les manuscrits doivent être aussi brefs que possible et la recherche décrite doit apporter une contribution impor-tante à la médecine vétérinaire. Les exposés complets, les communi-cations brèves et les comptes rendus sont bienvenus. Nous véri fierons le fonds scientifique et la forme de tous les manuscrits.

Les manuscrits doivent être conformes aux directives aux auteurs qui se trouvent sur le site Web de l’ACMV. La consultation d’un récent numéro de la revue est aussi profitable, afin de s’assurer qu’un manuscrit se conforme au style courant. Veuillez utiliser le site Web (www.veterinairesaucanada.net) pour soumettre les articles. Les auteurs doivent acquitter des frais de 95 $ par page publiée pour les membres de l’ACMV et 140 $ par page publiée pour les non-membres, 25 $ pour chaque tableau, 25 $ pour chaque illustration, et 155 $ pour chaque illustration en couleur. Un formulaire de commande de tirés à part accompagne l’en voi des épreuves d’un manuscrit. Les illustrations en couleurs ne seront publiées que si les auteurs en assument les frais supplémentaires.

Established in 1937 ISSN 0830-9000

Address all correspondence to/Adresser toute correspondance à :Canadian Journal of Veterinary Research Revue canadienne de recherche vétérinaire339, rue Booth Street, Ottawa, Ontario K1R 7K1Tel./Tél. : (613) 236-1162 — Fax/Télécopieur : (613) 236-9681e-mail/Courriel : [email protected]© Canadian Veterinary Medical Association 2020

L’Association canadienne des médecins vétérinaires 2020

Typesetting/Typographie AN Design Communications www.an-design.ca

Canadian Journal of Veterinary ResearchRevue canadienne de recherche vétérinaireEditor — Rédacteur Éva Nagy, Guelph, OntarioAssociate Editor — Rédacteur adjoint Faizal Careem, Calgary, AlbertaAssistant Editor — Assistant à la rédaction Serge Messier, Saint-Hyacinthe (Québec)Managing Editor — Directrice de la rédaction Heather Broughton, Ottawa, OntarioAssistant Managing Editor — Directrice adjointe à la rédaction Stella Wheatley, Ottawa, OntarioAdvertising and Sponsorship Consultant — Consultante, publicité et commandites Laima Laffitte, Wendover, OntarioEditorial Coordinator/Coordonnatrice de la rédaction Kelly Gray-Sabourin, Ottawa, Ontario

Abstracted and/or indexed in:/Résumé et/ou catalogué dans :Biological Abstracts Index VeterinariusChemical Abstracts Nutrition AbstractsCurrent Contents (Agric. Section) Science Citation IndexEMBASE/Excerpta Medica Veterinary BulletinIndex Medicus Derwent Veterinary Drug File

Subscription/Abonnement 2020Personal Subscription/ $185 $ (1 GST/HST as Abonnement personnel applicable; TPS/TVH en (Canada) vigueur)Personal Subscription/ $205 $ US/É.-U. Abonnement personnel (Foreign/Étranger)Institutional Subscription/ $255 $ (1 In Canada, GST/HST Abonnement institutionnel as applicable; Au Canada, TPS/TVH selon le cas)Institutional Subscription/ $255 $ US/É.-U. Abonnement institutionnel (Foreign/Étranger)

FOR PERSONAL USE ONLY

REVIEW ARTICLE/COMPTE RENDUCarbapenemase-producing Enterobacteriaceae in animals and current methodologies for their detectionRebecca E.V. Anderson, Patrick Boerlin . . . . . . . . . . . . . . . .3

ARTICLES

Microbiology

Prevalence and mutation analysis of the spike protein in feline enteric coronavirus and feline infectious peritonitis detected in household and shelter cats in western CanadaLaura A. McKay, Melissa Meachem, Elisabeth Snead, Terri Brannen, Natasha Mutlow, Liz Ruelle, Jennifer L. Davies, Frank van der Meer . . . . . . . . . . . . . . 18

Clinical Studies

Fluoroscopic and radiographic assessment of variations in tracheal height during inspiration and expiration in healthy adult small-breed dogsGrégoire Scherf, Isabelle Masseau, Anne-Sophie Bua, Guy Beauchamp, Marilyn E. Dunn . . . . . . . . . . . . . . . . . . 24

Comparison of 3 intraosseous catheter sites and methods of determining placement success in cadaver rabbitsChristopher R. Kennedy, Jay N. Gladden, Elizabeth A. Rozanski . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Effects of a single paracetamol injection on the sevoflurane minimum alveolar concentration in dogsPaula González-Blanco, Susana Canfrán, Rubén Mota, Ignacio A. Gómez de Segura, Delia Aguado . . . . . . . . . . 37

Correlation of activity data in normal dogs to distance traveledBishoy S. Eskander, Megan Barbar, Richard B. Evans, Masataka Enomoto, B. Duncan X. Lascelles, Michael G. Conzemius . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

Pathobiology

Differences in the relative counts of peripheral blood lymphocyte subsets in various age groups of pigsOlga Pietrasina, Julia Miller, Anna Rzasa . . . . . . . . . . . . 52

Determination of urokinase-type plasminogen activator serum levels in healthy and oncologic catsCláudia Viegas, Augusto J. de Matos, Liliana R. Leite-Martins, Inês Viegas, Rui R. F. Ferreira, Hugo Gregório, Andreia A. Santos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

Antioxidative enzyme activity and total antioxidant capacity in serum of dogs with degenerative mitral valve diseaseMarcin Michałek, Aleksandra Tabis, Alicja Cepiel, Agnieszka Noszczyk-Nowak . . . . . . . . . . . . . . . . . . . . . . 67

SHORT COMMUNICATIONS/COMMUNICATIONS BRÈVES

Pathobiology

Alterations in serum protein electrophoresis profiles during the acute phase response in dogs with acute pancreatitisJi-Seon Yoon, Suhee Kim, Jin-Hee Kang, Jinho Park, DoHyeon Yu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

Serum paraoxonase-1 activity in tail and mammary veins of ketotic dairy cowsRika Fukumori, Hanan K. Elsayed, Masahito Oba, Yasumitsu Tachibana, Ken Nakada, Shin Oikawa . . . . . 79

Acknowledgment of reviewers/translators (Volume 83 — 2019) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

JANUARY/JANVIER 2020, VOL. 84, NO. 1


Acknowledgment Témoignage

2 The Canadian Journal of Veterinary Research 2020;84:2

Acknowledgment of reviewers/TranslatorsRemerciement aux évaluateurs et aux traducteurs

The continued success of the Journal is due in no small measure to the willingness of reviewers to assist the editors by their evalua-tion of manuscripts. Their generous efforts contribute both to the quality of the Journal and to the quality of veterinary research. Reviewers of manuscripts published, rejected, or expired from November 2018 to October 2019 are listed below. The Editors and Editorial Board wish to thank these colleagues for the donation of their time and the sharing of their expertise.

Le succès continu de la Revue est attribuable en très grande par-tie à la collaboration des lecteurs qui appuient les rédacteurs en évaluant les manuscrits. Leurs généreux efforts contribuent tant à la qualité de la Revue qu’à celle de la recherche vétérinaire. Voici la liste des lecteurs des manuscrits qui ont été publiés ou rejetés pour la période de novembre 2018 à octobre 2019. Les rédacteurs et le Comité de rédaction désirent remercier ces collègues du don de leur temps et du partage de leur expertise.

The Editors and Editorial board wish to also thank the following translators for their excellent service.

Les rédacteurs at le Comité de redaction souhaitent aussi remercier les traducteurs ci-dessous pour leurs excellents services.

Abdul Careem, M. FaizalAbdul-Cader, SarjoonAgunos, AgnesAmbrisko, TamasAn, Dong JunArroyo, LuisAtalan, GültekinBajcsy, Árpád CsabaBeaudry, HeleneBissonnette, NathalieBoysen, SorenBrandao, JoaoCarson, CaroleeCassar, GlenCaswell, JeffChilds-Sanford, SarahCôte, EtienneCreighton, CateDavis, Margaret

Detmer, SusanDundas, JamesDunning, MarkEllis, JohnEstienne, MarkFox, Lawrence K.Gagnon, JeromeGivens, DanGodson, DaleGoodband, RobertJoseph, RubinKastelic, JohnKatholm, JorgenKeller, CharlotteKim, SukKrell, PeterKuhnert, PeterKwon, ByungjoonLescun, Tim

Linhares, DanielMacfarlane, JulieMakondo, KennedyMarcos Santana, AndréMateu, EnricMonostori, ÉvaMuckle, C. AnneMunoz Domon, MarcosMutharia, LucyNathues, HeikoNdou, SaymoreNykamp, StephanieOrtega Santana, CesarPapakonstantinou, StratosPaul, NarayanPayen, GuillaumePerret, JenniferPozzi, AntonioRisco,Carlos

Rozanski, ElizabethSalgado Miranda, CeleneSimon, BradleySlavic, DurdaSmith, BillySmith-Carr, SaralynStaffieri, FrancescoSun, HaiyanSuradhat, SanipaValverde, Alexvan der Meer, FrankWakshlag, JosephWilkins, WendyWood, DarrenWorthing, KateYan, Xing-RongYang, HanchunYason, CarmencitaZakhartchouk, Alexander

Isabelle Vallières • André Bisaillon • Serge Messier • Sophie Perreault


Review Article Compte rendu

2020;84:3–17 The Canadian Journal of Veterinary Research 3

I n t r o d u c t i o nCarbapenems are a class of antibiotics that have a wide spectrum

of activity against both Gram-positive and Gram-negative bacteria. They belong to the family b-lactams, which is also comprised of penicillins, cephalosporins, and monobactams. The first carbape-nem, thienamycin, was discovered in 1976 as a natural product of Streptomyces cattleya (1). Its structure served as a basis for the subsequent development of more stable and effective semisynthetic

carbapenems (Figure 1). The first of these new carbapenems is imipenem (Figure 2b), which presents a high affinity for penicillin-binding proteins (PBPs), although it must be coadministered with an inhibitor of dehydropeptidase-I to avoid the nephrotoxic effects of its degradation product (2).

Following the development of imipenem, further advancements were made in the production of more stable synthetic carbapen-ems, such as meropenem, ertapenem, doripenem, and biapenem

(Figure 2c) (2). While these new synthetic antibiotics have an additional methyl group, making them more stable, they have the same basic structure and mechanism of action (Figure 2a) (3). Carbapenems tend to be unstable in aqueous solution and have a short shelf life when not stored as dry powder (4), which may not only have clinical and pharmacological implications, but could also affect their use for research and diagnostic purposes.

Four main mechanisms have been described that confer resistance to carbapenems in Gram-negative bacteria. These include the pro-duction of carbapenemases, synergy between other b-lactamases and porin modifications, efflux pumps, and modifications to PBPs. These form the basis for the distinction between the broad category of carbapenem-resistant Enterobacteriaceae (CRE), which can be resistant to carbapenems by any of these mechanisms, and the more specific group of carbapenemase-producing Enterobacteriaceae (CPE or CP-CRE).

Carbapenemase-producing Enterobacteriaceae in animals and methodologies for their detection

Rebecca E.V. Anderson, Patrick Boerlin

A b s t r a c tCarbapenemase-producing bacteria are difficult to treat and pose an important threat for public health. Detecting and identifying them can be a challenging and time-consuming task. Due to the recent rise in prevalence of infections with these organisms, there is an increased demand for rapid and accurate detection methods. This review describes and contrasts current methods used for the identification and detection of carbapenemase-producing bacteria to help control their spread in animal populations and along the food chain. The methods discussed include cultures used for screening clinical samples and primary isolation, susceptibility testing, culture-based and molecular confirmation tests. Advantages and disadvantages as well as limitations of the methods are discussed.

R é s u m éLes bactéries productrices de carbapénèmases sont difficiles à traiter et représentent une menace importante pour la santé publique. Leur détection et identification peut être une tâche ardue et qui prend du temps. Étant donné l’augmentation récente de la prévalence des infections dues à ces microorganismes, il y a une demande accrue pour des méthodes de détection rapides et exactes. La présente revue décrit et met en contraste les méthodes actuelles utilisées pour l’identification et la détection des bactéries productrices de carbapénèmases afin d’aider à limiter leur dissémination dans les populations animales et dans la chaîne alimentaire. Les méthodes discutées incluent les cultures utilisées pour le tamisage d’échantillons cliniques et l’isolement primaire, les épreuves de sensibilité, ainsi que les tests de confirmation basés sur la culture et les méthodes moléculaires. Les avantages et inconvénients ainsi que les limitations des méthodes sont discutés.

(Traduit par Docteur Serge Messier)

Department of Pathobiology, Ontario Veterinary College, University of Guelph, 50 Stone Road East, Guelph, Ontario NIG 2W1.

Address all correspondence to Dr. Patrick Boerlin; telephone: (519) 824-4120, ext. 54647; fax: (519) 824-5930; e-mail: [email protected]

Received August 1, 2019. Accepted September 8, 2019.

Figure 1. Structure of thienamycin, the first described, naturally produced carbapenem [adapted from (2)].


4 The Canadian Journal of Veterinary Research 2000;64:0–00

Carbapenemase-producing Enterobacteriaceae are of particular concern and epidemiological relevance because carbapenemase genes are frequently located on mobile genetic elements, such as transposons, integrons, and plasmids, and can be transferred hori-zontally between cells (5), while the other carbapenem resistance mechanisms are less prone to such transfer. The active transfer of carbapenemase-encoding plasmids among bacterial strains, species, and genera has been documented repeatedly, both locally as part of hospital outbreaks (6,7) and more broadly at the regional or national level (8). As a result, the 2015 guidelines of the Centers for Disease Control recommended that bacterial isolates be screened specifically for CPE and not only for CRE (9).

As with other b-lactamases, carbapenemases hydrolyze the

b-lactam ring of penicillins, but they also hydrolyze cephalo-sporins, monobactams, and carbapenems (10). Carbapenemases

belong to 3 Ambler classes: Class A — serine carbapenemases; Class B — metallo-b-lactamases (MBL); and Class D — OXA b-lactamases (oxacillinases) (11,12). Class A consists mainly of non-metallo-carbapenemase (NMC), non-metallo-carbapenemase-A/imipenem resistant (IMI), Serratia marcescens enzyme (SME), Klebsiella pneumoniae carbapenemase (KPC), and Guiana extended-spectrum (GES) (12). Class B consists primarily of b-lactamase active on imi-penem (IMP), Verona integron-encoded metallo-b-lactamase (VIM), and New Delhi metallo-b-lactamase (NDM) (12). Finally, the main carbapenemases of the oxacillinase group are part of the OXA-48-like b-lactamases (13).

Carbapenemases can be encoded on either a plasmid or the chro-mosome (10,11). Although not formally carbapenemases, AmpC

b-lactamases, such as CMY (class C) and some extended-spectrum b-lactamase (ESBL), can also cause carbapenem resistance, especially when combined with other resistance mechanisms, such as porin

loss or efflux mechanisms (10,12). Porin modifications can cause a decrease in the diffusion rate of the antibiotics across the Gram-negative outer membrane at a rate that allows ESBLs and AmpC enzymes to sufficiently break down the remaining antibiotic to result in a fully resistant phenotype (14). Modifications in efflux pumps can cause an increase in antibiotic expulsion from the cell before the drug is able to bind to PBPs, thus rendering it incapable of harming the cell (14). Finally, modifications within the PBPs themselves prevent proper binding of the b-lactams to the targeted proteins, rendering them useless (14).

This diversity of carbapenem resistance mechanisms and car-bapenemases, as well as their variable effects on minimal inhibitory concentrations (MICs) in different bacterial host species (10,15,16), have made it difficult to detect CREs. Difficulty in reaching agree-ment on criteria for their detection and identification has resulted in variation across expert groups worldwide (15), particularly with regard to the specific carbapenem molecules and concentrations to use for screening for CRE and CPE and setting breakpoints for inter-preting susceptibility tests. For example, a recent study assessing the accuracy of CPE detection using the methodologies of the Clinical Laboratory Standards Institute (CLSI) and the European Committee on Antimicrobial Susceptibility Testing (EUCAST) showed that 13.8% and 1.6% of CPE isolates would have been missed when using their respective breakpoints (17). A study by Bulik et al (18) demonstrated the need for consistent breakpoints used across laboratories, e.g.,

Figure 2. Base structure of carbapenemase with a — b-lactam ring shown in blue, variable R1 region present in all carbapenemases shown in red, and variable R2 region present in meropenem, doripenem, and ertapenem shown in green; b — R1 group for imipenem; and c — R1 and applicable R2 groups for meropenem, ertapenem, doripenem, and biapenem. Molecular structures were adapted from (2).



CLSI versus US Food and Drug Administration, and the variability of accuracy of different tests, e.g., E-test, broth microdilution, and a variety of automated systems. These studies illustrate the challenges in detecting CPE and the need for accurate standardized detection methods that can be used worldwide.

Because of their broad spectrum of activity and the increasing frequency of resistance to other antimicrobial agents, carbapenems are considered to be critical for human medicine (19) and every attempt should be made to control the spread of CRE and of CPE in particular. The problem caused by the emergence of these bacteria is compounded by the frequent association of carbapenem resistance with multidrug resistance and the co-location of resistance genes for other antimicrobials on plasmids encoding CREs, thus leav-ing very few alternatives for treating an infection caused by these bacteria (20).

Although the prevalence of CRE and CPE is low in animals, it is increasing. They have now been found in companion animals, livestock, and even in wildlife and the environment. A more compre-hensive list of references on their occurrence in animals is included in Table I. Carbapenemase genes are found worldwide in bacteria from humans and animals (10), although there may be some geo-graphic variation in type and variant distribution (21). These data suggest that their prevalence in animals is increasing globally as is the potential for cross-species transfer.

To the best of our knowledge, no study has been published to date at the local level in Canada describing the presence of CPE in food animals. In the context of companion animals, we conducted a short study in Canada in 2018 with 64 fecal samples from dogs in Guelph and Toronto, Ontario. Approximately 1 g of feces was first enriched in tryptic soy broth containing 0.25 mg/mL of ertapenem and then plated onto CARBA SMART agar and MacConkey agar plates for detection of carbapenemase-producers. Isolates growing on these plates were tested with the modified Hodge test (MHT), using ertapenem disks to increase sensitivity. No CPE was detected with this approach, which suggests that they were still absent or infrequent (, 6% using an exact binomial 95% confidence interval) among dogs in southern Ontario at that time.

Carbapenemase-producing bacteria have been transferred between companion animals and humans in many cases and are a grow-ing public health concern (22–26). In this context, the case of a carbapenem- resistant Escherichia coli isolate resistant to all antimicro-bials except tigecycline and polymyxins reported recently in a young dog is of great concern (27). As part of the control measures to limit the spread and selection of CREs and CPEs, the use of carbapenems should be strictly restricted, if not avoided completely, in animals in general (28,29). Despite such measures, CRE and CPE may still occur in animals, in part because of the potential for co-selection by other antimicrobial agents. The detection of CPE carriers and CPE infec-tions in animals is therefore warranted and adequate methodologies for this purpose need to be further developed and agreed upon.

The relatively low carbapenem MICs of some CPEs (15,30) and the need for further tests, in addition to basic susceptibility testing to differentiate them among CREs, may make CPEs difficult to identify. For the same reasons, it is also difficult to detect CPE carriers for epidemiological and preventive purposes. If present in low numbers, CPEs can easily be overshadowed by other bacteria and missed with

standard detection procedures, thus allowing further transmission while remaining undetected (10). The second part of this review will therefore discuss the methodologies used to identify and detect CPEs and their challenges.

M e t h o d o l o g i e s a n d a p p l i c a t i o n s

Susceptibility testing methodsAntimicrobial susceptibility tests (ASTs) provide quantitative evi-

dence as to whether or not an isolate has reduced susceptibility to a particular carbapenem. The results of such tests are used primarily to guide treatment modalities and predict therapeutic success (10). Although these results can also be used to identify CREs for epide-miological and preventive purposes, additional tests are needed to reliably identify CPEs and to identify the type of carbapenemase responsible for the observed decreased susceptibility.

The main ASTs include disk diffusion, broth microdilution (BMD), agar dilution, gradient methods (E-test), and a variety of com-mercial automated systems. An important challenge with all these methods is choosing the adequate carbapenem molecule and clinical breakpoints or epidemiological cut-off values for the purpose under consideration. This choice would appear relatively straightforward for clinical diagnostic purposes, for which expert groups, such as the Clinical and Laboratory Standards Institute (CLSI) (31,32) and the European Committee on Microbial Susceptibility Testing [EUCAST (http://eucast.org)], among others, provide well-established and validated guidelines and standards.

This is more challenging for epidemiological and surveillance purposes in which it is important to detect any carbapenemase, even if associated with minor changes in MICs. For these purposes, it is critical to choose adequate molecule(s). The AST methodology chosen is critical and cut-off values to sensitively and specifically identify isolates with reduced susceptibility are more difficult to set. This is particularly challenging for OXA-48-like enzymes, which can be easily missed with traditional susceptibility testing techniques because of the relatively low MICs they afford (33).

Many studies have been conducted to compare the effectiveness of a variety of ASTs for identifying CPEs. Unfortunately, the diversity of objectives, populations under study, i.e., the distribution of different carbapenemases and bacterial species in the samples, and interpretation criteria used for these studies make them extremely difficult to compare and almost impossible to draw clear general conclusions on the respective performances of ASTs. Nevertheless, researchers seem to agree that tests based on ertapenem (and perhaps imipenem) are the most sensitive for detecting CPEs, although they present a relatively low specificity (34). Meropenem seems to present the best overall compromise in terms of sensitivity and specificity and is now recommended by both EUCAST (http://www.eucast.org/fileadmin/src/media/PDFs/EUCAST_files/Resistance_mechanisms/EUCAST_detection_of_resistance_mechanisms_170711.pdf) and CLSI (31,32) for screening isolates for reduced susceptibility to carbapenems and possible presence of carbapenemases. In order to reach maximum sensitivity, it is recommended that adequate species-specific cut-off values be used to differentiate CREs from wild-type organisms in the context of epidemiological investigations



Table I. Geographical location of carbapenemase-producing Enterobacteriaceae (CPE) isolated from animal clinical samples.

Animal Continent Year Gene Country Bacterial species species Clinical sample ReferenceNorth America 2008–2009 NDM-1 USA Escherichia coli Dog Wound (98) Dog Nose Dog Urine Cat Urine 2009–2013 OXA-48 USA Escherichia coli Dog Unknown clinical source (99) Cat Unknown clinical source

South America 2017 KPC-2 Brazil Escherichia coli Dog Urine (100)

Asia 2013 OXA-48 Lebanon Escherichia coli Fowl Rectum (101) 2013–2015 OXA-48 China Escherichia coli Dog Diagnostic sample (102) 2013 NDM-1 China Escherichia coli Dog Anus (27) 2014–2016 NDM-1 India Escherichia coli Swine Rectum (103) Unknown NDM-1 India Escherichia coli Dog Scrotum (104) 2014–2016 NDM-5 India Escherichia coli Swine Rectum (103) 2015 NDM-5 China Escherichia coli Dairy cattle Feces (105) Klebsiella pneumonia Dairy cattle Feces (106) Escherichia coli Duck Rectum (107) Cat Rectum (108) 2017 NDM-5 South Korea Escherichia coli Dog Rectum (109) 2015 NDM-17 China Escherichia coli Poultry Cloaca (110) 2015 VIM-2 China Escherichia coli Dairy cattle Feces (105)

Africa 2014 OXA-181 Egypt Escherichia coli Dairy cattle Rectum (111) 2014 OXA-48 Egypt Klebsiella pneumonia Poultry Organs (112) Escherichia coli Dairy cattle Rectum (111) 2014–2015 OXA-48 Algeria Escherichia coli Dog Rectum (113) Escherichia coli Cat Rectum 2014–2016 OXA-48 Algeria Escherichia. Coli Wild boar Rectum (114) Klebsiella pneumonia Wild boar Feces 2015 OXA-48 Algeria Escherichia coli Bird Feces (115) 2015–2016 OXA-48 Algeria Escherichia coli Dog Rectum (116) Enterobacter cloacae Bird Rectum Enterobacter cloacae Horse Rectum Enterobacter cloacae Dog Rectum Klebsiella pneumonia Cat Rectum 2014–2015 NDM-5 Algeria Escherichia coli Dog Rectum (113) 2015 NDM-5 Algeria Escherichia coli Dairy cattle Milk (117) 2014 NDM Egypt Klebsiella pneumonia Poultry Organs (112) 2014 KPC Egypt Klebsiella pneumoniae Poultry Organs (112)

Oceania 2012 IMP-38 Australia Citrobacter freundii Bird Cloaca (118) 2012 IMP-4 Australia Escherichia coli Bird Cloaca Escherichia fergusonii Bird Cloaca Klebsiella pneumoniae Bird Cloaca Kluyvera georgiana Bird Cloaca Enterobacter aerogenes Bird Cloaca Enterobacter cloacae Bird Cloaca Citrobacter braakii Bird Cloaca Proteus mirabilis Bird Cloaca Proteus penneri Bird Cloaca 2016 IMP-4 Australia Salmonella enterica Typhimurium Cat Rectum (85)

Europe 2009–2011 OXA-48 Germany Klebsiella pneumoniae Dog Diagnostic sample (119)



(17). These organism-specific epidemiological cut-off values can be found on the EUCAST website (https://mic.eucast.org/Eucast2/), although some caution may be needed in their use (http://www.eucast.org/mic_distributions_and_ecoffs/). Alternatively, they can be generated using software such as ECOFFinder (https://clsi.org/meetings/microbiology/ecoffinder/) (35).

For practical and financial reasons, only disk diffusion and BMD are usually used in veterinary medicine and in surveillance programs involving animals. Broth microdilution is often used as a reference method. Several studies using ertapenem (36) and meropenem (15) or variable combinations of carbapenems (37,38) for comparing ASTs suggest a good correlation between disk diffusion and BMD, thus supporting the use of disk diffusion for identifying CREs and CPEs in animals. The instability of carbapenems in aqueous solution

(3,4,39) further supports the use of disk diffusion, in which these drugs are more stable.

Confirmation tests for carbapenemase-producersAs previously stated, the inability of classical antimicro-

bial susceptibility tests (ASTs) to clearly differentiate between carbapenemase- producing Enterobacteriaceae (CPEs) and other types of carbapenem-resistant Enterobacteriaceae (CREs) has emphasized the need for alternative methods for identifying CREs. Multiple methods have been developed in response to this need with variable success. The main approaches to date include: i) tests based on carbapen-emase inhibitors; ii) growth tests based on inactivation of carbapen-ems; iii) tests based on detection of carbapenem hydrolysis products; iv) immunochromatographic tests to detect carbapenemases; and

Table I. Geographical location of carbapenemase-producing Enterobacteriaceae (CPE) isolated from animal clinical samples (continued).

Animal Continent Year Gene Country Bacterial species species Clinical sample Reference Enterobacter cloacae Dog Diagnostic sample Klebsiella pneumoniae Cat Diagnostic sample Klebsiella pneumoniae Horse Diagnostic sample 2009–2012 OXA-48 Germany Klebsiella pneumoniae Dog Diagnostic sample (23) Enterobacter cloacae Dog Diagnostic sample Escherichia coli Dog Diagnostic sample Klebsiella oxytoca Dog Diagnostic sample Klebsiella pneumoniae Cat Diagnostic sample Enterobacter cloacae Cat Diagnostic sample Escherichia coli Cat Diagnostic sample 2010 OXA-48 Germany Klebsiella pneumoniae Guinea pig Diagnostic sample (23) Klebsiella pneumoniae Rat Diagnostic sample Klebsiella pneumoniae Mouse Diagnostic sample Klebsiella pneumoniae Rabbit Diagnostic sample 2012 OXA-48 Germany Escherichia coli Dog Diagnostic sample (120) Klebsiella pneumoniae Dog Diagnostic sample 2015 OXA-48 France Escherichia coli Dog Rectum (121) 2015–2016 OXA-48 Spain Klebsiella pneumonia Bird Cloaca (122) Klebsiella pneumoniae Bird Cloaca Klebsiella pneumoniae Bird Cloaca Escherichia coli Bird Cloaca Enterobacter cloacae Bird Cloaca 2012 OXA-23 Belgium Acinetobacter sp. Horse Feces (123) 2016 OXA-181 Italy Escherichia coli Swine Feces (83) 2014 KPC-2 Spain Escherichia coli Bird Cloaca (124) 2011 VIM-1 Germany Salmonella Infantis Swine Feces (125) 2014 VIM-1 Spain Escherichia coli Bird Cloaca (124) 2014–2015 VIM-1 Spain Klebsiella pneumoniae Dog Rectum (24) 2015 VIM-1 Germany Escherichia coli Swine Colon (126) 2016 VIM-1 Germany Salmonella Infantis Swine Diagnostic sample (127) 2015–2016 NDM-5 UK Escherichia coli Dog Wound (128) 2018 NDM-5 Switzerland Escherichia coli Dog Wound (129) 2015 NDM Finland Escherichia coli Dog Ear (26) N/A NDM-1 Germany Salmonella Corvallis Bird Unknown (130)NDM — New Delhi metallo-b-lactamase; OXA — oxacillinases; KPC — Klebsiella pneumoniae carbapenemase; VIM — Verona integron-encoded metallo-b-lactamase.



Table II. Advantages and disadvantages of testing methods that confirm carbapenemase production in bacteria.

Testing method Example Advantages Disadvantages ReferenceCarbapenemase Boronic acid Inhibits class-A carbapenemases, Does not inhibit other classes (40) inhibitor-based tests derivatives such as KPC, commercially available of carbapenemases disks

Dipicolinic acid Inhibits MBLs, commercially available Does not inhibit other classes (41,42) and EDTA disks of carbapenemases

Temocillin Indicates the presence of OXA-48 like Not conclusive, optimal inhibitor (43) enzymes combinations have yet to be determined

Avibactam Inhibits OXA-48 like enzymes, can be Optimal inhibitor combinations have (44) used in addition to temocillin yet to be determined

Cloxacillin Inhibits AmpC enzymes, used to Optimal inhibitor combinations have (45) identify CREs with AmpC production yet to be determined and porin modifications

Growth tests based on Modified Hodge Results read after overnight Variable sensitivity and specificity (30,48,49) carbapenem inactivation test (MHT) incubation, visual identification of depending on agent used, inoculum carbapenemase production via clover and b-lactamase being identified, low leaf pattern, inexpensive sensitivity for MBLs and low specificity for CREs, best used in parallel with other methods

Carbapenem Results read after overnight Sensitivity and specificity of (48,51–53) inactivation incubation, visual identification of combination methods have yet method (CIM) carbapenemase production via lack to be elucidated of zone of inhibition, inexpensive, high sensitivity and specificity, can be combined with inhibitors for b-lactam type identification (e.g., cloxacillin, boronic acid, sodium mercaptoacetate)

Detection of carbapenem Carba NP test Results read after 2 h of incubation, Low sensitivity for detection of (48,51,55–58) hydrolysis products modifications have improved sensitivity OXA-48-like producers and maintained high specificity, variations of tests are commercially available and affordable

Immunochromatographic Tests identifying multiple High cost, unaffordable for routine (60,61) tests (ICT) carbapenemases have recently been screening of CPE and surveillance developed and validated, results are programs in animals read within minutes, high sensitivity and specificity, can be used with clinical samples, e.g., blood culture

Matrix-assisted laser Equipment present in diagnostic Struggles to identify OXA-48-like (65,66) desorption/ionization laboratories, detection of OXA-48 like producers, routine use for time-of-flight mass producers has been improved by surveillance laboratories is not spectrometry (MALDI-TOF) adding ammonium bicarbonate, practical due to lack of automation the use of inhibitors is strongly and standardization suggested to identify specific carbapenemase producers



v) Matrix-Assisted Laser Desorption/Ionization-Time of Flight Mass Spectrometry (MALDI-TOF). These methods are discussed here and their advantages and disadvantages are summarized in Table II.

Inhibitor-based methods — While inhibitor-based tests cannot identify all CRE classes, they provide useful information for at least some of them. These tests are used in conjunction with ASTs, mostly as part of disk diffusion tests, for which disks containing combinations of carbapenems and inhibitors are now commercially available. They can also be conducted in broth dilution methods, for which inhibitor tablets are also commercially available. There are 2 main types of inhibitors for carbapenemases. The first is composed of boronic acid derivatives that inhibit the activity of class-A carbapenemases and are used particularly to identify Klebseilla pneumonia carbapenenase (KPC) producers (40). The sec-ond includes dipicolinic acid and ethylenediaminetetraacetic acid (EDTA), which are chelating agents inhibiting metallo-b-lactamases (MBLs) (41,42).

The use of temocillin has been suggested to obtain more precise information about other types of carbapenemases and carbapenem resistance mechanisms since decreased susceptibility to this agent indicates, the presence of OXA-48-like enzymes, although not entirely conclusively (43). Avibactam specifically inhibits the activity of OXA-48-like enzymes on temocillin and has been suggested as an additional inhibition test to identify them (44). Cloxacillin, which inhibits AmpC enzymes, has also been recommended to identify CREs with combinations of AmpC production and porin modification (45). Although not fully validated and not always reliably identifying certain types of CPEs and CREs, combinations of these tests form the basis of some useful CPE confirmation schemes (46,47). See also the EUCAST guidelines available at: http://www.eucast.org/fileadmin/src/media/PDFs/EUCAST_

files/Resistance_mechanisms/EUCAST_detection_of_resistance_mechanisms_v1.0_20131211.pdf.

Growth tests based on inactivation of carbapenems — The 2 main tests for carbapenemases in this category are the modified Hodge test (MHT) (30) and the carbapenem inactivation method (CIM). For the MHT, a plate is first inoculated with a lawn of a carbapenem- susceptible indicator strain and a carbapenem- impregnated disk is placed in its center. Before overnight incubation, streaks of the strains to test for the presence of carbapenemase are drawn from the periphery of the plate to the disk. Results are read after overnight incubation. If a test strain is producing a carbapenemase, the enzyme will diffuse around the streak, allowing the indicator strain to grow around the streak closer to the carbapenem disk, which leads to the appearance of the “clover leaf” pattern shown in Figure 3a. Nothing will diffuse from CREs with resistance mechanisms other than car-bapenemases and the border of the inhibition zone of the susceptible strain will remain unchanged around the streak.

The effectiveness of the modified Hodge test (MHT) for detecting CPE has been examined in multiple studies and it has been shown to provide both false positive and false negative results under some conditions. Sensitivity and specificity vary among others as a function of the agent used for the test (meropenem vs. ertape-nem), inoculum used, and b-lactamases being investigated (48). Although the MHT is inexpensive and practical, both the sensitivity and specificity may be unsatisfactory. For example, problems may be encountered in its sensitivity in detecting MBLs such as NDM (30,48) and in its specificity in the presence of CREs resistant due to the combination of extended-spectrum b-lactamases and AmpC enzymes (48,49). Care should therefore be taken when using this test alone. It may be necessary to conduct other tests in parallel to compensate for these weaknesses (49). In evaluating the significance

Figure 3. Carbapenem-resistant Enterobacteriaceae (CRE) confirmation tests based on inactivation of carbapenemases: a — modified Hodge test using a meropenem disk (test strains counter-clockwise from top: NDM-producer, KPC-producer, OXA-48-producer, and CMY-producer); and b — carbapen-emase inactivation method (counter-clockwise from top right: NDM-producer, KPC-producer, OXA-48-producer, and CMY-producer).



of MHT results, it may also be helpful to know the geographical epidemiology of CPE and the local distribution of carbapenemase variants in order to predict whether or not it is identifying the most prevalent CPEs locally (49).

For the carbapenem inactivation method (CIM), a carbapenem disk, usually meropenem, is incubated with a suspension of the isolate under consideration and the disk is subsequently used for a disk diffusion test with a carbapenem-susceptible strain. If the isolate produces a carbapenemase, the antibiotic in the disk will be inactivated and no zone of inhibition will be observed around the disk in the subsequent disk diffusion test (Figure 3b). If the isolate is resistant to carbapenems because of a mechanism other than a carbapenemase, the disk will still contain enough carbapenem to induce the presence of an inhibition zone with the susceptible strain.

The CIM is a simple and inexpensive test with a relatively high sensitivity and specificity (48,50). Yamada et al (48) compared the effectiveness of 3 CPE detection methods (MHT, Carba NP, and CIM) and found that the CIM had the highest concordance rate to the results of the reference polymerase chain reaction (PCR). These results are corroborated by another study by Tijet et al (51), who also found a high sensitivity and specificity for CIM. The performance of CIM has recently been improved with different suspension media and incubation times (52,53). The CIM procedure can be further altered by adding inhibitors, such as cloxacillin, boronic acid, and sodium mercaptoacetate, to identify KPC and MBLs. While the sen-sitivity and specificity of these combination methods have yet to be assessed, they are worth further investigation (48).

Tests based on detection of carbapenem hydrolysis products — These tests consist of variations of the Carba NP test, which is a biochemical assay that detects hydrolysis of a carbapenem by bacte-rial extracts via a pH indicator (54). Modifications of the test have been proposed that either use different lysis procedures or entirely circumvent the need for bacterial extracts (55,56), while apparently improving its sensitivity and maintaining its high specificity (56,57).

Multiple studies have demonstrated satisfactory performances for this test, although one of its main weaknesses may be its relatively low sensitivity for detecting OXA-48-like producers (48,51,58). The main advantage of the Carba NP test and its derivatives over other phenotypic tests is that it can be read after 2 h of incubation. Multiple variants of the test are now commercially available and seem to rep-resent a fast and affordable method for identifying carbapenemase producers in clinical laboratories.

Immunochromatographic tests (ICTs) — Immunochromato-graphic tests are routinely used for point-of-care diagnostics in the form of lateral flow immunoassay. The principals of ICTs, including in veterinary medicine, are reviewed in Koczula and Gallotta (59). Although a number of ICTs for detecting a single carbapenemase or pairs have been described for characterizing bacterial cultures, they have only recently been developed and validated to detect a sufficiently broad panel of enzymes, i.e., NDM-, KPC-, IMP-, VIM-type, and OXA-48-like carbapenemases (60).

Immunochromatographic tests provide results within minutes and seem to present high levels of sensitivity, including for OXA-48-like enzymes, and specificity (60). They seem to even be applicable directly to blood culture for rapid human diagnostics (61). These

types of tests are still too expensive, however, for routine screening of CPEs and surveillance programs in animals.

Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF) — This type of spectrometry is now used routinely in diagnostic laboratories for rapidly identifying bacterial isolates. Recent studies have attempted to use it for detect-ing b-lactamases and carbapenemases in particular. Two approaches have been developed: one assesses the presence of specific peaks for the hydrolysis products of carbapenems when incubated together with the isolate under investigation (62,63) and the other is based on signature peaks for products of plasmids associated with car-bapenemases (64). Although it may have interesting applications for investigating outbreaks, this method is much narrower in scope and only the first approach will be discussed here.

As with other methods, the hydrolysis product detection method by MALDI-TOF has also struggled with OXA-48-like producers. This sensitivity issue was improved, however, by adding ammo-nium bicarbonate (NH4HCO3) to the reaction buffer (65). As with other tests, the use of inhibitors to further identify the group of car-bapenemases responsible for the observed hydrolysis has also been suggested for MALDI-TOF-based approaches (66). Despite these promising results, the routine use of MALDI-TOF for identifying carbapenemase-producers in diagnostic and surveillance laboratories may have to wait until automated and standardized protocols and instrument settings are established.

DNA-based methodsDeoxyribonucleic acid (DNA)-based methods are the gold stan-

dard for detecting CPEs (5,10,33). They are more commonly used for research purposes than for clinical diagnostics and surveillance, par-ticularly in animals, due to their high cost, frequent lack of standard-ization, and the need for specialized equipment and personnel (5). Although not necessarily able to detect new carbapenemase genes and variants (10), they are invaluable and essential procedures for characterizing resistant bacteria and can improve on current detec-tion methods by providing accurate and reliable comparative results.

Polymerase chain reaction (PCR) — This is the most commonly used tool to detect the presence of a gene of interest. Besides being fast, PCR is also able to detect carbapenemase genes that may be dif-ficult to detect otherwise, such as those for OXA-48-like enzymes (33). There are too many PCR protocols for carbapenemase genes in the literature to be described here. As for other antimicrobial resistance (AMR) genes, multiple variations and combinations of conventional (end-point), real-time, single, and multiplex protocols have been described, some of which are available commercially (67). These pro-tocols have targeted all the major groups of carbapenemases, includ-ing KPC, NDM, OXA-48 and OXA-48-like, VIM, IMP, SME, and GES.

Bialvaei et al (68) review PCR methods and protocols for car-bapenemase detection. Criteria and strain sets used to validate these numerous PCR tests are extremely variable and their performance in the real world should be carefully assessed before their use. Some official recommendations for primers for detecting the major carbapenemase genes are available, such as those from the European Union Reference Laboratory for Antimicrobial Resistance, avail-able at: https://www.eurl-ar.eu/CustomerData/Files/Folders/25-resourcer/459_primerliste-til-web-07-11-2018.pdf.



Polymerase chain reaction can be a first line test leading to fur-ther investigation, when, for example, testing bacterial isolates for surveillance (69,70). This method can also be used to directly screen clinical samples (71,72) with relatively low detection limits (73) and can also confirm results of a previous test that suggests the pres-ence of carbapenemase production (28). Finally, while PCR cannot necessarily identify the targeted carbapenemase gene to the variant level, sequencing of the PCR products can generally be used to identify them, thus providing much more detailed information for epidemiological investigations.

Microarrays — While providing acceptable, but slightly longer turnover times, microarrays have a major advantage over PCR-based approaches. Microarrays are able to test for the presence of a large number of genes at the same time and can also differentiate between variants on the basis of single nucleotide polymorphism (67). Multiple research groups have developed microarrays for the detection of AMR genes. Cross-validation against a collection of isolates with well-characterized carbapenemase genes has demonstrated excellent sensitivity and specificity (74,75). Microarray-based approaches have also been developed and assessed for direct testing of clinical samples (76,77). Due to the need to constantly adapt to new genes and gene variants, as well as for specialized equipment, however, only a few of these “homemade” and commercial platforms have withstood time.

New technologies combining DNA amplification and micro-array technologies have also been commercialized that promise to increase sensitivity of detection, while maintaining specificity (78). Unfortunately, these platforms remain relatively expensive and have not been widely adopted for surveillance of AMR in animal populations.

Genome sequencing — Whole genome sequencing (WGS) has become mainstream for epidemiological research and surveillance recently as a result of major technical progress. This has allowed bacterial strains to be characterized to a level unthinkable a decade

or 2 ago. It is now possible to investigate outbreaks and identify antimicrobial resistance genes, as well as assess virulence potential of a strain in almost real time through WGS (79). Despite many efforts in the development of dedicated bioinformatics tools, the short reads provided by earlier and current mainstream WGS plat-forms, e.g., https://www.illumina.com/, do not allow the consistent assembly of plasmids. Newer methods providing longer reads, such as the Oxford Nanopore platform (https://nanoporetech.com/) and the Pacific Biosciences platform (https://www.pacb.com/), are now providing the information needed to overcome the problems with long repeats in sequences that were hampering plasmid assembly.

This provides researchers with the tools to decipher the epidemi-ology of AMR and carbapenem resistance in particular at all levels, from the local (80) to global (21) spread of CRE and CPE clones, as well as the spread of CPE-encoding plasmids and associated CPE determinants (8). Future progress in the use of these same sequenc-ing platforms for metagenomics studies is bound to drastically deepen our understanding of AMR epidemiology, including of CREs and CPEs (81). Unfortunately, because of cost and the need for spe-cifically trained, highly qualified personnel, WGS is not yet a viable methodology for routine diagnostic and broad surveillance purposes. It is an essential tool, however, for understanding the epidemiology of antimicrobial resistance and the mobile genetic elements on which the carbapenemase genes are carried.

The first studies that used WGS for CPEs from animals have started trickling in and can be found in the literature. They have helped characterize Salmonella and E. coli CPE isolates and their plasmids encoding, for instance, KPC-4 (82), OXA-181 (83), VIM-1 (84), and IMP-4 (85). In one study, WGS allowed the authors to dem-onstrate the identity of NDM-5-carrying E. coli in dogs and a member of their owner’s family, thus suggesting a possible transfer between them (26). In another fascinating study, WGS allowed researchers to follow the persistence and transfer of an NDM-1-encoding plasmid

Table III. Comparison of selective culture media for isolating and detecting CPE.

Selective culture media Sensitivity (%) Specificity (%) Material ReferenceBrilliance CRE 75.9 98.1 Pure (90)Brilliance CRE 76.3 57.1 Pure (93)Brilliance CRE 78 to 82 60 to 66 Pure (94)Brilliance CRE 94 71 Pure (95)Brilliance CRE 1 Brilliance ESBL 98.1 67.6 Pure (90)ChomID ESBL 87.7 24.2 Pure (96)ChomID ESBL 96 to 97 6 to 19 Pure (94)SuperCARBA 96.5 76.3 Pure (93)SuperCARBA 95.6 82.2 Pure (96)CHROMagar KPC 100 98.4 Mixed (97)CHROMagar KPC 40.3 85.5 Pure (96)CHROMagar KPC 43 67.8 Pure (92)ChromArt CRE 100 55.8 Pure (90)BBL CHROMagar CPE 88.5 86.1 Pure (90)ChromID CARBA SMART 90.7 89.1 Pure (90)ChromID Carba 91 to 96 76 to 89 Pure (94)CPE — carbapenemase-producing Enterobacteriaceae; ESBL — extended-spectrum b-lactamase; KPC — Klebsiella pneumoniae carbapenemase.



in chickens (86). The authors were able to assess not only the dynam-ics of this plasmid in its original Salmonella Corvallis host, but also to follow its transfer to E. coli and Klebsiella pneumonia strains in vivo and its microevolution through recombination events (86). These studies on CPE in animals demonstrate the potential of WGS for the study of carbapenem resistance in bacteria from animals and for understanding its epidemiology in a more global context, including the animal/human/environment interface.

