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

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]. CanadianCanadian JournalJournal ofof VVeterinaryeterinary RResearchesearch RevueRevue CanadienneCanadienne dede RRechercheecherche VVétérinaireétérinaire

Articles An assessment of external biosecurity on southern Ontario swine farms and its application to surveillance on a geographic level Kate Bottoms, Zvonimir Poljak, Robert Friendship, Rob Deardon, Janet Alsop, Cate Dewey ...... 241

A retrospective study on the etiological diagnoses of diarrhea in neonatal piglets in Ontario, Canada, between 2001 and 2010 Gloria Chan, Abdolvahab Farzan, Josepha DeLay, Beverly McEwen, John F . Prescott, Robert M . Friendship . . . 254

Species-specificity of equine and porcine Lawsonia intracellularis isolates in laboratory animals Francesca Sampieri, Fabio A . Vannucci, Andrew L . Allen, Nicola Pusterla, Aphroditi J . Antonopoulos, Katherine R . Ball, Julie Thompson, Patricia M . Dowling, Don L . Hamilton, Connie J . Gebhart . . . . . 261

Analgesic and motor effects of a high- volume intercoccygeal epidural injection of 0.125% or 0.0625% bupivacaine in adult cows Eva Rioja, Luis M . Rubio-Martínez, Gabrielle Monteith, Carolyn L . Kerr . . . . . 273

The use of lung biopsy to determine early lung pathology and its association with health and production outcomes in feedlot steers Brandy A . Burgess, Steve H . Hendrick, Colleen M . Pollock, Sherry J . Hannon, Sameeh M . Abutarbush, Amanda Vogstad, G . Kee Jim, Calvin W . Booker ...... 281

Reversible chemical restraint of free-range cattle with a concentrated combination of tiletamine-zolazepam, ketamine, and detomidine Michela Re, Francisco J . Blanco-Murcia, José Maria San Miguel, Ignacio A . Gómez de Segura ...... 288

Effect of storage time and temperature on the total protein concentration and electrophoretic fractions in equine serum Daniela Alberghina, Stefania Casella, Claudia Giannetto, Simona Marafioti, Giuseppe Piccione ...... 293

October/Octobre 2013, Vol. 77 No. 4 table of contents continued on back cover/ la table des matières se poursuit sur la couverture arrière FOR PERSONAL USE ONLY Established in 1937 ISSN 0830-9000

Canadian Journal of Revue canadienne de Veterinary Research recherche vétérinaire Formerly CANADIAN JOURNAL OF COMPARATIVE MEDICINE Remplace la REVUE CANADIENNE DE MÉDECINE COMPARÉE The journal publishes the results of original research in veterinary La revue publie les résultats de recherches originales en médecine and comparative medicine. Manuscripts must be as concise as pos- vétérinaire et comparative. Les manuscrits doivent être aussi brefs que sible, and the research described must represent a significant con- possible et la recherche décrite doit apporter une contribution impor- tribution to knowledge in veterinary medicine. Full-length papers, tante à la médecine vétérinaire. Les exposés complets, les communi­ short communications, and review papers are welcome. All manu- cations brèves et les comptes rendus sont bienvenus. Nous vérifier­ ons scripts will be reviewed for scientific content and editorial accuracy. le fonds ­scientifique et la forme de tous les manuscrits. Manuscripts must conform to the Instructions for Authors found Les manuscrits doivent être conformes aux directives aux auteurs on the CVMA Web site. Consultation of a recent issue of the journal qui se trouvent sur le site Web de l’ACMV. La consultation d’un to ensure that the manuscript conforms to current style is also recom- récent numéro de la revue est aussi profitable, afin de s’assurer qu’un mended. Please use the Web site (www.canadianveterinarians.net) manuscrit se conforme au style courant. Veuillez utiliser le site Web to submit articles. A publication charge of $90 per published page, (www.veterinairesaucanada.net) pour soumettre les articles. Les plus a cost of $10 per table, $15 per illustration, and $150 per color auteurs doivent acquitter des frais de 90 $ par page publiée, 10 $ illustration is payable by contributors. Instructions for ordering pour chaque tableau, 15 $ pour chaque illustration, et 150 $ pour reprints are sent with the galley proofs. Color illustrations will only chaque illustration en couleur. Un formulaire de commande de be published if the author(s) agree to pay the extra cost. tirés à part accompagne l’en­voi des épreuves d’un manuscrit. Les illustrations en couleurs ne seront publiées que si les auteurs en assument les frais supplémentaires.

Canadian Journal of Veterinary Research Abstracted and/or indexed in:/Résumé et/ou catalogué dans : Revue canadienne de recherche vétérinaire Biological Abstracts Index Veterinarius Editor — Rédacteur Chemical Abstracts Nutrition Abstracts Éva Nagy, Guelph, Ontario Current Contents (Agric. Section) Science Citation Index Associate Editor — Rédacteur adjoint EMBASE/Excerpta Medica Veterinary Bulletin Deborah M. Haines, Saskatoon, Saskatchewan Index Medicus Derwent Veterinary Assistant Editors — Assistants à la rédaction Drug File Serge Messier, Saint-Hyacinthe (Québec) Jeffrey J. Wichtel, Charlottetown, Prince Edward Island Managing Editor — Directrice de la rédaction Heather Broughton, Ottawa, Ontario Assistant Managing Editor — Directrice adjointe à la ­rédaction Stella Wheatley, Ottawa, Ontario Advertising Manager — Gérante de la publicité Laima Laffitte, Wendover, Ontario Editorial Coordinator/Coordonnatrice de la rédaction Beverley Kelly, Ottawa, Ontario

Address all correspondence to/Adresser toute correspondance à : Canadian Journal of Veterinary Research Typesetting/Typographie Revue canadienne de recherche vétérinaire AN Design Communications 339, rue Booth Street, Ottawa, Ontario K1R 7K1 www.an-design.ca Tel./Tél. : (613) 236-1162 — Fax/Télécopieur : (613) 236-9681 e-mail/Courriel : [email protected] © Canadian Veterinary Medical Association 2013. L’Association canadienne des médecins vétérinaires 2013

Subscriptions/Abonnements — 2013 Published quarterly by the Canadian Veterinary Medical Association Canada $150 $ ( applicable GST/HST; Publication trimestrielle de l’Association canadienne des TPS/TVH en vigueur) médecins vétérinaires Foreign/Étranger $165 $ US/É.-U.

Return undeliverable Canadian addresses to: 339 rue Booth Street, Ottawa, Ontario K1R 7K1, e-mail: [email protected] FOR PERSONAL USE ONLY

Article

An assessment of external biosecurity on southern Ontario swine farms and its application to surveillance on a geographic level Kate Bottoms, Zvonimir Poljak, Robert Friendship, Rob Deardon, Janet Alsop, Cate Dewey

Abstract Risk-based surveillance is becoming increasingly important in the veterinary and public health fields. It serves as a means of increasing surveillance sensitivity and improving cost-effectiveness in an increasingly resource-limited environment. Our approach for developing a tool for the risk-based geographical surveillance of contagious diseases of swine incorporates information about animal density and external biosecurity practices within swine herds in southern Ontario. The objectives of this study were to group the sample of herds into discrete biosecurity groups, to develop a map of southern Ontario that can be used as a tool in the risk-based geographical surveillance of contagious swine diseases, and to identify significant predictors of biosecurity group membership. A subset of external biosecurity variables was selected for 2-step cluster analysis and latent class analysis (LCA). It was determined that 4 was the best number of groups to describe the data, using both analytical approaches. The authors named these groups: i) high biosecurity herds that were open with respect to replacement animals; ii) high biosecurity herds that were closed with respect to replacement animals; iii) moderate biosecurity herds; and iv) low biosecurity herds. The risk map was developed using information about the geographic distribution of herds in the biosecurity groups, as well as the density of swine sites and of grower-finisher pigs in the study region. Finally, multinomial logistic regression identified heat production units (HPUs), number of incoming pig shipments per month, and herd type as significant predictors of biosecurity group membership. It was concluded that the ability to identify areas of high and low risk for disease may improve the success of surveillance and eradication projects.

Résumé La surveillance basée sur le risque devient de plus en plus importante dans les domaines de la médecine vétérinaire et de la santé publique. Elle sert comme moyen d’augmenter la sensibilité de la surveillance et d’améliorer l’efficacité des coûts dans un environnement de plus en plus limité en ressources. Notre approche pour développer un outil pour la surveillance géographique basée sur le risque de maladies contagieuses porcines incorpore des informations sur la densité animale et les pratiques de biosécurité externes dans des troupeaux porcins du sud de l’Ontario. Les objectifs de cette étude étaient de grouper l’échantillon de troupeaux en groupes distincts de biosécurité, de développer une carte du sud de l’Ontario qui pourrait être utilisée comme outil dans la surveillance géographique basée sur le risque de maladies contagieuses porcines, et d’identifier des prédicteurs significatifs d’appartenance au groupe de biosécurité. Un sous-groupe de variables de biosécurité externes fut sélectionné pour une analyse d’agrégation en 2 étapes et une analyse des classes latentes (LCA). Il a été déterminé que le nombre 4 était le meilleur nombre de groupes pour décrire les données en utilisant les deux approches analytiques. Les auteurs ont désigné ces groupes : i) troupeaux à biosécurité élevée ouverts aux animaux de remplacement; ii) troupeaux à biosécurité élevée fermés aux animaux de remplacement; iii) troupeaux à biosécurité moyenne; et iv) troupeaux à biosécurité faible. La carte de risque a été développée en utilisant l’information sur la distribution géographique des troupeaux à l’intérieur des groupes de biosécurité, ainsi que la densité des sites porcins et des porcs en engraissement-finition dans la région à l’étude. Finalement, une régression logistique multinomiale a identifié les unités de production de chaleur (HPU), le nombre mensuel d’envois entrants de porcs, et le type de troupeau comme des prédicteurs significatifs de l’appartenance à un groupe de biosécurité. Il a été conclu que la facilité à identifier les zones à risque élevé et faible de maladie pourrait améliorer le succès des projets de surveillance et d’éradication. (Traduit par Docteur Serge Messier)

Introduction are important for various purposes, including detecting pathogens, demonstrating freedom from specific pathogens, and evaluating Risk-based surveillance systems are becoming increasingly impor- the efficacy of disease control and eradication efforts (4–7). They are tant in veterinary medicine. These systems serve as a means of also important in supporting trade agreements among countries and increasing surveillance sensitivity and improving cost-effectiveness ensuring consumer confidence in animal products. Early detection in an increasingly resource-limited environment (1–3). Such systems via surveillance systems can facilitate rapid response by animal

Department of Population Medicine, Ontario Veterinary College (Bottoms, Poljak, Friendship, Dewey) and Department of Mathematics and Statistics (Deardon), University of Guelph, Guelph, Ontario; Ontario Ministry of Agriculture, Food and Rural Affairs, Guelph, Ontario (Alsop). Address all correspondence to Kate Bottoms; telephone: (519) 824-4120, ext. 54873; fax: (519) 763-3117; e-mail: [email protected] Received July 6, 2012. Accepted December 2, 2012.

2013;77:241–253 The Canadian Journal of Veterinary Research 241 FOR PERSONAL USE ONLY

health authorities and allow emerging threats to be managed in a was conducted in order to determine which demographic variables timely manner. serve as significant predictors of biosecurity group membership. Risk-based surveillance activities are targeted at subpopulations that are at increased risk for the occurrence of disease or infection Study design due to the presence of known risk factors (3,4). This type of infor- The Ontario Porcine Reproductive and Respiratory Syndrome mation has been used to make recommendations for enhancing sur- (PRRS) Surveillance Survey was designed to assess risk factors asso- veillance systems (2,8,9). This risk-based concept can be applied at ciated with PRRSV-positivity and to describe the distribution of this different levels and risk factors can include animal-level, herd-level, virus on Ontario swine farms. The site-level portion of the survey and geographic or neighborhood characteristics. One recent study consists of 19 closed- and open-ended questions, including informa- developed a risk-based map for the incursion of H5N1 avian influ- tion about demographics and the biosecurity procedures used. In enza virus into the domestic poultry population in Great Britain (2). most cases, the same question was asked for all production classes, The risk map was based on holding-level factors and information including weaners, feeders, gilts, boars, market animals, and culls. about the abundance of wild birds known to carry the virus. The sampling strategy for the study has been reported in previ- The method in the current study incorporates information about ously published literature (13). In brief, herds in the original study external biosecurity practices and density of swine in order to iden- were included through 2 mechanisms. Sites that were PRRSV- tify high-risk geographical areas. Animal density is an important risk positive and were willing to participate in the study were obtained factor for the spread of disease, with the risk of disease transmission from the Animal Health Laboratory (AHL) at the University of increasing in areas of high density (9–11). The spread of disease Guelph, which is the largest diagnostic laboratory in the province. among farms and within regions can also be linked to the external Herds that were PRRSV-negative were obtained from veterinar- biosecurity practices that are in place. Farms can be defined in terms ians who routinely submitted samples to the AHL, as well as from of the risk of pathogen incursion, in part, on the basis of biosecurity members of the Ontario Association of Swine Veterinarians (OASV). practices employed on-farm. Thus, biosecurity information can be Staff at the Department of Population Medicine at the University of used as a component of risk-based surveillance. Guelph contacted herd owners and conducted a telephone interview The primary objective of this study was to develop a map of that took approximately 30 min to complete. The interviewer filled southern Ontario that can be used as a tool in risk-based geographi- out hard copies of the surveys and data were subsequently entered cal surveillance of contagious swine diseases. This was accomplished into a hierarchical database (Microsoft Access 2003; Microsoft, through several phases. Firstly, information about external biosecu- Redmond, Washington, USA). The data set consisted of 527 herds rity practices on a sample of swine herds was used to determine the from 392 unique sites. In some cases, sites had been interviewed at best number of groups to describe the data. The identified groups multiple points in time. For the purposes of this study, only the most were described in terms of their major characteristics and the vari- recent entry for each site was used. Boar stud and gilt pool sites were ables of importance in differentiating among them. To develop the excluded, as they are not representative of commercial swine opera- risk map, the geographic location of farms in each of the identified tions in southern Ontario. This reduced the sample size to 377 herds. groups was combined with information about herd and animal density in the study region, as well as with data from recent work Variables selected for analysis by this research group that described external biosecurity practices Two subsets of questions pertaining to demographic information, on a different sample of swine herds. Finally, a multivariable mul- geographic location, and external biosecurity practices were retained tinomial logistic regression model was used to determine which for analysis: 1 for cluster analysis and the other for regression demographic variables serve as significant predictors of biosecurity analysis. For the cluster analysis portion, 22 categorical variables group membership. pertaining to external biosecurity practices were selected (Table I). These variables addressed trucking practices and entrance sanita- Materials and methods tion procedures. Variables regarding demographic information, geographic location, and herd density were excluded from cluster Data for this study were obtained from a questionnaire that was analysis with the rationale that the identified biosecurity groups designed to investigate factors associated with porcine reproductive should reflect variables that can be changed at the herd level. Since and respiratory syndrome virus (PRRSV) on Ontario swine farms. this survey was designed to assess risk factors associated with A subset of variables pertaining to external biosecurity practices PRRSV-positivity, information about the PRRSV-status of the sample was selected for 2-step cluster analysis and latent class analysis farms was also retained. (LCA) in order to determine the best number of groups to describe Some variables concerning the frequency with which specific the data. Using kernel smoothing, maps were created in order to practices were implemented on-farm were modified before cluster visualize the spatial distribution of herds belonging to each of these analysis. These questions were all close-ended and, with the excep- groups. A risk map for the spread of contagious swine diseases was tion of 3 variables, the potential responses were i) always, ii) usually, developed using information about the spatial distribution of the iii) occasionally, iv) never, v) unknown, or vi) not applicable. The biosecurity groups identified here, in combination with the density differences between these response categories are subjective and they of pigs and swine sites in the study region as well as the distribution were condensed into 4 categories in order to simplify the conclu- of biosecurity groups identified in previous work conducted by this sions. Responses of “always” or “usually” became i) frequently and research group (12). Finally, multinomial logistic regression analysis responses of “occasionally” or “never” became ii) rarely. Responses

242 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Table I. External biosecurity variables used in 2-step cluster analysis and latent class analysis. Variables are presented in decreasing order of importance, as identified by 2-step cluster analysis, N = 377 All herds HB(open) HB(closed) MB LB Variable Responses (%) (%) (%) (%) (%) Are trucks that pick up culls cleaned and Frequently 39.3 70.9 35.4 0 27.7 disinfected before they arrive at your barn? Rarely 28.6 25.8 53.8 0 52.3 Unknown 6.6 3.3 10.8 0 20 Not applicable 25.5 0 0 100 0 Are trucks that deliver pigsa to your farm cleaned Frequently 55.2 76.2 0 66.7 44.6 and disinfected before arriving at your farm? Rarely 17.5 19.2 0 27.1 16.9 Unknown 10.3 4.6 3.1 5.2 38.5 Not applicable 17 0 96.9 1 0 Do trucks that deliver pigsa visit another farm Frequently 7.2 6 0 7.3 16.9 immediately before yours? Rarely 69.8 90.7 0 90.6 60 Unknown 6.1 3.3 3.1 1 23.1 Not applicable 17 0 96.9 1 0 Does the truck driver enter the barn to assist with Frequently 24.9 28.5 0 42.7 15.4 the unloading of pigsa? Rarely 57.6 71.5 0 56.3 84.6 Unknown 1.1 0 6.2 0 0 Not applicable 16.4 0 93.8 1 0 Do trucks that pick up culls ever visit farms other Yes 49.3 62.3 64.6 0 76.9 than the destination farm and yours? No 24.7 36.4 35.4 0 23.1 Unknown 0.5 1.3 0 0 0 Not applicable 25.5 0 0 100 0 Do trucks that deliver pigsa ever visit farms other Yes 50.9 56.3 0 59.4 76.9 than the source farm and yours? No 28.6 42.4 0 37.5 12.3 Unknown 2.4 1.3 0 2.1 7.7 Not applicable 18 0 100 1 3.1 Do trucks that pick up culls arrive empty? Frequently 63.1 88.7 81.5 0 78.5 Rarely 11.1 10.6 18.5 0 21.5 Unknown 0.3 0 0 1 0 Not applicable 25.5 0.7 0 99 0 When picking up culls, does the truck driver enter Frequently 18.8 19.9 36.9 0 26.2 the barn to assist with the loading of pigs? Rarely 54.9 78.1 63.1 0 73.8 Unknown 1.1 1.3 0 2.1 0 Not applicable 25.2 0.7 0 97.9 0 Are trucks that pick up culls driven to the barn Frequently 64.5 85.4 86.2 0 89.2 door? Rarely 9.3 13.2 13.8 0 9.2 Unknown 1.3 1.3 0 2.1 1.5 Not applicable 24.9 0 0 97.9 0 Are employees required to wear boots and Frequently 73.5 100 63.1 79.2 13.8 coveralls? Rarely 1.6 0 3.1 1 4.6 Unknown 0 0 0 0 0 Not applicable 24.9 0 33.8 19.8 81.5 Are employees required to wash hands and/or Frequently 44.6 70.2 27.7 44.8 1.5 shower before entering the barn? Rarely 30.2 29.8 38.5 34.4 16.9 Unknown 0.5 0 0 1 1.5 Not applicable 24.7 0 33.8 19.8 80 On the day of the visit, have your employees been Frequently 6.4 5.3 3.1 14.6 0 to another farm before yours? Rarely 67.6 93.4 61.5 65.6 16.9 Unknown 0.5 0 1.5 0 1.5 Not applicable 25.5 1.3 33.8 19.8 81.5

2000;64:0–00 The Canadian Journal of Veterinary Research 243 FOR PERSONAL USE ONLY

Table I. (continued)

Are visitors required to wear boots and coveralls? Frequently 69.2 81.5 60 84.4 27.7 Rarely 6.6 0 12.3 4.2 20 Unknown 0.3 0 1.5 0 0 Not applicable 23.9 18.5 26.2 11.5 52.3 Are visitors required to wash hands and/or shower Frequently 46.9 63.6 33.8 51 15.4 before entering the barn? Rarely 27.9 17.2 40 34.4 30.8 Unknown 1.6 0.7 0 3.1 3.1 Not applicable 23.6 18.5 26.2 11.5 50.8 On the day of the visit, have visitors been to Frequently 3.2 2.6 0 3.1 7.7 another farm before yours? Rarely 69.8 78.8 72.3 80.2 30.8 Unknown 3.2 0 1.5 5.2 9.2 Not applicable 23.9 18.5 26.2 11.5 52.3 Are trucks that pick up pigsb cleaned and Frequently 52.5 64.9 36.9 62.5 24.6 disinfected before they arrive at your barn? Rarely 40.6 32.5 53.8 33.3 56.9 Unknown 6.9 2.6 9.2 4.2 18.5 Not applicable 0 0 0 0 0 Do trucks that deliver bins of feed visit other farms Frequently 62.3 56.3 70.8 56.3 76.9 before yours? Rarely 26.5 35.8 21.5 24 13.8 Unknown 6.1 6 3.1 10.4 3.1 Not applicable 5 2 4.6 9.4 6.2 Does the truck driver enter the barn to assist with Frequently 2.9 0.7 1.5 6.3 4.6 the unloading of feed? Rarely 91.8 97.4 93.8 82.3 90.8 Unknown 0.5 0 0 2.1 0 Not applicable 4.8 2 4.6 9.4 4.6 Do trucks that pick up pigsb ever visit farms other Yes 69.2 59.6 67.7 77.1 81.5 than the destination farm and yours? No 30.2 39.7 30.8 22.9 18.5 Unknown 0.3 0.7 0 0 0 Not applicable 0.3 0 1.5 0 0 Are trucks that pick up pigsb driven to the barn Frequently 92.6 90.7 89.2 99 90.8 door? Rarely 7.2 9.3 10.8 1 7.7 Unknown 0.3 0 0 0 1.5 Not applicable 0 0 0 0 0 Do trucks that pick up pigsb arrive empty? Frequently 88.1 92.7 84.6 86.5 83.1 Rarely 11.7 7.3 15.4 12.5 16.9 Unknown 0.3 0 0 1 0 Not applicable 0 0 0 0 0 When picking up pigsb, does the truck driver Frequently 31.3 28.5 36.9 31.3 32.3 enter the barn to assist with the loading of pigs? Rarely 68.4 70.9 63.1 68.8 67.7 Unknown 0.3 0.7 0 0 0 Not applicable 0 0 0 0 0 HB(open) — high biosecurity herds that were open with respect to replacement animals. HB(closed) — high biosecurity herds that were closed with respect to replacement animals. MB — moderate biosecurity herds. LB — low biosecurity herds. a Weaned pigs, feeder pigs, and gilts. b Weaned pigs, feeder pigs, gilts, and market pigs.

244 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Table II. Continuous demographic variables that were investigated for an association with biosecurity group membership of southern Ontario swine herds in 2007 25th 50th 75th Variable Mean Range SD percentile percentile percentile N Current sow inventory 565.7 35, 3000 633.5 160 300 730 221 Capacity of the nursery barn 1397.5 50, 9600 1375.2 400 900 2000 218 Capacity of the finisher barn 1374.8 20, 7000 1051.1 600 1200 2000 213 Heat production units 444.8 20.4, 2346 325.3 226.3 357 601 377 Number of incoming pig shipments per month 1.9 0.08, 16 2.1 0.5 1.1 2.3 313 SD — standard deviation; N — total number. of “unknown” and “not applicable” were not changed. Trucking to convey information about the distribution of responses across variables related to the movement of pigs were addressed by ask- the groups and to provide names for each group. Following 2-step ing separate questions for each stage of production: weaned pigs, cluster analysis, LCA was carried out; both methods are well-suited feeder pigs, gilts, boars, and, if applicable, culls and market animals. to the purposes of this work and one is not inherently better than Questions pertaining to boars were excluded, with the rationale that the other. movement of boars between sites is not representative of commercial swine production in Ontario. Questions pertaining to the other stages Spatial analysis of production were condensed into 1 new variable that represented Mapping was accomplished with ArcMap (ArcGIS 9.3; Esri, each trucking practice. The poorest practice from all applicable stages Redlands, California, USA). Geographic information was represented was taken to be representative of this modified variable, with the using a projected coordinate system. All data were mapped using rationale that each aspect of trucking is only as good as its weakest the Universal Transverse Mercator (UTM) coordinate system (Zone link. Questions pertaining to culls were not grouped with other 17N), based on the World Geodetic System 1984 datum. Using this stages; they remained separate, as culled pigs are often moved by projection system, a 10 3 10 km grid was superimposed over a map trucking companies that deal exclusively with culls. These truck- of southern Ontario. ing practices may differ from those for healthy animals. Following Approximate farm locations were mapped using longitude and modifications, 22 categorical variables were selected for both 2-step latitude information available in the dataset. Geographic informa- cluster analysis and latent class analysis. tion could not be obtained for 21 farms; these farms were omitted Some variables were modified before the multinomial logistic from the mapping portion of the analysis, which reduced the sample regression model was developed. Heat production units (HPUs), size to 356 herds. Kernel smoothing methods are an effective way to which served as a proxy for herd size, were calculated using the for- represent spatial patterns among point data (18). Parametric kernel mula HPU = (0.17 3 capacity of the finisher barn) 1 (0.17 3 capacity smoothing with a bandwidth of 30 km was used to generate density of the nursery barn) 1 (0.45 3 sow inventory). This formula was maps for the entire study population and for each of the identified adapted from work by Zhuang et al (14). The number of incoming biosecurity groups. This particular bandwidth was selected in order pig shipments per month was calculated as the sum of the number of to visualize the variability in geographic distribution of the groups, times nursery pigs, feeder pigs, and gilts were delivered each month. without over-smoothing the data. Biosecurity groups identified using both 2-step cluster analysis and LCA were utilized. Rasters were Statistical and spatial analysis generated that represented the distribution of herds in each of the groups, as a proportion of the entire study population. Two-step cluster analysis and latent class analysis In order to determine the best number of groups to describe the data, Herd density mapping 22 external biosecurity variables were analyzed using 2 methods: 2-step Information about herd density and pig density in the study cluster analysis (SPSS Version 18.0; SPSS, Chicago, Illinois, USA) and region was provided through the Ontario Pork Marketing Board’s latent class analysis (SAS 9.2; SAS, Cary, North Carolina, USA). These database. This database contains geographic coordinates of swine procedures were selected because they are readily available methods sites in longitude and latitude, as well as information about sow, that use objective statistical criteria to determine the best number of grower, and finisher capacity for each premise. The database version groups to describe the data. The kappa statistic was used to determine used for this study contained 2955 swine sites, 2643 of which had how well the 2 methods agreed in their classification of herds. complete spatial information. These locations were used to estimate The 2-step clustering algorithm is advantageous because it is able the density of swine sites and grower-finisher pigs using parametric to handle large data sets consisting of numerous observations and kernel smoothing, with a bandwidth of 20 km. This smaller radius variables. In the first step of this procedure, the algorithm reduces was selected in order to better represent variability in areas with a the dataset into small pre-clusters of observations, while maintain- high density of pigs and pig sites. ing summary statistics for each group. In the second step, the pre- clusters are clustered hierarchically into larger groups. Bayesian Risk-based geographical analysis information criterion (BIC) is used to evaluate the best number of The risk-based map was produced by combining information clusters to describe the data (15–17). Descriptive statistics were used about the spatial distribution of the 2 high biosecurity groups

2000;64:0–00 The Canadian Journal of Veterinary Research 245 FOR PERSONAL USE ONLY

Figure 1. Kernel smoothed qualitative risk map representing areas of risk for the spread of conta- gious diseases of swine, presented on a 10 3 10 km grid. The map incorporates information about the density of swine and swine sites in the study region with information about the proportion of herds belonging to one of the high biosecurity groups. Risk has been redefined into quintiles.

­identified in the current study, with the rasters representing the graphic variables (Tables II and III). Groups identified using 2-step density of swine sites and grower-finisher pigs in the study region, cluster analysis were used for this purpose. A causal diagram was as well as the spatial distribution of the 2 high biosecurity groups developed in order to examine putative relationships between the identified in previous work by this research group (12). All quanti- variables of interest and biosecurity group membership. Univariable tative rasters were reclassified into quintiles, and each 10 3 10 km analysis was conducted for each predictor variable, with biosecurity grid space was ranked qualitatively in terms of risk of the spread of group as the outcome. A liberal P-value (P , 0.2) on the likelihood contagious swine diseases. Each raster was ranked from 1 through 5, ratio test (LRT) was used to determine which variables would be with a value of 5 representing the highest risk. High animal density further investigated for inclusion in the multivariable model. All pre- is a risk factor for the spread of disease; grid spaces with the high- dictors that were significant at this liberal P-value were assessed for est animal density were assigned a value of 5. As belonging to a collinearity using Spearman’s correlation coefficient; a value of 0.7 biosecurity group with good practices decreases the risk of disease or greater indicated collinearity existed. Serious collinearity issues incursion and spread, grid spaces with the highest proportion of existed between the nursery flow and finisher flow variables; in farms belonging to a high biosecurity group were assigned a value order to address this issue, a new dichotomous variable was created of 1. The overall risk map was developed using a weighted sum that indicated whether or not a farm was all-in, all-out by barn. A overlay function. The density rasters were each given a weight farm was assigned a value of 1 if both nursery and finisher flow were of 0.25 to provide half the weighted risk. Rasters representing the all-in, all-out by barn, or if 1 stage was all-in, all-out by barn and the probability of belonging to a high biosecurity group shared the other stage was not applicable. Otherwise, the farm was assigned remaining half of the weighted risk. The resulting quantitative risk a value of 0. Biologically plausible interactions were tested for sig- map was then reclassified into 5 qualitative categories, indicating nificance using the LRT. A manual backwards selection procedure “very high,” “high,” “moderate,” “low,” or “very low” risk for the was then used to determine which variables would remain in the spread of contagious pig diseases. final model, using a more conservative P-value (P , 0.05). Nested models were compared using the LRT and non-nested models were Multinomial logistic regression compared using BIC. Potential confounders were retained and exam- Multinomial logistic regression was used to investigate the rela- ined; a 25% change in the coefficient of interest provided sufficient tionship between biosecurity group membership and selected demo- evidence that confounding was present. Once the final model was

246 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Table III. Categorical demographic variables investigated for 40.1% of herds and will be referred to as “HB(open);” ii) high an association with biosecurity group membership of biosecurity herds that were closed with respect to replacement ani- southern Ontario swine herds in 2007 mals — this group represented 17.2% of herds and will be referred to as “HB(closed);” iii) moderate biosecurity herds that represented Percentage 25.5% of herds and will be referred to as “MB;” and iv) low bio- of herds security herds that represented 17.2% of herds and will be referred Variable Responses (%) to as “LB”. Descriptive information about the external biosecurity Herd type Farrow-to-finish 33.4 variables used, how the sample of swine producers responded to Farrow-to-grow 10.1 each question, as well as the breakdown of responses by biosecurity Farrow-to-wean 17.2 group is provided in Table I. Finisher 21 The 2-step cluster analysis output provides information about Nursery 13.8 the importance of each variable in differentiating among groups. Nursery and finisher 4.5 The 3 most important variables pertained to trucking practices and Nursery flow All-in-all-out by barn 3.7 included: whether trucks that picked up culls were cleaned and dis- All-in-all-out by room 41.9 infected before arriving at the farm; whether trucks that picked up Continuous flow 15.4 other pigs (weaned pigs, feeder pigs, gilts, and market animals) were Not applicable 39 cleaned and disinfected before arriving at the farm; and whether trucks that delivered pigs visited another farm immediately before Finisher flow All-in-all-out by barn 11.4 arriving at the farm of interest. All-in-all-out by room 19.6 Although not specifically related to our study objectives, it is Continuous flow 27.9 interesting to note how PRRSV cases and controls were distributed Not applicable 41.1 across the biosecurity groups. It is important to keep in mind that All-in-all-out by barn Yes 14.1 the farms were selected based on their PRRSV status, so farm-level No 85.9 prevalence is high across all groups. Overall, 85.9% of herds in our sample were PRRSV-positive. Stratified by group, 95.4% of herds in the LB group were PRRSV-positive, 89.4% of herds in the HB(open) determined, model fit was assessed using a series of binary logits, group were positive, 83.1% of herds in the HB(closed) group were where the outcome was membership to 1 specific biosecurity group positive, and 76% of herds in the MB group were positive. Previous versus membership to any of the other groups (19). Standardized work by this research group identified 3 biosecurity groups in a dif- residuals, leverage, and delta beta values were examined in order ferent sample of sow herds (12). In that study, 66.5% of herds were to identify potential outliers or influential observations. Any ques- PRRSV-positive. Stratified by biosecurity group, 82% of herds in the tionable observations were removed, one at a time, from the full LB group were PRRSV-positive, while 48% of herds in the HB(open) multinomial model and the impact on the magnitude and direction group and 74% of herds in the HB(closed) were PRRSV-positive (12). of the coefficients, as well as the significance of each predictor, were assessed. Once the multivariable multinomial logistic regression Spatial analysis model was finalized, predicted probability graphs were generated The final risk map is presented in Figure 1. This map consists of in order to examine how the probability of group membership information from 10 individual rasters: the spatial distribution of changed as each continuous predictor varied from its minimum to the 2 high biosecurity groups identified in this analysis, according to its maximum value. When considering the relationship between both clustering methodologies; the spatial distribution of the 2 high group membership and the number of incoming pig shipments per biosecurity groups identified in previous work by this research month, the model was run without the closed group before making group, according to both clustering methodologies (12); and the predictions, since these herds generally did not receive replacement rasters representing swine density and the density of grower-finisher animals from other sources. Categorical predictors were assessed by pigs in the study region. listing the coefficients for all pairs of outcome categories. Multinomial logistic regression Results Variables significant on univariable analysis (P , 0.2) were fur- ther investigated for inclusion in the multivariable model. These included: i) HPU; ii) herd type; iii) number of incoming pig ship- Biosecurity groups ments per month; and iv) whether pig flow was all-in, all-out by The sample of southern Ontario swine herds was best described by barn. Inclusion of the significant interaction terms in the full model 4 external biosecurity groups; this number of groups was identified resulted in either non-convergence or a higher BIC value than when using both clustering methods. Agreement was excellent; kappa = only the main effects were included. The inclusion of 1 interac- 0.96 [95% confidence interval (CI): 0.9, 1.0; P , 0.001]. Based on the tion term did result in a lower BIC value, but did not contribute characteristics of herds in each of the identified groups, the groups substantially to model interpretation. Herd type confounded the were named by the authors as: i) high biosecurity herds that were relationship between HPU and biosecurity group membership, as open with respect to replacement animals — this group represented well as between the number of incoming pig shipments per month

2000;64:0–00 The Canadian Journal of Veterinary Research 247 FOR PERSONAL USE ONLY

Table IV. Final multivariable multinomial logistic regression model for biosecurity group membership in southern Ontario swine herds in 2007 Biosecurity Relative risk Confidence group Risk factor ratio interval P-value LB (base outcome)

HB(open) Farrow-to-growa 6.10 1.909, 19.505 , 0.002 Farrow-to-weana 15.76 5.531, 44.892 , 0.0001 Finishera 6.94 2.607, 18.5 , 0.0001 Nurserya 4.37 0.954, 19.984 , 0.058 Nursery and finishera 4.89 0.459, 52.018 , 0.188 Heat production units 1.003 1.002, 1.005 , 0.0001 Number of pig deliveries per month 1.04 0.797, 1.356 , 0.777

HB(closed) Farrow-to-growa 3.36 0.147, 76.822 , 0.447 Farrow-to-weana 1.53 3 1026 0, ∞ , 0.991 Finishera 2.33 3 1027 0, ∞ , 0.990 Nurserya 0.002 0, ∞ , 0.994 Nursery and finishera 3.7 3 109 0, ∞ , 0.985 Heat production units 0.997 0.990, 1.003 , 0.323 Number of pig deliveries per month 7.59 3 1029 8.79 3 10215, 0.007 , 0.007 MB Farrow-to-growa 3.33 0.250, 44.411 , 0.362 Farrow-to-weana 3.27 0.252, 42.452 , 0.365 Finishera 97.49 18.96, 501.296 , 0.0001 Nurserya 173.49 24.874, 1210.017 , 0.0001 Nursery and finishera 127.0 9.368, 1721.7 , 0.0001 Heat production units 1.002 1.0005, 1.004 , 0.015 Number of pig deliveries per month 1.1008 0.828, 1.464 , 0.509

Log likelihood = 2211.3; LR(x2) 307.92; df = 21, P , 0.0001 HB(open) — high biosecurity herds that were open with respect to replacement animals. HB(closed) — high biosecurity herds that were closed with respect to replacement animals. MB — moderate biosecurity herds. LB — low biosecurity herds. a Reference is farrow-to-finish herds. and biosecurity group membership. The final multivariable multi- Discussion nomial model includes herd type, HPU, and number of incoming pig shipments per month (Table IV). Herds belonging to the HB(open) group were generally character- Predicted probability graphs are presented for selected herd types: ized by good trucking practices and strict entrance sanitation proce- farrow-to-wean; farrow-to-finish; nursery; and finisher. Figures 2 and dures for employees and visitors. Although some producers in the 3 demonstrate how the predicted probability of group membership HB(open) group did not report frequent washing and disinfection of changes as HPU and number of pig deliveries per month increase, trucks before arrival at the farm, a much larger proportion of the pro- respectively. Sow herds showed 1 general pattern, while non-sow ducers in this group did practice these high trucking standards than herds showed another. The effect of the categorical variable was in other groups. Only 65% of producers in the HB(open) reported assessed by examining the coefficients for all pairs of outcomes that trucks picking up pigs were frequently cleaned and disinfected (Table V). The odds of belonging to the high biosecurity group before arrival; this is an external biosecurity concern and should be rather than to the low biosecurity group were 6.1 times greater for addressed. However, this group did compare favorably to the other farrow-to-grow herds than for farrow-to-finish herds, holding the groups in the cleaning and disinfecting of trucks before arrival. continuous variables constant (95% CI = 1.9, 19.5; P = 0.002). The Herds in the HB(closed) group were characterized as such because odds of belonging to the moderate biosecurity group rather than the most producers in this group answered “not applicable” to questions high biosecurity group were 14 times greater for finisher herds than about pigs being delivered to the farm. This indicates that they were for farrow-to-finish herds (95% CI = 3.0, 65.7; P = 0.001). See Table V closed with respect to replacement animals. Direct contact between for all comparisons among the biosecurity groups. an infected, shedding pig and one that is susceptible to infection is

248 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Table V. Relative risk ratios representing the odds of belonging to 1 biosecurity group, relative to an alternative biosecurity group for a given herd type. These odds compare the given herd type to farrow-to-finish herds Relative risk Confidence Herd type Comparison ratio interval P-value Farrow-to-finish (base category) Farrow-to-grow HB(open) to HB(closed) 1.81 0.081, 40.487 , 0.707 HB(open) to MB 1.83 0.155, 21.692 , 0.631 HB(open) to LB 6.10 1.909, 19.505 , 0.002 HB(closed) to MB 1.01 0.021, 48.393 , 0.996 HB(closed) to LB 3.36 0.147, 76.822 , 0.447 MB to LB 3.33 0.250, 44.411 , 0.362

Farrow-to-wean HB(open) to HB(closed) 1.03 3 107 0, ∞ , 0.989 HB(open) to MB 4.82 0.415, 55.858 , 0.209 HB(open) to LB 15.76 5.531, 44.892 , 0.0001 HB(closed) to MB 4.67 3 1027 0, ∞ , 0.990 HB(closed) to LB 1.53 3 1026 0, ∞ , 0.991 MB to LB 3.27 0.252, 42.452 , 0.365

Finisher HB(open) to HB(closed) 2.98 3 107 0, ∞ , 0.989 HB(open) to MB 0.07 0.015, 0.334 , 0.001 HB(open) to LB 6.94 2.607, 18.5 , 0.0001 HB(closed) to MB 2.39 3 1029 0, ∞ , 0.988 HB(closed) to LB 2.33 3 1027 0, ∞ , 0.990 MB to LB 97.49 18.96, 501.296 , 0.0001 Nursery HB(open) to HB(closed) 2411.46 0, ∞ , 0.992 HB(open) to MB 0.03 0.005, 0.135 , 0.0001 HB(open) to LB 4.37 0.954, 19.984 , 0.058 HB(closed) to MB 1.0 3 1025 0, ∞ , 0.989 HB(closed) to LB 0.002 0, ∞ , 0.994 MB to LB 173.49 24.874, 1210.017 , 0.0001

Nursery and finisher HB(open) to HB(closed) 1.32 3 1029 0, ∞ , 0.986 HB(open) to MB 0.04 0.006, 0.252 , 0.001 HB(open) to LB 4.89 0.459, 52.018 , 0.188 HB(closed) to MB 2.91 3 107 0, ∞ , 0.989 HB(closed) to LB 3.70 3 109 0, ∞ , 0.985 MB to LB 127.00 9.368, 1721.7 , 0.0001 HB(open) — high biosecurity herds that were open with respect to replacement animals. HB(closed) — high biosecurity herds that were closed with respect to replacement animals. MB — moderate biosecurity herds. LB — low biosecurity herds.

the most important route of transmission for most infectious diseases allowing the truck driver to enter the barn and assist with unloading of swine (20–22). By not adding replacement animals from outside pigs, the MB group performed poorly compared to the low biosecu- sources, these farms were mitigating the risk posed by introducing rity group. Finally, herds in the LB group were characterized by poor new animals. Herds in this group did well in some areas of trucking, trucking and entrance sanitation procedures. Most of the producers but required improvement in other areas. Nearly 60% of producers in this group reported infrequent cleaning and disinfection of trucks that had employees rarely required them to wash their hands and/or picking up animals, including culls, weaned pigs, feeder pigs, gilts, shower before entering the facility. and market animals. Most of these farms did not have employees The MB group was defined as such because of the variation in and nearly 85% of those that did rarely required employees to wash external biosecurity practices. Herds in this group did very well in their hands and/or shower before entering the facility. some areas and poorly in others. These herds were comparable to Comparison of PRRSV-status across the biosecurity groups iden- those in the HB(open) group in some aspects of trucking, such as tified here and in previous work indicates that the low biosecurity use of dedicated trucks for pig deliveries. In other aspects, such as group had the highest proportion of PRRSV-positive herds. This

2000;64:0–00 The Canadian Journal of Veterinary Research 249 FOR PERSONAL USE ONLY

Figure 2. Predicted probability of biosecurity group membership for 4 herd types, as heat production units (HPUs) varies. Predictions are based on the final multivariable multinomial logistic regression model.

­suggests that herds belonging to the low biosecurity group are rity and density were weighted differently. Although we elected to indeed at increased risk for the presence of the PRRS virus. These assign equal weights to both sources of information, further research results should be interpreted with caution, however, since the dataset into risk-based geographical approaches may provide insight into used here was designed as a case-control study and the previous the relative importance of these respective issues. study (12) represented a convenience sample that was recruited The final multivariable multinomial logistic regression model pro- without a formal selection process. Additionally, we have no infor- vided insight into which demographic variables serve as significant mation about when these herds became PRRSV-positive or how long predictors of biosecurity group membership. There are conflicting the biosecurity practices of interest had been in use. reports of whether herd size itself is a risk factor for the presence of The concept of incorporating spatial information into risk-based different pathogens (24,25). Nonetheless, it plays an important role surveillance activities has gained interest in several recent publica- in predicting biosecurity level, as demonstrated by our model. For tions (2,8). In the current study, herd-level factors related to external both sow herds and non-sow herds, the probability of belonging to biosecurity practices were combined with geographic information the HB(open) group increased as HPUs increased. This likely reflects about the location of study herds and the density of animals in the the perceived risk associated with introducing pathogens into large study region to develop a risk map for the spread of contagious herds. As the consequences in terms of the welfare of animals and swine diseases. It is interesting to note that the risk map closely the financial impact would be more severe on larger farms, such approximates the density map (not shown); the areas of highest den- farms would benefit more from having stricter trucking and entrance sity correspond to the areas of highest risk. This supports previous sanitation procedures than smaller farms. When considering the vari- findings that herd and animal density serve as predictors for the risk ables that make up these biosecurity groups, there are other possible of disease transmission (9–11,23) and indicates that strict external explanations for this trend. While large farms may have the financial biosecurity practices are especially important in areas of high pig means to implement certain biosecurity protocols, such as the use of density. It is important to remember, however, that the risk map off-site isolation and acclimation facilities, acquiring such facilities presented here could have looked substantially different if biosecu- may not be economically justifiable or even feasible on smaller farms.

250 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Figure 3. Predicted probability of biosecurity group membership for 4 herd types, as the number of incoming pig shipments per month varies. Predictions are based on the final multivariable multinomial logistic regression model.