Media for CRE screeningSelective culture media are used to enhance the growth of bac-

teria that have a certain characteristic, while inhibiting the growth of unwanted bacteria. They can be used either as enrichment broth to increase the concentration of the bacteria of interest and increase sensitivity of detection or as a selective agar to more easily isolate and identify the target organism. Selective media often include indicators that allow colonies of specific organisms to be identified among those growing on a plate. Enrichment broth and selective agar can also be used sequentially to optimize sensitivity and specificity.

In the particular case of CRE, the instability of carbapenems in liquid solution has reduced the suitability of ready-made enrichment broth for CRE detection and has led to the use of carbapenem disks dipped into nutrient broth immediately before inoculation (87). The use of such enrichment broth has significantly increased the sensitivity of CRE-carrier detection in humans (88,89) and similar advantages can also be expected when assessing for the presence of carrier animals.

A variety of selective-indicator agar media for CRE is available on the market. Most of those currently available, e.g., CHROMagar, mSuperCARBA, CHROMID CARBA, Brilliance CRE Agar, can iden-tify E. coli and also provide putative differentiation of other CREs, such as fermenters from non-fermenter species, e.g., Pseudomonas spp. and Acinetobacter spp. While there are too many evaluations and comparative studies to cite here, a few examples are provided in Table III.

Challenges arise when selecting the correct carbapenem molecule and its concentration in the media, in the face of the diversity of MICs associated with the multiple types of carbapenemase-organism combinations present in the field (90,91). It seems that none of the commercially available media has been able to consistently achieve a combination of both sensitivity and specificity clearly above 90% when screening for CPE in general. This has led to the use of selec-tive agar to target specific types of CREs, such as CHROMagar KPC, CHROMID OXA-48, CHROMID CARBA (for KPC and MBLs), or combinations of media in a single plate, such as CHROMID CARBA SMART agar. It has recently been suggested that the new CHROMagar mSuperCARBA agar has significantly improved the detection of CRE from all the major carbapenemase-type producers at once (92). Further evaluation with clinical samples instead of pure cultures and on a larger scale is needed, however, to confirm these results for practical clinical applications (90).

In conclusion, there are many challenges in detecting CPEs and there is no single, cost-effective test that will detect every CPE. Promising strides have been made in identifying key features and combinations of tests to optimize the performance of detection methods, while reducing cost and time. These recent advances

make CPE detection and surveillance in animals more practicable and affordable.

Carbapenemase-producing Enterobacteriaceae (CPEs) have only recently been discovered in animals and may still be infrequent. Animals need to be screened for CPEs more frequently. Using some of the most specific and sensitive methods now available and described in this article, it is possible to thoroughly and reliably assess the current situation and monitor it in the future. This may be particularly important for assessing the dynamic transmission of these organisms between animals and humans, and vice versa, as well as their persistence once animals are colonized.

A c k n o w l e d g m e n tThe authors thank Gabhan Chalmers for reviewing the final ver-

sion of this article.

Re f e r e n c e s 1. Kahan JS, Kahan FM, Goegelman R, et al. Thienamycin, a new

b-lactam antibiotic. I. Discovery, taxonomy, isolation and physi-cal properties. J Antibiot (Tokyo) 1979;32:1–12.

2. Papp-Wallace KM, Endimiani A, Taracila MA, Bonomo RA. Carbapenems: Past, present, and future. Antimicrob Agents Chemother 2011;55:4943–4960.

3. Bryskier A. Antimicrobial Agents: Antibacterials and Antifun-gals. Washington DC: ASM Press, 2005:295.

4. Mcquade MS, VanNostrand V, Schariter J, Kanike JD, Forsyth RJ. Stability and compatibility of reconstituted ertapenem with commonly used i.v. infusion and coinfusion solutions. Am J Health Syst Pharm 2004;61:38–45.

5. Ansari M, Munir T, Saad N. Phenotypic identification, fre-quency distribution and antibiogram of carbapenemase produc-ing Enterobacteriaceae in clinical isolates. J Coll Physicians Surg Pak 2018;28:274–278.

6. Göttig S, Gruber TM, Stecher B, Wichelhaus TA, Kempf VA. In vivo horizontal gene transfer of the carbapenemase OXA-48 during a nosocomial outbreak. Clin Infect Dis 2015;60: 1808–1815.

7. Mathers AJ, Cox HL, Kitchel B, et al. Molecular dissection of an outbreak of carbapenem-resistant Enterobacteriaceae reveals intergenus KPC carbapenemase transmission through a pro-miscuous plasmid. MBio 2011;2:1–7.

8. Schweizer C, Bischoff P, Bender J, et al. Plasmid-mediated transmission of KPC-2 carbapenemase in Enterobacteriaceae in critically ill patients. Front Microbiol 2019;10:1–8.

9. Centers for Disease Control and Prevention. Laboratory pro-tocol for detection of carbapenem-resistant or carbapenemase- producing, Klebsiella spp. and E. coli from rectal swabs. Depart-ment of Health and Human Services 2009:1–6. Available from: https://www.cdc.gov/hai/pdfs/cre/CRE-guidance-508.pdf Last accessed October 6, 2019.

10. Levy Hara G, Gould I, Endimiani A, et al. Detection, treatment, and prevention of carbapenemase-producing Enterobacteriaceae: Recommendations from an International Working Group. J Chemother 2013;25:129–140.



11. Queenan AM, Bush K. Carbapenemases: The versatile b- lactamases. Clin Microbiol Rev 2007;20:440–458.

12. Patel G, Bonomo RA. “Stormy waters ahead:” Global emergence of carbapenemases. Front Microbiol 2013;4:48.

13. Poirel L, Potron A, Nordmann P. OXA-48-like carbapene-mases: The phantom menace. J Antimicrob Chemother 2012;67: 1597–1606.

14. Logan LK, Weinstein RA. The epidemiology of carbapenem-resistant Enterobacteriaceae: The impact and evolution of a global menace. J Infect Dis 2017;215:S28–S36.

15. Haldorsen B, Giske CG, Hansen DS, et al. Performance of the EUCAST disc diffusion method and two MIC methods in detection of Enterobacteriaceae with reduced susceptibil-ity to meropenem: The NordicAST CPE study. J Antimicrob Chemother 2018;73:2738–2747.

16. Potter RF, D’Souza AW, Dantas G. The rapid spread of carbapenem- resistant Enterobacteriaceae. Drug Resist Updat 2016;29:30–46.

17. Fattouh R, Tijet N, McGeer A, Poutanen SM, Melano RG, Patel SN. What is the appropriate meropenem MIC for screen-ing of carbapenemase-producing Enterobacteriaceae in low-prevalence settings? Antimicrob Agents Chemother 2015;60: 1556–1559.

18. Bulik CC, Fauntleroy KA, Jenkins SG, et al. Comparison of meropenem MICs and susceptibilities for carbapenemase-producing Klebsiella pneumoniae isolates by various testing methods. J Clin Microbiol 2010;48:2402–2406.

19. WHO, Critically Important Antimicrobials for Human Medicine. 6th revision. Geneva, Switzerland: World Health Organization, 2019.

20. Wang X, Wang Q, Cao B, et al. Retrospective observational study from a Chinese network of the impact of combination therapy versus monotherapy on mortality from carbapenem-resistant Enterobacteriaceae bacteremia. Antimicrob Agents Chemother 2018;63:1–13.

21. Peirano G, Matsumura Y, Adams MD, et al. Genomic epidemi-ology of global carbapenemase-producing Enterobacter spp., 2008–2014. Emerg Infect Dis 2018;24:1010–1019.

22. Ewers C, Klotz P, Leidner U, et al. OXA-23 and ISAba1-OXA-66 class D b-lactamases in Acinetobacter baumannii isolates from companion animals. Int J Antimicrob Agents 2017;49:37–44.

23. Pulss S, Stolle I, Stamm I, et al. Multispecies and clonal dis-semination of OXA-48 carbapenemase in Enterobacteriaceae from companion animals in Germany, 2009–2016. Front Microbiol 2018;9:1265.

24. González-Torralba A, Oteo J, Asenjo A, Bautista V, Fuentes E, Alós JI. Survey of carbapenemase-producing Enterobacteriaceae in companion dogs in Madrid, Spain. Antimicrob Agents Chemother 2016;60:2499–2501.

25. Chiotos K, Tamma PD, Flett KB, et al. Increased 30-day mortal-ity associated with carbapenem-resistant Enterobacteriaceae in children. Open Forum Infect Dis 2018;5:1–4.

26. Grönthal T, Österblad M, Eklund M, et al. Sharing more than friendship — Transmission of NDM-5 ST167 and CTX-M-9 ST69 Escherichia coli between dogs and humans in a family, Finland, 2015. Euro Surveill 2018;23:1–10.

27. Cui L, Lei L, Lv Y, et al. blaNDM-1-producing multidrug-resistant Escherichia coli isolated from a companion dog in China. J Glob Antimicrob Resist 2018;13:24–27.

28. EFSA Panel on Biological Hazards. Scientific opinion on car-bapenem resistance in food animal ecosystems. EFSA Journal 2013;11:1–70.

29. Abraham S, Wong HS, Turnidge J, Johnson JR, Trott DJ. Carbapenemase-producing bacteria in companion animals: A public health concern on the horizon. J Antimicrob Chemother 2014:69:1155–1157.

30. Girlich D, Poirel L, Nordmann P. Value of the modified Hodge test for detection of emerging carbapenemases in Enterobacteriaceae. J Clin Microbiol 2012;50:477–479.

31. CLSI. Performance Standards for Antimicrobial Disk Susceptibility Tests, 13th ed. Wayne: Clinical and Laboratory Standards Institute, 2018.

32. CLSI. Performance Standards for Antimicrobial Susceptibility Testing, 29th ed. CSLI supplement M100. Wayne: Clinical and Laboratory Standards Institute, 2019.

33. Koroska F, Göttig S, Kaase M, et al. Comparison of phenotypic tests and an immunochromatographic assay and development of a new algorithm for detection of OXA-48-like carbapen-emases. J Clin Microbiol 2017;55:877–883.

34. Nordmann P, Gniadkowski M, Giske CG, et al. Identification and screening of carbapenemase-producing Enterobacteriaceae. Clin Microbiol Infect 2012;18:432–438.

35. Turnidge J, Kahlmeter G, Kronvall G. Statistical characterisation of bacterial wild-type MIC value distributions and the determi-nation of epidemiological cut-off values. Clin Microbiol Infect 2006;12:418–425.

36. Lee M, Chung HS. Different antimicrobial susceptibility testing methods to detect ertapenem resistance in Enterobacteriaceae: VITEK2, MicroScan, Etest, disk diffusion, and broth microdilu-tion. J Microbiol Methods 2015;112:87–91.

37. Endimiani A, Hujer KM, Hujer AM, et al. Acinetobacter bauman-nii isolates from pets and horses in Switzerland: Molecular characterization and clinical data. J Antimicrob Chemother 2011;66:2248–2254.

38. Vading M, Samuelsen Ø, Haldorsen B, Sundsfjord AS, Giske CG. Comparison of disk diffusion, Etest and VITEK2 for detec-tion of carbapenemase-producing Klebsiella pneumoniae with the EUCAST and CLSI breakpoint systems. Clin Microbiol Infect 2011;17:668–674.

39. Jain JG, Sutherland C, Nicolau DP, Kuti JL. Stability of ertape-nem 100 mg/mL in polypropylene syringes stored at 25, 4, and −20°C. Am J Health Syst Pharm 2014;71:1480–1484.

40. Doi Y, Potoski BA, Adams-Haduch JM, Sidjabat HE, Pasculle AW, Paterson DL. Simple disk-based method for detection of Klebsiella pneumoniae carbapenemase-type beta-lactamase by use of a boronic acid compound. J Clin Microbiol 2008;46: 4083–4086.

41. Hansen F, Hammerum AM, Skov RL, Giske CG, Sundsfjord A, Samuelsen O. Evaluation of ROSCO Neo-Sensitabs for phenotypic detection and subgrouping of ESBL-, AmpC- and carbapenemase-producing Enterobacteriaceae. APMIS 2012;120:724–732.



42. Payne DJ, Cramp R, Bateson JH, Neale J, Knowles D. Rapid identification of metallo- and serine beta-lactamases. Antimicrob Agents Chemother 1994;38:991–996.

43. Hartl R, Widhalm S, Kerschner H, Apfalter P. Temocillin and meropenem to discriminate resistance mechanisms leading to decreased carbapenem susceptibility with focus on OXA-48 in Enterobacteriaceae. Clin Microbiol Infect 2013;19:E230–E232.

44. Huang TD, Berhin C, Bogaerts P, Glupczynski Y. Evaluation of avibactam-supplemented combination disk tests for the detec-tion of OXA-48 carbapenemase-producing Enterobacteriaceae. Diagn Microbiol Infect Dis 2014;79:252–254.

45. Giske CG, Gezelius L, Samuelsen Ø, Warner M, Sundsfjord A, Woodford N. A sensitive and specific phenotypic assay for detection of metallo-b-lactamases and KPC in Klebsiella pneu-moniae with the use of meropenem disks supplemented with aminophenylboronic acid, dipicolinic acid and cloxacillin. Clin Microbiol Infec 2011;17:552–556.

46. van Dijk K, Voets GM, Scharringa J, et al. A disc diffusion assay for detection of class A, B and OXA-48 carbapenemases in Enterobacteriaceae using phenyl boronic acid, dipicolinic acid and temocillin. Clin Microbiol Infec 2014;20:345–349.

47. Dortet L, Bernabeu S, Gonzalez C, Naas T. Evaluation of the Carbapenem Detection Set for the detection and characteriza-tion of carbapenemase-producing Enterobacteriaceae. Diagn Micr Infec Dis 2018;91:220–225.

48. Yamada K, Kashiwa M, Arai K, Nagano N, Saito R. Comparison of the Modified-Hodge test, Carba NP test, and carbapenem inactivation method as screening methods for carbapenemase-producing Enterobacteriaceae. J Microbiol Methods 2016;128: 48–51.

49. Carvalhaes CG, Picão RC, Nicoletti AG, Xavier DE, Gales AC. Cloverleaf test (modified Hodge test) for detecting carbapen-emase production in Klebsiella pneumoniae: Be aware of false positive results. J Antimicrob Chemother 2010;65:249–251.

50. van der Zwaluw K, de Haan A, Pluister GN, Bootsma HJ, de Neeling AJ, Schouls LM. The carbapenem inactivation method (CIM), a simple and low-cost alternative for the Carba NP test to assess phenotypic carbapenemase activity in Gram-negative rods. PLoS One 2015;10:1–13.

51. Tijet N, Boyd D, Patel SN, Mulvey MR, Melano RG. Evaluation of the Carba NP test for rapid detection of carbapenemase-producing Enterobacteriaceae and Pseudomonas aeruginosa. Antimicrob Agents Chemother 2013;57:4578–4580.

52. Pierce VM, Simner PJ, Lonsway DR, et al. Modified carbapenem inactivation method for phenotypic detection of carbapenemase production among Enterobacteriaceae. J Clin Microbiol 2017;55: 2321–2333.

53. Kuchibiro T, Komatsu M, Yamasaki K, et al. Evaluation of the modified carbapenem inactivation method for the detection of carbapenemase-producing Enterobacteriaceae. J Infect Chemother 2018;24:262–266.

54. Vasoo S, Cunningham SA, Kohner PC, et al. Comparison of a novel, rapid chromogenic biochemical assay, the carba NP test, with the modified Hodge test for detection of carbapenemase-producing Gram-negative bacilli. J Clin Microbiol 2013;51: 3097–3101.

55. Bakour S, Garcia V, Loucif L, et al. Rapid identification of carbapenemase-producing Enterobacteriaceae, Pseudomonas aeruginosa and Acinetobacter baumannii using a modified Carba NP test. New Microbes New Infect 2015;7:89–93.

56. Pasteran F, Tijet N, Melano RG, Corso A. Simplified protocol for Carba NP test for enhanced detection of carbapenemase producers directly from bacterial cultures. J Clin Microbiol 2015; 53:3908–3911.

57. Tamma PD, Opene BN, Gluck A, Chambers KK, Carroll KC, Simner PJ. Comparison of 11 phenotypic assays for accurate detection of carbapenemase-producing Enterobacteriaceae. J Clin Microbiol 2017;55:1046–1055.

58. Goire N, Harnett GB, O’Reilly LC, et al. The implications of endemic IMP-4 carbapenemase for clinical laboratory suscep-tibility testing. J Microbiol Methods 2016;124:10–12.

59. Koczula KM, Gallotta A. Lateral flow assays. Essays Biochem 2016;60:111–120.

60. Boutal H, Vogel A, Bernabeu S, et al. A multiplex lateral flow immunoassay for the rapid identification of NDM-, KPC-, IMP- and VIM-type and OXA-48-like carbapenemase- producing Enterobacteriaceae. J Antimicrob Chemother 2018;73: 909–915.

61. Bodendoerfer E, Keller PM, Mancini S. Rapid identification of NDM-, KPC-, IMP-, VIM- and OXA-48-like carbapenemase-producing Enterobacteriales from blood cultures by a multiplex lateral flow immunoassay. J Antimicrob Chemother 2019;74: 1749–1751.

62. Burckhardt I, Zimmermann S. Using matrix-assisted laser desorption ionization-time of flight mass spectrometry to detect carbapenem resistance within 1 to 2.5 hours. J Clin Microbiol 2011;49:3321–3324.

63. Hrabák J, Walková R, Študentová V, Chudácková E, Bergerová T. Carbapenemase activity detection by matrix-assisted laser desorption ionization-time of flight mass spectrometry. J Clin Microbiol 2011;49:3222–3227.

64. Lau AF, Wang H, Weingarten RA, et al. A rapid matrix-assisted laser desorption ionization-time of flight mass spectrometry-based method for single-plasmid tracking in an outbreak of carbapenem-resistant Enterobacteriaceae. J Clin Microbiol 2014;52:2804–2812.

65. Papagiannitsis CC, Študentová V, Izdebski R, et al. Matrix-assisted laser desorption ionization-time of flight mass spec-trometry meropenem hydrolysis assay with NH4HCO3, a reli-able tool for direct detection of carbapenemase activity. J Clin Microbiol 2015;53:1731–1735.

66. Monteferrante CG, Sultan S, ten Kate MT, et al. Evaluation of different pretreatment protocols to detect accurately clinical carbapenemase-producing Enterobacteriaceae by MALDI-TOF. J Antimicrob Chemother 2016;71:2856–2867.

67. Lupo A, Papp-Wallace KM, Sendi P, Bonomo RA, Endimiani A. Non-phenotypic tests to detect and characterize antibiotic resistance mechanisms in Enterobacteriaceae. Diagn Micr Infec Dis 2013;77:179–194.

68. Bialvaei AZ, Kafil HS, Asgharzadeh M, Yousef Memar M, Yousefi M. Current methods for the identification of carbapen-emases. J Chemother 2016;28:1–19.



69. Findlay J, Hopkins KL, Meunier D, Woodford N. Evaluation of three commercial assays for rapid detection of genes encod-ing clinically relevant carbapenemases in cultured bacteria. J Antimicrob Chemother 2015;70:1338–1342.

70. Weiß D, Engelmann I, Braun SD, Monecke S, Ehricht R. A mul-tiplex real-time PCR for the direct, fast, economic and simulta-neous detection of the carbapenemase genes blaKPC, blaNDM, blaVIM and blaOXA-48. J Microbiol Methods 2017;142:20–26.

71. Huang TD, Bogaerts P, Ghilani E, et al. Multicentre evalua-tion of the Check-Direct CPE® assay for direct screening of carbapenemase-producing Enterobacteriaceae from rectal swabs. J Antimicrob Chemother 2015;70:1669–1673.

72. Perry KA, Daniels JB, Reddy SC, et al. Direct detection of carbapenem-resistant organisms from environmental samples using the GeneXpert molecular diagnostic system. mSphere 2018;3:1–5.

73. Oviaño M, Torres I, González M, Bou G. Evaluation of a novel procedure for rapid detection of carbapenemase-producing Enterobacteriaceae (CPE) using the LightMix modular carbapen-emase kits. J Antimicrob Chemother 2016;71:3420–3423.

74. Braun SD, Monecke S, Thrümer A, et al. Rapid identification of carbapenemase genes in Gram-negative bacteria with an oligonucleotide microarray-based assay. PLoS One 2014;9:1–11.

75. Bernasconi OJ, Principe L, Tinguely R, et al. Evaluation of a new commercial microarray platform for the simultaneous detection of b-lactamase and mcr-1 and mcr-2 genes in Enterobacteriaceae. J Clin Microbiol 2017;55:3138–3141.

76. Dodémont M, De Mendonca R, Nonhoff C, Roisin S, Denis O. Performance of the verigene Gram-negative blood culture assay for rapid detection of bacteria and resistance determinants. J Clin Microbiol 2014;52:3085–3087.

77. Peter H, Berggrav K, Thomas P, et al. Direct detection and genotyping of Klebsiella pneumoniae carbapenemases from urine by use of a new DNA microarray test. J Clin Microbiol 2012; 50:3990–3998.

78. Torres Fink I, Tormo Palop N, Borrás Salvador R, Buesa Gómez J, Gimeno Cardona C, Navarro Ortega D. Evaluation of the DNA microarray “AMR Direct Flow Chip Kit” for detec-tion of antimicrobial resistance genes from Gram-positive and Gram-negative bacterial isolated colonies. Enferm Infecc Microbiol Clin 2019;37:454–457.

79. Kwong JC, Lane CR, Romanes F, et al. Translating genom-ics into practice for real-time surveillance and response to carbapenemase-producing Enterobacteriaceae: Evidence from a complex multi-institutional KPC outbreak. PeerJ 2018;6:1–32.

80. López-Camacho E, Paño-Pardo JR, Ruiz-Carrascoso G, et al. Population structure of OXA-48-producing Klebsiella pneumoniae ST405 isolates during a hospital outbreak characterised by genomic typing. J Glob Antimicrob Resist 2018;15:48–54.

81. Mu A, Kwong JC, Isles NS, et al. Reconstruction of the genomes of drug-resistant pathogens for outbreak investigation through metagenomic sequencing. mSphere 2019;4:1–13.

82. Daniels JB, Chen L, Grooters SV, et al. Enterobacter cloacae com-plex sequence type 171 isolates expressing KPC-4 carbapen-emase recovered from canine patients in Ohio. Antimicrob Agents Chemother 2018;62:1–6.

83. Pulss S, Semmler T, Prenger-Berninghoff E, Bauerfeind R, Ewers C. First report of an Escherichia coli strain from swine carrying an OXA-181 carbapenemase and the colistin resistance determinant MCR-1. Int J Antimicrob Agents 2017;50:232–236.

84. Roschanski N, Fischer J, Falgenhauer L, et al. Retrospective analysis of bacterial cultures sampled in German chicken-fattening farms during the years 2011–2012 revealed additional VIM-1 carbapenemase-producing Escherichia coli and a serologi-cally rough Salmonella enterica serovar infantis. Front Microbiol 2018;9:1–9.

85. Abraham S, O’Dea M, Trott DJ, et al. Isolation and plasmid char-acterization of carbapenemase (IMP-4) producing Salmonella enterica Typhimurium from cats. Sci Rep 2016;6:1–7.

86. Hadziabdic S, Fischer J, Malorny B, et al. In vivo transfer and microevolution of avian native IncA/C2 blaNDM-1-carrying plasmid pRH-1238 during a broiler chicken infection study. Antimicrob Agents Chemother 2018;62:1–16.

87. Darling LA, Evans AM, Stellrecht KA, Nattanmai SM, Montero CI. A triple-disk enrichment method for carbapenem- resistant Enterobacteriaceae (CRE) screening. J Clin Microbiol 2017;55:3557–3559.

88. Glaser L, Andreacchio K, Lyons M, Alby K. Improved surveil-lance for carbapenem-resistant Enterobacteriaceae using chroma-genic media with a broth enrichment. Diagn Microbiol Infect Dis 2015;82:284–285.

89. Peretz A, Pastukh N, Isakovich N, et al. Efficacy of an enrich-ment media for increasing threshold for carbapenem-resistant Enterobacteriaceae screening. J Clin Lab Anal 2016;30:563–566.

90. Hinic V, Amrein I, Stammler S, et al. Comparison of two rapid biochemical tests and four chromogenic selective media for detection of carbapenemase-producing Gram-negative bacteria. J Microbiol Methods 2017;135:66–68.

91. Nordmann P, Girlich D, Poirel L. Detection of carbapenemase producers in Enterobacteriaceae by use of a novel screening medium. J Clin Microbiol 2012;50:2761–2766.

92. Garcia-Quintanilla M, Poirel L, Nordmann P. CHROMagar mSuperCARBA and RAPIDEC Carba NP test for detection of carbapenemase-producing Enterobacteriaceae. Diagn Microbiol Infect Dis 2018;90:77–80.

93. Girlich D, Poirel L, Nordmann P. Comparison of the SUPERCARBA, CHROMagar KPC, and Brilliance CRE screening media for detection of Enterobacteriaceae with reduced susceptibil-ity to carbapenems. Diagn Microbiol Infect Dis 2013;75:214–217.

94. Wilkinson KM, Winstanley TG, Lanyon C, Cummings SP, Raza MW, Perry JD. Comparison of four chromogenic culture media for carbapenemase-producing Enterobacteriaceae. J Clin Microbiol 2012;50:3102–3104.

95. Cohen Stuart J, Voets G, Rottier W, et al. Evaluation of the Oxoid Brilliance CRE Agar for the detection of carbapenemase-producing Enterobacteriaceae. Eur J Clin Microbiol Infect Dis 2013;32:1445–1449.

96. Nordmann P, Girlich D, Poirel L. Detection of carbapenemase producers in Enterobacteriaceae by use of a novel screening medium. J Clin Microbiol 2012;50:2761–2766.

97. Samra Z, Bahar J, Madar-Shapiro L, Aziz N, Israel S, Bishara J. Evaluation of CHROMagar KPC for rapid detection of



carbapenem- resistant Enterobacteriaceae. J Clin Microbiol 2008;46:3110–3111.

98. Shaheen BW, Nayak R, Boothe DM. Emergence of a New Delhi metallo-b-lactamase (NDM-1)-encoding gene in clinical Escherichia coli isolates recovered from companion animals in the United States. Antimicrob Agents Chemother 2013;57:2902–2903.

99. Liu X, Thungrat K, Boothe DM. Occurrence of OXA-48 car-bapenemase and other b-lactamase genes in ESBL-producing multidrug resistant Escherichia coli from dogs and cats in the United States, 2009–2013. Front Microbiol 2016;7:1–10.

100. Sellera FP, Fernandes MR, Ruiz R, et al. Identification of KPC-2-producing Escherichia coli in a companion animal: A new challenge for veterinary clinicians. J Antimicrob Chemother 2018;73:2259–2261.

101. Al Bayssari C, Olaitan AO, Dabboussi F, Hamze M, Rolain J-M. Emergence of OXA-48-producing Escherichia coli clone ST38 in fowl. Antimicrob Agents Chemother 2015;59:745–746.

102. Liu X, Liu H, Li Y, Hao C. High prevalence of b-lactamase and plasmid-mediated quinolone resistance genes in extended-spectrum cephalosporin-resistant Escherichia coli from dogs in Shaanxi, China. Front Microbiol 2016;7:1–9.

103. Pruthvishree BS, Vinodh Kumar OR, Sinha DK, et al. Spatial molecular epidemiology of carbapenem-resistant and New Delhi metallo b-lactamase (blaNDM)-producing Escherichia coli in the piglets of organized farms in India. J Appl Microbiol 2017;122:1537–1546.

104. Pruthvishree BS, Kumar ORV, Sivakumar M, et al. Molecular characterization of extensively drug resistant (XDR), extended spectrum b-iactamases (ESBL) and New Delhi metallo b- lactamase-1 (blaNDM1) producing Escherichia coli isolated from a male dog — A case report. Vet Arhiv 2018;88:139–148.

105. He T, Wei R, Zhang L, et al. Characterization of NDM-5-positive extensively resistant Escherichia coli isolates from dairy cows. Vet Microbiol 2017;207:153–158.

106. He T, Wang Y, Sun L, Pang M, Zhang L, Wang R. Occurrence and characterization of blaNDM-5 -positive Klebsiella pneumoniae isolates from dairy cows in Jiangsu, China. J Antimicrob Chemother 2017;72:90–94.

107. Yang RS, Feng Y, Lv XY, et al. Emergence of NDM-5- and MCR-1-producing Escherichia coli clones ST648 and ST156 from a single Muscovy duck (Cairina moschata). Antimicrob Agents Chemother 2016;60:6899–6902.

108. Sun J, Yang RS, Zhang Q, et al. Co-transfer of blaNDM-5 and mcr-1 by an IncX3–X4 hybrid plasmid in Escherichia coli. Nat Microbiol 2016;1:1–4.

109. Hong JS, Song W, Park HM, et al. First detection of New Delhi metallo-b-lactamase-5-producing Escherichia coli from compan-ion animals in Korea. Microb Drug Resist 2019;25:344–349.

110. Liu Z, Wang Y, Walsh TR, et al. Plasmid-mediated novel blaNDM-17 gene encoding a carbapenemase with enhanced activ-ity in a sequence type 48 Escherichia coli strain. Antimicrob Agents Chemother 2017;61:1–9.

111. Braun SD, Ahmed MFE, El-Adawy H, et al. Surveillance of extended-spectrum beta-lactamase-producing Escherichia coli in dairy cattle farms in the Nile Delta, Egypt. Front Microbiol 2016;7:1–14.

112. Hamza E, Dorgham SM, Hamza DA. Carbapenemase-producing Klebsiella pneumoniae in broiler poultry farming in Egypt. J Glob Antimicrob Resist 2016;7:8–10.

113. Yousfi M, Touati A, Mairi A, et al. Emergence of carbapenemase-producing Escherichia coli isolated from companion animals in Algeria. Microb Drug Resist 2016;22:342–346.

114. Bachiri T, Bakour S, Lalaoui R, et al. Occurrence of carbapenemase- producing Enterobacteriaceae isolates in the wildlife: First report of OXA-48 in wild boars in Algeria. Microb Drug Resist 2018;24:337–345.

115. Bouaziz A, Loucif L, Ayachi A, Guehaz K, Bendjama E, Rolain JM. Migratory white stork (Ciconia ciconia): A potential vector of the OXA-48-producing Escherichia coli ST38 clone in Algeria. Microb Drug Resist 2018;24:461–468.

116. Yousfi M, Touati A, Muggeo A, et al. Clonal dissemination of OXA-48-producing Enterobacter cloacae isolates from com-panion animals in Algeria. J Glob Antimicrob Resist 2018;12: 187–191.

117. Yaici L, Haenni M, Saras E, Boudehouche W, Touati A, Madec JY. blaNDM-5-carrying IncX3 plasmid in Escherichia coli ST1284 isolated from raw milk collected in a dairy farm in Algeria. J Antimicrob Chemother 2016;71:2671–2672.

118. Dolejska M, Masarikova M, Dobiasova H, et al. High prevalence of Salmonella and IMP-4-producing Enterobacteriaceae in the silver gull on Five Islands, Australia. J Antimicrob Chemother 2016;71:63–70.

119. Schmiedel J, Falgenhauer L, Domann E, et al. Multiresistant extended-spectrum b-lactamase-producing Enterobacteriaceae from humans, companion animals and horses in central Hesse, Germany. BMC Microbiol 2014;14:1–13.

120. Stolle I, Prenger-Berninghoff E, Stamm I, et al. Emergence of OXA-48 carbapenemase-producing Escherichia coli and Klebsiella pneumoniae in dogs. J Antimicrob Chemother 2013;68:2802–2808.

121. Melo LC, Boisson MNG, Saras E, et al. OXA-48-producing ST372 Escherichia coli in a French dog. J Antimicrob Chemother 2016;72:1256–1258.

122. Oteo J, Mencía A, Bautista V, et al. Colonization with Enterobacteriaceae-producing ESBLs, AmpCs, and OXA-48 in wild avian species, Spain 2015–2016. Microb Drug Resist 2018;24:932–938.

123. Smet A, Boyen F, Pasmans F, et al. OXA-23-producing Acinetobacter species from horses: A public health hazard? J Antimicrob Chemother 2012;67:3009–3010.

124. Vergara A, Pitart C, Montalvo T, et al. Prevalence of ESBL and/or carbapenemase-producing Escherichia coli isolated from yellow-legged gulls from Barcelona, Spain. Antimicrob Agents Chemother 2016;61:1–5.

125. Fischer J, Rodríguez I, Schmoger S, et al. Salmonella enterica subsp. enterica producing VIM-1 carbapenemase isolated from livestock farms. J Antimicrob Chemother 2013;68:478–480.

126. Irrgang A, Fischer J, Grobbel M, et al. Recurrent detection of VIM-1-producing Escherichia coli clone in German pig produc-tion. J Antimicrob Chemother 2016;72:944–946.

127. Borowiak M, Szabo I, Baumann B, et al. VIM-1-producing Salmonella Infantis isolated from swine and minced pork meat in Germany. J Antimicrob Chemother 2017;72:2131–2133.



128. Reynolds ME, Phan HTT, George S, et al. Occurrence and char-acterization of Escherichia coli ST410 co-harbouring blaNDM-5, blaCMY-42 and blaTEM-190 in a dog from the UK. J Antimicrob Chemother 2019;74:1207–1211.

129. Peterhans S, Stevens MJA, Nüesch-Inderbinen M, Schmitt S, Stephan R, Zurfluh K. First report of a blaNDM-5-harbouring

Escherichia coli ST167 isolated from a wound infection in a dog in Switzerland. J Glob Antimicrob Resist 2018;15:226–227.

130. Fischer J, Schmoger S, Jahn S, Helmuth R, Guerra B. NDM-1 carbapenemase-producing Salmonella enterica subsp. enterica serovar Corvallis isolated from a wild bird in Germany. J Antimicrob Chemother 2013;68:2954–2956.


Article


I n t r o d u c t i o nFeline infectious peritonitis (FIP) is a fatal, multi-systemic dis-

ease for which no treatment or effective vaccine is currently avail-able (1). This disease can occur in cats following fecal-oral infection with feline enteric coronavirus (FeCV). An infection with FeCV will usually only result in mild clinical gastrointestinal signs. In approximately 5% of cats infected with FeCV, however, this virus mutates into an FIP-causing virus (FIPV) (2). It should be noted that 2 different FeCV serotypes, types 1 and 2, circulate in the feline

population (3). Type 1 is most prevalent in naturally infected popula-tions, whereas FeCV type 2 has emerged as a recombination of canine and feline coronaviruses and is relatively rare. Both serotypes can mutate and cause FIP (4).

In a single cat, FeCV and FIPV are genetically very closely related and only a few mutations are detectable when the enterocyte- infecting FeCV is compared to the monocyte- and macrophage- infecting FIPV. While the exact genetic change that mediates the increase in pathoge-nicity is unknown, there is strong evidence that specific mutations in the spike-protein (S) coding region play a major role in this transition

Prevalence and mutation analysis of the spike protein in feline enteric coronavirus and feline infectious peritonitis detected in household

and shelter cats in western CanadaLaura A. McKay, Melissa Meachem, Elisabeth Snead, Terri Brannen, Natasha Mutlow, Liz Ruelle,

Jennifer L. Davies, Frank van der Meer

A b s t r a c tFeline infectious peritonitis (FIP) is a fatal disease for which no simple antemortem diagnostic assay is available. A new polymerase chain reaction (PCR) test has recently been developed that targets the spike protein region of the FIP virus (FIPV) and can identify specific mutations (M1030L or S1032A), the presence of which indicates a shift from feline enteric coronavirus (FeCV) to FIPV. This test will only be useful in the geographical region of interest, however, if the FIP viruses contain these mutations. The primary objective of this study was to determine the presence of the M1030L or S1032A mutations in FeCV derived from stool samples from a selected group of healthy cats from households and shelters and determine how many of these cats excrete FeCV. The secondary objective was to evaluate how often these specific FIPV mutations were present in tissue samples derived from cats diagnosed with FIP at postmortem examination. Feline enteric coronavirus (FeCV) was detected in 46% of fecal samples (86/185), all were FeCV type 1, with no difference between household or shelter cats. Only 45% of the FIPV analyzed contained the previously reported M1030L or S1032A mutations. It should be noted that, as the pathological tissue samples were opportunistically obtained and not specifically obtained for PCR testing, caution is warranted in interpreting these data.

R é s u m éLa péritonite infectieuse féline (FIP) est une maladie fatale pour laquelle il n’existe pas de test diagnostique ante-mortem simple. Une nouvelle épreuve d’amplification en chaîne par la polymérase (PCR) a récemment été développée et qui vise la région de la protéine de spicule du virus FIP (FIPV) et peut identifier les mutations spécifiques (M1030L ou S1032A), la présence desquelles indique un glissement du coronavirus entérique félin (FeCV) vers le FIPV. Cette épreuve sera utile uniquement dans la région géographique d’intérêt, toutefois, si les virus FIP ont ces mutations. L’objectif premier de la présente étude était de déterminer la présence des mutations M1030L ou S1032A chez FeCV obtenu d’échantillons de fèces provenant d’un groupe sélectionné de chats en santé issus de maisonnée et refuges et de déterminer combien de ces chats excrètent FeCV. L’objectif secondaire était d’évaluer à quelle fréquence ces mutations spécifiques de FIPV étaient présentes dans des échantillons de tissu provenant de chats diagnostiqués avec FIP lors d’examen post-mortem. Le FeCV fut détecté dans 46 % des échantillons fécaux (86/185), tous de type FeCV 1, et aucune différence notée entre les chats de maisonnée ou de refuge. Seulement 45 % des FIPV analysés contenaient les mutations M1030L ou S1032A rapportées précédemment. Il faut noter que, étant donné que les échantillons de tissus pathologiques furent obtenus de manière opportuniste et non spécifiquement obtenus pour analyse par PCR, l’interprétation des résultats est à faire avec précaution.


Department of Ecosystem and Public Health (McKay, Davies, van der Meer) and Diagnostic Services Unit (Davies), Faculty of Veterinary Medicine, University of Calgary, Calgary, Alberta; Department of Veterinary Pathology (Meachem) and Department of Small Animal Clinical Sciences (Snead), Western College of Veterinary Medicine, University of Saskatchewan, Saskatoon, Saskatchewan; Meow Foundation, Calgary, Alberta (Brannen); Fish Creek Pet Hospital, Calgary, Alberta (Mutlow); Wild Rose Cat Clinic, Calgary, Alberta (Ruelle).

Address all correspondence to Dr. Frank van der Meer; e-mail: [email protected]

Received February 27, 2019. Accepted April 18, 2019.



(5–7). M1030L or S1032A mutations in the S-protein genomic region of this coronavirus have been identified in viruses from tissues of . 95% of FIP cats that were examined in Europe (5,6).

The clinical diagnosis of FIP is complex, especially in the early stages of the disease. While histopathology combined with immuno-histochemistry to detect FeCV antigens is the gold standard, this requires tissue obtained through biopsy or necropsy (8). Polymerase chain reaction (PCR) on blood or aspirates is an attractive alternative for diagnosing infectious diseases. A new PCR has recently been developed that could be a major diagnostic improvement (5,9). The primary objective of this study was to determine if the FIPV-specific M1030L and/or S1032A mutations can also be detected in healthy cats by analyzing the spike protein in stool samples obtained from shelter and household cats. The secondary objective was to assess the presence of these mutations in the S-protein genomic region in coronaviruses of cats diagnosed with FIP at postmortem examination in Saskatchewan and Alberta.

M a t e r i a l s a n d m e t h o d s

SamplesThis protocol was approved by the University of Calgary Animal

Care Committee (VSACC AC16-0268).Fresh fecal samples were collected from 185 domestic and stray

cats on intake to various shelters, veterinary clinics, and from pri-vate cat owners in Calgary. Of all fecal samples obtained, 130 were derived from cats that were offered to shelters and 55 were sampled in a veterinary clinic or at home by their owners. All cats were clinically healthy. With a conservative estimated prevalence of 33% to 50% in adult cats, we anticipated that this would result in 61 to 93 unique strains of FeCV (1,10).

We obtained 63 formalin-fixed, paraffin-embedded (FFPE) archived tissue samples, 4 frozen tissue samples, and 2 postmortem peritoneal fluid samples from the Diagnostic Services Unit at the University of Calgary Faculty of Veterinary Medicine and the Department of Pathology at the Western College of Veterinary Medicine. Tissue samples were selected on the basis of typical histopathological lesions that are indicative of FIPV infections (granulomatous inflammation and vasculitis, samples FIP 13, 14, and 17), preferably supported with a positive result on immunohistochemistry for the FeCV anti-gen (samples FIP 1 to 12 and 16 and 18). A few cases were included with only gross lesions as the basis of the FIP diagnosis (samples FIP 15, 19, and 20). Multiple tissue types could be present in a single paraffin block.

All samples came from a pool of spayed, neutered, or intact male and female cats, ranging in age from 3 mo to 21 y. Representative breeds included Siamese, Ragdoll, LaPerm, Persian, Himalayan, Sphynx, Russian blue, and Cornish rex, as well as a number of mixed breeds. Unfortunately, precise signalment was not available on many of the shelter cats, but they fell within the above-mentioned age range.

If known, the region, city, or quadrant of the city of Calgary where the samples were taken was included in the name of the sequences used for analysis, as indicated in Figures 2 and 3. All diagnosed FIP cases are indicated with ‘FIP’, whereas all positive fecal samples are indicated as ‘FeCV’.

RNA extractionViral ribonucleic acid (RNA) was extracted from the fecal samples

using either an EZNA Universal Pathogen Kit (Omega Bio-tek, Norcross, Georgia, USA) or a Norgen Stool Viral RNA Kit (Norgen Biotek, Thorold, Ontario) as per manufacturer’s instructions. The FFPE tissue was sectioned on a microtome and 3 to 5 10-mm scrolls were obtained from each sample. The Ambion RecoverAll Kit (ThermoFisher Scientific, Waltham, Massachusetts, USA) was used to extract viral RNA from the FFPE scrolls as directed. Viral RNA from frozen tissue and peritoneal fluid was extracted using Trizol reagent (Sigma Aldrich Canada, Oakville, Ontario). All RNA was assessed for quantity and quality with a Nanodrop spectrometer (ThermoFisher Scientific) and frozen at 220°C.

PCRFeces, FFPE tissue, peritoneal fluid, and frozen tissue were

screened to identify animals positive for FeCV by using a nested reverse transcription polymerase chain reaction (RT-PCR) targeting the highly conserved 39-untranslated region (39UTR) of the viral genome (11). Gene-specific complementary deoxyribonucleic acid (DNA) was synthesized using Superscript Reverse Transcriptase II (Invitrogen, Carlsbad, California, USA) and the antisense primer U211R (see Table I) as per manufacturer’s instructions. We con-ducted a nested PCR using Phusion High-Fidelity DNA Polymerase (Invitrogen) and specific primers (U211R and U204F) for the first reaction and specific primers (U276F and U205R) for the second reaction. PCR cycling conditions were 98°C for 30 s, followed by 35 cycles of 98°C for 30 s, 55°C for 30 s, and 72°C for 30 s, with an additional final elongation step at 72°C for 10 min. The expected product size was 223 basepair (bp) for the first reaction and 177 bp for the second reaction.

Samples positive for 39UTR were then reverse transcribed using Superscript Reverse Transcriptase II (Invitrogen) and the antisense primer S585R to produce gene-specific complementary DNA for the S protein region of interest. We conducted a nested PCR with cycling conditions identical to those previously described using specific primers (S585R and S866F) for the first reaction and specific primers (S876F and S1000R) for the second reaction. The expected product size was 598 bp for the first reaction and 142 bp for the second reaction (5).

Figure 1. Detection of FeCV in stool samples from healthy cats using PCR in Calgary and surrounding areas.



Amplified products were separated on a 1% agarose gel by elec-trophoresis, visualized with Safeview (Applied Biological Materials, Richmond, British Columbia), and the bands were cut out for DNA extraction using an EZNA DNA Gel Extraction Kit (Omega Bio-tek) as per manufacturer’s instructions. Extracted DNA was sent to the University of Calgary Core DNA Services in Calgary, Alberta or the TCAG DNA Sequencing Facility at Sick Kids Hospital in Toronto, Ontario for Sanger sequencing of the amplified products.

Phylogenetic analysisThe nucleotide fragments were translated into amino acid

sequences in Geneious 10.2.3 (Biomatters, Auckland, New Zealand) and aligned with Clustal Omega (12). Amino acid sequence number-ing was based on a type-1 feline coronavirus (GenBank accession number NC_002306). jModelTest (13) was used to determine the

most appropriate evolutionary model for our data. The generalized time-reversible model was selected to construct a maximum likeli-hood phylogenetic tree using PhyML (14).