Additionally, large herds would fill an entire truck, thus eliminating more detail. Similar conclusions can be drawn from the relative the need for transport vehicles to visit other farms to obtain a full risk ratios, which compare different herd types to farrow-to-finish load. It was also evident that a large proportion of herds in the low herds. Farrow-to-wean and farrow-to-grow herds were more likely biosecurity group did not have employees. If the owner is the only to belong to the HB(open) group than to the LB group, and nursery person entering the barn, entrance sanitation protocols may be more and finisher herds were more likely to belong to the MB group than relaxed than in larger farms, where many people have regular access. to the LB group. This could reflect the age of the swine facilities and The finding that sow herds were more likely to belong to the the monetary investments required to improve biosecurity level. HB(open) group as the number of pig shipments per month Generally, farrow-to-finish operations are older, family-owned increased is encouraging and shows that these producers were aware facilities. In contrast, the more modern approach to swine produc- of the biosecurity risks associated with increased contact rates. When tion involves multi-site systems. Newer, multi-site operations can the number of incoming pig shipments per month is high, trucking be built with specific biosecurity measures in mind, whereas older protocols in particular become more important. Farms with non-sow facilities are limited to the existing structures and require consider- herds were most likely to belong to the MB group, regardless of the able financial investments to meet today’s standards. number of pig shipments per month. We hypothesize that this find- This study had several limitations. Firstly, the data were collected ing could be due to a combination of 2 factors. Firstly, a considerable in 2007 and there have probably been improvements in more recent proportion of nursery and finisher sites have the capacity to operate years as a result of the increased focus on the importance of biosecu- on an all-in, all-out by site basis, which makes routine depopulation rity and the training and documentation that are now available to more feasible and practical than in sow units. This opportunity for producers and veterinarians. The National Farm-Level Biosecurity routine depopulation means that a newly introduced pathogen is Standard, published by the Canadian Swine Health Board, provides less likely to persist in the environment. Secondly, the consequences information about best management practices in the swine industry, of outbreaks in non-sow herds may have lower perceived cost than as well as training programs for veterinarians and producers (26). in the sow units, although this perception needs to be studied in Secondly, the selection process was based on the use of the AHL at

2000;64:0–00 The Canadian Journal of Veterinary Research 251 FOR PERSONAL USE ONLY

the University of Guelph and on OASV members. As a result, only oie.int/index.php?id=169&L=0&htmfile=chapitre_1.1.4.htm Last farms with veterinarians who submitted samples to the AHL or who accessed June 18, 2013. were members of the OASV were eligible for inclusion. Thirdly, as 5. Thacker SB. Historical development. In: Lee LM, Teutsch SM, with any questionnaire data, there is the possibility of misclassifica- Thacker SB, St. Louis ME, eds. Principles and Practice of Public tion bias that results in the herds being allotted to incorrect response Health Surveillance. 3rd ed. New York, New York: Oxford categories (27). This bias could have occurred during data collection, University Press, 2010:1–17. further modification of variables, or the analysis stage. Regarding the 6. National Farmed Animal Health and Welfare Strategy: A development of the overall risk map, any selection or information Framework for Industry and Government. Council of Chief bias would be reduced due to the inclusion of 2 data sets. Finally, Veterinary Officers (CCVO)/Farmed Animal Industry Joint some aspects of external biosecurity were not addressed by the Working Group, 2009. Canadian Animal Health Coalition survey and only a subset of the available variables was selected for [homepage on the Internet]. Available from: http://www.ani further analysis. This was necessary in order to limit the number of malhealth.ca/Projects/Detail.aspx?id=18 Last accessed June 18, variables to a manageable level. 2013. The practical application of the risk map is to identify geographic 7. Salman MD. Surveillance and monitoring systems for animal areas that are at risk for the spread of contagious swine diseases. health programs and disease surveys. In: Salman MD, ed. We recommend that it would be best to focus surveillance projects Animal Disease Surveillance and Survey Systems: Methods and in geographic areas that have been identified as high risk. The Applications. Ames, Iowa: Iowa State Press, 2003:3–13. combination of high animal density and a low proportion of herds 8. Benschop J, Stevenson MA, Dahl J, French NP. Towards incor- belonging to a group identified as having good biosecurity practices porating spatial risk analysis for Salmonella sero-positivity makes these areas ideal for targeted approaches to surveillance. into the Danish swine surveillance programme. Prev Vet Med Alternatively, the planning of disease control projects should take 2008;83:347–359. into consideration the geographic areas identified as being low risk; 9. Boender GJ, Nodelijk G, Hagenaars T, Elbers A, de Jong M. these areas have a low density of animals and a high proportion of Local spread of classical swine fever upon virus introduction herds belonging to a group identified as having good biosecurity into The Netherlands: Mapping of areas at high risk. BMC Vet practices. These findings may improve the success of surveillance Res 2008;4:9–20. and eradication projects. 10. Rose N, Madec F. Occurrence of respiratory disease outbreaks in fattening pigs: Relation with the features of a densely and a Acknowledgments sparsely populated pig area in France. Vet Res 2002;33:179–190. 11. Ortiz-Pelaez Á, Pfeiffer D. Use of data mining techniques to The authors thank Thomas Rosendal and Beth Young for their investigate disease risk classification as a proxy for compro- assistance with survey design and pretesting, Doug Wey for con- mised biosecurity of cattle herds in Wales. BMC Vet Res 2008;4: ducting telephone interviews, Karen Richardson for conducting 24–38. telephone interview and data entry, and William Sears for statistical 12. Bottoms K, Poljak Z, Dewey C, Deardon R, Holtkamp D, advice. We are grateful to the veterinarians and producers for their Friendship R. Evaluation of external biosecurity practices on participation. We also thank Ontario Pork for funding the initial southern Ontario sow farms. Prev Vet Med 2013;109:58–68. study and for providing information on herd and pig density, as 13. Rosendal T. The spread of porcine reproductive and respira- well as the Ontario Ministry of Agriculture, Food and Rural Affairs tory syndrome virus (PRRSV) by genotype and the associa- for financial support under the Emergency Management Research tion between genotype and clinical signs in Ontario, Canada Theme and the Natural Sciences and Engineering Research Council 2004–2007 [PhD dissertation]. Guelph, Ontario, University of for funding support. Guelph. 2012:20–28. 14. Zhuang Q, Barfod K, Wachmann H, et al. Risk factors for References Actinobacillus pleuropneumoniae serotype 2 infection in Danish genetic specific pathogen-free pig herds. Vet Rec 2007;160: 1. Alban L, Boes J, Kreiner H, Petersen JV, Willeberg P. Towards a 258–262. risk-based surveillance for Trichinella spp. in Danish pig produc- 15. The SPSS TwoStep Cluster Component: A Scalable Component tion. Prev Vet Med 2008;87:340–357. Enabling More Efficient Customer Segmentation. Technical 2. Snow LC, Newson SE, Musgrove AJ, Cranswick PA, Crick HQP, Report, 2001. IBM [homepage on the Internet]. Available from: Wilesmith JW. Risk-based surveillance for H5N1 avian influenza http://www.spss.ch/upload/1122644952_The SPSS TwoStep virus in wild birds in Great Britain. Vet Rec 2007;161:775–781. Cluster Component.pdf Last accessed June 18, 2013. 3. Stärk K, Regula G, Hernandez J, et al. Concepts for risk-based 16. Norušis M. Cluster analysis. In: IBM SPSS Statistics 19 Statistical surveillance in the field of veterinary medicine and veterinary Procedures Companion. Pearson Education, 2011:375–404. public health: Review of current approaches. BMC Health Serv 17. Schwarz G. Estimating the dimension of a model. Ann Stat Res 2006;6:20–27. 1978;6:461–464. 4. Terrestrial Animal Health Code. Chapter 1.4. Animal Health 18. Pfeiffer D, Robinson TP, Stevenson M, Stevens KB, Rogers DJ, Surveillance. World Organization for Animal Health (OIE) Clements ACA. Spatial Analysis in Epidemiology. New York, [homepage on the Internet]. 2011. Available from: http://www. New York: Oxford University Press, 2008.

252 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

19. Long JS, Freese J. Models for nominal outcomes with case- “neighbourhood infections” for different regions in Belgium. specific data. In: Regression Models for Categorical Dependent Prev Vet Med 2003;60:27–36. Variables using Stata. 2nd ed. College Station, Texas: Stata Press, 24. Baptista FM, Alban L, Nielsen LR, Domingos I, Pomba C, 2006:223–291. Almeida V. Use of herd information for predicting Salmonella 20. Amass SF, Clark LK. Biosecurity considerations for pork produc- status in pig herds. Zoonoses Public Health 2010;57:49–59. tion units. J Swine Health Prod 1999;7:217–230. 25. Dahl J. Cross-sectional epidemiological analysis of the relations 21. Dee SA. Biosecurity: A critical review of today’s practices. Proc between different herd factors and Salmonella sero-positivity. Amer Assoc Swine Vet Annual Meeting 2003:451–455. Proc 8th International Symposium Vet Epi Econ. Paris, France, 22. Good Biosecurity Practices for the Pig Sector: Issues and Options 1997:1–3. in Developing and Transition Countries. Food and Agriculture 26. National Swine Farm-Level Biosecurity Standard, 2010. Organization (FAO), Animal Production and Health Paper Canadian Swine Health Board (CSHB) [homepage on the No. 169, 2010. [homepage on the Internet]. Available from Internet]. Available from: http://www.swinehealth.ca/CSHB_ http://www.fao.org/docrep/012/i1435e/i1435e00.pdf Last Biosecurity_StandardE.pdf Last accessed June 18, 2013. accessed June 19, 2013. 27. Dohoo I, Martin W, Stryhn H. Veterinary Epidemiologic 23. Mintiens K, Laevens H, Dewulf J, Boelaert F, Verloo D, Koenen F. Research. 2nd ed. Charlottetown, Prince Edward Island: VER Risk analysis of the spread of classical swine fever virus through Inc, 2009.

2000;64:0–00 The Canadian Journal of Veterinary Research 253 FOR PERSONAL USE ONLY

Article

A retrospective study on the etiological diagnoses of diarrhea in neonatal piglets in Ontario, Canada, between 2001 and 2010 Gloria Chan, Abdolvahab Farzan, Josepha DeLay, Beverly McEwen, John F. Prescott, Robert M. Friendship

Abstract Laboratory surveillance data from the Animal Health Laboratory, University of Guelph, on the etiological diagnoses of neonatal diarrhea in piglets were analyzed to determine the relative importance and trends of different enteric pathogens in Ontario. A total of 237 cases, including live and dead 1- to 7-day-old piglets, were submitted for diagnosis of gastrointestinal illness between 2001 and 2010. The combined frequencies for cases of gastrointestinal illness involving Escherichia coli, Clostridium perfringens type A, rotavirus, and Clostridium difficile, either as single pathogens or a complex of pathogens, accounted for 56% of the total cases. In a total of 33% of cases of gastrointestinal illness, an etiological agent was not identified. The frequency of cases diagnosed with enterotoxigenic E. coli was decreased from 2007. Cases submitted in 2010 were more likely to be diagnosed with C. perfringens type A compared to cases submitted in 2002 to 2007 (P , 0.05). There was a significant trend for cases submitted in the winter to be diagnosed with C. perfringens type A, enterotoxigenic E. coli, rotavirus, and Cystoisospora suis (formerly Isospora suis) (P , 0.05). Enterotoxigenic E. coli was less likely diagnosed if C. difficile, C. perfringens, or rotavirus were detected (P , 0.05). Younger piglets were more likely to be diagnosed with C. perfringens type A (P , 0.05) and C. difficile (P , 0.05) than older piglets. This study shows that E. coli, C. perfringens type A, rotavirus, and C. difficile are enteric pathogens of concern for Ontario swine farrowing operations and further research is required to understand the reasons for the cases that are not diagnosed.

Résumé Les données de surveillance provenant du Animal Health Laboratory de l’University of Guelph sur les diagnostics étiologiques des diarrhées néonatales des porcelets ont été analysées afin de déterminer l’importance relative et les tendances des différents agents pathogènes entériques en Ontario. Entre 2001 et 2010, 237 cas de porcelets vivants et morts âgés de 1 à 7 jours ont été soumis pour un diagnostic de maladie gastro- intestinale. Les fréquences combinées pour les cas de maladies gastro-intestinales impliquant Escherichia coli, Clostridium perfringens type A, rotavirus, et Clostridium difficile, soit comme seul agent pathogène ou un complexe d’agents pathogènes, représentaient 56 % du total des cas. Pour 33 % des cas de maladies gastro-intestinales, aucun agent étiologique n’a été identifié. La fréquence des cas avec un diagnostic d’E. coli entérotoxigénique a diminué à compter de 2007. Les cas soumis en 2010 étaient plus susceptibles d’avoir un diagnostic d’infection à C. perfringens type A comparativement aux cas soumis de 2002 à 2007 (P , 0,05). Il y avait une tendance significative pour les cas soumis en hiver d’avoir un diagnostic avec C. perfringens type A, E. coli entérotoxigénique, rotavirus, et Cystoisospora suis (anciennement Isospora suis) (P , 0,05). Les E. coli entérotoxigéniques étaient moins souvent diagnostiquées si C. difficile, C. perfringens, ou du rotavirus étaient détectés (P , 0,05). Les porcelets plus jeunes étaient plus susceptibles d’être diagnostiqués avec du C. perfringens type A (P , 0,05) et C. difficile (P , 0,05) que les porcelets plus vieux. Cette étude démontre que E. coli, C. perfringens type A, rotavirus, et C. difficile sont des agents pathogènes d’intérêt pour les opérations de mise-bas en Ontario et des recherches supplémentaires sont requises pour comprendre les raisons pour lesquelles certains cas ne sont pas diagnostiqués. (Traduit par Docteur Serge Messier) Introduction diarrhea in piglets attributed to Clostridium difficile in a survey con- ducted in 2000 (1), and in Clostridium perfringens type A in diagnostic Neonatal piglet diarrhea is a major cause of pre-weaning mortal- laboratory submission data from 2004 to 2009 (2). These reports also ity, resulting in significant economic loss for swine producers. The indicate a relative decrease in cases attributed to enterotoxigenic relative importance of different diseases contributing to neonatal Escherichia coli (ETEC), transmissible gastroenteritis (TGE) virus, diarrhea in piglets appears to be changing, possibly because of and Clostridium perfringens type C compared to retrospective data changes in husbandry and management practices, advances in diag- from previous decades (1,2). nostic techniques, and/or the emergence of new diseases. Reports In recent years, clostridial infections as causes of neonatal diar- in the United States have indicated an increase in cases of neonatal rhea in piglets have gained prominence. Clostridium perfringens are

Department of Population Medicine (Chan, Farzan, Friendship); Animal Health Laboratory (DeLay, McEwen); and Department of Pathobiology (Prescott), University of Guelph, Guelph, Ontario. Address all correspondence and reprint requests to Dr. Robert M. Friendship; telephone: (519) 824-4120, ext. 54873; fax: (519) 763-3117; e-mail: [email protected] Received August 2, 2012. Accepted December 2, 2012.

254 The Canadian Journal of Veterinary Research 2013;77:254–260 FOR PERSONAL USE ONLY

divided into 5 toxinotypes (2,3), and some studies have found an enterocytes by FAT, or in feces or intestinal content by latex agglu- association between C. perfringens type A strains containing the tination test (LAT), which detected rotavirus A. From September beta2 toxin gene (cpb2) and neonatal diarrhea in piglets (4,5). The 2010 onwards, rotavirus infection was confirmed by multiplex current diagnostic criteria for C. perfringens type A enteritis is not real-time reverse-transcriptase polymerase chain reaction (real time specific and is generally based on a combination of isolation of RT-PCR) for rotavirus A and C, or gel-based RT-PCR for rotavirus B large numbers of C. perfringens type A from the intestinal contents, (6). Diagnosis of coccidiosis caused by Cystoisospora suis (formerly exclusion of other causes of diarrhea, and possibly the detection of Isospora suis) was based on histologic identification of Cystoisospora cpb2 through polymerase chain reaction (PCR) (3). Diarrhea associ- merozoites, schizonts, or oocysts in the cytoplasm of enterocytes, ated with C. difficile can occur in piglets between 1 to 7 d of age, often in combination with lesions of villus atrophy, or identification with some reports also including sudden death and scrotal edema of coccidial oocysts in feces by sucrose wet mount or fecal flotation. (2,3). Diagnosis of C. difficile-associated neonatal diarrhea in piglets Diagnosis of C. perfringens type A was based on the presence of is generally confirmed by histopathologic lesions in the colon and, long bacilli consistent with C. perfringens on the intestinal mucosal less often, the small intestine, in combination with toxin detection in surface, in combination with isolation of C. perfringens in large intestinal content via enzyme-linked immunosorbent assay (ELISA) numbers from intestinal content or, in some cases, genotyping of kits (1,3). C. perfringens strains through gel-based PCR, which detected the toxi- The purpose of this study was to use laboratory diagnostic data to notype (A, B, C, D, E) and cpb2. Confirmation of C. difficile-­associated identify the frequency and trends of different pathogens contributing enterocolitis was based on presence of typical histologic lesions of to neonatal diarrhea in piglets from Ontario swine farms between focal suppurative inflammation and necrosis (“volcano” lesions), 2001 and 2010. in combination with the detection of C. difficile toxins (A and B) in intestinal content or feces via ELISA, or isolation of C. difficile (the Materials and methods culture method was available from March 2008 onward) (7). Bacterial culture of the intestinal contents of the piglets was used to identify other bacterial pathogens, including Salmonella enterica serovars and Data collection Enterococcus durans. Cryptosporidium parvum was identified based on Data were provided by the Animal Health Laboratory (AHL) the presence of typical histological lesions, including the presence at the University of Guelph, and included laboratory submissions of appropriately sized and shaped protozoa adhered to the luminal from swine farms in Ontario from 2001 to 2010. The data included surface of the enterocytes. Concurrent porcine reproductive and cases in which . 1 live or dead piglets between 1 to 7 d of age were respiratory syndrome virus (PRRSV) infection in some pigs with submitted for a gastrointestinal tract (GIT) illness. A GIT case was GIT disease was determined by identification of PRRSV antigen in defined as a case submitted for one or more of the following reasons: lung by IHC, or detection of viral nucleic acid in lung using IHC enteritis, diarrhea, or scours. For farms with multiple submissions, or gel-based PCR (before June 2010) or multiplex real-time RT-PCR each submission was counted as 1 case. The data set for GIT cases (from June 2010 onward) (8). submitted from January 2001 to April 2007 was provided in text file, and the data set for GIT cases submitted from May 2007 to December Data analysis 2010 was provided in a spreadsheet (Microsoft Excel 2007; Microsoft The 2 data sets were merged in a spreadsheet and imported Corporation, Redmond, Washington, USA). for statistical analysis (Stata 10 Intercooled for Windows XP; StataCorp LP, College Station, Texas, USA). Multivariable logistic Diagnostic methods regression models were used to analyze the association between Necropsy and histopathologic examination were done in all cases. the diagnosis of ETEC, C. perfringens, C. difficile, rotavirus, and The etiological diagnoses of GIT cases were based on pathologists’ Cystoisospora suis, and independent variables; including age of piglet interpretation of gross and histopathologic lesions, in combination in days, year of submission, season of submission, and diagnosis with results of microbiologic tests. Ancillary microbiologic tests com- of other enteric pathogens. The seasons were categorized using mon to all cases with necropsy evidence of gastrointestinal disease the calendar dates for the equinoxes and solstices for each year in included bacterial culture from the small intestine and colon, and Canada. A manual stepwise procedure was used to build the models. rotavirus latex agglutination (RLA) testing. Additional tests done Univariable analysis for the association between diagnosis of the for each case were at the discretion of the individual pathologist enteric pathogen and independent variables was done using single assigned to the case, as directed by lesions indicative of specific logistic regression, the predictor variables where P , 0.20 were etiologic agents. A diagnosis of colibacillosis required histologic evi- selected for inclusion in multivariable analyses. Pair-wise correla- dence of colonization of the luminal surface of enterocytes by short tion coefficients were calculated between independent variables, and rod-shaped bacilli, isolation of E. coli from intestine, and identifica- coefficients with an absolute value greater than 0.8 were considered tion of ETEC strains through agglutination serotyping or genotyping collinear. If 2 independent variables were significantly correlated, the through gel-based PCR. Diagnosis of TGE was based on histologic more informative variable was used in the final model. A variable evidence of atrophic enteritis and identification of TGE virus antigen was identified as a confounder if it changed the coefficient of the by fluorescent antibody test (FAT) or immunohistochemistry (IHC). main effects by $ 20% when the potential confounder variable was A diagnosis of rotavirus enteritis required histologic evidence of removed. If a variable was determined to be a confounder, it was atrophic enteritis and identification of rotavirus antigen within included in the final model regardless of its statistical significance.

2000;64:0–00 The Canadian Journal of Veterinary Research 255 FOR PERSONAL USE ONLY

C. perfringens ETEC Rotavirus Cystoisospora suis C. difficile TGEV Diagnostic success rate (%) Proportion of total cases submitted

Year Year

Figure 1. Diagnostic success rate for gastrointestinal tract (GIT) cases Figure 2. Annual prevalence for gastrointestinal tract (GIT) cases in 1- to 7-day-old pigs submitted to Animal Health Laboratory (AHL) from submitted to Animal Health Laboratory (AHL) diagnosed with a single 2001 to 2010. etiological agent.

Interaction was evaluated between the detection of each pathogen. Clostridium perfringens was diagnosed as the cause of GIT disease Each interaction term was assessed for statistical significance with for 19 cases that involved a single etiological agent, and 9 cases that the main effects model, and the interaction terms with P , 0.05 were involved multiple etiological agents (12% of total cases; Table II). selected for inclusion in multivariable analysis. Interaction terms A total of 155 GIT cases were cultured for C. perfringens and the that were not significant in the final model were removed if removal organism was isolated in 133 (86%) cases. The GIT cases were less of the interaction term did not result in a significant change in the likely to be diagnosed with C. perfringens with increasing age of the likelihood ratio test. The logistic regression models were assessed piglets (P , 0.05) and if ETEC was detected (P = 0.065; Table III). for goodness-of-fit, which was indicated by P . 0.05 in the Pearson’s Clostridium perfringens was more likely diagnosed for GIT cases that chi-squared test. occurred in the winter compared to the fall and spring (P , 0.05; Table III), and more likely diagnosed for GIT cases that occurred Results in the summer compared to the fall (OR = 0.13, P = 0.022) and the spring (OR = 0.13, P = 0.029). The odds of a GIT case being diag- A total of 237 GIT cases involving the submission of live or dead nosed with C. perfringens was greater in 2010 compared to the years piglets, 1 to 7 d of age were submitted to the AHL, from 2001 to between 2002 and 2007 (P , 0.10). Clostridium perfringens isolates 2010. The number of these GIT cases submitted per year ranged from 40 cases were genotyped for major toxin genes and cpb2. from 10 to 39, with a mean of 23.7 6 7.97 cases per year. A successful All isolates from the 40 cases belonged to toxinotype A and cpb2 diagnosis was defined as a gastrointestinal case in which the etiologi- was detected in the C. perfringens isolates of 38 (95%) cases. Of the cal agent was identified. The diagnostic success rate for GIT cases 40 GIT cases where C. perfringens isolates were genotyped, 17 cases submitted per year is indicated in Figure 1. The annual prevalence were diagnosed with C. perfringens as the etiological agent (42.5%). for GIT cases diagnosed with single etiological agents is indicated Genotyping was not done on the isolates in 11 cases diagnosed with for ETEC, C. perfringens, rotavirus, Cystoisospora suis, C. difficile, and C. perfringens as the etiological agent. TGE virus (Figure 2). Rotavirus was diagnosed as the cause of GIT disease for 18 cases There were 79 (33%) GIT cases submitted to the AHL in which an that involved a single etiological agent, and 10 cases that involved etiological agent was not identified. There were a total of 51 (22%) multiple etiological agents (12% of total cases). Rotavirus was more GIT cases in the fall, 50 cases in the spring (21%), 46 (19%) cases likely diagnosed for GIT cases that occurred in the fall compared to in the summer, and 90 (38%) cases in the winter. The diagnosis of spring (OR = 0.14, P = 0.012) and the summer (OR = 0.15, P = 0.016). pathogens involved in each GIT case is given in Table I. Clostridium difficile was diagnosed as the cause of GIT disease for Enterotoxigenic E. coli infection was diagnosed as the cause of 10 cases that involved a single etiological agent and 11 cases that GIT disease for 63 cases that involved a single etiological agent, and involved multiple etiological agents (9% of total cases). Clostridium 10 cases that involved multiple etiological agents (31% of total cases; difficile was diagnosed less with increasing age of the piglets (OR = Table I). Enterotoxgenic E. coli was less likely to be recovered from 0.68, P = 0.031). Clostridium difficile was diagnosed less if ETEC was a GIT case if C. difficile, C. perfringens, or rotavirus were detected detected (OR = 0.085, P = 0.045), but diagnosed more if Salmonella (P , 0.05; Table II). Enterotoxgenic E. coli was more likely to be sp. was detected (OR = 20, P = 0.007). diagnosed for GIT cases that occurred in the winter compared to the Cystoisospora suis was diagnosed as the cause of GIT disease for spring and the summer (P , 0.05; Table II). There was a more likely 13 cases that involved a single etiological agent and 2 cases that tendency for a GIT case to be diagnosed with ETEC between years involved multiple etiological agents (6% of total cases). Coccidiosis 2001 and 2004 than in 2009 (P , 0.05; Table II). was diagnosed more in GIT cases that occurred in the summer

256 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Table I. Pathogens involved in gastrointestinal tract (GIT) disease cases for 1- to 7-day- old pigs submitted to the Animal Health Laboratory, University of Guelph, from 2001 to 2010 Number of cases Number of cases for multiple for single etiological etiological diagnosis diagnosis (% of total (percent of total Pathogens involved in GIT case GIT cases) GIT cases) Escherichia coli (n = 73) 63 (26.6) 10 (4.2) Clostridium perfringens (n = 28) 19 (8.0) 9 (4.3) Rotavirus (n = 28) 18 (7.6) 10 (4.2) Clostridium difficile (n = 21) 10 (4.2) 11 (4.6) Cystoisospora suis (n = 15) 13 (5.5) 2 (0.84) Transmissible gastroenteritis virus (n = 7) 6 (2.5) 1 (0.42) Salmonella sp. (n = 6) 2 (0.84) 4 (1.7) Porcine reproductive and respiratory 2 (0.84) 1 (0.42) syndrome virus (n = 3) Cryptosporidium parvum (n = 2) 1 (0.42) 1 (0.42) Enterococcus durans (n = 2) 1 (0.42) 1 (0.42)

Table II. Logistic regression comparing the association of enterotoxigenic Escherichia coli diagnosis and independent variables 95% Confidence Independent variables Odds ratio Standard error interval P-value Clostridium difficile detection No Referent Yes 0.11 0.125 0.013–.97 0.047 Clostridium perfringens detection No Referent Yes 0.20 0.148 0.049–0.85 0.029 Rotavirus detection No Referent Yes 0.29 0.170 0.95–0.91 0.034 Seasona Winter Referent Spring 0.38 0.171 0.16–0.92 0.032 Summer 0.37 0.177 0.15–0.94 0.037 Yeara 2009 Referent 2001 9.6 7.60 2.0–45 0.004 2002 9.6 7.56 2.1–45 0.004 2003 12 8.68 2.7–50 0.001 2004 5.0 3.95 1.1–24 0.042 2005 4.0 3.03 0.89–18 0.07 2007 4.3 3.39 0.93–20 0.062 a Parameters for fall, 2006, 2008, and 2010 are not shown (P . 0.1).

­compared to spring (OR = 0.26, P = 0.03) and winter (OR = 0.046, Porcine reproductive and respiratory syndrome was diagnosed as the P = 0.004). Transmissible gastroenteritis virus was diagnosed as the cause of GIT disease for 2 cases that involved a single etiological agent cause of GIT disease for 6 cases that involved a single etiological agent and 1 case that involved multiple etiological agents (0.01% of total and 1 case that involved multiple etiological agents (3% of total cases). cases).

2000;64:0–00 The Canadian Journal of Veterinary Research 257 FOR PERSONAL USE ONLY

Table III. Logistic regression comparing the association of Clostridium perfringens diagnosis and independent variables 95% Confidence Independent variables Odds ratio Standard error interval P-value Average age of submitted piglets (d) 0.69 0.106 0.51–0.93 0.016 Seasona Winter Referent Spring 0.13 0.116 0.024–0.73 0.021 Fall 0.13 0.108 0.028–0.65 0.013 ETEC detection No Referent Yes 0.27 0.192 0.068–1.1 0.065 Yeara 2010 Referent 2002 0.074 0.0967 0.0059–0.94 0.045 2003 0.089 0.0956 0.011–0.73 0.025 2004 0.056 0.0713 0.0046–0.68 0.024 2005 0.19 0.163 0.037–1.0 0.052 2006 0.040 0.0500 0.0033–0.47 0.011 2007 0.078 0.0840 0.0095–0.64 0.018 2009 0.17 0.159 0.029–1.0 0.056 a Parameters for 2001 (no cases of C. perfringens diagnosed), 2008, and summer (P . 0.1) are not shown. ETEC — enterotoxigenic E. coli.

Discussion was reported by the AHL to increase the diagnosis of rotavirus B and rotavirus C infection (12). The AHL reported that from 2010 to Cases of neonatal diarrhea were most frequent during winter, a April 2011, 17% of cases were positive by rotavirus latex agglutina- well-established finding in countries with harsh winters (9). The tion, whereas 68% of cases were positive by rotavirus group RT-PCR major known pathogens that contributed to neonatal diarrhea in tests. Of these, 6 cases were positive for rotavirus A, 8 cases were piglets from 2001 to 2010 were ETEC, rotavirus, C. perfringens, and positive for rotavirus B, 1 case was positive for rotavirus A and B, C. difficile. The combined GIT cases involving these pathogens con- and 1 case was positive for rotavirus A and C (12). tribute to more than half of the GIT cases submitted to the AHL. Studies conducted on the frequency of different swine enteric Interestingly, there was marked annual variation in specific etiologi- pathogens have also shown that rotavirus is a common pathogen cal diagnoses although the reason for this annual variation is unclear. isolated from diarrheic neonatal piglets (1,13,14). Our study reported Enterotoxigenic E. coli was a common cause of neonatal diarrhea that rotavirus was more likely to be diagnosed for GIT cases that from 2001 to 2005 in cases involving single etiological agents, but occurred in the fall compared to the spring and summer, suggesting there was a relative decrease in the number of cases diagnosed that rotaviral infection often occurred in the fall season. with this agent from 2006 onward. Enterotoxigenic E. coli was more Although C. perfringens type C infections in swine herds have been frequently diagnosed for GIT cases that occurred in the winter com- identified in some geographic areas (2), C. perfringens type C was not pared to the spring and summer. Post-weaning ETEC enteritis has identified in any Ontario GIT cases submitted to the AHL between been documented to occur most frequently in the fall and winter 2001 and 2010. Clostridium perfringens type A was the second most (10), and another study reported that cases of neonatal colibacil- frequent etiological diagnosis of GI illness in piglets within the first losis in piglets were least prevalent in the spring compared to other week of their life in this study. Analysis showed that the likelihood seasons (11). of C. perfringens diagnosis decreased with increasing piglet age. It is Rotavirus was an important pathogen contributing to neonatal thought that C. perfringens associated diarrhea usually occurs within diarrhea in piglets. The detection of rotavirus may be influenced by 48 h after birth (2). However, it is possible that diarrhea associated the sampling time of infected piglets during the course of disease, with C. perfringens type A is misdiagnosed, since the pathogenic with the highest viral load occurring at the acute stage. The AHL mechanism is unknown and current diagnostic methods are not has reported that approximately 50% of rotavirus cases were attrib- specific for C. perfringens-associated enteritis. In fact, C. perfringens uted to rotavirus group A (6); it is possible that rotavirus infection was considered as the etiological diagnosis for only 42.5% of cases caused by other groups (B or C) were underdiagnosed at the AHL, where genotyping was done on the isolates, without concurrent since PCR methods available for detecting these rotavirus groups histologic evidence of bacterial proliferation at the intestinal mucosal were only available from 2010 onward. The introduction of RT-PCR surface, positive cpb2 status, or both.

258 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Previous studies have indicated an association with cpb2-positive reduced submissions because of poor economic returns for Ontario C. perfringens type A and neonatal diarrhea in piglets (4,5). However, swine producers from from 2007 to 2010. Furthermore, no etiologi- a recent study showed that the beta2 toxin (CPB2) was detected in cal agent was determined in one third of the GIT cases submitted to the intestinal content of both healthy and diarrheic piglets, and that the AHL. The lack of diagnosis for some GIT cases is possibly due enumeration of cpb2-positive C. perfringens type A did not provide to inappropriate samples submitted to the diagnostic laboratory; a useful diagnosis for C. perfringens type A enteritis (14). Because piglets in the acute stage of disease have the highest pathogen load cpb2-positive C. perfringens were isolated from diarrheic piglets and, therefore, are more likely to yield useful diagnostic informa- where an etiological agent was not identified in that study, it was tion (2). It is also possible that there may be emerging pathogens suggested that C. perfringens type A enteritis was often misdiagnosed causing neonatal diarrhea in piglets that are not detected by current in cases where no other known pathogens except C. perfringens was diagnostic methods. Prevalence studies on infectious causes of diar- identified (14). rhea in piglets conducted in different countries could not identify Cases of neonatal diarrhea attributed to C. difficile appeared etiological agents in 17% to 58% of the pigs examined (1,13,17). sporadic and approximately half of the total cases associated with Similarly, a recent study conducted in Ontario did not identify the C. difficile involved a complex of other pathogens. For example, etiological agent in 38% of neonatal diarrheic pigs (14). It is possible C. difficile was more likely to be diagnosed for GIT cases when that some GIT cases had an etiological agent identified, but were Salmonella spp. was also isolated. Some studies indicate that not recorded electronically due to data entry error. Another reason C. difficile is an emerging pathogen in neonatal diarrhea (1,14). The for the lack of diagnosis in some GIT cases is that the producer may bacterial culture test for C. difficile became available at the AHL in choose not to pursue additional laboratory tests for cost reasons. 2008 (7); the difficulty of culturing this anaerobic organism sug- There may be other unknown pathogens involved in undiagnosed gests that cases attributed to C. difficile may be underdiagnosed (3). GIT cases submitted to the AHL and further investigation is required There is currently no consensus on the best method for the diagnos- to determine the current causes of diarrhea in piglets from Ontario ing C. difficile enteritis in humans, and guidelines for diagnosing swine farms. The results of the study may not represent the actual C. difficile enteritis in animals are unavailable (15). A recent study prevalences of pathogens causing neonatal diarrhea in piglets in the reported that the commercial tests available for detection of C. difficile field. toxins have lower sensitivity and specificity in piglets compared to This study identified several current pathogens involved in neo- humans (15). A 2-step algorithm was recommended for the detec- natal diarrhea for Ontario swine farms. Clostridium difficile appears tion of C. difficile in herds of swine, similar to that used in medical to be an emerging pathogen, and ETEC and rotavirus remain diagnostic laboratories. A first step suggested was the use of a test pathogens of concern for neonatal diarrhea in piglets. Further with high negative predictive value such as real-time PCR, followed research in the diagnostic method of these pathogens may be useful by confirmation of the positive results with toxigenic culture as in improving the diagnostic rate for GIT cases. The data suggested a second step (15). However, the study was more focused on the that C. perfringens type A may be an important pathogen for neonatal presence or absence of the organism in herds rather than specific diarrhea in piglets, but the current lack of specific diagnostic criteria diagnosis of C. difficile enteritis in piglets (15). Clostridium difficile for this pathogen made it difficult to determine the significance of toxins can be detected in healthy piglets and, therefore, its presence this pathogen. alone does not confirm the diagnosis. However, C. difficile may be an important pathogen causing subclinical infection in piglets (16). Acknowledgments Clearly, further work is required to develop a sensitive diagnostic approach for C. difficile enteritis in piglets. This project was supported by the OMAF and MRA-UG Cystoisospora suis was diagnosed as a single etiological agent in Agreement through the Animal Health Strategic Investment fund cases of neonatal diarrhea that occurred mainly from 2001 to 2005. (AHSI) managed by the Animal Health Laboratory of the University The general lack of diarrheal cases attributed to Cystoisospora suis of Guelph. from 2006 onwards suggests that it is currently not a major patho- gen of concern for 1- to 7-day-old piglets. This is in agreement with References reports that conclude coccidiosis is associated with a diarrhea that begins after 1 or 2 wk of age (13,17). Coccidiosis was diagnosed more . 1 Yaeger M, Funk N, Hoffman L. A survey of agents associated in GIT cases that occurred in the summer compared to the spring with neonatal diarrhea in Iowa swine including Clostridium dif- and the winter, in agreement with another report that indicated ficile and porcine reproductive and respiratory syndrome virus. neonatal piglet coccidiosis cases were most frequently identified J Vet Diag Invest 2002;14:281–287. in the summer (9). Cases neonatal diarrhea attributed to TGE virus 2. Schwartz KJ. Clostridium-associated diseases in swine. Proc Annu were sporadic and absent from 2008 onwards, indicating that it is Meet AASV. 2009:415–421. not a current concern for the Ontario swine industry. 3. Songer JG, Uzal FA. Clostridial enteric infections in pigs. J Vet The number of neonatal piglet GIT cases submitted to the AHL Diag Invest 2005;17:528–536. generally decreased in the 10-year period, which may reflect the 4. Garmory HS, Chanter N, French NP, Bueschel D, Songer JG, decrease in number of swine operations in Ontario during this Titball RW. Occurrence of Clostridium perfringens b2-toxin time, an actual decrease in prevalence of neonatal diarrhea due to amongst animals, determined using genotyping and subtyping improved farm management practices, or other factors, including PCR assays. Epidemiol Infect 2000;124:61–67.

2000;64:0–00 The Canadian Journal of Veterinary Research 259 FOR PERSONAL USE ONLY

. 5 Bueschel DM, Jost BH, Billington SJ, Trinh HT, Songer JG. 12. Carman S, Hazlett M, McEwen B, Rossov K. New Rotavirus Prevalence of cpb2, encoding beta2 toxin, in Clostridium perfrin- PCR test improves diagnostic rate for porcine diarrhea. AHL gens field isolates: Correlation of genotype with phenotype. Vet Newsletter 2011;15:13. Microbiol 2002;94:121–129. 13. Katsuda K, Kohmoto M, Kawashima K, Tsunemitsu H. Frequency 6. Carman S. New swine Rotavirus group A/C real-time RT-PCR of enteropathogen detection in suckling and weanling pigs with and swine Rotavirus group B RT-PCR tests available at the AHL. diarrhea in Japan. J Vet Diag Invest 2006;18:350–354. AHL Newsletter 2010;14:25. 14. Farzan A, Kircanki J, DeLay J, et al. An investigation into the 7. Slavic D. Clostridium difficile culture and toxin testing. AHL association between cpb2-encoding C. perfringens type A and Newsletter 2008;12:1. diarrhea in neonatal piglets. Can J Vet Res 2013;77:45–53. 8. Carman S, McEwen B, Fairles J. Tetracore Next generation 15. Keessen EC, Hopman NEM, van Leengoed LAMG, et al. NA/EU PRRSV multiplex real-time RTPCR is now available at Evaluation of four different diagnostic tests to detect Clostridium the AHL. AHL Newsletter 2010;14:15. difficile in piglets. J Clin Microbiol 2011;49:1816–1821. 9. Sanford SE, Josephson GKA. Porcine neonatal coccidiosis. Can 16. Yaeger MJ, Kinyon JM, Songer GJ. A prospective, case control Vet J 1981;22:282–285. study evaluating the association between Clostridium difficile 10. Amezcua R, Friendship RM, Dewey CE, Gyles C, Fairbrother JM. toxins in the colon of neonatal swine and gross and microscopic Presentation of postweaning Escherichia coli diarrhea in southern lesions. J Vet Diag Invest 2007;19:52. Ontario, prevalence of hemolytic E. coli serogroups involved, and 17. Wieler LH, Ilieff A, Herbst W, et al. Prevalence of enteropatho- their antimicrobial resistance patterns. Can J Vet Res 2002;66: gens in suckling and weanling suckling pigs with diarrhea in 73–78. southern Germany. J Vet Med B 2001;48:151–159. 11. Morin M, Turgeon D, Jolette J, et al. Neonatal diarrhea of pigs in Quebec: Infectious causes of significant outbreaks. Can J Comp Med 1983;47:11–17.

260 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Article

Species-specificity of equine and porcine Lawsonia intracellularis isolates in laboratory animals Francesca Sampieri, Fabio A. Vannucci, Andrew L. Allen, Nicola Pusterla, Aphroditi J. Antonopoulos, Katherine R. Ball, Julie Thompson, Patricia M. Dowling, Don L. Hamilton, Connie J. Gebhart Abstract Lawsonia intracellularis infection causes proliferative enteropathy (PE) in many mammalian species, with porcine and equine proliferative enteropathy (PPE and EPE) known worldwide. Hamsters are a well-published animal model for PPE infection studies in pigs. There is no laboratory animal model for EPE infection studies and it is not known whether there is species-specificity for equine or porcine isolates of L. intracellularis in animal models. The objective of this study was to determine whether it is possible to generate typical EPE lesions in hamsters after inoculation with an equine strain of L. intracellularis (EPE strain) and whether it is comparatively possible to generate PPE lesions in rabbits after inoculation with a porcine strain of L. intracellularis (PPE strain). In 2 separate trials, 4-week-old and 3-week-old weanling golden Syrian hamsters were challenged with EPE strains and compared to uninfected (both trials) and PPE-infected controls (Trial 2 only). Concurrently, 6 female New Zealand white juvenile rabbits were infected with PPE strain and observed concomitantly to 8 similar rabbits infected with EPE strain for a different experiment. Hamsters and rabbits were observed for 21 to 24 days post-infection (DPI), depending on the experiment. Neither infected species developed clinical signs. The presence of disease was assessed with diagnostic techniques classically used for pigs and horses: immune-peroxidase monolayer assay on sera; quantitative polymerase chain reaction (qPCR) detection of molecular DNA in feces; and hematoxylin and eosin (H&E) stain and immunohistochemistry (IHC) on intestinal tissues. Our results showed that EPE-challenged hamsters do not develop infection when compared with PPE controls (IHC, P = 0.009; qPCR, P = 0.0003). Conversely, PPE-challenged rabbits do not develop typical intestinal lesions in comparison to EPE-challenged rabbits, with serological response at 14 DPI being significantly lower (P = 0.0023). In conclusion, PPE and EPE strains appear to have different host-specificities for hamsters and rabbits, respectively. Résumé L’infection par Lawsonia intracellularis provoque une entéropathie proliférative chez de nombreuses espèces de mammifères; celle des porcins (EPP) et des équidés (EEP) sont connues mondialement. Les hamsters sont un modèle animal bien connu pour l’étude de l’EPP. Il n’existe pas de modèle animal de laboratoire pour étudier l’EEP, et on ne sait pas s’il y a spécificité d’espèce pour les isolats équins ou porcins de L. intracellularis dans des modèles animaux. L’objectif de la présente étude était de déterminer s’il est possible de générer des lésions typiques d’EEP chez les hamsters après inoculation d’une souche équine de L. intracellularis (souche EEP) et s’il est également possible de générer des lésions d’EPP chez des lapins après inoculation d’une souche porcine de L. intracellularis (souche EPP). Dans 2 essais séparés, des hamsters dorés syriens sevrés âgés de 4 semaines et de 3 semaines ont été inoculés avec des souches EEP, et ont été comparés à des témoins non infectés (les deux essais) et à des témoins infectés avec EPP (essai 2 seulement). Parallèlement, 6 jeunes lapines Nouvelle-Zélande ont été infectées par la souche EEP et observées de façon concomitante à 8 lapins similaires infectés par la souche EPP pour une expérience différente. Les hamsters et les lapins ont été observés pendant 21 à 24 jours après l’infection (JAI), en fonction de l’expérience. Aucune des espèces infectées n’a développé de signes cliniques. La présence de maladie a été évaluée par des techniques classiques de diagnostic utilisées pour les porcs et les chevaux : l’essai par immuno-peroxydase sur monocouche pour les sérums; la détection par réaction d’amplification en chaîne par la polymérase quantitative (qPCR) de l’ADN moléculaire dans les selles; la coloration hématoxyline-éosine et l’immunohistochimie (IHC) sur des tissus intestinaux. Nos résultats ont montré que les hamsters inoculés avec EEP ne développent pas d’infection comparativement aux EPP témoins (IHC P = 0,009; qPCR P = 0,0003). À l’inverse, les lapins inoculés avec EPP ne développent pas des lésions intestinales typiques comparativement aux lapins inoculés avec EEP, avec une réponse sérologique à 14 JAI significativement plus faible (P = 0,0023). En conclusion, les souches d’EPP et d’EEP semblent avoir des spécificités d’hôte différentes chez les hamsters et les lapins, respectivement. (Traduit par Dr. J.M. Dhillon) Introduction a wide range of domestic, wildlife, avian, and laboratory animal spe- cies (1–5). An intense proliferation is induced by intra-cytoplasmic Lawsonia intracellularis is a Gram-negative, obligate intracellular replication of L. intracellularis in the host’s enterocytes, principally in bacterium infecting the enterocytes of the large or small intestines of the jejunum and ileum, wherein the hyperplasic activity corresponds

Department of Veterinary Biomedical Sciences (Sampieri, Antonopoulos, Ball, Dowling, Hamilton) and Department of Veterinary Pathology (Allen), Western College of Veterinary Medicine, University of Saskatchewan, Saskatoon, Saskatchewan S7N 5B4; Department of Veterinary and Biomedical Sciences, College of Veterinary Medicine, University of Minnesota, Saint Paul, Minnesota 55108, USA (Vannucci, Gebhart); Department of Veterinary Medicine and Epidemiology, School of Veterinary Medicine, University of California, Davis, California 95616, USA (Pusterla); Canadian Light Source, Saskatoon, Saskatchewan (Thompson). Address all correspondence to Francesca Sampieri; telephone: (502) 727-3460; fax: (270) 767-7540; e-mail: [email protected] or [email protected] Received July 19, 2012. Accepted November 1, 2012.

2013;77:261–272 The Canadian Journal of Veterinary Research 261 FOR PERSONAL USE ONLY

to loss of function. Such lesions cause a variety of acute and chronic host-species to another will help researchers to choose appropri- clinical signs in affected individuals, ranging from malabsorption, ate animal models, thereby reducing costs, refining experimental diarrhea, depression, weight loss, abdominal pain, and even death methods and, ultimately, sparing animals. The specific objective of (6–8). It is known that L. intracellularis causes porcine and equine the present study was to determine whether it is possible to gener- proliferative enteropathies (PPE and EPE, respectively): the former ate typical EPE lesions in hamsters after inoculation with an equine represented a challenge for the swine industry for more than 50 y; strain of L. intracellularis (EPE strain) and whether it is comparatively the latter emerged more recently and, after sporadic and isolated possible to generate PPE lesions in rabbits after inoculation with a beginnings, is now diffused worldwide (1,8–10). Proliferative enter- porcine strain of L. intracellularis (PPE strain). opathies adversely affect both the porcine and equine industries and their impact goes beyond the presence of challenging clinical signs, Materials and methods as slow and long recoveries or severely reduced growth performance are common consequences (5,8,10). All the studies described were approved by the Animal Research Proliferative enteropathy has been studied experimentally in pigs Ethics Board of the University of Saskatchewan and conducted for years, using either porcine L. intracellularis isolates grown in vitro according to CCAC guidelines. or porcine tissues (11–12). Similarly, among laboratory animals, ham- sters have been known to be naturally and experimentally affected Hamsters by the disease since long before L. intracellularis was defined and In Trial 1, 29 pathogen-free golden Syrian hamster (Mesocricetus named in 1995 (13). Naturally affected hamsters undergo diarrhea auratus) weanlings (Strain 049, VAF Hamsters; Charles River Canada, (also known as “wet-tail”) and weight loss, with severe hyperpla- Pointe Claire, Quebec) were born at the Animal Care Unit, Western sia of the ileal segment in the small intestine, which is similar to College of Veterinary Medicine, University of Saskatchewan. The lesions observed in pigs (14). Hamsters have therefore been used hamsters originated from immuno-competent animal colonies moni- as an animal model for PPE studies, whether infection is produced tored by serology, polymerase chain reaction (PCR), bacteriology, through ileal homogenates harvested from naturally diseased pigs parasitology, or gross pathology tests for absence (pathogen-free or bacteria grown in pure cell culture (7,14–19). A few researchers and virus antibody-free, or VAF/Plus hamsters) of Reovirus, Sendai have attempted to infect other rodents (mice and rats) and chickens, virus, Lymphocytic choriomeningitis virus, Bordetella bronchiseptica, with variable success (20–22). Unlike PPE studies, the research on Streptococcus pneumonia, Campylobacter jejuni, Clostridium piliforme, EPE animal models is limited to a recently established foal model Corynebacterium kutscheri, Encephalitozoon cuniculi, Helicobacter spp., and to the preliminary results of a rabbit model, both obtained from Mycoplasma pulmonis, Salmonella spp., beta-hemolytic Streptococcus), the same virulent EPE strain (23–25). and Streptococcus spp. (serogroups A, B, and G); all intestinal hel- Due to the increasing costs of research for horses and pigs, which mints (Syphacia spp., Hymenolepis spp., etc.) and protozoa species are the main species affected, using laboratory animal models instead (Coccidia spp., Trichomonadaceae, Entamoeba spp., etc); dermathophy- of large animals could reduce expenses. Standardized welfare con- tosis and all arthropods ectoparasites (Strain 049, VAF Hamsters; ditions could still be maintained and environmental and infection Charles River Canada). control could be maximized over larger populations at one time, The hamsters were housed with their dams in rodent cages in a which could potentiate the statistical value of results. Although the Containment Level (CL)-1 Room until weaning (21 d from birth), “3Rs” tenet (replacement, reduction, and refinement) of the Canadian bedded on dust-free shavings, fed rodent chow pellets (5P00, Prolab Council for Animal Care (CCAC) is meant to responsibly control the RMH 3000, LabDiet; PMI Nutritional International, Brentwood, use of animals in research rather than reduce experimental costs, Missouri, USA) fresh water ad libitum, and maintained in standard animal models are advocated when in-vitro models have great husbandry conditions (12/12 h light/dark cycle and 20°C 6 2°C limitations, as in the case of L. intracellularis (26,27). The studies room temperature). Newborn hamsters were visually monitored described here were part of a larger project investigating EPE in once daily for the first 5 days of life and then twice daily until wean- relation to a novel pharmaceutical compound, which contains the ing. At weaning, they were weighed, ear-punched, and randomly post-transition metal gallium, whose environmental impact after assigned to 2 groups: uninfected controls (9 hamsters) and EPE- excretion has not been clarified (28). To that end, modeling EPE on strain-infected (20 hamsters). Hamsters were then separated into smaller surrogate animals was considered to be liberally in agree- pairs or triplets, according to group (controls versus EPE-infected) ment with the “3Rs” tenet. and gender. Hamsters were inoculated with L. intracellularis 1 wk Our interest was to verify whether L. intracellularis exhibits a after weaning. After inoculation, hamsters were housed in rodent species-restricted host-susceptibility or if one given strain could cages in a CL-2 Room, bedded on dust-free shavings in identical generate infection in multiple species (4,5). Understanding the role husbandry conditions, and provided with the same feed and water of cross-infection in the pathogenesis of proliferative enteropathy sources, as previously described. would also be epidemiologically useful in order to define which In Trial 2, an additional 24 pathogen-free and VAF golden Syrian species may amplify the bacterial shedding in the environment. Such hamsters (Strain HsdHan AURA; Harlan Laboratories, Indianapolis, an understanding could help to determine to what extent and for Indiana, USA) were born at the same Animal Care Unit facility how long these species would represent a challenge for epidemiol- described for Trial 1. For both experimental subjects and their dams, ogy boundaries in horse and pig farms. Even more importantly, housing, husbandry, and daily monitoring conditions were identical determining the ability of L. intracellularis to cross-infect from one to those in Trial 1. The health monitoring of the immuno-competent

262 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

originating hamster colonies included serology, gross pathology, and Inoculum preparation microscopic analysis tests to verify the absence of pathogens and Control inoculum (medium only) consisted of buffered sucrose/ related antibodies for Reovirus, Sendai virus, Lymphocytic chorio- phosphate/glutamate (SPG) solution and was administered to the meningitis virus, Simian Virus 5, Bordetella bronchiseptica, Clostridium uninfected hamster groups in both trials (controls), but not to the piliforme, Corynebacterium kutscheri, Salmonella spp., Campylobacter rabbits. jejuni, L. intracellularis, Helicobacter spp., Klebsiella spp., Mycoplasma EPE strain — Hamsters in Trials 1 and 2 and the EPE-infected pulmonis, Proteus mirabilis, Pseudomonas aeruginosa, Pneumocystis rabbits were challenged with the equine strain of L. intracellularis spp., Pasteurella pneumotropica, beta-hemolytic Streptococcus spp., E40504, prepared as described in previous studies (23,29). The Streptococcus pneumoniae, Staphyloccus aureus; ectoparasites, and L. intracellularis endoparasites, Demodex spp., and Encephalitozoon cuniculi (Strain infectious material originated from ileal mucosal HsdHan AURA; Harlan Laboratories). scrapings of a foal that succumbed to EPE. It was subsequently On day 21, hamsters were weaned, ear-punched, weighed, cultured on McCoy cells (ATCC CRL 1696) in a defined environ- and randomly assigned to 3 groups: controls — uninfected ment (27,30). Infectious virulent isolate from low passage pure (6 hamsters); PPE-strain group — inoculated with a porcine strain cell-cultures was prepared and suspended in buffered SPG medium of L. intracellularis (9 hamsters); and EPE-strain group — inoculated for all the EPE-infected animals, as described in previous studies with an equine strain of L. intracellularis (9 hamsters). Hamsters in (23,29). In the hamster trials, inoculum concentration was between 7 7 Trial 2 were inoculated at weaning to increase the impact of stress 4.5 3 10 bacteria/mL and 7.5 3 10 bacteria/mL (31). For the on disease reproduction. positive control, EPE-infected rabbits were inoculated at a dose of 5.5 mL/rabbit to give a final inoculum of between 1.3 and Rabbits 2.5 3 108 bacteria per rabbit (25). Six 4- to 5-week-old female New Zealand white, pathogen-free PPE strain — For the Trial 2 hamsters and PPE-infected rabbits, rabbits (Oryctolagus cuniculus) (Strain 052 VAF Rabbits; Charles inocula were prepared from PPE lesions harvested from an infected River Canada) were used. Animals originated from colonies tested pig’s ileal mucosa (PHE/MN1-00) and cultured on McCoy cells regularly by serology, gross and histo-pathology, bacteriology, and (ATCC CRL 1696) in a defined environment (27,29,30). Infectious parasitology tests for absence of pathogen agents and related anti- inocula were prepared as described previously from low passage bodies (VAF) [Lymphocytic choriomeningitis virus, Parainfluenza pure cell-culture isolate. The main difference between inocula virus 1 and 2, Rabbit hemorrhagic disease virus, Reovirus, Rotavirus, used for hamsters and rabbits was the final concentration and the Bordetella bronchiseptica, CAR-Bacillus, Clostridium piliforme, volume. For hamsters, the final concentration of the inocula was 7 Campylobacter jejuni, Corynebacterium kutscheri, Helicobacter spp., 3.2 3 10 bacteria/mL and the volume was as close as possible to Klebsiella spp., L. intracellularis, Mycoplasma pulmonis, Proteus mirabilis, 1 mL, whereas the volume was 8.5 mL per rabbit (target dose of 8 Pseudomonas aeruginosa, Pneumocystis spp., Pasteurella pneumotropica, inoculum was 4.25 3 10 bacteria per animal). Salmonella spp., beta-hemolytic Streptococcus spp., Streptococcus For hamsters in Trial 1, a total dose of 0.45 mL/hamster (approx. 7 pneumonia, Staphylococcus aureus, Encephalitozoon cuniculi, Giardia 3.375 3 10 bacteria/hamster), suspended in commercial fruit- spp., Toxoplasma gondii, Treponema cuniculi, pinworms, and several flavored pudding (Jell-O; Kraft Foods, Canada), was administered tests for typical rodent diseases, but used annually in rabbits by to each animal once on day 0. In Trial 2, an equal inoculum dose the provider (Charles River Canada) for internal control purposes], was given, but no palatable medium was added. Subjects were as described for the previous study on the L. intracellularis rabbit 1 wk younger and lighter in body weight (BW), thus the total target model (25). Also, 8 additional does of the same age, weight, and volume (1 mL) was divided into 3 administrations over 2 d (once on origin (Charles River Canada) were used as a comparative control day 0, twice on day 1). The dosing volume calculations were based to the PPE-strain study, as they had been EPE-infected for a different on a maximum volume of 20 mL/kg BW and repeated inoculation experiment. was used to maximize exposure (32). All challenge material was The 2 studies were purposefully conducted at the same time in stored in a 220°C freezer and thawed overnight at 4°C, 12 h before order to compare the lesions at the peak of infection (14 DPI). No the experiments. sham-challenged rabbits were used in this experiment, as it has Inoculation procedures previously been shown that unexposed rabbits do not develop infec- tion (25). On arrival, all rabbits underwent a health check and were Hamsters — The hamsters were inoculated using 2 different tech- identified with a permanent marker felt-pen on the right ear. Rabbits niques. In Trial 1, hamsters were fed the inocula through a syringe were group-housed in 2 separate and isolated pens in different areas because of the palatable medium added. In Trial 2, hamsters were within the same CL-2 Room. Pens were contained by stainless steel inoculated by oral gavage to ensure that the entire intended dose bars and polycarbonate panels. Both EPE- and PPE-infected rabbits was administered to the smaller animals. were inoculated concurrently and managed identically, with strict Rabbits — Rabbits were inoculated after a 1-week mandatory reverse isolation procedures maintained for the PPE-strain group. acclimation period (University of Saskatchewan AUC internal guide- All rabbits were fed rabbit pellets (Co-op Whole Earth Rabbit Ration; lines). Before infection, pen-pooled fecal samples were collected from Federated Co-op, Saskatoon, Saskatchewan) and water ad libitum each separate pen (PPE-infected and EPE-infected rabbits, respec- and maintained in standard husbandry conditions, as previously tively). Each rabbit was weighed, its general health was assessed, a described. blood sample was collected for serology analysis, and intragastric

2000;64:0–00 The Canadian Journal of Veterinary Research 263 FOR PERSONAL USE ONLY

administration of infective challenge was done through a nasogastric Sample analysis tube, as described in a previous study (25). Macroscopic examination — Regardless of group and trial, shortly Collection of samples after euthanasia the alimentary tract of both hamsters and rabbits was examined visually from stomach to rectum. The sampling pro- Hamsters — Changes in body weight (BW) were monitored daily tocol for hamsters consisted of collecting from the terminal jejunum, in hamsters for 21 DPI in Trial 1 and for 24 DPI in Trial 2. In both ileum, ileo-cecal valve, cecal ampulla, and colon, as reported in a trials, demeanor, gross appearance, fecal consistency and quality, as previous study by Vannucci and colleagues (7). Fecal pellets were well as self- and mutual grooming were monitored twice daily with collected from the terminal colon and rectum. The sampling protocol a 4-step grading system, with 0 corresponding to normal and 3 to severely abnormal findings (Table 1a). Treats of apples and small for rabbits was identical to that described in a previous study of our amounts of fruit-flavored pudding (Jell-O; Kraft Foods) were given research group (25). twice daily to check the appetite and awareness of each animal. Lack Immunohistochemistry — Formalin-fixed sections were paraffin- of response to treats and to researchers’ voices was considered a sign embedded, cut, and stained by the streptavidin method, including of decreased well-being (33). anti-L. intracellularis-specific mouse monoclonal antibody (34). The In both trials, pooled fecal samples were collected from the cages positive labeled L. intracellularis-specific antigen in the crypts was once weekly and on the predetermined days of euthanasia, when evaluated with a previously published 5-grade IHC scoring system individual samples were collected from the sacrificed animals. (7,34). For each animal, the negative control for each tissue section Blood samples for serologic evaluation were collected at the time consisted of a corresponding tissue section IHC-labeled, but omit- of euthanasia by intracardiac puncture. In Trial 1, 4 EPE-infected ting the primary antibody. Furthermore, pig ileal tissues that were hamsters were euthanized at 7, 11, 14, 17, and 21 DPI; 1 control was known to be negative and positive for L. intracellularis infection were euthanized at 7, 11, and 21 DPI; and 3 controls were sacrificed at labeled with the murine anti-L. intracellularis monoclonal antibody 14 and 17 DPI. In Trial 2, 8 hamsters (3 from EPE group, 3 from PPE to confirm the antibody’s specificity and sensitivity, respectively. group, and 2 controls) were euthanized at 17, 21, and 24 DPI. In both Serology analysis — Anti-L. intracellularis-specific immunoglobu- trials, animals were euthanized with an inhalatory anesthetic over- lin G (IgG) in serum was measured by an immuno-peroxidase mono- dose of isofluorane (Isoflurane, 99.9%; Halocarbon Products, River layer assay (IPMA), as reported in a previous study by Guedes and Edge, New Jersey, USA) in a glass chamber. Selection for euthanasia colleagues (35). Positive serum samples were end-point titrated from was randomized within each group. To evaluate the presence of pro- 1:30 up to 1:1920. Control samples consisted of serum from a rabbit liferative enteropathy lesions, sections of the ileum, jejunum, cecum, before (negative control) and after (positive control) hyperimmuni- and colon were harvested at necropsy, prepared within histology cas- zation with L. intracellularis, purified from cell culture. Also, serum settes, and immediately placed in 10% phosphate-buffered formalin total protein concentration was measured with the refractometer solution for histopathologic (hematoxylin and eosin — H&E) and method to investigate changes over time. immunohistochemical (IHC) examination (7). Quantitative PCR analysis — Quantitative polymerase chain Rabbits — Changes in BW were monitored daily. Demeanor, gross reaction (qPCR) analysis was conducted on fecal samples as appearance, fecal consistency and quality, appetite, and self- and reported in a previous study by Pusterla and colleagues on the mutual grooming were monitored twice daily with a 4-step grading EPE foal model (23). The purified DNA was analyzed by qPCR system, with 0 corresponding to normal and 3 to severely abnormal for the presence of L. intracellularis aspA gene copies (4,36). For findings, as previously described (Table 1b) (25). Lack of interest in each target gene, 2 primers and an internal, fluorescent-labeled treats (apples and carrots, in this case) was again considered a sign TaqMan probe [59end, reporter dye FAM (6-carboxyflourescein), of decreased well-being (33). 39end, quencher dye TAMRA (6-carboxytetramethylrhoda- Once weekly and at the time of euthanasia, blood samples (1 mL mine)] was designed using Primer Express software (Applied each) for serology were collected with a sterile needle and syringe Biosystems, Foster City, California, USA) The “L. intra sys- from the ear’s central artery, after a local block and disinfection were tem” forward primer and reverse primer were “bcL. intra-114f” applied, as described in a previous study (25). Pooled fecal samples (CACTTGCAAACAATAAACTTGGTCTTC) and “bcL. intra-263r” were also collected from the pen once weekly and individual fecal (CATTCATATTTGTACTTGTCCCTGCA), respectively, with the samples were collected at the time of euthanasia. Two PPE-infected aspartate ammonia lyase (aspA) gene as target and “bcL. intra- rabbits per week were humanely euthanized with an intravenous 201p” (TCCTTGATCAATTTGTTGTGGATTGTATTCAAGG) as overdose (720 mg/rabbit) of pentobarbital (Euthanyl; Bimeda-MTC probe. The “Pan-Bacteria system” forward primer and reverse Animal Health, Cambridge, Ontario). To observe for evidence of primer were “PB.283f” (GGATGATCAGCCACACTGGA) and lesions and to collect multiple samples (1.25 to 2.5 cm long) from “PB.352r” (CCAATATTCCTCACTGCTGCC), respectively, with the gastrointestinal tract (GIT) for histopathology and IHC examina- 16S ribosomal ribonucleic acid (rRNA) as a target and “PB.305” tion, selection for euthanasia was randomized (25). Samples were (CCCGTAGGAGTCTGGACCGTGTCTCA) as a probe. TaqMan PCR immediately placed in phosphate-buffered 10% formalin solution systems were validated using 2-fold dilutions of genomic deoxyribo- and prepared routinely for staining. At 14 DPI, EPE-infected rabbits nucleic acid (gDNA) testing positive for the target genes. Dilutions were euthanized by the same means as the PPE-infected rabbits and were analyzed in triplicate and a standard curve was plotted against samples were collected following the protocols reported in a previ- the dilutions. The slope of the standard curve was used to calculate ous study (25). amplification efficiencies using the formula E = 10 1/-s-1. Each ­system

264 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Table 1a. “Humane Intervention Monitoring Parameters” format used as a monitoring spreadsheet for hamsters enrolled in the infection animal model for porcine and equine proliferative enteropathy during the validation phase. This table was used to record clinical observations for each hamster twice daily, except for body weight, which was assessed only once daily Experimental day: Date: Hamster code: Variable Individual score Daily Feed Intake AM 0 Same or increased from previous day 1 Slight decrease from previous day 2 Marked decrease from previous day PM 3 Complete anorexia Daily Water Intake AM 0 Same or increased from previous day 1 Slight decrease from previous day 2 Marked decrease from previous day PM 3 No water taken in Body Weight (BW) (average weight gain for young hamsters is 1 to 2 g/d or ≈ 10 g/wk) AM 0 Increasing weight ~ 20%/wk 1 Increasing weight ~ 10%/wk (Calculated gain 2 Maintaining weight, no weight gain from last BW 3 Rapid weight loss (20% to 25%/wk) or losing weight for more than 2 d measurement) Stool AM 0 Normal feces 1 Feces formed but softer than usual 2 Feces not formed but not liquid PM 3 Liquid feces (for . 2 d), leading to emaciation Physical Appearance AM 0 Normal 1 Lack of grooming 2 Rough coat, nasal/ocular discharge PM 3 Very rough coat, abnormally hunched posture, distended abdomen (. 3 d) Behavior AM 0 Normal, alert, playful, comes to front of cage 1 Hyperactive, agitated 2 Decreased activity, minor depression, unwilling to move PM 3 Depressed, head turned to corner of cage, still moves when stimulated needed to be greater than 95% efficiency to be considered. The DNA sample in a final volume of 12 mL. The samples were placed detection limit for “L. intra system” is between 5 to 10 copies/uL of in a 384 well plate and amplified in an automated fluorometer (ABI DNA (determined with plasmid DNA). The “L. intra system” detects Prism7900 HTA Fast; Applied Biosystems). Applied Biosystem’s all L. intracellularis, regardless of the host species. Known positive standard amplification conditions were used: 2 min at 50°C, 10 min controls and no template controls were run on every plate and met at 95°C, 40 cycles of 15 s at 95°C, and 60 s at 60°C. Fluorescent signals previously established standardization values. The quality of DNA were collected during the annealing temperature and cycle threshold was determined by the “PanBacteria system,” with a control value (CT) values extracted with a threshold of 0.1 and baseline values of under 30, to pass quality control (37). A real-time PCR assay target- 3 to 12 for all samples using SDS Software, version 2.2.1 (Applied ing a universal sequence of the bacterial 16S rRNA gene was used as Biosystems). Absolute quantitation was calculated by a standard quality control, i.e., efficiency of DNA purification and amplification, curve and expressed as copy numbers of the L. intracellularis aspA and as an indicator of fecal inhibition (38). gene/g of feces. RT-reaction and real-time TaqMan PCR — Each PCR reaction con- tained 20x primer and probes for the respective TaqMan system with Statistics a final concentration of 400 nM for each primer and 80 nM for the For the hamster trials, BW findings were analyzed through a TaqMan probe and commercially available PCR Mastermix (TaqMan Mann-Whitney test, as Trial 1 had 2 uneven groups, and Kruskal- Universal PCR Mastermix; Applied Biosystems), containing 10 mM Wallis test with Dunn’s post hoc testing, as Trial 2 had 3 uneven

Tris-HCl (pH 8.3), 50 mM KCl, 5 mM MgCl2, 2.5 mM deoxynucleo- groups. Immunohistochemistry (IHC) results were analyzed through tide triphosphates, and 0.625 U AmpliTaq Gold DNA polymerase contingency tables with the Fisher Exact Test to compare the 2 cat- per reaction; 0.25 U AmpErase UNG per reaction; and 1 mL of the egorical outcomes (lesions versus no lesions) and to estimate odds

2000;64:0–00 The Canadian Journal of Veterinary Research 265 FOR PERSONAL USE ONLY

Table 1b. “Humane Intervention Monitoring Parameters” format used as a monitoring spreadsheet for the rabbits enrolled in the rabbit animal model for porcine and equine proliferative enteropathy during the validation phase. This table was used to record clinical observations for each rabbit twice daily, except for body weight, which was assessed only once daily Experimental day: Date: Rabbit code: Variable Individual score Daily Feed Intake AM 0 Same or increased from previous day 1 Slight decrease from previous day 2 Marked decrease from previous day PM 3 Complete anorexia Daily Water Intake AM 0 Same or increased from previous day 1 Slight decrease from previous day 2 Marked decrease from previous day PM 3 No water taken in Body Weight (BW) (average weight gain for young rabbits is 200 g/wk) AM 0 Increasing weight ~ 200 g/wk 1 Increasing weight ~ 100 g/wk (Calculated gain 2 Maintaining weight, no weight gain from last BW 3 Losing weight for more than 2 d measurement) Stool AM 0 Normal feces 1 Feces formed but softer than usual 2 Feces not formed but not liquid PM 3 Liquid feces Physical Appearance AM 0 Normal 1 Lack of grooming 2 Rough coat, nasal/ocular discharge PM 3 Very rough coat, abnormal posture, ears low Behavior AM 0 Normal, alert, comes to front of cage 1 Minor depression but still moves when stimulated 2 Grinding teeth, depressed, head turned to corner of cage PM 3 Very depressed, almost unresponsive ratio and relative risks. For analyses of qPCR results, a 1-way analy- which suggests abnormal behavior. There was no difference in daily sis of variance (ANOVA) was used, with Bonferroni’s post hoc testing. weight gain between infected and uninfected hamsters in either Only descriptive statistics were used for PPE-infected rabbits, as trial (P = 0.12 for Trial 1; 0.88 for Trial 2). There was no difference the sample number was insufficient for statistical analysis. When between 3- and 4-week-old hamsters (P = 0.14), either due to age the comparison was applicable to the EPE-infected animals, Mann- (3- versus 4-week-old) or breeding colony (hamsters from Charles Whitney, Kruskal-Wallis, and unpaired t-tests were used (BW and River Canada versus those from Harlan Laboratories), or in infected serum total protein concentration). For all statistical analyses in and uninfected hamsters (for EPE group: P = 0.77 and for EPE and hamsters and rabbits, commercial software (GraphPad Prism 5.4; PPE combined group: P = 0.6). Applied Biosystems) was used and alpha was set at 5%. No clinical abnormalities were noted in the PPE-infected rabbits during the study. Two does showed signs of aggression toward other Results pen-mates; 1 improved over time and 1 developed this behavior toward the end of the trial. No significant depression or suppression in BW gain was noted in any PPE-infected rabbit, not even at around Clinical appearance 14 DPI (peak of disease), as was observed in EPE-infected rabbits No hamsters showed signs of diarrhea throughout either trial (25). Three PPE-infected rabbits had negative BW gains at different and their enthusiastic interest in treats was never diminished (33). stages, which were unrelated to each other and quickly compensated The only behavioral change was a mild to moderate hyperactivity for (within 1 d) as the total BW gain of PPE-infected rabbits (7 DPI: that was noted after 14 DPI, as some PPE-infected animals rear- 171 to 309 g; 14 DPI: 506 to 667 g; 21 DPI: 781 to 871 g) was above ranged the entire cage’s content within a few hours in the daylight, the average of 200 g/wk that was recommended by the provider

266 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

0.15

0.10

0.05 Body weight gains (kg/day)

0.00 EPE strain PPE strain Figure 2. Example of the macroscopic appearance of a tract of jejunum, Figure 1. Comparison of body weight gain in 5-week-old rabbits approximately 3.5 cm long (bar scale = 1 cm) in rabbits infected with infected with the EPE strain and those infected with the PPE strain of the PPE strain of L. intracellularis. Note that the serosal edema and L. intracellularis. By 14 DPI, the PPE-infected rabbits gained more weight “cobblestone” appearance on the anti-mesenteric aspect of the sample overall than their EPE-infected counterparts (P = 0.04). The mean in the are similar to the macroscopic lesions of the EPE infection model in box plot is represented by a cross hair, whereas the whiskers represent rabbits, although such serosal appearance was not related to mucosal the 5 to 95 percentile interval. Note that the preliminary phases of the lesions in the PPE-infected rabbits. EPE infection model in rabbits show a trend of reduced growth perfor- mance around the infection peak at approximately 14 DPI, which is not matched by the PPE-infected rabbits. of inducing IHC detectable lesions in hamsters after PPE-strain inoculation was 3 times higher than after inoculating with EPE- (Charles River Canada). A comparison between daily average BW strain, with an odds ratio of 35.29 in favor of PPE-strain infection. gain in the EPE and PPE groups at 14 DPI (Figure 1) showed that The results of the challenge in Trial 2 are represented in Figure 4. EPE-does had a suppression in BW gain (P = 0.036). One last observation about EPE-strain is that there was no differ- ence (P = 0.28) found in either trial related to age (3-week-old versus Gross pathology 4-week-old), level of stress (inoculated a week after weaning versus at There were no gross lesions in the intestines of any hamsters in weaning), or method of inoculation (oral feeding versus oral gavage), either trial. In Trial 1, more abdominal fat was noted in older ani- even though the relative risk appears 3.3 times higher in younger, mals. Intestinal content appeared adequate for the tracts examined stressed hamsters. and the diameter of the intestinal sections (ilea) at the level of the In the PPE-infected rabbit experiment, no characteristic lesions ileo-cecal valve in Trial 1 controls was smaller (P = 0.0011) than the were detected through IHC labeling at any stage of the study. Only ilea in EPE-infected hamsters (1.5 to 2.5 mm versus 2 to 3 mm). In fragments of reactive DNA were noted at the level of the lamina Trial 2, however, the diameter of the jejunum and ileum in hamsters propria (Figures 5a and 5b). In the EPE-infected rabbits that were was the same in EPE-infected, PPE-infected, and control groups euthanized at the same time, IHC detectable lesions were found (P = 0.9 and 0.26, respectively). in several rabbits in at least 2 GIT sections (jejunum and ileum) In PPE-infected rabbits, mild edematous changes of the serosal (unpublished data), which mirrored the results of the initial model layer of the terminal portion of the jejunum and ileum were noted development (25). and were comparable to the serosal findings for EPE-infected rab- bits (Figure 2). No rugae were observed in the mucosa, however, and Serology by 14 DPI, only the serosal and mucosal edema were apparently Control samples for the IPMA procedure consisted of serum increased. from a rabbit before (negative control) and after (positive control) hyperimmunization with L. intracellularis, purified from cell culture. Immunohistochemistry No IPMA serology results consistent with immune response were In Trial 1, only 1 out of the 20 EPE-infected hamsters showed noted in hamsters in either trial, whether they were terminated at mild (grade 1) IHC labeling, which statistically was not different 21 or 24 DPI. (P = 1.0) than controls. The odds ratio showed 2.586 higher chances No antibody titers were evident in the PPE-infected rabbits until of infecting a hamster when it was challenged with the inoculum. 14 DPI. The IgG titers at that time (Figure 6) appeared significantly In Trial 2, a statistical difference (P = 0.013) was noted between the lower (P = 0.0023) in PPE-infected rabbits (range: 60 to 960) than in infected and uninfected hamsters, with greatly reduced (0.04) odds the EPE-infected group of rabbits (range: 480 to 1920) of the same ratios of having PE lesions in uninfected animals. Comparing the age, gender, litter, and housing conditions. Serum total protein con- presence of IHC-labeled lesions between PPE-infected hamsters and centration was measured to investigate changes, as decreased values uninfected controls, however, increased the statistical significance are reported in foals naturally and experimentally affected by EPE (P = 0.0003), whereas comparing PPE-strain and EPE-strain groups (5,23). No statistical difference was noted over time (P = 0.88, as it showed that the PPE-strain had higher ability than the EPE-strain ranged from 49 to 60 g/L on day 0; 45 to 53 g/L at 7 DPI; 40.5 to to infect hamsters (P = 0.009) (Figures 3a and 3b). The relative risk 49 g/L at 14 DPI; and 47.5 to 53.2 g/L at 21 DPI) in PPE-infected

2000;64:0–00 The Canadian Journal of Veterinary Research 267 FOR PERSONAL USE ONLY

b a

Figure 3. Comparison of IHC labeling in the EPE-infected and PPE-infected hamsters. Staining: streptavidin method, including anti-L. intracellularis-specific mouse monoclonal antibody, with 3-amino-9-ethylcarbazol (AEC) substrate-chromogen and counterstained with Mayer’s hematoxylin to observe for the presence of the antigen within the cells. Scale bar = 100 mm. a — Ileum of an EPE-infected hamster. Note that distribution of antigen (red dots) in the mucosal enterocytes of ileal villi is very limited and no antigen is detectable within the enteric crypts. b — Ileum of PPE-infected hamster. Note the much larger distribution of detectable antigen in the mucosal enterocytes.

Conversely, in Trial 2 (Figure 7b), despite an initial higher fecal lesions shedding of L. intracellularis for EPE-infected hamsters by 7 DPI, the EPE strain no lesions overall comparison of the EPE-strain and PPE-strain over time, from day 0 to 24 DPI, showed higher shedding in the PPE-strain group (P = 0.0003) for both time and infection factors (Figure 7c). A comparison between the detection of PPE-strain versus EPE- PPE strain strain L. intracellularis in rabbits is shown in Figure 8, which indicates higher fecal shedding in the EPE-infected rabbits (P = 0.0089). Type of exposure Type Discussion Controls These studies show that L. intracellularis cross-infection can occur in laboratory animals, but not with ease. The goal was to reproduce 0 1 2 3 4 5 6 7 8 9 10 PE disease, or lesions, from equine and porcine strain isolates in Number of hamsters order to individuate a surrogate species small enough to be cost- Figure 4. In Trial 2, the predominance of infection obtained in hamsters effective, yet large enough to endure humane research procedures, inoculated with the PPE-strain of L. intracellularis versus the EPE-strain and uninfected controls is shown. Note that all the PPE-challenged such as repeated blood samples. hamsters were showing IHC-detectable antigen in typical lesion sites The PPE infection model in hamsters was considered the term (enteric crypts and mucosal enterocytes), whereas only 3 out of 9 ham- of comparison for our hamster studies. Replication of infection sters showed mild EPE lesions (not in the intestinal crypts) after the EPE-strain challenge. in hamster weanlings from a pure cell culture of PPE-strain of L. intracellularis that was isolated from naturally diseased pigs rabbits, although a decreased concentration was noted in at least was in agreement with previous publications, which showed that 4 rabbits at around 14 DPI, which was soon recovered by the 2 sur- inoculation of hamsters yielded a typical PE infection (7,14,16,17). In vivor rabbits by 21 DPI. theses studies, no clinical signs were noted in the infected hamsters, as is commonly reported with pure cell culture inoculation, both in Quantitative PCR hamsters and pigs (7,17,30). The only exception was some degree of On day 0 of Trial 1, 5 hamsters from the EPE-infected group hyperactivity during daylight hours, which is atypical for hamsters showed positive fecal qPCR, 1 being quite high (693609.1 aspA as they are nocturnal animals. In addition, intestinal lesions were gene copies/g feces — see Figure 7a). For Trial 1, qPCR results detected by H&E and confirmed with IHC labeling. Quantitative showed no difference (P = 0.051) over time, when compared to the PCR showed that L. intracellularis DNA was shed in feces in expo- “day 0” group, although at the limits of significance (Figure 7a). By nentially increasing concentrations throughout the study (24 DPI), 21 DPI, however, 3 out of 4 infected hamsters were still shedding which demonstrated active replication of the bacteria in the host’s L. intracellularis. After infection with EPE-strain L. intracellularis, intestine (7,14,18,19,39). a comparison between 3- and 4-week-old hamsters showed The leading hypothesis supporting the idea of the EPE-strain of that younger hamsters have a higher likelihood of shedding L. intracellularis being capable of infecting hamsters was based not L. intracellularis in feces (P , 0.0001). only on the similarities of the bacterial genomic characteristics (98%

268 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

b

a

Figure 5. Comparison of IHC labeling in EPE- and PPE-infected rabbits. Staining: strepta- vidin method, including anti-L. intracellularis-specific mouse monoclonal antibody, with 3-amino-9-ethylcarbazol (AEC) substrate-chromogen and counterstained with Mayer’s hematoxylin to observe for typical proliferative lesions of the intestinal epithelium and for the presence of the antigen within the cells. Scale bar = 100 mm. a — Example of ileum of an EPE-infected rabbit at 14 DPI. Note the wide distribution of IHC-detectable antigen in the mucosal enterocytes over ileal villi and within enteric crypts. b — Example of cecum of a PPE-infected rabbit at 14 DPI. An ileal tissue counterpart for the PPE strain was unavailable as no antigen was detected anywhere, except in the cecum. Note that not only is the distribution of detectable antigen in the mucosal enterocytes minimal to none, but also the antigenic material appears scattered at the level of the lamina propria, where it is destroyed after being engulfed and processed by macrophages.

2500 the 20 inoculated hamsters became contaminated (see Figure 7a, day 0), despite rigorous experimental procedures being followed in handling each subject. Although such an outcome was unexpected, 2000 potential contamination of the feces with fresh inoculum was a

L. intracellularis plausible explanation. Interestingly, L. intracellularis DNA material

1500 reached values higher than those detected in the first 11 DPI, with maximum values ranging from 2.0 3 105 aspA gene/g. In the entire experiment, the DNA concentration never rose above the order of 1000 magnitude of 105, which is in contrast to the concentrations detected in PPE-infected hamsters (largely over 2 orders of magnitude higher at wk 2 and 3 post-infection). 500 In Trial 2, the component of stress, which was carefully avoided in Trial 1, as well as young age was added as hamsters were infected

IPMA titers for antibodies versus immediately after weaning. Despite 2 attempts, cross-infection of 0 EPE strain PPE strain hamsters with the EPE strain was unsuccessful, apart from a mild Figure 6. Immuno-peroxidase monolayer assay (IPMA) findings indi- and inconsistent fecal shedding, which even at its maximum value cating antibody response at 14 DPI in rabbits infected with the EPE was never as high as the shedding in PPE-infected hamsters (at dis- strain and those infected with the PPE strain of L. intracellularis. The ease peak, from 17 to 24 DPI). This is particularly striking compared serological response at 14 DPI in EPE-infected rabbits is visibly differ- ent (P = 0.0023) than in PPE-infected rabbits. The whiskers on the bar to the textbook-like development of fecal shedding and lesions in indicate standard deviations. PPE-infected hamsters. Although our results originated from a small population, the fact that PPE-challenged hamsters have over between PPE and EPE strains), but also on how easily the PPE-strain 35 times more chance to develop lesions than the EPE-challenged can cross-infect from pigs to hamsters (7,17,19,39). Furthermore, ones is an obvious limitation. This should discourage the further use similarities in the GIT of hamsters, rabbits, and horses, such as hind- of hamsters to replace horses in EPE infection models. On a wider gut fermentation and a proportionally large cecum, were considered perspective, these results constitute evidence to reconsider the ability advantageous, even in the absence of evolutionary correlations (40). of the EPE-strain of L. intracellularis to cross-infect and cause lesions In Trial 1, we attempted to cross-infect young, susceptible hamsters in other rodents, unlike what was seen for the PPE-strain, whether (4-week-old) with the EPE-strain, starting from a pure cell culture it depends on their resistance to infection or on specific bacterial isolate that is virulent for both horses and rabbits. Since this failed tropism for specific hosts (7,21). to generate detectable IHC lesions, it was tried again on younger As no serologic response was noted in either hamster trial, it and potentially more vulnerable subjects (3-week-old) (24,25). It is arguable that hamsters may need a longer time to develop a should be noted that on day 0 in Trial 1, the fecal material of 5 of detectable immune response to L. intracellularis infection. This

2000;64:0–00 The Canadian Journal of Veterinary Research 269 FOR PERSONAL USE ONLY

a b 1.0 3 108

7 EPE strain 1.0 3 106 8.0 3 10 PPE strain

8.0 3 105 6.0 3 107

6.0 3 105 4.0 3 107

4.0 3 105

2.0 3 107 2.0 3 105 L. intracellularis aspA gene copies/g of feces

L. intracellularis aspA gene copies/g of feces 0 0 0 7 11 14 17 21 0 0 7 7 14 14 17 17 21 21 24 24 Day Day

c 5.0 3 107

4.0 3 107

3.0 3 107

2.0 3 107

1.0 3 107 L. intracellularis aspA gene copies/g of feces

0 Controls EPE strain PPE strain Figure 7. Detection of L. intracellularis DNA gene copies (aspA gene) through qPCR analysis in feces (grams) of experimentally infected weanling ­hamsters. For all 3 graphs, the whiskers on the bar indicate the standard deviations. a — Fecal shedding for EPE-strain L. intracellularis during Trial 1, shown from day 0 to 21 DPI. On day 0, some unexplained technical contamination occurred while orally feeding the infectious inocula and this is responsible for the measurable aspA gene in the material collected on inoculation day. At 7 DPI, the shedding was minimal and then increased over time. Because of the technical contamination observed, the method of inoculation was changed for Trial 2. Note that the concentration of the aspA gene is barely in the range 105 to 106, particularly when compared to the PPE-infected hamsters (107 to 108 range). b — Fecal shedding for EPE-strain and PPE-strain L. intracellularis during Trial 2, shown from day 0 to 24 DPI. While the PPE-strain fecal shedding increases exponentially over time, the EPE-strain fecal shedding is highest at 14 DPI and then decreases significantly over time (P = 0.003) to the end of the experiment (24 DPI). c — Overall comparison between strain group (uninfected controls, EPE, and PPE). The different amount of shedding among groups is visible (P = 0.033). is partially due to the privileged location of the L. intracellularis, performance or IHC lesions, were observed in our PPE-challenged remote from processing, opsonizing, and exposing units of the rabbits (25). A transient IHC labeling was noted in the lamina propria reticulo-endothelial system and partly because weanling hamsters of only 1 sample out of 6 whole GIT sets of samples (42 samples may not be as immuno-competent as the adults. In other infection total), which is consistent with destruction of the bacteria, but not models, older subjects have developed an IgG response in a much with typical active lesions, which should be located at the apex of more timely fashion (1 wk post-infection) (41). Based on lack of enterocytes in the intestinal crypts. This shows that the PPE-strain of seroconversion and limited fecal shedding, the question now is L. intracellularis is eliminated swiftly by rabbits, despite inoculation whether EPE-challenged rodents can become healthy short-term with almost twice the dose used in the EPE-challenged rabbits (25) carriers of the disease and in what capacity they contribute to the A recent study showed that the EPE-strain of L. intracellularis amplification of infection. is unable to cross-infect age-susceptible pigs and the PPE-strain In this context, failing to reproduce a viable PPE-infection in is unable to cross-infect age-susceptible foals, which makes the rabbits is significant, as no clinical signs, not even reduced growth results of this study on laboratory animals even more relevant (29).

270 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

8.0 3 1015 Acknowledgments The authors thank Dr. C. Wheler at the Animal Resources Centre and the support staff of the Animal Care Unit at Western College 6.0 3 1015 of Veterinary Medicine, University of Saskatchewan, S. Mapes at University of California at Davis for the PCR explanation, and Dr. J.M. Dhillon for the French translation of the abstract. These studies were funded by a grant from the Equine Health 4.0 3 1015 Research Fund of the Western College of Veterinary Medicine. F. Sampieri and KR Ball are fellows of the CIHR-THRUST (Canadian Institutes of Health Research — Training in Health Research Using Synchrotron Techniques Grant). F.A. Vannucci was supported by 15 2.0 3 10 the Brazilian Government sponsoring agency “Conselho Nacional de Desenvolvimento Cientifico e Tecnologico” (CNPq). L. intracellularis aspA gene copies/g of feces

0 References PPE strain EPE strain . 1 Lawson GH, Gebhart CJ. Proliferative enteropathy. J Comp Figure 8. Comparison of L. intracellularis aspA gene copies/g of feces Pathol 2000;122:77–100. obtained through qPCR between EPE-infected and PPE-infected rabbits (P = 0.0089). The whiskers on the bar indicate the standard deviation. 2. McOrist S, Gebhart CJ, Boid R, Barns SM. Characterization of Note that the concentration of the aspA gene in the feces of EPE-infected Lawsonia intracellularis gen. nov., sp. nov., the obligately intracel- rabbits is so high that the values for the PPE-infected group do not lular bacterium of porcine proliferative enteropathy. Int J Syst appear in the graph, as they remained several orders of magnitude lower (106 or 107, at the most). Bacteriol 1995;45:820–825. 3. Hotchkiss CE, Shames B, Perkins SE, Fox JG. Proliferative enteropathy of rabbits: The intracellular Campylobacter-like organism is closely related to Lawsonia intracellularis. Lab Anim The inability of the EPE-strain to cause demonstrable infection in Sci 1996;46:623–627. hamsters mirrors the results of cross-infection to pigs, although it is 4. Pusterla N, Mapes S, Rejmanek D, Gebhart C. Detection of clearly capable of infecting and inducing clinical disease and lesions Lawsonia intracellularis by real-time PCR in the feces of free- in young horses, as well as cross-infecting rabbits, with mild clinical living animals from equine farms with documented occur- discomfort, typical intestinal lesions, limited fecal shedding, and rence of equine proliferative enteropathy. J Wildl Dis 2008;44: high serologic response (25,29). On the contrary, the PPE-strain of 992–998. L. intracellularis is capable of infecting and causing overt (in some 5. Lavoie JP, Drolet R, Parsons D, et al. Equine proliferative enter- cases) disease in pigs and hamsters, but does not cause infection in opathy: A cause of weight loss, colic, diarrhoea and hypopro- rabbits or horses, although a limited immune-response and fecal teinaemia in foals on three breeding farms in Canada. Equine shedding were detected in rabbits as well, similar to some exposure Vet J 2000;32:418–425. results in sentinel rabbits used in piggeries (7,30,42). Perhaps a sec- 6. Wong DM, Alcott CJ, Sponseller BA, Young JL, Sponseller BT. ondary host-adaptation, or species-specificity, exists for different Impaired intestinal absorption of glucose in 4 foals with Lawsonia L. intracellularis strains (25,29,43). In hamsters cross-infected with intracellularis infection. J Vet Intern Med 2009;23:940–944. PPE-strain, the most severe clinical signs were caused by inocula- 7. Vannucci FA, Borges EL, de Oliveira JS, Guedes RM. Intestinal tion with homogenates of porcine mucosal ileal scrapings (7,15–17). absorption and histomorphometry of Syrian hamsters Future studies may therefore focus on whether the role of passage (Mesocricetus auratus) experimentally infected with Lawsonia of the infectious material through cell cultures also hampers the intracellularis. Vet Microbiol 2010;145:286–291. virulence in EPE-strains, as we know that higher passages of PPE- 8. Pusterla N, Gebhart C. Equine proliferative enteropathy strain in cell cultures definitely immortalize L. intracellularis, but caused by Lawsonia intracellularis. Equine Vet Educ 2009;21: reduce its virulence. 415–419. In conclusion, although only a small number of animals were 9. Gebhart CJ, Barns SM, McOrist S, Lin GF, Lawson GH. Ileal used, our investigations help to determine whether different strains symbiont intracellularis, an obligate intracellular bacterium of of L. intracellularis have different secondary host-specificities. Such porcine intestines showing a relationship to Desulfovibrio species. findings should help to exclude hamsters and rabbits from modeling Int J Syst Bacteriol 1993;43:533–538. EPE-strain and PPE-strain infection, respectively. As EPE modeling 10. Frazer ML. Lawsonia intracellularis infection in horses: 2005–2007. has been achieved in rabbits, however, 2 important concepts may J Vet Intern Med 2008;22:1243–1248. be extrapolated from our results: studying the infection caused by 11. Guedes RM, Gebhart CJ. Onset and duration of fecal shedding, L. intracellularis may necessitate the use of more than 1 animal model cell-mediated and humoral immune responses in pigs after chal- and different wildlife species could potentially be a connecting link lenge with a pathogenic isolate or attenuated vaccine strain of between differently adapted infectious strains (4). Lawsonia intracellularis. Vet Microbiol 2003;91:135–145.

2000;64:0–00 The Canadian Journal of Veterinary Research 271 FOR PERSONAL USE ONLY

12. Boutrup TS, Boesen H, Boye M, Agerholm JS, Jensen TK. Early Mycobacterium avium subsp. paratuberculosis in neonatal calves. pathogenesis in porcine proliferative enteropathy caused by J Vet Intern Med 2011;25:1152–1155. Lawsonia intracellularis. J Comp Pathol 2010;143:101–109. 29. Vannucci FA, Pusterla N, Mapes SM, Gebhart C. Evidence of 13. Stills HF, Jr, Hook RR, Jr, Sprouse RF. Utilization of monoclonal host adaptation in Lawsonia intracellularis infections. Vet Res antibodies to evaluate the involvement of Campylobacter jejuni 2012;43:53. in proliferative ileitis in Syrian hamsters (Mesocricetis auratus). 30. Guedes RM, Gebhart CJ. Comparison of intestinal mucosa Infect Immun 1987;55:2240–2246. homogenate and pure culture of the homologous Lawsonia 14. Stills HF, Jr. Isolation of an intracellular bacterium from hamsters intracellularis isolate in reproducing proliferative enteropathy (Mesocricetus auratus) with proliferative ileitis and reproduction of in swine. Vet Microbiol 2003;93:159–166. the disease with a pure culture. Infect Immun 1991;59:3227–3236. 31. Guedes RM, Winkelman NL, Gebhart CJ. Relationship between 15. Stills HF, Jr, Hook RR, Jr. Experimental production of prolifera- the severity of porcine proliferative enteropathy and the infec- tive ileitis in Syrian hamsters (Mesocricetus auratus) by using an tious dose of Lawsonia intracellularis. Vet Rec 2003;153:432–433. ileal homogenate free of Campylobacter jejuni. Infect Immun 32. Morton DB, Jennings M, Buckwell A, et al. Refining procedures 1989;57:191–195. for the administration of substances. Report of the BVAAWF/ 16. Jasni S, McOrist S, Lawson GH. Experimentally induced prolif- FRAME/RSPCA/UFAW joint working group on refinement. erative enteritis in hamsters: An ultrastructural study. Res Vet Lab Anim 2001;35:1–41. Sci 1994;56:186–192. 33. Johnson-Delaney C. Anatomy and physiology of the rabbit and 17. Jasni S, McOrist S, Lawson GH. Reproduction of proliferative rodent gastrointestinal system. Proc Assoc Avian Vet 2006:9–17. enteritis in hamsters with a pure culture of porcine ileal symbi- 34. Guedes RM, Gebhart CJ. Preparation and characterization of ont intracellularis. Vet Microbiol 1994;41:1–9. polyclonal and monoclonal antibodies against Lawsonia intracel- 18. McOrist S, Lawson GH, Rowland AC, Macintyre N. Early lularis. J Vet Diagn Invest 2003;15:438–446. lesions of proliferative enteritis in pigs and hamsters. Vet Pathol 35. Guedes RM, Gebhart CJ, Deen J, Winkelman NL. Validation of an 1989;26:260–264. immunoperoxidase monolayer assay as a serologic test for por- 19. Cooper DM, Swanson DL, Gebhart CJ. Diagnosis of prolifera- cine proliferative enteropathy. J Vet Diagn Invest 2002;14:528–530. tive enteritis in frozen and formalin-fixed, paraffin-embedded 36. Pusterla N, Jackson R, Wilson R, Collier J, Mapes S, Gebhart C. tissues from a hamster, horse, deer and ostrich using a Lawsonia Temporal detection of Lawsonia intracellularis using serology intracellularis-­specific multiplex PCR assay. Vet Microbiol and real-time PCR in Thoroughbred horses residing on a farm 1997;54:47–62. endemic for equine proliferative enteropathy. Vet Microbiol 20. Collins AM, Love RJ, Jasni S, McOrist S. Attempted infection of 2009;136:173–176. mice, rats and chickens by porcine strains of Lawsonia intracel- 37. Leutenegger CM, Mislin CN, Sigrist B, Ehrengruber MU, lularis. Aust Vet J 1999;77:120–122. Hoffman-Lehmann R, Lutz H. Quantitative real-time PCR 21. Murakata K, Sato A, Yoshiya M, et al. Infection of different strains for the measurement of feline cytokine mRNA. Vet Immunol of mice with Lawsonia intracellularis derived from rabbit or por- Immunopathol 1999;71:291–305. cine proliferative enteropathy. J Comp Pathol 2008;139:8–15. 38. Mapes S, Rhodes DM, Wilson WD, Leutenegger CM, Pusterla N. 22. Collins AM, Fell S, Pearson H, Toribio JA. Colonisation and shed- Comparison of five real-time PCR assays for detecting virulence ding of Lawsonia intracellularis in experimentally inoculated rodents genes in isolates of Escherichia coli from septicaemic neonatal and in wild rodents on pig farms. Vet Microbiol 2011;150:384–388. foals. Vet Rec 2007;161:716–718. 23. Pusterla N, Wattanaphansak S, Mapes S, et al. Oral infection of 39. Cooper DM, Swanson DL, Barns SM, Gebhart CJ. Comparison of weanling foals with an equine isolate of Lawsonia intracellularis, the 16S ribosomal DNA sequences from the intracellular agents agent of equine proliferative enteropathy. J Vet Intern Med of proliferative enteritis in a hamster, deer, and ostrich with the 2010;24:622–627. sequence of a porcine isolate of Lawsonia intracellularis. Int J Syst 24. Pusterla N, Vannucci FA, Mapes S, et al. Efficacy of an avirulent Bacteriol 1997;47:635–639. live vaccine against Lawsonia intracellularis in the prevention of 40. Argenzio R. General functions of the gastrointestinal tract proliferative enteropathy in experimentally infected weanling and their control and integration. In: Dukes HH, Swenson JM, foals. Am J Vet Res 2012;73:741–746. eds. Dukes’ Physiology of Domestic Animals. 10th ed. Ithaca, 25. Sampieri F, Allen AL, Pusterla N, et al. The rabbit as an infec- New York: Cornell University Pr, 1984:326–328. tion model for equine proliferative enteropathy. Can J Vet Res 41. Zuerner RL, Alt DP, Palmer MV. Development of chronic and 2013;77:110–119. acute golden Syrian hamster infection models with Leptospira 26. Fenwick N, Griffin G, Gauthier C. The welfare of animals used borgpetersenii serovar Hardjo. Vet Pathol 2012;49:403–411. in science: How the “Three Rs” ethic guides improvements. Can 42. Duhamel GE, Klein EC, Elder RO, Gebhart CJ. Subclinical prolif- Vet J 2009;50:523–530. erative enteropathy in sentinel rabbits associated with Lawsonia 27. Lawson GH, McOrist S, Jasni S, Mackie RA. Intracellular bacteria intracellularis. Vet Pathol 1998;35:300–303. of porcine proliferative enteropathy: Cultivation and mainte- 43. Friedman M, Bednar V, Klimes J, Smola J, Mrlik V, Literak I. nance in vitro. J Clin Microbiol 1993;31:1136–1142. Lawsonia intracellularis in rodents from pig farms with the occur- 28. Fecteau ME, Whitlock RH, Fyock TL, McAdams SC, Boston RC, rence of porcine proliferative enteropathy. Lett Appl Microbiol Sweeney RW. Antimicrobial activity of gallium nitrate against 2008;47:117–121.