Statistical analysisThe occurrence of FeCV in household and shelter cats was ana-

lyzed using Chi-squared tests with P , 0.05 considered significant.

Re s u l t s

Prevalence of FeCV in samples from Calgary and surrounding areas

Overall, 46% of cats sampled (86/185) tested positive for feline enteric coronavirus (FeCV) in their feces. The percentage of shelter

Figure 2. Unrooted nucleic acid maximum likelihood tree of the partial spike protein sequence of FeCV and FIPV as detected in western Canada.



Figure 3. Amino acid sequence alignment of the partial FeCV and FIPV spike proteins. Agreements with the consensus are gray; disagreements with the consensus are highlighted. Numbering on top is based on the amino acid positions in the spike protein of the reference genome: GenBank acces-sion number NC_002306.



cats that were FeCV-positive (43%, 56/130) did not differ from the percentage of positive household cats (55%, 30/55), X2 (1, n = 185) = 2.0434, P = 1.58 (Figure 1). No difference was found between the per-centage of FeCV-positive cats originating from different quadrants of the city and those from outside city limits, X2 (5, n = 185) = 7.298, P = 0.199.

Phylogenetic analysisAll sequences were of the FeCV type 1. We found no difference

between FeCV and FIPV samples, as the clustering and distribution appeared to be randomly dispersed in the phylogenetic tree. No clear transmission or relation between samples could be determined (Figure 2). Many of the FIPV and FeCV strains showed a very similar amino acid sequence (Figure 3). As most of the mutations observed in the nucleotide sequence did not result in an amino acid mutation (silent mutation), the variability in the amino acid sequence is limited and only the alignment is presented.

Mutation analysisWe were able to amplify the S-protein region of interest from

35/86 39UTR positive FeCV samples and 20/49 39UTR positive FIPV samples (Figure 3). The M1030 and S1032 loci were analyzed for the presence of mutations. None of the 35 FeCV samples had a mutation at either location. In the samples derived from FIP-positive cats, however, 8/20 of the S-protein gene regions sequenced had a M1030L mutation and 1/20 had an S1032A mutation. In 3/20 FIP samples and 1/35 FeCV samples, a T1034S mutation was detected, whereas 8/20 FIP samples had no mutation at any of these locations. One other locus, D1014, could be identified that contained several different mutations in both FIPV and FeCV samples, including D1014G (n = 2), D1014A (n = 3), D1014S (n = 3), and D1014Q (n = 3).

D i s c u s s i o nFeline coronavirus type 1 was the only serotype we could detect

in the fecal and tissue samples collected. This is consistent with previous studies that found FeCV type 1 to be the most commonly isolated serotype in North American and European populations of cats (4,7,10). Our observed number of positive cats (46%) is compa-rable to the results of previous reports (1,10). We expected a higher prevalence of FeCV-infected cats in shelters than in household cats. Interestingly, this was not the case, which may be explained by the high proportion of household cats that were either origi-

nally obtained from a rescue or were from multi-cat households in our sample. Additionally, we collected feces on intake to the shelter before the cat could be infected with strains circulating in the shelter. It has consistently been found that FeCV infection is significantly more prevalent in multi-cat households, catteries, and shelters (10,15).

While the FIP virus is a mutated variant of FeCV, it is not yet fully understood how this mutation arises or where it is located in the genome. Phylogenetic analysis of our FeCV and FIPV samples showed no clustering due to the geographic origin or FeCV versus FIPV infection. This supports the internal mutation theory of FIP infection that is favored by most researchers. In this theory, the mutation causing FIP arises de novo in each cat and cannot be trans-mitted (15,16). The alternative theory postulates that circulating virulent virus (FIP) and non-virulent strains (FeCV) are present (17), although we were unable to detect a separate FIPV lineage in FIP cats in this study.

We carried out a mutation analysis on the specific region of the S-protein gene targeted by the IDEXX FIP Virus RealPCR test (9). Chang et al (5) reported that . 95% of FIP-positive cats in Europe could be identified by mutations in 1 of 2 amino acids (M1030L or S1032A). Lewis et al (6) also reported the M1030L mutation in all 3 of the FIP genomes they sequenced. In contrast, we found that only 40% of FIP-positive cats in this study had the M1030L mutation and only 1 out of 20 FIP-derived sequences had the S1032A mutation. In our samples, 40% had no mutations in this region, while 15% had a T1034S mutation. We expected the majority of our FIP samples to have 1 or both mutations. Five of these no-mutation samples may have contained intestinal tissue and we could have amplified wildtype FeCV. Six pathological samples did not contain intestines, however, and did not have either mutation present. Another explana-tion could be that the virus presents itself as a quasispecies and we amplified the non-mutated form of the virus (18). Next-generation sequencing would be required to answer this question. A third option is that FIPV found in western Canada may have additional or alternative virulence sites that have not yet been identified. It has been postulated that the furin cleavage site within the S protein is a possible virulence region (7), which would be worthwhile investi-gating. If this is the case, caution would be advised in interpreting negative results on IDEXX’s FIP Virus RealPCR test.

While there was a region of increased variability in both FeCV and FIP samples at position D1014 of the S-protein genome, these were all conservative mutations that resulted in amino acids with

Table I. Primers used for FeCV and FIPV detection and characterization.

Primer name Reaction Primer sequence Product sizeP204 Forward UTR1 59-CACTAGATCCAGACGTTAGCTC-39 223 bpP211 Reverse UTR1 59-GCTCTTCCATTGTTGGCTCGTC-39 223 bpP276 Forward UTR2 59-CCGAGGAATTACTGGTCATCGCG-39 177 bpP205 Reverse UTR2 59-GGCAACCCGATGTTTAAAACTGG-39 177 bpS866 Forward S1 59-CAATATTACAATGGCATAATGG-39 598 bpS585 Reverse S1 59-CCCTCGAGTCCCGCAGAAACCATACCTA-39 598 bpS877 Forward S2 59-GGCATAATGGTTTTACCTGGTG-39 142 bpS1000 Reverse S2 59-TAATTAAGCCTCGCCTGCACTT-39 142 bp



similar biochemical properties, which are unlikely to change the functionality, structure, or antigenicity of the S-protein.

In our study, 45% of the FIP viruses derived from clinical cases in western Canada contained 1 of the previously described M1030L or S1032A mutations. While this supports the idea that these mutations play an important role in the virulence of FIPV, it suggests that a large proportion of FIPV is lacking either mutation.

A c k n o w l e d g m e n t sThis study was funded by the University of Calgary, Alberta. The

authors extend special thanks to the University of Calgary Veterinary Medicine’s Diagnostic Services Unit for assistance with sample collection, to Dr. Grace Kwong from the University of Calgary for assistance with statistical analysis, and to Dr. Christian Leutenegger, Dr. Wen Chang, and Dr. Herman Egberink for helpful discussion and Tammy Mazubert from the Calgary Humane Society for assistance in taking samples.

Re f e r e n c e s 1. Kahn CM, Line S, eds. The Merck Veterinary Manual. 10th ed.

Whitehouse Station, New Jersey: Merck, 2010:7072718. 2. Kipar A, Meli ML. Feline infectious peritonitis: Still an enigma?

Vet Pathol 2014;51:505–526. 3. Kummrow M, Meli ML, Haessig M, et al. Feline coronavirus

serotypes 1 and 2: Seroprevalence and association with disease in Switzerland. Clin Diagn Lab Immunol 2005;12:1209–1215.

4. Pedersen NC. An update on feline infectious peritonitis: Virology and immunopathogenesis. Vet J 2014;201:123–132.

5. Chang HW, Egberink HF, Halpin R, Spiro DJ, Rottier PJ. Spike protein fusion peptide and feline coronavirus virulence. Emerg Infect Dis 2012;18:1089–1095.

6. Lewis CS, Porter E, Matthews D, et al. Genotyping coronavi-ruses associated with feline infectious peritonitis. J Gen Virol 2015;96:1358–1368.

7. Licitra BN, Millet JK, Regan AD, et al. Mutation in spike protein cleavage site and pathogenesis of feline coronavirus. Emerg Infect Dis 2013;19:1066–1073.

8. Giori L, Giordano A, Giudice C, Grieco V, Paltrinieri S. Perfor-mances of different diagnostic tests for feline infectious perito-nitis in challenging clinical cases. J Small Anim Pract 2011;52: 152–157.

9. IDEXX Laboratories. Diagnostic update April 2015. Available from: https://idexxcom-live-b02da1e51e754c9cb292133b-9c56c33.aldryn-media.com/filer_public/dd/81/dd8192c9-ae9b-4d1b-841a-06748adc75ac/feline-infectious-peritonitis-virus.pdf Last accessed September 22, 2019.

10. Pedersen NC, Sato R, Foley JE, Poland AM. Common virus infections in cats, before and after being placed in shelters, with emphasis on feline enteric coronavirus. J Feline Med Surg 2004; 6:83–88.

11. Herrewegh AA, de Groot RJ, Cepica A, Egberink HF, Horzinek MC, Rottier PJ. Detection of feline coronavirus RNA in feces, tissues, and body-fluids of naturally infected cats by reverse-transcriptase PCR. J Clin Microbiol 1995;33:684–689.

12. Sievers F, Wilm A, Dineen D, et al. Fast, scalable generation of high-quality protein multiple sequence alignments using Clustal Omega. Mol Syst Biol 2011;7:5392544.

13. Posada D. jModelTest: Phylogenetic model averaging. Mol Biol Evol 2008;25:1253–1256.

14. Guindon S, Dufayard JF, Hordijk W, Lefort V, Gascuel O. PhyML: Fast and accurate phylogeny reconstruction by maximum likeli-hood. Infect Genet Evol 2009;9:384–385.

15. Pedersen NC, Liu H, Dodd KA, Pesavento PA. Significance of coronavirus mutants in feces and diseased tissues of cats suffer-ing from feline infectious peritonitis. Viruses 2009;1:166–184.

16. Chang HW, de Groot RJ, Egberink HF, Rottier PJ. Feline infec-tious peritonitis: Insights into feline coronavirus pathobiogenesis and epidemiology based on genetic analysis of the viral 3c gene. J Gen Virol 2010;91:415–420.

17. Brown MA, Troyer JL, Pecon-Slattery J, Roelke ME, O’Brien SJ. Genetics and pathogenesis of feline infectious peritonitis virus. Emerg Infect Dis 2009;15:1445–1452.

18. Domingo E, Sheldon J, Perales C. Viral quasispecies evolution. Microbiol Mol Biol Rev 2012;76:159–216.


Article


I n t r o d u c t i o nAirway collapse is a common contributing factor to chronic cough

in dogs, particularly in middle-aged small- and toy-breed dogs (1). This disease is multifactorial and often involves degeneration of the tracheal cartilage rings, which leads to dorsoventral flattening of the trachea and laxity of the dorsal tracheal membrane. Collapse may be focal, regional, or generalized and can affect the cervical region, thoracic inlet or intrathoracic region, large bronchi, or small

airways (2). Based on the severity of variations in tracheal luminal diameter assessed by bronchoscopy, the following 4 grades of tra-cheal collapse have been established: grade 1 — variation in tracheal luminal diameter of 25% to 30%; grade 2 — variation of 30% to 60%; grade 3 — variation of 60% to 90%; and grade 4 — variation of 90% to 100% (3,4).

Given the dynamic nature of the disease and the small size of some affected airways, airway collapse is difficult to definitively diagnose and presents a daily challenge to small animal veterinary

Fluoroscopic and radiographic assessment of variations in tracheal height during inspiration and expiration in healthy adult small-breed dogs

Grégoire Scherf, Isabelle Masseau, Anne-Sophie Bua, Guy Beauchamp, Marilyn E. Dunn

A b s t r a c tThe objective of this study was to document tidal variations in tracheal height during normal respiration in 19 healthy adult (. 1 y old) small-breed dogs (, 10 kg) using fluoroscopy and radiography. Each dog underwent tracheal fluoroscopic examination on inspiration and expiration while in a standing position (F-S) and in right lateral recumbency (F-RL), followed by radiographic projections obtained in right lateral recumbency. The percent variation in tracheal height during maximal inspiration and expiration was determined at 3 different locations [cervical region (CR), thoracic inlet (TI), and intrathoracic (IT) region]. When all imaging procedures and sites of measurement were considered, tracheal height varied during physiologic inspiration and expiration from 0% to 21.1%, with a mean of 4.5%. The mean percent variation in tracheal height was not significantly different among imaging modalities (F-S versus F-RL versus radiography) (P = 0.16) or measurement sites (CR versus TI versus IT) (P = 0.89). The body condition score (BCS) (P = 0.96), age (P = 0.95), and breed (P = 0.19) did not significantly influence the mean percent variation in tracheal height. The average variation in tracheal height during maximal physiological inspiration and expiration is small (, 6%) in most healthy adult small-breed dogs as assessed by fluoroscopy and radiography, although tracheal height may vary by as much as 21.1% in some healthy individuals. Inspiratory and expiratory radiographs acquired in right lateral recumbency provide an accurate assessment of tracheal height as an alternative to fluoroscopy.

R é s u m éL’objectif de la présente étude était de documenter les variations de la hauteur de la trachée durant la respiration normale chez 19 chiens adulte en santé (. 1 an) de petites races (, 10 kg) à l’aide de la fluoroscopie et de la radiographie. Chaque chien a été soumis à un examen fluoroscopique de la trachée lors de l’inspiration et de l’expiration alors qu’il était en position debout (F-S) et en décubitus latéral droit (F-RL), suivi d’images radiographiques obtenues en décubitus latéral droit. Le pourcentage de variation de la hauteur de la trachée durant l’inspiration et l’expiration maximales fut déterminé à trois endroits différents [région cervicale (CR), l’entrée thoracique (TI), et la région intrathoracique (IT)]. Lorsque toutes les procédures d’imagerie et les sites de mesure étaient considérés, la hauteur de la trachée variait durant l’inspiration et l’expiration physiologique de 0 % à 21,1 %, avec une moyenne de 4,5 %. Le pourcentage de variation moyen de la hauteur de la trachée n’était pas significativement différent parmi les différentes modalités d’imagerie (F-S versus F-RL versus radiographie) (P = 0,16) ou les sites de mesure (CR versus TI versus IT) (P = 0,89). Le score de condition corporelle (BCS) (P = 0,96), l’âge (P = 0,95), et la race (P = 0,19) n’influençaient pas significativement le pourcentage de variation moyen de la hauteur de la trachée. La variation moyenne de la hauteur de la trachée durant l’inspiration et l’expiration physiologique maximale est petite (, 6 %) chez la plupart des chiens adultes de petites races en santé telle qu’évalué par fluoroscopie et radiographie, bien que la hauteur de la trachée puisse varier jusqu’à 21,1 % chez certains individus en santé. Les radiographies à l’inspiration et à l’expiration obtenues en décubitus latéral droit fournissent une évaluation précise de la hauteur de la trachée comme alternative à la fluoroscopie.


Tufts University Cummings School of Veterinary Medicine, Clinical Sciences, 200 Westboro Road, North Grafton, Massachusetts 01536, USA (Scherf); Department of Clinical Sciences, Faculty of Veterinary Medicine, University of Montreal, 3200 rue Sicotte, Saint-Hyacinthe, Quebec J2S 2M2 (Masseau, Bua, Beauchamp, Dunn).

Address all correspondence to Dr. Grégoire Scherf; telephone: (508) 839-5302; fax: (508) 839-7951; e-mail: [email protected]

The authors declare that there was no conflict of interest and no off-label use of antimicrobials involved in this study.

Received October 8, 2018. Accepted January 16, 2019.



practitioners. Clinical signs such as stridor, wheezing, and cough-ing may guide the clinician toward respiratory disorders affecting the trachea, airways, or pulmonary parenchyma. A wide array of imaging modalities is available to document morphologic changes affecting these structures. These include radiography, fluoroscopy, bronchoscopy, and computed tomography (CT) scan, each of which has specific advantages and limitations.

Radiography has been used extensively to evaluate the trachea and lungs as it is non-invasive, does not require general anesthesia, and most veterinary clinics are equipped with X-ray units. While standard radiography is variably helpful in depicting tracheal col-lapse, with 60% to over 90% diagnostic success, it tends to underesti-mate the degree of tracheal collapse (5). As extrathoracic airways are more likely to collapse on inspiration and intrathoracic airways tend to collapse during expiration (2), both inspiratory and expiratory lateral radiographs are often recommended to increase the likelihood of detecting airway collapse. It is also more difficult to diagnose bronchial collapse on radiographs, especially in small-breed dogs.

Fluoroscopy is non-invasive, does not require anesthesia or deep sedation, and can be done in real time while the patient is resting. Carrying out a fluoroscopy after eliciting a cough by tracheal palpa-tion or short bouts of exercise may help identify dynamic disorders (2,6). It is limited to evaluating the trachea and its bifurcation, as well as the principal bronchi in small breeds, however, and provides a poor assessment of small airways and pulmonary parenchymal changes.

Bronchoscopy provides excellent visualization of the luminal aspect of the trachea, principal bronchi, and lobar bronchi, but it does not allow assessment of small airways and pulmonary parenchyma, is invasive and requires general anesthesia, and provides subjec-tive results (2). Despite this, some studies refer to bronchoscopy as the diagnostic test of choice for identifying airway collapse (7,8). According to a recent study, however, fluoroscopy while eliciting coughing is considered more specific than radiography and bron-choscopy for diagnosing tracheal collapse (2).

In recent years, modern computed tomography (CT) units with faster acquisition capabilities have been increasingly used to inves-tigate disorders affecting large and small airways and pulmonary parenchyma. Unlike radiography, fluoroscopy, and tracheobronchos-copy, it provides excellent non-invasive, cross-sectional visualization of airways of various caliber that in turn can be easily measured. While it can be carried out without chemical restraint in patients at risk for anesthesia, motion artefact is best avoided with ventilator-controlled anesthesia.

The tracheal cross-sectional area of healthy dogs can change by up to 24% (mean: 5.5%) from inspiration to expiration (9). However, variations in tracheal dimensions between phases of respiration are often evaluated by inducing positive inspiratory pressure and simulated end expiration. The largest percent variation in tracheal dimensions from inspiration to expiration seems to occur in tracheal height in the cervical region (9).

The main issue with the various diagnostic modalities is that the degree of dynamic variation in tracheal luminal diameter during tidal breathing in clinically normal dogs has not been determined. In clinically normal humans, computed tomography (CT) has revealed that the tracheal diameter can change by 12% to 32% (10) from

maximal inspiration to expiration. This range is likely higher than would be expected in dogs because forced inspiration and expiration cannot be easily achieved during radiography or fluoroscopy and there would probably be less variation in tracheal diameter between inspiration and expiration.

Fluoroscopy appears to be the best imaging modality for evalu-ating tidal variations in tracheal height in dogs, as tracheal height can be evaluated dynamically, without any influence of sedation or anesthesia. Moreover, it has not been described in the literature whether fluoroscopic examinations in a standing position allow better visualization of bronchial collapse and whether the measure-ments are similar between F-S and F-RL, which are frequently carried out in practice.

The objective of this study was to document tidal variations in tracheal height in adult small-breed dogs with no clinical signs or recent history of respiratory disease, using fluoroscopy [standing fluoroscopy (F-S) and right lateral recumbency fluoroscopy (F-RL)] and inspiratory and expiratory radiographs (right lateral recum-bency). First, we wanted to specifically determine the percent varia-tions in tracheal height during inspiration and expiration at 3 sites of measurement [cervical region (CR), thoracic inlet (TI), and intra-thoracic (IT) region]. Second, we evaluated whether these variations differed significantly depending on the type of examination or site of measurement and third, whether the tracheal height measurements (mm) were significantly different between the 2 types of fluoroscopic positioning (F-S or F-RL) at each site of measurement. Fourth, the influence of body condition score (BCS), age, or breed on the tidal variations of tracheal height was assessed. Finally, we wanted to estimate whether the magnification induced in our imaging proce-dures (fluoroscopy and radiography) was significantly different in the 2 types of radio-opaque markers used, i.e., catheter and ruler.

We hypothesized that variation in tracheal height from inspiration to expiration, as evaluated by fluoroscopy and radiography, would not exceed 24% (9). Moreover, we speculated that neither the fluo-roscopic position (F-S, F-RL) during acquisition nor the site of mea-surement (CR, TI, IT) would impact variation in tracheal height from inspiration to expiration and from fluoroscopy to radiography (right lateral recumbency, images taken at inspiration and expiration).

We expected that the percent variation in tracheal height would be influenced by the body condition score (# 5/9 or . 5/9), age (# 6 y old or . 6 y old), and breed of the dogs in our study, depending on the type of imaging procedure and/or the site of measurement. More precisely, we hypothesized that the percent variation in tracheal height would be significantly greater in older dogs (. 6 y old), as tracheal collapse is a progressive disease, dogs with a body condition score of . 5/9, as obesity is a well-known risk factor for tracheal collapse (1), or dogs belonging to a breed at risk.


Patient selectionThe study was conducted at the Faculty of Veterinary Medicine of

the University of Montreal from November 1, 2016 to May 30, 2017. This study was approved and conducted in accordance with the



Committee for the Ethical Use of Animals of the Faculty of Veterinary Medicine, University of Montreal.

Dogs recruited to participate in the study belonged to students, staff, and clients from the general practice. Dogs were enrolled if they were . 1 y old, weighed , 10 kg, and had no recent history (, 1 mo) of respiratory disease and no respiratory signs, e.g., cough-ing, excessive panting or reverse sneezing, sneezing, and intolerance to exercise.

A 10-kg cutoff weight was determined in order to recruit adult dogs belonging to breeds at risk of tracheal collapse. Dogs were excluded from the study if respiratory anomalies were detected dur-ing physical examination, if they were too agitated to allow examina-tion without chemical restraint, or if a thoracic lesion (mediastinal, pleural, alveolar, structured interstitial, or heavy unstructured interstitial) was observed on radiographic projections.

Owners provided written consent before enrolling in the study. A thorough physical examination was carried out before imaging. On the day of imaging, owners were asked about their dog’s medical history, vaccination status, and current therapy.

Imaging protocolFluoroscopic and radiographic images were acquired with a high

frequency Siemens AXIOM Iconos R200 (Siemens, Mississauga, Ontario). A cineloop encompassing 3 respiratory cycles was first recorded with a horizontal beam while the dog was standing (F-S)

and then with a vertical beam after placing the dog in right lateral recumbency (F-RL) (Figure 1). Each cineloop was reviewed to ensure that the caudal aspect of the larynx and the entire diaphragm was included during all 3 respiratory cycles.

Digital radiographic (DR) projections using a vertical beam were then acquired in inspiration and expiration while in right lateral recumbency, followed by ventrodorsal and left lateral recumbent projections taken in inspiration. A marker catheter (Infiniti Medical, Redwood City, California, USA) was fixed with small clips along the spinous processes or the sternum to provide a calibration tool for measuring tracheal height A radio-opaque ruler was also placed directly on the table along the sternum of each dog and for each imaging procedure, in order to compare the magnification induced by the imaging procedures of the 2 radio-opaque mark-ers (Figure 1A–D) The focal detector distance was kept constant throughout the study.

No chemical restraint was administered for any of the imag-ing examinations. A research project number attributed through the hospital identification system (HIS) was used as the patient identification number for all dogs, thereby preserving anonymity. Each acquisition was identified with a unique 6-digit requisition number obtained by a number generator, providing a total of 6 dif-ferent acquisitions for each dog. This information was kept by the radiology technologist and was revealed only after all measure-ments were completed. Fluoroscopic and radiographic images were

Figure 1. Tracheal height assessment was performed on images obtained during fluoroscopic and radiographic acquisitions in right lateral recumbency (A, B) and in a standing position (C, D). A — Right lateral recumbent images were obtained with a vertical X-ray beam. B — Close-up image of A illus-trating placement of the radio-opaque catheter marker (*) along the sternum using a small clip (black clip on the dog’s hair). The radio-opaque ruler (black arrow) was secured on the table using tape immediately ventral to the dog’s sternum. C — The X-ray beam was rotated horizontally to obtain left to right lateral projection of the thorax in a standing position. D — Close-up image of C illustrating placement of the radio-opaque catheter marker (*) disposed along the dorsal spinous processes of the vertebral column, whereas the radio-opaque ruler is secured directly on the examination table ventral to the dog’s head and neck.



transferred to a picture archiving and communication system (PACS) and reviewed using the digital radiography imaging system AGFA Impax 6 (Agfa, Toronto, Ontario). All measurements were taken over a period of 1 mo after the study was completed, in a blind randomized manner, and reviewed by a Board-certified veterinary radiologist (DACVR).

Fluoroscopic assessment and measurementsEach cineloop was reviewed using a frame-by-frame mode allow-

ing selection of each frame demonstrating maximal inspiration and expiration, respectively, based on movements of the diaphragm, during 3 respiratory cycles for a total of 6 frames selected per acquisition. Selected frames from one respiratory cycle performed during fluoroscopic examination are shown in Figure 2 (A–D). Those frames show the movement of the diaphragm during spontaneous breathing, allowing definition of maximal inspiration and expiration. Maximal inspiration was defined as the maximal caudal excursion of the diaphragm (Figure 2A). Maximal expiration was defined as the maximal cranial excursion of the diaphragm (Figure 2C). For each of the 6 selected frames, tracheal height, defined as the distance in

Figure 2. Selection of fluoroscopic images for maximal inspiration and expiration based on excursion of the diaphragm in a healthy small breed dog. A–D — Selected frames from one respiratory cycle performed during fluoroscopic examination showing the movement of the diaphragm during spontaneous breathing. A — Maximal caudal excursion of the left crus of the diaphragm (large arrow) during one respiratory cycle reaching the landmark 2 was designated as maximal inspiration. B — Intermediate excursion of the left crus of the diaphragm (large arrow) corresponding to the half-way location between landmark 1 and 2. C — Maximal cra-nial excursion of the left crus of the diaphragm (large arrow) during one respiratory cycle reaching the landmark 1 was designated as maximal expiration. D — After expiration, the left crus of the diaphragm moves caudally, reaching an intermediate position between landmark 1 and 2 as seen on image B. * — indicates location of the larynx; T9 — 9th thoracic vertebra.

Figure 3. Measurement sites of tracheal height at maximal inspiration (A) and expiration (B). First, lines (white lines) were traced at the level of C4-C5 (1), C7-T1 (2) and T3-T4 (3) intervertebral disc spaces. Using these lines as guides, tracheal height was then measured for the cervical (CR), thoracic inlet (TI), and intra-thoracic (IT) regions by tracing a line (black lines) perpendicular to the dorsal and ventral luminal borders of the trachea as close as possible to the guideline. The diaphragm (arrow) and caudal end of the larynx (*) were included on each image. The distance between two marks on the catheter marker (d1) and the ruler (d2) was measured on each image [for image clarity purposes, these distances are only illustrated on the top image (A)]. Each calibration device is identified on the bottom image (B). T10, represents the vertebral body of the 10th thoracic vertebra.



millimeters between the dorsal and the ventral borders of the tra-cheal lumen, was measured at the following 3 locations: the level of the C4-C5 intervertebral disc space (CR); C7-T1 (TI); and T3-T4 (IT), as illustrated in Figure 3 (A–B). Three measurements of tracheal height at maximal inspiration and 3 measurements at maximal expi-ration were determined successively at each site and then averaged.

Radiographic assessment and measurementsDigital radiography (DR) projections were used to assess each

dog’s airways and pulmonary parenchyma on the day of acquisi-tion and to identify a redundant dorsal tracheal membrane when present. Only right lateral recumbent DR projections were used for measurements. Tracheal height was measured at the same 3 loca-tions described for the fluoroscopic images on both inspiratory and expiratory DR projections (Figure 3A–B).

Measurements using the marker catheter and the radio-opaque ruler

Tracheal height measurements were adjusted, taking into account the magnification induced by the imaging procedures, by first mea-suring the distance between 2 radio-opaque marks on the marker catheter, which should correspond to 10 mm. The tracheal height measured in absolute value (mm) was then adjusted according to the magnification obtained. For example, on most images, the distance between the 2 marks was equivalent to 11 mm, corresponding to a magnification of 10% (Figure 3A–B), which was then subtracted from the measured tracheal height in millimeters to account for magnifica-tion, e.g., a measured tracheal height of 11 mm and a magnification of 10% = adjusted tracheal height of 10 mm.

Measurements were also taken between 2 marks on the ruler, which was placed directly on the table, in order to identify if there was a significant difference in terms of magnification between the ruler, which is more convenient to use as it is placed directly on the table, and the marker catheter.

The mean tracheal height for each location, position, and respira-tory phase was used to determine the percent tracheal height varia-

tion during maximal inspiration (maxinsp) and expiration (maxexp), according to the modality used, using the following equation:

(mean tracheal height at maxinsp 2 mean tracheal height at maxexp)

% Tracheal height = 3 100 highest mean between maxinsp and maxexp

Statistical analysisFluoroscopic measurements were analyzed using a mixed-

linear model with position (F-S versus F-RL) and measurement site (CR, TI, IT) as within-subject factors, including the interaction between the 2 factors. In a further elaboration of this model, potential risk factors were added as additional fixed factors, including body condition score, age, or breed. A paired t-test was used to compare tracheal heights between the 2 fluoroscopic acquisitions (F-S, F-RL).

A similar model was used to analyze the radiographic results and to compare them with those obtained with the 2 fluoroscopic positions.

Re s u l t sTwenty-three dogs were recruited for the study, but 4 were

excluded because of excessive agitation or anxiety during the imag-ing procedures. Of the 19 dogs included in the study, 13 were females and 6 were males. The median age was 6 y 4 mo (1 to 15 y) and the median weight was 5.4 kg (1.7 to 9.1 kg). Two dogs had a body con-dition score (BCS) of 4/9, 12 dogs had a BCS of 5/9, and 5 dogs had a BCS of 6/9. Eleven breeds were represented, including Yorkshire terrier (n = 3), Chihuahua (n = 3), Dachshund (n = 3), Cavalier King Charles spaniel (n = 2), Jack Russell terrier (n = 2), and 1 each of poodle, schnauzer, beagle, shih tzu, cocker spaniel, and Maltese.

Two dogs presented occasional coughing and 4 dogs occasional reverse sneezing at home. These dogs were not excluded from the study since the clinical signs were occasional and had not been observed in the month before the study. All but 1 dog received routine vaccinations, whereas only 11 dogs were vaccinated against Bordetella bronchiseptica. Three dogs had a left apical heart murmur (grade 2/6) during physical examination and 8 dogs coughed on palpation of the trachea. The clinical signs and results of physical examinations for all dogs are presented in Table I. A prolapse of the dorsal tracheal membrane was diagnosed in 6 dogs. This anomaly was either diagnosed on fluoroscopy (2/6), radiography (6/6), or both (2/6). Only half of these 6 dogs coughed on tracheal palpa-tion. Eight dogs had a mild bronchial pattern exclusively noted on thoracic radiographs, which was considered age-related.

Measurements of tracheal heightTracheal height measurements at all 3 sites (CR, TI, and IT) for

both fluoroscopic and radiographic examinations are shown in Table II. In our study, all types of examinations considered (F-S, F-RL, radiography), tracheal height measurements were greater at inspiration than at expiration at TI and IT (67% and 70%, respec-tively), which supports dynamic narrowing of the intrathoracic air-ways during expiration. Measurements of the cervical tracheal height were significantly greater on F-RL than those measured on F-S, at

Table I. Clinical signs and physical examination findings of dogs studied.

Number of Percentage dogs (19) (%)Anamnesis Reverse sneezing 4 21 Vaccination status 18 95 Deworming prevention 14 74 NSAID when neededa 2 11

Physical examination Body condition score 5 26 Triggerable cough 8 42 Heart murmur 3 16a NSAID (non-steroidal anti-inflammatory drug) was administered for disc herniation in 1 dog and for coxo-femoral dysplasia in another dog when the owner felt it was necessary.



both inspiration (P = 0.01) and expiration (P = 0.015). No statistically significant difference was found for tracheal height measurements at other sites between F-RL and F-S or between phases of respiration.

Percent variation in tracheal height during maximal inspiration and expiration

The percent variation in tracheal height at CR, TI, and IT during maximal inspiration and expiration and for both fluoroscopic and radiographic examinations are given in Table III. The mean percent change of tracheal height variation between maximal inspiration and expiration during fluoroscopic examinations was not higher than 4.7% during F-S and 4% for F-RL (Table III). In F-S, the maxi-mal percent variation was obtained at TI, reaching 21%, whereas IT was associated with the maximal percent change in tracheal height variation during F-RL with a value of 13.5% (Table III).

Position during acquisition (F-S versus F-RL) did not significantly impact (P = 0.41) the mean percent of tracheal height variation or did the 3 sites of measurements (P = 0.98). No significant interaction was found between these 2 factors (position during acquisition and sites of measurement) (P = 0.82).

The mean percent change of tracheal height variation between maximal inspiration and expiration during radiographic examina-tions was not higher than 5.2% (Table III). The maximal percent variation on radiographic examinations was obtained at IT, reaching 16.4% (Table III). The mean percent of tracheal height variation was not affected by the site of measurement along the trachea (P = 0.89) for all imaging procedures or by the imaging modality or position (F-S versus F-RL versus radiography) (P = 0.16). No significant inter-action was found between the 2 factors (position during acquisition and sites of measurement) (P = 0.85).

Several subjects (n = 8 dogs) did not demonstrate variation of their tracheal height during maximal inspiration and expiration at certain sites, which explains the zero percent variation of tracheal

height frequently obtained, either on fluoroscopic or radiographic examinations. This was more frequently observed at TI and at IT (n = 6 dogs each) for the fluoroscopic examinations and at CR for the radiographic projections (n = 3).

Influence of BCS, age, and breedThe mean percentage of variation in tracheal height was not

associated with BCS (P = 0.96), age (P = 0.95), or breed (P = 0.19), regardless of the modality (F-S, F-RL, right lateral recumbency DR) or sites of measurement (CR, TI, IT).

Magnification (marker catheter versus ruler)For F-RL and right lateral recumbent radiographs (at inspiration

and expiration), the magnification induced was significantly greater with the marker catheter than with the ruler (P = 0.032). There was no significant difference, however, in terms of magnification between the marker catheter and the ruler during F-S (P = 0.13). Absolute tracheal height values measured on images were therefore adjusted considering the magnification calculated on the catheter marker, which was believed to more accurately reflect tracheal measurements since the marker was placed at the same height as and parallel to the trachea (Figure 1).

D i s c u s s i o nThis study evaluated variations in tracheal height during inspira-

tion and expiration in adult small-breed dogs with no clinical signs or recent history of respiratory disease, using fluoroscopy (F-S and F-RL) and radiography (right lateral recumbency). The population included 19 dogs of various small breeds: 8 dogs were from breeds considered at risk of tracheal collapse (Yorkshire terrier, Maltese, poodle, Chihuahua, shih tzu) (1); and 11 dogs belonged to breeds with no particular known risk for tracheal anomalies. We chose this

Table II. Range and mean of tracheal height measurements obtained from fluoroscopic and radiographic examinations according to the dog’s position during the imaging procedure, site of measurement, and phase of respiration (inspiration versus expiration).

Fluoroscopy Digital radiography Standing Right lateral Right lateral Position recumbency recumbency Site of measurement (mm) (mm) (mm)CR Inspiration 5.0 to 14.0 (9.1)a 4.8 to 13.4 (10.1)a 6.9 to 14.2 (11.0) Expiration 4.8 to 12.4 (9.2)a 5.5 to 13.6 (10.0)a 6.9 to 14.1 (10.8)

TI Inspiration 4.0 to 11.7 (7.9) 3.8 to 11.2 (7.8) 4.5 to 12.0 (8.8) Expiration 4.0 to 11.1 (7.8) 3.8 to 10.5 (7.6) 4.5 to 11.6 (8.5)

IT Inspiration 4.2 to 12.5 (8.1) 4.4 to 12.5 (8.3) 5.5 to 13.1 (9.6) Expiration 4.0 to 11.7 (7.9) 3.8 to 12.1 (8.1) 5.4 to 13.1 (9.3)In each column, the range of tracheal height measurements is first displayed followed by the mean in between parentheses.a P , 0.05.



healthy small-breed dog population specifically as they are most often presented for evaluation of airway collapse by fluoroscopy and radiography. The goal of this study was to provide a better understanding of the normal variation in tracheal height found in healthy small-breed dogs, in order to help differentiate between airway collapse and normal variations in tracheal diameter in dogs presented for evaluation of airway collapse.

Four dogs were excluded from our study, as it was not possible to carry out fluoroscopy in 3 complete respiratory cycles and radi-ography without significant restraint and sedation. Since patients presented for suspicion of tracheal collapse are often evaluated without the use of any chemical restraint, including sedation and anesthetics, and our desire was to design a study for which results could be extrapolated to routine evaluation of tracheal height done in the clinic, the use of sedatives was avoided (11,12). We also wanted to limit any possible effect of sedatives or anesthetics on tidal volume and diaphragmatic movement, as has been described in previous studies (11,12).

In accordance with our hypothesis, the percent of variation in tracheal height obtained in our study (0% to 21.1%, mean of 4.5%) was slightly lower than that reported in a previous study of dogs that underwent general anesthesia and evaluation of tracheal height by CT (9). Our results may be explained by the fact that we measured tracheal height during spontaneous tidal breathing using fluoroscopy and radiography, observing movements of the trachea in real time when the patient is at rest and not influenced by seda-tion or mechanical ventilation forces, which therefore more closely reproduces daily tracheal examination in dogs suspected of having tracheal collapse. In healthy dogs, normal voluntary inspiration creates negative intrathoracic pressure. The intratracheal pressure is therefore negative rather than positive in spontaneously breath-ing dogs during inspiration. It is reasonable to think that applying positive intratracheal airway pressure during inspiration in an anes-thetized animal will affect the tracheal measurements compared to those obtained in spontaneously breathing dogs (9,13).

Despite the relatively low mean percent of variation in tracheal height during inspiration and expiration in our study (4.5%), some individuals (n = 4) presented greater variation (21.1%, 16.4%, 13.5%, 13.3%). Only 4 dogs had variations in tracheal height greater than 10%. The dog with the highest value (21.1% at TI, F-S) was an 11-year-old Maltese with a prolapse of the dorsal tracheal mem-brane identified on thoracic radiographs. The dog with the second highest values was a 15-year-old cocker spaniel with a percent variation in tracheal height of 16.4% (IT, right lateral recumbency radiography). This individual coughed on tracheal palpation during the physical examination and had a mild diffuse bronchial pattern on thoracic radiographs. It is possible that these 2 dogs had a sub-clinical tracheal collapse, with no clinical signs or recent history of respiratory disease. These higher variations in height were noted on only 1 imaging procedure for each dog, which makes them even more difficult to interpret. It may also point to the fact that varia-tion in tracheal height during spontaneous breathing is a dynamic phenomenon that may occasionally manifest different patterns. It would be interesting to demonstrate whether intermittent dynamic tracheal collapse is underestimated during a single static evaluation on either fluoroscopy or radiography compared to various other positions and modalities.

Our results demonstrate that radiographs taken in right lateral recumbency offer a good alternative to fluoroscopy for evaluating variations in tracheal height during inspiration and expiration in healthy adult small-breed dogs as no significant difference was found between modalities and/or positions during acquisition (F-S versus F-RL versus radiography). Inspiratory and expiratory radio-graphs carried out in right lateral recumbency may aid recognition of normal variations in tracheal height in adult small-breed dogs for practitioners without access to fluoroscopy. Since dogs included in our study were healthy and with no clinical evidence of tracheal collapse, we were not able to compare the suitability of radiography with fluoroscopy for diagnosing tracheal collapse in affected dogs, as has been evaluated in previous studies (2,5).

Table III. Percent variation in tracheal height during maximal inspiration and expiration according to the dog’s position during fluoroscopic and digital radiography acquisitions and to site of measurement

Range of variation in tracheal Mean Median Type of examination height (%) (%) (%)Sternal fluoroscopy Cervical region 0 to 11.9 4.1 3 Thoracic inlet region 0 to 21.1 4.7 3.5 Intrathoracic region 0 to 10.5 4.5 3.4

Right lateral recumbency fluoroscopy Cervical region 0 to 9.5 4 3.9 Thoracic inlet region 0 to 8.7 3.7 4.2 Intrathoracic region 0 to 13.5 3.6 3.3

Right lateral recumbency digital radiography Cervical region 0 to 13.3 5.1 4.9 Thoracic inlet region 0 to 10.3 4.6 4 Intrathoracic region 0.8 to 16.4 5.8 4.9



The greater cervical tracheal height measurements in right lateral recumbency fluoroscopy (F-RL) compared to standing fluoroscopy (F-S), at both inspiration and expiration, was not expected. This implies that measurements taken standing and in right lateral recumbency are not identical for the cervical region, which needs to be taken into account when conducting longitudinal studies on an individual. This result could suggest that F-RL should be used if a more accurate measurement of cervical tracheal diameter is required, i.e., in a patient needing a tracheal stent, as the measurements of cervical tracheal height were greater in our study in F-RL than in F-S.

It is sometimes tempting for general practitioners to diagnose a tracheal collapse or redundant dorsal tracheal membrane based solely on history and physical examination, especially if cough is elicited on tracheal palpation. In our study, a redundant dorsal tracheal membrane was visible in 6 dogs on thoracic radiographs and/or fluoroscopy. As cough was elicited on tracheal palpation in only 3 of these dogs, it seems inappropriate to diagnose a tracheal collapse or a redundant dorsal tracheal membrane based only on eliciting cough on tracheal palpation during physical examination. Imaging assessment of the trachea is essential to diagnose airway collapse.

Our study had several limitations. First, its statistical power is relatively low (19 dogs) and a significant difference of mean percent variation in tracheal height may have been identified in a larger sample of dogs. Specifically, 52 dogs would have been necessary to identify a significant difference in mean percent variation in tracheal height between fluoroscopy in right lateral recumbency (F-RL) and right lateral recumbent radiographs, 80% of the time. Hundreds of dogs would have been necessary to identify a signifi-cant difference in mean percent variation in tracheal height, 80% of the time, between the other types of imaging procedures or sites of measurement.

Moreover, variations in tracheal height were not related to BCS, age, or breed, as would have been expected. This may also be due to the small sample size of our study. As variations in tracheal height tended to be greater in breeds at risk of tracheal collapse, it is pos-sible that a significant difference between dogs belonging to a breed at risk or not would have been identified with a larger sample size. As tracheal collapse is a slowly progressive degenerative disorder of the tracheal and bronchial cartilaginous rings, it is reasonable to think that tracheal height would vary more in older dogs than in younger dogs (1).

Approximately 25% of dogs included in this study had a BCS of . 5/9. Although obesity is a well-known risk factor for tracheal col-lapse (1), none of these dogs demonstrated a significantly increased variability in tracheal height during inspiration and expiration, which suggests that other factors may also play a role in manifest-ing tracheal collapse in overweight dogs. Only 2 of these dogs were breeds considered at risk of tracheal collapse (1 Yorkshire terrier and 1 poodle), however, which makes the comparison of overweight dogs to dogs that are not overweight unsuitable in our study. It would have been interesting to know whether being overweight would increase the variability in tracheal height in a population of dogs all belonging to breeds at risk of tracheal collapse compared to dogs with an optimal BCS.

The results concerning magnification were not conclusive in our study. Our conflicting results could be explained by the greater

mobility of the marker catheter placed along the spinous processes or sternum of each animal, from 1 imaging procedure to the other. Even if the marker catheter is a less convenient tool than the table ruler, it is advantageous in order to obtain the most accurate mea-surements of tracheal height, as it is placed parallel to the trachea and at the same height, which allows the veterinarian to obtain the most accurate measurements of tracheal height despite the magni-fication induced by the imaging procedure.

The main objective of this study was to document variations in tra-cheal height at inspiration and expiration in healthy adult small-breed dogs with no clinical signs or recent history of respiratory disease using fluoroscopy in a standing position and in right lateral recumbency (F-S and F-RL) and radiography in right lateral recumbency. Despite its limitations, this study improves our understanding of the degree of fluctuations in tracheal diameter during spontaneous breathing in clinically normal dogs. The assessment of tracheal height in healthy small-breed dogs appears to be similar in fluoroscopy and radiogra-phy, allowing general practitioners without access to fluoroscopy to estimate the tracheal height in small-breed dogs using radiography. When diagnosing tracheal collapse, given its dynamic nature, it may be beneficial to use a combination of various imaging modalities.

A c k n o w l e d g m e n t sThe authors thank the medical imaging technicians from the

Faculty of Veterinary Medicine at the University of Montreal in Saint-Hyacinthe for their excellent technical support. Funding was provided in part by the Companion Animal Healthcare Fund, University of Montreal.

Re f e r e n c e s 1. Tappin SW. Canine tracheal collapse. J Small Anim Pract 2016;

57:9–17. 2. Johnson LR, Singh MK, Pollard RE. Agreement among radio-

graphs, fluoroscopy, and bronchoscopy in documentation of airway collapse in dogs. J Vet Intern Med 2015;29:1619–1626.

3. Johnson LR, Pollard RE. Tracheal collapse and bronchomalacia in dogs: 58 cases. J Vet Intern Med 2010;24:298–305.

4. Tangner CH, Hobson HP. A retrospective of 20 surgically man-aged cases of collapsed trachea. Vet Surg 1982;11:146–149.

5. Macready DM, Johnson LR, Pollard RE. Fluoroscopic and radiographic evaluation of tracheal collapse in dogs: 62 cases (2001–2006). J Am Vet Med Assoc 2007;230:1870–1876.

6. Hayl JE. Tracheobronchial collapse during cough. Radiology 1965;85:87–92.

7. Hedlund CS. Surgery of the upper respiratory system. In: Fossum TW, ed. Small Animal Surgery. 3rd ed. St. Louis, Missouri: Mosby Elsevier, 2007:817–866.

8. Sura PA, Durant AM. Trachea and bronchi. In: Tobias KM, Johnston SA, eds. Veterinary Surgery: Small Animal. 1st ed. St. Louis, Missouri: Elsevier/Saunders, 2012:1734–1751.

9. Leonard CD, Johnson LR, Bonadio CM, Pollard RE. Changes in tracheal dimensions during inspiration and expiration in healthy dogs as detected via computed tomography. Am J Vet Res 2009;70:986–991.



10. Aquino SL, Shepard JA, Ginns LC, et al. Acquired tracheomala-cia: Detection by expiratory CT scan. J Comput Assist Tomogr 2001;25:394–399.

11. Jacobson JD, Mcgrath CJ, Smith EP. Cardiorespiratory effects of four opioid-tranquilizer combinations in dogs. Vet Surg 1994; 23:299–306.

12. Ko JC, Bailey JE, Pablo LS, Heaton-Jones TG. Comparison of sedative and cardiorespiratory effects of medetomidine and

medetomidine-butorphanol combination in dogs. Am J Vet Res 1996;57:535–540.

13. Opdyke DF, Brecher GA. Effect of normal and abnormal changes of intrathoracic pressure on effective right and left atrial pres-sures. Am J Physiol 1950;160:556–566.


Article


I n t r o d u c t i o nIntraosseous (IO) catheterization may be used to obtain vascular

access in rabbits. For debilitated rabbits presenting with circula-tory collapse, or those at imminent risk of cardiopulmonary arrest, IO catheters may be considered as an alternative or in preference to, intravenous catheters.

Any long bone can be used for IO catheterization in small mammals. The ideal IO placement site is unknown, although the humerus, femur, and tibia have all been recommended (1). Despite

a wealth of anecdotal recommendations, there is little supporting lit-erature related to IO catheters in veterinary medicine and much less in exotic animals. A recent canine cadaver study demonstrated that placement of a humeral IO catheter compared favorably to jugular venous catheters and suggested IO use in emergent dogs (2). A report of successful cardiopulmonary resuscitation in a geriatric chinchilla included the use of an IO catheter placed in the humerus (3).

Intraosseous catheters were successfully placed in the craniome-dial aspect of the tibial plateau of rabbits in a study comparing IO versus IV propofol administration (4). Insertion of a bone marrow

Comparison of 3 intraosseous catheter sites and methods of determining placement success in cadaver rabbits

Christopher R. Kennedy, Jay N. Gladden, Elizabeth A. Rozanski

A b s t r a c tThe study goals were to determine if intraosseous (IO) catheters can be placed with greater success into the humerus, femur, or tibia of cadaver rabbits, and to evaluate the accuracy of perceived success (PS) and objective clinical success (OCS) criteria against true intramedullary catheterization confirmed by fluoroscopy. This was a prospective study utilizing 12 rabbit cadavers. Twenty-two participants attempted IO catheter placement at 3 sites. Perceived success, OCS, and fluoroscopic true success (FTS) were recorded. A Fisher’s exact test was used to compare PS, OCS, and FTS, and FTS rates between sites (P , 0.05). A Wilcoxon test was used to compare speed of placement (P , 0.05). Overall, of 66 attempts, PS was reported in 86.4%, OCS was documented in 62.1%, FTS was confirmed in 43.9%. Perceived success and OCS overestimated FTS (P # 0.01 and P = 0.027, respectively). Confirmation of FTS occurred in 10/22 (45.5%) humeral, 5/22 (22.7%) femoral, and 14/22 (63.6%) tibial (P = 0.03) attempts.

Median time until placement for the humerus was 37.5 seconds (range: 15 to 125 seconds); the femur 135 seconds (range: 91 to 148 seconds); the tibia 49 seconds (range: 19 to 150 seconds). The humerus and tibia were faster to catheterize than the femur (P = 0.01 and 0.03, respectively). Participant PS and OCS criteria overestimated FTS. The humerus or tibia may be more successful and are faster to catheterize.

R é s u m éLes objectifs de la présente étude étaient de déterminer si des cathéters intra-osseux (IO) peuvent être placés avec plus de succès dans l’humérus, le fémur ou le tibia de cadavres de lapins, et d’évaluer la précision des critères du succès perçu (PS) et du succès clinique objectif (OCS) versus le cathétérisme intramédullaire réel confirmé par fluoroscopie. Il s’agissait d’une étude prospective utilisant 12 cadavres de lapin. Vingt-deux participants ont tenté le placement des cathéters IO aux trois sites. Le PS, l’OCS et le succès réel par fluoroscopie (FTS) furent notés. Un test exact de Fisher fut utilisé pour comparer PS, OCS, et FTS, et les taux de FTS entre les sites (P , 0,05). Un test de Wilcoxon a été utilisé pour comparer la vitesse de placement (P , 0,05). Globalement, des 66 essais, PS a été rapporté dans 86,4 % des cas, OCS a été documenté dans 62,1 % des cas, et FTS a été confirmé dans 43,9 % des cas. Le PS et l’OCS surestimaient le FTS (P # 0,01 et P = 0,027, respectivement). La confirmation de FTS s’est produite dans 10/22 (45,5 %) des essais sur l’humérus, 5/22 (22,7 %) des essais sur le fémur, et 14/22 (63,6 %) des essais sur le tibia (P = 0,03).

Le temps médian du placement pour l’humérus était de 37,5 secondes (écart : 15 à 125 secondes); pour le fémur de 135 secondes (écart : 91 à 148 secondes); et pour le tibia de 49 secondes (écart : 19 à 150 secondes). Le cathétérisme de l’humérus et du tibia étaient plus rapides que celui du fémur (P = 0,01 et 0,03, respectivement). Les critères pour le PS et l’OCS des participants surestimaient le FTS. L’humérus et le tibia sont plus rapides à cathétériser et le taux de succès est meilleur.


Department of Clinical Sciences, Cummings School of Veterinary Medicine at Tufts University, 55 Willard Street N, North Grafton, Massachusetts 01536, USA.

Address all correspondence to Dr. Jay Gladden; telephone: 262-542-3241; fax: 262-542-0805; e-mail: [email protected]

Dr. Gladden’s current address is Wisconsin Veterinary Referral Center, 360 Bluemound Road, Waukesha, Wisconsin 53188, USA.

Dr. Kennedy’s current address is Canada West Veterinary Specialists, 1988 Kootenay Street, Vancouver, British Columbia V5M 4Y3.

Dr. Kennedy’s new credentials are BSc, BVetMed, DACVECC, MRCS.

Received October 24, 2018. Accepted February 18, 2019.



needle was reported to be well-tolerated in these animals after infiltration of a local anesthetic. Clinical recommendations for the ideal IO catheter placement site vary amongst clinicians, with the tibial crest site often recommended for rabbits and most other small mammals (5). Based on extrapolations from IO recommendations in neonatal dogs and cats, some clinicians favor the proximal femur. Knowing which bone may be catheterized the fastest and most suc-cessfully, especially for the novice placer, will likely help optimize patient care and reduce the impact of restraint on critically ill rabbits.

No current studies have evaluated the ideal IO catheter placement site in rabbits. Confirmation of successful urgent IO catheter place-ment has not been previously evaluated. Currently bedside clinical indicators of successful IO catheterization include loss of resistance when the cortex is penetrated, the perception of firm seating of the catheter at the insertion site with the ability to manipulate the limb via the catheter, a slight resistance to catheter flushing, and the absence of post-flush subcutaneous fluid accumulation (1). No studies have evaluated the accuracy of these criteria. Although radiographic confirmation is often recommended, it may not be practical in an impending cardiopulmonary arrest situation. Once stable, radiographs are important for confirmation of placement and further patient diagnostics. Validating bedside clinical indicators of IO catheter placement success would be useful for clinicians treating emergent rabbits.

The goals of this study were to determine if IO catheters can be placed faster and with greater success into the humerus, femur, or tibia of cadaver rabbits, and to evaluate the accuracy of perceived success and commonly used clinical criteria for success, against true intramedullary catheterization confirmed by fluoroscopy.

M a t e r i a l s a n d m e t h o d sTwenty-two participants were recruited (veterinary students and

house officers) with limited (# 5 prior attempts at IO catheter place-ment) experience. Participants underwent a 15 min training session on IO catheter placement techniques immediately before the study.

Twelve donated rabbit cadavers were obtained from a combina-tion of an owner-initiated donation program (6) and the department of laboratory animal medicine. Each participant attempted timed placement of a hypodermic needle into the humerus, femur, and tibia of a cadaver rabbit. A choice of 18- to 20-gauge hypodermic needles was provided for catheterization. Each placement attempt was timed and evaluated by a single assessor (CK). Each participant was allowed up to 3 min per site. Time until perceived success (PS) was recorded. Perceived success was recorded if the participant believed the catheter to be correctly placed. Once PS was noted by the participant, the IO catheters were further assessed by a single investigator (CK) using the following objective clinical success (OCS) criteria: perception of firm seating of the catheter at the insertion site; ability to manipulate the limb via the catheter; a slight resistance to catheter flushing (i.e., in excess of subcutaneous placement); and the absence of post-flush subcutaneous fluid accumulation after infusing 0.5 mL saline in the IO catheter. Meeting all 4 criteria was required to claim OCS. After being assessed for these criteria, the IO catheters were further assessed via fluoroscopy by a second investigator (ER), blinded to the previous results. Fluoroscopic True Success (FTS) was

determined by infusion of iohexol (Omnipaque, GE Healthcare, Oslo, Norway) and subsequent identification of intramedullary flow via fluoroscopy (Fluoroscopy/DR Unit [Model# UD150L], Shimadzu, Kyoto, Japan) (Figure 1). Presence of contrast material within the medullary cavity represented truly successful intramedullary catheterization. Data collected included: site of placement, time (seconds), perceived success (PS), objective clinical success (OCS), and fluoroscopic true success (FTS). A Fisher’s exact test was used to compare success between the 3 placement sites: humerus, femur, and tibia; (P , 0.05) and the different methods of success determina-tion: PS, OCS, FTS (P , 0.05). Individual 2 3 2 Fisher’s exact tests were used to compare variables when significance was found with a 3 3 3 test (P , 0.05). A Mann-Whitney test was used to compare speed of placement between sites where applicable.

Re s u l t sThe median weight of the 12 rabbit cadavers was 2.26 kg (range:

1.04 to 5.44 kg). There was a total of 66 placement attempts; 29 (43.9%) were true intramedullary catheterizations confirmed by fluoroscopy. Participants reported PS in 57/66 (86.4%) attempts, although only 41/66 (62.1%) attempts met the criteria for OCS. Two catheters placed were assessed as truly successful via fluoroscopy, but clinically unsuccessful based on an initial inability to flush (i.e., did not meet the all of the OCS criteria). Of the 2 catheters that were incorrectly assessed by OCS, both placers had reported PS. One attempt at a tibial catheter was successfully placed in the dis-tal femur but counted as a failure as it was not the intended bone. There was an overall significant difference between PS, OCS, and FTS (P # 0.01) for all sites together. Both participant PS and OCS overestimated true intramedullary placement of catheters (P # 0.01 and P = 0.027, respectively).

Confirmation of FTS occurred in 10/22 (45.5%) humeral, 5/22 (22.7%) femoral, and 14/22 (63.6%) tibial (P = 0.03) attempts. These results were significantly different between attempted sites (P = 0.03). The tibia was more successful than the femur (P # 0.01); the tibia was not different from the humerus (P = 0.18); the humerus was not different from the femur (P = 0.1).

For catheters that had confirmed FTS, median time until place-ment for the humerus (n = 10) was 37.5 s (range: 15 to 125 s); the femur (n = 5) 135 s (range: 91 to 148 s); the tibia (n = 14) 49 s (range: 19 to 150 s). There was a significant difference in time until place-ment between the humerus versus the femur (P = 0.01). There was a

Figure 1. A — Correctly placed intramedullary tibial IO catheter. B — Incorrectly placed humeral IO catheter with accumulation of radiopaque contrast material outside of the medullary cavity.



significant difference between the tibia versus the femur (P = 0.03). There was no significant difference in speed of placement between the humerus and the tibia (P = 0.9).

Considering only catheters with confirmed FTS, only 1 participant was successful at 0 sites. Fifteen (68.2%) participants were success-ful at 1 site. Four (18.2%) participants were successful at 2 sites. Two (9.1%) participants were successful at all 3 sites. See Table I for a complete summary table outlining methods of determining IO catheter placement success by catheter site reported as success/total attempts.

D i s c u s s i o nThis study confirms that inexperienced clinicians can place IO

catheters in rabbits with moderate success. A recent canine cadaver study did not find that experience influenced success rates nor impacted time to placement with IO catheters (2). In humans, the superiority of IO catheterization over peripheral IV catheterization was demonstrated amongst inexperienced dental medicine students, supporting the role of the IO catheter for the novice placer in emer-gency settings (7).

The results of this study indicate that either the humerus or the tibia should be used for emergent IO catheterization in rabbits. The femur should not be used in the emergent setting by the novice placer. Superiority of the tibia over the radius has been demonstrated in a neonatal foal study (8). Both the humerus and the tibia were desirable for IO catheterizations in humans (9). Tibial IO catheters had less longevity than humeral IO catheters in goats (10). However, these studies were performed in larger animals with significant morphological differences. A feline IO catheterization study showed similar results to the study herein, with no difference in success between the humerus versus the tibia (11). In obese human patients, the tibia has been recommended over the humerus (12); whether this is also true for obese rabbits is unclear.

Speed of placement is of particular interest in the emergency set-ting. This study showed no difference in speed of placement between the humerus versus the tibia, again supporting use of either in the emergent rabbit. The femur took significantly longer to catheterize. A recent cadaveric study in dogs demonstrated faster placement of IO catheters versus central venous catheters by cut-down venotomy (2) and recommended early employment of IO catheters. Similar superiority of IO catheters over central venous catheters has been demonstrated in human emergency settings (13,14). Intraosseous catheters in general can be rapidly placed, with human studies reporting short placement times: # 20 s (9); , 60 s (7), and , 72 s

(14). Importantly, human IO placement is often facilitated by use of a motorized placement device. A study comparing IO catheterization in humans showed the tibia was catheterized faster with respect to the humerus, though time until placement was significantly longer relative to other studies (median: 4.6 min and 7.0 min, respectively) (15). Faster catheterization of the tibia (mean: 33 s) versus the radius (mean: 63 s) was shown in neonatal foals (8). In a canine cadaver models median time until IO catheterization was 54 s (range: 15 to 153 s), showing similar results to this study in rabbits (2).

This study documented that commonly employed clinical or bedside criteria used to identify successful intraosseous catheteriza-tion significantly overestimated truly successful catheterization, i.e., intramedullary placement, as assessed by fluoroscopy. Additionally, participants perceived their catheterizations to be twice as success-ful as they truly were; this is clinically relevant, as administering drugs or fluids by a periosteal route (i.e., outside of the medullary cavity) would be of limited or no benefit. The perceived success, however, may have been impacted by the relative inexperience of the participants.

While establishing correct catheter placement is essential, both fluoroscopy and radiography may be impractical in the critically ill rabbit. In humans, bedside confirmation of successful intraosseous catheter placement has been described by the following 5 criteria: loss of resistance on entering the medullary cavity; a stable, self-standing needle; easy aspiration of bone marrow or blood via the catheter; administration of 2 mL of saline without subcutaneous tissue swelling; and administration of 8 mL of saline without resis-tance (12).

In dogs, aspiration of 2 mL of bone marrow via the catheter has been used to confirm intramedullary placement (16). Collection of 3 mL of bone marrow from the femur of anesthetized rabbits has been shown to be possible, suggesting that this criterion may be useful in confirming intramedullary catheter placement (17); this might be difficult in rabbits which have suffered cardiovascular collapse. As defrosted cadaver rabbits were used, no effort was made to aspirate bone marrow in this study. More recently, the use of a bubble study to confirm correct placement of central venous catheters in humans has been described (18). This technique has not been evaluated for IO catheters, but may be useful as a point-of-care determination of success.

This study had a number of limitations. The use of previously fro-zen cadavers may not accurately represent the tissue characteristics of a living or recently deceased rabbit. These tissue changes may have affected the ability to evaluate success of catheter placement, including perception of success and the ability to manipulate the leg and evaluate for subcutaneous fluid accumulation. The sample size was small. Furthermore, the low number of truly successful catheterizations limits the ability to draw statistically meaningful conclusions, particularly when looking for significance between the humerus and the tibia. The weight of rabbits in this study was vari-able. Rabbits were a collection of New Zealand white and domestic pet rabbits. New Zealand white rabbits may be larger than rabbits encountered in practice. Intraosseous catheters may be easier to place in larger rabbits due to larger landmarks. Conversely, cath-eters may be more easily placed in small rabbits due to less cortical resistance to a hypodermic needle. The goal of this study was not to

Table I. Methods of determining IO catheter placement success by catheter site (reported success/total attempts).

Objective Perceived clinical Fluoroscopic IO catheter site success success true successHumerus 20/22 15/22 10/22Femur 17/22 10/22 5/22Tibia 20/22 16/22 14/22Total 57/66 41/66 29/66



evaluate the effect of rabbit size on successful catheterization and as such was underpowered to draw conclusions; however, in human patients, obesity is known to complicate IO catheter placement. This study did not evaluate the effect of learning. Each participant only attempted each site once. Possibly, more experience would result in greater success. The results of this study may be more applicable to a novice placer.

A key consideration when using hypodermic needles as IO cath-eters is the risk of a bone core plugging the needle. This can result in an unusable catheter. Bone cores were thought to have occurred in the 2 catheters that were assessed as successful by fluoroscopy, but did not meet clinical criteria, as they did not flush initially. A simple way around this potential complication is to remove the plugged catheter and replace it with a new one: new catheters tend to easily find the previous hole. However, this increases time until successful catheterization. Additionally, catheters placed in previously made holes in the bone tend to leak infusate around the catheter more readily. Based on the criteria used in this study, this may indicate failure of placement. Spinal needles or commercial IO needles readily overcome bone coring due to the presence of a stylet. Spinal needles or commercial IO needles readily overcome bone coring issues due to the presence of a stylet. The typical length of many available spinal needles can predispose them to bending though during placement and placement techniques have to be adjusted to account for this potential issue making rapid placement more challenging. Compared to spinal needles, most commercial IO needles are more appropriate in length; however, they are often only available in larger gauges (15G). Commercial IO needles are also considerably more expensive than hypodermic needles. Hypodermic needles were used in this study, as they are more readily available in practice in a variety of gauges and more appropriate lengths.

In conclusion, IO catheters can be placed in rabbits by inexpe-rienced (novice) clinicians with moderate success. Perceptions of success and selected objective clinical success criteria both overesti-mated truly successful intramedullary placement of catheters. Better performing clinical criteria for determining correct placement of IO catheters are required. We recommend that either the humerus or the tibia be used for IO catheterization in the emergent rabbit.

A c k n o w l e d g m e n tThis study was supported by the Tucker and Smeagle Critical

Care Research fund.

Re f e r e n c e s1. Zehnder A. Intraosseous catheter placement in small mammals.

Lab Anim (NY) 2008;37:351–352.2. Allukian AR, Abelson AL, Babyak J, Rozanski EA. Comparison

of time to obtain intraosseous versus jugular venous catheteriza-tion on canine cadavers. J Vet Emerg Crit Care 2017;27:506–511.

3. Fernandez CM, Peyton JL, Miller M, Johnson EG, Kovacic JP. Successful cardiopulmonary resuscitation following cardiopul-monary arrest in a geriatric chinchilla. J Vet Emerg Crit Care 2013;23:657–662.

4. Mazaheri-Khameneh R, Sarrafzadeh-Rezaei F, Asri-Rezaei S, Dalir-Naghadeh B. Comparison of time to loss of conscious-ness and maintenance of anesthesia following intraosseous and intravenous administration of propofol in rabbits. J Am Vet Med Assoc 2012;241:73–80.

5. Huynh M, Boyeaux A, Pignon C. Assessment and care of the critically ill rabbit. Vet Clin North Am Exot Anim Pract 2016; 19:379–409.

6. Kumar AM, Murtaugh R, Brown D, et al. Client donation program for acquiring dogs and cats to teach veterinary gross anatomy. J Vet Med Educ 2001;28:737–7.

7. Goldschalt C, Doll S, Ihle B, Kirsch J, Mutzbauer TS. Peripheral venous or tibial intraosseous access for medical emergency treat-ment in the dental office? Br Dent J 2015;218:E16.

8. Golenz MR, Carlson GP, Madigan JE, Craychee T. Preliminary report: The development of an intraosseous infusion technique for neonatal foals. J Vet Intern Med 1993;7:377–382.

9. Ong ME, Chan YH, Oh JJ, Ngo AS. An observational, prospective study comparing tibial and humeral intraosseous access using the EZ-IO. Am J Emerg Med 2009;27:8–15.

10. Jackson EE, Ashley TC, Snowden KF, et al. Performance and longevity of a novel intraosseous device in a goat (Capra hircus) model. J Am Assoc Lab Anim Sci 2011;50:365–373.

11. Bukoski A, Winter M, Bandt C, Wilson M, Shih A. Comparison of three intraosseous access techniques in cats. J Vet Emerg Crit Care 2010;20:393–397.

12. Petitpas F, Guenezan J, Vendeuvre T, Scepi M, Oriot D, Mimoz O. Use of intra-osseous access in adults: A systematic review. Crit Care 2016;20:102.

13. Chreiman KM, Dumas RP, Seamon MJ, et al. The intraosseous have it: A prospective observational study of vascular access success rates in patients in extremis using video review. J Trauma Acute Care Surg 2018;84:558–563.

14. Lee PM, Lee C, Rattner P, Wu X, Gershengorn H, Acquah S. Intraosseous versus central venous catheter utilization and performance during inpatient medical emergencies. Crit Care Med 2015;43:1233–1238.

15. Santos D, Carron PN, Yersin B, Pasquier M. EZ-IO(®) intraosse-ous device implementation in a pre-hospital emergency service: A prospective study and review of the literature. Resuscitation 2013;84:440–445.

16. Olsen D, Packer BE, Perrett J, Balentine H, Andrews GA. Evaluation of the bone injection gun as a method for intraos-seous cannula placement for fluid therapy in adult dogs. Vet Surg 2002;31:533–540.

17. Eça LP, Ramalho RB, Oliveira IS, et al. Comparative study of technique to obtain stem cells from bone marrow collection between the iliac crest and the femoral epiphysis in rabbits. Acta Cir Bras 2009;24:400–404.

18. Duran-Gehring PE, Guirgis FW, McKee KC, et al. The bubble study: Ultrasound confirmation of central venous catheter place-ment. Am J Emerg Med 2015;33:315–319.


Article


I n t r o d u c t i o nParacetamol, also known as acetaminophen, is a synthetic non-

opiate analgesic drug derived from p-aminophenol. Despite wide-spread use in the human clinical setting, its mechanism of action is complex and not fully understood (1). Paracetamol participates in the central and peripheral inhibition of cyclooxygenases (COX) (2) and probably affects centrally acting COX-3, also known as COX-1b (3). In addition, it may interact with the endogenous serotonergic, opioid, and cannabinoid systems (4,5). Paracetamol produces cen-trally acting analgesic and antipyretic effects, but unlike nonsteroi-dal anti-inflammatory drugs (NSAIDs), it has limited peripheral anti-inflammatory properties (1,6). As a result, gastrointestinal and renal side effects are rare and platelet function is not affected (4,7).

Concerns regarding adverse side effects of opioid drugs (e.g., respiratory and cardiovascular depression, altered thermoregula-tion, sedation, ileus, opioid dependence) have led to rising inter-est in other analgesic drugs (8,9). Adverse effects of paracetamol have been described in veterinary medicine, especially in cats (10). However, it has a high margin of safety in dogs (11) and its reported toxicity is associated with doses higher than those used in pharmacokinetic studies. Dogs administered up to 20 mg/kg body weight (BW) of paracetamol intravenously (IV) do not have side effects (12).

Few studies have assessed the efficacy of paracetamol in the post-operative period (13,14) and to the best of the authors’ knowledge, there are no reports assessing the effects of its intraoperative IV administration in dogs.

Effects of a single paracetamol injection on the sevoflurane minimum alveolar concentration in dogs

Paula González-Blanco, Susana Canfrán, Rubén Mota, Ignacio A. Gómez de Segura, Delia Aguado

A b s t r a c tThis study aimed to determine the effect of a single injection of paracetamol on the sevoflurane minimum alveolar concentration (MAC) response to noxious mechanical stimulation. Seven healthy adult beagles were enrolled in a prospective, randomized, blinded, crossover experimental study. Anesthesia was induced with propofol [11.6 6 2.4 mg/kg body weight (BW)] and maintained with sevoflurane. The MAC was determined before (MAC-1) and after (MAC-2) treatment with 15 mg/kg BW of intravenous (IV) paracetamol or saline over 15 minutes. Samples for plasma paracetamol determination were collected immediately after IV treatment administration and following MAC-2 determination (123 6 27 minutes after starting paracetamol administration). The MAC-1 was similar between treatments (1.7% 6 0.4%). There were no differences between control and paracetamol groups at MAC-2 (2.0% 6 0.4% and 1.7% 6 0.5%, respectively; P = 0.285). Paracetamol plasma concentrations after paracetamol administration were 34.5 6 9.9 mg/mL, decreasing at the end of the procedure (8.5 6 4.2 mg/mL). In conclusion, 15 mg/kg BW of IV paracetamol did not significantly reduce sevoflurane MAC in healthy dogs.

R é s u m éLa présente étude visait à déterminer l’effet d’une injection unique de paracétamol sur la réponse de la concentration alvéolaire minimale (MAC) de sévoflurane à une stimulation mécanique nocive. Sept chiens adultes en santé de race Beagle participèrent à une étude croisée prospective, randomisée, et à l’aveugle. L’anesthésie fut induite avec du propofol [11,6 6 2,4 mg/kg de poids corporel (BW)] et maintenue avec du sévoflurane. La MAC fut déterminée avant (MAC-1) et après (MAC-2) traitement par voie intraveineuse (IV) avec 15 mg/kg BW de paracétamol ou de saline sur une période de 15 minutes. Des échantillons pour déterminer le paracétamol plasmatique furent prélevés immédiatement après l’administration IV du traitement et suivant la détermination de MAC-2 (123 6 27 minutes après le début de l’administration de paracétamol). La valeur de MAC-1 était similaire entre les traitements (1,7 % 6 0,4 %). Il n’y avait pas de différence entre les groupes témoins et paracétamol à MAC-2 (2,0 % 6 0,4 % et 1,7 % 6 0,5 %, respectivement; P = 0,285). Les concentrations plasmatiques de paracétamol après l’administration de paracétamol étaient de 34,5 6 9,9 mg/mL, et diminuaient à la fin de la procédure (8,5 6 4,2 mg/mL). En conclusion, 15 mg/kg de BW de paracétamol par voie IV n’a pas réduit de manière significative la MAC de sévoflurane chez des chiens en santé.


Department of Animal Medicine and Surgery, School of Veterinary Medicine, Complutense University of Madrid, Av. Puerta de Hierro, s/n, 28040 Madrid, Spain (González-Blanco, Canfrán, Gómez de Segura, Aguado); Comparative Medicine Unit, Fundación Centro Nacional de Investigaciones Cardiovasculares, Instituto de Salud Carlos III, 3 Melchor Fernández Almagro St., 28029 Madrid, Spain (Mota).

Address all correspondence to Dr. Delia Aguado; telephone: 134 91 394 3858; e-mail: [email protected]

Dr. Paula González-Blanco was the recipient of a co-funded contract from the Council of Education (Madrid) and the European Social Fund, included in the Youth Employment Initiative.

Received July 3, 2018. Accepted April 12, 2019.



The perioperative efficacy of analgesic drugs under general anesthesia is generally assessed through reductions in the minimum alveolar concentration (MAC) of inhalational anesthetics (15,16). Several analgesic drugs commonly used intraoperatively such as opioids (17) or NSAIDs (18,19) may reduce the MAC in dogs. Previous studies in rats have provided contradictory results and whilst a single paracetamol dose did dose-dependently reduce the sevoflurane MAC by as much as 29% during several hours (20), a lack of effect has also been reported (4). Thus, the aim of this study was to determine the effect of a single paracetamol dose on the sevoflurane MAC in dogs. We hypothesized that paracetamol would decrease sevoflurane MAC in dogs.

M a t e r i a l s a n d m e t h o d sThe study was approved by the Institutional Animal Care and

Use Committee (Proex 032/16). The Animals in Research: Reporting In Vivo Experiments guidelines were followed.

AnimalsSeven healthy adult beagles (6 males and 1 female) aged 6.6 6 0.2 y

and weighing 14.4 6 2.1 kg were used. Dogs were obtained from the institutional colony, originally purchased from an authorized breeder (Isoquimen, Barcelona, Spain). Dogs were housed in groups of 2 to 3 animals per cage (3 3 6 m) using environmental enrichment with natural light at a relative humidity of 40% to 70% and 21°C 6 2°C ambient temperature. They were fed twice daily (Advanced Fitness Adult Medium; Hill’s Pet Nutrition, Madrid, Spain) and water was provided ad libitum. The dogs were judged to be in good health based on physical examination, complete blood (cell) count (CBC), and serum biochemical analysis.

Experimental designThe dogs were anesthetized to determine MAC on 2 occasions in

a prospective, randomized, blinded, crossover design, with a mini-mum 2-week washout period between treatments. The treatment was a single IV injection of 15 mg/kg BW paracetamol (Paracetamol Kabi 10 mg/mL; Fresenius Kabi, Barcelona, Spain) administered over 15 min (6 mL/kg BW per hour) (PRC group) or the equiva-lent volume of saline (FisioVet 1.5 mL/kg BW, NaCl 0.9%; Braun, Barcelona, Spain), also administered over 15 min (CTL group) under general anesthesia. Doses were selected according to previous stud-ies in dogs (12,21,22). Randomization of treatment allocation was

performed with a random number generator (Excel 2007, Microsoft Office, Redmond, Washington, USA).

Anesthetic protocolFood, but not water, was withheld beginning 12 h before the

experiment. All studies were performed in the afternoon. A 20-gauge (G), 1 ¼ inch IV catheter (Surflo IV catheter; Terumo, Madrid, Spain) was placed in the cephalic vein and fluid therapy with lactated Ringer’s solution (Lactato-RingerVet; Braun) was started at a rate of 5 mL/kg BW per hour using an SK-600II infusion pump (Mindray Bio-Medical Electronics, Guangdong, China). Dogs were preoxygen-ated with 100% oxygen via facemask while anesthesia was induced with propofol IV titrated to effect (Propofol-Lipuro 10 mg/mL; Braun). The dose and time of propofol administration were recorded. Orotracheal intubation was performed using an appropriately sized orotracheal tube; dogs were connected to an anesthesia machine (Julian Anesthetic Workstation; Dräger, Madrid, Spain) and placed in left lateral recumbency. The dorsal pedal artery was catheterized with a 22G, 1-inch catheter (Surflo IV Catheter; Terumo) to measure invasive blood pressure. Anesthesia was maintained with sevo-flurane (SevoFlo; Esteve, Barcelona, Spain). Initially set at 2.0% to 2.4%, end-tidal concentration of sevoflurane (ETSEV) was delivered in a gas mixture of oxygen and air (FiO2 50%) in a continuous flow of 3 L/min, via a circle anesthetic rebreathing system. Dogs were allowed to breathe spontaneously but were mechanically venti-lated when necessary (i.e., volume control using a tidal volume of 10 mL/kg, a 20% inspiratory pause, and a positive end-expiratory pressure of 5 cm H2O) to maintain normocapnia [i.e., end-tidal car-bon dioxide (ETCO2) between 4.7 and 6 kPa). Mechanical ventilation was stopped when dogs started fighting the ventilator, usually after the application of the noxious stimuli.

MonitoringDuring the procedure, systolic, diastolic, and mean arterial pres-

sures, arterial oxygen hemoglobin saturation (SpO2), heart rate using lead II electrocardiogram, respiratory rate (RR), ETCO2, ETSEV, and body temperature were continuously monitored using a monitoring system (Datex-Ohmeda S/5 Anesthesia Monitor; General Electric, Helsinki, Finland). The blood pressure was monitored invasively; alternatively, when arterial catheterization was not possible, the oscillometric method would be used by placing the cuff on the dog’s dorsal pedal artery.

Figure 1. Experimental design.MAC — minimum alveolar concentration; IV — intravenous.

IV catheter

5 min 20 min 20 min

15 min

60–120 min 60–120 min

Propofol Paracetamol Saline

Orotraqueal intubation

Paracetamol plasma

concentration

Paracetamol plasma

concentration

MAC-1 MAC-2



Airway gas samples were obtained from a sampling port located between the proximal end of the endotracheal tube and the breathing system and directed into an infrared gas analyzer (Datex-Ohmeda S/5 Anesthesia Monitor; General Electric) to monitor RR, ETCO2, and ETSEV. The gas analyzer was calibrated at the start of the experiment using a reference gas (Quick Cal Calibration Gas; General Electric). Body temperature was monitored using an esophageal probe and maintained between 37.0°C and 38.5°C with a circulating warming water blanket (Heat Therapy Pump, Model TP-220; Gaymar, New York, USA) and a warm air blanket (Bair Hugger Model 505; 3M Health Care, Neuss, Germany), if necessary. Data were recorded 60 s before each tail clamping and values from the 4 clamping procedures employed for MAC determination were averaged to provide each MAC value.

Sevoflurane MAC determinationAfter a 20-minute equilibration period with ETSEV kept at 2.0% to

2.4%, MAC-1 determination was initiated.A noxious stimulus was applied employing the tail clamp method

as previously described (16) using a Doyen intestinal clamp attached to the first ratchet lock onto the tail for 60 s or until a positive response immediately after acquiring the monitored data recording. The tail was always stimulated at different sites starting 12 cm distal to the tail base in an attempt to prevent potential pain tolerance or sensitization. A positive response was considered a gross purpose-ful movement of the head, limbs, or body. A negative response was considered to be the lack of movement or progressive withdrawal movement of limbs, swallowing, or tail flick. When a positive response was seen, ETSEV was increased by 0.20% step increases until the positive response became negative. Similarly, when a negative response was seen, ETSEV was reduced in decrements of 0.20% until the negative response became positive (23,24).

Thus, MAC was determined as the concentration mid-way between the highest concentration that permitted movement in response to the stimulus and the lowest concentration that prevented such movement. The MAC was calculated in duplicate and the mean value was considered the MAC value for that animal. Times from anesthetic induction with propofol until MAC determination

as well as the duration of MAC determination were recorded. The same researcher (PGB), blinded to the treatment, was responsible for applying the noxious stimuli and assessing the response in all instances.

Drug administration and experimental procedureAfter MAC-1 determination, there was a 20-minute equilibration

period, whereby ETSEV was maintained at that individual MAC-1 value. Dogs were randomly administered a single IV injection of either 15 mg/kg BW of 1% paracetamol over 15 min (6 mL/kg BW per hour) (PRC group) or 1.5 mL/kg BW of saline over 15 min (CTL group). Both treatments were administered using a syringe pump (Infusomat fmS; Braun). After paracetamol or saline administration, MAC was re-calculated (MAC-2).

At the end of each experiment, sevoflurane administration was discontinued, and dogs recovered from anesthesia (Figure 1). A CBC and serum biochemical analysis was repeated twice in each dog, 1 wk after each treatment.

Blood serum detection and quantification of paracetamol

Venous blood samples (6 mL) were drawn in sterile serum sepa-rator tubes (Corning 15 mL polypropylene centrifuge tubes; Merck KGaA, Madrid, Spain) from a peripheral vein twice: 1 to 2 min immediately after administering the treatment and just after deter-mining MAC-2 (over 2 h after starting paracetamol administration). Blood was refrigerated and centrifuged to obtain the serum and stored at 280°C in Eppendorf tubes. Samples for plasma paracetamol determination were collected immediately after IV administration of treatment and following determination of MAC-2.

The Dimension RxL Max Integrated Chemistry System (Siemens Healthcare Diagnostics, Newark, Delaware, USA) and its ACTM Flex reagent cartridge were used to detect and quantify the level of paracetamol in 100 mL serum samples. The protocol assay is based on enzymatic hydrolysis-producing acetate and p-aminophenol. The level of p-aminophenol is determined colorimetrically by reac-tion with o-cresol and ammoniacal copper sulphate, producing indophenol with a wavelength absorbance of 600 nm. The amount

Table I. Monitored parameters from 7 healthy beagles during minimum alveolar concentration (MAC) determination. Data are expressed as mean 6 standard deviation.

Treatment Paracetamol Saline MAC-1 MAC-2 MAC-1 MAC-2Systolic arterial pressure (mmHg) 123 6 17 130 6 11 118 6 12 122 6 22Diastolic arterial pressure (mmHg) 72 6 15 78 6 13 80 6 10 73 6 5Mean arterial pressure (mmHg) 88 6 11 94 6 10 93 6 10 88 6 9Heart rate (beats/min) 114 6 23 117 6 30 108 6 16 115 6 22Respiratory rate (breaths/min) 42 6 28 32 6 16 29 6 16 25 6 14ETCO2 (kPa) 4.3 6 1.0 4.4 6 0.5 4.7 6 0.5 4.5 6 0.2SpO2 (%) 95 6 3 96 6 2 97 6 2 98 6 3Esophageal temperature (°C) 37.5 6 0.6 37.5 6 0.6 37.2 6 0.5 37.4 6 0.5ETCO2 — end-tidal carbon dioxide; SpO2 — arterial oxygen hemoglobin saturation.



of p-aminophenol produced and estimated by the analyzer is pro-portional to the blood concentration of acetaminophen. Data are expressed in mg/mL.

Statistical analysisThe sample size calculation was performed for a 2-tailed Student’s

t-test with a power of 80% and an alpha error of 0.05 to detect a dif-ference of 0.5% in sevoflurane MAC. This determined a minimum of 7 dogs per group.

Data were tested for normality employing the Kolmogorov-Smirnov test. Monitored parameters at MAC-1 and MAC-2 included arterial blood pressure, heart and respiratory rates, ETCO2, SpO2, esophageal temperature, and time from propofol administration to MAC-1 and MAC-2 determination. Propofol induction dose, time required to determine MAC-1 and MAC-2, and paracetamol plasma concentrations were analyzed using the 2-tailed Student’s paired t-test. To assess the effect of the treatment on MAC, the Student’s t-test was also used to compare the MAC-1 and MAC-2 from each treatment group. The correlation between baseline MAC and time from propofol administration was analyzed with the Pearson test.

Data are shown as mean 6 standard deviation and a P-value , 0.05 was set to indicate statistical significance. All statistical analyses were performed using SPSS for Windows (IBM SPSS Statistics V22.0).

Re s u l t sThe MAC-1 was similar in both groups (1.7% 6 0.5% and 1.6% 6

0.3% in CTL and PRC groups, respectively; P = 0.525). The MAC-2 value was 13% lower in the PRC group than in the CTL group, but there were no statistically significant differences in MAC-2 between treatments (2.0% 6 0.4% and 1.7% 6 0.5% in CTL and PRC groups, respectively; P = 0.285). The MAC-2 was significantly higher than MAC-1 in the CTL group (P = 0.025) and no significant differences were observed between MAC-1 and MAC-2 in the PRC group (P = 0.444).

No adverse effects were noted when administering paracetamol to the dogs. There were no significant differences in arterial blood pressure, heart and respiratory rates, ETCO2, SpO2, and temperature between treatment groups (Table I).

The time needed to determine MAC-1 was shorter in the PRC than in the CTL group (83 6 18 min and 122 6 38 min, respectively; P = 0.037) but similar for MAC-2 (86 6 24 min and 94 6 29 min, respectively; P = 0.590).

The propofol dose required for intubation in all non-premedicated dogs was 11.6 6 2.4 mg/kg BW and did not differ between treatment groups (11.7 6 2.7 mg/kg BW and 11.4 6 2.4 mg/kg BW in PRC and CTL groups, respectively; P = 0.811). The time from propofol induction until MAC determination was similar in both groups (MAC-1: 125 6 17 min and 165 6 42 min in PRC and CTL groups, respectively; P = 0.051 and MAC-2: 255 6 33 min and 295 6 35 min in PRC and CTL groups, respectively; P = 0.051). A positive cor-relation between the MAC and time to propofol administration in MAC-1 was observed (P = 0.027) (Figure 2).

Paracetamol plasma concentrations immediately after the administration of paracetamol were significantly higher in the PRC

group (34.5 6 9.9 mg/mL) than the CTL group (3.8 6 1.7 mg/mL) (P = 0.000). These concentrations decreased when MAC-2 was deter-mined (123 6 27 min from starting paracetamol administration) and were significantly higher than in the CTL group (8.5 6 4.2 mg/mL and 1.6 6 0.9 mg/mL in the PRC and CTL groups, respectively; P = 0.027).

D i s c u s s i o nParacetamol 15 mg/kg BW, IV was not successful at reducing

the sevoflurane MAC in dogs. Previous studies had determined a sevoflurane MAC reduction of 30% in rats for at least 4 h following a single intraperitoneal dose (300 mg/kg BW) (20). Different species, but also pharmacokinetics, dose, and administration route, may account for these apparently contradictory results. However, another study using a single dose of 300 mg/kg BW, IV in rats reported a lack of effect on the MAC and no MAC reduction of isoflurane (4). Different routes of administration, anesthetic gases used, and equili-bration time from drug administration to MAC determination, may account for these apparently contradictory results.

The MAC reduction produced by paracetamol (13%) here was not statistically significant and may reflect a limited effect by paracetamol, similar to that produced by other drugs, including NSAIDs. Robenacoxib, a selective COX-2 inhibitor, decreased the sevoflurane MAC for blunting adrenergic response by 17% in dogs (18). Similarly, carprofen and meloxicam decreased the sevoflurane MAC by 11% and 13%, respectively (19). However, a lack of effect from orally administered carprofen was reported in the isoflurane MAC, in which a non-statistically significant MAC reduction of 9% was detected (25).

Although paracetamol might not have a relevant anesthetic- sparing effect, if any, it may still provide clinically useful peri-operative analgesia (13,14). The reduced side effects of paracetamol could support its perioperative administration in patients with gastrointestinal or renal impairment and when NSAIDs are contra-indicated. All data from the physical examination and hematological

Figure 2. Positive correlation between sevoflurane MAC-1 and time from propofol induction (P = 0.027).MAC — minimum alveolar concentration.

Sevo

flur

ane

MA

C-1

Time from propofol induction (min)

3.0

2.5

2.0

1.5

1.0

0.5

0.00 50 100 150 200 250



and biochemical analyses in this study showed values within the physiological range and no side effects were noted.

The baseline sevoflurane MAC (MAC-1) is lower than in previ-ous studies that reported values between 2.4% 6 0.3% (15) and 2.7% 6 0.2% (21), and similar to MAC studies employing propofol as the anesthetic induction agent (1.8% 6 0.1%) (26). Factors affecting variability in MAC values include the type of noxious stimulus, the anatomical site of stimulation, and subjectivity in the interpretation of purposeful movement. Other factors such as ETCO2, body tem-perature, and arterial blood pressure were controlled in this study (27). Finally, the age of the dogs included in the study (6 y) may also affect the MAC value, since the sevoflurane MAC is 17% lower in older dogs (8 to 10 y) than in dogs under 2 y (24). Sevoflurane MAC of 1.9% 6 0.3% has been reported in dogs aged 8 to 10 y (28). Additionally, the effects of an anesthetic induction using propofol on sevoflurane MAC should be considered. Propofol clearance is slower in unpremedicated elderly dogs, perhaps due to age-related physiological changes such as an increase in the lipid compartment, where lipophilic drugs like propofol distribute, which would result in slower body clearance (29).

At the beginning of MAC-1 determination, 30 to 60 min after anesthetic induction, propofol plasma concentrations of at least 0.95 mg/mL were expected (29). Lower values of at least 0.3 mg/mL were expected once MAC-1 was determined, 2 to 3 h after anesthetic induction (29). Propofol decreases MAC in a dose-dependent man-ner; thus, the propofol blood levels in our study may have reduced the MAC by at least 23% and 16%, at the beginning and end of MAC-1 determination, respectively, as suggested by studies using propofol under a continuous infusion rate (30). We found a positive correlation between MAC-1 and the time required for its determina-tion, since longer times favored propofol clearance and thus a higher MAC value. A longer equilibration time before the MAC baseline determination could have reduced this effect. In addition, residual propofol in plasma may explain why MAC-2 increased significantly in the CTL group, which had received no other co-administered drugs. However, confirming this requires assessing propofol plasma concentrations, which was not done here. To prevent the interference of propofol or any other drug in the MAC, dogs should ideally use an anesthetic mask for inhalational induction. There are 3 reasons for using a propofol induction: i) it produces a less stressful anesthetic induction, ii) propofol administration better mimics the clinical situation where mask induction is rarely employed, and iii) induc-tion with propofol avoids atmospheric contamination inside the operating room. Similarly, previous experimental (26,31,32) and clinical (33,34) MAC studies determining the inhalational anesthetic sparing action of analgesics, anesthetics, and sedatives used propofol to induce anesthesia in dogs; although, lower induction doses or longer equilibration times were used in those studies.