272 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Article

Analgesic and motor effects of a high-volume intercoccygeal epidural injection of 0.125% or 0.0625% bupivacaine in adult cows Eva Rioja, Luis M. Rubio-Martínez, Gabrielle Monteith, Carolyn L. Kerr

Abstract The objectives of this study were to determine the analgesic and motor effects of a high-volume intercoccygeal epidural injection of bupivacaine at 2 concentrations in cows. A prospective, randomized, blinded, crossover trial was conducted on 6 adult cows. An indwelling epidural catheter was placed in the first intercoccygeal space and advanced 10 cm cranially. All the cows received 3 treatments with a washout period of 48 h: saline (control), 0.125% bupivacaine (high dose), or 0.0625% bupivacaine (low dose), at a final volume of 0.15 mL per kilogram of body weight, infused manually into the epidural space over a period of 15 min. The anal and tail tone and motor deficits of the pelvic limbs were evaluated in 5 of the cows with use of a numerical rating scale and a visual analogue scale (VAS). Sensory block was assessed in 4 of the cows by the response to needle pricks in different regions with the use of a VAS. Measurements were obtained before and at different time points after injection, up to 360 min. Analysis of variance for repeated measures and post-hoc Tukey’s and Dunnett’s tests were used. Differences were considered significant when the P-value was # 0.05. One cow became recumbent 6 h after injection. Anal and tail tones were significantly decreased and motor deficits of the pelvic limbs were significantly increased after bupivacaine treatment compared with control treatment. The overall mean VASpain scores 6 standard deviation were 66 6 8 after control treatment, 52 6 5 after low-dose bupivacaine treatment, and 43 6 5 after high-dose bupivacaine treatment. The pain scores were significantly lower in caudal regions up to the saphenous nerve after high-dose bupivacaine treatment compared with control treatment and significantly lower in the anus, vulva, and tail after low-dose bupivacaine treatment compared with control treatment. Thus, analgesia with moderate motor deficits of the pelvic limbs may be obtained with 0.125% bupivacaine administered epidurally.

Résumé Les objectifs de la présente étude étaient de déterminer chez la vache les effets analgésiques et moteurs d’une injection épidurale inter- coccygienne d’un volume important de bupivacaïne à deux concentrations. Une étude croisée prospective, randomisée, et à l’aveugle a été réalisée chez 6 vaches adultes. Un cathéter épidural à demeure a été placé dans le premier espace inter-coccygien et avancé cranialement de 10 cm. Toutes les vaches ont reçu 3 traitements avec une période d’évacuation de 48 h : saline (témoin), 0,125 % de bupivacaïne (dose élevée) ou 0,0625 % de bupivacaïne (faible dose), à un volume final de 0,15 mL par kilo de poids corporel, infusé manuellement dans l’espace épidural sur une période de 15 min. Le tonus anal et de la queue ainsi que les déficits moteurs des membres pelviens ont été évalués chez 5 des vaches au moyen d’une échelle numérique de pointage et une échelle analogue visuelle (VAS). Le bloc sensitif a été évalué chez 4 des vaches par la réponse à des piqûres d’aiguille dans différentes régions avec l’utilisation d’une VAS. Les mesures ont été obtenues avant et à différents temps après l’injection, jusqu’à 360 min. Une analyse de variance pour mesures répétées et les tests post-hoc de Tukey et de Dunnett ont été utilisés. Les différences étaient considérées significatives lorsque la valeur de P était # 0,05. Une vache est demeurée couchée 6 h après l’injection. Le tonus anal et de la queue était réduit de manière significative et les déficits moteurs des membres pelviens étaient significativement augmentés après le traitement à la bupivacaïne comparativement au traitement témoin avec la saline. Dans l’ensemble les scores moyens 6 l’écart-type de VASdouleur étaient 66 6 8 après le traitement témoin, 52 6 5 après le traitement à faible dose de bupivacaïne, et 43 6 5 après le traitement avec la dose élevée de bupivacaïne. Les scores de douleur étaient significativement plus faibles dans les régions caudales jusqu’au nerf saphène après le traitement avec les doses élevées de bupivacaïne comparativement au traitement témoin et significativement plus faibles au niveau de l’anus, la vulve et la queue après le traitement avec les faibles doses de bupivacaïne comparativement au traitement témoin. Ainsi, une analgésie avec des déficits moteurs modérés des membres pelviens peut être obtenue avec de la bupivacaïne à 0,125 % administrée par voie épidurale. (Traduit par Docteur Serge Messier)

Department of Clinical Studies, Ontario Veterinary College, University of Guelph, Guelph, Ontario. Dr. Rioja’s current address is School of Veterinary Science, Leahurst Campus, University of Liverpool, Neston, CH647TE, United Kingdom. Address all correspondence to Dr. Eva Rioja; e-mail: [email protected] Received April 18, 2012. Accepted October 26, 2012.

2013;77:273–280 The Canadian Journal of Veterinary Research 273 FOR PERSONAL USE ONLY

Introduction Materials and methods Epidural administration of local anesthetics in the first intercoc- The study was approved by the Animal Care Committee of the cygeal space is frequently used to provide anesthesia and analgesia University of Guelph, and the animals were maintained in accor- for caudal surgical and obstetric procedures in large animals. The dance with the Canadian Council on Animal Care guidelines (20). short-acting local anesthetic lidocaine is typically used for this The 7 adult nonlactating Holstein Friesian cows were determined purpose (1,2). However, local anesthetics administered epidurally to be healthy by physical examination before entry to the study and can affect motor nerve function in addition to sensory nerve fibers, daily physical examination throughout the study period. Cows with which can lead to serious complications secondary to ataxia, muscle obvious musculoskeletal disease were excluded. No cows were preg- weakness, or paralysis of the pelvic limbs. Volume of administration nant, as determined by rectal ultrasonographic examination before has been shown to be a major factor influencing cranial distribu- entry to the study. During the study period the cows were placed tion of local anesthetics within the epidural space (3); therefore, the in individual stables, restrained in a head gate, and offered hay and general recommendation for intercoccygeal epidural administra- water ad libitum. During the time between treatments they were kept tion of 2% lidocaine in large animals is 1 mL per 100 kg of body in multiple indoor bovine stables and fed only hay. weight (4). Unfortunately, this limits the anesthesia to the most One of the cows was used in a pilot study, carried out to deter- caudal dermatomes; specifically, the perineum, tail, anus, vulva, mine the volume and concentration of bupivacaine that could be and vagina (4). Administration of higher volumes of lidocaine in injected epidurally without causing pronounced dysfunction of the the sacrococcygeal or first intercoccygeal space has been used clini- pelvic limbs. The other 6 cows were used in a prospective, random- cally in calves with minimal untoward cardiovascular effects and ized, blinded, crossover study. All the cows received 3 treatments, excellent anesthesia up to the level of the umbilicus (5,6). However, administered in a random order with a washout period of at least this technique caused paralysis of the pelvic limbs and recum- 48 h. The treatments were as follows: sterile isotonic saline (control), bency, and therefore its use is limited to calves, small ruminants, bupivacaine (Marcaine, 0.5%; Hospira, Montreal, Quebec) diluted and pigs. in sterile saline to a concentration of 0.0625% (low dose), and Other techniques to desensitize the flank of cattle and horses bupivacaine diluted in sterile saline to a concentration of 0.125% without affecting the motor function of the pelvic limbs have been (high dose). All treatments were administered at a dose of 0.15 mL described; these include paravertebral thoracolumbar anesthesia (4,7) per kilogram of body weight and infused manually over a period and segmental dorsolumbar epidural anesthesia (8–11). However, of 15 min via an indwelling epidural catheter placed into the first these techniques do not provide analgesia to the pelvic limbs. intercoccygeal space. Bupivacaine is a longer-lasting local anesthetic with potent The epidural catheter (Med-Rx Epidural Anesthesia Kit; Benlan, analgesic action. Concurrent intrathecal administration of opi- Oakville, Ontario) was always placed by the same investigator (ER) oids and bupivacaine decreased the opioid requirements and on the first treatment day with sterile technique. The catheter was the development of opioid tolerance compared with intrathecal inserted into the first intercoccygeal space after the area had been administration of opioids alone (12). When administered epidur- clipped and aseptically prepared with clorhexidine soap, ethanol, ally at low concentrations (, 0.125%) in humans, bupivacaine and clorhexidine solution. A small skin incision was made with a produces analgesia while minimally affecting motor function no. 15 scalpel blade after subcutaneous infiltration of 2 mL of 2% (13–17). Epidural coadministration of methadone, ketamine, and lidocaine. A Tuohy needle (17-gauge, 3.5-in, Huber point) was then low-concentration bupivacaine provided good analgesia without inserted through the skin and directed into the epidural space at a motor deficits in a cow with complex regional pain syndrome 30° angle to the contour of the rump. Correct placement of the tip (18). In dogs, the analgesic and motor effects of 3 concentrations of the needle into the epidural space was confirmed by the “hang- of bupivacaine administered epidurally were compared (19): the ing drop” technique with the use of sterile saline. The catheter lowest concentration of bupivacaine studied (0.25%) produced less (19-gauge) was introduced into the epidural space through the analgesia but also less motor deficit than the higher concentrations Tuohy needle and advanced cranially for 10 cm, up to approximately (0.5% and 0.75%). To our knowledge, no controlled studies have the level of the 4th sacral vertebra. Correct catheter placement was been conducted on the analgesic and motor effects of epidural assumed if there was no resistance to insertion. The catheter was administration of high volumes and low concentrations of bupi- cut to a length of approximately 20 cm and a filter (0.22 mm) placed vacaine in large animals. The possibility of achieving analgesia at the end of the catheter and sutured to the skin. The catheter was without motor function deficits of the pelvic limbs would have fixed with cyanoacrylate glue and the entry site covered with gauze important clinical implications in the analgesic management of impregnated with povidone iodine solution and a drape (Opsite; painful conditions associated with the abdomen and/or pelvic Smith and Nephew Medical Limited, Hull, England). The catheter limbs in large animals. was left in place for the duration of the study and was flushed The objectives of this study were to evaluate, in adult cows, the every 8 h with 3 mL of sterile heparinized saline (50 IU/mL). It was analgesic and motor effects of 2 low concentrations of bupivacaine removed after the 3 treatments had been completed (minimum 7 d, administered epidurally into the first intercoccygeal space at high maximum 10 d) and visually inspected for any signs of kinking or volume. The hypothesis was that low concentrations would produce contamination. analgesia of caudal dermatomes as well as of the pelvic limbs and Cardiorespiratory variables, motor tone, and analgesia were evalu- flanks without interfering with motor function. ated and recorded at the following time points: before the epidural

274 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Table I. Cardiorespiratory values for 5 cows before and at different time points (overall value) after an intercoccygeal epidural injection of saline (control), 0.0625% bupivacaine (low dose) or 0.125% bupivacaine (high dose), final volume 0.15 mL/kg in all treatments Mean 6 standard deviation (SD); treatment group 5 min before Average in the 6 h after injection Parameter injection Control Low dose High dose HR (beats/min) 57 6 8 55 6 7 54 6 7 55 6 7 RR (breaths/min) 31 6 9 32 6 8 33 6 10 35 6 12 SAP (mmHg) 119 6 12 122 6 10 126 6 13 127 6 12 DAP (mmHg) 56 6 11 58 6 10 63 6 8 58 6 13 MAP (mmHg) 77 6 11 78 6 11 83 6 9 80 6 13 Figure 1. Regions used for needle-prick stimulation in 4 cows to evaluate HR — heart rate; RR — respiratory rate; SAP — systolic arterial blood analgesia in response to an intercoccygeal epidural injection of saline pressure; DAP — diastolic arterial blood pressure; MAP — mean arte- or bupivacaine [0.0625% (low dose) or 0.125% (high dose)] (0.15 mL/kg rial blood pressure. in all treatments). Modified from Budras and Habel (21). 1 — anus; 2 — vulva; 3 — perineum; 4 — udder caudal; 5 — tail; 6 — pudendal nerve; 7 — tibial nerve; 8 — common peroneal nerve; 9 — saphenous nerve; 10 — lateral cutaneous femoral nerve; 11 — middle clunial nerve; 12 — cranial clunial nerve; 13 — udder middle; 14 — udder cranial; Muscle pricks were done only if the animal did not respond to the 15 — second lumbar vertebral nerve; 16 — first lumbar vertebral nerve. skin pricks in that region or if the animal’s response was question- able (i.e., slight movement or movement possibly related to another injection (time −5 min; baseline), upon completion of the injection cause). A VAS for avoidance behavior in response to the needle (time 0 min), and 15, 30, 45, 60, 90, 120, 180, 240, 300, and 360 min pricks (VASpain) was used: 0 mm meant no response and 100 mm after completion of the injection. meant an aggressive response (i.e., kicking at the investigator doing Cardiorespiratory variables recorded included heart rate (deter- the pricks). Needle pricks were done up to at least 60 min after mined by direct auscultation of the heart for 1 min), respiratory rate epidural injection in all cows. When a region was clearly positive (determined by counting chest movements for 1 min), and arterial to the pricks (i.e., the animal showed a clear avoidance response) at blood pressure (AP; measured noninvasively by oscillometric tech- all time points up to 60 min after epidural injection or at any time nique). The oscillometric monitor (Cardell Veterinary Monitor 9402 point after 60 min, no more pricks were done in that region, and

BP/SpO2; Sharn Veterinary, Tampa, Florida, USA) was connected to a the last VASpain value obtained was recorded for the subsequent large-animal cuff (model SV10; Sharn Veterinary), which was placed time points. The maximum VASpain value obtained in each region around the base of the tail. Three consecutive readings of the oscil- at each time point with either skin or muscle pricks was used for lometric systolic (SAP), diastolic (DAP), and mean (MAP) arterial statistical analysis. blood pressures were collected at each time point, and the average For statistical analysis commercial software (SAS, version 9.1.3; value was calculated and used for statistical analysis. SAS Institute, Cary, North Carolina, USA) was used, and the follow- Tail and anal tone as well as motor deficits of the pelvic limbs were ing were done: a Shapiro–Wilk test to assess normal distribution of scored by a blinded evaluator (LMR) on a numerical rating scale the data; log transformation when appropriate; a logit transforma- (NRS; Appendix). To evaluate for motor deficits of the pelvic limbs, tion applied to the VAS scores; analysis of variance, general linear the investigator pushed the cow on the hindquarters and pulled model for repeated measures, accounting for random effects (cow the tail sideways while the cow remained restrained in a head gate, and period) and fixed effects (treatment, time, and treatment–time and this was recorded on video. Additionally, the degree of motor interaction); and post-hoc analysis with Tukey’s and Dunnett’s tests. deficit of the pelvic limbs was scored from the video recordings by A P-value # 0.05 was considered significant. a blinded investigator (ER) using a visual analogue scale (VAS) for motor tone (VASmotor), which consisted of a 100-mm-long line with Results no marks, where 0 mm meant absence of ataxia and 100 mm meant extreme ataxia, with the animal falling down. The mean weight of the cows was 658 6 81 [standard devi- The degree of analgesia was determined by assessing the animal’s ation (SD)] kg. The total volume of epidural treatments was response to the application of standard noxious stimuli to skin and 102.8 6 7.5 mL. One cow did not respond to any of the treatments. muscle, which consisted of pricks with a 25-gauge, 1-in hypodermic This treatment failure was possibly due to misplacement of the needle and a 22-gauge, 1.5-in needle, respectively, applied sequen- epidural catheter. Data for this cow were therefore not used for tially at different regions innervated by specific nerves, as shown in statistical analysis. Another cow was used only for motor and car- Figure 1 (21). The pricks were done bilaterally from caudal to cranial, diorespiratory evaluation owing to behavioral limitations. Therefore, always by the same blinded investigator (LMR), and were recorded 5 cows were used for motor and cardiorespiratory evaluation and on video for later evaluation by a different blinded investigator (ER). 4 cows for evaluation of analgesia.

2000;64:0–00 The Canadian Journal of Veterinary Research 275 FOR PERSONAL USE ONLY

Table II. Scores for anal and tail motor tone and motor Control Low dose High dose deficits of the pelvic limbs for the 5 cows at different time points (overall value) after the injections Median score (and range) or Scoring scale; mean 6 SD; treatment group parameter Control Low dose High dose NRS: anal tone 1 (1–1) 2 (1–4)a 2 (1–4)a NRS: tail tone 1 (1–2) 2 (1–4)a 3 (1–4)a,b NRS: motor deficits of pelvic limbs 1 (1–2) 1 (1–3) 2 (1–3)a,b VASmotor 0.4 6 1.9 11.3 6 15.5a 37.2 6 24.4a,b a Significantly different from the value for the control treatment (P , 0.001). b Significantly different from the value for the low-dose treatment Figure 2. Mean scores 6 standard deviation for motor deficits of the (P , 0.001). pelvic limbs in 5 cows, determined from video recordings on a visual NRS — numerical rating scale (see Appendix); VASmotor — visual analogue scale for motor tone 5 min before and at different time points analogue scale for motor tone. after the control, low-dose, or high-dose injection. * Significantly different from the pretreatment value. † Significantly different from the control value at the same time point. ‡ Significantly different from the high- dose value at the same time point.

The cardiorespiratory parameters were not significantly different among treatments or within a treatment over time relative to the The regions that showed a significant treatment effect were tail baseline values and remained within normal limits (Table I). (P = 0.007), anus (P , 0.001), vulva (P , 0.001), perineum (P , 0.001), All the cows remained standing during the 6-h study period. One udder caudal (P , 0.001), pudendal nerve (P = 0.026), tibial nerve cow became recumbent immediately after being released from the (P , 0.001), common peroneal nerve (P , 0.001), saphenous nerve head gate 6 h after epidural injection of the high dose of bupivacaine. (P = 0.008), and udder middle (P , 0.001) on the left side and tail The cow remained recumbent for a few minutes and then stood up (P , 0.001), anus (P , 0.001), perineum (P , 0.001), and tibial unassisted without problems. nerve (P = 0.002) on the right side. The differences in scores are The NRS scores for anal and tail tone and for motor deficits of the summarized in Table III. The regions that showed a significant pelvic limbs are summarized in Table II. The effects of treatment, time, treatment–time interaction were anus (P = 0.035), vulva (P = 0.047), and treatment–time interaction were significant for anal and tail tone perineum (P = 0.002), udder caudal (P = 0.047), and middle clunial (P , 0.001). Compared with baseline the anal scores were significantly nerve (P = 0.015) on the left side and anus (P = 0.035) and middle increased from 0 to 240 min after high-dose bupivacaine treatment clunial nerve (P , 0.001) on the right side. and from 0 to 180 min after low-dose treatment, and compared with With high-dose bupivacaine treatment the onset and duration the control treatment the anal scores were significantly increased from of analgesia were variable among regions: between 0 and 30 min 0 to 300 min after high-dose bupivacaine treatment and from 0 to for onset and between 300 and 360 min for duration. The VASpain 180 min after low-dose treatment. Compared with both baseline and scores were significantly decreased below baseline from 0 to 300 min control treatment the tail scores were significantly increased from 0 to in the anus and vulva, at 30 min in the region of the middle clunial 360 min after high-dose bupivacaine treatment and from 0 to 300 min nerve, and from 30 to 360 min in the perineum and udder caudal after low-dose treatment. There was also a significant effect of treat- on the left side, as well as from 0 to 300 min in the anus and from ment (P , 0.001), time (P , 0.001), and treatment–time interaction 0 to 90 min in the region of the middle clunial nerve on the right side. (P = 0.008) for the NRS scores for motor deficits of the pelvic limbs: The scores were significantly lower than with the control treatment the scores were significantly increased from baseline and compared from 30 to 360 min in the anus and perineum, from 15 to 360 min in with control treatment from 0 to 360 min after high-dose bupivacaine the vulva, at 30 min in the region of the middle clunial nerve, and treatment and at 15 and 45 minutes after low-dose treatment. from 60 to 360 min in the udder caudal on the left side, as well as The VASmotor scores are summarized in Table II and Figure 2. from 30 to 360 min in the anus and at 60 min in the region of the The effects of treatment, time, and treatment–time interaction were middle clunial nerve on the right side. significant (P , 0.001). The motor deficits of the pelvic limbs were At the time of removal, no catheters had signs of kinking or evident immediately after epidural injection of low-dose and high- contamination. dose bupivacaine and lasted for 3 and up to 6 h, respectively. The overall VASpain scores in response to needle pricks were Discussion higher for each region on the left side than on the right side, but no statistical comparisons were made between sides. The overall This study showed that administration of 0.15 mL/kg of 0.125% mean scores (6 SD) for all regions together over time, excluding bupivacaine 10 cm cranial to the first intercoccygeal space of adult baseline, were 66 6 8, 52 6 5, and 43 6 5 for the control, low-dose cows produced complete or partial sensory block in caudal regions bupivacaine, and high-dose bupivacaine treatments, respectively. up to the level of the cutaneous area innervated by the saphenous

276 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Table III. Scores for avoidance behavior in response to needle tered epidurally, with minimal motor dysfunction (13–17). On the pricks in 4 cows at different time points (overall value) after other hand, consistent analgesia with epidural administration of the injections; only regions that showed a significant 0.0625% bupivacaine was not obtained in this study, in accordance treatment effect (P , 0.05) for analgesia are included with the study in humans by Brennum et al (16), who observed that epidural administration of 0.075% bupivacaine induced only a slight Mean score 6 SD; and nonsignificant attenuation in pain ratings of brief argon laser, treatment group mechanical, and electrical stimuli. In human obstetric clinical studies, Region; side Control Low dose High dose analgesia for labor was considered insufficient with 0.0625% bupi- Anus vacaine administered epidurally; supplemental bupivacaine doses Left 66 6 18 31 6 23 11 6 20a,b were necessary (14,15,24). However, other studies in humans have Right 51 6 10 14 6 19a 7 6 16a,b demonstrated a good analgesic effect of 0.0625% bupivacaine when Vulva, left 66 6 21 35 6 26a 9 6 15a,b administered epidurally in combination with fentanyl or sufentanil during labor (25–27). The reasons for the discrepancy among studies Perineum using 0.0625% bupivacaine could be multiple, but most likely the Left 71 6 15 54 6 21 29 6 28a,b concomitant use of opioids is the reason, as it has been shown that Right 65 6 22 31 6 26 11 6 20a,b the epidural use of opioids combined with local anesthetics results in Udder caudal, left 56 6 37 43 6 32 28 6 34a,b more intense analgesia than individual administration of the drugs owing to their synergistic antinociceptive interaction at the level of Tail the spinal cord (28). Left 79 6 12 35 6 26a 16 6 24a Response to needle pricks was used in the present study to Right 72 6 6 39 6 25a 19 6 27a,b evaluate analgesia, as it is a common method in research studies Pudendal nerve, left 81 6 8 57 6 21 47 6 32a in humans and animals (23,29,30) and is frequently used clinically to test whether surgery can be done after a local block in standing Tibial nerve large animals. Pinpricks are considered mechanical stimuli of brief Left 85 6 11 68 6 22a 66 6 27a duration that activate fast Ad nociceptors (31), which might not be Right 70 6 12 66 6 22 55 6 29a,b the same type of stimulus induced by surgery or painful condi- Common peroneal nerve, left 67 6 11 61 6 24 46 6 31a,b tions, such as inflammation. Hypoalgesia for such brief stimuli is obtained before hypoalgesia for stimuli of greater spatial and tem- Saphenous nerve, left 65 6 11 54 6 18 47 6 31a poral dimensions, which may explain the frequent clinical scenario Udder middle, left 55 6 14 49 6 25a 26 6 21a,b in which insufficient surgical analgesia is encountered in spite of a Significantly different from the value for the control treatment good pinprick analgesia (16,29). A poor correlation between lack of (P , 0.001). pinprick sensation and absence of postoperative pain in humans has b Significantly different from the value for the low-dose treatment also been demonstrated (32). In the study by Brennum et al (16) in (P , 0.001). humans, epidural administration of 0.075% bupivacaine selectively induced hypoalgesia for heat but not for mechanical or electrical nerve and caused mild to moderate motor deficits of the tail, anus, stimuli, whereas the 0.125% concentration induced hypoalgesia and pelvic limbs, with no untoward cardiorespiratory effects. for heat and mechanical stimuli but not electrical stimuli, and only However, administration of 0.0625% bupivacaine did not produce bupivacaine 0.5% induced total anesthesia, with lack of pain in consistent sensory block and caused mild motor deficits of the tail, response to the 3 types of stimuli. It is possible that an analgesic anus, and pelvic limbs, with no untoward cardiorespiratory effects. effect of 0.0625% bupivacaine administered epidurally to animals Lumbosacral administration of high volumes of 2% lidocaine could be demonstrated with a different type of stimulus, such as (0.18 to 0.24 mL/kg BW) in combination with xylazine has been thermal, or by administering the bupivacaine in combination with reported to provide adequate analgesia to permit umbilical surgery opioids. Alternatively, including a greater number of cows in the in calves (22). Also in calves, sacrococcygeal epidural administra- study might have increased the power to reveal an analgesic effect tion of a high volume (0.4 mL/kg) of 2% lidocaine had no clinically of 0.0625% bupivacaine administered epidurally. significant cardiovascular effects and provided good conditions for The scale used to evaluate response to needle pricks (VAS for surgery in the abdomen up to the level of the umbilicus (5,6). In avoidance behavior) was selected because it is commonly used to buffalo calves the use of epidural 0.75% ropivacaine in the lumbo- evaluate pain in animals (33) and allows evaluation on the basis of sacral space at doses between 0.05 and 0.1 mL/kg BW induced good an animal’s individual behavior (i.e., the animal can serve as its own analgesia up to the thorax and recumbency (23). In all these studies, control). Moreover, because it is a continuous scale, statistical analy- paralysis of the pelvic limbs was obtained by injecting high volumes sis is facilitated and statistical power increased. Most studies using of local anesthetics epidurally. pinpricks in animals use categorical scales (11,19,30). Nevertheless, In the present study, consistent analgesia was observed on the left there is currently no “gold standard” pain scale, and all scales have side up to the level of the saphenous nerve after epidural injection limitations, as they are based on subjective interpretation of animal of 0.125% bupivacaine. Previous studies in humans have also found behavior. Some pain scales also use objective parameters such as consistent analgesia at this concentration of bupivacaine adminis- heart and respiratory rates (34), but these parameters may be affected

2000;64:0–00 The Canadian Journal of Veterinary Research 277 FOR PERSONAL USE ONLY

by many other factors unrelated to pain. After careful evaluation of With the volume of bupivacaine used in this study, we expected the video recordings, we decided to use the VAS scale for avoidance analgesia up to the flank. However, it was obtained only up to the response, as some cows had low responses at baseline. Also, with region of the saphenous nerve with 0.125% bupivacaine. The limited this scale it is possible to evaluate behavioral changes on the basis craniad epidural spread could be due to the rate of administration of the individual, which is not possible with predefined categorical (0.01 mL/kg BW per minute), as a slow rate would be expected to scoring systems. The assignment of a positive or negative response generate lower epidural pressures, which could limit the spread when pinpricks are done is not always straightforward, as some (36). However, studies in humans showed that rapid injection of animals might feel pain but not respond owing to fear, or they might local anesthetics into the epidural space produced greater initial have hypoalgesia but not complete sensory block and therefore pressures and a faster onset of sensory block than slow injection, might still respond to the stimulus although the painful perception but there was no difference in the final extent and duration of the is decreased. Also because of learned behavior some animals might block (37,38). Multiple factors can affect the spread of the epidural respond even when there is no associated pain. Randomization of block: in humans these factors include age, weight, dose of local the treatments should have minimized any possible effect of learned anesthetic, addition of opioids, site of injection, and body height behavior on our results. (39). In dogs, body position after epidural administration affected the Sympathetic B fibers are readily blocked by local anesthetics, spread of a methylene blue solution (40). Even though total sensory inducing cutaneous vasodilation and a rise in skin temperature. Skin block was not obtained in the present study, hypoalgesia in more temperature has been used in some studies in cattle in an attempt to cranial regions might have been present and not demonstrated with evaluate success or failure of sensory blockade produced by epidural the methods used. administration of local anesthetics (11). Skin temperature was not A washout period of at least 48 h was chosen from epidural measured in the present study to evaluate sensory blockade as it pharmacokinetic data in isoflurane-anesthetized sheep: the reported has been shown to be a nonsensitive indicator of sympathectomy elimination half-life of bupivacaine was 142.5 min (41). The pharma- induced by epidurally administered anesthetics in humans (35). cokinetic parameters of bupivacaine administered epidurally have Also a skin temperature change was not observed after epidural not been reported in cattle to our knowledge. administration of lidocaine in cows (11), and the authors suggested In the present study, the left side had lower VASpain scores than that the thickness of cattle skin may prevent marked changes in skin the right side in most dermatomes. The side difference could have temperature. In addition, skin temperature may be influenced by been due to lateralization of the epidural catheter towards the left other factors, such as environmental temperature. side in the studied cows. However, such lateralization could not be The onset of motor deficits of the anus, tail, and pelvic limbs was confirmed, as no radiographs were obtained and no postmortem immediate with both bupivacaine concentrations, but the deficits examinations conducted. One cow did not respond to any of the lasted longer with 0.125% bupivacaine (up to 6 h for tail and pelvic treatments, and the treatment failure was probably due to misplace- limbs). The onset of analgesia was immediate in the caudal regions ment of the catheter out of the epidural space, but this could not be and later in the more cranial regions. The duration of analgesia varied confirmed. among regions, lasting up to 360 min in some of them with the high One limitation of the present study was the small sample size concentration of bupivacaine. In dogs the duration of motor effects and thus the low power to detect more significant effects of the exceeded the duration of sensory block after epidural administra- treatments. The study was initially designed with 6 cows, similar to tion of bupivacaine at concentrations of 0.25%, 0.5%, and 0.75% (19). previous crossover studies in cattle (1,11,30), but unfortunately the In humans the duration of motor block was similar to the duration sample size was reduced to 5 cows for motor and cardiorespiratory of hypoesthesia for mechanical and electrical stimuli after epidural evaluation and 4 cows for evaluation of analgesia during the course administration of 0.25% and 0.5% bupivacaine (16). In the present of the study. None the less, post-hoc power calculations revealed that study, measurements were done up to 6 h after injection, and the the power was 71% and 99% to detect differences in motor tone and complete offset of sensory and motor effects could not be determined. analgesia, respectively, between the use of a high dose of bupivacaine In humans, epidural administration of bupivacaine at concentra- and the use of saline. tions of 0.075% and 0.125% did not induce motor blockade, as assessed In conclusion, high-volume intercoccygeal epidural administration by the knee-extension strength test (16), and the blockade was consid- of 0.125% bupivacaine provided partial or complete sensory block ered mild after administration of 0.125% bupivacaine, as evaluated by in most caudal regions and in the pelvic limbs of cows. However, the ability of the patient to move the lower extremities (17). However, high-volume intercoccygeal epidural administration of 0.0625% bupi- neither of these studies evaluated the ability of the person to stand or vacaine did not provide consistent sensory block with the methods walk without ataxia. In the present study the motor effects were evalu- used in this study. This technique should be used with caution in ated with the cows standing restrained in a head gate, with sideways large-animal clinical practice as it causes mild to moderate motor movement allowed; however, they were not walked to evaluate the deficits of the pelvic limbs, which may cause recumbency if the degree of ataxia because in the pilot study these doses of epidural animal is moved or walked. bupivacaine induced mild to moderate motor deficits of the pelvic limbs and therefore the animals could fall if they were walked. In Acknowledgments fact, 1 cow fell down after being released from the head gate 6 h after epidural administration of 0.125% bupivacaine. Therefore, this tech- The Department of Clinical Studies of the Ontario Veterinary nique should be used with caution in large-animal clinical practice. College, University of Guelph, Guelph, Ontario, funded the study.

278 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Appendix: Numerical rating scale for evaluating motor 11. DeRossi R, Bertoni RA, Ruzzon RH, Verde-Selva AB, Frazilio FO. function Segmental dorsolumbar epidural analgesia via the caudal approach using multiple port catheters with ketamine or lido- Tail tone when evaluator moves the tail up and down caine or in combination in cattle. Vet Anaesth Analg 2010;37: 1 — Strong tone: presence of brisk voluntary movements 451–459. 2 — Mildly relaxed: slow voluntary movements 12. van Dongen RT, Crul BJ, van Egmond J. Intrathecal coadminis- 3 — Moderately relaxed: no voluntary movements tration of bupivacaine diminishes morphine dose progression 4 — Absent tone: no voluntary movements during long-term intrathecal infusion in cancer patients. Clin J Anal tone when evaluator pinches the anus Pain 1999;15:166–172. 1 — Strong 13. Bleyaert A, Soetens M, Vaes L, Van Steenberge AL, Van der 2 — Mildly relaxed Donck A. Bupivacaine, 0.125 per cent, in obstetric epidural 3 — Moderately relaxed analgesia: Experience in three thousand cases. Anesthesiology 4 — Absent tone 1979;51:435–438. Motor deficits of pelvic limbs when evaluator pushes the 14. Cohen SE, Yeh JY, Riley ET, Vogel TM. Walking with labor epi- ­hindquarters and pulls the tail sideways dural analgesia: The impact of bupivacaine concentration and a 1 — No ataxia: strong lidocaine–epinephrine test dose. Anesthesiology 2000;92:387–392. 2 — Mild ataxia 15. Sanchez-Pereles MC, Uribarri FJ, Gragera I. [0.0625% bupiva- 3 — Moderate ataxia caine compared with 0.125% bupivacaine continuously perfused 4 — Severe ataxia: falling epidurally during vaginal delivery]. Rev Esp Anestesiol Reanim 1993;40:9–11. References 16. Brennum J, Nielsen PT, Horn A, Arendt-Nielsen L, Secher NH. Quantitative sensory examination of epidural anaesthesia and . 1 DeRossi R, Zanenga NF, Alves OD, Carneiro RP, Ossuna MR, analgesia in man; dose–response effect of bupivacaine. Pain Jorge TP. Effects of caudal epidural ketamine and/or lidocaine 1994;56:315–326. on heifers during reproductive procedures: A preliminary study. 17. Owen MD, D’Angelo R, Gerancher JC, et al. 0.125% ropiva- Vet J 2010;185:344–346. caine is similar to 0.125% bupivacaine for labor analgesia using 2. Mosure WL, Meyer RA, Gudmundson J, Barth AD. Evaluation of patient-controlled epidural infusion. Anesth Analg 1998;86: possible methods to reduce pain associated with electroejacula- 527–531. tion in bulls. Can Vet J 1998;39:504–506. 18. Bergadano A, Moens Y, Schatzmann U. Continuous extradural 3. Freire CD, Torres ML, Fantoni DT, Cavalcanti RL, Noel-Morgan J. analgesia in a cow with complex regional pain syndrome. Vet Bupivacaine 0.25% and methylene blue spread with epidural Anaesth Analg 2006;33:189–192. anesthesia in dog. Vet Anaesth Analg 2010;37:63–69. 19. Gomez de Segura IA, Menafro A, Garcia-Fernandez P, Murillo S, 4. Skarda RT, Tranquilli WJ. Local and regional anesthetic and Parodi EM. Analgesic and motor-blocking action of epidurally analgesic techniques: Ruminants and swine. In: Tranquilli WJ, administered levobupivacaine or bupivacaine in the conscious Thurmon JC, Grimm KA, et al, eds. Lumb & Jones Veterinary dog. Vet Anaesth Analg 2009;36:485–494. Anesthesia and Analgesia. 4th ed. Ames, Iowa: Blackwell 20. Olfert ED, Cross BM, McWilliam AA, eds. Guide to the Care Publishing, 2007:643–681. and Use of Experimental Animals. 2nd ed. Volume 1. Ottawa, 5. Meyer H, Starke A, Kehler W, Rehage J. High caudal epidural Ontario: Canadian Council on Animal Care, 1993. Available from anaesthesia with local anaesthetics or alpha(2)-agonists in calves. www.ccac.ca/Documents/Standards/Guidelines/Experimental_ J Vet Med A Physiol Pathol Clin Med 2007;54:384–389. Animals_Vol1.pdf Last accessed June 16, 2013. 6. Meyer H, Kastner SB, Beyerbach M, Rehage J. Cardiopulmonary 21. Budras K-D, Habel RE. Bovine Anatomy. An Illustrated Text. effects of dorsal recumbency and high-volume caudal epi- Hannover, Germany: Schlutersche GmbH & Company, 2003:20. dural anaesthesia with lidocaine or xylazine in calves. Vet J 22. Lewis CA, Constable PD, Huhn JC, Morin DE. Sedation with 2010;186:316–322. Epub 2009 Sep 18. xylazine and lumbosacral epidural administration of lidocaine 7. Moon PF, Suter CM. Paravertebral thoracolumbar anaesthesia and xylazine for umbilical surgery in calves. J Am Vet Med Assoc in 10 horses. Equine Vet J 1993;25:304–308. 1999;214:89–95. 8. Skarda RT, Muir WW. Segmental epidural and subarachnoid 23. Amarpal, Kinjavdekar P, Aithal HP, et al. Comparison of analgesia in conscious horses: A comparative study. Am J Vet two doses of ropivacaine for lumbosacral epidural analgesia in Res 1983;44:1870–1876. buffalo calves (Bubalus bubalis). Vet Rec 2007;160:766–769. 9. Lee I, Yamagishi N, Oboshi K, Ayukawa Y, Sasaki N, Yamada H. 24. Li DF, Rees GA, Rosen M. Continuous extradural infusion of Comparison of xylazine, lidocaine and the 2 drugs combined for 0.0625% or 0.125% bupivacaine for pain relief in primigravid modified dorsolumbar epidural anaesthesia in cattle. Vet Rec labour. Br J Anaesth 1985;57:264–270. 2004;155:797–799. 25. Fernandez-Guisasola J, Serrano ML, Cobo B, et al. A comparison 10. Hiraoka M, Miyagawa T, Kobayashi H, Takahashi T, Kishi H, of 0.0625% bupivacaine with fentanyl and 0.1% ropivacaine with Lee I. Successful introduction of modified dorsolumbar epidural fentanyl for continuous epidural labor analgesia. Anesth Analg anesthesia in a bovine referral center. J Vet Sci 2007;8:181–184. 2001;92:1261–1265.

2000;64:0–00 The Canadian Journal of Veterinary Research 279 FOR PERSONAL USE ONLY

26. Russell R, Reynolds F. Epidural infusions for nulliparous 34. Firth AM, Haldane SL. Development of a scale to evaluate post- women in labour. A randomised double-blind comparison of operative pain in dogs. J Am Vet Med Assoc 1999;214:651–659. fentanyl/bupivacaine and sufentanil/bupivacaine. Anaesthesia 35. Ginosar Y, Weiniger CF, Meroz Y, et al. Pulse oximeter perfusion 1993;48:856–861. index as an early indicator of sympathectomy after epidural 27. Russell R, Quinlan J, Reynolds F. Motor block during epidural anesthesia. Acta Anaesthesiol Scand 2009;53:1018–1026. infusions for nulliparous women in labour: A randomized double- 36. Usubiaga JE, Wikinski JA, Usubiaga LE. Epidural pressure and blind study of plain bupivacaine and low dose bupivacaine with its relation to spread of anesthetic solutions in epidural space. fentanyl. Int J Obstet Anesth 1995;4:82–88. Anesth Analg 1967;46:440–446. 28. Kaneko M, Saito Y, Kirihara Y, Collins JG, Kosaka Y. Synergistic 37. Kanai A, Suzuki A, Hoka S. Rapid injection of epidural mepi- antinociceptive interaction after epidural coadministration of vacaine speeds the onset of nerve blockade. Can J Anaesth morphine and lidocaine in rats. Anesthesiology 1994;80:137–150. 2005;52:281–284. 29. Adrendt-Nielsen L, Oberg B, Bjerring P. Quantitative assess- 38. Cardoso MM, Carvalho JC. Epidural pressures and spread of 2% ment of extradural bupivacaine analgesia. Br J Anaesth 1990;65: lidocaine in the epidural space: Influence of volume and speed 633–638. of injection of the local anesthetic solution. Reg Anesth Pain Med 30. De Rossi R, Bucker GV, Varela JV. Perineal analgesic actions of 1998;23:14–19. epidural clonidine in cattle. Vet Anaesth Analg 2003;30:64–71. 39. Curatolo M, Orlando A, Zbinden AM, Scaramozzino P, Venuti FS. 31. Mackenzie RA, Burke D, Skuse NF, Lethlean AK. Fibre func- A multifactorial analysis of the spread of epidural analgesia. tion and perception during cutaneous nerve block. J Neurol Acta Anaesthesiol Scand 1994;38:646–652. Neurosurg Psychiatry 1975;38:865–873. 40. Gorgi AA, Hofmeister EH, Higginbotham MJ, Kent M. Effect of 32. Curatolo M, Kaufmann R, Petersen-Felix S, Arendt-Nielsen L, body position on cranial migration of epidurally injected methy- Scaramozzino P, Zbinden AM. Block of pinprick and cold sen- lene blue in recumbent dogs. Am J Vet Res 2006;67:219–221. sation poorly correlate with relief of postoperative pain during 41. Ratajczak-Enselme M, Estebe JP, Rose FX, et al. Effect of epi- epidural analgesia. Clin J Pain 1999;15:6–12. nephrine on epidural, intrathecal, and plasma pharmacoki- 33. Dobromylskyj P, Flecknell PA, Lascelles BD, Livingston A, netics of ropivacaine and bupivacaine in sheep. Br J Anaesth Taylor P, Waterman-Pearson A. Pain assessment. In: Flecknell PA, 2007;99:881–890. Waterman-Pearson A, eds. Pain Management in Animals. 1st ed. London: Saunders Elsevier, 2000:53–80.

280 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Article

The use of lung biopsy to determine early lung pathology and its association with health and production outcomes in feedlot steers Brandy A. Burgess, Steve H. Hendrick, Colleen M. Pollock, Sherry J. Hannon, Sameeh M. Abutarbush, Amanda Vogstad, G. Kee Jim, Calvin W. Booker

Abstract The objectives of this study were to determine if percutaneous lung biopsy can be used to characterize early pathologic changes in bovine lung associated with bovine respiratory disease (BRD), to determine if specific infectious respiratory pathogens can be identified in association with these changes, and to determine whether pulmonary pathology at arrival and at the time of initial diagnosis are associated with health and production outcomes. One hundred auction-market derived crossbred steer calves from a commercial feedlot in southern Alberta were included in this study. A percutaneous lung biopsy technique was used to obtain lung samples from the right middle lung. Steers were sampled 295 times yielding 283 samples with 210 (74%) containing lung tissue. Overall, histopathological changes were observed in 20 (9.5%) of lung biopsy samples. There were too few samples with pathology to reveal an association between lung pathology and subsequent health events. In general, percutaneous lung biopsy can be done safely on feedlot steers in a commercial feedlot setting with few clinical side effects. This technique did not prove useful as a diagnostic tool or prognostic indicator for early BRD. However, it may be useful for the diagnosis of BRD in targeted populations of commercial feedlot steers.