Paracetamol plasma concentrations were 34.5 6 9.9 mg/mL before the second MAC determination, approximately 2 h after propofol anesthetic induction. Time required to determine MAC-2 could have affected the paracetamol plasma concentration and therefore, the MAC reduction observed in each dog. Lower plasma concentrations of 7 to 8 mg/mL have been observed in greyhounds 1 h after 8 to 23 mg/kg of paracetamol was administered orally (22,35). A mean maximum serum concentration of approximately 45 mg/mL has

been determined after an oral dose of 100 mg/kg BW and 90 mg/mL after a dose of 200 mg/kg BW (11). The differences may reflect the administration route of the drug. Previous studies have not reported on paracetamol plasma concentrations using IV administration nor have they correlated the plasmatic concentration of paracetamol with analgesic efficacy in dogs. A similar correlation has been observed in humans with postoperative dental pain (36).

As expected, the paracetamol plasma concentration decreased at the end of the study. Paracetamol is rapidly absorbed and elimi-nated within 1 h in dogs (22,35) and humans, with a rapid onset of analgesic effect within 5 min (1). Therefore, a higher MAC reduction might be expected just after paracetamol administration and an intra-operative continuous infusion would have provided an improved effect. As a limitation of the study, it should be noted that low levels of paracetamol were measured in the CTL group, probably as a consequence of phenolic compound cross-detection by the analytic method employed (37). The resulting measured levels were too low (4 6 2 mg/mL) to have modified the observed results.

There are additional limitations to this study. First, the reduction in the sevoflurane MAC does not necessarily reflect an analgesic effect, although this has been suggested (36), as with opioids. Second, the lack of a paracetamol effect on MAC may be due to the low statistical power; a higher sample size could have made a significant difference. To achieve a power of 80%, a larger sample size of 8 to 10 dogs per group (1- and 2-tailed, respectively) would have provided higher MAC variation and made a statistically significant difference. Furthermore, additional paracetamol doses could have been studied to determine a potential dose-dependent sevoflurane-sparing effect. Finally, the use of propofol might have modified the sevoflurane MAC and affected paracetamol plasma measurements. Longer equilibration times before the MAC baseline determination and lower propofol doses could have been used to minimize this effect. However, it was decided to use an anesthetic induction agent such as propofol since it is widely used and better mimics a clinical setting, whilst inhalational anesthetic induction is rarely used and is banned in many clinical practices because of personnel safety concerns.

Despite the limitations to the study, the relatively small anesthetic sparing effect of paracetamol is of limited clinical value when MAC reductions above 20% are considered more desirable. This decrease is higher than the expected inter-individual variability of MAC (27), making it unnecessary to increase the number of dogs.

In conclusion, 15 mg/kg BW of IV paracetamol did not signifi-cantly reduce sevoflurane MAC in healthy dogs.

A c k n o w l e d g m e n t sThe authors acknowledge technical support provided by

Alejandro Sánchez, Rocío Bustamante, and Mario Arenillas, and the support from the Department of Pharmacology, Faculty of Veterinary Medicine, Complutense University of Madrid.

Re f e r e n c e s 1. Jahr JS, Lee VK. Intravenous acetaminophen. Anesthesiol Clin

2010;28:619–645.



2. Smith HS. Potential analgesic mechanisms of acetaminophen. Pain Physician 2009;12:269–280.

3. Chandrasekharan NV, Dai H, Roos KL, et al. COX-3, a cyclooxygenase-1 variant inhibited by acetaminophen and other analgesic/antipyretic drugs: Cloning, structure, and expression. Proc Natl Acad Sci U S A 2002;99:13926–13931.

4. Chavez JR, Ibancovichi JA, Sanchez-Aparicio P, Acevedo-Arcique CM, Moran-Muñoz R, Recillas-Morales S. Effect of acetamino-phen alone and in combination with morphine and tramadol on the minimum alveolar concentration of isoflurane in rats. PLoS One 2015;10:e0143710.

5. Graham GG, Davies MJ, Day RO, Mohamudally A, Scott KF. The modern pharmacology of paracetamol: Therapeutic actions, mechanism of action, metabolism, toxicity and recent pharma-cological findings. Inflammopharmacology 2013;21:201–232.

6. Epstein ME, Rodanm I, Griffenhagen G, et al. 2015 AAHA/AAFP pain management guidelines for dogs and cats. J Feline Med Surg 2015;17:251–272.

7. Hinz B, Brune K. Antipyretic analgesics: Nonsteroidal antiin-flammatory drugs, selective COX-2 inhibitors, paracetamol and pyrazolinones. Handb Exp Pharmacol 2007:65–93.

8. Sultana A, Torres D, Schumann R. Special indications for Opioid Free Anaesthesia and Analgesia, patient and procedure related: Including obesity, sleep apnoea, chronic obstructive pulmonary disease, complex regional pain syndromes, opioid addiction and cancer surgery. Best Pract Res Clin Anaesthesiol 2017;31: 547–560.

9. White DM, Mair AR, Martinez-Taboada F. Opioid-free anaesthe-sia in three dogs. Open Vet J 2017;7:104–110.

10. McConkey SE, Grant DM, Cribb AE. The role of para- aminophenol in acetaminophen-induced methemoglobinemia in dogs and cats. J Vet Pharmacol Ther 2009;32:585–595.

11. Savides MC, Oehme FW, Nash SL, Leipold HW. The toxicity and biotransformation of single doses of acetaminophen in dogs and cats. Toxicol Appl Pharmacol 1984;74:26–34.

12. Serrano-Rodríguez JM, Mengual C, Quirós-Carmona S, et al. Comparative pharmacokinetics and a clinical laboratory evalua-tion of intravenous acetaminophen in Beagle and Galgo Español dogs. Vet Anaesth Analg 2019;46:226–235.

13. Mburu DN. Evaluation of the anti-inflammatory effects of a low dose of acetaminophen following surgery in dogs. J Vet Pharmacol Ther 1991;14:109–111.

14. Mburu DN, Mbugua SW, Skoglund LA, Lökken P. Effects of paracetamol and acetylsalicylic acid on the post-operative course after experimental orthopaedic surgery in dogs. J Vet Pharmacol Ther 1988;11:163–170.

15. Docquier MA, Lavand’homme P, Ledermann C, Collet V, De Kock M. Can determining the minimum alveolar anesthetic concentration of volatile anesthetic be used as an objective tool to assess antinociception in animals? Anesth Analg 2003;97: 1033–1039.

16. Eger EI 2nd, Saidman LJ, Brandstater B. Minimum alveolar anesthetic concentration: A standard of anesthetic potency. Anesthesiology 1965;26:756–763.

17. Credie RG, Teixeira Neto FJ, Ferreira TH, Aguiar AJ, Restitutti FC, Corrente JE. Effects of methadone on the minimum alveo-

lar concentration of isoflurane in dogs. Vet Anaesth Analg 2010;37:240–249.

18. Tamura J, Itami T, Ishizuka T, et al. Sparing effect of robenacoxib on the minimum alveolar concentration for blunting adrenergic response (MAC-BAR) of sevoflurane in dogs. J Vet Med Sci 2014; 76:113–117.

19. Yamashita K, Okano Y, Yamashita M, Umar MA, Kushiro T, Muir WW. Effects of carprofen and meloxicam with or without butorphanol on the minimum alveolar concentration of sevoflu-rane in dogs. J Vet Med Sci 2008;70:29–35.

20. Benito J, Aguado D, Abreu M, Gómez de Segura I. Abstracts presented at the 10th World Congress of Veterinary Anaesthesia, 31st August–4th September 2009, Glasgow, UK. Veterinary Anaesthesia and Analgesia 2010;37:1–88.

21. Mathews K, Kronen PW, Lascelles D, et al. Guidelines for recog-nition, assessment and treatment of pain: WSAVA Global Pain Council members and co-authors of this document. J Small Anim Pract 2014;55:E10–68.

22. KuKanich B. Pharmacokinetics and pharmacodynamics of oral acetaminophen in combination with codeine in healthy Greyhound dogs. J Vet Pharmacol Ther 2016;39:514–517.

23. Aguado D, Abreu M, Benito J, Garcia-Fernandez J, Gómez de Segura IA. Effects of naloxone on opioid-induced hyperalgesia and tolerance to remifentanil under sevoflurane anesthesia in rats. Anesthesiology 2013;118:1160–1169.

24. Yamashita K, Iwasaki Y, Umar MA, Itami T. Effect of age on minimum alveolar concentration (MAC) of sevoflurane in dogs. J Vet Med Sci 2009;71:1509–1512.

25. Ko JC, Weil AB, Inoue T. Effects of carprofen and morphine on the minimum alveolar concentration of isoflurane in dogs. J Am Anim Hosp Assoc 2009;45:19–23.

26. Moran-Muñoz R, Ibancovichi JA, Gutierrez-Blanco E, et al. Effects of lidocaine, dexmedetomidine or their combination on the minimum alveolar concentration of sevoflurane in dogs. J Vet Med Sci 2014;76:847–853.

27. Quasha AL, Eger EI 2nd, Tinker JH. Determination and applica-tions of MAC. Anesthesiology 1980;53:315–334.

28. Itami T, Kawase K, Tamaru N, et al. Effects of a single bolus intravenous dose of tramadol on minimum alveolar concen-tration (MAC) of sevoflurane in dogs. J Vet Med Sci 2013;75: 613–618.

29. Reid J, Nolan AM. Pharmacokinetics of propofol as an induction agent in geriatric dogs. Res Vet Sci 1996;61:169–171.

30. Singsank-Coats J, Seddighi R, Rohrbach BW, et al. The anesthetic interaction of propofol and sevoflurane on the minimum alveolar concentration preventing motor movement (MACNM) in dogs. Can J Vet Res 2015;79:95–100.

31. Acevedo-Arcique CM, Ibancovichi JA, Chavez JR, et al. Lidocaine, dexmedetomidine and their combination reduce isoflurane mini-mum alveolar concentration in dogs. PLoS One 2014;9:e106620.

32. Ueyama Y, Lerche P, Eppler CM, Muir WW. Effects of intrave-nous administration of perzinfotel, fentanyl, and a combination of both drugs on the minimum alveolar concentration of isoflu-rane in dogs. Am J Vet Res 2009;70:1459–1464.

33. Aguado D, Benito J, Gómez de Segura IA. Reduction of the minimum alveolar concentration of isoflurane in dogs using a



constant rate of infusion of lidocaine-ketamine in combination with either morphine or fentanyl. Vet J 2011;189:63–66.

34. Monteiro ER, Coelho K, Bressan TF, Simões CR, Monteiro BS. Effects of acepromazine-morphine and acepromazine-methadone premedication on the minimum alveolar concentration of iso-flurane in dogs. Vet Anaesth Analg 2016;43:27–34.

35. KuKanich B. Pharmacokinetics of acetaminophen, codeine, and the codeine metabolites morphine and codeine-6-glucuronide in healthy Greyhound dogs. J Vet Pharmacol Ther 2010;33:15–21.

36. Seymour RA, Rawlins MD. Pharmacokinetics of parenteral paracetamol and its analgesic effects in post-operative dental pain. Eur J Clin Pharmacol 1981;20:215–218.

37. Navarro I, Gonzalez-Arjona D, Roldan E, Rueda M. Determina-tion of paracetamol in tablets and blood plasma by differential pulse voltammetry. J Pharm Biomed Anal 1988;6:969–976.


Article


Correlation of activity data in normal dogs to distance traveledBishoy S. Eskander, Megan Barbar, Richard B. Evans, Masataka Enomoto,

B. Duncan X. Lascelles, Michael G. Conzemius

A b s t r a c tThe objective of this study was to explore the mathematical relationships between independent variables (patient morphometrics and treadmill speed) and dependent variables (accelerometer or pedometer output) when evaluating data from accelerometers and pedometers in dogs. Twenty dogs took part in 3 randomized activities, consisting of exercise on a treadmill at 1.0, 1.5, and 2.0 m/s for a total distance of 1 km at each speed. Dogs simultaneously wore both an accelerometer and a pedometer. Statistical analysis used multiple regression models to discover the relationships between independent and dependent variables. A formula was developed to predict the distance traveled by a dog based on its morphometrics and activity monitor output. Shoulder height had stronger correlations to accelerometer and pedometer outputs than other morphometric variables. As shoulder height increased, all accelerometer and pedometer outputs decreased. As treadmill speed increased, both accelerometer and pedometer step counts decreased, while accelerometer activity counts increased. According to a formula derived to predict the total distance traveled using patient shoulder height and accelerometer or pedometer output, pedometer steps were the most accurate predictor of distance traveled. Accelerometer steps were less accurate when using the same model. Accelerometer activity counts did not reveal a meaningful predictive formula. The results of this study indicate that patient morphometrics and treadmill speed (as a measure of intensity) influenced accelerometer and pedometer data. The pedometer data more precisely and accurately estimated the distance traveled based on step counts and patient shoulder height. In normal dogs, accelerometer and pedometer steps may reasonably estimate distance traveled.

R é s u m éL’objectif de la présente étude était d’explorer les relations mathématiques entre des variables indépendantes (données morphométriques du patient et vitesse du tapis d’exercice) et des variables dépendantes (accéléromètre ou données du podomètre) lors de l’évaluation des données provenant d’accéléromètres et de podomètres chez des chiens. Vingt chiens ont pris part à trois activités randomisées, consistant en des exercices sur un tapis roulant à 1,0, 1,5, et 2,0 m/s pour une distance totale de 1 km à chaque vitesse. Les chiens portaient simultanément un accéléromètre et un podomètre. Les analyses statistiques utilisèrent des modèles de régression multiple afin de découvrir les relations entre les variables indépendantes et dépendantes. Une formule fut développée afin de prédire la distance parcourue par un chien sur la base de sa morphométrie et les données des moniteurs d’activité. La hauteur à l’épaule avait la plus forte corrélation aux données de l’accéléromètre et du podomètre comparativement aux autres variables morphométriques. À mesure que la hauteur à l’épaule augmentait, toutes les données de l’accéléromètre et du podomètre diminuaient. Avec l’augmentation de la vitesse du tapis d’exercice, il y eu une diminution du nombre de pas mesuré par l’accéléromètre et le podomètre, alors qu’il y avait une augmentation du compte d’activité de l’accéléromètre. Selon la formule dérivée pour prédire la distance totale parcourue en utilisant la hauteur à l’épaule du patient et les données de l’accéléromètre ou du podomètre, le nombre de pas du podomètre était le prédicteur le plus précis de la distance parcourue. Le nombre de pas avec l’accéléromètre était moins précis en utilisant le même modèle. Le dénombrement des activités par l’accéléromètre n’ont pas permis de déterminer une formule prédictive significative. Les résultats de cette étude indiquent que les données morphométriques des patients et la vitesse du tapis d’exercice (comme mesure d’intensité) influencent les résultats de l’accéléromètre et du podomètre. Les données du podomètre ont estimé avec plus de justesse et de précision la distance parcourue en se basant sur le nombre de pas et la hauteur à l’épaule du patient. Chez les chiens normaux, le nombre de pas mesuré par un accéléromètre et un podomètre peut raisonnablement estimer la distance parcourue.


University of Minnesota, College of Veterinary Medicine, Department of Veterinary Clinical Sciences, 1365 Gortner Avenue, St. Paul, Minnesota 55108, USA (Eskander, Evans, Conzemius); Comparative Pain Research Laboratory, Department of Clinical Sciences, College of Veterinary Medicine, North Carolina State University, 1060 William Moore Drive, Raleigh, North Carolina 27607, USA (Barbar, Enomoto, Lascelles); Center for Pain Research and Innovation, UNC School of Dentistry, 385 South Columbia Street, Chapel Hill, North Carolina 27599, USA (Lascelles); Center for Translational Pain Research, Department of Anesthesiology, Duke University, 134 Research Drive, Durham, North Carolina 27710, USA (Lascelles).

Address all correspondence to Dr. Bishoy S. Eskander; telephone: (952) 942-8272; fax: (952) 829-4089; e-mail: [email protected]

Dr. Eskander’s current address is 7717 Flying Cloud Drive, Eden Prairie, Minnesota 55344, USA.

Received October 14, 2018. Accepted January 18, 2019.



I n t r o d u c t i o nAssessments of the response of veterinary patients to many

therapies remain largely subjective, especially when considering the patient’s response to treatment at home. An objective outcome measure that effectively characterizes chronic pain, e.g., severity and frequency, and potentially the response to therapy in veterinary patients in the home environment remains largely elusive. The lack of such a measure limits investigations addressing naturally occur-ring chronic pain in veterinary patients and makes it more challeng-ing to identify drugs for treating chronic pain. This difficulty may help explain why FDA-approved treatments for chronic pain in dogs have used owner survey responses (in blinded, placebo-controlled protocols) as primary outcome measures (1,2).

There has recently been increased interest in the use of accel-erometers and pedometers to measure the response of diseased canine patients to various therapies (3–8) and to correlate data from these devices with the actual behavior and movements of the animals (9–12). Briefly, multiaxis accelerometers measure accelera-tion in the x-, y-, and z-axis and report not only when acceleration has changed in each axis but also the magnitude of the change. In contrast, piezoelectric pedometers generally measure acceleration in a single axis and report only when a predetermined threshold in acceleration, e.g., 1/3 G, has been exceeded. Research evaluating the response of canine patients with osteoarthritis (OA) to non-steroidal anti-inflammatory drug (NSAID) therapy using accelerometers has been conducted with some success (3–6). Previous evalua-tions of accelerometers in relation to movements of patients with OA have primarily focused on activity counts generated from an omnidirectional accelerometer. Limitations of this device include decreased precision over longer distances (9), lack of qualification of the intensity of motion by the accelerometer, and difficulty in identifying true active movements or distance traveled compared to “mobility” due to other reasons such as pruritus or grooming (13,14). Despite these limitations, a Spearman correlation coefficient of 0.78 has been reported between accelerometer activity counts and distance traveled documented by videography (7,12). However, one of these studies described similarly sized dogs walking short distances. While that report includes valuable information, without video confirmation of the patient’s mobility, we are unaware of an effective means to describe distance traveled based on accelerometer and/or pedometer data.

In the human medical field, accelerometers and pedometers have been extensively researched with a similar goal of evaluating patient response to therapies at home between hospital evaluations (15–17). Much of the literature highlights the use of accelerometers and pedometers to monitor physician-prescribed activity for vari-ous medical diseases, such as evaluating changes in patient activity after discharge from the hospital for managing chronic obstructive pulmonary disease (COPD) (18–20).

The objective of this study was to explore the mathematical rela-tionships between independent variables (patient morphometrics and treadmill speed) and dependent variables (accelerometer or pedometer output) in normal dogs wearing an accelerometer and a pedometer. Our null hypothesis was that there was no relationship between the independent and dependent variables.

M a t e r i a l s a n d m e t h o d sThe Institutional Animal Care and Use Committee at all partici-

pating sites approved this study (NC State: 16-086-O; Minnesota: 1605-33721A). Informed client consent was obtained before the screening process. Investigators at the 2 sites met before the study began to ensure that study procedures were identical.

Inclusion criteriaAll dogs participating in the study were older than 1 y of age

and were evaluated to ensure they could walk well on a leash. A general physical examination was carried out to ensure that dogs were in good general health, were not pregnant or in heat, and had no orthopedic or neurologic abnormalities. Ten Labrador retrievers and 10 beagles were studied, 5 of each breed at each site.

Monitors evaluatedThe accelerometer evaluated in this study (Actical Z-series

Physical Activity Monitor; Philips Respironics Mini Mitter Division, Bend, Oregon, USA) is an omnidirectional accelerometer that is designed to detect body movement in humans. It is built from a cantilevered piezoelectric bimorph plate and seismic mass and is sensitive to movement in all directions, but is most sensitive to movement in the direction parallel with the longest dimension of the case. The output from this accelerometer is activity counts and/or step counts (21,22). The pedometer evaluated [Omron Pedometer (HJ-720ITC); Omron Healthcare, Lake Forest, Illinois, USA] is a piezoelectric pedometer with dual-axis acceleration sensors that counts steps when placed horizontally or vertically.

The accelerometer and pedometer were secured to generic nylon collars that were applied to the dog approximately 2 cm apart from each other. The Actical accelerometer has a metal case with slit holes so that bands or a collar can be placed through it, with the long axis of the Actical case lying along the length of the collar, as described in a previous study (7). The pedometer comes with a holder clip that was fastened to the collar with zip ties and secured upright, perpendicular to the long axis of the collar, as recommended by the manufacturer (Figure 1). The accelerometer and pedometer were positioned on the ventral aspect of the neck and tightened to pro-vide a space equivalent to 2 finger widths between the collar and the dog’s neck. The collars were applied by a single person at each participating site.

Subject measurementsPatient morphometrics were measured after enrollment and

included shoulder height (floor to acromion process), stance dis-tances (distances between paws or feet), body weight, and body condition score (BCS). Since stride length influences the relation-ship between steps taken and distance traveled, stance distances were measured to determine the strength of their relationship with the accelerometer and pedometer output. Shoulder height and stance distance were measured with the dog standing square. Measurements included: i) the distance between the front limbs [medial aspect of the left front (LF) foot to the medial aspect of the right front (RF) foot]; ii) the distance between the hind limbs [medial aspect of the left hind (LH) foot to the medial aspect of the right hind



(RH) foot]; iii) the distance between the caudal aspect of the RF foot and the cranial aspect of the RH foot; iv) the distance between the caudal aspect of the LF foot and the cranial aspect of the LH foot; and v) the inter-limb distance (the diagonal distance from the cau-dal aspect of a front foot to the cranial aspect of the contralateral hind foot) and the standing hypotenuse (the distance between the acromion process and the caudal aspect of the contralateral hind foot, which was calculated using existing measurements and the Pythagorean theorem). For dogs that were not standing square, opposing measurements were averaged.

ProtocolAlthough both the accelerometer and pedometer were placed

simultaneously around each dog’s neck with a single collar (Figure 1), the order in which each device was arranged on the collar (which was on right or left on the collar) was randomized for each dog in each breed group. The order of the activities was similarly randomized for each dog within each breed group. The randomization of the order in which the monitors were placed on the collars and the order of activities for all 10 subjects of each breed was determined using a web-based program (Research Randomizer, Version 4.0; retrieved on June 15, 2016 from http://www.randomizer.org/). This randomization was then split between the 2 study sites.

Since walking on a treadmill was included in the study, dogs were acclimated to the treadmill (L880 Rehabilitation Treadmill;

Landice, Randolph, New Jersey, USA) at the initial enrollment visit. Acclimation was achieved by first placing the dog on the treadmill while it was not moving and encouraging the dog with treats. The dog then remained on the treadmill while it was started at a low speed (1.0 m/s) for a few minutes, then slowed and stopped, and the dog was removed from the treadmill. The process was repeated 3 to 4 times over the course of 30 min, with increasing treadmill speeds each time (working up to 2.0 m/s). This acclimation was typically achieved by an assistant providing treats at the front of the treadmill, while another person maintained the dog on the treadmill. All dogs enrolled in the study were successfully acclimated to the treadmill during the initial enrollment visit (23,24). The dogs then returned for 2 separate activity sessions, at least 1 d apart. The same treadmill, pedometer, and accelerometer models were used at each testing site throughout the study.

The single collar that held both the accelerometer and the pedom-eter around each dog’s neck was put on and removed before and after each session. Each dog completed 3 activities during an activ-ity session day, consisting of walking/trotting on a treadmill at 3 different speeds (1.0 m/s, 1.5 m/s, and 2.0 m/s) for a total distance of 1 km at each speed. The dogs were given a 30-minute rest period between each activity session. The same order of the accelerometer and pedometer on the collars, as well as the same order of events, were repeated for each dog during both sessions. Data were collected from both the accelerometer (step counts and activity counts) and the pedometer (step counts).

Statistical analysisIn this exploratory study, our statistical approach involved 2 parts.

The first part was to understand the statistical and mathematical relationships between the independent variables (patient morpho-metrics, treadmill speed) and dependent variables (accelerometer or pedometer output) in 2 breeds of normal dogs. The second part was to develop a formula to predict the distance traveled by a dog based on its independent and dependent variables.

For the first part, we investigated a number of multiple regres-sion models (including correlations, linear models, and higher-order linear models, e.g., quadratic models), transformations, subset analyses, and plots, with the goal of discovering how independent and dependent variables were related. When the more complicated model was not statistically significant, the results were reported as correlation coefficients. Additional initial statistical investigations included random effects models (dog, session day) to test the effect of day and study site on the dependent variables. Statistical signifi-cance was set at P , 0.05.

For the second part of the analysis, we developed an estima-tion equation for the distance traveled by a dog using the simple formula: (number of steps) 3 (stride length). The number of steps came from the accelerometer or pedometer, but the stride length was estimated from morphometric data using a regression equa-tion. The dependent variable (stride length) and the independent variables (morphometric data) were used to develop the regression equation. When distance traveled is known, a dog’s average stride length can be estimated with (distance traveled) ÷ (number of steps taken). Therefore, using data from session 1 with a known distance traveled, we regressed the morphometric data using that estimate

Figure 1. Mounting arrangement of the accelerometer and pedometer.



of stride length. The resulting regression equation allowed us to use morphometric data to predict the stride length for any dog within the general size range of the study dogs.

The stride length regression equation from session 1 data was then validated using session 2 data by estimating the distance each dog traveled in session 2 (using the equation from session 1) and compar-ing it to the known distance for session 2. This comparison was done independently for Omron pedometer steps, Actical accelerometer steps, and Actical accelerometer activity counts. To estimate each dog’s distance traveled, the dog’s stride length was estimated from its morphometrics only (using the equation found in session 1). The estimated stride length was then multiplied by the number of steps from the accelerometer or pedometer to yield the estimated distance traveled. Finally, the estimated distance traveled for session 2 was compared to the actual distance traveled in session 2 using descrip-tive statistics.

Note that, since treadmill speed influences accelerometer and pedometer output and stride length, the regression equation for stride length in session 1 used the total distance traveled (3.0 km) and combined all 3 velocities (1.0 m/s, 1.5 m/s, and 2.0 m/s). We elected to use all 3 velocities together because we thought it was reasonable to assume that a dog in its natural environment would carry out activities with various levels of rigor. Also note that the final formulas were generated from multiple breeds, velocities, days, and institutions. We elected this format because we thought it best represented data that might come from a clinical population. The final formula reports a value in meters/step, which allows the simple calculation of meters, since steps is reported by the accelerometer and pedometer.

Re s u l t sTwenty dogs were enrolled in the study, including 10 beagles

and 10 Labrador retrievers. The average age (4.86 y, 2 to 8.5 y) and BCS (5.2/9, 4 to 6/9) of the beagles did not differ significantly from

the average age (4.9 y, 1.4 to 9.3 y) and BCS (4.8/9, 4 to 6/9) of the Labrador retrievers. The mean body weight and morphometric measurements were significantly different between breeds. The mean body weight for the beagles was 11.34 kg (7.9 to 14.5 kg) while the mean body weight of the Labrador retrievers was 28.32 kg (21.1 to 40 kg). The mean shoulder height for the beagles was 26.63 cm (21 to 32 cm) and mean shoulder height for the Labrador retrievers was 44.28 cm (32.8 to 53.5 cm). All enrolled dogs completed the study without complication.

Early inspection, e.g., correlation coefficients, of the dependent variables (patient morphometrics and treadmill speed) identified a univariate relationship with the accelerometer and pedometer out-put. Regardless of the activity, shoulder height had slightly greater correlations with accelerometer and pedometer outputs than other morphometric variables, e.g., body weight and standing hypotenuse and all correlation coefficients had a P , 0.01. Across all dogs and all activities, the correlation coefficient (R2) between shoulder height and pedometer steps was 0.47, for accelerometer steps was 0.41, and for accelerometer activity counts was 0.38. The correlation between body weight and pedometer steps was 0.41, for accelerometer steps was 0.35, and for accelerometer activity counts was 0.31. The cor-relation between the standing hypotenuse and pedometer steps was 0.42, for accelerometer steps was 0.31, and for accelerometer activity counts was 0.31.

A relationship was also found between treadmill speed and accelerometer/pedometer outputs. In general, as the treadmill speed increased, both pedometer and accelerometer step counts decreased, while accelerometer activity counts increased (Figures 2 to 4).

When variation between session days was evaluated, no sig-nificant differences were found, with the exception of accelerometer activity counts at the fastest treadmill speed (P = 0.03). No differ-ences were found between institutional sites.

A mathematical relationship between the morphometric variables and data from the accelerometer and pedometer was investigated in order to develop a formula to estimate the total distance traveled.

Figure 2. The relationship between pedometer steps and treadmill speed for all dogs is shown. For this box plot, the median (middle line), 1st and 3rd quartiles, and minimum and maximum values are repre-sented. Each activity at a different speed was carried out for a distance of 1 km. Mean pedometer steps decreased with increased speed for the distance traveled.

Treadmill speed (m/s)

Om

ron

pedo

met

er s

teps

3000

2500

2000

1500

1000

500

0 1.0 1.5 2.0

Figure 3. The relationship between accelerometer steps and treadmill speed for all dogs is shown. For this box plot, the median (middle line), 1st and 3rd quartiles, and minimum and maximum values are repre-sented. Each activity at a different speed was carried out for a distance of 1 km.


Act

ical

acc

eler

omet

er s

teps

2500

2000

1500

1000

500

0 1.0 1.5 2.0



Shoulder height was the only morphometric variable that signifi-cantly impacted the formula output. The formula derived for the pedometer was (0.33769340 1 0.0112636838 3 Shoulder Height) 3 pedometer steps, with the results in meters per step. For this study population, the pedometer equation reported an average distance traveled of 6258 m [95% confidence interval (CI): 5821 to 6695 m], while the dogs actually traveled 6000 m (Figure 5).

When this same model was investigated to estimate the total steps taken using the accelerometer step count, the equation reported a traveled average of 5748 m (95% CI: 4839 to 6657 m). Accelerometer activity count output did not reveal a predictive formula that was meaningful (confidence intervals were greater than the mean).

D i s c u s s i o nIn this study of normal beagles and Labrador retrievers, we found

that step counts collected from the tested pedometer or accelerometer provided a reasonable estimate of the distance traveled by dogs on a treadmill. We therefore reject our null hypothesis that there is no relationship between independent variables (morphometrics and treadmill speed) and dependent variables (accelerometer and pedometer output). While the errors in the estimated distance traveled using step counts from the pedometer, which were over-estimated by 4.3%, and from the accelerometer, which were under-estimated by 4.2%, were similar, the 95% CI from the pedometer (6 437 m) was much narrower than the 95% CI of the accelerometer-predicted distance (6 909 m).

Morphometrics (shoulder height, standing hypotenuse, and body weight) were inversely but significantly related to all accelerometer and pedometer outputs. This finding supports those of a previous report that identified stride frequency as inversely related to animal body weight (25). We measured several geometric proportions and body weight on the dogs enrolled in an effort to identify their cor-

relation with accelerometer and pedometer output. We investigated the standing hypotenuse because we thought a measurement of both patient height and length might heavily influence the stride length, which would in turn influence the step count. It is not completely understood why shoulder height was slightly more predictive. Possible reasons may be that there was not enough variation in patient conformation in our study population for other morpho-metric aspects to become important or that the other morphometric variables were simply not as important as patient shoulder height. While a correlation between morphometric variables and acceler-ometer/pedometer output has been identified in a previous study (26), this study revealed the importance of accounting for shoulder height when estimating distance traveled.

A relationship was also found between treadmill speed and accelerometer/pedometer outputs. In general, as speed increased, both pedometer and accelerometer steps decreased, while acceler-ometer activity counts increased. The authors attribute this inverse difference in output from the different monitors to the fewer steps required per distance traveled with increased treadmill speed, while the accelerometer continued to pick up constant changes in accelera-tion of the subject.

There is a need for improved outcome measures for dogs with diseases that affect activity at home, such as osteoarthritis (OA), in order to accurately assess interventions. For OA, a more objective evaluation of a treatment response could be obtained by using force platform gait analysis (1,27). While gait analysis is objective and can accurately measure limb function, its evaluation is intermittent, it is not widely available, it is best used for patients with only a single affected limb, and the interpretation and reporting of results have varied in the veterinary literature. While these limitations could be overcome for studies involving groups of patients in a regula-tory study, they remain obstacles for the widespread evaluation of individual dogs. Accelerometers and pedometers potentially

Figure 5. The relationship between actual (6000 m) and predicted distance traveled for the pedometer and accelerometer step counts is shown. For this box plot, the median (middle line), 1st and 3rd quar-tiles, and minimum and maximum values are represented. Data from all 3 speeds and both replicates (activity session data) of each test were combined. Comparing the box plots reveals that the pedometer was more accurate and precise than the accelerometer.

Omron pedometer Actical accelerometer

Pre

dict

ed d

ista

nce

trav

eled

(m)

8000

7000

6000

5000

4000

3000

2000

1000

0

Figure 4. The relationship between accelerometer activity counts and treadmill speed for all dogs is shown. For this box plot, the median (middle line), 1st and 3rd quartiles, and minimum and maximum values are represented. Each test at a different speed was carried out for a distance of 1 km. Mean accelerometer activity counts increased with increased treadmill speed.


Act

ical

acc

eler

omet

er a

ctiv

ity c

ount

s

80 000

70 000

60 000

50 000

40 000

30 000

20 000

10 000

0 1.0 1.5 2.0



provide insight into the activity and behavior of veterinary patients at home (28). These devices are small and lightweight, can be added to an existing collar, and can hold a charge for long periods of time without changing or recharging the battery, allowing data to be collected for longer periods of time at home in the patients’ natural environments.

Several limitations of the most widely studied accelerometer have previously been reported (7,9). One limitation is that the accuracy and precision of activity counts over longer distances have not been reported. Hansen et al (7) reported the relationship between Actical activity counts and short distances traveled and the data showed a clear dispersion of the data points when dogs walked more than 100 m. While this is an important finding, as many dogs likely travel several kilometers in a day, that study reported only activity counts, not step counts (7). This distinction is relevant given the finding from our study that accelerometer step counts could reasonably estimate the distance traveled, while activity counts did not.

The Actical accelerometer provides an output in activity counts. Step counts is an output that can be selected and is created by attempting to filter the signal and pick out occurrences that may resemble a step. It is important to note that, to the best of our knowl-edge, the filters used are not available for review and have not been validated for use in animals. The Actical accelerometer was not designed to detect distance traveled, which may explain why the Actical step counts varied from the pedometer step counts.

Neither the accelerometer nor the pedometer studied quantifies the intensity of the movement or magnitude of the acceleration. For example, many accelerometers have the capacity to measure and report the magnitude of acceleration (positive and negative) in m/s2 in the x-, y-, and z-planes. For our patients, while it is advanta-geous to know whether the distance traveled has increased, it might be just as important to identify whether the dog has increased the frequency of jumping up and down or rapidly accelerating and stopping. These more intense physical activities would involve a greater magnitude of acceleration than walking and may provide useful information about the comfort of the patient. This option was not available for the accelerometer and pedometer evaluated in this study. While the exact details were not made available by the manufacturer, it appears that the accelerometer has a preset accelera-tion threshold to detect and report “activity counts.” Importantly, proper acceleration and the number of activity counts over a period of time measure different behaviors. A previous study described the cut-off points for activity counts to estimate activity intensity (11). While we found a correlation between activity intensity (defined by the differing treadmill speeds) and activity counts, it can be seen in Figures 2 to 4 that pedometer step counts varied the least, while accelerometer activity counts varied the most. This difference may indicate that pedometer step counts would provide a better estimate of the intensity of activity.

Understanding the relationship between data from an acceler-ometer and/or pedometer and the distance traveled by a canine patient in its home environment is clinically relevant when accel-erometers and/or pedometers are used to evaluate treatment out-comes. For example, a dog may register 100 000 counts/day before an intervention and 120 000 counts/day after an intervention. To fully understand the clinical effect of the intervention, however, we

must know whether the 20 000 counts/day difference is equivalent to 50 m or 5000 m. The ability to estimate distance traveled would allow veterinarians to prescribe and/or monitor a distance traveled each day for a pet. Measuring the distance traveled could provide a clinically relevant, objective outcome measure for clinical studies where increased or decreased patient activity at home is important.

Accelerometers and pedometers have been used in humans to monitor patient activity or recovery at home from certain diseases (18,20–22,29). People can often input data such as height, body weight, and stride length into the accelerometer or pedometer and it reports an estimated distance traveled. The ability to incorporate this information may greatly improve the usefulness of accelerometers and/or pedometers in veterinary medicine. There are limitations, however, when translating how accelerometers and pedometers work in bipeds to their performance in quadrupeds. While humans always have a predictable step pattern or gait (left-right-left-right), dogs do not always follow a predictable step pattern, with different foot-fall patterns in different gaits. Depending on the species and gait velocity, the type of gait may vary from a walk, trot, pace, gal-lop, etc., in quadrupeds. This inconsistent and unpredictable gait likely has some influence on the formula derived based on stride distance, although this could not be easily identified and controlled for in the present study.

It is likely that individual dogs adapt the length of their stride differently to varying velocities, as has been shown in humans (30). Deriving a function that translates the distance traveled based on activity will therefore always be somewhat variable. Changes in the average dogs’ pedometer steps followed a reasonably predictable mathematical model, however, which may allow a clinical inter-pretation of accelerometer and pedometer data when a dog is in its natural home environment. That said, the ability to translate our findings in a controlled setting to pets moving in their natural home environment with varying speeds and gaits on the ground, rather than on a treadmill, needs to be evaluated (31). This information must be used carefully when interpreting a therapeutic interven-tion. The intervention must have a biological explanation to alter the symptoms associated with the disease beyond simply stimulating the patient. For example, in a dog with symptoms of OA, a drug such as caffeine could induce increased activity, but the drug has no known predictable analgesic activity.

The information reported in this study is based on evaluating the effect of morphometry and velocity on activity/step outputs from both the Omron pedometer and the Actical accelerometer. It is important to note that each monitoring system should be assessed independently for interactions between patient/subject morphom-etry data and velocity/distance traveled and this formula should not be applied to another system without doing similar assessments.

In conclusion, we identified and developed a mathematical model to accurately estimate the distance traveled based on the patient’s shoulder height and step counts from an accelerometer and a pedom-eter in a population of normal beagles and Labrador retrievers. We could not, however, devise a useful mathematical model using activity counts. For several reasons, these data must be translated carefully to a clinical trial of dogs with OA pain or pain of a differ-ent etiology. First, the mathematical model should be tested in a less controlled environment with more patient heterogenicity. Second,



while it may be assumed that normal dogs travel further, i.e., take more steps, than arthritic dogs, the total number of steps between the 2 populations could be similar. If this is the case, the pattern of behavior of the populations could be different, e.g., more jumping documented in patients with less pain by measuring rate of accel-eration, which could make aspects of the accelerometer more useful than simply for steps taken. Additionally, we do not know how the location (focal or multifocal joints involved) and severity of the OA will influence patient activity. Further testing outside of a laboratory setting is warranted to determine whether this mathematical model can accurately distinguish between various populations, e.g., normal and in pain, steps taken by normal and clinically affected arthritic dogs, and the ability of this mathematical model to accurately distin-guish between these 2 populations. Only then can this body of work potentially be clinically relevant for evaluating affected patients and their responses to various medical therapies.

A c k n o w l e d g m e n tThe authors thank the staff at the University of Minnesota Clinical

Investigation Center for supporting this project.

Re f e r e n c e s 1. Hanson PD, Brooks KC, Case J, et al. Efficacy and safety of

firocoxib in the management of canine osteoarthritis under field conditions. Vet Ther 2006;7:127–140.

2. Rausch-Derra L, Huebner M, Wofford J, Rhodes L. A prospective, randomized, masked, placebo-controlled multisite clinical study of Grapiprant, an EP4 prostaglandin receptor antagonist (PRA), in dogs with osteoarthritis. J Vet Intern Med 2016;30:756–763.

3. Brown DC, Boston RC, Farrar JT. Use of an activity monitor to detect response to treatment in dogs with osteoarthritis. J Am Vet Med Assoc 2010;237:66–70.

4. Knazovicky D, Tomas A, Motsinger-Reif A, Lascelles BD. Initial evaluation of nighttime restlessness in a naturally occurring canine model of osteoarthritis pain. PeerJ 2015;3:e772.

5. Lascelles BD, Knazovicky D, Case B, et al. A canine-specific anti-nerve growth factor antibody alleviates pain and improves mobility and function in dogs with degenerative joint disease-associated pain. BMC Vet Res 2015;11:101.

6. Wernham BG, Trumpatori B, Hash J, et al. Dose reduction of meloxicam in dogs with osteoarthritis-associated pain and impaired mobility. J Vet Intern Med 2011;25:1298–1305.

7. Hansen BD, Lascelles BD, Keene BW, Adams AK, Thomson AE. Evaluation of an accelerometer for at-home monitoring of spontaneous activity in dogs. Am J Vet Res 2007;68:468–475.

8. Yashari JM, Duncan CG, Duerr FM. Evaluation of a novel canine activity monitor for at-home physical activity analysis. BMC Vet Res 2015;11:146.

9. Brown DC, Michel KE, Love M, Dow C. Evaluation of the effect of signalment and body conformation on activity monitoring in companion dogs. Am J Vet Res 2010;71:322–325.

10. Dow C, Michel KE, Love M, Brown DC. Evaluation of optimal sampling interval for activity monitoring in companion dogs. Am J Vet Res 2009;70:444–448.

11. Michel KE, Brown DC. Determination and application of cut points for accelerometer-based activity counts of activities with differing intensity in pet dogs. Am J Vet Res 2011;72:866–870.

12. Lascelles BD, Hansen BD, Thomson A, Pierce CC, Boland E, Smith ES. Evaluation of a digitally integrated accelerometer-based activity monitor for the measurement of activity in cats. Vet Anaesth Analg 2008;35:173–183.

13. Schwab-Richards R, Prost C, Steffan J, Seewald W, Nenci C, Roosje P. Use of activity monitors for assessment of pruritus in an acute model of canine atopic dermatitis. Vet Dermatol 2014;25:441–e69.

14. Nuttall T, McEwan N. Objective measurement of pruritus in dogs: A preliminary study using activity monitors. Vet Dermatol 2006;17:348–351.

15. Holbrook EA, Barreira TV, Kang M. Validity and reliability of Omron pedometers for prescribed and self-paced walking. Med Sci Sports Exerc 2009;41:670–674.

16. Saunders TJ, Gray CE, Borghese MM, et al. Validity of SC-StepRx pedometer-derived moderate and vigorous physical activity dur-ing treadmill walking and running in a heterogeneous sample of children and youth. BMC Public Health 2014;14:519.

17. Montoye AH, Pfeiffer KA, Suton D, Trost SG. Evaluating the responsiveness of accelerometry to detect change in physical activity. Meas Phys Educ Exerc Sci 2014;18:273–285.

18. Benzo R. Activity monitoring in chronic obstructive pulmonary disease. J Cardiopulm Rehabil Prev 2009;29:341–347.

19. Prieto-Centurion V, Bracken N, Norwick L, et al. Can com-mercially available pedometers be used for physical activity monitoring in patients with COPD following exacerbations? Chronic Obstr Pulm Dis 2016;3:636–642.

20. Alahmari AD, Kowlessar BS, Patel ARC, et al. P44 Validation of a pedometer to measure daily physical activities in COPD patients. Thorax 2011;66:A86.

21. Freedson P, Pober D, Janz KF. Calibration of accelerometer out-put for children. Med Sci Sports Exerc 2005;37:S523–530.

22. Freedson PS, Miller K. Objective monitoring of physical activ-ity using motion sensors and heart rate. Res Q Exerc Sport 2000;71:S21–29.

23. Gustas P, Pettersson K, Honkavaara S, Lagerstedt AS, Byström A. Kinematic and temporospatial assessment of habituation of Labrador retrievers to treadmill trotting. Vet J 2013;198:e114–119.

24. Fanchon L, Grandjean D. Habituation of healthy dogs to tread-mill trotting: Repeatability assessment of vertical ground reac-tion force. Res Vet Sci 2009;87:135–139.

25. Heglund NC, Taylor CR, McMahon TA. Scaling stride fre-quency and gait to animal size: Mice to horses. Science 1974;186: 1112–1113.

26. Griffin TM, Main RP, Farley CT. Biomechanics of quadrupe-dal walking: How do four-legged animals achieve inverted pendulum-like movements? J Exp Biol 2004;207:3545–3558.

27. Budsberg SC, Johnston SA, Schwarz PD, DeCamp CE, Claxton R. Efficacy of etodolac for the treatment of osteoarthritis of the hip joints in dogs. J Am Vet Med Assoc 1999;214:206–210.

28. Chan CB, Spierenburg M, Ihle SL, Tudor-Locke C. Use of pedom-eters to measure physical activity in dogs. J Am Vet Med Assoc 2005;226:2010–2015.



29. Kambas A, Venetsanou F, Avloniti A, et al. Pedometer deter-mined physical activity and obesity prevalence of Greek children aged 4–6 years. Ann Hum Biol 2015;42:231–236.

30. Salo AI, Bezodis IN, Batterham AM, Kerwin DG. Elite sprinting: Are athletes individually step-frequency or step-length reliant? Med Sci Sports Exerc 2011;43:1055–1062.

31. Torres BT, Moëns NM, Al-Nadaf S, Reynolds LR, Fu YC, Budsberg SC. Comparison of overground and treadmill-based gaits of dogs. Am J Vet Res 2013;74:535–541.


Article


I n t r o d u c t i o nMaturation of the immune system and age-dependent differences

in basic peripheral blood lymphocyte subpopulations in humans and animals have been widely discussed in the literature (1–11). Owing to anatomical and physiological similarities shared with humans, pigs are promoted as a model for biomedical studies. Piglets constitute a unique material to study the development of the postnatal immune response, especially in antigen-free environments (12–15). Unique aspects of T-cell subset diversity in pigs are also worth highlight-ing. Although some phenomena, like the remarkable percentage of peripheral double-positive CD41CD81 T-lymphocytes and the large proportion of gd T-cells have been widely described, few data concerning changes in these parameters according to age are avail-able (10,16–18).

The specific identification of lymphocyte subpopulations allows for a more precise and in-depth analysis of cellular immune responses to infections (e.g., porcine reproductive and respiratory

syndrome, swine influenza virus, porcine circovirus, Actinobacillus pleuropneumoniae) or environmental factors (e.g., gut microflora changes, stress caused by excessive concentration). Currently, phe-notypic analysis of leukocytes is improved by the development of monoclonal antibodies (mAb) against well-established surface mol-ecules called clusters of differentiation. Owing to these markers, it is possible to identify the cell type, its activity, and the differentiation stage (12,19). A number of factors must be taken into consideration as potentially associated with changes in lymphocyte subsets, includ-ing antigenic stimulation, hormonal changes, or immune system maturation (8,20–30). Thus, it is essential to expand the analysis of immune blood parameters in pigs at different ages, particularly mature ones, since reports to date mainly describe changes occurring in the neonatal period (1,8,16,31–33).

The aim of this study was to characterize age-dependent changes in the porcine immune system by determining relative counts of porcine peripheral blood lymphocyte subsets through the analysis of phenotypic expression of surface antigens.

Differences in the relative counts of peripheral blood lymphocyte subsets in various age groups of pigs

Olga Pietrasina, Julia Miller, Anna Rzasa

A b s t r a c tThe aim of the study was to determine age-related changes in peripheral blood lymphocytes in pigs. Previous studies looking at age-related differences in lymphocyte subsets in porcine blood have not established reference ranges for these parameters. Moreover, most studies have concentrated on the dynamic changes in the first months of life, failing to continue observations in older animals. Therefore, in the present study, relative counts of various lymphocyte subpopulations (cytotoxic and helper T-cells, B-cells, and gd T-cells) were evaluated to characterize the development of the cellular immune system at 28, 35, 135, and 200 days of age in growing pigs and adult sows (i.e., first and subsequent parity). In all examined groups, CD31 cells constituted the largest percentage of cells. A statistically significant higher percentage of TCRgd1CD31 was noted in fatteners and gilts in comparison to other age groups. These results may be a reflection of antigenic pressure and show an immune response to viral or bacterial agents/environmental microbism.

R é s u m éL’objectif de la présente étude était de déterminer les changements associés à l’âge dans les lymphocytes du sang périphérique chez les porcs. Des études antérieures examinant les différences associées à l’âge dans les sous-populations de lymphocytes dans le sang porcin n’ont pas établi des écarts de référence pour ces paramètres. De plus, la plupart des études se sont concentrées sur les changements dynamiques dans les premiers mois de vie, omettant de continuer les observations chez les animaux plus âgés. Ainsi, dans la présente étude, les dénombrements relatifs des différentes sous-populations lymphocytaires (cellules-T cytotoxiques et helper, cellules-B, et cellules-T gd) furent évalués afin de caractériser le développement du système immunitaire cellulaire à 28, 35, 135, et 200 jours d’âge chez des porcs en croissance et des truies adultes (i.e. première parité ainsi que les suivantes). Dans tous les groupes examinés, les cellules CD31 constituaient le pourcentage le plus élevé de cellules. Un pourcentage significativement plus élevé de TCRgd1CD31 était noté chez les porcs en croissance et les cochettes comparativement aux autres groupes d’âge. Ces résultats pourraient être un reflet de la pression antigénique et montre une réponse immunitaire à des agents viraux ou bactériens du microbisme environnemental.


Department of Immunology, Pathophysiology and Veterinary Preventive Medicine, Wroclaw University of Environmental and Life Sciences, Norwida 25, 50-375 Wroclaw, Poland.

Address all correspondence to Dr. Anna Rzasa; telephone: 71-320-5240; e-mail: [email protected]

Received August 16, 2018. Accepted December 2, 2018.




AnimalsForty-seven commercially farmed crossbred pigs (Polish Landrace

3 Polish Large White) were included in the study. Blood samples were collected from pigs and divided into age groups as follows: i) piglets (28 d; n = 6); ii) weaner piglets (35 d; n = 6); iii) fatteners (135 d; n = 9); iv) gilts (200 d; n = 10); v) primiparous lactating sows (n = 7); and vi) multiparous lactating sows (n = 9). The animals were subject to routine care and veterinary treatment at the farm.

HousingDuring the lactation period (28 d), piglets were housed with sows

in traditional slatted farrowing crates. After weaning, nursery pigs and later fatteners were housed in concrete grate pens in 15- and 30-head groups, respectively. Breeding gilts were housed in 12-head group pens on solid floors with straw. After weaning, sows were moved to individual pens. Four weeks later (after confirming insemi-nation efficiency), they were moved to 12-head group pens where they stayed until day 108 of pregnancy, when they were moved back to farrowing pens.

Prophylactic proceduresHalf a milliliter of penicillin with dihydrostreptomycin (Shotapen

L.A.; Virbac, Carros, France) was administered metaphylactically to piglets on the first day of life during routine procedures, such as teeth clipping and tail cauterization. One milliliter of iron dextran (Ferran 200; Vet-Agro, Lublin, Poland) was injected twice on day 5 (during castration) and 1 wk later.

Piglets were vaccinated against Mycoplasma hyopneumoniae at 7 d and again at 28 d (weaning), with the M1PAC vaccine (Intervet International BV, Boxmeer, The Netherlands). On weaning day, pro-phylactic vaccination against porcine circovirus-2 (PCV-2) (Porcilis PCV; Intervet International BV, Boxmeer, The Netherlands) was carried out. Gilts were vaccinated 8 wk antepartum with Porcilis AR-T DF and Porcilis Porcoli Diluvac Forte (Intervet International BV, Boxmeer, The Netherlands) (progressive atrophic rhinitis and Escherichia coli prophylaxis, respectively) and 4 wk later, sows and gilts were vaccinated with Ingelvac CircoFLEX (Boehringer Ingelheim Vetmedica GmbH; Ingelheim/Rhein; Germany) and Respiporc FLU3 (IDT Biologika GmbH; Dessau-Rosslau, Germany) (PCV-2 and influenza prophylaxis, respectively). One week later, parturition vaccination with Parvoruvax (against porcine parvovi-rus and Erysipelothrix rhusiopathiae) (Ceva Santé Animale, Libourne, France) and Ingelvac CircoFLEX (Boehringer Ingelheim Vetmedica GmbH, Germany), was performed. Antiparasitic treatment was car-ried out by administration of ivermectin (Ivomec Premix for Swine 0.6%, Boehringer Ingelheim Vetmedica GmbH, Germany) to gilts and sows in feed 1 wk before moving to the farrowing sector.

According to the II Local Ethics Committee at the University of Environmental and Life Sciences, Wroclaw, Poland, the present study did not require review and approval by an ethics committee.

Blood samplingBlood samples were collected via venipuncture of the vena cava

cranialis in 15 mL tubes containing heparin as an anticoagulant. In each age group, blood was collected from 10 individuals; however, laboratory analysis was not performed in samples with a visible clot formation, which explains the different numbers in groups.

Flow cytometryPhenotypic analysis of peripheral blood lymphocytes was per-

formed in whole blood by flow cytometry using fluorochrome-labelled monoclonal antibodies (mAbs). Twenty-five mL of whole blood was mixed with 25 mL of PBS and incubated with mAb against surface molecules. The cells were double-stained for CD3/CD21 and CD4/CD8 using the following mAbs: mouse anti-porcine CD3-FITC (catalogue no. 4510-02; SouthernBiotech, Birmingham, Alabama, USA), mouse anti-porcine CD21-PE (catalogue no. 4530-09; SouthernBiotech), APC rat anti-pig gd T-lymphocytes clone MAC320 (catalogue no. 561482; BD Biosciences, San Jose, California, USA),

Figure 1. Gating strategy for the evaluation of basic lymphocyte sub-populations (CD31, CD211, and T gd1 cells). The frequency of cells is expressed as a percentage of positive cells in a lymphocyte gate.

Figure 2. Gating strategy for the evaluation of T-helper and T-cytotoxic cells. The frequency of cells is expressed as a percentage of positive cells in gated CD31 cells.



Alexa Fluor 647 mouse anti-pig CD4a clone 74-12-4 (catalogue no. 561472; BD Biosciences), and PE mouse anti-pig CD8a clone 76-2-11 (catalogue no. 559584; BD Biosciences). After incubation (15 min at room temperature), erythrocytes were lysed with 1 mL of High Yield Lyse buffer (HYL-250; Invitrogen, Carlsbad, California, USA) and the samples were analyzed after 15 min using a FACSCalibur (BD Biosciences). On the basis of the forward- and side-scatter dot plot, a peripheral blood mononuclear cells (PBMC) gate was set (Figure 1) and the percentage of subpopulations was evaluated by Weasel 2.0 software (Weasel Software Oy, Turku, Varsinais-Suomi, Finland). The results were expressed as percentages of the gated PBMCs.

Figure 1 shows the gating strategy used in the experiment and representative dot plots of CD21, CD3 for the detection of T-cells (CD31CD212), and CD211 B-cells. The proportions of ab T-cells and gd T-cells were determined as the percentages of CD31TCRgd and CD31TCRgd1 gated lymphocytes, respectively. Then, to determine the occurrence of changes in specific T-cell populations, anti-CD3, -CD4, and -CD8alpha antibodies were used to detect CD41 (helper cells), CD81 (cytotoxic cells), and CD41CD81 (memory T-cells) (33). The gating strategy is presented in Figure 2.

Statistical analysisTreatment comparisons were made using the analysis of vari-

ance (ANOVA) procedures for completely randomized design in STATISTICA 12.5 (Stat Soft, Tulsa, Oklahoma, USA). The differences between the means were determined by the post-hoc Tukey test for different numbers at 0.05 and 0.01 levels. The Pearson correlation between analyzed parameters was calculated. The results are pre-sented as a mean value and standard deviation (mean 6 SD) and correlation coefficient (r).

Re s u l t sThe results of flow cytometric analyses, including signifi-

cant differences between age groups, are summarized in Table I. Additionally, T-cell subset changes are presented in Figure 4. Relative numbers of B (CD211CD32) and T-lymphocytes (CD212CD31) varied in different groups during pigs’ lives and some statistically significant differences were noted. CD31-lymphocytes constituted the largest population of cells, with the highest result among all

groups noted in gilts around 200 d of life (group 4). Starting from the weaning day (day 28), a decreasing tendency in the relative number of B-lymphocytes (CD211) was observed, with the lowest frequency noted in group 4. In older pigs, the percentage increased slightly and reached a similar level in both sow groups (5 and 6). Statistically significant differences (P # 0.05) were noted between piglet groups (28 and 35 d), weaned piglet and sow groups, and fattener and gilt groups. Also, the gd T-cell subpopulation analysis showed some statistical differences between examined groups. Relative counts of gd1 T-cells were lowest in weaning piglets, the result being approximately fourfold lower in comparison with the highest values that occurred in gilts and fatteners (Figure 3). The average noted in fatteners and gilts varied significantly (P # 0.01) from the rest of the results obtained in the study.

Additionally, we assessed whether T-cells express CD4, CD8, or both coreceptors (double-positive CD41CD81 cells). The highest rela-tive values of CD41 cells were observed in piglet groups (1 and 2) around weaning; results differed significantly from those obtained in other groups (P # 0.01). Statistically significant differences were confirmed in CD42CD81 percentages between groups of fatteners and primiparous sows (P # 0.05). The highest average values of double-positive CD41CD81 lymphocytes were observed in 2 groups of sows and the difference between primiparous and weaned piglets turned out to be statistically significant (P # 0.05).

Representative dot plots showing CD4 and/or CD8 expression in model samples obtained from animals in different age groups are presented in Figure 5. In the study herein, the CD41:CD81 ratio exceeded 1 in both piglet groups. The lowest ratio was noted in gilts and primiparous sows. The analysis of CD41:CD81 ratios showed statistically significant differences between weaned piglets and both gilts and primiparous sows (P # 0.01), as well as between weaned piglets and fatteners (P # 0.05).

In terms of the correlation coefficient, a significant interaction with age was found for CD211CD32, CD212CD31, CD41CD82, and CD41CD81 subpopulations as well as for the CD41:CD81 ratio (Table II).

D i s c u s s i o nStudies assessing cellular immune status changes in clinically

healthy pigs belonging to various production groups, such as the

Table I. Percentage of lymphocyte subpopulations in blood.

Age Lymphocyte subpopulationsgroup n CD211CD32 CD212CD31 TCR1CD31 CD41CD82 CD41CD81 CD42CD81 CD4:CD81 6 23.48A,a 6 3.83 59.75B,b 6 5.30 8.98B 6 3.02 32.53A 6 10.77 17.73 6 6.20 33.58 6 12.77 1.19 6 0.772 6 16.33C,b,c 6 6.77 69.96a,d 6 7.61 13.30B 6 3.78 32.46A 6 8.40 14.31B,b 6 2.41 28.10 6 6.46 1.24A,a 6 0.533 9 12.69B,e 6 1.82 72.42A,f 6 3.59 32.29A 6 9.49 14.46B 6 2.64 10.63B,D 6 3.00 23.78b 6 4.63 0.62b 6 0.094 10 6.59B,D,f 6 1.87 81.35A,C,c,e 6 2.70 39.60A 6 8.25 12.39B 6 3.36 8.55B,D 6 2.55 28.49 6 4.66 0.44B 6 0.145 7 9.06B,d 6 2.12 76.67A 6 4.87 15.87B 6 7.95 15.28B 6 2.54 25.31C,a 6 8.00 35.47a 6 5.49 0.44B 6 0.116 9 9.03B,d 6 4.24 70.96D,a 6 8.76 13.27B 6 6.7 16.39B 6 4.93 28.63A 6 9.77 28.11 6 9.97 0.67 6 0.33Total 47 12.0 6 6.4 72.7 6 8.5 22.4 6 13.7 19.1 6 9.7 17.2 6 9.7 29.2 6 8.1 0.72 6 0.46Statistical differences are marked in pairs:P # 0.05 — a and b, c and d, e and f.P # 0.01 — A and B, C and D.



Figure 3 a — Representative dot plots showing the percentage of T gd1 cells in model samples obtained from animals in different age groups. b — Percentage of CD3 1 TCR gd1 lymphocytes.

a

b Percentage of CD31/TCR gd1 lymphocytes



one carried out by our team, are rare. In the study conducted by Pomorska-Mól and Markowska-Daniel (10) on pigs from birth to slaughter, CD31 lymphocytes constituted the largest percentage of cells and more than a half of them were CD42CD81. Moreover, mean percentages of CD31 cells remained stable between 47.58% and 60.93% from birth to 5 mo of age. Despite the fact that our analysis also included older animals, CD31 values from day 28 in both studies are comparable (57.92 6 6.03 in the cited study versus 59.75 6 5.3 in our trial). The results obtained at other similar time points were approximately 20% lower in comparison to our trial. According to Pomorska-Mól and Markowska-Daniel (10), percent-ages of CD41CD82 cells in peripheral blood were highest between 14 and 21 d of age and there was a significant positive correlation between the age of pigs and the mean percentages of double positive cells in blood. In the present study, the highest values were noted in groups 1 and 2. Moreover, CD41CD82 percentages noted around 140 d were very similar for both experiments. Cited authors under-line an evident positive correlation between the age of pigs and the mean percentages of double positive cells in blood, which was also confirmed in our study, whereby the highest percentages of double positive cells were noted in both sow groups. However, the results obtained in group 1 were around twofold higher than in the cited publication. Besides different timepoints of blood collection, the dif-ferences between experiments may be due to differences in animal breeds, environmental microbisms, or gating strategies. The authors of the cited study did not explain the gating strategy for the calcula-tion of the CD41 and/or CD81 cell percentage.

Peripheral double positive T-cells (TCRab1CD41CD8aa1 and TCRgd1CD21CD81) were absent or rare in fetal and germ-free pigs, while an expansion of these subsets occurred in piglets reared under conventional conditions in a previous study described by Zuckermann et al (34). The author also highlighted that all double positive T-cells could be restimulated in vitro with a recall antigen, which strongly supports the hypothesis that peripheral double posi-tive lymphocytes represent a population of memory helper cells. It might confirm our investigation, wherein a higher percentage of double positive CD41CD81 cells was observed in the oldest groups (sows). The highest percentages of CD41CD81 noted in the oldest

groups of animals are partially comparable to the results reported by Yang and Parkhouse (32). The hypothesis that extrathymic CD41CD81 T-cells increase with age and could be considered to be mature antigen-experienced memory/effector cells, was also emphasized by Pescovitz et al (12) and Talker et al (11).

Prenatal and early postnatal immune system development in minipigs was studied by Sinkora et al (23). Several lymphocyte subsets were shown to be recruited upon contact of neonatal piglets with live microflora (i.e., intestinal colonization with some E. coli strains). Studies in gnotobiotic piglets revealed that the appearance of CD41CD81 is absolutely dependent on the contact of immune system with live viruses and bacteria, respectively. Pescovitz et al (12) state that approximately 25% of the CD41 and CD81 cells in swine co-express both markers. In the study carried out by our team, this result was lower (around 20%) in younger groups of animals (1, 2, 3, and 4) and approximately twofold higher in both sow groups. Considering that this population reflects memory cells, these slightly lower values may result from genetic differences in animals or differences in the experimental design (e.g., housing, environmental pressure, etc.).

Many authors trying to establish an indicator of immune status mention the CD41:CD81 ratio (3,7,12,36–38). In our investigation, the CD41:CD81 ratio was above 1 in groups 1 and 2 of pigs and below 1 in groups 3, 4, 5, and 6. The researchers report that the expression of CD81 in swine is greater than that of CD41, thus the CD41:CD81 ratios are reversed from those reported in humans (1.5 to 2.0). These observations were also confirmed by Becker and Misfeldt (1), whose investigations consistently showed a higher per-centage of CD81 than of CD41 in the porcine peripheral blood and spleen. However, in the same paper, the authors stated that the per-centages of identified populations varied among genetically different animals. In the literature, it was also noted that the value of the ratio decreased with growing pigs’ age, owing to the immune system maturation, which is also true for our findings (17,18,25,35,36). Higher ratios in groups 1 and 2 in the present study may also be an effect of high immune stimulation during the peri-weaning period or a developmental stage of the immune system (17,36).

T-lymphocytes play a central role in the antigen-specific immune response against various pathogens. Characteristics of T-lymphocytes and their immune response against antigens were previously investigated in the context of bacterial (Actinobacillus pleu-ropneumoniae, Lawsonia intracellularis, and Brachyspira hyodysenteriae) and viral (porcine reproductive and respiratory syndrome, African swine fever, and Aujeszky’s disease) infections and vaccinations (20–23,25–28,37,38,39). Gamma delta T-lymphocytes participate in anti-infective and anti-cancer responses as well as in the regulation of the immune response. They secrete cytokines, such as IFN-g and IL-17. These cells recognize heat shock proteins, phospho-antigens, and alkylamines produced by many bacteria, as well as MICA and MICB molecules produced by virus-infected cells. The frequency of gd T-cells was shown to increase in response to infections in mice, cattle, chickens, and humans (2,40242). They are also believed to play an important role in the epithelial immunity and immune responses of young pigs (11). We suspect that a significantly higher percentage of CD31TCRgd1 observed in our study in fattener and gilt sectors might be a result of antigenic pressure and reflect the

Figure 4. Mean percentage of T-cell subsets in different age groups (mean 6 SD).

1 2 3 4 5 6Group

(%)

-



immunological response to viral or bacterial agents. It should also be taken into consideration that immune parameters may be influ-enced by sexual maturation as well. In order to underline the higher percentage of gd1 T-cells in fattener and gilt groups, representative dot plots of animals in each age group are presented in Figure 3a.

Many studies report a higher number of B-cells in children than in adults (43,44). Similar findings have also been described in other species, such as dogs (45). A decreasing number of B-cells in grow-ing animals may be associated with a lower bone marrow output, although according to Sinkora et al (29), pig lymphogenesis occurs in bone marrow throughout life, even in adulthood.

Beneficial information for the prevention of mycoplasmal pneu-monia in swine (MPS) was shown in research authored by Shimazu et al (28). In order to evaluate immunological changes in peripheral blood in pigs that were genetically selected for their improved resistance to M. hyopneumoniae, the percentages of B-cells and CD41 T-cells in total leukocytes were evaluated. The results indicate that the selection of pigs on the basis of MPS resistance affects their immunophenotype. The results obtained in our study are compa-rable to those reported by Shimazu et al (28) in the unselected swine line. In both experiments, immunological status was assessed 7 d after the second vaccination. However, in comparison to the cited study, our research noted lower CD41 and CD81 scores as well

as double positive cells 1 wk after the second vaccination. The presented results are also in opposition to those of Kick et al (46), whereby the percentage of CD41 T-cells after similar vaccinations was higher in vaccinated pigs than in unvaccinated control animals. However, the conclusion was that it might simply be due to age and rapid expansion of markers during the early postnatal period. We suspect that observed dissimilarities may arise from differences in the breeds, vaccine features (even though both were inactivated, but different adjuvants were used), and vaccination protocols (vaccina-tion in peri-weaning versus finishing period in our study and the cited one, respectively).

Studying the development and changes in cellular immune response may be crucial for the improvement of biomedical studies with respect to cross-species similarities. It seems that large-scale studies describing individual lymphocyte subpopulations may also contribute to improve swine disease control. Immaturity and/or failure of the immune system caused by a commonly occurring dis-ease can be conducive to weakness or death. Acquired knowledge may allow appropriate modulation of animals’ immune system (e.g., new generations of vaccines, serum administration, use of immu-nomodulators). The analysis of the cellular immune response could help in planning a swine vaccination calendar or suggest changes to the routine weaning time (19). Finally, considering immunological

Figure 5. Representative dot plots showing the expression of CD4 and/or CD8 in model samples obtained from animals in different age groups.



characteristics when animal breeding is feasible and may help to cut down economic losses by reducing the incidence of infection (28).

The assessment of infection-induced cellular immune status changes in pigs may be difficult because of age-related differences in leukocyte subsets in porcine peripheral blood. It must be empha-sized that relative counts of various porcine blood leukocyte subsets will vary depending on the genetic properties of animals and the epizootic status of the farms. Establishing reference intervals of mentioned parameters in pigs is important for both practitioners and researchers. It is vital to collect as much data as possible from differ-ent areas in order to suggest a proper reference range, which is why the authors are open to criticism and suggestions of the presented data set. It is important to accurately interpret laboratory results in view of the variation within individual pigs, production groups, and farms. These differences are related, though not limited to, genetics, growth rate, nutrition, health status, season, and area (47).

The results presented complement the existing knowledge and fill the gap in available data. They are useful for the identification of many health or development disorders, especially in subclinical states. Nevertheless, the relationship between lymphocyte subpopu-lations in different age groups requires further research in association with epidemiological or specific pathogen-free studies.

A c k n o w l e d g m e n tThis work was supported by the Wroclaw University of Environ-

mental Life and Sciences (project number: B030/0068/17).

Re f e r e n c e s1. Becker BA, Misfeldt ML. Evaluation of the mitogen-induced

proliferation and cell surface differentiation antigens of lympho-cytes from pigs 1 to 30 days of age. J Anim Sci 1993;71:2073–2078.

2. Rothwell L, Gramzinski RA, Rose ME, Kaiser P. Avian coccid-iosis: Changes in intestinal lymphocyte populations associated with the development of immunity of Eimeria maxima. Parasite Immunol 1995;17:525–533.

3. Comans-Bitter WM, de Groot R, van den Beemd R, et al. Immunophenotyping of blood lymphocytes in childhood. Reference values for lymphocyte subpopulations. J Pediatr 1997; 130:388–393.

4. Winnicka A, Klucinski W, Kawiak J, et al. Lymphocyte subpopu-lations, null cells and MHC II positive cells in peripheral blood of goats at different ages. Small Rumin Res 1999;33:247–253.

5. Faldyna M, Levá L, Knötigová P, Toman M. Lymphocyte sub-sets in peripheral blood of dogs — A flow cytometric study. Vet Immunol Immunopathol 2001;82:23–37.

6. Heaton PR, Blount DG, Mann SJ, et al. Assessing age-related changes in peripheral blood leukocyte phenotypes in domes-tic shorthaired cats using flow cytometry. J Nutr 2002;132: 1607–1609.

7. Toman M, Faldyna M, Knotigova P, Pokorova D, Sinkora J. Postnatal development of leukocyte subset composition and activity in dogs. Vet Immunol Immunopathol 2002;87:321–326.

8. Borghetti P, De Angelis E, Saleri R, et al. Peripheral T lympho-cyte changes in neonatal piglets: Relationship with growth hor-mone (GH), prolactin (PRL) and cortisol changes. Vet Immunol Immunopathol 2006;110:17–25.

9. Brown DC, Maxwell CV, Erf GF, Davis ME, Singh S, Johnson ZB. Ontogeny of T lymphocytes and intestinal morphological characteristics in neonatal pigs at different ages in the postnatal period. J Anim Sci 2006;84:567–578.

10. Pomorska-Mól M, Markowska-Daniel I. Age-dependent changes in relative and absolute size of lymphocyte subsets in the blood of pigs from birth to slaughter. Bull Vet Inst Pulawy 2011; 55:305–310.

11. Talker SC, Käser T, Reutner K, et al. Phenotypic maturation of porcine NK- and T-cell subsets. Dev Comp Immunol 2013;40: 51–68.

12. Pescovitz MD, Lunney JK, Sachs DH. Preparation and charac-terization of monoclonal antibodies reactive with porcine PBL. J Immunol 1984;133:134–137.

13. Piriou-Guzylack L, Salmon H. Membrane markers of the immune cells in swine: An update. Vet Res 2008;39:54.

14. Meurens F, Summerfield A, Nauwynck H, Saif L, Gerdts V. The pig: A model for human infectious diseases. Trends Microbiol 2012;20:50–57.

15. Swindle MM, Makin A, Herron AJ, Clubb FJ Jr, Frazier KS. Swine as models in biomedical research and toxicology testing. Vet Pathol 2012;49:344–356.

16. Juul-Madsen HR, Jensen KH, Nielsen J, Damgaard BM. Ontogeny and characterization of blood leukocyte subsets and serum proteins in piglets before and after weaning. Vet Immunol Immunopathol 2010;133:95–108.

17. Overgaard NH, Jung JW, Steptoe RJ, Wells JW. CD41/CD81 double-positive T cells: More than just a developmental stage? J Leukoc Biol 2015;97:31–38.

18. Saalmüller A, Reddehase MJ, Bühring H, Jonjic S, Koszinowski UH. Simultaneous expression of CD4 and CD8 antigens by substantial proportion of resting porcine T lymphocytes. Eur J Immunol 1987;17:1297–1301.

19. Jonasson R, Johannisson A, Jacobson M, Fellström C, Jensen-Waern M. Differences in lymphocyte subpopulations and cell counts before and after experimentally induced swine dysentery. J Med Microbiol 2004;53:267–272.

20. Casal I, Enhuanes L, Viñuela E. Porcine leukocyte cellular subsets sensitive to African swine fever virus in vitro. J Virol 1984;52:37–46.

Table II. Correlations between pigs’ age and analyzed lymphocyte subpopulations.

Age correlation Lymphocyte subpopulation coefficient (r)CD211CD32 20.698*CD212CD31 0.412*TCR1CD31 0.048CD41CD82 20.592*CD41CD81 0.469*CD42CD81 0.010CD4:CD8 20.471** P # 0.01.



21. Shimizu M, Yamada S, Kawashima K, Ohashi S, Shimizu S, Ogawa T. Changes of lymphocyte subpopulations in pigs infected with porcine reproductive and respiratory syndrome (PRRS) virus. Vet Immunol Immunopathol 1996;50:19–27.

22. Sinkora J, Rehakova Z, Sinkora M, Cukrowska B, Tlaskalova-Hogenova H. Early development of immune system in pigs. Vet Immunol Immunopathol 2002;87:301–306.

23. Sinkora M, Sinkorova J. B cell lymphogenesis in swine is located in the bone marrow. J Immunol 2014;193:5023–5032.

24. Saalmüller A, Pauly T, Höhlich BJ, Pfaff E. Characterization of porcine T lymphocytes and their immune response against viral antigens. J Biotechnol 1999;73:223–233.

25. Saalmüller A, Werner T, Fachinger V. T-helper cells from naive to committed. Vet Immunol Immunopathol 2002;87:137–145.

26. Díaz I, Darwich L, Pappaterra G, Pujols J, Mateu E. Immune responses of pigs after experimental infection with a European strain of porcine reproductive and respiratory syndrome virus. J Gen Virol 2005;86:1943–1951.

27. Cordes H, Riber U, Jensen TK, Jungersen G. Cell-mediated and humoral immune responses in pigs following primary and challenge-exposure to Lawsonia intracellularis. Vet Res 2012;43:9.

28. Shimazu T, Borjigin L, Katayama Y, et al. Immunological char-acterization of peripheral blood leukocytes using vaccine for mycoplasmal pneumonia of swine (MPS) in swine line selected for resistance to MPS. Anim Sci J 2013;84:683–692.

29. Sinkora M, Butler JE, Lager KM, Potockova H, Sinkorova J. The comparative profile of lymphoid cells and the T and B cell spec-tratype of germ-free piglets infected with viruses SIV, PRRSV or PCV2. Vet Res 2014;45:91.

30. Czyzewska-Dors E, Kwit K, Pejsak Z, Pomorska-Mól M. Changes in circulating lymphocyte subpopulations in pigs receiving therapeutic doses of ceftiofur and tulathromycin. J Vet Res 2016; 60:481–487.

31. Joling P, Bianchi AT, Kappe AL, Zwart RJ. Distribution of lymphocyte subpopulations in thymus, spleen, and peripheral blood of specific pathogen free pigs from 1 to 40 weeks of age. Vet Immunol Immunopathol 1994;40:105–117.

32. Yang H, Parkhouse RM. Differential expression of CD8 epitopes amongst porcine CD8-positive functional lymphocyte subsets. Immunology 1997;92:45–52.

33. Terzic S, Sver L, Valpotic I, et al. Immunophenotyping of leuko-cyte subsets in peripheral blood and palatine tonsils of prefat-tening pigs. Vet Res Commun 2002;26:273–283.

34. Zuckermann FA, Husmann RJ. Functional and phenotypic analysis of porcine peripheral blood CD4/CD8 double-positive T cells. Immunology 1996;87:500–512.

35. Talker M, Koinig H, Stadler M, et al. Magnitude and kinetics of multifunctional CD41 and CD8b1 T cells in pigs infected with swine influenza A virus. Vet Res 2015;46:52.

36. Fangman TJ, Tubbs RC, Becker BA, Allee GL, Misfeldt ML, Henningsen-Dyer K. Evaluation of segregated early weaning investigating performance, immunologic indicators (CD4, CD8), and herd health status. Swine Health Prod 1996;4:217–222.

37. Watson JL, Stott JL, Blanchard MT, et al. Phenotypic character-ization of lymphocyte subpopulations in horses affected with chronic obstructive pulmonary disease and in normal controls. Vet Pathol 1997;34:108–116.

38. Binns RM. The Null/gamma delta TCR1 T cell family in the pig. Vet Immunol Immunopathol 1994;43:69–77.

39. Takamatsu HH, Denyer MS, Stirling C, et al. Porcine gammad-elta T cells: Possible roles on the innate and adaptive immune responses following virus infection. Vet Immunol Immunopathol 2006;112:49–61.

40. Koets A, Rutten V, Hoek A, et al. Progressive bovine paratuber-culosis is associated with local loss of CD4(1) T cells, increased frequency of gamma delta T cells, and related changes in T-cell function. Infect Immun 2002;70:3856–3864.

41. Munk ME, Gatrill AJ, Kaufmann SH. Target cell lysis and IL-2 secretion by gamma/delta T lymphocytes after activation with bacteria. J Immunol 1990;145:2434–2439.

42. Hiromatsu K, Yoshikai Y, Matsuzaki G, et al. A protective role of gamma/delta T cells in primary infection with Listeria mono-cytogenes in mice. J Exp Med 1992;175:49–56.

43. Carver JD, Pimentel B, Wiener DA, Lowell NE, Barness LA. Infant feeding effects on flow cytometric analysis of blood. J Clin Lab Anal 1991;5:54–56.

44. Erkeller-Yuksel FM, Deneys V, Yuksel B, et al. Age-related changes in human blood lymphocyte subpopulations. J Pediatr 1992;120:216–222.

45. Greeley EH, Kealy RD, Ballam JM, Lawler DF, Segre M. The influence of age on the canine immune system. Vet Immunol Immunopathol 1996;55:1–10.

46. Kick AR, Tompkins MB, Hammer JM, Routh PA, Almond GW. Evaluation of peripheral lymphocytes after weaning and vac-cination for Mycoplasma hyopneumoniae. Res Vet Sci 2011;91: e68–e72.

47. Humann-Ziehank E, Ganter M. Pre-analytical factors affecting the results of laboratory blood analyses in farm animal veteri-nary diagnostics. Animal 2012;6:1115–1123.


Article


I n t r o d u c t i o nCompanion animals and humans are exposed to an increasing

number of environmental carcinogens that may contribute to the development of tumors (1). Early cancer detection offers the best chance for successful treatment and several cancer biomarkers con-tinue to be studied to achieve this goal (1). Many are valuable for the diagnosis, staging, and prognosis of cancers, and can also be mark-ers of progression and therapeutic response (1). Cancer biomarkers are scarcely described in veterinary medicine and available studies address dogs more than cats.

The urokinase plasminogen activator system (uPAS) is one of the extracellular matrix proteolytic systems involved in several physi-ological and pathological conditions concerning the basal membrane

and extracellular matrix (ECM) remodeling, such as angiogenesis, cancer invasion, and metastasis (1). This system comprises the serine protease urokinase plasminogen activator (uPA), its cell membrane receptor (uPAR), the tissue plasminogen activator and its 2 main inhibitors, and the plasminogen activator inhibitors PAI-1 and PAI-2 (2). The binding of uPA to its receptor triggers the conversion of plasminogen into plasmin, disintegrating the various components of the ECM and facilitating cell invasion, migration, and dissemination (2). Plasmin also promotes tumor cell invasion by converting pro-metalloproteases (pro-MMPs) to enzymatically active MMPs, as well as tumor cell proliferation by activating latent growth factors (2).

The uPA and other members of the system have been studied in a wide range of human malignancies, such as mammary, prostatic, and pancreatic cancers (2–4). Early studies demonstrate that high

Determination of urokinase-type plasminogen activator serum levels in healthy and oncologic cats

Cláudia Viegas, Augusto J. de Matos, Liliana R. Leite-Martins, Inês Viegas, Rui R. F. Ferreira, Hugo Gregório, Andreia A. Santos

A b s t r a c tThe urokinase plasminogen activator system (uPAS) has been poorly investigated in veterinary oncology. The aim of this study was to determine uPA serum concentrations in healthy and oncologic cats to understand the potential value of uPA as a cancer biomarker. Serum samples were collected from 19 healthy cats and 18 cats with spontaneous malignant neoplasms and uPA was measured through a specific enzyme-linked immunosorbent assay kit. The differences between uPA values and their relation with intrinsic factors and clinicopathological parameters were analyzed using an analysis of variance (ANOVA) and independent t-test. The average serum concentration of uPA in cancerous cats (0.54 6 0.22 ng/mL) differed from that of healthy cats (1.10 6 1.16 ng/mL) but was not significantly influenced by cats’ clinicopathological parameters or by the presence of metastases. This study describes, for the first time, the serum concentrations of uPA in cats and proposes directions for future studies to uncover the relevance of uPAS in feline carcinogenesis.

R é s u m éLe système activateur de plasminogène de type urokinase (uPAS) a été peu étudié en oncologie vétérinaire. L’objectif de la présente étude était de déterminer les concentrations sériques d’uPA chez des chats en santé et oncologiques afin de comprendre la valeur potentielle d’uPA comme marqueur de cancer. Des échantillons de sérum furent prélevés de 19 chats en santé et de 18 chats avec des néoplasmes malins spontanés et l’uPA fut mesuré à l’aide d’une trousse immuno-enzymatique. Les différences entre les valeurs d’uPA et leur relation avec des facteurs intrinsèques et des paramètres clinico-pathologiques furent analysées par analyse de variance (ANOVA) et test de t indépendant. La concentration moyenne d’uPA chez les chats avec cancer (0,54 6 0,22 ng/mL) différait de celle des chats en santé (1,10 6 1,16 ng/mL) mais n’était pas influencée de manière significative par les paramètres clinico-pathologiques des chats ou la présence de métastases. Cette étude décrit, pour la première fois, les concentrations sériques d’uPA chez les chats et propose des orientations pour des études ultérieures afin de révéler la pertinence d’uPAS dans la carcinogénèse chez les chats.


Department of Veterinary Clinics, Faculty of Veterinary Medicine, Universidade Lusófona de Humanidades e Tecnologias, Campo Grande 376, Lisbon, Portugal (Viegas C, Viegas I); Animal Blood Bank, R. de João de Deus 741, Porto, Portugal (Ferreira); Department of Veterinary Clinics of the Biomedical Sciences Institute of Abel Salazar, University of Porto, R. Jorge Viterbo Ferreira 228, Porto, Portugal (de Matos, Leite-Martins, Santos); Animal Science and Study Centre/Food and Agrarian Sciences and Technologies Institute, P. Gomes Teixeira, Porto, Portugal (de Matos, Santos); Centro Hospitalar Veterinário, Porto, Portugal (Gregório).

Address all correspondence to Dr. Andreia A. Santos; e-mail: [email protected]

Received January 22, 2019. Accepted March 12, 2019.



levels of uPA catalytic activity are correlated with larger tumor sizes, lymph node involvement, and shorter disease-free survival, com-pared to lower activity levels (2). Moreover, node-negative patients with lower levels of uPA and PAI-1 do not require systemic adjuvant treatment, while those with elevated levels of uPA and PAI-1 present a higher risk of relapse associated with poor prognosis and require adjuvant chemotherapy (5,6).

The uPA has also been studied in the serum of human cancer patients, including head and neck squamous cell carcinomas (7), as well as breast (8), ovarian (9), colorectal (10), and prostatic cancers (3). Those studies establish a correlation between the levels of uPAS members, tumor aggressiveness, and patient survival, suggesting that they play a crucial role in the cancer’s evolution, by promoting angiogenic and metastatic phenomena, making them a promis-ing therapeutic target (2).

In contrast with human medicine, uPA has rarely been investi-gated in veterinary medicine and serum uPA levels have only been described in dogs (11). In all other veterinary studies, tumor expres-sion of uPA has been investigated via immunohistochemical methods in canine mammary (12–14), prostatic (15), and vascular endothelial neoplasms (16,17), with only 1 study in cats with giant cell tumors of the bone (18). Santos et al (12,13) studied uPA expression in canine mammary neoplasms and demonstrated that a high expression by the tumor stroma is significantly associated with several poor prognostic factors, such as higher histological grade, higher Ki-67 indices, higher metastatic rate, and shorter overall and disease-free survival. Other authors obtained similar results in the same type of neoplasms (14,19), as well as in canine hemangiosarcomas (16) and prostatic proliferative disorders (15). In domestic cats, Leonardi et al (18) also demonstrated an overexpression of stromal uPAS members in a giant cell bone tumor.

Considering these findings and the scarcity of animal serum can-cer biomarkers, this study was developed to determine the concen-tration of circulating uPA in healthy cats and assess its significance as a diagnostic feline cancer biomarker. The relationships between uPA levels and clinicopathological parameters such as tumor type, tumor size, presence of inflammation, and metastasis were also analyzed.


Ethical approvalThe study protocol was approved by the Ethics Committee of

the Faculty of Veterinary Medicine at the Lusófona University of Humanities and Technologies (FMV-ULHT) (N69/2015).

Inclusion and exclusion criteriaThis prospective case-control study included a control group com-

posed of healthy cats and a case group of cats with confirmed cancer.The inclusion criteria for the control group were adult healthy

cats, aged more than 2 y without evidence of disease according to anamnesis, physical examination, and blood work results.

The inclusion criteria for cases were cats older than 2 y affected by any type of malignant neoplasm confirmed by cytology or histo-pathology. Cats with inflammatory diseases or previously subjected to oncologic treatments were excluded.

Collection of samplesSamples were obtained from 2 veterinary teaching hospitals

(FMV-ULHT and the Veterinary Hospital of the University of Porto) and 1 veterinary hospital (CHV — Veterinary Hospital in Porto), whereas most of the controls were obtained from the BSA Animal Blood Bank in Porto.

Data regarding breed, age, gender, weight, reproductive sta-tus, past medical history, clinical signs, and tumor characteristics (e.g., location, size, presence of inflammation, and adhesion to skin or adjacent tissues) were registered from each cancer patient. Cytological and histological diagnoses were obtained by the Laboratory of Clinical Analysis and Histopathology of ULHT in Lisbon, the INNO Veterinary Laboratory, and the Veterinary Pathology Laboratory of the University of Porto. All cancer cases were staged through clinical examination, 3-view thoracic radio-graphs, abdominal ultrasound and cytology, and/or biopsy of organs with suspected metastatic involvement.

Five mL of blood was collected from all controls and cases at the time of diagnosis into a serum separator tube and allowed to clot for 2 h at room temperature or overnight for an 8 h period at 2°C to 8°C and then centrifuged at 3000 rpm for 15 min. The serum was removed and divided into 0.5 mL aliquots and stored at 280°C until processing. All samples with moderate to intense hemolysis (based on visual inspection) were excluded. The serum supernatant hemoglobin was measured in all samples with mild hemolysis to investigate if its presence could influence the uPA serum concentra-tion levels, using quantitative spectrophotometry (Plasma/Low Hb System; HemoCue, Lake Forest, California, USA), according to the manufacturer’s protocol.

uPA measurementThe serum levels of uPA were measured using a feline-specific uPA

sandwich enzyme-linked immunosorbent assay (cat. no. MBS070832; MyBiosource, San Diego, California, USA) performed according to the manufacturer’s instructions. All samples were run in duplicate and optical densities (OD) were read spectrophotometrically at 450 nm in a microplate reader. The average OD of the kit standard solution was calculated and a standard curve was achieved using the CurveExpert Professional 2.4.0 software (MyBiosource), from which the uPA serum concentrations were obtained.

Statistical analysisAll data for the determination and comparison of uPA levels in

cats with oncologic disease and healthy cats were registered through Microsoft Excel 2015.

Statistical analysis was performed using SPSS Statistics 22 (IBM, Armonk, New York, USA). Taking into account the size of the stud-ied sample (. 30) and the normality of the data (20), we opted for the appropriate parametric tests.

Student’s t-test was used to compare the differences in uPA serum concentrations between 2 groups of data, while an analysis of variance (ANOVA) was used for comparisons among more than 2 groups. The calculations of averages, medians, and standard deviations (SD) were performed using descriptive statistics and the significance level was defined as P , 0.05.



Re s u l t sThis study included 37 cats, 19 controls, and 18 oncologic cats.

Twenty-two were females (59.5%) and 15 were males (40.5%). The group of cats with cancer included 10 carcinomas (7 grade II and III mammary carcinomas and 3 adenocarcinomas localized in the esophagus, intestine, and salivary gland), 3 interscapular fibrosarco-mas, and 5 hematopoietic neoplasms (4 lymphomas and 1 leukemia).

The mean serum uPA concentration of healthy cats (6 SD) was 1.10 6 1.16 ng/mL, while oncologic cats presented a lower mean concentration of 0.54 6 0.22 ng/mL (Table I). To ensure that the low levels of hemolysis found in some samples did not influence results, hemoglobin values were measured and related to uPA. No significant influence was detected (Table I).

Mean uPA serum values were not significantly affected by breed, gender, reproductive status, weight, or age (Tables II and III). In cats with cancer, mean uPA values were slightly higher in females (0.61 ng/mL) than males (0.49 ng/mL), in non-sterilized cats (0.59 ng/mL) than sterilized cats (0.53 ng/mL), and in cats aged $ 11 y (0.65 ng/mL) than cats aged , 7 y (0.54 ng/mL) and

7 to 11 y (0.46 ng/mL) (Table III); although, these differences did not achieve statistical significance.

Tumor locations (cutaneous/mammary, internal organs, and hematopoietic neoplasms) and histological types (carcinomas, fibrosarcomas, and lymphoid neoplasms) were found to have no influence on circulating uPA values (Table IV).

Cats with cutaneous/mammary tumors, which represented more than 50% of the population, did not show significant variations in the mean serum concentration of uPA when compared to cats with other tumor types. Furthermore, cutaneous/mammary tumor size was not significantly related to circulating uPA values. There was a high number of cases with tumor-associated inflammation and those had slightly higher uPA concentrations (0.60 ng/mL 6 0.27), although the difference was not statistically significant (Table IV).