Résumé Les objectifs de la présente étude étaient de déterminer si une biopsie pulmonaire obtenue par voie transcutanée pouvait être utilisée afin de caractériser dans les poumons bovins les changements pathologiques hâtifs associés aux maladies respiratoires bovines (MRB), de déterminer si des agents infectieux pathogènes spécifiques au système respiratoire peuvent être identifiés en association avec ces changements, et de déterminer si les pathologies pulmonaires à l’arrivée et au moment du diagnostic initial sont associées avec les résultats de production et de santé. Cent bouvillons de race croisée issus d’encans et élevés dans un parc d’engraissement commercial du sud de l’Alberta ont été inclus dans cette étude. Une technique de biopsie pulmonaire transcutanée a été utilisée pour obtenir des échantillons de poumon du lobe pulmonaire médial droit. Les bouvillons ont été échantillonnés 295 fois produisant 283 échantillons avec 210 (74 %) contenant du tissu pulmonaire. Des changements histopathologiques ont été observés dans 20 (9,5 %) des échantillons de biopsie pulmonaire. Il y avait trop peu d’échantillons avec des pathologies pour démontrer une association entre une pathologie pulmonaire et des conséquences subséquentes sur la santé. En général, la biopsie pulmonaire transcutanée peut être faite de manière sécuritaire sur des bouvillons d’embouche en parc d’engraissement commercial avec peu d’effets cliniques secondaires. Cette technique ne s’est pas avérée utile comme outil diagnostique ou indicateur de pronostic pour les MRB hâtives. Toutefois, elle pourrait être utile pour le diagnostic de MRB dans des populations ciblées de bouvillons d’embouche en parc d’engraissement. (Traduit par Docteur Serge Messier) Introduction complex pathogenesis. In feedlot cattle, BRD often manifests as bronchopneumonia secondary to the stress of shipping, transpor- Bovine respiratory disease (BRD) continues to be an economically tation, passing through auction markets, and co-mingling (4,5). important disease complex in feedlot production that accounts for Lesions typically develop within 7 to 14 d in the cranioventral lung 10% to 61% of mortality in western Canadian feedlots (1,2), 86.8% lobes and are frequently bilateral in distribution (4). The most com- of mortality in Ontario feedlots (3), and 44% to 67% of mortality mon bacterial isolates from fibrinosuppurative bronchopneumonia in United States (US) feedlots (3,4). Prevention and control of BRD include Mannheimia haemolytica, Pasteurella multocida, and Histophilus continues to be a challenge due to its multifactorial nature and somni (3,4).

Departments of Large Animal Clinical Sciences, Western College of Veterinary Medicine, University of Saskatchewan, Saskatoon, Saskatchewan (Burgess, Hendrick); Feedlot Health Management Services, Ltd, Okotoks, Alberta (Pollock, Hannon, Abutarbush, Vogstad, Jim, Booker). Address all correspondence to Dr. Brandy Burgess; e-mail: [email protected] Dr. Burgess’ current address is Department of Clinical Sciences, Colorado State University, Fort Collins, Colorado, USA. Dr. Abutarbush’s cur- rent address is Department of Veterinary Clinical Sciences, Faculty of Veterinary Medicine, Jordan University of Science and Technology, Irbid, Jordan. Dr. Pollock’s current address is Alberta Agriculture and Rural Development, Airdrie, Alberta, Canada. Dr. Vogstad’s current address is School of Veterinary Medicine and Biomedical Sciences, University of Nebraska–Lincoln, Lincoln, Nebraska, USA. This report was previously presented in part. Burgess BA, Hendrick SH, Pollock CM, et al. The use of lung biopsy to assess early lung pathology and its association with health and production outcomes in high-risk feedlot steers. In: Proc 90th Ann Meet CRWAD 2009:139. Received March 16, 2012. Accepted October 26, 2012.

2013;77:281–287 The Canadian Journal of Veterinary Research 281 FOR PERSONAL USE ONLY

Early detection and recognition of BRD are necessary for effective ear tag as a unique identifier. Each animal was also administered top- treatment and prevention. Typically, BRD detection in feedlot cattle ical parasiticide (Vetomectin pour-on; Vétoquinol N.A., Buena, New is based on the observation of depression, coughing, decreased feed Jersey, USA), a hormonal growth implant (Synovex Choice; Wyeth intake, standing away from the group, and/or an increased rectal Animal Health, Guelph, Ontario), and an injection of a metaphy- temperature. Many of these criteria are subjective and lend them- lactic antimicrobial: an intramuscular (IM) injection of long-acting selves to individual variation and misdiagnosis. In South African oxytetracycline (Oxymycin LA; Wyeth Animal Health), 30 mg/kg feedlots, of the 42.8% of animals with lung lesions at slaughter, BW, or a subcutaneous (SC) injection of tulathromycin (2.5 mg/kg 69.5% had never been treated for BRD (6). In a US study, only 35% BW if rectal temperature . 408C, “sick on arrival;” Draxxin Injectable of steers from a US feedlot had been treated for respiratory disease Solution; Pfizer Canada). Re-implantation of a hormonal growth even though 72% of animals at slaughter had evidence of pulmo- implant occurred at . 130 d into the feeding period. Owner consent nary lesions (7). These findings suggest that continued research was obtained for the animals used in this study. into detection methods for BRD are warranted, and may improve individual animal health and productivity. In addition, the study Experimental design of early pathogenesis of BRD may also give insight into prevention, Animals were enrolled in 3 different groups: sick on arrival detection, and intervention strategies. (ARR-SA), pen-pull with undifferentiated fever (PP-UF), and pen- The objectives of this project were: i) to determine if a previously pull with no fever (PP-NF) with controls for each group. At the time described percutaneous lung biopsy technique could be used to of study enrollment, every animal had an ear notch taken for BVD characterize early pathologic changes in the bovine lung associated virus testing, as well as an ear tag placed to designate them as a with BRD, ii) to determine if specific infectious respiratory pathogens study animal. All study animals were followed from enrollment until could be identified in association with early pathological changes, feedlot exit (death or shipment to slaughter). All study animals were and iii) to determine whether pulmonary pathology characterized evaluated at the time of enrollment by a study veterinarian (BAB or by percutaneous lung biopsy at arrival and at the time of initial BRD CMP) to ensure fulfillment of study enrollment criteria. diagnosis were associated with health and production outcomes of Animals with BRD enrolled at feedlot arrival — Cases allocated feedlot steers in a commercial feedlot. as sick on arrival (ARR-SA; n = 27) had a rectal temperature . 408C at the time of processing, and no abnormal clinical signs referable to Materials and methods body systems other than the respiratory system (i.e., nasal discharge and/or coughing). See above for the antimicrobial given to sick animals at arrival. Allocated arrival controls (ARR-CT; n = 13) were Study facilities healthy animals with no clinical signs of illness and a rectal tem- This study took place in a commercial feedlot with a 30 000 perature , 39.48C at the time of processing. Controls were identified animal capacity in southern Alberta, Canada, in the fall of 2007. at the same time, and from the same processing group, as enrolled Animals were housed in side-by-side, open-air, dirt-floor pens with ARR-SA animals. See above for the antimicrobial given to controls a central feed alley and 20% porosity wood-fence windbreaks. Each at arrival. No more than 4 case and 2 control animals were allocated pen housed approximately 250 animals. This feedlot has a process- to the study from each processing group (a group of animals that ing facility and 2 hospital facilities. The study period was from represents one buyer). November 2007 to September 2008. Animals with BRD enrolled after arrival — Pen-pulls were identified by experienced feedlot personnel based on clinical signs Study animals of BRD including depression, decreased feed intake, nasal discharge The source population for this study consisted of 5059 animals and/or coughing, and lack of clinical signs referable to body systems distributed within 20 pens. Candidate animals for this study were other than the respiratory system. In addition, only those animals auction-market derived crossbred steer calves from western Canada between 3 and 30 d on feed (DOF) and with no previous history that were at high risk of developing BRD. One hundred animals that of treatment were eligible for enrollment in the study. There were met the inclusion criteria (outlined in the Experimental design sec- 3 groups of steers enrolled as pen-pulls. Pen-pulls with undifferenti- tion) participated in the study. The mean weight (6 standard devia- ated fever (PP-UF; n = 26) were those animals with a rectal tempera- tion) of the study animals was 286 kg (6 27 kg) with a minimum of ture . 40.58C. All PP-UF animals were treated with tulathromycin 211 kg to a maximum of 377 kg at study enrollment. (Draxxin Injectable Solution; Pfizer Canada), 2.5 mg/kg BW, SC, Within 18 h of feedlot arrival, all cattle received a modified live once; at the time of identification and returned to their home pens virus vaccine (Pyramid FP 5 Vaccine; Wyeth Animal Health, Guelph, (original pen). Pen-pulls with no fever (PP-NF; n = 14) were those Ontario) for bovine rhinotracheitis, bovine viral diarrhea (BVD) animals with a rectal temperature . 39.48C and , 40.58C. All PP-NF virus (types I and II), bovine parainfluenza-3, and bovine respiratory animals were treated with long-acting oxytetracycline (Oxymycin syncytial virus, a Mannheimia haemolytica toxoid (Presponse SQ bac- LA; Wyeth Animal Health), 20 mg/kg BW, IM (once) at the time of terin toxoid; Wyeth Animal Health, Guelph, Ontario) and a bacterin identification and returned to their home pens. Pen-pull controls toxoid (Ultrabac 7/ Somubac; Pfizer Canada, Kirkland, Quebec) for (PP-CT; n = 20) were healthy animals with a rectal temperature Clostridium chauvoei, C. septicum, C. novyi, C. sordelli, C. perfringens , 39.48C, that had no clinical signs of illness, and were identified at types C, and D, and Histophilus somni. In addition, all animals had the same time as an enrolled PP-UF or PP-NF. One control animal their body weight and rectal temperature recorded and received an was enrolled for every 2 cases enrolled in either the PP-UF or PP-NF

282 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

groups, and the control animal was selected from the same original interstitial reaction (i.e., damage or inflammatory cell infiltration of or home pen as the most recent case in each group. No more than the alveolar septa), syncytial cells, hyaline membranes, and fibrosis], 4 cases and 2 control animals were allocated from one pen after associated with the bronchoalveolar associated lymphoid tissue arrival; designated as pen-pulls. (BALT) (i.e., hyperplasia), associated with the pleura (i.e., neutrophilic infiltration and fibrinous pleuritis), and associated with vessels (i.e., Percutaneous lung biopsy of the right middle neutrophilic infiltration and fibrinoid necrosis). All study animals that lung lobe of feedlot steers in a field setting died or were euthanized were weighed at the time of death. A study A lung biopsy sample was taken within 24 h of initial diagnosis of veterinarian performed a field postmortem (9) and a cause of death BRD and again on weeks 2 and 4 using a previously described method was determined for each animal based on gross postmortem examina- (8). Briefly, the study animal was placed in a squeeze chute and tion. A digital image was obtained of the opened right thoracic cavity, sedated with xylazine (Rompun; Healthcare Animal Health Division, as well as a cross section of the lung tissue. Tissue samples from the Toronto, Ontario), 0.05 mg/kg BW, administered in the epaxial mus- cranioventral, middle, and caudal dorsal lung lobes were collected culature. An area on the right side of the thorax extending from the from each study animal that died and submitted to PDS for IHC and third intercostal space (ICS) to ICS6 and from mid-scapula extending histopathological evaluation, as described. 1 inch (2.5 cm) below the elbow was aseptically prepared. Lidocaine with epinephrine (Lidocaine; Lido2 with epinephrine, Rafter Products, Data collection and analysis Calgary, Alberta), 6 mL of 20 mg/mL, SC and IM, was administered to Individual animal data collected, using proprietary feedlot soft- locally block the musculature and subcutaneous tissues at the biopsy ware (FHARM software; Feedlot Health Management Services, site in ICS4 1 inch proximal of the elbow. A stab incision was made Okotoks, Alberta), included treatment history, health events through- using a #15 blade. The biopsy was taken with an automatic biopsy out the feeding period and, where applicable, cause of death. The instrument using an 8-gauge biopsy needle (Bard Magnum Biopsy entry weight and allocation weight data were analyzed (STATA/ Instrument; C.R. Bard, Inc., Covington, Georgia, USA). The needle IC version 10; Statacorp LP, College Station, Texas, USA) using a was inserted perpendicular to the body wall, angled slightly caudally one-way analysis of variance (ANOVA) as the data was normally to avoid the heart, and advanced to a depth of 7 to 10 cm for sample distributed based on the Shapiro-Wilks test for normality and the collection. As the animals gained weight the required depth for sample groups had equal variances using Scheffe’s test for equal variances. collection increased to . 10 cm. The stab incision was closed using The average daily gain (ADG), DOF, and final weight data were tissue staples (Appose ULC35; Tyco Healthcare, Norwalk, Connecticut, not normally distributed and were, therefore, analyzed using the USA). Study animals were observed by study personnel for epistaxis, Kruskal-Wallis equality of populations rank test. A P-value of  0.05 hemoptysis, and dyspnea for 1 h after the procedure. The study proto- was accepted as significant. col was approved by the institutional animal care and use committee of the University of Saskatchewan. Results Sample processing Lung samples were removed from the biopsy instrument using a Study animals 20-gauge needle and placed into individual wells within a sample A total of 100 steers were enrolled in this study; ARR-SA group cartridge. Up to 5 samples were placed into a cartridge with a con- consisted of 27 study animals and the ARR-CT group included 13 trol tissue (BVD virus positive bovine kidney). The control tissue control animals, the PP-UF group consisted of 26 study animals served as a reference for individual sample identification within the and the PP-CT group included 13 control animals, and the PP-NF cartridge, as well as a positive control for BVD virus immunohis- group consisted of 14 study animals and the PP-CT group included tochemistry. The cartridge was placed in a 10% formalin solution 7 control animals. All enrolled study animals were included in data and shipped daily to a diagnostic laboratory for processing [Prairie analysis. During the study period 4 steers died or were euthanized Diagnostic Services (PDS), Saskatoon, Saskatchewan]. Ear notches for humane reasons and 1 steer was railed (i.e., sold early for salvage were placed in peptone buffered saline for weekly processing. A slaughter) due to chronic bloat (at 177 DOF). Three of the 4 steers BVD virus antigen capture enzyme-linked immunosorbent assay died or were euthanized early in the feeding period ( 37 DOF). On (ELISA; HerdChek BVD Antigen ELISA Ear-Notch/Serum Test Kit; postmortem examination, 1 steer each was diagnosed with myocar- IDEXX Laboratories, Toronto, Ontario) was used in-house for BVD ditis (PP-UF), chronic fibrotic pleuritis (PP-UF), and chronic pneu- virus testing. Immunohistochemistry (IHC) was done at PDS for monia with arthritis (ARR-SA). In addition, 3 control steers were BVD virus, M. bovis, H. somni, and M. haemolytica. All lung biopsy randomly selected and euthanized for postmortem comparison to the samples were evaluated for the presence of lung tissue, as well as case animals that died; 2 PP-CT and 1 ARR-CT. No gross lesions of pathological changes by the same board certified veterinary patholo- BRD were found on postmortem examination of the control animals. gist in a blinded manner. This evaluation was based on a standard Later in the feeding period (at 149 DOF), 1 more steer died with form developed for the purposes of this study to assess the presence chronic pneumonia (PP-UF). As this death occurred much later in the of lung pathology associated with the bronchioles (i.e., neutrophilic feeding period, no control animal was euthanized for comparison. infiltration, abscessation, bronchiectasis, bronchiolar necrosis, syncy- The overall case fatality rate for this study was 4% (4/100) with an tial cells, and bronchiolitis obliterans), associated with the alveoli [i.e., ARR-SA case fatality rate of 3.7% (1/27), a PP-UF case fatality rate neutrophilic infiltration, abscessation, necrosis and fibrin, oat cells, of 11.5% (3/26), and a PP-NF case fatality rate of 0% (0/14).

2000;64:0–00 The Canadian Journal of Veterinary Research 283 FOR PERSONAL USE ONLY

Table I. Histopathologic findings of percutaneous lung biopsy samples obtained from feedlot steers in a commercial feedlot setting Allocation biopsy Week 2 biopsy Week 4 biopsy Treatment group Steer number (n = 100 steers) (n = 99 steers) (n = 96 steers)

ARR-SA 1 AIR no pathology no pathology

ARR-SA 2 ANI, BNI no pathology no pathology

PP-UF 3 AAM, AIR, ANI no biopsy ANI, AIR, AF

PP-UF 4 AAM, AIR, ANI, BNI, BO no pathology no pathology

PP-UF 5 ANI no lung no lung

PP-UF 6 ANI no pathology no pathology

PP-UF 7 ANI, BNI no pathology no pathology

PP-UF 8 ANI no pathology no pathology

PP-UF 9 ANI no lung no lung PP-UF 10 ANI no pathology no pathology PP-NF 11 BO BNI, ANI no lung PP-NF 12 ANI, BNI no pathology no pathology ARR-C 13 no pathology no pathology ANI, AAM ARR-C 14 ANI no pathology no pathology PP-C 15 ANI no pathology no pathology PP-C 16 no lung PNI n/a PP-C 17 ANI no pathology no pathology PP-C 18 no lung ANI no pathology PP-C 19 BNI no pathology no pathology AAM — alveolar activated macrophages; AF — alveolar fibrosis; AIR — alveolar interstitial reaction; ANI — alveolar neutrophilic infiltration; ARR-C — control for sick on arrival group; ARR-SA — sick on arrival; BNI — bronchiolar neutrophilic infiltration; BO — bronchiolitis obliterans; CT — control; n/a — not applicable (steer was euthanized as control); PNI — pleural neutrophilic infiltration; PP-C — control for pen-pull groups; PP-NF — pen-pull no fever; PP-UF — pen-pull with an undifferentiated fever.

Lung biopsies There were only 2 complications noted during the biopsy proce- Study animals were examined in the chute 3 times; at enrollment dures; 1 steer developed cardiac tamponade and died as a result of (n = 100), week 2 (n = 99) and week 4 (n = 96). A total of 295 biopsies the procedure, and 1 steer developed mild self-limiting epistaxis. were attempted with 283 (95.9%) successfully obtaining tissue. Of the 283 different biopsy samples, 210 (74%) contained lung tissue Immunohistochemistry and 73 (34.8%) contained tissue other than lung, such as liver or All study animals tested negative for BVD virus by capture skeletal muscle. Of the 210 lung biopsy samples obtained, 5 (2.4%) ELISA (HerdChek BVD Antigen ELISA Ear-Notch/Serum Test contained all of the expected lung structures i.e., bronchioles, alveoli, Kit; IDEXX Laboratories) on ear notches, indicating that none BALT, pleura, and vessels, and 137 (65.2%) contained only bronchi- of the study animals were persistently infected with BVD virus. oles and alveoli. Immunohistochemistry was done on every lung biopsy for M. bovis, Overall, histopathological changes were observed in 9.5% (20/210) H. somni, BVD virus, and M. haemolytica. One lung biopsy sample, of lung biopsy samples; representing 19 different steers. More from a steer in the PP-UF group, was positive for both M. bovis and lung lesions were identified at allocation than at week 2 or week 4 M. haemolytica. All other lung biopsy samples were negative for the (Table I). Lung lesions were identified in 17.9% (12/67; 1 case had antigens tested. 2 biopsies with histopathologic changes) of cases and 21.2% (7/33) of controls. Of the cases with lung lesions, 16.7% (2/12) were ARR-SA, Health events 66.7% (8/12; one case had 2 biopsies with histopathologic changes) Of the 100 steers enrolled in the study, 12 had at least 1 additional were PP-UF, and 16.7% (2/12) were PP-NF. There were 7 different health event recorded (by trained feedlot personnel or a feedlot histopathological changes noted in this study; alveolar neutrophilic veterinarian) throughout the feeding period, including BRD (no infiltration (ANI) was the most common, especially among PP-UF fever), BRD (undifferentiated fever), diphtheria, bloat, arthritis, and at allocation, followed by bronchiolar neutrophilic infiltration (BNI) inappetence. One steer (PP-NF) was found to be sick and identified (Table I). There were too few samples with lung pathology to reveal with a recurrence of BRD 22 d after enrollment, 5 d after the second any association between lung biopsy pathology and allocation group. lung biopsy was done. This animal was treated according to standard However, a higher percentage of animals with lung lesions were feedlot protocols and did not have another recurrence. This steer’s identified in the PP-UF group (9/20, 45%) compared to the other second biopsy revealed ANI and BNI, although the 1st and 3rd biop- case and control groups (Table I). sies did not show evidence of pathology. A second steer (ARR-CT)

284 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Table II. Body weights of enrolled commercial feedlot steers Mean arrival weight Mean enrollment weight Mean re-implantation weight Number in group kg (± SD) kg (± SD) kg (± SD) ARR-SA 27 297 (6 33)a 292 (± 27)b 521 (± 86) ARR-CT 13 294 (± 23) 288 (± 25) 520 (± 70) PP-UF 26 271 (± 29)a 273 (± 22)b,c 511 (± 83) c PP-CTUF 13 283 (± 32) 267 (± 24) 521 (± 87) PP-NF 14 264 (± 36)a 275 (± 30)d 529 (± 37) d PP-CTNF 7 288 (± 28) 305 (± 34) 556 (± 31) a,b,c,d SD — Significant difference (P # 0.05) between groups. ARR-SA — sick on arrival; ARR-CT — arrival controls; PP-UF — pen-pulls with undifferentiated fever; PP-CTUF — pen- pull controls with undifferentiated fever; PP-NF — pen-pulls with no fever; PP-CTNF — pen-pull controls with no fever. was found to be sick and diagnosed with BRD (undifferentiated revealed BNI, ANI, alveolar necrosis with fibrin, oat cells, fibrinous fever) 24 d after enrollment, 4 d after the 3rd biopsy was done. This pleuritis, and vascular neutrophilic infiltration. steer’s 3rd biopsy revealed ANI and activated alveolar macrophages Animals euthanized for comparison — Three control animals (AAM), while the first 2 biopsies did not show evidence of pathol- were randomly selected for euthanasia: 2 PP-CT and 1 ARR-CT. One ogy. This steer was randomly selected for postmortem comparison PP-CT had 2 successful biopsies done before euthanasia, biopsy 1 and was euthanized 13 d later. This steer had no visible pulmonary and 2. Biopsy 1 was of good quality without evidence of pathology; lesions on gross postmortem examination. There were too few biopsy and contained bronchioles, alveoli, pleura, and vessels. The second samples with lung pathology to reveal an association between lung biopsy was also of good quality without evidence of pathology, pathology and subsequent health events. and bronchioles, alveoli, BALT, pleura and vessels were observed. Histopathological evaluation of postmortem samples revealed ANI, Postmortem findings AIR, AAM, and BH. The second PP-CT animal euthanized had During the study period, 4 study animals died or were euthanized 2 biopsies taken with the 1st biopsy yielding no lung tissue. The 2nd due to poor treatment response or for humane reasons. The post- biopsy was of moderate quality, and contained alveoli and vessels mortem diagnoses included one each due to myocarditis, fibrotic without evidence of pathology and pleural neutrophilic infiltra- pleuritis, chronic pneumonia, and pneumonia with arthritis. tion (PNI). Histopathological changes on postmortem evaluation Animal with postmortem diagnoses of myocarditis — This ani- revealed ANI, AIR, AAM, BNI, and alveolar necrosis with oat cells. mal was from the PP-UF group. This steer had 1 biopsy done prior One ARR-CT was the 3rd control animal euthanized, and had 3 to death, yielding no lung tissue. Histopathological evaluation of successful biopsies taken prior to euthanasia. All 3 samples were of postmortem samples revealed bronchiolar necrosis (BN), alveolar good quality and contained bronchioles, alveoli, and vessels without interstitial reaction (AIR) and pleural hemorrhage, in addition to evidence of pathology. The 2nd biopsy sample also contained pleura, the myocarditis. while the 3rd contained pleura and ANI. Histopathologic examina- Animal with postmortem diagnosis of chronic fibrotic pleuritis — tion following postmortem evaluation revealed ANI, AIR, and AAM. This animal was from the PP-UF group. This steer had 2 lung biop- sies that yielded no lung tissue. Histopathological evaluation of Individual animal weight and average daily gain postmortem samples revealed bronchiectasis, bronchiolar fibrosis, Study animal weights at feedlot entry were significantly differ- AIR, and alveolar fibrosis (AF). ent between the ARR-SA and both the PP-UF and PP-NF groups Animal with postmortem diagnosis of pneumonia and arthritis — (P , 0.05; Table II). Additionally, study animal weights at alloca- This animal was from the ARR-SA group. This steer had 3 lung tion were significantly different between the ARR-SA group and biopsies done, yielding 1 sample of lung tissue (biopsy 1). On his- the PP-UF group, between the PP-UF group and their controls, topathologic examination, this sample was of good quality without and between the PP-NF and their controls (P , 0.05). At the time evidence of pathology and contained bronchioles, alveoli, and ves- of re-implantation (130 d into the feeding period) there were no sels. Histopathological evaluation of postmortem samples revealed significant differences (P = 0.69) between treatment group weights. bronchiolar abscessation, bronchiectasis, bronchiolar fibrosis, and At the time of re-implantation, average daily gain (ADG) was not diffuse AIR, in addition to arthritis. significantly different (P = 0.35) between groups. The mean (6 SD) Animal with postmortem diagnosis of pneumonia — This animal ADG for ARR-SA group (n = 27) was 1.56 kg (6 0.93 kg), and its was from the PP-UF group. This steer had 3 biopsies done, yield- control group (n = 13) was 1.80 kg (6 0.50 kg). The mean ADG for ing 2 lung tissue samples (biopsy 2 and 3). Biopsy 2 was of good PP-UF group (n = 26) was 1.75 kg (6 0.56 kg) and its control group quality without evidence of pathology and contained bronchioles, (n = 13) was 1.86 kg (6 0.55 kg). The mean ADG of the PP-NF group alveoli, BALT, and vessels. Biopsy 3 was also of good quality without (n = 14) was 1.91 kg (6 0.25 kg), and its control group (n = 7) was evidence of pathology, and bronchioles, alveoli, and vessels were 1.95 kg (6 0.09 kg). Additionally, the presence of lung biopsy histo- observed. Histopathological evaluation of postmortem samples pathology was not associated with ADG (P = 0.33).

2000;64:0–00 The Canadian Journal of Veterinary Research 285 FOR PERSONAL USE ONLY

Discussion indicate possible respiratory disease (e.g., fever, depression, and inappetence) could relate to other diseases such as gastrointestinal This percutaneous lung biopsy technique was effective at obtain- disease. In general, feedlot personnel are very skilled at determining ing lung samples of variable quality from feedlot steers in a com- which animals are becoming sick, however, a number of personnel mercial feedlot. With the exception of 1 fatal complication and 1 steer were used to evaluate the cattle over the course of this study. Some with self-limiting epistaxis, no abnormal clinical signs related to the feedlot personnel may have been more adept at selecting very acute biopsy procedure were observed in the remaining 98 steers. cases, while others may have required more obvious clinical signs In this study, attempts per biopsy sampling were limited to 2 before cattle were pulled for BRD. based on human literature which found no significant gain in diag- In this study, 20 lung samples from 19 steers were found to have nostic information with greater than 2 biopsies (10). In addition, a histopathological changes on evaluation. The small number of previous study on mature dairy cattle demonstrated that, on average, samples with pathological lesions limited our ability to find associa- only 1.2 samples were required when using an automatic biopsy tions between lung pathology and outcomes of interest (subsequent instrument (11). In our study, under field conditions, we obtained health events, average daily gain, enrollment group, or postmortem lung tissue 74% of the time from the right middle lung lobe, which findings). There were only 2 instances in which the same animal had is lower than other reports (97% to 98%) when using the same lung pathology on both the lung biopsy sample and on the postmor- instrumentation to sample the caudodorsal lung lobes of horses (12) tem sample, limiting our ability to make comparisons between lung and cows (11). This difference may be due to sampling location with biopsy samples and postmortem samples. However the postmortem the middle lung lobe being more difficult to sample given its close evaluations revealed more significant histopathological lesions proximity to the heart or due to a less controlled field environment than did the lung biopsies. This suggests that the percutaneous compared to a more controlled hospital environment. lung biopsy technique may be an inadequate sampling method for This study targeted steers considered to be at high-risk for the detecting early BRD in feedlot calves. development of BRD. As such, we had expected to demonstrate lung Interestingly, body weights were found to be significantly differ- pathology in more of the steers, especially in the PP-UF group and ent between ARR-SA, PP-UF, and PP-NF groups at study allocation, PP-NF groups at the time of enrollment. Overall, lung lesions were with the ARR-SA group having higher weights. One explanation may demonstrated by lung biopsy in 17.9% of cases and 30% of controls be that animals in the pen-pull groups had been sick for a longer with most of the lung lesions observed at the time of enrollment, duration, resulting in greater weight loss. Additionally, targeted 16.4% of cases and 17.4% of controls. The yield dropped considerably enrollment of cattle based on signs of BRD may have resulted in the with the 2nd and 3rd biopsies. selection of heavier weight cattle for the ARR-SA group strictly by The exact reasons for the low incidence of BRD based on histo- chance alone. Finally, small sample sizes likely affected the weight pathologic examination of lung biopsy samples are not known. One distributions. explanation may be that the sampling method was inadequate for Immunohistochemistry, a technique used to visualize the distribu- elucidating early lung pathology, which is typically focal in nature tion of antigens in tissue sections, is a useful technique for detecting and tends to localize in the cranioventral lung lobes. The focal nature microorganisms that are difficult to isolate, difficult to culture, or are of BRD has been demonstrated in a previous study assessing lung not viable. In this study, only 1 sample was positive for both M. bovis biopsy as a diagnostic tool (13), as well as in experimental infections and M. haemolytica using IHC. It is possible that no organisms were with M. haemolytica serotype A1 (14). In our study, lung biopsies present in the tissues, that the number of organisms present was were obtained blind (without the guidance of ultrasound or other below the level of detection for this technique, or that the distribu- lesion visualizing techniques) and, therefore, a lung lesion could tion of lung lesions, as well as microorganisms, may have been focal have been missed upon sampling with the small 8-gauge biopsy in nature (17,18). Previous research reported that vascular lesions instrument. The use of a blind biopsy technique was by design, as associated with BVD virus infection or BVD virus antigens were not our intention was to evaluate a previously developed lung biopsy present in all postmortem sections evaluated (18). Additionally, the technique that did not require additional diagnostic methods. This administration of antimicrobial therapy at arrival or for treatment research targeted the right middle lung lobe based on this previous may have affected the distribution of organisms in the 2nd or 3rd work done in preparation for the study reported here, in which the biopsy samples. In some cases, it can be difficult to demonstrate right cranial lung lobe proved difficult to consistently sample (8). antigen in tissue sections taken at postmortem in chronic cases of It is possible that the right middle lung lobe is less affected early in BRD (17). Our study targeted animals with acute disease (early the disease process compared to the right cranial lung lobe, mak- BRD) and attempted to use very small lung biopsy tissue sections to ing it less likely to obtain a lung biopsy with pathology within the identify disease. In addition, the biopsies were done without the use first 4 wk of the feeding period. The low incidence of BRD may also of ultrasound or other techniques as guidance for the biopsy instru- demonstrate the effectiveness of detection, prevention, and control ment. As a result, the use of IHC was likely limited by the size of the strategies by the feedlot. All steers received an injection of an antimi- tissue sample and the non-uniform distribution of lesions within the crobial as metaphylaxis at arrival. This may have resulted in delayed lung and of organisms/antigens within the tissue. development of respiratory disease beyond the duration of this The percutaneous lung biopsy procedure utilized in this study study, as has been shown in other research (15,16). These findings did not appear to cause long-lasting adverse effects to participating could also have been related to the specificity of our case definition animals, although expertise in the technique is essential to prevent or detection methods for BRD. Many of the clinical signs used to complications such as cardiac tamponade. The case fatality rate

286 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

observed in the study animals was comparable to, and actually 4. Andrews GA, Kennedy GA. Respiratory diagnostic pathology. less than, the feedlot’s overall fatality rate for fall placed calves. Vet Clin North Am Food Anim Pract 1997;13:515–547. The overall case fatality rate for this study was 4% with an ARR-SA 5. Wikse SE. Feedlot cattle pneumonia. Vet Clin North Am Food case fatality rate of 3.7%, a PP-UF case fatality rate of 11.5%, and a Anim Pract 1985;1:289–310. PP-NF case fatality rate of 0%. This differs slightly from the feedlots 6. Thompson PN, Stone A, Schultheiss WA. Use of treatment overall case fatality rates for respiratory disease. From 2007 to 2008, records and lung lesion scoring to estimate the effect of respira- the feedlot reported a case fatality rate of 2.86% for fall placed calves tory disease on growth during early and late finishing periods diagnosed as sick on arrival with BRD, 12.92% for calves diagnosed in South African feedlot cattle. J Anim Sci 2006;84:488–498. as pen-pulls with BRD and undifferentiated fever, and 9.86% for 7. Wittum TE, Woollen NE, Perino LJ, et al. Relationships among calves diagnosed as pen-pulls with BRD and no fever. The reasons treatment for respiratory tract disease, pulmonary lesions evi- for this are not known, although it is possible that study animals dent at slaughter, and rate of weight gain in feedlot cattle. J Am were inadvertently monitored more closely by feedlot personnel than Vet Med Assoc 1996;209:814–818. other animals in the feedlot. In addition, the sample sizes were small. 8. Burgess BA, Hendrick SH, Pollock CM, et al. The development The results of this study indicate that in general, percutaneous of a novel percutaneous lung biopsy procedure for use on feedlot lung biopsy can be done safely on feedlot steers in a commercial steers. Can J Vet Res 2011;75:254–260. feedlot with few clinical side effects. Despite our ability to recover 9. Booker CW, Janzen ED. Postmortem examination of feed- lung tissue that could be evaluated for histopathologic signs of BRD lot animals. Proceedings of the 24th Annual Convention, (bronchioles and alveoli), this technique did not prove useful either American Association of Bovine Practitioners. Orlando, Florida, as a diagnostic tool or prognostic indicator for cases with early BRD. 1991:131–133. However, this lung biopsy technique may be a useful diagnostic tool 10. Curley FJ, Johal JS, Burke ME, et al. Transbronchial lung biopsy: on a case by case basis, with image guidance, for BRD diagnosis in Can specimen quality be predicted at the time of biopsy? Chest a targeted population of cattle with chronic disease (more likely to 1998;113:1037–1041. sample an affected area). 11. Braun U, Estermann U, Feige K, et al. Percutaneous lung biopsy While the biopsy technique evaluated in this study was not use- in cattle. J Am Vet Med Assoc 1999;215:679-681. ful for characterizing early BRD in feedlot cattle under commercial 12. Venner M, Schmidbauer S, Drommer W, et al. Percutaneous conditions, additional studies are required to determine if other lung biopsy in the horse: Comparison of two instruments and techniques, such as ultrasound guided percutaneous biopsy, could be repeated biopsy in horses with induced acute interstitial pneu- useful for this purpose. The lack of sensitivity and specificity related mopathy. J Vet Intern Med 2006;20:968–973. to detection methods for BRD continues to be of critical importance 13. Sydler T, Braun U, Estermann U, et al. A comparison of biopsy in commercial feedlot production, and, as such, further research into and post-mortem findings in the lungs of healthy cows. J Vet the pathogenesis of BRD and diagnostic techniques to identify early Med A Physiol Pathol Clin Med 2004;51:184–187. BRD should continue to be supported. 14. Allan EM, Gibbs HA, Wiseman A, et al. Sequential lesions of experimental bovine pneumonic pasteurellosis. Vet Rec Acknowledgment 1985;117:438–442. 15. Morck DW, Merrill JK, Thorlakson BE, et al. Prophylactic efficacy This project was supported by Alberta Beef Producers, Calgary, of tilmicosin for bovine respiratory tract disease. J Am Vet Med Alberta, Canada. Assoc 1993;202:273–277. 16. Schumann FJ, Janzen ED, McKinnon JJ. Prophylactic tilmicosin References medication of feedlot calves at arrival. Can Vet J 1990;31:285–288. 17. Haines DM, Martin KM, Clark EG, et al. The immunohistochemi- . 1 Church TL, Radostits OM. A retrospective survey of diseases of cal detection of Mycoplasma bovis and bovine viral diarrhea virus feedlot cattle in Alberta. Can Vet J 1981;22:27–30. in tissues of feedlot cattle with chronic, unresponsive respiratory 2. Ribble CS, Meek AH, Jim GK, et al. The pattern of fatal fibrinous disease and/or arthritis. Can Vet J 2001;42:857–860. pneumonia (shipping fever) affecting calves in a large feedlot in 18. Shahriar FM, Clark EG, Janzen E, et al. Coinfection with bovine Alberta (1985–1988). Can Vet J 1995;36:753–757. viral diarrhea virus and Mycoplasma bovis in feedlot cattle with 3. Gagea MI, Bateman KG, van Dreumel T, et al. Diseases and chronic pneumonia. Can Vet J 2002;43:863–868. pathogens associated with mortality in Ontario beef feedlots. J Vet Diagn Invest 2006;18:18–28.

2000;64:0–00 The Canadian Journal of Veterinary Research 287 FOR PERSONAL USE ONLY

Article

Reversible chemical restraint of free-range cattle with a concentrated combination of tiletamine–zolazepam, ketamine, and detomidine Michela Re, Francisco J. Blanco-Murcia, José Maria San Miguel, Ignacio A. Gómez de Segura

Abstract The aim of this study was to determine the efficacy of a concentrated combination of tiletamine–zolazepam [TZ, 0.53 mg/kg body weight (BW)], ketamine (Ket, 0.53 mg/kg BW), and detomidine (Det, 0.04 mg/kg BW) in the immobilization of free-range cattle for clinical procedures. The combination was administered intramuscularly to 53 animals. Anesthesia was reversed with the a2-adrenoceptor antagonist atipamezole. Locoregional anesthesia was provided with lidocaine when required. The TZKD combination induced suitable immobilization for minor surgical procedures or medical treatments. Anesthetic onset was rapid, taking a mean of 6.1 min [standard deviation (SD) 2.8 min]. The duration of anesthesia depended on the time of administration of the antagonist; the animals recovered in the standing position in 12.9 6 8.9 min after the administration of atipamezole. The quality of anesthesia and analgesia were satisfactory. In conclusion, this TZKD combination can be used for both immobilization and minor surgical procedures in free-range cattle.

Résumé L’objectif de la présente étude était de déterminer l’efficacité d’une combinaison concentrée de tiletamine-zolazepan [TZ, 0,53 mg/kg de poids corporel (BW)], kétamine (Ket, 0,53 mg/kg de BW), et detomidine (Det, 0,04 mg/kg BW) pour l’immobilisation de bovins libres au pâturage aux fins de procédures cliniques. La combinaison fut administrée à 53 animaux par voie intramusculaire. L’anesthésie a été renversée avec l’atipamézole, antagoniste des adrénocepteurs-a2. Au besoin, l’anesthésie locorégionale a été induite avec de la lidocaïne. La combinaison TZKD a induit une immobilisation adéquate pour des procédures chirurgicales mineures ou des traitements médicaux. L’initiation de l’anesthésie était rapide, avec une moyenne de 6,1 minutes [écart-type (SD) 2,8 min]. La durée de l’anesthésie dépendait du temps de l’administration de l’antagoniste; les animaux récupérant en position debout en 12,9 6 8,9 min après l’administration de l’atipamézole. La qualité de l’anesthésie et de l’analgésie était satisfaisante. En conclusion, la combinaison TZKD peut être utilisée pour l’immobilisation et la réalisation de procédures chirurgicales mineures chez les bovins au pâturage. (Traduit par Docteur Serge Messier)

Introduction determine the clinical efficacy of TZKD used in free-range cattle under field conditions. Chemical restraint of free-range ruminants in the field has been obtained with a combination of the opioid etorphine and the tran- Materials and methods quilizer acepromazine (Immobilon; Novartis Animal Health UK Limited, Frimley, United Kingdom), which can be used in asso- Animals ciation with an a2-adrenoceptor agonist (1–3). Immobilon has been used extensively as an immobilization agent in ungulates. In Spain, A prospective observational clinical study from July 2007 and in several other countries, Immobilon has been banned for until September 2011 was conducted with free-range cattle from safety reasons or has legal restrictions as a scheduled substance. 8 ranches. Free-range bovines requiring chemical restraint for Pharmacologic alternatives include tiletamine and zolazepam (TZ) diagnostic or therapeutic (surgical or medical) reasons were plus ketamine (Ket), combined with an a2-adrenoceptor agonist included in the study independently of breed, gender, and age. (4–6). However, little information is available regarding the use of No animals were excluded because of administered treatment or these drugs in free-range cattle (7–10). Recently, a combination of health status. However, animals in which restraint was clearly not TZK with detomidine (Det) was tested in calves, and its efficacy as achieved, most likely owing to a failure in drug administration an immobilization combination agent producing anesthesia and (such as broken darts), were excluded, and their data were not analgesia was confirmed (11). The objective of this study was to recorded.

Department of Animal Medicine and Surgery, Veterinary Faculty, University Complutense, Avda, Puerta de Hierro s/n. 28040, Madrid, Spain (Re, Blanco-Murcia, Gómez de Segura); Pfizer Animal Health, Alcobendas, Spain (San Miguel). Address all correspondence to Dr. Ignacio A. Gómez de Segura; telephone: 134 3943858; fax: 134 3943808; e-mail: [email protected] Received May 31, 2012. Accepted October 24, 2012.

288 The Canadian Journal of Veterinary Research 2013;77:288–292 FOR PERSONAL USE ONLY

Drug combination and administration Table I. Procedures requiring immobilization and anesthesia with tiletamine–zolazepam, ketamine, and detomidine (TZKD) For the drug combination, 500 mg of 10% Ket (Imalgene 1000, in 53 free-range bovines 5 mL; Mérial Laboratorios, Barcelona, Spain) and 40 mg of 1% Det (Domosedan, 4 mL; Pfizer Animal Health, Alcobendas, Spain) Category of were added to a vial containing 250 mg each of lyophilized T and procedure Locoregional Z (Zoletil 100; Virbac, Esplugues de LLobregat, Spain). The final (number of Procedure technique solution contained 52.6 mg/mL each of TZ and Ket and 4.2 mg/mL animals) (number of animals) (number of animals) of Det; the total volume was 9.5 mL. Ophthalmologic (31) Subconjunctival Topical infiltration All of the animals were given the same dose, selected on the infiltration (31) of ocular globe (31) basis of pilot field studies, in a volume of 1 mL/100 kg of body Surgical (20) Orchiectomy (11) Testicular weight (BW): 0.53 mg/kg BW of TZ, 0.53 mg/kg BW of Ket, and Goring repair (3) infiltration (7) 0.04 mg/kg BW of Det. An additional dose (50% of the initial dose) Abscess Local wound was given to those animals whose procedures lasted longer than the treatment (3) infiltration (9) expected duration of a single dose and was administered when the Abdominal first spontaneous pinna movement was observed. The additional herniorrhaphy (2) dose was calculated from clinical criteria and in relation to the esti- Seroma mated time for completion of the clinical procedure. Body weight treatment (1) was estimated visually. Weight tapes were not considered, because the correlation coefficients have not been determined for these breeds Farm transfer (1) — — (12). When needed, locoregional anesthesia was provided with 2% Lameness (1) Joint infiltration — lidocaine plus epinephrine. The method of TZKD administration depended on the distance from the animal as well as the material and facilities available. Statistical analysis Remote intramuscular injection with a blowpipe 15 mm in diam- Means and standard deviations (SD) were calculated for the eter and high-performance 5-mL darts (Telinject USA; Agua Dulce, quantitative variables. Additionally, a decision tree was constructed California, USA) or a syringe was used whenever possible. The with the induction and recovery times considered as independent injection site depended on the field conditions but was chosen in variables. All of the analyses were done with SPSS software, ver- the following order of preference: back of the thigh, rump, loin, and sion 15 for Windows (SPSS, Chicago, Illinois, USA). neck. The effect of the combination was reversed in all of the animals Results with the a2-adrenoceptor antagonist atipamezole (Antisedan; Pfizer Animal Health, Alcobendas, Spain), 0.02 to 0.06 mg/kg BW, given For the 53 animals (50 males and 3 females) whose data were intravenously. The time of administration varied according to pro- included in the study the age ranged from 5 d to 15 y (mean 3.1 6 3.1 y) cedural requirements. and the body weight from 35 to 650 kg (mean 335 6 156 kg). Of these Data collection animals 42 were fighting bulls and 11 of the local red mottled strain. These animals were anesthetized for the reasons presented in Table I. The following times were recorded: onset of anesthesia (defined Most of the procedures lasted less than 1 h and were considered to be as the time from administration of the drug combination until the minimally or moderately painful. The procedure took longer for the animal became recumbent, either sternally or laterally), duration of repair of 2 abdominal hernias and 1 goring, and additional doses of the anesthesia (defined as the time from the onset of recumbence to the combination were administered to prolong the anesthesia. administration of atipamezole), recumbence time (defined as the The mean dose of the anesthetic combination was 1.07 6 0.13 mL/ amount of time that the animal remained recumbent, with its head 100 kg BW (Table II). Animals that received an additional dose on the ground), and recovery time (defined as the time at which the (half of the initial dose) did so an average of 42 6 3 min after the animal regained a standing position). initial dose. In 1 of the herniorrhaphy cases, 3 additional doses were Analgesia was evaluated according to the response of the animal required to prolong anesthesia, the total volume being 13.3 mL to a needle prick in the rump and was rated as excellent (absence of (2.9 mL/100 kg BW). The combination was administered by blow- response), fair (mild response, with mild muscle fasciculation), or pipe (n = 39), syringe (n = 11), or pole syringe (n = 3) to the back poor (very responsive, with moderate to severe muscle fasciculation) of the thigh (n = 21), rump (n = 18), loin (n = 6), neck (n = 4) and, (4). The quality of the anesthesia was rated as good (good anesthesia accidentally, tail (n = 1), shoulder blade (n = 1), rib cage (n = 1), or and analgesia), fair (good anesthesia and fair analgesia), or poor (fair head (n = 1). The analgesic response to a needle prick was excellent anesthesia and analgesia). in all cases. The quality of anesthesia was rated as good in 51 cases The volume of the anesthetic combination administered, the dose and fair in 2 cases (orchiectomy without locoregional anesthesia). and route of administration of the antagonist, the estimated weight Table III presents mean data for the anesthetic variables measured. of the animal, the method of administration, and the site of injection The duration of anesthesia ranged from 5 to 119 min. The decision- were also recorded. tree analysis showed no relationship between the recumbence time

2000;64:0–00 The Canadian Journal of Veterinary Research 289 FOR PERSONAL USE ONLY

Table II. Doses of the TZKD components and of the a2-adrenoceptor antagonist atipamezole, and the ratio of these agents, administered to the 53 animals Mean dose (6 SD) Mean ratio (6 SD) Units Tiletamine–zolazepam Ketamine Detomidine Atipamezole Atipamezole/detomidine mg/kg 0.56 6 0.07 0.56 6 0.07 0.045 6 0.005 0.03 6 0.01 0.7 6 0.2 mL/100 kg 1.07 6 0.13 0.5 6 0.2 BW — body weight; SD — standard deviation.

Table III. Times measured after administration of the anesthetic combination and the antagonist Animal group (number); mean time (6 SD), min Variable Total (53) Single dose (50) More than 1 dose (3) From TZKD administration until Onset of anesthesia 6.1 6 2.8 6.2 6 2.8 4.7 6 1.6 Head drop 7.9 6 3.4 8.0 6 3.5 6.3 6 2.5 Atipamezole administration 23.6 6 19.4 19.5 6 8.1 91.7 6 28.0 Duration of anesthesia 17.5 6 19.4 13.3 6 7.5 86.9 6 27.7 From atipamezole administration until Head lift 5.6 6 5.0 5.6 6 5.1 5.3 6 2.5 Standing 12.9 6 8.9 12.1 6 8.4 23.0 6 14.4 SD — standard deviation.

and the animal’s breed, farm, age, gender, or body weight, the dose/ In this observational study, the dose used in the free-range cattle, volume of TZKD, or the mode or area of administration. Of the 1 mL/100 kg BW, easily allowed the immobilization of bulls weigh- 53 animals, 26 lay in sternal recumbency and 27 in lateral recum- ing up to 650 kg. The volume of anesthetic drugs has been a major bency, 19 of the latter dropping their head to the ground. According limitation and has conditioned the choice of drugs and dosages and to the procedural requirements the animals were maintained in prevented the use of commercial formulations. In the present study, sternal (n = 33), lateral (n = 18), or dorsal (n = 2) recumbency. In all a solution of 10% ketamine and 1% detomidine was used as a solvent cases the head was positioned to prevent regurgitation. to dilute lyophilized tiletamine–zolazepam. This mixture allowed a The mean dose of atipamezole was 0.9 to 3.1 times higher than reduction in the total volume of the combination while maintaining the dose of detomidine (Table II). The antagonist was administered the required dose of each drug. 6 to 41 min after the TZKD except in the 3 animals that required Rapid induction of anesthesia is a key issue for chemical restraint additional doses; in those cases the antagonist was administered in wild animals (17) and reduces the likelihood of complications. 75 to 124 min after the initial dose of TZKD. The recovery time was Most animals receive anesthesia under stressful conditions and independent of the total dose/volume of TZKD or atipamezole and in some instances after a period of increased muscular activity of the atipamezole/detomidine dose ratio. before the drugs can be administered by dart. With the TZKD combination the time required for the induction of anesthesia Discussion was independent of the mode of administration (method and site), indicating its effectiveness. The mean induction time was The anesthetic combination TZKD produced adequate immobili- 6.1 min, similar to that found in a randomized controlled study of zation and anesthesia to allow all the procedures to be done in the the same drug combination and doses in calves (11). Furthermore, free-range cattle under field conditions. considering that the combination was administered to animals of The doses of drugs used in the anesthetic combinations adminis- different ages and breeds and in very different conditions (11), the tered to free-range cattle and wild ruminants are within a relatively anesthetic effects appear to be highly consistent. Similar induction narrow range, but high variability has been observed in their effects times (6 to 9 min) have also been observed in different breeds of (8,13–15), and not all combinations of dissociative anesthetics cattle (8). The induction time was 7.5 min in banteng (Bos javanicus) and a2-adrenoceptor agonists produce adequate anesthesia. Most restrained with detomidine, tiletamine, and zolazepam (15) and in anesthetic combinations cannot be used in wild ruminants or free- wood bison restrained with a combination of medetomidine, tilet- range cattle owing to the volume that would usually be required amine, and zolazepam (16). Although times may exceed 10 min (15), for administration by dart and several injections may be required the ideal induction time would be shorter to allow the use of these to reach the dose necessary to produce immobilization (14–16), combinations under field conditions (15) and to prevent capture which may preclude the use of these agents in field conditions. myopathy (14).

290 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Reversal of anesthesia is a requirement under field conditions combination, is the most likely reason for the lack of sedation in the to limit the possibility of attacks from surrounding animals. In this excluded animals. study the antagonist was administered within 30 min except in An important limitation of the study was that oxygenation was 6 animals, for which administration was 30 to 40 min after the dart not determined. In a previous study in calves, the most important was used, suggesting that the TZKD combination induced anesthesia adverse effect of this anesthetic combination was respiratory depres- lasting at least this long. Moreover, animals that required a longer sion leading to potentially dangerous hypoxia (11). Although oxygen anesthesia time received the first additional dose no earlier than supplementation could be suggested, it is not a realistic option under 40 min after initial administration of the combination. In fact, the field conditions; therefore, it is highly advisable to adequately posi- anesthesia time was slightly longer in our study than that observed tion the animal in right recumbency with the head lifted to avoid in young calves (30 6 8 min) (11). regurgitation and to monitor the mucous membranes. Analgesia was not tested or measured, although the level was ade- In conclusion, low volumes of the TZKD combination adminis- quate to perform all procedures, including the surgical procedures. tered with darts allowed the immobilization and anesthesia of wild Nevertheless, locoregional techniques, including local infiltration in cattle. The induction of anesthesia was rapid, and the duration was the incision, were used in the cases of hernia repair, orchiectomy, and sufficient to perform all of the procedures. However, additional ophthalmologic procedures (altogether in 47 of the 53 cases), and the doses could be administered to prolong the anesthesia time when local anesthesia would have improved the analgesia. necessary. In all cases, the effects of the TZKD combination were The animals were not fasted, although this factor does not seem clinically reversed with the antagonist atipamezole. to have resulted in side effects such as regurgitation or bloat, which have often been described in ruminants (18). There was increased References production of saliva, which is normally observed in ruminants under injectable anesthesia (19). However, the saliva production . 1 Blix AS, Lian H, Ness J. Immobilization of muskoxen (Ovibos was not considered to be profuse and did not cause sufficient accu- moschatus) with etorphine and xylazine. Acta Vet Scand 2011; mulation in the oral cavity to require aspiration or endotracheal 53:42. intubation. 2. Howard LL, Kearns KS, Clippinger TL, Larsen RS, Morris PJ.

The anesthesia was reversed with atipamezole, an a­ 2-adrenoceptor Chemical immobilization of rhebok (Pelea capreolus) with antagonist. Although tiletamine, zolazepam, and ketamine were carfentanil–xylazine or etorphine–xylazine. J Zoo Wildl Med not antagonized, clinical reversal of the combination is easily 2004;35:312–319. obtained and also allows the reversal of respiratory depression (20). 3. McMahon CR, Bradshaw CJ. To catch a buffalo: Field immobilisa- Phencyclidines (21,22) are N-methyl-D-aspartate receptor antago- tion of Asian swamp buffalo using etorphine and xylazine. Aust nists and cannot be antagonized. The dose of atipamezole used in Vet J 2008;86:235–241. this study (0.02 to 0.06 mg/kg BW) allowed the reversal of anes- 4. Thurmon JC, Lin HC, Benson GJ, Tranquilli WJ, Olson WA. thesia without causing excitation. The relatively low atipamezole/­ Combining Telazol and xylazine for anesthesia in calves. Vet detomidine dose ratio of 0.7:1 used in this study was chosen on Med 1989;84:824–830. the basis of the results of pilot studies in which rapid reversal of 5. Lin HC, Thurmon JC, Tranquilli WJ, Benson GJ, Olson WA. anesthesia was achieved with the lowest possible dose. Dose ratios Hemodynamic response of calves to tiletamine–zolazepam–­ between 1.1:1 and 1.6:1 have also been used in nondomestic rumi- xylazine anesthesia. Am J Vet Res 1991;52:1606–1610. nants (15,23), although higher dose ratios, between 3:1 and 6:1, are 6. Waterman AE. Preliminary observations on the use of a combina- commonly administered (10). Although unlikely, a severe resedation tion of xylazine and ketamine hydrochloride in calves. Vet Rec effect in the animals in this study was not assessed, and therefore 1981;109:464–467. it is possible that higher dose ratios might be necessary to ensure 7. Hofkes LM, Hoyer MJ, van DP, Overgaauw PA. Immobilization full antagonism. of cattle and bison with a combination of xylazine, zolazepam– High variability in recovery time has been observed after admin- tiletamine and ketamine. Tijdschr Diergeneeskd 2005;130: istration of the a2-adrenoceptor antagonist. Differences between 268–272. species and breeds, anesthetic drugs and doses, chosen antagonist 8. Hoyer MJ, Hofkes L, Dijk P, Van Pieterse MC. Zolazepam–­ and dose, and route of administration may account for the observed tiletamine–ketamine–xylazine anaesthesia: An effective, safe differences. The antagonist reverses only the effects of agonists of and reliable alternative for the use of etorphine in feral cattle a2-adrenergic receptors; thus, the recovery time and its quality can and bison. In: Erkrankungen der Zootiere. Verhandlungsbericht be influenced by residual effects of other drugs used concurrently. des 41 Internationalen Symposiums über die Erkrankungen der The present study excluded animals that failed to respond as Zoo- und Wildtiere [41st International Symposium on Diseases expected to the anesthetic combination. Failure of anesthesia deliv- of Zoo and Wild Animals]; 2003 May 28–June 01; Rome, Italy; ery was determined when animals were not recumbent within the Dortmund, Akademie Verlag Berlin, 2003:367. first 10 min after drug administration. No record was made of the 9. Caulkett NA, Arnemo JM. Chemical immobilization of free- excluded animals, although they could have provided additional ranging terrestrial mammals In: Tranquilli WJ, Thurmon JC, information related to the rate of administration failure. Owing to Grimm KA, et al, eds. Lumb & Jones Veterinary Anesthesia the consistency of the results obtained from the animals included in and Analgesia. 4th ed. Ames, Iowa: Blackwell Publishing, 2007: the study, administration failure, rather than variability of the TZKD 807–831.

2000;64:0–00 The Canadian Journal of Veterinary Research 291 FOR PERSONAL USE ONLY

10. Curro TG. Non-domestic cattle. In: West G, Heard D, Caulkett N, 18. Carroll GL, Hooper RN, Slater MR, Hartsfield SM, Matthews NS. eds. Zoo Animal and Wildlife Immobilization and Anesthesia. Detomidine–butorphanol–propofol for carotid artery transloca- Ames, Iowa: Blackwell Publishing, 2007:635–642. tion and castration or ovariectomy in goats. Vet Surg 1998;27: 11. Re M, Blanco-Murcia FJ, Gomez de Segura IA. Chemical restraint 75–82. and anaesthetic effects of a tiletamine–zolazepam/ketamine/ 19. Valverde A, Doherty TJ. Anesthesia and analgesia of ruminants. detomidine combination in cattle. Vet J 2011;190:66–70. In: Fish R, Danneman P, Brown M, Karas A, eds. Anesthesia and 12. Enevoldsen C, Kristensen T. Estimation of body weight from Analgesia in Laboratory Animals. 2nd ed. Amsterdam: Elsevier body size measurements and body condition scores in dairy Academic Press, 2008:385–411. cows. J Dairy Sci 1997;80:1988–1995. 20. Campbell KB, Klavano PA, Richardson P, Alexander JE. 13. Bush M, Citino SB, Tell L. Telazol and Telazol/Rompun anes- Hemodynamic effects of xylazine in the calf. Am J Vet Res 1979; thesia in non-domestic cervids and bovids. In: Proceedings 40:1777–1780. of the Joint Conference of the American Association of Zoo 21. Ko JC, McGrath CJ. Effects of atipamezole and yohimbine on Veterinarians and the American Association of Wildlife medetomidine-induced central nervous system depression Veterinarians, 1992 Nov 15–19; Oakland, California:224–225. and cardiorespiratory changes in lambs. Am J Vet Res 1995;56: 14. Wilson SC, Armstrong DL, Simmons LG, Morris DJ, Gross TS. 629–632. A clinical trial using three regimens for immobilizing gaur 22. Rioja E, Kerr CL, Enouri SS, McDonell WN. Sedative and cardio- (Bos gaurus). J Zoo Wildl Med 1993;24:93–101. pulmonary effects of medetomidine hydrochloride and xylazine 15. Bradshaw CJ, Traill LW, Wertz KL, White WH, Gurry IM. hydrochloride and their reversal with atipamezole hydrochloride Chemical immobilisation of wild banteng (Bos javanicus) in in calves. Am J Vet Res 2008;69:319–329. northern Australia using detomidine, tiletamine and zolazepam. 23. Laricchiuta P, De Monte V, Campolo M, et al. Evaluation of Aust Vet J 2005;83:616–617. a butorphanol, detomidine, and midazolam combination for 16. Caulkett NA, Cattet MR, Cantwell S, Cool N, Olsen W. Anesthe­ immobilization of captive Nile lechwe antelopes (Kobus maga- sia of wood bison with medetomidine–zolazepam/tiletamine ceros). J Wildl Dis 2012;48:739–746. and xylazine–zolazepam/tiletamine combinations. Can Vet J 2000;41:49–53. 17. Arnemo JM, Soli NE. Immobilization of free-ranging cattle with medetomidine and its reversal by atipamezole. Vet Res Commun 1995;19:59–62.

292 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Article

Effect of storage time and temperature on the total protein concentration and electrophoretic fractions in equine serum Daniela Alberghina, Stefania Casella, Claudia Giannetto, Simona Marafioti, Giuseppe Piccione

Abstract Serum protein electrophoresis (SPE) is a technique that could be considered one of the most useful diagnostic aids available to the clinician. The effect of storage time and temperature on the total proteins and electrophoretic fractions (albumin, a1-, a2-, b1-, b2-, and g-globulins) was assessed in 24 healthy horses. All samples, collected by jugular vein puncture, were centrifuged and divided into 4 aliquots. The 1st aliquot was analyzed within 3 h from collection (time 0), the 2nd was refrigerated at 14°C for 24 h, the 3rd was refrigerated at 14°C for 48 h, and the last was frozen at 220°C for 48 h. One-way repeated-measures analysis of variance (ANOVA) showed a significant effect (P , 0.05) of the different storage conditions on the concentrations of all the parameters studied and significant variations in the percentages of albumin, a1-globulins, a2-globulins, and g-globulins. Compared with time 0 the total protein concentration increased significantly after 48 h at 220°C, the albumin percentage decreased after 48 h at

220°C, the a1-globulin percentage increased after 24 h at 14°C, the a2-globulin percentage increased after 48 h at 14°C and at 220°C, and the g-globulin percentage increased after 48 h at 220°C. The results should help veterinary practitioners handle and store equine serum samples appropriately. Further investigations at different storage times and temperatures could be useful.

Résumé L’électrophorèse des protéines sériques (SPE) est une technique qui pourrait être considérée comme un des outils diagnostiques les plus utiles au clinicien. L’effet du temps et de la température d’entreposage sur les protéines totales et les fractions électrophorétiques (albumine, a1-, a2-, b1-, b2-, et g-globulines) a été évalué chez 24 chevaux en santé. Tous les échantillons ont été prélevés par ponction de la veine jugulaire, centrifugés et divisés en quatre aliquots. Le premier aliquot a été analysé en dedans de trois heures du moment de la collecte (temps 0), le deuxième a été réfrigéré à 4 °C pour 24 h, le troisième a été réfrigéré à 4 °C pour 48, et le dernier a été congelé à 220 °C pendant 48 h. Une analyse de variance unidirectionnelle sur des mesures répétées (ANOVA) a montré un effet significatif (P , 0,05) des différentes conditions d’entreposage sur les concentrations de tous les paramètres étudiés et des variations significatives dans les pourcentages d’albumine, d’a1-globuline, d’a2-globuline, et de g-globuline. Comparativement au temps 0, la concentration de protéines totales a augmenté significativement après 48 h à 220 °C, le pourcentage d’albumine a diminué après 48 h à 220 °C, le pourcentage d’a1-globuline a augmenté après 24 h à 4 °C, le pourcentage d’a2-globuline a augmenté après 48 h à 4 °C et à 220 °C, et le pourcentage de g-globuline a augmenté après 48 h à 220 °C. Ces résultats devraient aider les vétérinaires praticiens à manipuler et entreposer de manière appropriée les échantillons de sérum équin. Des études ultérieures sur différents temps et températures d’entreposage seraient utiles. (Traduit par Docteur Serge Messier) Introduction pain, and polyuria are clinical signs for which evaluation of the serum protein fractions is recommended (1). For diagnostic value, The serum total proteins represent the sum of numerous differ- the measurements for all fractions must be reliable and correctly ent proteins, many of which vary independent of each other. Since interpreted. To maximize the diagnostic value of SPE for clinical measurement of the serum total protein concentration in blood is laboratories it is essential to have access to well-established relative useful to evaluate, diagnose, and monitor a variety of diseases and and absolute reference values (3). Ideally reference intervals for conditions, it is one of the most frequent routine analyses done to serum or plasma parameters should be established in each labora- investigate electrolyte disorders, inflammatory or infectious diseases, tory; however, especially in veterinary laboratories, the establish- colostrum intake, and tumors (1). Its routine determination is also ment of reference intervals is expensive, and the ideal may not be a prerequisite of serum protein electrophoresis (SPE) (2), the most feasible for all laboratories (4). common means of fractionating serum proteins. If the results are Standard guidelines for blood sample handling state that plasma properly interpreted, SPE could be considered one of the most useful or serum should be separated from cells as soon as possible (within diagnostic aids available to the clinician. This technique is used in 20 to 30 min) after clot formation is complete to avoid clot-induced equine medicine for diagnosis, monitoring, and prognosis of many changes in the concentration of serum analytes (5). Inadequate stor- diseases that cause changes in albumin and globulin concentra- age of biologic samples, a potential source of preanalytic error, may tions (1). Poor performance, depression, fever, diarrhea, abdominal markedly affect the concentrations of biochemical variables (6,7).

Dipartimento di Scienze Veterinarie, Università degli Studi di Messina, Polo Universitario dell’Annunziata, 98168 Messina, Italy. Address all correspondence to Prof. Giuseppe Piccione; telephone: 139 903503584; fax: 139 903503975; e-mail: [email protected] Received June 26, 2012. Accepted September 5, 2012.

2013;77:293–296 The Canadian Journal of Veterinary Research 293 FOR PERSONAL USE ONLY

Table I. Average values and significance of differencesa for the total protein concentration, the percentage and concentration of protein fractions, and the albumin/globulin (A/G) ratio obtained by electrophoresis of serum from 24 healthy horses analyzed under different experimental conditions Experimental condition; mean 6 standard deviation Within 3 h after collection After 24 h After 48 h After 48 h Parameter (time 0) at 14°C at 14°C at 220°C Total proteins (g/L) 51.90 6 3.00 49.70 6 3.20 52.30 6 3.00 67.50 6 8.70b Albumin (%) 56.07 6 5.18 55.08 6 5.49 54.26 6 4.22 53.56 6 7.38b Albumin (g/L) 30.00 6 4.00 30.30 6 7.10 28.30 6 2.10 36.20 6 7.30b,c,d b a1-globulins (%) 1.76 6 0.39 2.12 6 0.53 1.83 6 0.33 1.82 6 0.47 b,d a1-globulins (g/L) 1.00 6 0.20 1.10 6 0.30 1.00 6 0.20 1.20 6 0.40 b b a2-globulins (%) 7.14 6 1.46 7.56 6 1.55 7.78 6 1.35 8.02 6 1.60 b,c,d a2-globulins (g/L) 3.80 6 1.10 4.10 6 1.00 4.10 6 0.73 5.40 6 1.20 b1-globulins (%) 11.98 6 2.16 12.09 6 1.89 12.23 6 1.89 12.49 6 2.08 b,c,d b1-globulins (g/L) 6.50 6 1.90 6.50 6 1.3 6.40 6 1.10 8.40 6 1.80 b2-globulins (%) 6.12 6 2.27 6.25 6 2.67 6.18 6 2.14 6.17 6 2.40 b,c,d b2-globulins (g/L) 3.30 6 1.20 3.40 6 1.5 3.30 6 1.10 4.20 6 1.60 g-globulins (%) 16.50 6 3.94 16.50 6 4.64 17.70 6 3.15 17.90 6 4.54b g-globulins (g/L) 9.00 6 3.40 9.60 6 3.40 9.30 6 2.00 12.20 6 3.80b,c,d A/G ratio 1.28 6 0.24 1.28 6 0.33 1.20 6 0.20 1.22 6 0.46 a Significantly different (P , 0.05) from the values, bat time 0, cafter 24 hours at 14°C, and dafter 48 hours at 14°C by the Bonferroni multiple-comparison test.

Therefore, changes in gas composition and acid–base values (8) and standard ration comprised hay (first-cut meadow hay, sun-cured, clotting parameters (9) under different storage conditions have been late-cut, 8 kg/d; 6.9% crude protein on average) and a 50:50 mixture evaluated. However, the influence of storage conditions on total of oats and barley (approximately 3.5 kg/d). The daily ration was proteins and electrophoretic fractions in veterinary medicine is less offered 3 times a day: at 7 am, 1 pm, and 6 pm. The mean composi- well-documented. In the veterinary literature the effect of storage on tion values of the ration were 87% dry matter and 13% moisture. The colorimetric serum total protein stability has been limited to dog, rat, dry matter contained 9.1% horse-digestible protein, 12.1% crude pro- and avian samples (6,10,11), and the effect of storage conditions on tein, 20.7% crude fiber, and 3.4% ether extract, as well as 0.80 horse the stability of serum protein fractions has been shown for cattle (12). feed units per 7 kg. Water was available ad libitum. In horses the storage conditions for measurement of serum or plasma Blood was collected by jugular vein puncture into 10-mL protein concentrations have not yet been sufficiently standardized Vacutainer tubes (Terumo Corporation, Japan) without antico- (8,9). Considering that in equine clinical practice there is a need to agulant and then centrifuged at 755 3 g for 10 min. The obtained carry the samples from stable to laboratory, optimal management of serum, removed with a plastic pipette and transferred into gradu- the preanalytic phase is important (13). Thus, the aim of this study ated Eppendorf microtubes (LP Italiana Spa, Milano, Italy), was was to evaluate the effect of short-term storage (analysis within 3 h divided into 4 aliquots and transported to the diagnostic laboratory after collection, after 24 h at 14°C, after 48 h at 14°C, and after 48 h for assessment of the total proteins and electrophoretic fractions 2 a g at 20°C) on total proteins and electrophoretic fractions (albumin (albumin and a1-, 2-, b1-, b2-, and -globulins). The 1st aliquot a g and a1-, 2-, b1-, b2-, and -globulins) in healthy horses. was analyzed within 3 h after collection, the 2nd was refrigerated at 14°C for 24 h, the 3rd was refrigerated at 14°C for 48 h, and the Materials and methods last was frozen at 220°C for 48 h. The values obtained within 3 h after collection were considered the initial concentrations, at time 0. Blood samples were obtained from 24 clinically healthy Italian Before the analyses the samples were thawed at room temperature Saddle Horses aged 6 to 10 y with a mean body weight of 530 6 20 (18°C to 20°C). (SD) kg. The health status of the horses was checked by physical The total protein concentration was measured at 125°C by the biu- examination. The animals’ vaccinations were up-to-date, and the ret method with the use of commercially available reagents and an horses were free from internal and external parasites. All housing automated-analyzer ultraviolet spectrophotometer according to the and care conformed to the standards recommended by the Guide procedures described by the manufacturer. This method is the most for the Care and Use of Laboratory Animals and Directive 86/609 widely used colorimetric method for the determination of the total CEE (14). All horses were fed standard rations, calculated to fulfill protein concentration in serum because of its simplicity, precision, their nutritional requirements according to the specifications of the and accuracy (2). The method is highly accurate for the range of total Institut national de la recherche agronomique, Paris, France (15). The proteins likely to be found in serum (10 to 100 g/L), with linearity

294 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

(a) (b)

Figure 1. Cellulose acetate electrophoretograms of serum proteins of a clinically healthy horse obtained (a) within 3 h after sample collection and (b) after storage for 48 h at 220°C. Bands were visualized by Red Ponceau S staining and densitometer scanning. of 0.1 g/L and sensitivity of 0.05 g/L. The amount of protein in the Results serum was then quantified by measuring the absorbance at 540 nm and comparing it with the absorbance of a solution containing a The total protein concentrations ranged from 44.60 to 57.30 g/L known concentration of a standard (bovine albumin, 52.8 g/L). The within 3 h after collection, from 45.00 to 55.50 g/L after 24 h at absorbance of each sample was measured in duplicate. 14°C, from 44.70 to 56.60 g/L after 48 h at 14°C, and from 51.70 to The protein fractions were measured by an automated system 78.90 g/L after 48 h at 220°C. A significant effect (P , 0.001) of the according to the procedures described by the manufacturer. For experimental conditions was observed for the total protein concen- each sample, 25 μL of serum was applied to numbered wells. Each tration (F[2,46] = 89.48), which significantly increased after 48 h at holder accommodated up to 24 samples. The proteins migrated in 220°C with respect to the previous experimental conditions (Table I). an electric field of 450 V for about 30 min. Electrophoresis was done A significant effect (P , 0.05) of the different storage conditions was on a supporting medium of cellulose acetate strips 76 3 60 mm. also observed for the concentrations of all the protein fractions, The buffering solution used was Tris-hippurato and distilled water whereas ANOVA showed a significant effect on the percentages , (1:10). After the electrophoresis the strips were colored with Red of only albumin (F[2,46] = 3.24; P 0.05), a1-globulins (F[2,46] = 4.10; Ponceau S for about 3 min. For destaining, a solution of orthophos- P , 0.01), a2-globulins (F[2,46] = 7.99; P , 0.001), and g-globulins phoric acid and distilled water (1:20) was applied for about 10 min. (F[2,46] = 4.31; P , 0.01). The Bonferroni multiple-comparison test Finally, the strips were analyzed in a scanner densitometer, in which revealed a significant effect (P , 0.05) of the experimental conditions electrophoretic curves with related specific protein concentrations on the albumin percentage, which decreased after 48 h at 220°C for each sample were displayed. The major protein fractions were versus time 0, and consequently the globulin percentages increased a , 1 divided from cathode to anode as albumin and a1-, 2-, b1, b2-, and significantly (P 0.05): a1-globulins after 24 h at 4°C versus time g-globulins, respectively. All samples were assayed in duplicate by 0, a2-globulins after 48 h at 14°C and after 48 h at 220°C versus time the same person each time. Samples exhibited displacement parallel 0, and g-globulins after 48 h at 220°C versus time 0. to the standard curve; the intra-assay coefficient of variation was Electrophoretograms for samples from the same horse analyzed , 7% and the interassay coefficient of variation , 9% for all the within 3 h after collection and after 48 h at 220°C are shown in Figure 1. parameters measured. The relative concentrations of the protein fractions were determined as a percentage of the optical absorbance Discussion and as a concentration expressed in grams per liter. All results were expressed as mean 6 SD. All data were normally The results showed that time and temperature have a statistically distributed (P , 0.05; Kolmogorov–Smirnov test). One-way repeated- significant effect on the total proteins and their fractions in equine measures analysis of variance (ANOVA) was used to determine the serum. The fractions and the A/G ratio were within the range of significance of differences related to storage conditions (time and values obtained in horses by electrophoresis in other studies (3). As temperature); P-values , 0.05 were considered statistically signifi- previously found for other parameters in equine samples (16), the cant. The Bonferroni multiple-comparison test was used for post- total protein values increased after 48 h of freezing. In fact, storage hoc comparison. Data were analyzed with the Statistica 8 software of serum for up to 48 h resulted in a significant increase in total package (StatSoft, Tulsa, Oklahoma, USA). proteins only after 48 h at 220°C. Modifications of the total ­protein

2000;64:0–00 The Canadian Journal of Veterinary Research 295 FOR PERSONAL USE ONLY

concentration as an effect of storage conditions could be due to 3. Riond B, Wenger-Riggenbach B, Hofmann-Lehmann R, Lutz H. species-specific differences (6,10,11) or simply due to the high num- Serum protein concentrations from clinically healthy horses ber of samples used in this study, which could influence the results determined by agarose gel electrophoresis. Vet Clin Pathol 2009; of statistical analysis. 38:73–77. It is important to establish a range for serum albumin values 4. Stockham SL. Interpretation of equine serum biochemical profile because albumin has several important roles: a contribution to colloid results. Vet Clin North Am Equine Pract 1995;11:391–414. oncotic pressure, maintenance of vascular permeability, modulation of 5. Young DS, Bermes EW. Specimen collection and processing: coagulation, and as a carrier of many substances. In comparison with Sources of biological variation. In: Burtis CA, Ashwood ER, eds. the results of other studies, in our study the albumin and g-globulin Tietz Textbook of Clinical Chemistry. Philadelphia, Pennsylvania: a WB Saunders, 1999:42–72. percentages were higher (15–17) and the a1- and 2-globulin values lower (3,16–18). Thus, the results of this study suggest that the storage 6. Cray C, Rodriguez M, Zaias J, Altman NH. Effects of storage of equine serum for 24 h and 48 h at 14°C does not have a significant temperature and time on clinical biochemical parameters from effect on total proteins, albumin, and g-globulins but has a significant rat serum. J Am Assoc Lab Anim Sci 2009;48:202–204. effect on a-globulins, and that the storage of serum for 48 h at 220°C 7. Ehsani A, Afshari A, Bahadori H, Mohri M, Seifi HA. Serum con- considerably modifies all these parameters. stituents analyses in dairy cows: Effects of duration and temper- These findings demonstrate that serum protein fractions can be ature of the storage of clotted blood. Res Vet Sci 2008;85:473–475. assessed within 48 h of blood collection when samples are stored at 8. Piccione G, Bertolucci C, Grasso F, Giudice E. Changes in gas 14°C because refrigeration does not change the results of analysis composition and acid–base values of venous blood samples except for a-globulins. Thus, short-term storage could influence the stored under different conditions in 4 domestic species. Vet Clin Pathol 2007;36:358–360. electrophoretic fractions and the A/G ratio. The latter is of special 9. Casella S, Giannetto C, Fazio F, et al. Assessment of prothrombin interest to pathologists because it allows for systematic classification time, activated partial thromboplastin time, and fibrinogen con- of the electrophoretic profile and identification of dysproteinemias centration on equine plasma samples following different storage (10). However, there is a discrepancy in the veterinary literature conditions. J Vet Diagn Invest 2009;21:674–678. regarding values for the physiological A/G ratio in horses (3,16–18). 10. Hawkins MG, Kass PH, Zinkl JG, Tell LA. Comparison of bio- Equine blood samples should be transported to the diagnostic labo- chemical values in serum and plasma, fresh and frozen plasma, ratory within a few hours and analyzed within 48 h at 14°C. The and hemolyzed samples from orange-winged Amazon parrots results found for equine species agree with those previously reported (Amazona amazonica). Vet Clin Pathol 2006;35:219–225. for bovine species (12) and are probably due to a different effect of 11. Thoresen SI, Tverdal A, Havre G, Morberg H. Effects of storage temperature and time of storage on degradation of the molecular con- and freezing temperature on clinical chemical parameters from figuration of albumin with respect to globulins. Almost all proteins canine serum and heparinized plasma. Vet Clin Pathol 1995;24: are glycoproteins, whereas albumin contains no carbohydrate (2). 129–133. Albumin has several molecular forms and exists as a monomer and in 12. Tóthová C, Nagy O, Seidel H, Kovacs G. The effect of storage 1 higher aggregate states. The monomer is stable for 5 months at 8°C, on the protein electrophoretic pattern in bovine serum. Iranian whereas the dimer and the tetramer dissociate into smaller forms J Vet Sci Technol 2010;2:77–84. within a few hours (19). The decrease of albumin found in equine 13. Braun JP, Trumel C, Bézille P. Clinical biochemistry in sheep: species agrees with results obtained in cows (12). The influence of A selected review. Small Rumin Res 2010;92:10–18. storage on the electrophoretic distribution of globulin fractions in 14. Anon 1986. Directive 86/609/EEC of 24 November 1986 on the mammalian species of veterinary interest is still little documented. approximation of laws, regulations and administrative provi- In conclusion, storage conditions appear to have a statistically sions of the Member States regarding the protection of animals significant effect, directly correlated to temperature and time, on used for experimental and other scientific purposes. Off J Eur total proteins and their fractions in horse serum. Further investiga- Communities L 358:1–29. tions, at different times and temperatures, are necessary to confirm 15. Martin-Rosset W. L’alimentation des chevaux. Paris, France: these findings and to define storage guidelines for equine samples. Institut national de la recherche agronomique, 1990. Reliable information about preanalytic alterations should help vet- 16. Carapeto V, Barrera M, Cinta Mañe R, Zaragoza MC. Serum erinary practitioners handle and store samples appropriately, so as a-globulin fraction in horses is related to changes in the acute to have good sample quality for correct diagnosis. phase proteins. J Equine Vet Sci 2006;26:120–127. 17. Cavalcante PH, Silva AC, Sakamoto S, Soto-Blanco B. Serum References protein fractions in Brazilian-breed donkeys using agarose gel electrophoresis. Turk J Vet Anim Sci 2012;36:9–12. 1. Colahan PT, Merritt AM, Moore JN, Mayhew G. Equine Medicine 18. Mitruka BM, Rawnsley HM. Clinical biochemistry. In: Clinical and Surgery. 5th ed. Vol. II. St. Louis, Missouri: Mosby, 1999: Biochemical and Hematological Reference Values in Normal 987–1989. Experimental Animals. New York, Masson Publishing USA, 1977: 2. Kaneko JJ. Serum proteins and the dysproteinemias. In: 117–245. Kaneko JJ, Harvey JW, Bruss ML, eds. Clinical Biochemistry 19. Atmeh RF, Arafa IM, Al-Khateeb M. Albumin aggregates: of Domestic Animals. 5th ed. San Diego, California: Academic Hydrodynamic shape and physico-chemical properties. Jordan Press, 1997:117–137. J Chem 2007;2:169–182.

296 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Article

Evaluation of cerebrospinal fluid lactate and plasma lactate concentrations in anesthetized dogs with and without intracranial disease Deanne Caines, Melissa Sinclair, Darren Wood, Alexander Valverde, Doris Dyson, Luis Gaitero, Stephanie Nykamp

Abstract The objectives of this study were to establish a reference interval for canine cerebrospinal fluid lactate (CSFL) and to compare CSFL and plasma lactate (PL) concentrations in anesthetized dogs with and without intracranial disease. Using a prospective study, canine blood and cerebrospinal fluid were collected for lactate analysis in 11 dogs with intracranial disease after undergoing magnetic resonance imaging (MRI) (Group ID-MRI), in 10 healthy dogs post-MRI (Group H-MRI), and in 39 healthy dogs after induction of anesthesia (Group H-Sx). Dogs were anesthetized for the procedures using different anesthetic protocols. Neurological scores (NS) and sedation scores (SS) were assessed pre-anesthesia in ID-MRI dogs. The CSFL reference interval [90% confidence interval (CI) for lower and upper limits] was 1.1 (1.0 to 1.2) to 2.0 (2.0 to 2.1) mmol/L. Mean 6 SD CSFL concentrations were: ID-MRI, 2.1 6 0.8; H-MRI, 1.6 6 0.4; and H-Sx, 1.6 6 0.2 mmol/L. There was a tendency for higher CSFL in dogs in the ID-MRI group than in those in the H-MRI or H-Sx groups (P = 0.12). There was agreement between CSFL and PL in ID-MRI dogs (P = 0.007), but not in dogs in H-MRI (P = 0.5) or H-Sx (P = 0.2). Of the ID-MRI dogs, those with worse NS had higher CSFL (r2 = 0.44). The correlation between CSFL and PL in dogs with intracranial disease and between worse NS and higher CSFL warrants further investigation into the use of CSFL and PL for diagnostic and prognostic purposes.

Résumé Les objectifs de la présente étude étaient d’établir un intervalle de référence pour la valeur du lactate du liquide céphalo-rachidien (CSFL) chez le chien et de comparer les concentrations du lactate plasmatique (PL) chez des chiens anesthésiés avec et sans maladie intracrânienne. Lors d’une étude prospective du sang et du liquide céphalo-rachidien ont été prélevés pour analyse du lactate chez 11 chiens avec maladie intracrânienne après un examen d’imagerie par résonnance magnétique (MRI) (Groupe ID-MRI), chez 10 chiens en santé post-MRI (Groupe H-MRI), et chez 39 chiens après induction de l’anesthésie (Groupe H-Sx). Les chiens ont été anesthésiés pour les procédures en utilisant différents protocoles d’anesthésie. Les scores neurologiques (NS) et les scores de sédation (SS) furent évalués pré-anesthésie chez les chiens ID-MRI. L’intervalle de référence pour le CSFL [intervalle de confiance 90 % (CI) pour les limites inférieure et supérieure] était 1,1 (1,0 à 1,2) à 2,0 (2,0 à 2,1) mmol/L. Les concentrations moyennes 6 écart-type du CSFL étaient : ID-MRI, 2,1 6 0,8; H-MRI, 1,6 6 0,4; et H-Sx 1,6 6 0,2 mmol/L. Il y avait tendance à observer des valeurs de CSFL plus élevées chez les chiens du groupe ID-MRI que chez les chiens des groupes H-MRI ou H-Sx (P = 0,12). Il y avait concordance entre le CSFL et le PL chez les chiens ID-MRI (P = 0,007), mais pas chez les chiens H-MRI (P = 0,5) ou H-Sx (P = 0,2). Parmi les chiens ID-MRI, ceux avec les pires NS avaient un CSFL élevé (r2 = 0,44). La corrélation entre le CSFL et le PL chez les chiens avec une maladie intracrânienne et entre le pire NS et un CSFL élevé mérite une investigation plus poussée sur l’utilisation du CSFL et du PL pour fin de diagnostic et de pronostic. (Traduit par Docteur Serge Messier)

Introduction (CSFL) in human medicine provides additional information about a variety of intracranial and metabolic disorders (2–4). Increased Analysis of cerebrospinal fluid (CSF) is a basic diagnostic tool CSFL has been identified in conjunction with several neurological that is often used in veterinary patients with neurological disease disorders, including traumatic brain injury, inflammatory disease, and is the primary method of diagnosing central nervous system neoplasia, and epilepsy (4,5) and has frequently been investigated (CNS) inflammatory disease (1). Potential information gathered for its potential ability to distinguish between bacterial and aseptic from evaluation of this fluid includes white blood cell count and meningitis in humans (3,6,7). In cats with experimental brain injury, differential, total protein, and serologic testing and culture. The CSFL concentrations were found to increase in proportion to the measurement of the concentration of cerebrospinal fluid lactate severity of brain injury (5).

Department of Clinical Studies (Caines, Sinclair, Valverde, Dyson, Gaitero, Nykamp) and Department of Pathobiology (Wood), Ontario Veterinary College, University of Guelph, Guelph, Ontario N1G 2W1. Address all correspondence to Dr. Deanne Caines; telephone: (416) 247-8387; fax: (416) 287-3642; e-mail:[email protected] This paper was presented at the 18th International Veterinary Emergency and Critical Care Symposium and Annual Conference of the American College of Veterinary Anesthesiologists in San Antonio, Texas, 2012. Received August 30, 2012. Accepted October 29, 2012.

2013;77:297–302 The Canadian Journal of Veterinary Research 297 FOR PERSONAL USE ONLY

Lactate is a by-product of metabolism within all body tissues. It Materials and methods is produced in anerobic conditions and can be increased in cases of decreased oxygen delivery or increased oxygen demand, which frequently occur in critically ill patients (8). In humans, the basal Animals lactate production rate has been determined to be 0.8 mmol/kg Three groups of dogs were studied: 11 client-owned dogs with body weight (BW) per hour (8), which results in normal arterial pre-existing intracranial disease that were scheduled for anes- blood lactate values of approximately 0.6 mmol/L and slightly thesia to facilitate a diagnostic MRI and CSF collection (ID-MRI); higher venous lactate values of 1.0 mmol/L (9,10). In dogs, lactate 10 healthy dogs that were anesthetized for MRI as part of another production has been estimated at 1.4 to 2.4 mmol/kg BW per hour, study (H-MRI); and 39 healthy dogs that were scheduled for surgical depending on the method of determination (11), and the accepted exercises for undergraduate veterinary students (H-Sx). The latter upper reference limit for plasma lactate (PL) (arterial or venous) is 2 groups were used as study controls and to establish a reference 2.5 mmol/L (12,13). interval. All dogs underwent a full physical examination before Lactate found in CSF can be an important indicator of cerebral general anesthesia. Pre-anesthetic stabilization of the neurological ischemia. As lactic acid is fully ionized at physiologic pH, the trans- disease for dogs in the ID-MRI group was conducted as dictated fer of plasma lactate across an intact blood brain barrier (BBB) is by the patient’s clinical signs, and at times included the adminis- limited (14). As such, CSFL concentrations are relatively independent tration of steroids, opioid analgesics, or hyperosmolar agents. All of those found in plasma of healthy individuals (14,15). In several procedures were approved by the Animal Care Committee of the pathologic CNS conditions, however, the integrity of the BBB is University of Guelph and followed the guidelines established by the disrupted (16–18) and CSFL concentrations are influenced by lactate Canadian Council of Animal Care. For client-owned dogs, informed concentrations in the blood. owner consent was obtained before anesthesia. The potential for CSFL concentration as a prognostic indica- tor in veterinary medicine is intriguing. Although blood lactate Anesthesia and instrumentation concentration is routinely measured in veterinary medicine, CSFL Dogs in groups ID-MRI and H-MRI were fasted for approximately measurement is not assessed as often as it is in human medicine. 6 to 12 h before anesthesia. A cephalic venous catheter (Insyte-W; However, interest is growing in the significance of increases in Becton Dickinson, Sandy, Utah, USA) was placed without pre- CSFL and PL in dogs with neurological disease (19,20). Sullivan medication and propofol (Diprivan 1%; AstraZeneca, Mississauga, et al (19) retrospectively evaluated blood lactate concentrations in Ontario), 4 to 6 mg/kg BW was administered by IV to effect for dogs with intracranial disease compared to dogs with intervertebral induction of anesthesia. All dogs were intubated and mechani- disk disease (IVDD) and found that dogs with meningioma and cally ventilated (10 to 20 mL/kg BW tidal volume) (Multiflow 2002 hydrocephalus had significantly higher blood lactate concentra- Anesthesia Ventilator; Hallowell EMC, Pittsfield, Massachusetts, tions, although only dogs with meningioma had lactate concentra- USA) with 100% oxygen to an end-tidal carbon dioxide (ETCO2) of tions that were outside of the reference range. Witsberger et al (20) 30 to 35 mmHg immediately after induction. Anesthesia in ID-MRI found no correlation between CSFL concentrations in dogs with dogs was randomly maintained with either a continuous infusion IVDD and long-term outcome, whereas in dogs with age-related of propofol (267 to 344 mg/kg BW per minute) or with isoflurane cognitive dysfunction, CSFL was significantly higher in dogs with (IsoFlo; Abbott Animal Health, Abbott Park, Illinois, USA) at an severe cognitive dysfunction than in unaffected and mildly affected end-tidal isoflurane (ETIso) of 0.66% to 1.17%. dogs (21). Anesthesia for H-MRI dogs was maintained with isoflurane only. Dogs with intracranial disease may be considered high-risk anes- Due to requirements in minimizing artifact on MRI scans in the thetic candidates due to the possibility of increases in intracranial H-MRI group undergoing the concurrent study, atracurium (Sandoz pressure, which could be exacerbated by the effects of some anes- Canada, Boucherville, Quebec) was administered by IV to those dogs thetic drugs, leading to cerebral ischemia or fatal brain herniation. at a mean dose 6 SD of 0.13 6 0.02 mg/kg BW to provide neuromus- Due to the added cost and risk of general anesthesia when analyzing cular blockade. Rectal or esophageal temperature, ECG, and lithium CSF in these patients, establishing a correlation between CSFL and dilution cardiac output were assessed before and after the MRI due PL concentrations could improve the accuracy of diagnostic evalu- to lack of MRI-compatible equipment for these parameters. Pulse ation of canine patients with intracranial disease. oximetry, ETCO2, ETIso, heart rate, and systolic, mean, and diastolic The objective of this study was to evaluate PL and CSFL concen- arterial pressure were measured continuously before, during, and trations in anesthetized dogs with and without intracranial disease. after the MRI and recorded at 5 to 15 min intervals. Once MRI was The goals were 3-fold: i) to establish a reference interval for CSFL in completed, cisternal CSF collection and venipuncture were carried healthy anesthetized dogs; ii) to evaluate whether correlation exists out for lactate determinations before recovery from anesthesia. between CSFL concentrations and neurological status; and iii) to Food was withheld for 12 h before anesthesia for dogs in the investigate the correlation between CSFL and PL concentrations in H-Sx group, while water was provided ad libitum. Premedication dogs with intracranial disease undergoing magnetic resonance imag- with 0.05 mg/kg BW of acepromazine (Atravet; Wyeth Animal ing (MRI) and healthy dogs, with or without an MRI. Our hypotheses Health, Guelph, Ontario) and 0.05 mg/kg BW of hydromorphone were that CSFL concentration would correlate with PL concentration (Sandoz Canada) by IM was followed by cephalic vein catheter- and that CSFL concentration would be higher in dogs with evidence ization and induction of anesthesia with either 4 to 6 mg/kg BW of intracranial disease. of propofol (Diprivan 1%; AstraZeneca) or 5 to 10 mg/kg BW of