Finally, cats with cancer which developed tumor metastases (n = 8; 44.5%) had similar uPA values to those without tumor dissemination (Table IV). The metastatic group included 4 mammary carcinomas, 1 intestinal adenocarcinoma, 2 intestinal lymphomas, and 1 patient with acute leukemia which was included due to the disseminated nature of the neoplasm.

Table I. uPA serum concentrations and hemolysis levels in healthy cats and oncologic feline patients. Data are represented as mean 6 standard deviation.

Animals 95% Confidence interval n % uPA serum values (ng/mL) Minimum Maximum P-valueN 37 100Cases 18 49 0.54 6 0.22 0.44 0.65 0.05Controls 19 51 1.10 6 1.16 0.54 1.66 Hemolysis levels (gl/dL)Cases 18 49 0.05 6 0.04 0.03 0.07 0.06Controls 19 51 0.09 6 0.05 0.06 0.10

Table II. Relationship between uPA serum concentrations and intrinsic parameters in healthy cats. Data are represented as mean 6 standard deviation.

Animals uPA serum 95% Confidence interval N % values (ng/mL) Minimum Maximum P-valueBreed European shorthair 17 89.5 1.16 6 1.22 0.53 1.78 0.86 Persian 1 5.25 0.61 6 0 0 0 Siamese 1 5.25 0.68 6 0 0 0

Gender Female 13 68.4 1.34 6 1.36 — — 0.70 Male 6 31.6 0.59 6 0.04 — —

Weight (kg) $ 2.5 to 4 11 57.9 0.77 6 0.42 — — 0.4 $ 4.1 to 10 8 42.1 1.55 6 1.68 — —

Age (y) , 7 14 73.7 0.94 6 0.52 0.64 1.24 0.60 7 to 11 3 15.8 0.68 6 0.20 0.17 1.18 $ 11 2 10.5 2.88 6 3.69 30.2 35.9



D i s c u s s i o nIn human medicine, uPA system components show altered expres-

sion patterns in several common malignancies, which make them ideal diagnostic and prognostic markers, as well as therapeutic targets to reduce cancer-associated morbidity and mortality (21).

In dogs, uPA has been studied by immunohistochemical methods (13–16), and recently, serum levels were determined for the first time (11). However, there are very few studies regarding uPA in cats with cancer and none have evaluated uPA values in serum.

Therefore, the first aim of our study was to determine the normal serum values of uPA in healthy cats. We found that their mean uPA concentration (1.10 6 1.16 ng/mL) was within the described range for healthy humans (0.7 6 0.42 ng/mL to 1.1 6 0.38 ng/mL) (4,9,22,23), suggesting that the physiologic values of uPA remain constant between species. The presence of mild hemolysis in some serum samples included in our study did not influence uPA values. Supporting this statement, Dimeski (24) mentioned that hemolysis rarely influences immunoassays.

The other important finding in this study, similar to what has been reported in human studies, was that feline serum uPA values seemed to be independent of intrinsic parameters such as breed, gender, weight, reproductive status, and age (23,25).

The second objective was to compare-circulating uPA values in healthy cats with those of cats with cancer to evaluate uPA’s potential as a cancer biomarker in the species. Surprisingly, uPA values were significantly lower in cats with cancer (0.54 6 0.22 ng/mL), but still similar to some values reported in humans with cancer, such as patients with sarcomas (0.66 ng/mL) (26) and prostatic carcinomas (0.68 ng/mL) (27).

Several human reports described a higher uPA concentration, since this protease is upregulated in a wide variety of carcinogenic processes (21). It has also been associated with a poor prognosis in

patients with distinct types of malignant neoplasms and is used in women with lymph node-negative breast cancer to recommend chemotherapy (8,19). However, in some human oncology reports, uPA serum concentrations varied according to tumor type (3,22). For instance, Maguire et al (28), studied the serum levels of uPA in different types of skin tumors and found that the concentrations varied according to the tumor histology and were positively associ-ated with tumor aggressiveness. In this study, we included cats with several malignant tumors; therefore, we may hypothesize that the different cell types segregate uPA in distinct quantities and that may have contributed to the unexpected lower values of uPA in cancerous cats compared to healthy ones. Also supporting this hypothesis are the very different concentrations of uPA that have been reported in the blood of human patients with ovarian carcinomas (1.9 ng/mL) (9), pancreatic ductal adenocarcinomas (3.23 ng/mL) (4), bladder transition cell carcinomas (0.32 ng/mL) (29), and different sarcomas (0.66 ng/mL) (26). Future studies in feline cancer patients should evaluate uPA in more specific tumor groups.

The absent association between uPA levels and the studied clinicopathological parameters may be a consequence of the small sample size, the heterogeneity of the tumor types, or an effective and fast uPA metabolization/elimination. Other authors also failed to find associations between uPA serum levels and clinicopathologi-cal factors in human patients with prostatic cancer and sarcomas, including age (26,28), tumor histological grade (26), lymph node status (28), clinical stage (26,28), and survival (26,28), indicating that the relevance of uPA as a tumor marker or prognostic factor may be restricted to specific subsets of patients.

Serum uPA values were unrelated to the size of cutaneous/ mammary neoplasms, in opposition to some reports in human breast cancer and canine mammary tumors that established a relationship between uPA serum levels and the dimensions of the neoplasms (13,14,30,31).

Table III. Relationship between uPA serum concentrations and intrinsic parameters in oncologic cats. Data represented as mean 6 standard deviation.

Animals uPA serum 95% Confidence interval n % values (ng/mL) Minimum Maximum P-valueBreed European shorthair 5 27.8 0.57 6 0.24 0.27 0.87 0.56 Persian 1 5.6 0.23 6 0 0 0 Siamese 1 5.6 0.5 6 0 0 0 Mixed breed 11 61.1 0.57 6 0.22 0.43 0.72

Gender Female 9 50 0.61 6 0.22 — — 0.29 Male 9 50 0.49 6 0.22 — —

Reproductive status Non-sterilized 5 27.8 0.59 6 0.35 — — 0.73 Sterilized 13 72.2 0.53 6 0.17 — —

Age (y) , 7 5 27.8 0.54 6 0.16 0.34 0.74 0.23 7 to 11 7 38.9 0.46 6 0.14 0.32 0.59 $ 11 6 33.3 0.67 6 0.30 0.35 0.98



Although the difference was not statically significant, we found more cases of cats with cutaneous/mammary tumors with inflam-mation and those had slightly higher uPA values than those without inflammation. Indeed, uPA may be secreted not only by neoplastic cells but also by tumor-associated fibroblasts and inflammatory cells (13,19). The presence of these cells may therefore increase the concentration of several matrix proteolytic enzymes, includ-ing MMPs (MMP-2 and MMP-9), and uPA (32–34). The anchorage of uPA to its cell surface receptor not only favors the degrada-tion of the extracellular matrix, but also regulates cell migration, adhesion, proliferation, and angiogenesis, thus influencing the development of inflammatory and immune responses in the tumor microenvironment (2,31,35). Since our study sample was small, this potential relationship between uPA values and inflammation needs to be confirmed with further studies with a higher number of cases.

Finally, cats with metastasis had similar serum uPA values to those without tumor dissemination, which has also been reported in humans with prostatic cancer (27). We may hypothesize that either the metastatic process of some tumors included in this study is more dependent on other proteases or the serum values do not correlate to the uPA levels in the tumor microenvironment. Both hypotheses deserve to be further studied by comparing serum and tissue expres-

sions of uPA and other proteases in the same cats. In addition, as recently reviewed (21), the clinical value of the plasminogen activator system, as tumor biomarker and prognosticator, seems to increase when determining uPA-uPAR and uPA-PAI-1, rather than uPA alone, both in tumor tissue and serum.

In conclusion, the urokinase plasminogen activator system is still very understudied in veterinary medicine, especially in feline oncology compared with studies in human oncology. This study intended to establish and compare, for the first time, the mean serum concentrations of uPA in healthy cats and cats with cancer. These preliminary results were not strong enough to support uPA as a useful serum biomarker to screen cats with cancer or as a prog-nostic factor. However, it is the starting point to understand the relevance of uPA in cats with cancer and to propose directions for future studies, namely focusing on specific neoplasms (e.g., with similar histopathological characteristics) that may rely differently on the uPAS during the carcinogenic process.

A c k n o w l e d g m e n t sWe gratefully thank the Biomedical Sciences Institute of Abel

Salazar, University of Porto, for providing financial support to carry out this research project; the Veterinary Hospital of the University

Table IV. uPA serum concentrations according to clinicopathological parameters in cats with cancer. Data represented as mean 6 standard deviation.

Animals uPA serum 95% Confidence interval n % values (ng/mL) Minimum Maximum P-valueLocation Cutaneous 10 55.5 0.57 6 0.23 0.40 0.73 0.66 Internal organs 5 28 0.58 6 0.25 0.27 0.89 Hematopoietic 3 16.5 0.44 6 0.18 0.01 0.89

Neoplasms Non cutaneous 8 44.5 0.53 6 0.22 0.37 0.68 0.70 Cutaneous 10 55.5 0.57 6 0.23 0.43 0.71

Largest diameter in cutaneous neoplasms (cm) # 3 0 0 0 0 0.13 3 to 5 6 60 0.66 6 0.25 0.40 0.93 $ 5 4 40 0.43 6 0.09 0.28 0.58

Inflammation in cutaneous neoplasms Yes 7 70 0.60 6 0.27 0.40 0.80 0.28 No 3 30 0.48 6 0.04 0.44 0.83

Histologic type Carcinoma 10 55.5 0.57 6 0.23 0.41 0.75 0.47 Fibrosarcoma 3 16.5 0.40 6 0.09 0.18 0.62 Hematopoietic 5 28 0.58 6 0.25 0.27 0.88

Metastasis Metastatic 8 44.5 0.54 6 0.26 0.33 0.77 0.96 Non-metastatic 10 55.5 0.55 6 0.19 0.38 0.69



of Porto, the CHV — Veterinary Hospital in Porto, and the BSA Animal Blood Bank, for providing some of the blood samples and clinical data; the Laboratory of Clinical Analysis and Histopathology of ULHT, and the Laboratory of Immunology of ICBAS for helping with the sample conservation and analysis.

Re f e r e n c e s1. Falco M, Palma G, Rea D, et al. Tumour biomarkers: Homeostasis

as a novel prognostic indicator. Open Biol 2016;6:160254.2. Ulisse S, Baldini E, Sorrenti S, D’Armiento M. The urokinase

plasminogen activator system: A target for anti-cancer therapy. Curr Cancer Drug Targets 2009;9:32–71.

3. Shariat SF, Roehrborn CG, McConnell JD, et al. Association of the circulating levels of the urokinase system of plasminogen activation with the presence of prostate cancer and invasion, progression and metastasis. J Clin Oncol 2007;25:349–355.

4. Winter K, Szczesniak P, Bulska M, et al. Serum level of uroki-nase plasminogen activator (uPA) correlates with the survival of patients with pancreatic ductal adenocarcinoma (PDAC). Pancreat Disord Ther 2015;5:163.

5. Jänicke F, Prechtl A, Thomssen C, et al. Randomized adjuvant chemotherapy trial in high-risk, lymph node-negative breast cancer patients identified by urokinase-type plasminogen acti-vator and plasminogen activator inhibitor type 1. J Natl Cancer Inst 2001;93:913–920.

6. Harbeck N, Kates RE, Look MP, et al. Enhanced benefit from adjuvant chemotherapy in breast cancer patients classified high-risk according to urokinase-type plasminogen activator (uPA) and plasminogen activator inhibitor type 1 (n = 3424). Cancer Res 2002;62:4617–4622.

7. Strojan P, Budihna M, Šmid L, Vrhovec I, Škrk J. Urokinase-type plasminogen activator (uPA) and plasminogen activator inhibitor type 1 (PAI-1) in tissue and serum of head and neck squamous cell carcinoma patients. Eur J Cancer 1998;34:1193–1197.

8. Minisini AM, Fabbro D, Di Loreto C, et al. Markers of the uPA system and common prognostic factors in breast cancer. Am J Clin Pathol 2007;128:112–117.

9. Zhang W, Ling D, Tan J, Zhang J, Li L. Expression of urokinase plasminogen activator and plasminogen activator inhibitor type-1 in ovarian cancer and its clinical significance. Oncol Rep 2013;29:637–645.

10. Märkl B, Hardt J, Franz S, et al. Tumor budding, uPA, and PAI-1 in colorectal cancer: Update of a prospective study. Gastroenterol Res Pract 2017;2017:6504960.

11. Ramos SC, de Matos AJ, Ribeiro JN, et al. Serum levels of urokinase- type plasminogen activator in healthy dogs and oncologic canine patients. Vet World 2017;10:918–923.

12. Santos AA, Lopes CC, Ribeiro JR, et al. Identification of prognos-tic factors in canine mammary malignant tumours: A multivari-able survival study. BMC Vet Res 2013;9:1–11.

13. Santos A, Lopes C, Marques RM, et al. Immunohistochemical analysis of urokinase plasminogen activator and its prognostic value in canine mammary tumours. Vet J 2011;189:43–48.

14. Golshahi H, Tavasoly A, Rezaie A, Farhadi M. Urokinase plas-minogen activator expression in canine malignant mammary

tumours by immunohistochemical study. Bull Vet Inst Pulawy 2013;57:587–591.

15. Faleiro MR, Toledo DC, Rodrigues MP, et al. uPAR expression in canine normal prostate and with proliferative disorders. Ci Anim Bras 2013;14:237–244.

16. Anwar S, Yanai T, Sakai H. Immunohistochemical detection of urokinase plasminogen activator and urokinase plasminogen activator receptor in canine vascular endothelial tumours. J Comp Pathol 2015;153:278–282.

17. Schappa JT, Frantz AM, Gorden BH, Dickerson EB, Vallera DA, Modiano JF. Hemangiosarcoma and its cancer stem cell subpopulation are effectively killed by a toxin targeted through epidermal growth factor and urokinase receptors. Int J Cancer 2013;133:1936–1944.

18. Leonardi L, Quattrini I, Roperto F, Benassi MS. Protease expres-sion in giant cell tumour of bone: A comparative study on feline and human samples. Res Vet Sci 2013;95:310–315.

19. Hildenbrand R, Schaaf A, Dorn-Beineke A, et al. Tumor stroma is the predominant uPA-, uPAR-, PAI-1 expressing tissue in human breast cancer: Prognostic impact. Histol Histopathol 2009;24:869–877.

20. Ghasemi A, Zahediasl S. Normality tests for statistical analysis: A guide for non-statisticians. Int J Endocrinol Metab 2012;10: 486–489.

21. Mahmood N, Mihalcioiu C, Rabbani SA. Multifaceted role of the urokinase-type plasminogen activator (uPA) and its receptor (uPAR): Diagnostic, prognostic, and therapeutic applications. Front Oncol 2018;8:24.

22. Cobos E, Jumper C, Lox C. Pretreatment determination of the serum urokinase plasminogen activator and its soluble receptor in advanced small-cell lung cancer or non-small-cell lung cancer. Clin Appl Thromb Hemost 2003;9:241–246.

23. Gustafsson A, Ajeti V, Ljunggren L. Detection of suPAR in the saliva of healthy young adults: Comparison with plasma levels. Biomark Insights 2011;6:119–125.

24. Dimeski G. Interference testing. Clin Biochem Rev 2008;29: S43–S48.

25. Fuessel S, Erdmann K, Taubert H, et al. Prognostic impact of urokinase-type plasminogen activator system components in clear cell renal cell carcinoma patients without distant metasta-sis. BCM Cancer 2014;14:974.

26. Taubert H, Würl P, Greither T, et al. Co-detection of members of the urokinase plasminogen activator system in tumour tis-sue and serum correlates with a poor prognosis for soft-tissue sarcoma patients. Br J Cancer 2010;102:731–737.

27. Al-Janabi O, Taubert H, Lohse-Fischer A, et al. Association of tissue mRNA and serum antigen levels of members of the urokinase-type plasminogen activator system with clinical and prognostic parameters in prostate cancer. Biomed Res Int 2014;2014:972587.

28. Maguire T, Chin D, Soutar D, Duffy MJ. Low levels of urokinase plasminogen activator components in basal cell carcinoma of the skin. Int J Cancer 2000;85:457–459.

29. Shariat SF, Monoski MA, Andrews B, Wheeler TM, Lerner SP, Slawin KM. Association of plasma urokinase-type plasmino-gen activator and its receptor with clinical outcome in patients



undergoing radical cystectomy for transitional cell carcinoma of the bladder. Urology 2003;61:1053–1058.

30. Lampelj M, Arko D, Cas-Sikosek N, et al. Urokinase plasmino-gen activator (uPA) and plasminogen activator inhibitor type-1 (PAI-1) in breast cancer — Correlation with traditional prognos-tic factors. Radiol Oncol 2015;49:357–364.

31. Dublin E, Hanby A, Patel NK, Liebman R, Barnes D. Immuno-histochemical expression of uPA, uPAR, and PAI-1 in breast carcinoma. Fibroblastic expression has strong associations with tumor pathology. Am J Pathol 2000;157:1219–1227.

32. Zhang J, Sud S, Mizutani K, Gyetko MR, Pienta KJ. Activation of urokinase plasminogen activator and its receptor axis is essential

for macrophage infiltration in a prostate cancer mouse model. Neoplasia 2011;13:23–30.

33. Liu J, Lin PC, Zhou BP. Inflammation fuels tumor progress and metastasis. Curr Pharm Des 2015;21:3032–3040.

34. Obeid E, Nanda R, Fu YX, Olopade OI. (2013). The role of tumor-associated macrophages in breast cancer progression (review). Int J Oncol 2013;43:5–12.

35. Del Rosso M, Margheri F, Serratì S, Chillà A, Laurenzana A, Fibbi G. The urokinase receptor system, a key regulator at the intersection between inflammation, immunity, and coagulation. Curr Pharm Des 2011;17:1924–1943.


Article


I n t r o d u c t i o nDegenerative mitral valve disease (DMVD) is the most common

acquired cardiovascular disease in dogs and accounts for approxi-mately 75% of cases of chronic heart failure (1,2). The macroscopic lesions on the atrioventricular valves in the course of DMVD consist of small thickened nodules on the leaflets and rupture of the chordae tendineae (3). This leads to poor coaptation of the mitral valve, its

insufficiency, and eventually the development of congestive heart failure. This disease is most prevalent in older, small-breed dogs, although it may occasionally occur in larger breeds (4). Cavalier King Charles spaniels are particularly predisposed to this condition and develop degenerative lesions at an earlier age (5).

The disease is characterized by a long asymptomatic period that progresses into a clinical stage observed only in some dogs. The goal of therapy is to prolong the asymptomatic stage of the disease, which

Antioxidative enzyme activity and total antioxidant capacity in serum of dogs with degenerative mitral valve disease

Marcin Michałek, Aleksandra Tabis, Alicja Cepiel, Agnieszka Noszczyk-Nowak

A b s t r a c tThis study was designed to evaluate the antioxidative status of serum by measuring its total antioxidant capacity, as well as the antioxidant enzyme activity (superoxide dismutase, catalase, and glutathione reductase), in dogs with various stages of degenerative mitral valve disease (DMVD) compared to healthy controls. In total, 71 client-owned dogs in different stages of DMVD, which included healthy controls, took part in the study. Following an anamnesis, clinical examination, standard transthoracic echocardiograpic examination, chest X-ray, complete blood (cell) count, and serum biochemistry, dogs were divided into 2 study groups. Blood was drawn from each dog once at the time of presentation and selected antioxidant parameters were measured using commercially available assay kits. The activity of superoxide dismutase gradually decreased in the more advanced stages of DMVD, while the activity of catalase was significantly higher in the group of dogs with asymptomatic DMVD compared to healthy controls and dogs with symptomatic DMVD. No significant changes were noted in total antioxidant capacity and the activity of glutathione reductase. Results suggested that DMVD has a significant impact on the activity of superoxide dismutase and catalase in the serum of the tested dogs. Knowledge of changes in the activity of antioxidative enzymes may warrant further studies, possibly to evaluate the potential role of compounds with antioxidative properties in the clinical outcome of dogs with DMVD.

R é s u m éLa présente étude a été conçue afin d’évaluer le statut antioxydant du sérum en mesurant sa capacité antioxydante totale, ainsi que l’activité antioxydante enzymatique (superoxyde dismutase, catalase, et glutathion réductase), chez des chiens avec des degrés divers de maladie dégénérative de la valvule mitrale (DMVD) comparativement à des témoins en santé. Au total, 71 chiens appartenant à des clients à différents stades de DMVD, qui incluaient des témoins en santé, ont pris part à cette étude. À la suite de la prise d’anamnèse, d’un examen clinique, d’un examen échocardiographie transthoracique standard, de radiographie thoracique, d’un comptage cellulaire sanguin complet, et d’analyse biochimique sérique, les chiens étaient séparés en deux groupes d’étude. Du sang fut prélevé de chaque chien une fois au moment de la présentation et les paramètres antioxydants sélectionnés furent mesurés à l’aide d’une trousse disponible commercialement. L’activité de la superoxyde dismutase diminuait graduellement dans les stades plus avancés de DMVD, alors que l’activité de la catalase était significativement plus élevée dans le groupe de chiens avec une DMVD asymptomatique comparativement aux témoins en santé et aux chiens avec une DMVD symptomatique. Aucun changement significatif n’était noté dans la capacité antioxydante totale et dans l’activité de la glutathion réductase. Les résultats suggèrent que la DMVD a un impact significatif sur l’activité de la superoxyde dismutase, et de la catalase dans le sérum des chiens testés. Des connaissances sur les changements dans l’activité des enzymes antioxydantes pourraient justifier des études additionnelles, possiblement pour évaluer le rôle potentiel de produits avec des propriétés antioxydantes dans le devenir clinique de chiens avec DMVD.


Department of Internal Medicine and Clinic of Diseases of Horses, Dogs and Cats (Michałek, Cepiel, Noszczyk-Nowak) and Department of Food Hygiene and Consumer Health (Tabis), Faculty of Veterinary Medicine, Wrocław University of Environmental and Life Sciences, Wroclaw, Poland.

Address all correspondence to Dr. Marcin Michałek; telephone: 148 71 320 1011; e-mail: [email protected]

Received December 26, 2018. Accepted April 12, 2019.



is currently achieved using pharmacotherapy (6). There are single reports suggesting that the pathophysiology of DMVD might be associated with oxidative stress and that administering antioxidants may positively affect the patient clinical outcome. Significant differ-ences were found between dogs with chronic heart failure (CHF) and healthy controls in plasma 8-F2a

-isoprostane and total antioxidant capacity (7). Moreover, higher erythrocytic glutathione peroxidase activity was reported in dogs with CHF secondary to DMVD than in dogs with dilated cardiomyopathy (DCM) (8). It was also found that serum paraoxonase-1 activity was lower in dachshunds with asymptomatic DMVD than in those with a symptomatic stage of the disease and healthy controls (9). On the other hand, no differences were found either in the plasma coenzyme Q10 concentration or the erythrocyte superoxide dismutase (SOD) and whole blood glutathi-one peroxidase activity among dogs with different classes of heart failure (10). Furthermore, it was found that plasma malondialdehyde, oxidized low-density lipoprotein, and vitamin E were not associated with a clinical stage of DMVD (7,11).

Oxidative stress has recently been defined as “an imbalance between oxidants and antioxidants in favor of the oxidants, leading to a disruption of redox signalling and control and/or molecular damage” (12). The most common reactive oxygen species (ROS) include the superoxide anion radical (O2•−), hydrogen peroxide (H2O2), the hydroxyl radical (HO•), as well as singlet oxygen (12). An uncontrolled increase in the levels of ROS may cause structural cellular injury, damaging deoxyribonucleic acid (DNA), proteins, and lipids. Biological compounds, such as carotenoids, vitamins C and E, uric acid, glutathione, and a number of enzymes that catalyze the degradation of reactive forms of oxygen provide protection against ROS. The superoxide anion radical is neutralized to H2O2 by super-oxide dismutase, while H2O2 is converted into water by catalase and glutathione peroxidase. The latter reaction uses glutathione, which serves as an electron donor, converting it to its oxidized form, glu-tathione disulfide. Finally, glutathione reductase (GR) reduces the oxidized glutathione to the monomeric molecule, completing the cycle (13). These enzymes act intra- and extracellularly and play an important role in protecting the organism from oxidative injury.

To date, no reports exist on the activity of major antioxidative enzymes in serum during the course of DMVD. Therefore, the aim of this study was to evaluate the antioxidative status of serum by measuring its total antioxidant capacity as well as the antioxidative enzyme activity in dogs with various stages of degenerative mitral valve disease compared to healthy controls.

M a t e r i a l s a n d m e t h o d sA total of 71 dogs were included in the study. All dogs underwent

a detailed anamnesis, clinical examination, standard transthoracic echocardiograpic examination (Aloka F36 or Aloka Arietta V60; Hitachi-Aloka, Tokyo, Japan), chest X-rays (GIERTH HF 200A; Gierth X-Ray International, Riesa, Germany), as well as a complete blood (cell) count (CBC) (LaserCyte Dx; IDEXX Laboratories, Westbrook, Maine, USA), and serum chemistry analysis (Konelab Prime 30ISE; Thermo Scientific, Waltham, Massachusetts, USA). Only animals with no changes in the echocardiogram (M-mode, B-mode, color doppler) and no degenerative lesions typical for DMVD, as well

as no abnormalities found during the various examinations, were included in the control group (group A). Blood samples used as controls were collected from healthy animals presenting to the Faculty’s Small Animal Teaching Hospital for preventive screening. An additional cardiac examination was offered to these individuals at no further cost.

The main inclusion criterion for the study group was the presence of an echocardiographically confirmed mitral valve insufficiency (or an additional tricuspid valve insufficiency) formed secondary to degenerative lesions of its leaflets, diagnosed using color doppler as well as continuous-wave doppler. Group B contained animals in stage B1 or B2 of DMVD according to the American College of Veterinary Internal Medicine (ACVIM) classification, including dogs with DMVD and echocardiographic features of cardiomegaly or without cardiomegaly, without clinical signs of chronic heart failure (14). Group C included animals with stage Cc disease according to the ACVIM classification. These were dogs with compensated chronic heart failure caused by DMVD. Animals with acute DMVD were excluded from the study. In addition, animals with abnormali-ties in their CBC or serum biochemistry (up to a 2-fold increase in the reference range values of aminotransferase activity and urea was considered acceptable) and clinical, ultrasound, or X-ray signs of any other disease apart from chronic DMVD were excluded from the study.

Blood from all the studied animals was drawn from a peripheral vein into serum and ethylenediaminetetraacetic acid (EDTA) tubes. All the animals fasted for at least 12 h before the examinations. The blood collected for serum was centrifuged after 15 min at room temperature (2000 3 g, 15 min, 4°C) and transferred to a laboratory for biochemical analysis, while the remaining blood was divided and frozen at 280°C until analysis. All samples were collected within a 4-month period and stored no longer than 6 mo in total. Written informed consent was obtained from all the owners.

Determination of activity of antioxidant enzymes in serum

The activity of extracellular superoxide dismutase-3 (SOD-3), catalase (CAT), and glutathione reductase (GR) was measured in the serum of the studied animals using commercial kits (Cayman Chemical, Ann Arbor, Michigan, USA and BioVision, Milpitas, California, USA) (SOD: 706002; CAT: 707002; GR: #K761-200). One unit of SOD was defined as the amount of enzyme needed to exhibit 50% dismutation of the superoxide anion radical. The activity of CAT was defined as the amount of enzyme leading to the formation of 1 nmol of formaldehyde per minute at room temperature. One unit of GR was defined as the amount of enzyme that generates 1.0 mmol of 5-thio-2-nitrobenzoic acid (TNB) per minute at room temperature.

Determination of total antioxidant capacityThe total antioxidant capacity of serum was measured using the

cupric-reducing antioxidant capacity (CUPRAC) method that is based on a single electron transfer (SET) mechanism. A commercially available reagent kit was used to carry out the assay (OxiSelect Total Antioxidant Capacity; Cell Biolabs, San Diego, California, USA). The results of the total antioxidant capacity were expressed as copper-reducing equivalents (CRE). The absorbance of the colored enzymatic



reaction products was measured using a multi-plate spectrophotom-eter (Tecan Spark 10M; Tecan, Austria). All the samples were run in duplicate and analyzed simultaneously by 1 examiner (AT) who was blinded to the sample origin, while the final result was an average value of 2 measurements.

Statistical analysisThe data underwent statistical analysis using the GraphPad

Prism 5.0 (GraphPad Software, San Diego, California, USA) and Statistica 13 software. The Kolmogorov-Smirnov normality test was used to assess the normally distributed data. The oxidative stress parameters with a skewed distribution were natural logarithm (ln) transformed to normality. Based on the data distribution, either a 1-way analysis of variance (ANOVA) with the Bonferroni post-hoc test or the Kruskal-Wallis with the post-hoc Dunns test were used to

compare more than 2 groups of data. Two groups of data were com-pared using the unpaired t-test or the nonparametric Mann-Whitney test. As the groups differed in terms of age, multivariable linear regression analyses, followed by analyses of covariance (ANCOVA), were conducted to evaluate potential bias. Correlations among the groups were determined using the Spearman correlation coefficient with P , 0.05 considered statistically significant.

Re s u l t sIn total, 71 dogs were included in the study. The mean (6 stan-

dard deviation) body weight of the animals was 10.2 6 3.7 kg and the mean age was 10.7 6 2.7 y. Forty of the dogs were male and 31 were female. The control group (group A) consisted of 16 healthy dogs, 9 male and 7 female, with a mean age of 8.1 6 2.6 y and a

Table I. Results of complete blood count and blood chemistry of dogs.

Variable A B C P-valueWhite blood cells (109/L) 7.25 6 2.28 7.98 6 2.45 8.88 6 2.61 0.08Red blood cells (1012/L) 6.85 6 0.52 7.24 6 0.82 7.14 6 0.81 0.4Hemoglobin (g/L) 162.3 6 13.6 169.3 6 15.6 162.4 6 19.9 0.33Hematocrit (%) 48 6 6 51 6 5 49 6 5 0.25AST (U/L) 27.58 6 6.73 26.96 6 7.05 28.72 6 8.8 0.72ALT (U/L) 42.08 6 18.99C 58.43 6 29C 91.72 6 53.2A,B 0.001Urea (mmol/L) 5.63 6 1.76C 5.96 6 2.07C 8.68 6 4.02A,B 0.002Creatinine (mmol/L) 91.5 6 24.02 71.26 6 31.17C 93 6 30.81B 0.03Total protein (g/L) 60.83 6 5.44 59.74 6 4.76 61.78 6 7.03 0.49Albumin (g/L) 30.25 6 2.42 30.61 6 2.93 32.19 6 2.84 0.06Magnesium (mmol/L) 0.7 6 0.09 0.74 6 0.09 0.74 6 0.11 0.51Sodium ion (mmol/L) 147.44 6 2.9C 144.4 6 3.09 143.62 6 3.86A 0.01Potassium (mmol/L) 4.43 6 0.2 4.45 6 0.35 4.39 6 0.36 0.79Chloride (mmol/L) 111.02 6 1.3 110.7 6 2.69 109.24 6 3.23 0.09Calcium ion (mmol/L) 1.36 6 0.07C 1.29 6 0.08 1.27 6 0.08A 0.02Glucose (mmol/L) 5.46 6 0.44 6.05 6 0.67 5.69 6 0.8 0.13Data are presented as mean 6 SD. Statistical significance among the groups is shown in the upper indexes (A,B,C).AST — aspartate aminotransferase; ALT — alanine aminotransferase.

Table II. Echocardiographic and treatment characteristics of study groups.

Variable A B C P-valueLA/Ao 1.33 6 0.14C 1.51 6 0.3C 2.39 6 0.49A,B 0.0001LVIDd 30.93 6 3.23C 33.37 6 4.34C 40.87 6 7.23A,B 0.0001LVIDdN 1.54 6 0.13C 1.7 6 0.21C 2.11 6 0.28A,B 0.0001Heart rate 111 6 12.25C 103.1 6 18.89C 159.5 6 46.51A,B 0.0004TR 0/16 1/23 14/32Cardiac treatment 0/16 6/23 22/32Benazepril 0/16 6/23 20/32Spironolactone 0/16 1/23 17/23Pimobendan 0/16 0/23 17/23Furosemide 0/16 0/23 18/23Data are presented as mean 6 SD. Statistical significance among the groups is shown in the upper indexes (A,B,C).LA/Ao — left-atrium-to-aorta ratio; LVIDd — left ventricular inner diameter in diastole; LVIDdN — normalized left ventricular end-diastolic diameter (34); TR — tricuspid regurgitation.Mode heart dimensions were measured in the subvalvular region.



mean body weight of 11.2 6 3 kg and included 7 mixed-breed dogs, 3 beagles, 3 miniature schnauzers, a Cavalier King Charles spaniel, a dachshund, and a Nova Scotia duck tolling retriever. Group B included 23 dogs with asymptomatic mitral valve insufficiency with or without cardiomegaly, 13 males and 10 females, with a mean age of 10.3 6 2.6 y and a mean body weight of 10 6 3.5 kg. There were 8 mixed-breed dogs, 4 dachshunds, 3 miniature schnauzers, 2 Shih Tzus, 2 Cavalier King Charles spaniels, 1 bichon frise, 1 fox terrier, 1 Yorkshire terrier, and an English cocker spaniel. Group C con-sisted of 32 dogs with chronic heart failure caused by DMVD, with a mean age of 11.9 6 1.7 y and a mean weight of 9.95 6 4 kg. There were 18 male and 14 female dogs, which included 16 mixed-breed dogs, 4 Cavalier King Charles spaniels, 4 dachshunds, 2 miniature schnauzers, 2 Pekingese, 2 Shih Tzus, 1 bull terrier, and a medium poodle. The groups did not differ with respect to body weight, although the dogs with heart failure were significantly older than the healthy dogs (P , 0.001). The results of the CBC and serum chemistry of the dogs are included in Table I, while the results of the heart ultrasound are presented in Table II.

The study found a statistically significant difference in the activity of the antioxidant enzymes among the groups. The activ-ity of SOD-3 was significantly lower in the dogs from group C (mean value: 1.33 6 0.71 U/mL) compared to group B (mean value: 1.86 6 0.69 U/mL) and the healthy dogs (mean value: 2.15 6 1.1 U/mL) (P , 0.01) (Figure 1). The serum activity of catalase was significantly higher in group B (mean value: 9.98 6 3.32 nmol/min/mL) than in group A (mean value: 5.89 6 1.46 nmol/min/mL) and group C (mean value: 7.69 6 3.62 nmol/min/mL) (P , 0.01) (Figure 2). The activity of glutathione reductase in group A was 40.2 6 8.26 mU/mL, 42.5 6 9.38 mU/mL in group B, and 43.9 6 10.6 mU/mL in group C, and this value did not differ significantly among the groups (P = 0.57) (Figure 3). The total antioxidant capac-ity of serum in group A was 579 6 54 copper-reducing equivalents

(CRE), 623 6 99 CRE in group B, and 601 6 106 CRE in group C and did not differ significantly among the groups (P = 0.61) (Figure 4). There were no associations between age and studied oxidative stress parameters in the multivariable linear regression model. The following intra-assay coefficients of variation (CV) measured for canine samples were obtained: CAT — 23.5%; total oxidant capacity (TAC) — 13.3%; GR — 20.2%; and SOD — 22%.

In addition, the symptomatic group of dogs was divided into 2 subgroups: dogs treated pharmacologically (n = 22) and dogs before pharmacological treatment (n = 10) (Table II). There were no significant differences in any of the parameters between the groups.

The correlation analysis revealed a moderate positive correlation between the CAT activity and TAC (r = 0.42, P = 0.001). The activity of catalase weakly correlated with age (r = 0.31, P = 0.012).

D i s c u s s i o nThis study proved that the serum activity of superoxide dismutase

and catalase differs between healthy animals and those with asymp-tomatic and symptomatic DMVD. To the authors’ knowledge, this is the first study to focus on the serum activity of antioxidant enzymes in dogs with DMVD.

Most dogs in this study were small and miniature breeds due to a higher prevalence of DMVD in those breeds than in larger breeds. Despite the fact that Cavalier King Charles spaniels are particularly predisposed to DMVD, they constituted a very small group of subjects due to their low popularity in Poland (15). The mean age of dogs in the groups increased with more advanced heart failure. This is consistent with the authors’ expectations, as the prevalence and severity of the disease increase with age (14).

Mammals have developed enzymatic mechanisms that directly protect the organism from excessive oxidative stress. In the current study, the activity of extracellular isoform of superoxide dismutase

Figure 1. The superoxide dismutase (SOD) activity of serum in healthy dogs and dogs with degenerative mitral valve disease (DMVD). * P , 0.05; ** P , 0.01. A — healthy controls; B — asymptomatic DMVD; C — symptomatic, stable DMVD.

Figure 2. The catalase (CAT) activity of serum in healthy dogs and dogs with degenerative mitral valve disease (DMVD). ** P , 0.01. A — healthy controls; B — asymptomatic DMVD; C — symptomatic, stable DMVD.



(SOD) decreased gradually and significantly. In veterinary litera-ture, there are conflicting reports of SOD activity in dogs with car-diovascular disease. For example, studies on dogs with surgically induced mitral valve insufficiency found a lower cellular activity of SOD in the cardiac tissue from the left ventricle (16). Another study assessing the activity of SOD in an erythrocyte lysate from dogs with idiopathic dilated cardiomyopathy did not confirm this relationship (17). A study on human patients with congestive heart failure found a lower catalytic activity of this enzyme compared to healthy subjects and a negative correlation between its activity and the serum concentration of uric acid, which has an antioxidative effect (18). It may therefore be assumed that the gradual decrease in extracellular SOD activity is a result of declining efficacy of the antioxidant mechanisms caused by an increased oxidative assault. Measuring the SOD activity in blood serum may thus serve as an indicator for evaluating antioxidant status in the circulation of patients with DMVD.

Catalase, which is another parameter analyzed in this study, is considered one of the main intracellular antioxidant enzymes, although it was also reported to catalyze reactions associated with H2O2 metabolism in blood serum (19). Despite the fact that most studies focus on the intracellular activity of this enzyme, mainly in an erythrocyte lysate, there are single reports of its activity in the blood serum of cardiac patients (20). In humans with ischemic heart disease, the activity of serum catalase decreased with the increas-ing number of vessels affected by atherosclerosis (21). Our findings may suggest therefore that the increase in catalase activity in the asymptomatic stage of the disease was caused by an activation of compensatory mechanisms in response to an excessive formation of free radicals. These mechanisms were most likely exhausted in the symptomatic stage of the disease.

Glutathione reductase (GR) plays a key role in regenerating oxi-dized glutathione, which is a tripeptide with antioxidant properties, helping it to maintain its function. A study carried out on humans revealed that the activity of that enzyme significantly increases in patients with unstable angina or myocardial infarction, although there was no difference between patients with stable angina and the control group (22). This observation was confirmed by another study that reported increased GR activity in the blood plasma of patients admitted to hospital with unstable angina compared to patients at discharge and in the control group (23). Those results suggest that an acute stage of cardiovascular disease significantly disrupts gluta-thione metabolism. This is in accordance with our findings as we did not find significant differences in the GR activity among the studied groups of dogs. The patients with symptomatic DMVD in our study were in a stable stage of the disease, which may have significantly affected the activity of glutathione reductase.

The total antioxidant capacity (TAC) is measured in order to determine the ability of the organism to neutralize free radicals (24). The cupric-reducing antioxidant capacity method (CUPRAC) used in our study enables the measurement of the total antioxidant capacity formed by thiol-group antioxidants, vitamins (ascorbic acid, a-tocopherol, b-carotene), bilirubin, albumin, and uric acid (25). It is also known that TAC is not associated with age and sex in healthy dogs (26). This parameter has been repeatedly analyzed in human and canine cardiovascular diseases, with conflicting results. Dogs with chronic heart failure secondary to DMVD and DCM have sig-nificantly higher plasma antioxidant capacity than healthy dogs (7). In another study, dogs with cardiac disease (DMVD and DCM) had a significantly higher TAC, measured based on the ferric-reducing abil-ity of the plasma, than the control group, although an analysis using the ABTS method did not reveal differences between the 2 study

Figure 3. The glutathione reductase (GR) activity of serum in healthy dogs and dogs with degenerative mitral valve disease (DMVD). A — healthy controls; B — asymptomatic DMVD; C — symptomatic, stable DMVD.

Figure 4. The total antioxidant capacity (TAC) of serum in healthy dogs and dogs with degenerative mitral valve disease (DMVD). A — healthy controls; B — asymptomatic DMVD; C — symptomatic, stable DMVD.

-



groups (27). There were no significant differences in the TAC values between dogs with DMVD and DCM in either of the 2 studies (7,27).

Interestingly, another recent study reported a significantly lower plasma TAC in dogs with an asymptomatic phase of spontaneous cardiovascular disease compared to healthy dogs, although no dif-ferences were found between dogs with chronic heart failure and the control group (10). In humans, a strong correlation was found between TAC and the degree of dilated cardiomyopathy expressed in ejection fractions (28). In our study, there was no significant dif-ference in the total antioxidant capacity of serum among the groups and this parameter did not correlate with the stage of the disease [expressed in left-atrium-to-aorta ratio (LA/Ao), left ventricular inner diameter in diastole (LVIDd)]. In addition, there was no cor-relation of this parameter with fractional shortening (FS). This may be caused by the presence of stable and compensated heart failure in the dogs included in our study. Each method used to measure TAC is distinctive, however, and determines different components of the antioxidant defence, which is reflected in the different study results (24).

In this study, there were no differences between the group of dogs with heart failure treated pharmacologically and those without treatment, which is surprising. Several studies reported that the drugs used in standard therapy for chronic heart failure in dogs, such as angiotensin-converting enzyme inhibitors, spironolactone, beta-blockers, or calcium channel blockers, alleviate oxidative stress (29–32). Each drug represents its unique mechanism of free-radical scavenging action, which is usually linked to its chemical structure (33). The results of our study are in accordance with other studies carried out on dogs. For example, no significant difference was found between treated and non-treated dogs with DMVD and with DCM (8). In our study, however, the subgroups were relatively small, the dogs did not receive uniform pharmacotherapy as it was tailored to the needs of each dog, and the treatment duration was not taken into consideration. As a result, the effect of pharmacotherapy for heart failure on oxidative stress in dogs still needs to be studied.

This study had several limitations. In order to obtain practical results, the dogs were divided into groups based on the current ACVIM classification system for DMVD. Dogs with various sever-ity levels of the disease were therefore included in the symptomatic DMVD group, although all were in stable condition. Group B1 and B2 of the ACVIM classification were combined to increase the sample size for the statistical analysis, which should also be con-sidered a limitation. The analysis of a non-standardized group of dogs in terms of the pharmacological treatment may have affected the results. Although all dogs were fed commercial diets with no additional supplementation with antioxidants, the influence of antioxidant additives that might have been included in these diets cannot be excluded.

In summary, degenerative mitral valve disease had a significant impact on the activity of superoxide dismutase and catalase in the serum of the tested dogs. The activity of superoxide dismutase gradually decreased in the more advanced stages of DMVD, while the activity of catalase was significantly higher in the group of dogs with asymptomatic DMVD. Further research should focus on substances that increase the antioxidant capacity in sick animals, which could positively affect the clinical outcome. Other biomark-

ers of oxidative stress and antioxidative enzymes in other biological material of dogs with DMVD warrant further study.

A c k n o w l e d g m e n tThis work was supported by the Wrocław University of

Environmental and Life Sciences (Poland) as PhD research program ‘Innowacyjny Doktorat,’ no. D220/0003/17.

Re f e r e n c e s1. Borgarelli M, Häggström J. Canine degenerative myxomatous

mitral valve disease: Natural history, clinical presentation and therapy. Vet Clin North Am Small Anim Pract 2010;40: 651–663.

2. Häggström J, Hoglund K, Borgarelli M. An update on treatment and prognostic indicators in canine myxomatous mitral valve disease. J Small Anim Pract 2009;50:25–33.

3. Whitney JC. Observations on the effect of age on the severity of heart valve lesions in the dog. J Small Anim Pract 1974;15: 511–522.

4. Borgarelli M, Zini E, D’Agnolo G, et al. Comparison of primary mitral valve disease in German Shepherd dogs and in small breeds. J Vet Cardiol 2004;6:27–34.

5. Malik R, Hunt GB, Allan GS. Prevalence of mitral valve insuf-ficiency in cavalier King Charles spaniels. Vet Rec 1992;130: 302–303.

6. Boswood A, Gordon SG, Häggström J, et al. Longitudinal analy-sis of quality of life, clinical, radiographic, echocardiographic, and laboratory variables in dogs with preclinical myxomatous mitral valve disease receiving pimobendan or placebo: The EPIC study. J Vet Intern Med 2018;32:72–85.

7. Freeman LM, Rush JE, Milbury PE, Blumberg JB. Antioxidant status and biomarkers of oxidative stress in dogs with congestive heart failure. J Vet Intern Med 2005;19:537–541.

8. Verk B, Nemec Svete A, Salobir J, Rezar V, Domanjko Petric A. Markers of oxidative stress in dogs with heart failure. J Vet Diagn Invest 2017;29:636–644.

9. Kulka M, Garncarz M, Parzeniecka-Jaworska M, Klucinski W. Serum paraoxonase 1 activity and lipid metabolism param-eter changes in Dachshunds with chronic mitral valve disease. Assessment of its diagnostic usefulness. Pol J Vet Sci 2017;20: 723–729.