298 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Table I. Summary statistics for 3 groups of dogs that underwent cerebrospinal fluid (CSF) and blood collection for plasma lactate (PL) measurement under general anesthesia Weight CSFL (mmol/L) PL (mmol/L) R2 Bias for Precision for Group N-CSFL N-PL (kg) concentration concentration (P-value) rhoc PL-CSFL PL-CSFL ID-MRI 11 10 25.3 6 14.4 2.1 6 0.8 1.7 6 1.0 0.7 (0.007) 0.8 20.3 20.6 to 0.03 H-MRI 10 4 15.3 6 10.4 1.6 6 0.4 1.5 6 0.2 20.06 (0.5) 0.3 0.2 20.2 to 0.6 H-Sx 39 10 14.6 6 10.9 1.6 6 0.2 1.6 6 0.5 0.1 (.2) 0.3 0.09 20.1 to 0.3 Data presented as mean 6 standard deviation (SD). CSFL — Cerebral spinal fluid lactate; PL — plasma lactate; N-CSFL — number of CSFL samples; N-PL — number of PL samples; Wt — weight, CSFL — cerebrospinal fluid lactate concentration, PL — plasma lactate concentration, R2 — adjusted R squared for linear regression; rhoc — Concordance correlation. thiopental (Vétoquinol Canada, Lavaltrie, Quebec) by IV to effect. Results Dogs were intubated and maintained on isoflurane in 100% oxygen via spontaneous ventilation. Immediately after induction, once Group ID-MRI consisted of 11 dogs with a mean 6 SD weight of dogs were deemed stable and at an appropriate anesthetic depth, 25.3 6 14.4 kg (Table I). This group included 3 Labrador retrievers, cisternal CSF collection and venipuncture were carried out for lactate 2 soft-coated wheaton terriers, and 1 each of the following breeds: determinations. Shih tzu, mixed-breed, Border collie, Newfoundland, Pomeranian, All CSF samples were taken from the cerebellomedullary cistern and German shepherd. Six dogs (2 males, 4 females) received propo- with the patient in lateral recumbency, following a standard tech- fol maintenance anesthesia and 5 dogs (3 males, 2 females) received nique (22). Whole blood was collected from a portion of study dogs isoflurane maintenance. Disease distribution within this group was (ID-MRI: 10 dogs, H-MRI: 4 dogs, H-Sx: 10 dogs). Blood was collected as follows: 2 with intracranial masses; 4 with inflammatory disease; from either a peripheral vein or from an arterial catheter previously and 1 each of the following diagnoses: parasitic migration; diffuse inserted in the dorsal pedal artery and placed in sodium heparin lesion in the brain stem/thalamocortex; cerebrovascular accident; tubes (Vacutainer; Becton Dickinson, Franklin Lakes, New Jersey, focal edema; and normal MRI/CSF but with clinical signs strongly USA). Blood was immediately centrifuged with a Dynac II centrifuge suggesting intracranial disease. Mean 6 SD pre-anesthesia SS and at 2000 rpm for 8 min and plasma was removed. All samples were NS scores were 1.9 6 0.9 and 11.3 6 2.5, respectively. One of these placed in cryogenic vials (Corning Incorporated, Corning, New York, dogs required administration of 0.25 mg/kg BW of dexamethasone USA) and frozen at 280°C until analyzed approximately 2 to 6 mo (Dexamethasone 2; Vétoquinol Canada) during the anesthetic period after collection. To our knowledge, the stability of lactate with long- to treat its disease. term storage at 280°C has not been evaluated. However, this method Group H-MRI consisted of 10 dogs (3 males, 7 females), of collection and storage has been used in a previous study (21). 15.3 6 10.4 kg, which included 8 research beagles, plus 1 Labrador Sedation scores (SS) and neurological status (NS) scores were retriever and 1 golden retriever that were intended for the ID-MRI assigned to dogs in the ID-MRI group pre-anesthesia using simple group, but were excluded due to normal MRI and CSF analysis and semiquantitative descriptive scales (Appendices 1 and 2). inadequate clinical signs of neurological disease. Group H-Sx consisted of 39 dogs, 14.6 6 10.9 kg, which included Lactate analysis 10 hounds, 3 mixed breed dogs, and 26 beagles. Genders were not Lactate was analyzed with an L-lactate oxidase-based colorimetric recorded. assay (Cobas 6000 c501; Roche, West Sussex, UK). The measurement The time that elapsed between induction and CSF collection was range for this assay is 0.2 to 15.5 mmol/L. The reference interval 2.2 6 0.25 h for ID-MRI and 2.5 6 0.45 h for H-MRI. Anesthesia was was calculated with a computer software program (Analyse-it, ver- stopped immediately after CSF collection was completed. Cerebral sion 2.24; Analyse-it Software, Leeds, UK). spinal fluid (CSF) was collected for H-Sx dogs within the first 15 min post-induction. Statistical analysis Concentrations of CSF lactate were normally distributed in all Statistical analysis was done with standard computer software groups. The CSFL reference interval [90% confidence interval (CI) (SAS OnlineDoc 9.2; SAS Institute, Cary, North Carolina, USA). for lower and upper limits] established from H-MRI and H-Sx Normality of CSFL concentration distribution was tested with dogs was 1.1 (1.0 to 1.2) to 2.0 (2.0 to 2.1) mmol/L. Mean 6 SD Shapiro-Wilk. Difference in means between groups was tested CSFL concentrations were 2.1 6 0.8 mmol/L in the ID-MRI group, with analysis of variance (ANOVA). Agreement between CSFL and 1.6 6 0.4 mmol/L in the H-MRI group, and 1.6 6 0.2 mmol/L in the PL concentrations for each group was tested with concordance cor- H-Sx group. These were not significantly different among the groups relation and by determining bias and precision using Bland-Altman (P = 0.12), although dogs with intracranial disease (ID-MRI) tended plots and paired t-test. Regression analysis was used to test for linear to have higher CSFL concentrations (Table I). There was greater relationships between CSFL and PL. Pearson correlation was used variability in the CSFL values in the ID-MRI group than in the H-Sx to evaluate correlation between CSFL concentration and NS and SS. group (range: 1.0 to 3.4 mmol/L, SD = 0.84 versus 1.1 to 2.1 mmol/L, Significance was set at P , 0.05. SD = 0.19, respectively). There was agreement between CSFL and

2000;64:0–00 The Canadian Journal of Veterinary Research 299 FOR PERSONAL USE ONLY

PL concentrations in ID-MRI dogs (P = 0.007), but not in the other dogs undergoing an abdominal MRI eliminated the potential for the groups (Table I). The measures of concordance correlation were 0.8, MRI itself to alter CSFL concentrations. Magnetic resonance imag- 0.3, and 0.3 for the ID-MRI, H-MRI, and H-Sx groups, respectively. ing (MRI) is known to have physiologic effects on some patients, No significant bias was detected for either group (P = 0.07, 0.2, and e.g., hyperthermia due to absorption by the body of radiofrequency 0.4 for the ID-MRI, H-MRI, and H-Sx groups, respectively). radiation has been documented in children (26–28). It is possible Mean (SD) neurologic score for the ID-MRI group was 11.3 (2.5). that a potential increase in body temperature could influence CSFL Dogs in the ID-MRI group with worse NS had significantly higher concentrations. Hyperthermia was not noted in any of the ID-MRI CSFL concentrations (r2 = 0.44, P = 0.04). No correlation was found dogs and conversely, temperatures tended to decrease. Furthermore, between CSFL concentration and sedation score (r2 = 0.09). our finding that CSFL concentrations did not differ between dogs in the H-MRI group and those in the H-Sx group decreases the likeli- Discussion hood of that possibility. There was greater variability in the CSFL values in the ID-MRI The first objective of this study was to establish a reference than in the H-Sx group. This is likely due to the range of diseases and interval for canine CSFL in healthy dogs. The reference interval severity of neurological status of some dogs in the ID-MRI group. established from our data was from 49 dogs (groups H-MRI and Although CSFL concentrations did not differ significantly among H-Sx) and was 1.1 to 2.0 mmol/L, which is comparable to reference groups, ID-MRI dogs had a wider range of values and tended to values of 1.8 6 0.06 mmol/L for CSFL reported in human medicine have higher mean values. This degree of variability reflects the range (23). Although guidelines suggest that . 120 would be an ideal of disease diagnoses and could also have been influenced by drugs number of samples to attain statistical significance, it is generally (systemic corticosteroids, hyperosmolar agents, and opioid analge- accepted that this is often not feasible in veterinary medicine (24). sics) administered in the pre-anesthetic period. As dexamethasone Our number of samples was considered acceptable for the reference has been shown to decrease CSFL concentrations in rabbits (29), it interval established. is possible that CSFL concentrations could also have been affected Anesthetic management varied in the 3 groups in accordance with in some dogs in the ID-MRI group that received dexamethasone as concurrent research projects. The dogs in ID-MRI were maintained part of their pre-anesthetic management. The number of dogs in the on either isoflurane (5 dogs) or propofol (6 dogs) for the duration of ID-MRI group was small and it is possible that the lack of signifi- their anesthesia. This differed from the other 2 groups, which were cance in the difference between CSFL concentrations among groups maintained only on isoflurane. Similarly, the dogs in H-Sx were pre- is due to lack of power. It is also interesting to note that, although medicated with acepromazine and hydromorphone, while dogs in the mean CSFL concentration for the ID-MRI group (2.1 mmol/L) the other 2 groups were not premedicated. The degree to which this was not statistically different from the control groups, it exceeded may have affected our results is uncertain. Horn and Klein (25) found the upper limit of the reference interval (2.0 mmol/L) established in that anesthesia with 2% isoflurane increased CSFL concentrations in our study using the Cobas instrument and L-lactate oxidase method. mice by 118%, while propofol had no effect on CSFL. Similar studies Dogs in the H-MRI group received atracurium to reduce motion that compare anesthetic protocols and their effects on CSFL are not artifact on abdominal MRI images. Atracurium does not appear to available for dogs. If the effects in dogs are similar to those in mice, affect cerebral hemodynamics (30) and should not have had any one might surmise that the dogs in this study that were maintained impact on the results. on isoflurane might have higher CSFL concentrations than those The third objective of this study was to evaluate whether there is maintained on propofol. This may have narrowed the gap between agreement between CSFL and PL concentrations in healthy dogs and CSFL concentrations in ID-MRI and the control groups. This variation also in dogs with intracranial disease. Blood samples were collected in anesthetic protocol reflects the actual clinical scenario, in which from only 4 dogs in the H-MRI group and only 10 dogs in the H-Sx propofol and isoflurane are both acceptable as maintenance regimes. group, which weakened the comparison between CSFL and PL for Another factor to consider in our study design is the variation these groups. Despite the low number of paired samples, a very good in sample times among groups. Dogs in the ID-MRI group were correlation was found between CSFL and PL concentrations for dogs sampled after the brain MRI at approximately 2.2 h after the start in the ID-MRI group, but not for those in the H-MRI or H-Sx groups. of anesthesia, whereas dogs in the H-MRI group were sampled Due to the limited power of our study and the multiple factors that after an abdominal MRI at approximately 2.5 h after the start of may affect lactate concentrations, this data should be interpreted anesthesia and samples were collected from dogs in the H-Sx group cautiously. The ability to infer a CSFL concentration from plasma immediately after induction, once anesthetic depth under isoflurane concentrations in diseased patients is an attractive prospect, how- was adequate for the procedure. It is unlikely that this variation in ever, as venipuncture is much less invasive than collecting CSF. It has sample time among groups affected results, since it has been shown been documented that severity of brain injury influences both CSFL that CSFL concentrations rose immediately after isoflurane was and PL concentrations in cats (5). Given the tendency for dogs with administered (25) and in our study CSFL concentrations were very worse NS scores in this study to have higher CSFL concentrations, it similar in both control groups. is possible that, in the absence of other diagnostics, PL may be used The second objective of this study was to evaluate the relation- to predict severity of disease while limiting invasiveness. Due to the ship between CSFL concentrations and intracranial disease. Dogs multiple factors that can affect PL levels as a result of hypoxia and in the ID-MRI group had naturally occurring intracranial disease hypoperfusion, however, it is important that a thorough physical and its comparison to the H-MRI group that consisted of healthy exam and diagnostics are completed in each patient.

300 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

One possible factor of this study that may have affected our results between these values was analyzed, however, it was considered is that blood was collected from either a lateral saphenous vein or insufficient to allow prediction of arterial hyperlactetemia in the a dorsal pedal artery catheter. Evidence regarding the agreement face of elevated venous lactate (33). between arterial and peripheral venous lactate concentrations is In conclusion, this study established a reference interval for canine conflicting (9,10,29–33). While some studies suggest that blood from CSFL concentration in healthy anesthetized dogs. Concentrations of these sample sites can be used interchangeably (13,31,32), others CSFL were significantly higher in clinical canine cases with worse claim that the discrepancy is too great (9,10,33). When Hughes and NS scores using the methodology in our study. Further investigation Drobatz (13) evaluated lactate concentrations from various venous is warranted into the evaluation of CSFL concentrations in dogs and and arterial venipuncture sites in dogs, they concluded that the the potential to infer prognostic information in clinical cases with discrepancies between the sample sites were clinically insignificant. intracranial disease or trauma. Gallagher et al (33), however, advise caution in the substitution of arterial and venous lactate. In their study, which measured arterial Acknowledgments and venous lactate concentrations in human patients, the differences between the mean values for each group were determined to be The authors thank Gabrielle Monteith and William Sears for sta- not clinically significant (mean arterial lactate and venous lactate tistical assistance. This work was funded in part by a grant from the were 2.8 mmol/L and 3.0 mmol/L, respectively). When agreement Ontario Veterinary College Pet Trust.

Appendix 1. Neurological scoring system used pre-anesthesia Appendix 2. Simple descriptive scale for pre-anesthetic for dogs with intracranial disease anesthetized for diagnostic sedation for dogs with intracranial disease anesthetized with magnetic resonance imaging isoflurane or propofol undergoing anesthesia for magnetic resonance imaging Parameter Score Mentation/Response to stimuli (circle one): Pre-anesthesia sedation • BAR 13 0 Excitable with no apparent sedation or depression (excited, • Depressed 12 ambulatory, difficult to restrain for catheter placement, very • Stuporous 11 interactive) • Coma 0 1 Calm with no apparent sedation or depression (alert, ambulatory, mild to moderate resistance to restraint for Menace Response (regardless of the side) catheter placement, moderately interactive) • Present 12 2 Apparent sedation/depression (quiet, ambulatory with • Decreased 11 difficulty/ataxia, minimal restraint required for catheter • Absent 0 placement, mild response to voice or touch) Pupil Size (regardless of the side) 3 Profound sedation/depression (dull, does not respond to voice • Normal 12 or touch, non-ambulatory, able to place IV catheter without • Dilated/Constricted 11 resistance) Pupil Symmetry (regardless of the side) • Symmetrical 11 • Asymmetrical 0 References PLR (regardless of the side) 1. Tipold A. Diagnosis of inflammatory and infectious diseases of • Normal 12 the central nervous system in dogs: A retrospective study. J Vet • Decreased 11 Int Med 1995;9:304–314. • Absent 0 2. Hutchesson A, Preece MA, Gray G, Green A. Measurement of lactate in cerebrospinal fluid in investigation of inherited meta- Nystagmus bolic disease. Clin Chem 1997;43:158–161. • Physiological 12 3. Cuhna BA. The diagnostic usefulness of cerebrospinal fluid lactic • Pathological 11 acid levels in central nervous system infections. Clin Infect Dis (horizontal, rotary, vertical, changing in phase) 2004;39:1250–1251. • Absent 0 4. Chow SL, Rooney ZJ, Cleary MA, Clayton PT, Leonard JV. The Strabismus (regardless of the side) significance of elevated CSF lactate. Arch Dis Child 2005;90: • Absent (Normal) 11 1188–1189. • Present (Abnormal) 0 5. Inao S, Marmarou A, Clarke G, Andersen B, Faouros PP, Young HF. Production and clearance of lactate from brain tissue, cere- Nasal Septum Stimulation (regardless of the side) brospinal fluid, and serum following experimental brain injury. • Present 12 J Neurosurg 1988;69:736–744. • Decreased/Unilateral 11 6. Lieb SL, Boscacci R, Gratzl O, Zimmerli W. Predictive value of • Absent 10 cerebrospinal (CSF) lactate level versus CSF/blood glucose ratio

2000;64:0–00 The Canadian Journal of Veterinary Research 301 FOR PERSONAL USE ONLY

for the diagnosis of bacterial meningitis following neurosurgery. 21. Pugliese M, Carrasco JL, Andrade C, Mas E, Mascort J, Mahy N. Clin Infect Dis 1999;29:69–74. Severe cognitive impairment correlates with higher cerebrospinal 7. Quagliarello V, Scheld WM. Bacterial meningitis: Pathogenesis, fluid levels of lactate and pyruvate in a canine model of senile pathophysiology, and progress. New Engl J Med 1992;327: dementia. Prog Neuro-Psychoph 2005;29:603–610. 864–872. 22. Terlizzi RD, Simon RP. The function, composition and analysis 8. Fall PJ, Szerlip HM. Lactic acidosis: From sour milk to septic of cerebrospinal fluid in companion animals: Part II — Analysis. shock. J Intensive Care Med 2005;20:255–271. Vet J 2009;180:15–32. 9. Huckabee WE. Abnormal resting blood lactate. I. The signifi- 23. Genton B, Berger JP. Cerebrospinal fluid lactate in 78 cases of cance of hyperlactatemia in hospitalized patients. Am J Med adult meningitis. Intens Care Med 1989;16:196–200. 1961;30:833–839. 24. ASVCP Quality Assurance and Laboratory Standards Committee 10. Huckabee WE. Abnormal resting blood lactate. II. Lactic acidosis. (QALS) Guidelines for the Determination of Reference Intervals Am J Med 1961;30:840–848. in Veterinary Species and Other Related Topics: SCOPE. 11. Forbath N, Kenshole AB, Hetenyi G. Turnover of lactic acid in American Society for Veterinary Clinical Pathology [homepage normal and diabetic dogs calculated by two tracer methods. Am on the Internet]. 2011. Available from: http://www.asvcp.org/ J Physiol 1967;212:1179–1184. pubs/qas/index.cfm Last accessed June 10, 2013. 12. Hughes D. Lactate measurement: Diagnostic, therapeutic, and 25. Horn T, Klein J. Lactate levels in the brain are elevated upon prognostic implications. In: Bonagura JD, ed. Kirk’s Current exposure to volatile anesthetics: A microdialysis study. Veterinary Therapy: Small Animal Practice. Vol XIII. Philadelphia, Neurochem Int 2010;57:940–947. Pennsylvania: WB Saunders, 1999:112–116. 26. Machata AM, Willschke H, Kabon B, Prayer D, Marhofer P. Effect 13. Hughes D, Drobatz KJ. Comparison of plasma lactate concentra- of brain magnetic resonance imaging on body core temperature tions from cephalic, jugular, and femoral arterial blood samples in sedated infants and children. Br J Anaesth 2009;102:385–389. in normal dogs. J Vet Emerg Crit Care 1996;6:115. 27. Bryan YF, Templeton TW, Nick TG, Szafran M, Tung A. Brain 14. Posner JB, Plum F. Independence of blood and cerebral spinal magnetic resonance imaging increases core body temperature in fluid lactate. Arch Neurol 1967;16:492–496. sedated children. Anesth Analg 2006;102:1674–1679. 15. Pang DS, Boysen S. Lactate in veterinary critical care: Pathophys­ 28. Kussman BD, Mulkern RV, Holzman RS. Iatrogenic hyperther- iology and management. J Am Anim Hosp Assoc 2007;43: mia during cardiac magnetic resonance imaging. Anesth Analg 270–279. 2004;99:1053–1055. 16. Zhu XP, Li KL, Kamaly-Asl ID, et al. Quantification of endo- 29. Täuber MG, Khayam-Bashi H, Sande MA. Effects of ampicillin thelial permeability, leakage space, and blood volume in brain and corticosteroids on brain water content, cerebrospinal fluid tumors using combined T1 and T2 contrast-enhanced dynamic pressure, and cerebrospinal fluid lactate levels in experimental MR imaging. J Magn Reson Imaging 2000;11:575–585. pneumococcal meningitis. J Infect Dis 1985;151:528–534. 17. Turgut M, Erdogan S, Ergin K, Serter M. Melatonin ameliorates 30. Giffin JP, Litwak B, Cottrell JE, Hartung JH, Capuano C. blood-brain barrier permeability, glutathione, and nitric oxide Intracranial pressure, mean arterial pressure, and heart rate levels in the choroid plexus of the infantile rats with kaolin- after rapid paralysis with atracurium in cats. Can Anaesth Soc induced hydrocephalus. Brain Res 2007:117–125. J 1985;32:618–621. 18. Lobetti R, Pearson J. Magnetic resonance imaging in the diag- 31. Younger JG, Falk JL, Rothrock SG. Relationship between arte- nosis of focal granulomatous meningoencephalitis in two dogs. rial and peripheral venous lactate levels. Acad Emerg Med Vet Radiol Ultrasoun 1996;37:424–427. 1996;3:730–734. 19. Sullivan LA, Campbell VL, Klopp LS, Rao S. Blood lactate con- 32. Valverde A, Gianotti G, Rioja-Garcia E, Hathway A. Effects of centrations in anesthetized dogs with intracranial disease. J Vet high-volume, rapid-fluid therapy on cardiovascular function and Intern Med 2009;23:488–492. hematological values during isoflurane-induced hypotension in 20. Witsberger TH, Levine JM, Fosgate GT, et al. Associations healthy dogs. Can J Vet Res 2012;76:99–108. between cerebrospinal fluid biomarkers and long-term neuro- 33. Gallagher EJ, Rodriguez K, Touger M. Agreement between logical outcome in dogs with acute intervertebral disk hernia- peripheral venous and arterial lactate levels. Ann Emerg Med tion. J Am Vet Med Assoc 2012;240:555–562. 1997;29:479–483.

302 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Article

Sandwich-dot enzyme-linked immunosorbent assay for the detection of canine distemper virus Zhi Li, Yanlong Zhang, Huiguo Wang, Jinhua Jin, Wenzhe Li

Abstract A sandwich-dot enzyme-linked immunosorbent assay (dot ELISA) was developed for the detection of canine distemper virus (CDV). In 56 dogs suspected to have CD the rates of detection of CDV antigen in samples of blood lymphocytes and palpebral conjunctiva by dot ELISA and ELISA were, respectively, 91% (49/54) and 81% (44/54) for the lymphocyte samples and 88% (28/32) and 75% (24/32) for the conjunctival samples. The CDV detection limits were 10 ng/50 mL for dot ELISA and 40 ng/50 mL for ELISA. The reliability of dot ELISA relative to electron microscopy was 96% with 22 samples: all 21 samples in which CDV particles were observed by electron microscopy yielded positive results with dot ELISA; the single sample in which particles were not observed yielded false-positive results with dot ELISA. The results indicate that the dot ELISA developed can serve as a reliable rapid diagnostic test in suspected cases of CD and also be useful for epidemiologic surveillance of the disease.

Résumé Une épreuve immuno-enzymatique sandwich par point (dot ELISA) a été développée afin de détecter le virus du distemper canin (CDV). Chez 56 chiens suspectés d’avoir le CD, les taux de détection d’antigène du CDV dans des échantillons de lymphocytes sanguins et de la conjonctive palpébrale par dot ELISA et ELISA étaient, respectivement, 91 % (49/54) et 81 % (44/54) pour les échantillons de lymphocytes et 88 % (28/32) et 75 % (75/32) pour les échantillons de conjonctive. Les limites de détection de CDV étaient 10 ng/50 mL pour le dot ELISA et 40 ng/50 mL pour l’ELISA. La fiabilité du dot ELISA relativement au microscope électronique était de 96 % avec 22 échantillons : les 21 échantillons à partir desquels des particules de CDV furent observées ont donné des résultats positifs au dot ELISA; le seul échantillon à partir duquel aucune particule ne fut observée a donné un résultat faussement positif au dot ELISA. Les résultats indiquent que l’épreuve dot ELISA développée peut servir en tant que test diagnostique rapide et fiable lors de cas suspectés de CD et peut également être utile pour la surveillance épidémiologique de la maladie. (Traduit par Docteur Serge Messier)

Introduction Furthermore, many puppies with maternal antibodies and vac- cinated dogs may have high titers of neutralizing antibodies (5). Canine distemper (CD) is a highly contagious disease that affects Therefore, detection of the neutralizing antibodies is not fully dogs of all ages. It has high morbidity and mortality rates and occurs reliable for a diagnosis of CD, and more attention has been paid worldwide. Canine distemper virus (CDV), a member of the family to detecting CDV antigens. The most reliable method of detecting Paramyxoviridae, genus Morbillivirus (1), causes acute generalized CDV in infected dogs is virus isolation (9); however, the method is infection or chronic localized and persistent infection in the central time-consuming and frequently unsuccessful when the infection nervous system (2). Infected dogs have either the catarrhal form of is not in an acute stage (8). Other laboratory tools, such as stain- distemper or epileptiform convulsions in the initial stages of the ing for inclusion bodies (10) and fluorescent antibody testing (8), disease. Since the virus shows strong infectivity and the infection also produce a negative result in subacute or chronic cases. An has a high mortality rate, most dog breeders suffer serious economic enzyme-linked immunosorbent assay (ELISA) using protein A and losses with CDV infection (1,3). monoclonal antibody (11), an immunochromatographic assay (12), Several serologic assays for evaluating antibody status have been and an immunocapture ELISA (13) were developed to detect CDV used to confirm clinical CD (4–7). However, only a low antibody in cell cultures and clinical specimens. They all have high specific- response can be detected in the first few weeks after infection (8). ity and sensitivity; however, they require an ELISA reader and have

Clinical Laboratory, Dalian Municipal Central Hospital, 826-Xinan Road, Shahekou District, Dalian city, Liaoning 116033, China (Zhi Li); Department of Wildlife Medicine, College of Wildlife Resources, Northeast Forestry University, 26-Hexing Road, Harbin, Heilongjiang 150040, China (Zhang); Institute of Immunology, College of Life Science and Technology, Dalian University, 10-Xuefu Avenue, Dalian Economical and Technological Development Zone, Liaoning 116622, China (Wang, Jin, Wenzhe Li). Address all correspondence to Dr. Zhi Li or Dr. Wenzhe Li; e-mail: [email protected] or [email protected] Zhi Li and Yanlong Zhang contributed equally to this work. Received April 17, 2012. Accepted November 1, 2012.

2013;77:303–308 The Canadian Journal of Veterinary Research 303 FOR PERSONAL USE ONLY

A B Figure 1. Vero cells grown for 24 h and then inoculated (A) with canine distemper virus (CDV) or not inoculated (B). After 72 h about 90% of the CDV- inoculated Vero cells showed cytopathic effects. Images were acquired with a Leica DM IRB microscope (Leica Microsystems, Heerbrugg, Germany) at a magnification of 3 200. limitations for field studies (13,14). Recently, methods to detect the and medium were then centrifuged at 650 3 g for 20 min to remove CDV nucleocapsid protein gene by means of reverse-transcription cellular debris. The supernatant was filtered through 0.22-mm filters polymerase chain reaction (PCR) (10,15) and real-time PCR (16) have (Millipore Corporation, Bedford, Massachusetts, USA). The partially been developed. These methods can be carried out only in equipped purified virus was placed on a discontinuous sucrose gradient of 30%, laboratories but are highly sensitive and useful. Rapid and sensitive 45%, and 60% (w/v) and centrifuged for 18 h at 54 000 3 g to separate laboratory and field tests for the diagnosis of CDV infection are CDV particles from cellular proteins responsible for cross-reactivity essential for CD control. with the virus. Among the 6 major fractions obtained, infectivity in Sandwich-dot ELISA (dot ELISA) is a sensitive and specific tech- Vero cell culture was greatest for fraction 4, at a 50% tissue culture 5.0 nique for detecting various virus antigens that has wide clinical diag- infective dose (TCID50) of 10 . This fraction was therefore used as nostic applications (9,17,18). The aim of this study was to establish a the antigen for specific antibody production and for the dot ELISA. rapid and sensitive laboratory and field test for the diagnosis of CDV infection. Our results indicate that the monoclonal-antibody-based Preparation of polyclonal IgG against CDV dot ELISA has virtues such as reliability, simplicity of performance, In brief, rabbits were given a subcutaneous injection of 500 mg of and good reproducibility. gradient-purified CDV (fraction 4) Gradient purified CDV antigen of fraction 4 used for preparation of the monoclonal IgG.mixed with an Materials and methods equal volume of complete Freund’s adjuvant (1,3). Two weeks later a booster dose, 500 mg of purified CDV mixed with an equal volume of incomplete Freund’s adjuvant, was injected subcutaneously. Two Cells and virus strains weeks later 500 mg of purified CDV was injected intraperitoneally. Vero cells were seeded into 75-cm2 cell-culture flasks. After Blood was drawn 10 d after the last injection and 33% ammonium being grown for 24 h in modified Eagle’s medium supplemented sulfate added to precipitate the protein. The partially purified IgG with 100 U/mL of penicillin G, 100 mg of streptomycin (GIBCO was dialyzed overnight against 0.02 M phosphate-buffered saline BRL, Carlsbad, California, USA), and 5% fetal bovine serum (Sigma (PBS), pH7.2, and then the IgG was isolated by affinity chromatog-

Chemical Company, St. Louis, Missouri, USA) at 37°C in 5% CO2, raphy with a HiTrap Protein G HP column (GE Healthcare, Fairfield, monolayer cultures that were 80% to 90% confluent were infected Connecticut, USA). with CDV (19). Noninfected Vero cells were used as controls. Yongchun Jin (Yanbian University, Jilin, China) kindly provided Preparation of monoclonal IgG against CDV 20 CDV (YJ-IV) strains, 4 canine parvovirus (CPV) (JIN-C-4) strains, Monoclonal antibodies were produced as described previously 3 infectious canine hepatitis virus (ICHV) (TS-25) strains, and (20,21) with some modifications. Briefly, BALB/C mice (male, 8 wk 3 rabies virus (RV) (RU-34) strains, to be used as indicator viruses old) were injected subcutaneously with the gradient-purified whole in the dot ELISA. CDV antigen of fraction 4. Splenic cells obtained from the mice were fused with SP2/0 myeloma cells with the use of polyethylene glycol Purification of virus antigen 4000. The resulting hybridomas (Figure 2) were screened by ELISA, When approximately 75% of the CDV-inoculated Vero cells showed and those that produced CDV-specific monoclonal antibodies were cytopathic effects (CPE) (Figure 1, bottom panel), 72 h after inocu- subcloned 3 times from single cells by the limiting-dilution method. lation, the cell-associated viruses were harvested by freezing and The purified CDV was coated on 96-well plates and then incubated thawing the cells 3 times in serum-free medium. The resultant cells with 100 mL of the IgG at different dilutions (1:100 to 1:102 400).

304 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

A B

Figure 2. Development of hybridoma from spleen cells obtained from CDV-inoculated BALB/c mice and fused with SP2/0 myeloma cells. A — day 1. B — day 7. The magnification is 3 200.

Optical density was measured at 492 nm with a computer-interfaced strips were cut into the squares and placed in the microwells. The microplate reader (Bio-Rad Laboratories, Hercules, California, USA). remainder of the protocol was the same as for the ELISA described. Of the 12 CDV-specific antibodies screened, 1 antibody, named 9C11, A substrate solution of diaminobenzidine in 0.1 M Tris-HCl buffer, was selected for its strong immunoreactivity against CDV and little pH 7.4, with 0.01% H2O2 was used to color the NC membranes. cross-reactivity with other proteins in immunoblots. The 9C11 cells were inoculated intraperitoneally into pristane-primed BALB/c Determination of analytic specificity and mice, and IgG was isolated from the resulting ascites fluid by affinity sensitivity chromatography with the HiTrap column. The specificity of the dot ELISA was tested with the 20 supplied CDV strains isolated from dog species and the 10 supplied non- The ELISA procedures CDV virus strains (CPV, ICHV, and RV) along with 100 mL of PBS The rabbit polyclonal IgG against CDV was diluted 1:500 in containing the purified CDV strains. Non-CDV viruses were used 50 mM carbonate–bicarbonate buffer, pH 9.6, and coated on 96-well to eliminate false-positive results. No cross-reactivity was observed plates. The plates were left at 4°C overnight, incubated for 60 min in with any of the non-CDV viruses tested. The sensitivity of the dot PBS containing 3% bovine serum albumin (BSA), and then washed ELISA was determined with the YJ-IV CDV strain diluted serially with PBS containing 0.1% Tween-20, pH7.4. A 50-mL aliquot of the from 10 mg/50 mL to 1 ng/50 mL. purified CDV antigen and clinical samples was distributed into each well and the plate incubated at 37°C for 2 h. Each plate had Clinical samples CDV-positive and CDV-negative control antigens. The plates were A Yanbian University animal hospital provided 86 specimens washed and then incubated at 37°C for 1 h with 50 mL of the mouse (54 swabs of palpebral conjunctiva and 32 samples of blood lympho- monoclonal IgG against CDV diluted 1:1000 in PBS containing l% cytes) from 56 dogs suspected to have CD. The dogs demonstrated BSA. After being washed the plates were incubated at 37°C with goat mainly acute and systemic clinical signs, such as fever, lack of appe- IgG against rabbit antigen conjugated with horseradish peroxidase tite, vomiting, diarrhea, and dehydration. Conjunctival epithelial (HRP) diluted 1:5000 in PBS containing 1% BSA. The HRP activity on cells, obtained by vigorous swabbing with a cotton swab, were the immunoplate was detected with the use of O-phenylenediamine frozen and thawed in 1 mL of PBS, the suspensions were centrifuged

(12) and H2O2 as enzyme substrates. Color development was stopped at 2000 3 g for 20 min, and the supernatant was stored at −20°C until with 2 M H2SO4 and the absorbance measured at 492 nm with tested. Blood lymphocytes were separated from 2 mL of peripheral the microplate reader. The results were considered positive if the blood containing heparin (100 units/mL) by centrifugation with absorbance was greater than that of the negative control. Titers Ficoll-Paque separation fluid. The samples were diluted from 1:5 to were expressed as a reciprocal of the highest dilution of the sample 1:80 with PBS for testing by ELISA and dot ELISA. showing a positive signal. The palpebral conjunctival secretions and blood lymphocytes The dot ELlSA was done according to a previously described were inoculated onto Vero cells and incubated at 37°C for 2 d. When method (9) with minor modification. The nitrocellulose (NC) mem- approximately 75% of the monolayer showed CPE the viruses were brane strips were divided into squares 0.3 3 0.3 cm with a hard harvested by freezing and thawing the cells 3 times. Thereafter they lead pencil, and 5-mL aliquots of rabbit polyclonal IgG against CDV were collected on carbon-coated grids by touching the grids against diluted 1:500 in PBS were dotted on separate squares. The strips the samples. The grids were blot-dried by touching the edge of the were allowed to dry, and then the protein-binding sites were blocked grids to a filter paper. All of blood lymphocytes were then stained with a solution of 2% BSA in PBS. After a washing with PBS, the with 2% (w/v) sodium phosphotungstate, pH 7.0, for 2 min and

2000;64:0–00 The Canadian Journal of Veterinary Research 305 FOR PERSONAL USE ONLY

Figure 5. Specificity of sandwich-dot ELISA in detecting purified CDV and the CDV in CDV-infected blood lymphocytes and not cross-reacting with canine parvovirus (CPV) (strain JIN-C-4), canine hepatitis virus (ICHV) (strain TS-25), or rabies virus (RV) (strain RU-34). The experiment was performed twice.

1.3 10 TCID50, respectively (Figure 3). The protein concentration of fraction 4 was 1.36 mg/mL, whereas that of fractions 3 and 5 was 0.33 and 0.43 mg/mL, respectively. The purified CDV virions were morphologically identical to morbilliviruses by electron microscopy, Figure 3. Infective titers of CDV purified by centrifugation on a sucrose gradient of 30%, 45%, and 60% (w/v) for 18 h at 54 000 3 g. Six major and the virus diameter was approximately 200 nm. Fraction 4 was fractions were collected and the 50% tissue culture infective dose therefore used as the antigen for specific antibody production and

(TCID50) determined. for the dot ELISA. The titer of the polyclonal antibody against CDV was measured by ELISA as 1:102 400 (data not shown). Monoclonal antibody 9C11 was selected from among the 12 CDV-specific monoclonal antibodies screened by ELISA because of its strong immunoreactivity against CDV (Figure 4) and little cross-reactivity with other proteins in immunoblots; its titer was 1:51 200, suggesting its high sensitivity for recognizing the CDV protein. Comparison of the results for dot ELISA and ELISA showed that 1:20 dilutions of samples, 1:1000 dilutions of monoclonal IgG against CDV, and 1:2000 dilutions of polyclonal IgG against CDV gave opti- mal results. The CDV detection limits of the dot ELISA and ELISA were 10 ng/50 mL and 40 ng/50 mL, respectively. Moreover, the dot ELISA showed no cross-reactivity with the other viruses: CPV, ICHV, and RV (Figure 5), indicating its high specificity. Among the clinical specimens from 56 dogs suspected to have CD, the rates of detection of CDV antigen by the dot ELISA and Figure 4. Titers of monoclonal anti-CDV IgG (9C11) as detected by the ELISA differed significantly (0.01 , P , 0.05): 91% and 81%, enzyme-linked immunosorbent assay (ELISA). Purified CDV was coated respectively, for the 32 blood-lymphocyte specimens and 88% and on 96-well plates and then incubated with 100 mL of the IgG at different dilutions. Optical density was measured at 492 nm. A positive signal was 75%, respectively, for the conjunctival-swab samples (Table I). Of not detected in serum from unimmunized mice. *P < 0.05; **P < 0.01. the 22 blood-lymphocyte samples that were positive by dot ELISA in which CDV particles were sought by electron microscopy, CDV particles were observed in 21; 17 were positive and 5 negative by observed with a JEOL JEM-1200EX electron microscope (JEOL, ELISA. The particles configurationally resembled morbilliviruses. Peabody, Massachusetts, USA). The overall reliability of dot ELISA relative to electron microscopy was therefore 96%, with only 1 sample yielding false-positive results Statistical analysis by dot ELISA. The results are expressed as means 6 standard deviation (SD). Statistical analyses were carried out using the Student’s t-test. A P-value of less than 0.05 was considered statistically significant and Discussion less than 0.01 extremely significant. Monoclonal antibody testing is a powerful tool for the detection of CDV antigens (22). In recent outbreaks of CD in Yanbian, Jilin, Results China, cross-reaction with the YJ-IV strain of CDV was prominent. To develop a more sensitive test for rapid diagnosis, we referred to In purifying the CDV, sucrose-gradient centrifugation allowed the previous successes in the purification of CDV (14) and used sucrose- separation of CDV particles from cellular proteins that are respon- gradient centrifugation to isolate the fraction with greatest infectivity sible for cross-reactivity with the virus. Of the 6 fractions obtained, in Vero cell culture. We then prepared monoclonal and polyclonal fraction 4 showed the greatest infectivity on Vero cell culture (105.0 antibodies against fraction 4 and used them in a dot ELISA to detect 1.1 TCID50); the infectivity of fractions 3 and 5 was 10 TCID50 and CDV. In this study the dot ELISA had greater sensitivity than the

306 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Table I. Detection by enzyme-linked immunosorbent assay (ELISA) and sandwich- dot ELISA of canine distemper virus in specimens from dogs with suspected infection Dot-ELISA results;b Number of specimens Blood lymphocytes Palpebral conjunctiva (n = 54) (n = 32) Total number Results of ELISAa Positive Negative Positive Negative of specimens Positive 44 0 24 0 68 Negative 5 5 4 4 18 Total 49 5 28 4 86 a Rate of positive results: 81% (44/54) for blood lymphocytes and 75% (24/32) for palpebral conjunctiva. b Rate of positive results: 91% (49/54) for blood lymphocytes and 88% (28/32) for palpebral conjunctiva, both rates significantly higher (0.01 , P , 0.05) than those for the ELISA.

ELISA: the lowest limits of detection of purified CDV antigen were Foundation of China (30972675 and 31270864) and the Science and 10 ng/50 mL and 40 ng/50 mL, respectively. Technology Planning Project of Dalian City, China (2010J21DW011). The porous structure of the NC membrane provides a higher binding capacity than does the solid polystyrene surface of an References ELISA well (9,17,18,23). In addition, no apparent difference in color intensity was observed on NC membranes freshly prepared or stored 1. Appel MJ. Pathogenesis of canine distemper. Am J Vet Res for 6 mo (data not shown). Accordingly, the use of such membranes 1969;30:1167–1182. could greatly facilitate the reproducibility and field applicability of 2. Dorsey WA. Comments on canine distemper outbreak. J Am Vet the dot ELISA (17,18). Med Assoc 2005;226:32. Because CDV has been shown to multiply in the monocytes and 3. Pawar RM, Raj GD, Gopinath VP, et al. Isolation and molecular lymphocytes of the host (1,24) and can be detected with fluorescent characterization of canine distemper virus from India. Trop Anim antibody techniques in the cytoplasm of epithelial cells and neutro- Health Prod 2011;43:1617–1622. phils in conjunctival or genital smears (8), we used ELISA and dot 4. Gemma T, Miyashita N, Shin YS, et al. Serological survey of ELISA to detect CDV antigen in blood lymphocytes and palpebral canine distemper virus infection using enzyme-linked immu- conjunctival secretions from dogs suspected of having CD. Our nosorbent assay. J Vet Med Sci 1995;57:761–763. data showed that lymphocytes are the most suitable clinical speci- 5. Litster A, Nichols J, Volpe A. Prevalence of positive antibody mens from live dogs, the rates of detection of CDV antigen being test results for canine parvovirus (CPV) and canine distemper 91% for the dot ELISA and 81% for the ELISA. From a diagnostic virus (CDV) and response to modified live vaccination against viewpoint, one of the most reliable methods for the diagnosis of CPV and CDV in dogs entering animal shelters. Vet Microbiol CD is isolation of CDV from affected animals or detection of CDV 2012;157:86–90. Epub 2011 Dec 30. antigen in their tissues or cells (11,25). Comparison of the results of 6. Soma T, Ishii H, Hara M, et al. Detection of canine distemper dot ELISA and electron microscopy for detection of CDV yielded a virus antigen in canine serum and its application to diagnosis. relative reliability of 96% for dot ELISA: only 1 of 21 samples had Vet Rec 2003;153:499–501. false-positive results. 7. Waner T, Mazar S, Nachmias E, et al. Evaluation of a dot ELISA In this study, using the YJ-IV CDV strain, we successfully estab- kit for measuring immunoglobulin M antibodies to canine par- lished a monoclonal-antibody-based dot ELISA to detect CDV infec- vovirus and distemper virus. Vet Rec 2003;152:588–591. tion in dogs. This rapid test detected infection in clinical samples 8. Fairchild GA, Wyman M, Donovan EF. Fluorescent antibody from different regions and years. Because the dot ELISA proved technique as a diagnostic test for canine distemper infection: to be nearly as sensitive and specific as electron microscopy while Detection of viral antigen in epithelial tissues of experimentally being simpler and more rapid, it would be an adequate screening infected dogs. Am J Vet Res 1967;28:761–768. test for suspected CDV and useful for epidemiologic surveillance of 9. Abdel-Moneim AS, El-Kady MF, Ladman BS, Gelb J, Jr. S1 gene CD infections in the field. sequence analysis of a nephropathogenic strain of avian infec- tious bronchitis virus in Egypt. Virol J 2006;3:78. Epub 2006 Acknowledgments Sep 20. 10. Shin Y, Mori T, Okita M, et al. Detection of canine distemper We thank Professor Yongkui Jin (Yanbian University, Jilin, China) virus nucleocapsid protein gene in canine peripheral blood for help with the electron microscopic analyses and Professor mononuclear cells by RT-PCR. J Vet Med Sci 1995;57:439–445. Yongchun Jin (Yanbian University) for the clinical samples. This 11. Potgieter LN, Ajidagba PA. Quantitation of canine distemper work was supported by grants of the National Nature Science virus and antibodies by enzyme-linked immunosorbent assays

2000;64:0–00 The Canadian Journal of Veterinary Research 307 FOR PERSONAL USE ONLY

using protein A and monoclonal antibody capture. J Vet Diagn 19. Del Puerto HL, Martins AS, Braz GF, et al. Vero cells infected Invest 1989;1:110–115. with the Lederle strain of canine distemper virus have increased 12. An DJ, Kim TY, Song DS, et al. An immunochromatography Fas receptor signaling expression at 15 h post-infection. Genet assay for rapid antemortem diagnosis of dogs suspected to have Mol Res 2011;10:2527–2533. canine distemper. J Virol Methods 2008;147:244–249. 20. Saliki JT, Lehenbauer TW. Monoclonal antibody-based com- 13. Gemma T, Iwatsuki K, Shin YS, et al. Serological analysis of petitive enzyme-linked immunosorbent assay for detection of canine distemper virus using an immunocapture ELISA. J Vet morbillivirus antibody in marine mammal sera. J Clin Microbiol Med Sci 1996;58:791–794. 2001;39:1877–1881. 14. Bernard SL, Shen DT, Gorham JR. Antigen requirements and 21. Sugiyama M, Minamoto N, Kinjo T, et al. Characterization of specificity of enzyme-linked immunosorbent assay for detection monoclonal antibodies against four structural proteins of rin- of canine IgG against canine distemper viral antigens. Am J Vet derpest virus. J Gen Virol 1989;70(Pt 10):2605–2613. Res 1982;43:2266–2269. 22. Masuda M, Sato H, Kamata H, et al. Characterization of mono- 15. Di Francesco CE, Di Francesco D, Di Martino B, et al. Detection clonal antibodies directed against the canine distemper virus by hemi-nested reverse transcription polymerase chain reac- nucleocapsid protein. Comp Immunol Microbiol Infect Dis tion and genetic characterization of wild type strains of canine 2006;29(2–3):157–165. distemper virus in suspected infected dogs. J Vet Diagn Invest 23. Tsai SJ, Hutchinson LJ, Zarkower A. Comparison of dot immu- 2012;24:107–115. nobinding assay, enzyme-linked immunosorbent assay and 16. Scagliarini A, Dal Pozzo F, Gallina L, et al. TaqMan based real immunodiffusion for serodiagnosis of paratuberculosis. Can J time PCR for the quantification of canine distemper virus. Vet Vet Res 1989;53:405–410. Res Commun 2007;31(Suppl 1):261–263. 24. Krakowka S, Cockerell G, Koestner A. Effects of canine distemper 17. He F, Soejoedono RD, Murtini S, Goutama M, Kwang J. virus infection on lymphoid function in vitro and in vivo. Infect Complementary monoclonal antibody-based dot ELISA for Immun 1975;11:1069–1078. universal detection of H5 avian influenza virus. BMC Microbiol 25. Kai C, Ochikubo F, Okita M, et al. Use of B95a cells for isola- 2010;10:330. Epub 2010 Dec 30. tion of canine distemper virus from clinical cases. J Vet Med Sci 18. Domingues HG, Campalans J, Almeida RS, et al. Dot-enzyme 1993;55:1067–1070. linked immunosorbent assay as an alternative technique for the detection of bovine respiratory syncytial virus (BRSV) antibod- ies.Vet Res 2002;33:397–404.

308 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Short Communication Communication brève

Needle-free delivery of an inactivated avian influenza H5N3 virus vaccine elicits potent antibody responses in chickens Oladele Ogunremi, John Pasick, Yohannes Berhane

Abstract A needle-free delivery system was assessed as a route for providing quick, safe, and effective vaccination against avian influenza (AI). Two groups of chickens were vaccinated with a commercially available inactivated H5N3 virus vaccine delivered either with a needle-free device or with the conventional syringe-and-needle method recommended by the vaccine manufacturer. The kinetic aspects of seroconversion, peak antibody levels, and antibody titers were measured by a combination of an indirect enzyme-linked immunosorbent assay and the hemagglutination-inhibition test and were all found to be similar in the 2 groups of chickens. We conclude that the needle-free delivery system could result in effective immunization against H5N1 AI epidemics and pandemics in chickens.

Résumé Un système d’injection sans aiguille a été évalué comme méthode de vaccination rapide, sécuritaire et efficace contre l’influenza aviaire (AI). Deux groupes de poulets ont été vaccinés avec un vaccin inactivé disponible commercialement contenant le virus H5N3 administré soit au moyen d’un appareil sans aiguille ou de manière conventionnelle avec une seringue et une aiguille tel que recommandé par le manufacturier du vaccin. L’aspect cinétique de la séroconversion, les niveaux maximaux d’anticorps, et les titres d’anticorps ont été mesurés par une combinaison d’épreuve immuno-enzymatique indirecte et le test d’inhibition de l’hémagglutination et ont toutes été trouvés similaires dans les 2 groupes de poulet. Nous concluons que le système d’inoculation sans aiguille entrainerait une immunisation efficace contre H5N1 lors d’épidémies et de pandémie d’AI chez les poulets. (Traduit par Docteur Serge Messier)

The highly pathogenic avian influenza (AI) H5N1 virus remains H5N1 have to be administered parenterally by handling and inject- a threat to public health and to the poultry industry worldwide. ing individual birds, which is time-consuming, labor-intensive, and After introduction of the Eurasian lineage of the virus in geese in potentially cost-prohibitive. This limitation could eventually result the Guangdong province of China in 1996, the virus gained access in low vaccine coverage (9) and a defeat of vaccination as a strategy. to domestic chickens in Hong Kong in 1997 (1–3). To date, over Needle-free delivery technology could overcome some of the chal- 400 million birds have been killed as a result of virus exposure or risk lenges associated with the current syringe-and-needle parenteral of exposure, with an estimated economic loss of $20 billion (4). In delivery procedure. The needle-free device, with the use of metal contrast to the seasonal flu virus strains, which are highly contagious springs, compressed air, or carbon dioxide, forces a vaccine suspen- but cause illness of low mortality in the wider human population, sion at high speed through an orifice about the diameter of a human the H5N1 strain is less contagious but causes illness with a high hair. When the device is held against the skin the ultrafine stream case-fatality rate, ~60% (5). The potential exists, however, that this of high-pressure fluid penetrates the skin, and, depending on the Eurasian H5N1 virus could acquire genetic mutations that alter its pressure setting, the vaccine is delivered at an intradermal, subcu- behavior and result in widespread infection in humans (6). taneous, or intramuscular location (10). Optimal use of a needle-free Vaccination remains the cornerstone of influenza control in delivery system could enable a single operator to deliver a vaccine humans (7), but the procedure is not as widely used for poultry. Less parenterally to up to 3000 birds per day. than 50% of countries that have experienced H5N1 AI outbreaks in Using a commercially available inactivated AI virus vaccine and poultry have developed the requisite legislative authority to use vac- 3-week-old white leghorn chickens, we investigated the use of a cination, and less than 30% have actually used the procedure to any needle-free device to deliver the vaccine safely and to generate a degree as a control strategy (8). It is imperative to examine factors protection-correlating antibody response. A water-in-oil emulsion of preventing the wide adoption of vaccination as a control strategy for inactivated, recombinant H5N3 AI virus suspended in a proprietary the H5N1 virus. Vaccines currently available for protecting against adjuvant mixture (Poulvac Flufend I AI H5N3 RG) was purchased

Ottawa Laboratory Fallowfield, Canadian Food Inspection Agency, 3851 Fallowfield Road, Ottawa, Ontario K2H 8P9 (Ogunremi); National Centre for Foreign Animal Diseases, Canadian Food Inspection Agency, Winnipeg, 1015 Arlington Street, Winnipeg, Manitoba R3E 3M4 (Pasick, Berhane). Address all correspondence to Dr. Oladele Ogunremi; telephone: 343-212-0235, ext. 4844; e-mail: [email protected] Received June 6, 2012. Accepted December 2, 2012.