10. Svete AN, Verk B, Seliškar A, Tomsic K, Križman PJ, Petric AD. Plasma coenzyme Q10 concentration, antioxidant status, and serum N-terminal pro-brain natriuretic peptide concentration in dogs with various cardiovascular diseases and the effect of cardiac treatment on measured variables. Am J Vet Res 2017;78: 447–457.

11. Reimann MJ, Häggström J, Møller JE, Lykkesfeldt J, Falk T, Olsen LH. Markers of oxidative stress in dogs with myxomatous mitral valve disease are influenced by sex, neuter status, and serum cholesterol concentration. J Vet Intern Med 2017;31:295–302.

12. Gutteridge JMC, Halliwell B. Mini-review: Oxidative stress, redox stress or redox success? Biochem Biophys Res Commun 2018;502:183–186.



13. Lei XG, Zhu JH, Cheng WH, et al. Paradoxical roles of anti-oxidant enzymes: Basic mechanisms and health implications. Physiol Rev 2016;96:307–364.

14. Atkins C, Bonagura J, Ettinger S, et al. Guidelines for the diag-nosis and treatment of canine chronic valvular heart disease. J Vet Intern Med 2009;23:1142–1150.

15. Garncarz M, Parzeniecka-Jaworska M, Jank M, Łój M. A ret-rospective study of clinical signs and epidemiology of chronic valve disease in a group of 207 Dachshunds in Poland. Acta Vet Scand 2013;55:52.

16. Prasad K, Gupta JB, Kalra J, Lee P, Mantha SV, Bharadwaj B. Oxidative stress as a mechanism of cardiac failure in chronic volume overload in canine model. J Mol Cell Cardiol 1996;28: 375–385.

17. Freeman LM, Brown DJ, Rush JE. Antioxidant status in dogs with idiopathic dilated cardiomyopathy. J Nutr 1998;128:2768S–2770S.

18. Alcaino H, Greig D, Chiong M, et al. Serum uric acid correlates with extracellular superoxide dismutase activity in patients with chronic heart failure. Eur J Heart Fail 2008;10:646–651.

19. Leff JA, Oppegard MA, Terada LS, McCarty EC, Repine JE. Human serum catalase decreases endothelial cell injury from hydrogen peroxide. J Appl Physiol (1985) 1991;71:1903–1906.

20. Flores-Mateo G, Carrillo-Santisteve P, Elosua R, et al. Antioxidant enzyme activity and coronary heart disease: Meta-analyses of observational studies. Am J Epidemiol 2009;170:135–147.

21. Serdar Z, Aslan K, Dirican M, Sarandöl E, Yesilbursa D, Serdar A. Lipid and protein oxidation and antioxidant status in patients with angiographically proven coronary artery disease. Clin Biochem 2006;39:794–803.

22. Zuzak E, Horecka A, Kielczykowska M, et al. Glutathione level and glutathione reductase activity in serum of coronary heart disease patients. J Pre-Clin Clin Res 2017;11:103–105.

23. Sapira V, Cojocaru IM, Socoliuc G, et al. Glutathione reductase levels in patients with unstable angina. Rom J Intern Med 2011;49:197–201.

24. Rubio CP, Hernández-Ruiz J, Martinez-Subiela S, Tvarijonaviciute A, Ceron JJ. Spectrophotometric assays for total antioxi-

dant capacity (TAC) in dog serum: An update. BMC Vet Res 2016;12:166.

25. Apak R, Güçlü K, Ozyürek M, Karademir SE, Altun M. Total antioxidant capacity assay of human serum using copper (II)-neocuproine as chromogenic oxidant: The CUPRAC method. Free Radic Res 2005;39:949–961.

26. Tomsic K, Seliškar A, Lukanc B, Nemec Svete A. Plasma total antioxidant capacity and activities of blood glutathione per-oxidase and superoxide dismutase determined in healthy dogs by using commercially available kits. Acta Veterinaria-Beograd 2016;66,534–548.

27. Hetyey CS, Manczur F, Dudás-Györki Z, et al. Plasma antioxi-dant capacity in dogs with naturally occurring heart diseases. J Vet Med A Physiol Pathol Clin Med 2007;54:36–39.

28. Demirbag R, Yilmaz R, Erel O, Gultekin U, Asci D, Elbasan Z. The relationship between potency of oxidative stress and sever-ity of dilated cardiomyopathy. Can J Cardiol 2005;21:851–855.

29. Chandran G, Sirajudeen KN, Yusoff NS, Swamy M, Samarendra MS. Effect of the antihypertensive drug enalapril on oxidative stress markers and antioxidant enzymes in kidney of spontane-ously hypertensive rat. Oxid Med Cell Longev 2014;2014:608512.

30. Queisser N, Happ K, Link S, et al. Aldosterone induces fibrosis, oxidative stress and DNA damage in livers of male rats indepen-dent of blood pressure changes. Toxicol Appl Pharmacol 2014; 280:399–407.

31. Nakamura K, Murakami M, Miura D, et al. Beta-blockers and oxidative stress in patients with heart failure. Pharmaceuticals (Basel) 2011;4:1088–1100.

32. Anjaneyulu M, Chopra K. Diltiazem attenuates oxidative stress in diabetic rats. Ren Fail 2005;27:335–344.

33. Weglicki WB, Mak IT, Simìc MG. Mechanisms of cardiovascular drugs as antioxidants. J Mol Cell Cardiol 1990;22:1199–1208.

34. Cornell CC, Kittleson MD, Della Torre P, et al. Allometric scaling of M-mode cardiac measurements in normal adult dogs. J Vet Intern Med 2004;18:311–321.


Short Communication Communication brève


Serum proteins serve many different functions, including the transport of lipids, hormones, vitamins, and minerals for the regu-lation of cellular activity (1). They are also involved in regulating the immune response, inflammation, protecting against infection, and repairing damaged tissue. The profiles of serum proteins can be altered by several factors, including the acute phase response, which is an early defense system activated by trauma, infection, and inflammation (2). The quantification of serum proteins is, therefore, a useful tool for the diagnosis, prognosis, and monitoring of diseases that involve changes in the concentrations of serum proteins. Serum protein electrophoresis (SPE) separates proteins by size and electrical charge that are broadly fractionized as albumin and alpha, beta, and gamma (a, ß, and g) globulins. These fractions are known to contain various proteins that function as part of the acute phase and acquired immune responses (1).

Acute phase proteins (APPs) are blood proteins that are produced by the acute phase response to infection, inflammation, or trauma.

They play a role in both a pro- and anti-inflammatory effect on balance between the 2 functions (3). Blood concentrations of APPs have been used as diagnostic and prognostic markers in humans and animals (4). Acute phase proteins (APPs) are classified into categories according to the severity of the acute phase response, corresponding to a major, moderate, and minor response.

Major APPs in dogs include C-reactive protein (CRP) and serum amyloid A (SAA) (3,5). C-reactive protein (CRP) plays a role in the induction of cytokines, inhibition of chemotaxis, and modulation of neutrophil function (3), and is frequently used as a marker for systemic inflammation. Its serum concentration can increase rapidly up to 100-fold as part of the response to a number of infectious and inflammatory diseases in dogs (3,6). Serum amyloid (SAA) also induces chemotactic recruitment of inflammatory cells to sites of inflammation and increases during the acute phase response in dogs, as previously reported for infectious diseases, such as parvo-virus infection and leishmaniosis and inflammatory diseases (3–5).

Alterations in serum protein electrophoresis profiles during the acute phase response in dogs with acute pancreatitisJi-Seon Yoon, Suhee Kim, Jin-Hee Kang, Jinho Park*, DoHyeon Yu*

A b s t r a c tThe quantification of serum proteins is a useful tool for diagnosing and monitoring various diseases that involve changes in the concentrations of these proteins. As canine acute pancreatitis (AP) accompanies the systemic inflammatory response syndrome, serum proteins such as C-reactive protein (CRP) have been used as inflammatory markers for dogs with AP. The goal of this study was to investigate the overall profiles of serum proteins by serum protein electrophoresis (SPE) and to determine the concentration of acute phase proteins (APPs) in dogs with AP in order to better understand serum protein profiles as diagnostic markers in these dogs. Decreased levels of albumin and increased levels of alpha-2 globulin were observed in dogs with AP by SPE. Among APPs, elevated concentrations of CRP, serum amyloid A (SAA), and haptoglobin were detected. The concentration of SAA was positively correlated with that of CRP, which suggests that SAA could be a sensitive marker of inflammation in dogs with AP, similar to CRP.

R é s u m éLa quantification des protéines sériques est un outil utile pour diagnostiquer et suivre différentes pathologies qui impliquent des changements dans les concentrations de ces protéines. Comme la pancréatite aiguë (AP) accompagne le syndrome de réponse inflammatoire systémique, les protéines sériques telles que la protéine C-réactive (CRP) ont été utilisées comme marqueurs d’inflammation chez les chiens avec AP. L’objectif de la présente étude était d’examiner les profils globaux des protéines sériques par électrophorèse des protéines sériques (SPE) et de déterminer les concentrations des protéines de phase aiguë (APPs) chez les chiens avec AP afin de mieux comprendre les profils de protéines sériques comme marqueurs diagnostiques chez ces chiens. Des niveaux diminués d’albumine et des niveaux augmentés de globuline alpha-2 furent observés par SPE chez des chiens avec AP. Parmi les APPs, des concentrations élevées de CRP, d’amyloïde sérique A (SAA), et d’haptoglobine furent détectées. La concentration de SAA était corrélée positivement avec celle de CRP, ce qui suggère que SAA pourrait être un marqueur sensible d’inflammation chez les chiens avec AP, de manière similaire à la CRP.


College of Veterinary Medicine, Chonbuk National University, Iksan 54596, Korea (Yoon, Kang, Park); Gyeongsang National University Hospital, Jinju 52727, Korea (Kim); College of Veterinary Medicine, Gyeongsang National University, Jinju 52828, Korea (Yu).

Address all correspondence to Dr. DoHyeon Yu; telephone: 182-55-772-2368; fax: 182-55-772-2308; e-mail: [email protected]

*These authors contribute equally in this study.

Received January 25, 2019. Accepted April 19, 2019.



Moderate APPs in dogs include haptoglobin and acid glycopro-tein. Haptoglobin acts as a natural antagonist for receptor-ligand activation of the immune system and also inhibits granulocyte chemotaxis and phagocytosis (3). Elevations in haptoglobin levels have been reported in dogs with Cushing disease, leishmaniosis, surgical trauma, and inflammatory diseases such as pyometra and pancreatitis (4,5).

Acute pancreatitis (AP) is a relatively common disorder in dogs, occurring predominantly in middle-aged, obese, female dogs. As the clinical condition of canine AP varies from mild to severe, it is important to determine whether the condition of a patient is severe or mild for an accurate diagnosis and appropriate treatment (7). As acute pancreatitis (AP) is accompanied by sudden inflammation of the pancreas, with the adjunctive tissue involved to varying extents (7), there have been attempts to determine whether inflammatory markers such as CRP can be used as diagnostic and prognostic mark-ers for canine pancreatitis. Few studies, however, have investigated the overall profiles of serum proteins in dogs with pancreatitis (1).

The goal of the present study was to investigate the concentrations of overall serum proteins in dogs with AP in order to better under-stand serum protein profiles as diagnostic markers in these dogs. We identified serum protein fractions differentially represented in SPEs of dogs with AP and further analyzed the concentrations of single APPs, including CRP, SAA, and haptoglobin.

The study was conducted on the serum samples of 13 dogs (6 females and 7 males, 2 to 12 y old) diagnosed with AP by physi-cal examinations (with clinical signs of anorexia, diarrhea, and vomiting), lab analyses, including the SNAP canine pancreas-specific lipase (cPL) kit (IDEXX, Westbrook, Maine, USA), and diagnostic imaging. Four Maltese, 3 Yorkshire terriers, 3 Shih Tzus, 1 Pomeranian, and 2 mongrels were enrolled in the study. Increases of canine pancreas-specific lipase were observed in all enrolled dogs. In addition, 6 beagles (3 females and 3 males, 4 to 8 y old) were enrolled as control dogs. At initial presentation to the hospital, 3-mL blood samples were collected by direct jugular venipuncture, placed

in a tube, and centrifuged at 1300 3 g for 10 min. Serum samples were then frozen immediately at 280°C until assayed.

Total proteins were evaluated by chemistry analyzer (FUJI DRI-CHEM 7000i; Fujifilm, Tokyo, Japan) and SPEs were conducted by agarose gel electrophoresis (Hydrasys2; SEBIA, Lisses, France) with a protein electrophoresis reagent kit [Hydragel protein(e) 15/30; SEBIA] following the manufacturer’s instructions; the fractions were subsequently identified by electrophoretograms. The results of the serum protein electrophoresis gels were reviewed and interpreted by a laboratory expert.

As the fractions in SPE contain various proteins including APPs, the concentrations of single APPs, such as CRP, SAA, and hapto-globin, were also measured by commercial colorimetric kits that are validated for dogs (8). Briefly, CRP and SAA were measured using an immunoassay kit (Tridelta Development, Kildare, Ireland) according to the manufacturer’s instructions. In addition, hap-toglobin was measured by kits for the peroxidase activity of the haptoglobin-hemoglobin complex (Tridelta Development) according to the manufacturer’s instructions. The absorbance of samples was measured on a microtiter plate reader (BioTek, Winooski, Vermont, USA) at 450 nm using 630 nm as reference. The values of dogs with AP were compared to those of control dogs using Student’s t-test, with a P-value of , 0.05 considered statistically significant. In order to compare each marker, correlations between CRP, SAA, and hap-toglobin were also evaluated by Pearson’s correlation. Correlation coefficients (r) of , 20.2 and . 0.2 were considered as displaying weak negative and positive correlations, respectively, and r of , 20.4 and . 0.4 were considered to display significant negative and posi-tive correlations, respectively.

By SPE, 5 fractions of albumin and a1, a2, ß, and g globulins were identified in both control dogs and dogs with AP. The percentage of the fractions in SPE was then multiplied by total protein concen-tration to quantify values for each fraction (Figure 1). Total protein levels were significantly decreased in dogs with AP. In addition, the levels of albumin were significantly decreased and the levels

Figure 1. Concentrations of serum proteins separated by serum protein electrophoresis (SPE). Total protein levels were significantly decreased in dogs with acute pancreatitis (AP). In addition, significantly decreased levels of albumin and significantly increased levels of a2 globulin were observed in dogs with AP. The albumin/globulin (A/G) ratio was also significantly decreased in dogs with AP (**P , 0.01).

Con

cent

ratio

ns o

f ser

um p

rote

ins

(g/d

L) 7

6

5

4

3

2

1

0

Controls

Dogs with AP

Total Albumin a-1 a-2 b- g A/G protein Globulin Globulin Globulin Globulin ratio



of a2 globulin were significantly increased in dogs with AP. The albumin/globulin ratio (A/G ratio) was also significantly decreased in dogs with AP (Figure 1).

In addition, in order to determine which acute phase proteins were elevated in dogs with AP, the amounts of single APPs were investigated by colorimetric kit. The quantities of CRP, haptoglobin, and SAA in the serum of healthy dogs and dogs with AP are shown in Figure 2. Significant differences in the concentrations of CRP, SAA, and haptoglobin were observed in dogs with AP. C-reactive protein (CRP), which is known as a major APP in dogs, was elevated an aver-age of 20-fold compared to control dogs (Figure 2A). Another major APP, SAA, was elevated in all enrolled dogs with AP, displaying an average elevation of 50-fold compared to controls (Figure 2B). The moderate APP, haptoglobin, was elevated 2-fold in dogs with AP (Figure 2C). Furthermore, for the comparison of each marker, the correlations between the concentrations of CRP, SAA, and hapto-globin were also investigated. A positive correlation was observed between CRP and SAA (Figure 2D, P = 0.013), while haptoglobin was negatively correlated with CRP (Figure 2E, P = 0.027). A weak negative correlation was also observed between haptoglobin and SAA (Figure 2F, P = 0.335).

In this study, the concentration of total protein was significantly decreased in dogs with AP. Decreased albumin synthesis (negative

acute phase protein) might contribute to decreased total protein. In addition, increasing protein loss due to gastrointestinal tract or proteinuria because of type-III hypersensitivity glomerulonephritis in AP might be associated with low total protein. The concentration of serum albumin is usually lower in dogs with AP (7), as was also observed in the present study. Decreased albumin levels in dogs with AP are associated with numerous factors, including the preferential synthesis of APP in the liver during the acute phase response (3). When an acute phase response occurs, the liver preferentially syn-thesizes positive APPs, whereas the production of other proteins, including albumin, is reduced.

By SPE analysis, an elevation of a2 globulin levels was also observed in dogs with AP. As a2 globulin includes haptoglobin, ceru-loplasmin, and a2 macroglobulin (5), elevation of haptoglobin con-centration might be one of the factors that contributed to increased a2 globulin concentration in dogs with AP. Meanwhile, there was no significant difference in ß globulin, which includes SAA and CRP. It has been reported that increased globulin fractions are observed in SPE when concentrations of APPs such as haptoglobin are increased in the serum (5). Acute phase proteins (APPs) with lower concentra-tions, such as SAA, may not induce increases in globulin fractions in SPE, even though they are significantly increased in the serum (5). Therefore, SPE analysis in dogs with AP represented characteristic

Figure 2. Concentrations of acute phase proteins and correlations for each protein. Significantly increased concentrations of C-reactive protein (CRP) (A) and serum amyloid A (SAA) (B) and moderately increased haptoglobin (C) were observed in dogs with acute pancreatitis (AP) (**P , 0.01). In addi-tion, the concentrations of CRP were positively correlated with those of SAA (D). However, a significant negative correlation with CRP (E) and a weak negative correlation with SAA (F) were observed for the concentrations of haptoglobin in dogs with AP.

Con

cent

ratio

ns o

f CR

P (m

g/L)

Con

cent

ratio

ns o

f SA

A (m

g/L)

Con

cent

ratio

ns o

f SA

A (m

g/L)

Con

cent

ratio

ns o

f hap

togl

obin

(mg/

dL)

Con

cent

ratio

ns o

f hap

togl

obin

(mg/

dL)

Con

cent

ratio

ns o

f hap

togl

obin

(mg/

dL)

150

100

50

0

200

150

100

50

0

200

150

100

50

0

1500

1000

500

0

1600

1200

800

400

0

1500

1000

500

0

CRP SAA Haptoglobin

r = 0.6664 r = 20.6119 r = 20.2904

Control dogs Dogs with AP

0 50 100 150 0 50 100 150 0 50 100 150 200

Concentrations of CRP (mg/L) Concentrations of CRP (mg/L) Concentrations of SAA (mg/L)

[A]

[D]

[B]

[E]

[C]

[F]

Control dogs Dogs with AP Control dogs Dogs with AP



features in acute phase response, including decreased levels of albumin and increased levels of a2 globulin, although it may be less sensitive in detecting elevations of specific APPs.

In the present study, serum CRP levels were increased up to 20-fold in dogs with pancreatitis. Increased CRP levels in dogs with pancreatitis have been well-described and CRP has been used as a sensitive marker of systemic inflammation in dogs (7,9). However, there are relatively few studies on SAA levels in dogs with pancre-atitis (10). It has been reported in humans that SAAs are potentially comparable to CRPs as biomarkers for systemic inflammation (11). In dogs, SAAs were found to be limited as a biomarker for acute phase response, as they were measured with time-consuming methods such as enzyme-linked immunoassays (ELISA) (10), while CRPs could be measured by various point-of-care tests.

A recent study developed a routine measurement of canine SAA using latex agglutination turbidimetric immunoassay (LAT) (10,12) and revealed that significantly higher concentrations of SAA were detected in dogs with systemic inflammation, displaying positive correlation with the concentrations of CRP (12). Both markers showed comparable diagnostic capacity and excellent agreement in clinical classification, indicating that both SAA and CRP were useful diagnostic markers of systemic inflammation in dogs (12). In accordance with that, the present study revealed that increased SAA levels in all enrolled dogs with AP displayed an average elevation of 50-fold, which was positively correlated with CRP concentrations. Pearson’s correlation coefficient between canine SAA and CRP was also comparable to those of a previous study (12). These findings therefore suggest that SAA, in addition to CRP, could be a sensitive marker for the detection of inflammation in dogs with AP. It would be expected that SAA could be an alternative to CRP for evaluating the acute phase response of dogs with pancreatitis in future routine clinical practice.

Haptoglobin, a moderate APP, also showed increased levels in dogs with AP compared to controls. The elevation of haptoglobin levels in canine pancreatitis has been reported previously (13,14). Interestingly, the concentrations of haptoglobin were negatively correlated with the concentrations of CRP and SAA in this study. Previous stud-ies that experimentally induced pancreatitis in dogs reported that haptoglobin levels increased slowly compared to CRP levels (9,14). Plasma CRP levels were significantly increased compared to basal values after 3 h of ligation of the pancreatic ducts and peak values were observed after 24 h. The CRP concentrations slowly declined thereafter, but remained high in the plasma until the 96-hour time points (9). Significantly increased levels of haptoglobin were detected, however, after 48 h and peak levels were observed 96 h after liga-tion of the pancreatic ducts (14). Therefore, if the blood samples in this study were collected after CRP and SAA levels had peaked, the latter declined slowly thereafter. Concurrently, the concentration of haptoglobin began to increase, which may manifest as a negative cor-relation between concentrations of haptoglobin and those of CRP and SAA. Therefore, CRP and SAA are earlier markers for inflammation in dogs with AP, while haptoglobin increases only slowly.

Pancreatitis needs to be diagnosed using a combination of inter-pretation of clinical signs in the context of laboratory findings and imaging. While simply measuring serum protein as an indicator of pain is not enough, it may be helpful for monitoring disease

progression, response to treatment, and prognosis (7). One of the potential limitations of our study was that concentration of APPs was only measured at its initial presentation. The changes in APP concentrations after dogs have been treated needs to be further investigated to better understand the use of APPs as a prognostic marker in dogs with AP.

Overall, the analysis of protein profiles in inflammatory disease could be a useful tool for diagnosing the disease and predicting prog-nosis. The SPE pattern of dogs with AP revealed the characteristic features of acute phase response and analysis of APP concentration revealed increased concentrations of CRP, SAA, and haptoglobin. In particular, concentration of SAA in dogs with AP was significantly increased and showed positive correlation with that of CRP. These findings suggest that SAA could be a sensitive marker for inflam-mation in dogs with AP, similar to CRP. Future studies investigating SAAs in disease states of differing severities and responses to treat-ment will further increase our understanding of SAAs as diagnostic and prognostic markers in dogs with AP.

A c k n o w l e d g m e n t sThis research was supported by the Basic Science Research

Program of the National Research Foundation of Korea (NRF) and funded by the Ministry of Science, ICT and Future Planning (NRF-2017R1D1A1B03034904) and the Cooperative Research Program for Agriculture Science and Technology Development (PJ01284305), Rural Development Administration.

Re f e r e n c e s1. Tappin SW, Taylor SS, Tasker S, Dodkin SJ, Papasouliotis K,

Murphy KF. Serum protein electrophoresis in 147 dogs. Vet Rec 2011;168:456.

2. Cray C, Zaias J, Altman NH. Acute phase response in animals: A review. Comp Med 2009;59:517–526.

3. Ceron JJ, Eckersall PD, Martýnez-Subiela S. Acute phase proteins in dogs and cats: Current knowledge and future perspectives. Vet Clin Pathol 2005;34:85–99.

4. Tothova C, Nagy O, Kovac G. Serum proteins and their diag-nostic utility in veterinary medicine: A review. Vet Med 2016;61: 475–496.

5. Eckersall PD, Bell R. Acute phase proteins: Biomarkers of infec-tion and inflammation in veterinary medicine. Vet J 2010;185: 23–27.

6. Nakamura M, Takahashi M, Ohno K, et al. C-reactive protein concentration in dogs with various diseases. J Vet Med Sci 2008; 70:127–131.

7. Sato T, Ohno K, Tamamoto T, et al. Assessment of severity and changes in C-reactive protein concentration and various bio-markers in dogs with pancreatitis. J Vet Med Sci 2017;79:35–40.

8. Jitpean S, Holst BS, Höglund OV, et al. Serum insulin-like growth factor-I, iron, C-reactive protein, and serum amyloid A for pre-diction of outcome in dogs with pyometra. Theriogenology 2014; 82:43–48.

9. Lazarov L, Georgieva TM, Simeonova G, Zapryanova D, Nikolov J, Simeonov R. Markers of inflammation in experimentally



induced pancreatitis in dogs (Part I): C-reactive protein and white blood cell counts. Revue Méd Vét 2011;162:118–122.

10. Christensen MB, Langhorn R, Goddard A, et al. Canine serum amyloid A (SAA) measured by automated latex agglutination turbidimetry is useful for routine sensitive and specific detection of systemic inflammation in a general clinical setting. J Vet Med Sci 2013;75:459–466.

11. Lange U, Boss B, Teichmann J, Klör HU, Neeck G. Serum amy-loid A — An indicator of inflammation in ankylosing spondylitis. Rheumatol Int 2000;19:119–122.

12. Christensen MB, Langhorn R, Goddard A, et al. Comparison of serum amyloid A and C-reactive protein as diagnostic markers of systemic inflammation in dogs. Can Vet J 2014;55:161–168.

13. Feldman BF, Attix EA, Strombeck DR, O’Neill S. Biochemical and coagulation changes in a canine model of acute necrotizing pancreatitis. Am J Vet Res 1981;42:805–809.

14. Georgieva TM, Lazarov L, Simeonova G, Zapryanova D, Goranov N, Nikolov J. Markers of inflammation in experimen-tally induced pancreatitis in dogs (Part II): Correlation between clinical parameters and haptoglobin. Revue Méd Vét 2011;162: 72–75.


Short Communication Communication brève


Dairy cows often experience negative energy balance during the calving transition, and mobilize a substantial amount of body fat in early lactation. Excessive uptake of non-esterified fatty acids (NEFA) by the liver results in incomplete oxidation of NEFA and the formation of b-hydroxybutyrate (BHBA) (1). Oxidation of NEFA also leads to lipid peroxidation, and greater oxidative stress during the calving transition period is associated with health disorders and poor productivity (2). Paraoxonase-1 (PON1), synthesized in the liver, is an enzyme associated with high-density lipoprotein (HDL) and plays a critical role in protecting lipoproteins from oxidative stress (3). It has been reported that serum PON1 activity is reduced during the calving transition period (4) and in cows with fatty liver (5). Oxidative stress is also known to facilitate the production of malondialdehyde (MDA), an indicator of lipid peroxidation, in the mammary gland (6). However, the relationship between serum PON1 and MDA is unclear for ketotic dairy cows in the early lactat-

ing stage. To the authors’ knowledge, there is also no report that has evaluated serum PON1 and MDA in both tail and mammary veins. Therefore, the primary objective of this study was to investigate the association between ketonemia and PON1, MDA, and other blood serum components related.

Forty-two Holstein dairy cows on commercial dairy farms in the Saroma area of Hokkaido that had low feed intake and for which sampling from both tail and mammary veins was possible before treatment by a single veterinarian, were used in this study. The sampling was carried out from June 2013 through September 2014. Blood samples collected from both veins were immediately aliquoted into a sodium heparin tube (for BHBA and NEFA analyses), a sodium fluoride tube (for glucose analysis), and a plain tube (for the other analyses). Plasma and serum were obtained by centrifugation (2000 3 g for 20 min at 4°C) and stored at 240°C until the analysis. Cows were divided into HIGH ketonemia ($ 1.2 mM) and LOW

Serum paraoxonase-1 activity in tail and mammary veins of ketotic dairy cows

Rika Fukumori, Hanan K. Elsayed, Masahito Oba, Yasumitsu Tachibana, Ken Nakada, Shin Oikawa

A b s t r a c tThe objective of this study was to evaluate the association between ketonemia and serum paraoxonase-1 (PON1), malondialdehyde (MDA), and other blood components in tail and mammary veins of dairy cows. Forty-two Holstein dairy cows with decreased feed intake were divided into HIGH ($ 1.2 mM; n = 31) and LOW (, 1.2 mM; n = 11) groups based on the b-hydroxybutyrate concentration in plasma collected from the tail vein. The HIGH group had a significantly greater plasma non-esterified fatty acid (NEFA) concentration, but significantly lower serum PON1 activity and phospholipid concentration, and a tendency to have a lower cholesterol ester concentration than the LOW group. Serum PON1 activity was not correlated with the MDA concentration but was positively correlated with serum concentrations of cholesterol esters and phospholipids, and negatively correlated with the plasma NEFA concentration. These results suggest that serum PON1 activity is reduced by hyperketonemia and the relevance of PON1 to MDA seems to not be direct, though it is involved.

R é s u m éL’objectif de la présente étude était d’évaluer l’association entre l’acétonémie et la paraoxonase-1 (PON1), le malondialdéhyde (MDA), et d’autres composés du sang dans les veines caudale et mammaire de vaches laitières. Quarante-deux vaches laitières de race Holstein présentant une diminution de l’ingestion d’aliments furent divisées en groupes ÉLEVÉ ($ 1,2 mM; n = 31) et BAS (, 1,2 mM; n = 11) basés sur la concentration de b-hydroxybutyrate de plasma prélevé de la veine caudale. Le groupe ÉLEVÉ avait une concentration plasmatique significativement plus grande d’acides gras non-estérifiés (NEFA), mais le sérum présentait une activité PON1 et une concentration de phospholipides significativement réduite, et une tendance à avoir une concentration d’esters de cholestérol plus faible que le groupe BAS. L’activité de PON1 sérique n’était pas corrélée avec la concentration de MDA mais était corrélée positivement avec les concentrations sériques d’esters de cholestérol et de phospholipides, et corrélée négativement avec la concentration plasmatique de NEFA. Ces résultats suggèrent que l’activité de PON1 sérique est réduite par l’hypercétonémie et la pertinence de PON1 envers MDA ne semble pas être directe, bien qu’elle semble impliquée.


Department of Veterinary Herd Health, School of Veterinary Medicine, Rakuno Gakuen University, Ebetsu, Hokkaido 069-8501, Japan (Fukumori, Nakada, Oikawa); Department of Animal Medicine, Faculty of Veterinary Medicine, Assiut University, Assiut Governate 71515, Egypt (Elsayed); Department of Agricultural, Food and Nutritional Science, University of Alberta, Edmonton, Alberta T6G 2P5 (Oba); Saroma Veterinary Clinical Center, Okhotsk Agricultural Mutual Aid Association, Saroma, Hokkaido 093-0507, Japan (Tachibana).

Address all correspondence to Dr. Shin Oikawa; e-mail: [email protected]

Received January 28, 2019. Accepted May 13, 2019.



ketonemia (, 1.2 mM) groups based on the plasma BHBA concentra-tion in the tail vein, with 31 cows [3.13 6 0.34 parity; 17.7 6 2.1 days in milk (DIM); 3.13 6 0.07 body condition score] in the HIGH group, and 11 (3.45 6 0.57 parity; 18.5 6 3.5 DIM; 3.25 6 0.12 body condition score) in the LOW group. The BHBA concentrations of the HIGH and LOW groups were 0.88 6 0.56 mM (range: 0.62 to 1.14 mM) and 4.04 6 0.33 mM (range: 1.42 to 9.43 mM), respectively. Serum MDA concentrations and PON1 activity were determined by colorimetric assays using commercially available kits (Colorimetric Assay for Lipid Peroxidation; Oxis Research, Portland, Oregon, USA and Arylesterase Assay Kit; Japan Institute for the Control of Aging, Nikken Seil, Shizuoka, Japan) as previously reported (7). Serum concentrations of acetoacetate, triglycerides, total cholesterol, free cholesterol, and phospholipids were analyzed using commercial kits (Wako Pure Chemical, Osaka, Japan) (8). The concentrations of BHBA, NEFA, and glucose were measured using an automatic analyzer (Bio Majesty JCA-BM2250; JEOL, Tokyo, Japan) with their kits (N-assay NEFA; Nittobo Medical, Tokyo, Japan, 3-OHBA-TR; Kishimoto Clinical Laboratory, Sapporo, Japan, and GLU-TR; Kishimoto Clinical Laboratory) (8). The serum concentration of cholesterol esters was estimated by subtracting the concentration of free cholesterol from that of total cholesterol. Data were analyzed by the FIT model procedure of JMP (Version 13.0 for Windows; SAS Institute, Cary, North Carolina, USA) using a model including the fixed effects of groups (HIGH and LOW), sampling site (tail vein or mammary vein), and their interactions. The relationship between PON1 and other variables was calculated using Spearman’s correla-tion coefficient. Significance was declared at P , 0.05, and tendency was discussed at P , 0.10.

The HIGH group had a greater plasma NEFA concentration (P , 0.001; Table I), but less serum PON1 activity (P = 0.049), a lower phospholipid concentration (P = 0.041), and a tendency to have a lower cholesterol ester concentration than the LOW group (P = 0.065). There were no significant differences in MDA, glucose, triglycerides, total cholesterol, and free cholesterol concentrations. The concentrations of glucose (P = 0.002) and triglycerides (P = 0.014)

were lower, and the serum concentration of MDA (P = 0.062) tended to be lower for samples collected from the mammary vein than in those from the tail vein. Serum PON1 activity was not correlated with the MDA concentration, but positively correlated with serum concentrations of cholesterol esters (r = 0.377, P , 0.001), free choles-terol (r = 0.314, P = 0.006), and phospholipids (r = 0.374, P , 0.001), and was negatively correlated with the plasma NEFA concentration (r = 20.391, P , 0.001).

We found that the HIGH group had lower serum PON1 activity and lower serum concentrations of phospholipids and cholesterol esters, which are the primary components of HDL, and that all these response variables were positively correlated. These results sug-gest that the hepatic function to synthesize PON1 and HDL might be impaired in cows with hyperketonemia. Consistent with our observations, de Campos et al (9) demonstrated that the intravenous injection of lipopolysaccharide decreased serum PON1 activity and the HDL concentration. In addition, cows with fatty liver had lower serum PON1 activity (5). We did not find a correlation between BHBA and PON1 activity, which suggests that higher BHBA might not reduce PON1 activity directly.

The knowledge obtained from the current study and previous reports suggests that liver damage associated with inflammation or excessive NEFA oxidation may reduce PON1 production in the liver. The blood cholesterol concentration follows a pattern that parallels dry matter intake in periparturient dairy cows, and it is enhanced by fat feeding (10) because the intestine is considered the main site of de novo synthesis of cholesterol in ruminants. This cholesterol is then secreted as a constituent of chylomicrons and HDL (11). Therefore, the lower serum cholesterol concentration in HIGH cows might potentially be a result of lower feed intake. Unfortunately, it was difficult to identify the factors that decreased cholesterol and HDL in HIGH cows, strictly because the field investigation could not evaluate feed intake.

In the current study, HIGH cows had a greater plasma NEFA concentration than LOW cows, but the serum MDA concentration did not significantly differ between the HIGH and LOW groups,

Table I. Plasma and serum metabolites of dairy cows (LSM 6 SEM) categorized as LOW and HIGH based on b-hydroxybutyrate (BHBA) concentration in plasma collected from the tail vein.

Tail Mammary P-value Sampling LOW HIGH LOW HIGH site Ketonemia InteractionPON1 (U/L) 115 6 13.8 93.4 6 8.43 116 6 13.2 92.3 6 8.27 0.989 0.049 0.910MDA (mM) 2.08 6 0.489 2.45 6 0.309 1.35 6 0.489 1.63 6 0.303 0.062 0.431 0.904Glucose (mg/dL) 56.6 6 4.84 50.2 6 3.03 43.2 6 4.84 37.6 6 3.03 0.002 0.140 0.920BHBA (mM) 0.88 6 0.556 4.04 6 0.331 0.69 6 0.556 3.87 6 0.331 0.696 , 0.001 0.983Acetoacetate (mM) 0.063 6 0.156 0.686 6 0.098 0.060 6 0.156 0.850 6 0.098 0.535 , 0.001 0.522NEFA (mEq/L) 0.47 6 0.186 1.35 6 0.111 0.44 6 0.186 1.05 6 0.111 0.285 , 0.001 0.363Triglycerides (mg/dL) 4.09 6 0.583 3.42 6 0.347 2.36 6 0.583 2.74 6 0.347 0.014 0.761 0.277Total cholesterol (mg/dL) 112 6 12.4 94.8 6 7.41 112 6 12.4 95.6 6 7.41 0.099 0.956 0.984Cholesterol esters (mg/dL) 90.2 6 9.45 75.5 6 5.63 90.4 6 9.45 76.0 6 5.63 0.966 0.065 0.985Free cholesterol (mg/dL) 21.9 6 3.19 19.3 6 1.90 22.1 6 3.19 19.6 6 1.90 0.929 0.337 0.984Phospholipids (md/dL) 130 6 13.3 107 6 7.92 130 6 13.3 108 6 7.92 0.942 0.041 0.988LSM — least square mean; SEM — standard error of the mean; PON1 — paraoxonase-1; MDA — malondialdehyde; NEFA — non-esterified fatty acids.



which is in contrast with previous studies that reported higher serum MDA concentrations in ketotic cows (12,13). In addition, although the serum PON1 activity was negatively correlated with the plasma NEFA concentration, it was not correlated with the serum MDA concentration in the current study, suggesting that reduced PON1 activity might not directly increase the MDA concentration in lactating dairy cows.

A novel aspect of our study was the comparisons of lipid metabo-lite concentrations between the mammary vein and the tail vein. The lower concentrations of glucose and triglycerides for blood samples collected from the mammary vein than in those from the tail vein, indicate substantial uptakes of these metabolites by the mammary gland. The extensive oxidative metabolism of high-producing cows can increase oxidative stress and the generation of MDA in the mammary gland (6) and it is possible that PON1 is inactivated in the mammary gland when it protects lipoprotein from oxidative stress (14). However, we found that serum PON1 activity did not differ by sampling location, indicating that PON1 might not be used in the mammary gland. In addition, a tendency for a lower MDA con-centration in the mammary vein than in the tail vein was observed. Bouwstra et al (15) reported that the MDA concentration in milk was much higher than that in blood and that they were positively correlated. These findings suggest that milk MDA may primarily come from uptake of blood MDA rather than MDA generation in the mammary gland. Furthermore, the lack of a relationship between PON1 and MDA seems to reflect the difference of metabolism and excretion in the mammary gland.

In conclusion, serum PON1 activity is reduced in ketotic cows, but this may not necessarily increase the serum MDA concentration in lactating dairy cows, possibly because of its uptake by the mammary gland and subsequent excretion in milk.

A c k n o w l e d g m e n tThe authors are grateful to Dr. K. Barrymore for his critical read-

ing of the manuscript.

Re f e r e n c e s1. Drackley JK, Dann HM, Douglas N, et al. Physiological and

pathological adaptations in dairy cows that may increase sus-ceptibility to periparturient diseases and disorders. Ital J Anim Sci 2005;4:323–344.

2. Abuelo A, Hernández J, Benedito JL, Castillo C. The importance of the oxidative status of dairy cattle in the periparturient period: Revisiting antioxidant supplementation. J Anim Physiol Anim Nutr (Berl) 2015;99:1003–1016.

3. Mackness MI, Durrington PN. HDL, its enzymes and its poten-tial to influence lipid peroxidation. Atherosclerosis 1995;115: 243–253.

4. Turk R, Juretic D, Geres D, Svetina A, Turk N, Flegar-Mestric Z. Influence of oxidative stress and metabolic adaptation on PON1 activity and MDA level in transition dairy cows. Anim Reprod Sci 2008;108:98–106.

5. Farid AS, Honkawa K, Fath EM, Nonaka N, Horii Y. Serum paraoxonase-1 as biomarker for improved diagnosis of fatty liver in dairy cows. BMC Vet Res 2013;9:73.

6. Kapusta A, Kuczynska B, Puppel K. Relationship between the degree of antioxidant protection and the level of malondialde-hyde in high-performance Polish Holstein-Friesian cows in peak of lactation. PLoS One 2018;13:e0193512.

7. Senoh T, Oikawa S, Nakada K, Tagami T, Iwasaki T. Increased serum malondialdehyde concentration in cows with subclinical ketosis. J Vet Med Sci 2019;81:817–820.

8. Oikawa S, Saitoh-Okumura H, Tanji M, Nakada K. Relevance of serum concentrations of non-esterified fatty acids and very low-density lipoproteins in nulli/primiparous and multiparous cows in the close-up period. J Vet Med Sci 2017;79:1656–1659.

9. de Campos FT, Rincon JAA, Acosta DAV, et al. The acute effect of intravenous lipopolysaccharide injection on serum and intrafol-licular HDL components and gene expression in granulosa cells of the bovine dominant follicle. Theriogenology 2017;89:244–249.

10. van Knegsel AT, van den Brand H, Graat EA, et al. Dietary energy source in dairy cows in early lactation: Metabolites and metabolic hormones. J Dairy Sci 2007;90:1477–1485.

11. Grummer RR, Carroll DJ. A review of lipoprotein cholesterol metabolism: Importance to ovarian function. J Anim Sci 1988; 66:3160–3173.

12. El-Deeb WM, El-Bahr SM. Biochemical markers of ketosis in dairy cows at post-paturient period: Oxidative stress biomarkers and lipid profile. Am J Biochem Mol Biol 2017;7:86–90.

13. Li Y, Ding HY, Wang XC, et al. An association between the level of oxidative stress and the concentrations of NEFA and BHBA in the plasma of ketotic dairy cows. J Anim Physiol Anim Nutr (Berl) 2016;100:844–851.

14. Aviram M, Rosenblat M, Billecke S, et al. Human serum para-oxonase (PON 1) is inactivated by oxidized low density lipo-protein and preserved by antioxidants. Free Radic Biol Med 1999;26:892–904.

15. Bouwstra RJ, Goselink RM, Dobbelaar P, Nielen M, Newbold JR, van Werven T. The relationship between oxidative damage and vitamin E concentration in blood, milk, and liver tissue from vitamin E supplemented and nonsupplemented periparturient heifers. J Dairy Sci 2008;91:977–987.


canadianveterinarians.net veterinairesaucanada.net

In collaboration with | En collaboration avec :

CVMA Convention | July 9–12, 2020CANADIAN VETERINARY MEDICAL ASSOCIATION

City Ville de

Joignez-vous à vos pairs et collègues et laissez-vous enivrer par le charme et l’hospitalité du Vieux-Québec.

Assistez aux événements de marque de l’ACMV, à la formation continue exceptionnelle et plus encore!

Join peers and colleagues and be exhilarated by old Quebec charm and hospitality.

Attend CVMA signature events, stellar continuing education, AND MUCH MORE!

Exciting! Inspiring! Motivating! Excitant! Inspirant! Motivant!

CVMA CONVENTION • CONGRÈS DE L’ACMVJuly 9-12 juillet, 2020

Registration Opens January 15, 2020 L’inscription ouvrira le 15 janvier 2020

20190708305 CMVA – Bilingual Convention Ad 8 ½ x 11.indd 1 19-09-26 9:05 AM


Supplies and services Produits et services

RENOUVELEZ VOTRE ABONNEMENT

DÈS MAINTENANT!

RENEW YOUR SUBSCRIPTION

NOW!

Four issues of the Canadian Journal of Veterinary Research for 2020 (online)

Quatre numéros de la Revue canadienne de recherche vétérinaire pour 2020

Subscription/Abonnement 2020Personal Subscription/ $185 $ (1 GST/HST as Abonnement personnel applicable; TPS/TVH en (Canada) vigueur)

Personal Subscription/ $205 $ US/É.-U. Abonnement personnel (Foreign/Étranger)

Institutional Subscription/ $255 $ (1 In Canada, GST/HST Abonnement institutionnel as applicable; Au Canada, TPS/TVH selon le cas)

Institutional Subscription/ $255 $ US/É.-U. Abonnement institutionnel (Foreign/Étranger)

Contact/Communiquez avec : Kelly Gray-Sabourin Editorial Coordinator/Coordonnatrice de la rédactionCanadian Journal of Veterinary ResearchRevue canadienne de recherche vétérinaire339 rue Booth Street Ottawa, Ontario K1R 7K1Canada

Tel./Tél. : (613) 236-1162, x117Fax/Télécopieur : (613) 236-9681E-mail/Courriel : [email protected]

w w w . c a n a d i a n v e t e r i n a r i a n s . n e t

w w w . v e t e r i n a i r e s a u c a n a d a . n e t

Tables of contentsComing events

Article abstractsClassified advertising

Subscription informationInstructions for authors

Tables des matières Événements à venirRésumés d’articles Annonces classées Information sur l’abonnement Directives à l’intention des auteurs

Canadian Journal of Veterinary Research

available online now!

Articles are now easily accessible over the Internet at

PubMed Centralwww.pubmedcentral.com

Reaching Canada’s VeterinariansGet your message into

The Canadian Veterinary JournalFor more information contact:Laima LaffitteAdvertising and Sponsorship ConsultantTel.: (613) 673-2659Fax: (613) 673-2462e-mail: [email protected]


CANADIAN JOURNAL OF VETERINARY RESEARCH / REVUE CANADIENNE DE RECHERCHE VÉTÉRINAIRE


Provided for non-commercial research and education use ... · van der Meer, Frank Wakshlag, Joseph...

Documents

Transcript of Provided for non-commercial research and education use ... · van der Meer, Frank Wakshlag, Joseph...