2013;77:309–313 The Canadian Journal of Veterinary Research 309 FOR PERSONAL USE ONLY

Needle-free immunization ELISA index (%)

Weeks post-immunization

Syringe and needle immunization ELISA index (%)

Weeks post-immunization

Figure 1. Anti-hemagglutinin (H5) antibody responses in chickens vaccinated with inactivated H5N3 virus vaccine against avian influenza as measured by an indirect enzyme-linked immunosorbent assay (ELISA) done on serum samples obtained before and after vaccination. At the start of the experi- ment, 39 birds were vaccinated with a needle-free injector device and 38 birds with a hypodermic syringe and needle. The horizontal lines indicate the cut-off point for positive values. from Fort Dodge Animal Health (Fort Dodge, Iowa, USA). Thirty- Medical Corporation, Somerset, New Jersey, USA) and the syringe nine chickens (group 1) were injected subcutaneously with the vac- attached to the injector, which was set to deliver 0.5 mL of vac- cine, delivered by a needle-free injector device [Agro-Jet; Medical cine at a pressure of 40 psi (276 kPa) into a subcutaneous location International Technologies (MIT Canada), Montreal, Quebec]. The on top of the breast muscle. For comparative purposes 38 birds vaccine preparation was aspirated into a 60-mL syringe (Terumo (group 2) received the same amount of the AI vaccine, delivered

310 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

ating curve, samples with $ 10% ELISA units were considered positive (11). The serum hemagglutination-inhibition (HI) antibody titer was determined (12) in serum samples obtained at week 3 after vaccina- tion to assess whether the delivered vaccine was sufficiently potent. The samples were prescreened for the presence of natural hemag- glutinins, and all were found to be negative. For the titer measure- ment, 4 HA units of A/chicken/Vietnam/14/2005 H5N1 virus was combined with 2-fold serial dilutions of serum beginning with 1 in 2. Chicken erythrocytes (0.5% in 0.1 M phosphate-buffered saline, pH 7.4; v/v suspension) were then added and the HI end-points read. Chicken antiserum to A/duck/British Columbia/26-2/2005 (H5N2) used as a positive control produced end-point titers of 1/80 to 1/160. Negative-control serum did not inhibit the virus-induced hemagglutination (titer , 1:4). The detection of HI antibody titers of Figure 2. Hemagglutination-inhibition (HI) antibody titers in the serum 1:32 or higher in 80% of vaccinated chickens infers sufficient vaccine 3 wk after vaccination. potency according to field (13) and empirical (8,14) observations. Differences in the speed of seroconversion, peak antibody levels, and median HI antibody titers were assessed for significance by use ­subcutaneously with a hypodermic syringe and needle. Two control of a statistical software package (MedCalc Software; Mariakerke, groups of chickens were injected with saline by either the needle-free Belgium). The week-to-week antibody levels and the mean ELISA delivery procedure (group 3, n = 28) or with a syringe and needle values were compared in the needle-free and needle-and-syringe (group 4, n = 29). All the chickens were bled before vaccination groups by computing F-statistics with 1-way analysis of variance (on day 0) and weekly after vaccination for 4 weeks. For the follow- (MedCalc). Differences were considered significant if the P-value ing 2 weeks, only a random selection of birds were bled (n = 5 to 12). was , 0.05. At the end of the experiment (week 16), 19 birds from each group No bleeding or redness was observed at the site of injection in any were bled. Institutional animal-care guidelines were followed in the of the vaccinated chickens, nor were there any clinical signs attribut- maintenance, care, and handling of the chickens. able to systemic impairment during and immediately after vaccina- Immediately after vaccination the injection site on the chicken was tion: general body condition, appetite, and mobility were normal for observed for redness, bleeding, or any other untoward sign. Each each vaccinated bird. All the vaccinated chickens appeared normal bird was observed for 5 min after vaccination and at 1 h, 6 h, or every during vaccination, immediately after vaccination (up to 6 h), and 24 h after vaccination until the end of week 6. Observations were each day afterwards until the end of the experiment. In chickens dis- recorded on general appearance, appetite, mobility, and lameness. sected 24 h after vaccination with the needle-free device, no tissue For evaluation of visible tissue damage caused by the needle-free damage was visible at the site of injection. Remnants of the injected system a separate group of 8 birds vaccinated in this manner were vaccine were visible to the naked eye as whitish streaks, smears, or sacrificed 24 h after injection and dissected to check for early signs deposits at subcutaneous, intermuscular, and sometimes intramus- of necrosis or gross inflammation at the injection site. To determine cular sites during postmortem examinations conducted at week 5 whether the vaccine persisted for any length of time at the subcuta- (in 4 vaccinated birds) and week 6 (in 2 vaccinated birds). No vaccine neous site of needle-free vaccination (e.g., by the time the bird was remnants or any discoloration was visible in the dissected saline- ready for consumption), the vaccinated birds and saline controls injected chickens at week 5 (n = 4) or week 6 (n = 2) after injection. inoculated in this manner were randomly picked and sacrificed at The kinetic aspects of antibody production, as measured by week 5 (4 vaccinated birds, 4 saline controls) and week 6 (2 vac- ELISA, were similar in the 2 groups of vaccinated chickens. The cinated birds, 2 saline controls). The subcutaneous inoculation site speed of seroconversion, measured from the time of vaccination as well as nearby intramuscular and intermuscular locations were to when antibody was first detectable, was similar in the 2 groups: carefully dissected and grossly examined. 2.57 6 0.50 weeks in the needle-free group and 2.46 6 0.56 weeks The immune response of the chickens after vaccine administration in the needle-and-syringe group (P . 0.05). Anti-HA antibody was monitored weekly by measuring the serum antibody response responses were first detected by week 2 after vaccination in 41% (16) to hemagglutinin (HA) with the use of a new indirect enzyme-linked of the 39 birds in group 1 (needle-free) compared with 55% (21) of immunosorbent assay (ELISA), the details of which are the subject the 38 birds in group 2 (syringe and needle) (P . 0.05). By week 3, of a separate communication (11). The assay used a baculovirus- 87% (34/39) of the chickens in group 1 were positive compared with expressed HA protein fragment of A/turkey/ON/773266 H5N9 97% (37/38) of the chickens in group 2. The antibody profiles in the AI virus as the coating antigen and an alkaline-phosphatase con- 2 vaccinated groups of chickens remained similar at 6 wk after vac- jugate of rabbit IgG against chicken IgG heavy and light (Jackson cination (Figure 1). The peak antibody level was 80% 6 19% ELISA ImmunoResearch Laboratories, West Grove, Pennsylvania, USA). units (mean 6 standard deviation) in group 1 and 81% 6 20% ELISA Serum anti-HA antibody levels were expressed as % ELISA units. units in group 2 (P . 0.05). From week 4 to week 6, changes in anti- On the basis of a cut-off value determined from a receiver oper- body levels were not significantly different among chickens in the

2000;64:0–00 The Canadian Journal of Veterinary Research 311 FOR PERSONAL USE ONLY

same group (group 1: F-ratio = 1.15, P = 0.32; group 2: F-ratio = 0.499, the trigger failed to discharge the vaccine, and this may have been P = 0.61). Similarly, no significant differences were found when the the case with the chickens in which antibodies were not detectable test results from each group were pooled and compared with those by ELISA or HI testing. In one other instance, only a partial dose was of the other group (F-ratio = 0.127, P = 0.72). All serum samples delivered, and the antibody response was reduced. obtained from the control chickens injected with saline, regardless We used HI testing to verify the potency of the antibody response of injection system, were negative (0 to 6% ELISA units). because of strong consensus among vaccinologists that HI antibodies Evaluation of the potency of the vaccination procedures at week 3 confer anti-AI protection (18). Vaccination failure with the needle- after vaccination showed that the HI antibody titer was $ 32 in 85% free device was infrequent, occurring in only 2 of 39 birds, and could of the chickens vaccinated with the needle-free delivery system and probably be overcome with more experience and training for better 77% of those vaccinated by syringe and needle. The median HI anti- proficiency with the device. Also, an effort is still required by the body titer for the latter group was, however, higher, at 1:128, than handler in choosing a proper site of administration with the needle- the median of 1:64 for the needle-free group, although the difference free device. Although we did not observe any bleeding among the was statistically insignificant according to the Mann–Whitney test vaccinated chickens, it is still possible that the pressure used in (Z = 0.6194, P = 0.54; Figure 2). delivering the vaccine could inadvertently damage an adjacent blood Many endemically infected countries introduced poultry AI vessel, with fatal consequences in the case of a major vessel. In this vaccination to stem high and unacceptable economic losses and study the lowest pressure setting resulted in the delivery of vaccine human infections. However, the results have not always met expecta- to the subcutis; however, on occasion the vaccine penetrated into the tions: disease transmission has continued during mass vaccination muscle, and vaccine remnants were visible upon dissection of birds campaigns. In Indonesia, large-scale vaccination was eventually 5 to 6 wk after vaccination. Despite the pressurized delivery of the abandoned, although many farms still carry out vaccination (15). vaccine, tissue damage was not visible at the site of injection, prob- In Egypt, where mass culling of birds has been done, large-scale ably because of the small size of the delivery orifice (less than that vaccination has been ineffective because of inappropriate use of of a 25-gauge needle) and the minimal shearing effect on adjoining AI vaccines (15). When vaccination has not fared well, the prob- cells. The similar antibody titers in groups 1 and 2 suggest that tissue lem could not be attributed to the vaccine. Indeed, commercially damage may not have been any more severe in the needle-free group. available vaccines, including the vaccine used in this study, have Our results show that the needle-free device is safe for both chick- been convincingly shown under experimental conditions to protect ens and human handlers, notwithstanding the need to exercise cau- chickens and ducks and to eliminate virus shedding (16,17). Rather, tion in equipment handling and the choice of injection site to avoid failure of or ineffective vaccination is largely due to poor veterinary damage to major blood vessels. The equipment is not accident-prone, infrastructure in the endemically infected countries. Examples of is easy to use, and leads to the development of a potent antibody known problems have included inadequate dosage, improper age response. The speed of delivery of the vaccine is an advantage over of vaccinated poultry, and inadequate coverage (15). current syringe-and-needle methods, which require time-consuming, Adoption of a needle-free device that provides safe, fast, and effi- meticulous handling of individual birds. Although the protective cacious delivery of vaccine against AI could help to improve vaccine immune responses generated by the 2 methods are comparable, we coverage in poultry through optimal use of manpower resources. To conclude that a needle-free device could improve anti-AI vaccination date, at least 10 countries have successfully vaccinated birds against coverage, and the effectiveness of the approach might encourage the the H5N1 virus, and this achievement underscores the benefits to adoption of vaccination as a control strategy in future epidemic or be gained from properly instituted anti-influenza vaccination cam- pandemic H5N1 outbreaks. paigns. The continuing profitability of the poultry industry world- wide is contingent on an overall management system that is able Acknowledgments to keep down the cost of operations, including that of vaccines and vaccination. Therefore, for AI vaccination to be adopted by farmers The authors acknowledge the following staff members of the and producers as a disease control strategy, labor and other costs Canadian Food Inspection Agency (CFIA), Ottawa Laboratory associated with vaccination need to be kept low. A needle-free device Fallowfield: Hilary Kelly, Andrée Ann Dupras, Sebastien Belanger, permits the automated discharge of a preset volume of a vaccine Qigao Fu, Linru Wang, Leith Kealey, Sandra McIntosh, Peter Neave, formulation into a specific location on the target animal, as dictated Diane Ayres (deceased), Doug Gowenlock, Dr. Jim Algire, and Dr. Cyril by the adjustable pressure setting. The current study showed that the Lutze-Wallace. They also acknowledge Dr. Ming Yang and Tim Salo needle-free delivery system was safe and effective in generating a of the CFIA National Centre for Foreign Animal Diseases, Winnipeg. potent antibody response after vaccination of chickens with AI H5N3 vaccine. Other advantages of using a needle-free device include References the elimination of accidental needle sticks for handlers as well as elimination of biohazard waste in large-scale vaccination campaigns. . 1 Shortridge KF, Zhou NN, Guan Y, et al. Characterization of avian At the same time, some caution is warranted in the use of a H5N1 influenza viruses from poultry in Hong Kong. Virology needle-free delivery system according to lessons learned during this 1998;252:331–342. study. On occasion the vaccine failed to penetrate the skin layer and 2. Tang X, Tian G, Zhao J, Zhou KY. Isolation and characterization was left on the skin surface. This may have been as a result of a lack of prevalent strains of avian influenza viruses in China. Chin J of direct contact between the nozzle and the skin. In a few instances Anim Poult Infect Dis 1998;20:1–5.

312 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

. 3 Xu SK, Subbarao K, Cox N J, Guo Y. Genetic characteriza- gene of Asian H5N1 avian influenza virus generated a protective tion of the pathogenic influenza A/goose/Guangdong/1/96 antibody response in chickens against a North American virus (H5N1) virus: Similarity of its hemagglutinin gene to those of strain. Clin Vaccine Immunol 2013;20:491–500. Epub 2013 Jan 30. H5N1 viruses from the 1997 outbreaks in Hong Kong. Virology 12. World Organisation for Animal Health. Avian influenza 1999;261:15–19. (Chapter 2.3.4 [version adopted in May 2012]). In: Manual of 4. Food and Agriculture Organization. Bird flu rears its head again. Diagnostic Tests and Vaccines for Terrestrial Animals 2012. Increased preparedness and surveillance urged against variant Available from: http://www.oie.int/international-standard- strain [press release]. 2011 Aug 29. Available from: http://www. setting/terrestrial-manual/access-online/ Last accessed June 6, fao.org/news/story/en/item/87196/icode/ Last accessed 2013. February 2012;2011:1–20. June 6, 2013. 13. Ellis TM, Sims LD, Wong HK, et al. Use of avian influenza vac- 5. World Health Organization. Joint WHO-FAO-OIE assessment of cination in Hong Kong. Dev Biol (Basel) 2006;124:133–143. community-level risk of zoonotic avian influenza H5N1 infec- 14. Toma B, Dufour B, Sanaa M, et al. Disease control strategies. In: tions http://www.who.int/influenza/human_animal_interface/ Applied Veterinary Epidemiology and the Control of Disease in Joint_WHO_FAO_OIE_project_report_Oct12.pdf Last accessed Populations. Maisons-Alfort, France: Association for the Study July 30, 2013. of Epidemiology and Animal Diseases, 1999:275–306. 6. Webster RG. H5 influenza viruses. In: Kawaoka Y, ed. Influenza 15. Food and Agriculture Organization, Animal Production and Virology: Current Topics. Norfolk, England: Caister Academic Health. Approaches to Controlling, Preventing and Eliminating Press, 2006:281–298. H5N1 Highly Pathogenic Avian Influenza in Endemic Countries. 7. Amorij JP, Hinrichs WLJ, Frijlink HW, Wilschut JC, Huckriede A. Rome, Italy: FAO, 2011, paper 171:59–69. Available from: http:// Needle-free influenza vaccination. Lancet Infect Dis 2010;10: www.fao.org/docrep/014/i2150e/i2150e.pdf Last accessed 699–711. June 6, 2013. 8. Peyre M, Fusheng G, Desvaux S, Roger F. Avian influenza vac- 16. Webster RG, Webby RJ, Hoffman E, et al. The immunogenic- cines: A practical view in relation to their application in the field ity and efficacy against H5N1 challenge of reverse genetics- with a focus on the Asian experience. Epidemiol Infect 2009; derived H5N3 influenza vaccine in ducks and chickens. Virology 137:1–21. 2006;351:303–311. 9. Fuchs W, Römer-Oberdörfer A, Veits J, Mettenleiter TC. Novel 17. Middleton D, Bingham J, Selleck P, et al. Efficacy of inactivated avian influenza virus vaccines. Rev Sci Tech 2009;28:319–332. vaccines against H5N1 avian influenza infection in ducks. 10. Landau S. Needle-less injection system. US Patent 6,506,177 B2. Virology 2007;359:66–71. Alexandria, Virginia: United States Patent Office, 2003. 18. Kumar M, Chu H-J, Rodenberg J, Krauss S, Webster RG. 11. Ogunremi O, Pasick J, Kobinger GP, et al. A single electropora- Association of serologic and protective responses of avian tion delivery of a DNA vaccine containing the hemagglutinin influenza vaccines in chickens. Avian Dis 2007;50:481–483.

2000;64:0–00 The Canadian Journal of Veterinary Research 313 FOR PERSONAL USE ONLY

Short Communication Communication brève

The effects of dietary omega fatty acids on pregnancy rate, plasma prostaglandin metabolite levels, serum progesterone levels, and milk fatty-acid profile in beef cows Gavin F. Richardson, Mary A. McNiven, Hélène V. Petit, John L. Duynisveld

Abstract The objectives were to determine the effects of feeding supplements rich in omega-6 or omega-3 fatty acids (FA) during the late gestation to the early postpartum and breeding periods on reproduction and milk FA profile in beef cows. For each of two years, at the beginning of period 1 (mid-December), 72 beef cows, calving in January or February, were assigned to diets supplemented with roasted flaxseed (Flax) or roasted soybean (Soybean). For each of two years, after 11 wk (end of period 1), 18 cows of 36 in the Flax group were switched to the soybean supplement and 18 cows of 36 in the Soybean group were switched to the flax supplement (start of Period 2). Cows were bred by timed artificial insemination (TAI) in week 5 of period 2. The FA composition of the milk reflected the FA profile of the oilseed supplements. There were no differences in pregnancy rates among the 4 groups. The treatments had no effect on plasma prostaglandin metabolite levels or ratios at 4 to 11 d postpartum. At 5 to 6 d post- TAI, pregnant cows fed Flax in period 1 had lower (P , 0.05) plasma prostaglandin F metabolite (PGFM) levels and PGFM to prostaglandin E metabolite (PGEM) ratio than cows fed Soybean, but there were no significant differences at 19 to 20 d post-TAI. Cows pregnant from TAI and fed Flax in period 2 had higher (P , 0.05) serum progesterone levels at 5 to 6 d post-TAI than cows fed Soybean, but there was no difference at 19 to 20 d post-TAI. The dietary treatments had no effect on pregnancy rates, but there were some effects on plasma PGFM levels, PGFM to PGEM ratios, and serum progesterone levels. The FA supplements influenced the FA composition of milk.

Résumé Les objectifs de la présente étude étaient de déterminer les effets de suppléments alimentaires riches en acides gras (FA) omega-6 ou omega-3 lors de la période fin de gestation au début du postpartum et lors des périodes d’accouplement sur la reproduction et les profils de FA chez les vaches d’embouche. Pour chacune des deux années, au début de la période 1 (mi-décembre), 72 vaches d’embouche, devant vêler en janvier ou février, ont été assignées à des rations supplémentées avec de la graine de lin rôtie (Flax) ou des graines de soya rôties (Soya). Pour chacune des deux années, après 11 semaines (fin de la période 1), 18 des 36 vaches dans le groupe Flax ont été changées au supplément de soya et 18 des 36 vaches du groupe soya ont été changées pour le groupe Flax (Début de la période 2). Les vaches ont été saillies par insémination artificielle minutée (TAI) lors de la semaine 5 de la période 2. La composition en FA du lait représentait le profil de FA des suppléments alimentaires. Il n’y avait pas de différence dans les taux de gestation parmi les 4 groupes. Le traitement n’avait pas d’effet sur les niveaux ou ratios plasmatiques des métabolites des prostaglandines du jour 4 au jour 11 postpartum. Aux jours 5 à 6 post-TAI, les vaches gestantes nourries au Flax durant la période 1 avaient des niveaux significativement (P , 0,05) plus bas de métabolite de la prostaglandine F (PGFM) et des ratios de PGFM au métabolite de la prostaglandine E (PGEM) que les vaches nourries avec le Soya, mais il n’y avait pas de différence significative aux jours 19 à 20 post-TAI. Les vaches gestantes suite à la TAI et nourries avec Flax durant la période 2 avaient des niveaux sériques de progestérone plus élevés (P , 0,05) aux jours 5 à 6 post-TAI que les vaches nourries au Soya, mais il n’y avait plus de différence aux jours 19–20 post TAI. Les traitements alimentaires n’avaient aucun effet sur les taux de gestation, mais il y avait des différences sur les niveaux plasmatiques de PGFM, les ratios PGFM/PGEM, et les niveaux sériques de progestérone. Les FA des suppléments ont influencé la composition en FA du lait. (Traduit par Docteur Serge Messier)

Feeding cows supplemental fats high in omega-6 (n-6) fatty acids in the early postpartum period (1–3). Feeding omega-6 FA to dairy (FA) (e.g., linoleic acid, C18:2n-6) can increase tissue arachidonic cows in late pregnancy and early lactation has been reported to acid levels (1). High arachidonic acid levels can, in turn, increase increase PGF2a production, enhance immune competency, reduce uterine prostaglandin-F synthesis (1–3). Prostaglandin F2a (PGF2a) uterine infections, and assist normal uterine involution in the post- plays an important role in enhancing uterine defence mechanisms partum period, with consequent beneficial effects on fertility. In

Department of Health Management, Atlantic Veterinary College, University of Prince Edward Island, Charlottetown, Prince Edward Island (Richardson, McNiven); Dairy and Swine Research and Development Centre, Agriculture and Agri-Food Canada, Stn Lennoxville, Sherbrooke, Québec (Petit); and Agriculture and Agri-Food Canada Research Farm, Nappan, Nova Scotia (Duynisveld). Address all correspondence to Dr. Mary A. McNiven; telephone: (902) 566-0817; fax: (902) 566-0823; e-mail: [email protected] Received March 7, 2012. Accepted August 23, 2012.

314 The Canadian Journal of Veterinary Research 2013;77:314–318 FOR PERSONAL USE ONLY

addition, there is ­evidence of positive effects from feeding these FA blocks were available in each pen. The soybean supplement consisted on follicular growth, embryo quality, and pregnancy rates during of 2 kg of whole roasted soybean (containing linoleic acid, C18:2n-6, this period (1–3). 57.7%). The flax supplement consisted of 1 kg of roasted flaxseed During most of the gestation period, pregnancy in cattle is main- (containing linolenic acid, C18:3n-3, 58.5%) and 1 kg of soybean meal tained by progesterone, which is secreted by the corpus luteum (CL) to give similar lipid and protein contents. The oilseeds were roasted on the ovary. Prostaglandin F2a is the main hormone responsible for in a fluidized bed roaster (Sweet Manufacturing, Springfield, Ohio, luteolysis (4). In contrast, prostaglandin E2 (PGE2) is luteotropic (5) USA). Supplements were top dressed for each cow daily. For each and increased production by the uterus would support pregnancy. year, cows on each treatment were assigned randomly to 6 pens for

The relative production of PGF2a and PGE2 by the uterus may be a total of 6 cows per diet in each pen. One week after breeding, cows more important in the maintenance of pregnancy than the absolute were group fed according to the dietary supplement in 2 larger pens production of each prostaglandin (6). Feeding supplements high in to allow for breeding with bulls. The basal silage was mixed with omega-3 (n-3) FA (e.g., linolenic acid, C18:3n-3) during the breeding the dietary supplement at a rate of 2 kg per head. season and early gestation has been reported to reduce uterine PGF2a For each of two years, after 11 weeks on the experimental diets production and improve embryo quality and pregnancy mainte- (end of Period 1), 18 of the 36 cows in the Flax group, and 18 of nance. Pregnancy rates may be increased by enhanced progesterone the 36 cows in the Soybean group were switched to the Soybean production and decreased embryo mortality (2,3,6). and Flax diets, respectively (start of Period 2). This resulted in Dairy cattle have been used in most studies on the dietary effects four treatment groups: Flax-Flax, Flax-Soybean, Soybean-Flax, and of FA on bovine reproduction. Studies on beef cattle have yielded Soybean-Soybean. On the 11th wk of periods 1 and 2, calves were variable and inconsistent results (7,8). In addition, to the authors’ separated from cows for 12 h, returned to nurse to remove all avail- knowledge, there is no information on sequential feeding using able milk, and then removed until after milk samples were collected different fats during the late gestation to the early postpartum and for FA analysis 24 h later. For the period 2 milking, milk samples breeding periods in beef cows. were also collected from all cows for milk fat and protein analysis. The effects of dietary supplementation with FA on milk compo- Milk yield was also recorded at this time. sition has been studied extensively (9,10). The FA composition of The percentage of milk fat and protein was determined (FOSS the diet can affect milk FA profile, which in turn can influence calf Milko Scan FT 6000; FOSS, Eden Prairie, Minnesota, USA). For plasma and adipose tissue profiles (11). Fatty acids are important FA analysis, milk samples were centrifuged at 4°C, 15300 3 g in metabolic regulatory functions, vigor, and immune response in for 30 min (Eppendorf centrifuge model 5804R, Eppendorf rotor calves (11–15). model F45-30-11; Eppendorf Canada, Mississauga, Ontario) and The objectives of this study were to determine the effects of a weighed aliquot of the milk fat layer was removed for meth- feeding supplements high in omega-6 or omega-3 FA during the ylation and gas chromatography analysis (Agilent 5890, Agilent late gestation to the early postpartum and breeding periods on Technologies Canada, Mississauga, Ontario) (17). pregnancy rates, plasma prostaglandin metabolite levels and ratios, In weeks 4 and 5 of period 2, estrus was synchronized by treat- serum progesterone levels, and milk FA profile in beef cows. Our ing each cow with gonadotropin-releasing hormone (GnRH, 100 mg hypotheses were that increasing omega-6 FA in the diet would result Cystorelin IM; Mérial Canada, Baie d’Urfé, Québec) and a proges- in increased PGF2a production by the uterus, and that increasing terone releasing intravaginal insert (CIDR 1380, 1.38 g progesterone omega-3 FA in the diet would result in decreased PGF2a production. USP; Pfizer Canada, Kirkland, Quebec) on day 0, prostaglandin These could affect reproductive efficiency and progesterone produc- (Estrumate, 500 mg cloprostenol, IM; Schering Plough Canada, tion, as described previously. We also hypothesized that dietary FA Pointe-Claire, Quebec) on day 7, and a second GnRH injection at would be reflected in the milk FA profile. the time of insemination. All cows were bred by timed artificial The 144 cows used in this study were British-breed crossbred beef insemination (TAI) 53 to 56 h after CIDR removal. Inseminations cows in good body condition housed at the Agriculture and Agri-Food were done by 5 and 4 experienced technicians in year 1 and year 2, Canada Research Farm, Nappan, Nova Scotia. The mean 6 standard respectively, using frozen-thawed semen from sires with proven deviation (SD) body condition score of the cows was 5.7 6 0.9 in year 1 fertility. Technicians and semen were balanced as much as possible and 5.6 6 0.7 in year 2, on a 9-point scale, and the mean 6 SD body across treatments. Bulls were put in with the cows 2 d after TAI. weight was 708 6 95 kg in year 1 and 704 6 66 kg in year 2. Seventy- The cows were pregnancy tested by transrectal palpation or two cows (mean 6 SD age of 5.0 6 1.5 y in both years) were used for ultrasonography (Aloka SSD-500V with a 5.0 MHz linear array each of the 2 y and all cows calved in January or February. All animals transducer; IMAGO, Vaudreuil, Dorion, Quebec) 68 to 69 d after in this study were cared for in accordance with the guidelines sug- TAI. Cows were categorized into those bred by TAI (68 to 69 d of gested by the Canadian Council on Animal Care (16). gestation), those bred by the bull (50 d of gestation or less), and In each year, on December 15 (beginning of period 1), 72 cows those with no detectable pregnancy. The occasional cow may have were randomly assigned to 1 of 2 dietary treatment groups: Flax or been bred by a bull shortly after TAI, but this applied equally to all Soybean. The basal diet of first-cut timothy (70%, dry matter basis) groups so the results would not be affected. and red clover (30%, dry matter basis) silage (dry matter 32%) was Blood was collected for plasma prostaglandin metabolites fed free choice using Calan gates (American Calan, Northwood, [13,14-dihydro-15-keto-PGF2a (PGFM) and 13,14-dihydro-15-keto

New Hampshire, USA) for individual feeding, with a weighback prostaglandin E2 (PGEM)] levels 4 to 11 d after calving, and 5 to 6 d maintained of about 5% (as fed basis). Free-choice minerals and salt and 19 to 20 d post-TAI. Blood was drawn from the jugular vein

2000;64:0–00 The Canadian Journal of Veterinary Research 315 FOR PERSONAL USE ONLY

Table I. Percentage of selected fatty acids (FA) (g/100 g of total FA) in milk fat of beef cows at the end of periods 1 and 2. Minor FA have been excluded from table, but are included in totals Period 1 Period 2

Fatty acids Flax Soybean S¯x P Flax Soybean S¯x P C 10:0 2.36 2.28 0.051 0.306 2.36 2.21 0.057 0.054 C 12:0 2.76 2.62 0.064 0.118 2.71 2.54 0.067 0.089 C 14:0 9.06 8.31 0.143 0.0003 9.42 8.61 0.157 0.0004 C15:0 1.58 1.40 0.019 0.0001 1.37 1.25 0.015 0.0001 C16:0 22.6 22.3 0.260 0.461 22.4 22.1 0.328 0.426 C18:0 16.4 16.6 0.300 0.693 14.8 14.0 0.275 0.051 C18:1n-9 23.5 22.9 0.280 0.167 23.8 24.2 0.288 0.400 C18:2n-6 1.06 2.36 0.051 0.0001 1.12 2.30 0.057 0.0001 C18:1n-7t 2.43 3.62 0.079 0.0001 3.21 4.34 0.100 0.0001 c9, t11-CLA 0.79 1.12 0.022 0.0001 1.12 1.64 0.033 0.0001 C20:4n-6 0.09 0.11 0.003 0.0001 0.09 0.11 0.002 0.0001 C18:3n-3 0.96 0.97 0.012 0.628 1.04 1.02 0.016 0.245 C20:5n-3 0.12 0.09 0.002 0.0001 0.11 0.09 0.002 0.0001 C22:5n-3 0.17 0.13 0.004 0.0001 0.14 0.13 0.003 0.0001 C22:6n-3 0.03 0.02 0.002 0.009 0.03 0.02 0.001 0.288 SFA 60.3 58.9 0.34 0.007 58.3 55.8 0.40 0.0001 MFA 29.1 29.8 0.31 0.149 30.5 32.0 0.32 0.001 PUFA 3.30 4.95 0.077 0.0001 3.75 5.46 0.092 0.0001 n-3 FA 1.28 1.22 0.015 0.007 1.32 1.27 0.017 0.024 n-6 FA 1.23 2.60 0.052 0.0001 1.31 2.55 0.060 0.0001 n-3/n-6 1.09 0.49 0.019 0.0001 1.06 0.52 0.021 0.0001 SFA — saturated FA; MUFA — monounsaturated FA; PUFA — polyunsaturated FA; n-6 — omega-6 FA; n-3 — omega-3

FA; c9, t11-CLA — conjugated linoleic acid. S¯x — standard error of the mean. into 10 mL K2 EDTA tubes (Vacutainer tubes; Becton Dickinson, fat and protein) from samples collected during period 2, ANOVA Mississauga, Ontario). The samples were cooled on ice and then with period 1 treatment and period 2 treatment as main effects was centrifuged at 1000 3 g for 25 min. The plasma was transferred into done, as well as the interaction between the 2 treatment periods. In 2.0 mL cryovials (Fisher Scientific, Whitby, Ontario) and frozen on addition, pregnancy rates were analyzed using Fisher’s exact test. dry ice for transport to the laboratory where they were stored at There were no significant differences among treatments in milk 280°C until analyzed. yield, and percentages of milk fat and milk protein at the end of Plasma samples were assayed in triplicate for PGEM and PGFM by period 2 (P . 0.05; data not shown). This was expected because all enzyme immunoassay using commercial kits (No. 514531 and 516671, diets contained similar levels of oil. respectively; Cayman Chemical Company, Ann Arbor, Michigan, The results of the milk FA analysis are presented in Table I. USA). Inter-assay coefficient of variation for PGFM was 17.2%. Intra Because there was no significant interaction between dietary treat- and inter-assay coefficients of variation for PGEM were 9.1% and ments and periods 1 and 2, the results were pooled for each period. 15.3%, respectively. The sensitivities of the PGEM and PGFM assays In general, the FA composition of the milk reflected the major FA in were 2 and 8.2 pg/mL of plasma, respectively. the oilseed supplements with the milk from Flax-fed cows containing Blood was collected for serum progesterone levels 5 to 6 d more omega-3 FA, and the milk of the Soybean-fed cows containing and 19 to 20 d after TAI. Blood was drawn from the jugular vein more omega-6 FA. into 10 mL tubes without anticoagulant (Vacutainer tubes; Becton With respect to the omega-6 series of FA, there were significant Dickinson). After clotting, samples were centrifuged and the serum differences in the levels of linoleic acid (C18:2n-6, the major FA in was kept on ice for transport to the laboratory. Serum was stored at soybean), conjugated linoleic acid (CLA), and vaccenic acid, the 220°C until analysis for progesterone concentration, which was done precursor of CLA. This resulted in a significantly higher level of using a sequential competitive chemiluminescent enzyme immuno- total omega-6 FA in the milk for cows fed the Soybean supplement. assay (Immulite Progesterone; Diagnostic Products Corporation, Los For the omega-3 series of FA, there were no significant differences Angeles, California, USA). in linolenic acid (C18:3n-3) level between the flax and the soybean Results (PGFM, PGEM, PGFM to PGEM ratios, FA composition) supplemented cows. However, there were marked differences in from samples collected during period 1 were analyzed using a 2-way the longer chain omega-3 FA, i.e., C20:5n-3 (EPA), C22:5n-3 (DPA), analysis of variance (ANOVA) with dietary treatment and pen as the C22:6n-3 (DHA). Since none of these FA are found in plant material, main effects. For the results (PGFM, PGEM, PGFM to PGEM ratios, it is obvious that the metabolism of dietary linolenic acid (C18:3n-3) FA composition, progesterone, milk yield, and percentages of milk was efficient, and the production of these longer omega-3 FA was,

316 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

Table II. Pregnancy rates of cows fed 1 of 4 sequential dietary treatments as expressed as absolute numbers and percentages Pregnant by TAI Pregnant by bull Not pregnant at Supplement n (%) n (%) 68 to 69 d post-TAI (%) Flax-Flax (n = 32) 13 (40.6) 14 (43.8) 5 (15.6) Flax-Soybean (n = 34) 17 (50.0) 12 (35.3) 5 (14.7) Soybean-Flax (n = 32) 16 (50.0) 15 (46.9) 1 (3.1) Soybean-Soybean (n = 35) 20 (57.1) 10 (28.6) 5 (14.3) TAI — Timed artificial insemination.

Table III. Mean 6 S¯x plasma PGFM and PGEM levels (pg/mL) and PGFM to PGEM ratios of cows fed 1 of 4 sequential dietary treatments and pregnant at 5 to 6 and 19 to 20 d post-TAI 5 to 6 d post-TAI 19 to 20 d post-TAI Supplement PGFM PGEM PGFM/PGEM PGFM PGEM PGFM/PGEM Flax-Flax 156.0 6 39.8 36.2 6 3.6 4.8 6 1.2 212.5 6 49.8 36.2 6 3.6 6.2 6 1.4 Flax-Soybean 157.6 6 35.2 38.1 6 3.2 4.6 6 1.0 218.4 6 44.0 41.1 6 3.1 5.7 6 1.3 Soybean-Flax 233.8 6 37.2 34.0 6 3.4 9.0 6 1.1 287.2 6 46.6 34.7 6 3.3 9.3 6 1.3 Soybean-Soybean 227.4 6 32.5 42.7 6 3.0 5.9 6 1.0 270.9 6 40.7 36.8 6 2.9 7.5 6 1.2

PGFM — Prostaglandin F metabolites; PGEM — Prostaglandin E metabolites; TAI — Timed artificial insemination. S¯x — standard error of the mean. perhaps, limited by the amount of linolenic acid supplied in the cows fed Flax in period 1 compared with those fed Soybean, had supplement. significantly lower plasma PGFM levels (157.5 6 24.2 pg/mL and Unlike other studies (9,10), we did not have enough animals to 230.9 6 22.5 pg/mL, respectively; P , 0.05) and PGFM to PGEM have a control treatment with no fat supplement. However, it is obvi- ratio (4.7 6 0.7 and 7.4 6 0.8, respectively; P = 0.02), but there were ous from the milk FA profile that much of the dietary FA were able no significant differences in PGFM levels at 19 to 20 d post-TAI. to bypass the rumen and escape biohydrogenation due to the heat There was a tendency (P = 0.08) for cows pregnant 19 to 20 d post- treatment of the oilseed supplements. This could positively affect TAI that were fed Flax in period 1 to have a lower plasma PGFM to calf health and vigor (11–15). PGEM ratio than cows fed Soybean during this period. Thus, Flax The results of pregnancy diagnosis are presented in Table II. No appeared to have a long-lasting effect on prostaglandin metabolite significant differences in pregnancy rates were found among the levels, which was not affected by the FA supplement fed in period 2. 4 treatment groups (Fisher’s exact test, P = 0.43). This may be the The serum progesterone levels at 5 to 6 d and 19 to 20 d post-TAI result of an insufficient number of cows to demonstrate differences are presented in Table IV. Values at 19 to 20 d for nonpregnant cows or the type of dietary fat may not influence pregnancy rates in a are not presented as their stage of the estrous cycle would be uncer- well-managed herd. Recent studies involving this herd with cows tain. Cows pregnant from TAI and fed Flax in period 2 had higher on 1 of 3 levels of postpartum metabolizable energy intake and (P , 0.05) serum progesterone levels at 5 to 6 d post-TAI than cows using 2 methods of estrus synchronization have demonstrated good fed Soybean in period 2 (2.95 6 0.15 ng/mL and 2.52 6 0.14 ng/mL, pregnancy rates (53% to 57%) from TAI (18) regardless of which respectively). This effect was not seen in the nonpregnant cows at diet was fed. 5 to 6 d post-TAI or in pregnant cows at 19 to 20 d post-TAI.

Mean 6 S¯x plasma levels of PGFM taken 4 to 11 d after calv- Cows that were 19 to 20 d pregnant would be well past maternal ing were 1787.1 6 164.2 pg/mL and 1546.3 6 169.8 pg/mL for recognition of pregnancy, which occurs at 15 to 16 d of gestation (4), cows on the Flax and Soybean diets, respectively (P . 0.05). and the corpora lutea would be well established by this time.

Mean 6 S¯x levels of PGEM taken at the same time were 148.4 6 18.0 However, the differences in progesterone results between pregnant and 131.7 6 18.6 pg/mL for cows on the Flax and Soybean diets, and nonpregnant cows at 5 to 6 d post-TAI are difficult to explain, respectively (P . 0.05). In addition, there was no significant differ- as the embryos of the pregnant cows would not be hatched from ence in the plasma PGFM to PGEM ratios between the 2 treatments the zona pellucida and would be in the oviduct or tip of the uterine (data not shown). horn (4). The relatively low plasma PGFM and PGFM to PGEM Plasma levels of PGFM and PGEM, and PGFM to PGEM ratios of ratio in cows pregnant 5 to 6 d post-TAI and fed Flax in period 1 cows that were pregnant at 5 to 6 and 19 to 20 d post-TAI are pre- may partially explain these progesterone results, but this does not sented in Table III. There were no significant differences (P . 0.05) in account for the relatively high serum progesterone levels in the these levels in the nonpregnant cows (data not shown). In addition, Soybean-Flax group. the stage of the estrous cycle would be uncertain for nonpregnant Our hypotheses were partially supported by the results. The dietary cows 19 to 20 d post-TAI. treatments used in this trial had little effect on factors affecting preg- The combined effect of supplements fed in periods 1 and 2 had nancy rates in this well-managed beef herd. However, feeding flax, as no significant effect (P . 0.05) on plasma PGFM or PGEM levels, opposed to soybean, up to a month before breeding decreased plasma or PGFM to PGEM ratios. However, at 5 to 6 d post-TAI, pregnant PGFM levels and the PGFM to PGEM ratio in pregnant cows at 5 to 6 d

2000;64:0–00 The Canadian Journal of Veterinary Research 317 FOR PERSONAL USE ONLY

Table IV. Mean 6 S¯x serum progesterone levels (ng/mL) at 5 to 6 and 19 to 20 d post-TAI of cows fed 1 of 4 sequential dietary treatments Progesterone Progesterone levels at 5 to 6 d levels at Cows not 19 to 20 d pregnant Cows pregnant Cows pregnant Supplement from TAI from TAI from TAI Flax-Flax 2.24 6 0.24 2.99 6 0.21a 7.94 6 0.53 Flax-Soybean 2.04 6 0.25 2.69 6 0.19a,b 8.03 6 0.48 Soybean-Flax 2.87 6 0.25 2.92 6 0.20a 8.66 6 0.50 Soybean-Soybean 2.38 6 0.25 2.35 6 0.19b 8.08 6 0.47

TAI — Timed artificial insemination. S¯x — standard error of the mean. a,b Means with different superscripts within the same column are significantly differ- ent (P , 0.05). post-TAI. Feeding flax a month before TAI resulted in higher levels of 7. Fuston RN. Fat supplementation and reproduction in beef serum progesterone in cows 5 to 6 d pregnant compared with cows fed females. J Anim Sci 2004;82:E154–E161. soybean for the same period. In addition, the FA supplements in this 8. Hess BW, Lake SL, Scholljegerdes EJ, et al. Nutritional controls study influenced the FA composition of milk, which could positively of beef cow reproduction. J Anim Sci 2005;83:E90–E106. affect calf health and vigor. The effect of changes in FA composition 9. Glasser F, Ferlay A, Chilliard Y. Oilseed lipid supplements and of milk on calves requires further research. fatty acid composition of cow milk: A meta-analysis. J Dairy Sci 2008;91:4687–4703. Acknowledgments 10. Petit HV. Review: Feed intake, milk production and milk com- position of dairy cows fed flaxseed. Can J Anim Sci 2010;90: This study was supported by the Agriculture Research Fund, 115–127. Agricultural Resource Division, Prince Edward Island Department 11. Lake SL, Scholljegerdes EJ, Weston TR, Rule DC, Hess BW. of Agriculture. The authors thank the staff of the Agriculture and Postpartum supplemental fat, but not maternal body condition Agri-Food Canada Research Farm, Nappan, Nova Scotia for their score at parturition, affects plasma and adipose tissue fatty acid excellent cooperation and assistance. We greatly appreciate Brian profiles of suckling beef calves. J Anim Sci 2006;84:1811–1819. Andrews’ invaluable help with sample collection and recording. We 12. Lake SL, Scholljegerdes EJ, Small WT, et al. Immune response also thank Alfred Mitchell for his technical support and Dr. Bronwyn and serum immunoglobulin G concentrations in beef calves Crane for help with pregnancy diagnosis. suckling cows of differing body condition score at parturition and supplemented with high-linoleate or high-oleate safflower References seeds. J Anim Sci 2006;84:997–1003. 13. Petit HV, Berthiaume R. Effect of feeding different sources of fat . 1 Thatcher WW, Bilby TR, Bartolome JA, Silvestre F, Staples CR, during gestation and lactation on reproduction of beef cows and Santos JEP. Strategies for improving fertility in the modern dairy calf performance. Can J Anim Sci 2006;86:235–243. cow. Theriogenology 2006;65:30–44. 14. Ballou MA, Cruz GD, Pittroff W, Keisler DH, DePeters EJ. 2. Santos JEP, Bilby TR, Thatcher WW, Staples CR, Silvestre FT. Modifying the acute phase response of Jersey calves by supple- Long chain fatty acids of diet as factors influencing reproduction menting milk replacer with omega-3 fatty acids from fish oil. in cattle. Reprod Dom Anim 2008,Suppl2;43:23–30. J Dairy Sci 2008;91:3478–3487. 3. Silvestre FT, Carvalho TSM, Francisco N, et al. Effects of dif- 15. Ballou MA, DePeters EJ. Supplementing milk replacer with ferential supplementation of fatty acids during the peripartum omega-3 fatty acids from fish oil on immunocompetence and and breeding periods of Holstein cows: I. Uterine and metabolic health of Jersey calves. J Dairy Sci 2008;91:3488–3500. responses, reproduction, and lactation. J Dairy Sci 2011;94: 16. Canadian Council on Animal Care. CCAC guidelines on: The 189–204. care and use of farm animals in research, teaching and testing. 4. Senger PL. Pathways to Pregnancy and Parturition. 2nd ed. CCAC, Ottawa, Ontario, 2009. Pullman, Washington: Current Conceptions Inc., 2003:288–294,316. 17. McNiven MA, Duynisveld JL, Turner T, Mitchell AW. Ratio 5. Pratt RR, Butcher RI, Inskeep EK. Antiluteolytic effect of the of n-6/n-3 in the diets of beef cattle: Effect on growth, fatty

conceptus and of PGE2 in ewes. J Anim Sci 1977;45:784–781. acid composition, and taste of beef. Animal Feed Science and 6. Petit HV, Small JA, Palin MF, Giguère A, Santos GTD. Effects of Technology 2011;170:171–181. flaxseed supplementation on endometrial expression of ISG17 18. Wichtel JJ, Charmley E, Richardson GF, Duynisveld JL, Lofstedt R. and intrauterine prostaglandin concentrations in primiparous Effects of postpartum energy intake on pregnancy rates in beef dairy cows submitted to GnRH-based synchronized ovulation. cattle subjected to GnRH- or CIDR-based timed artificial insemi- Can J Anim Sci 2007;87:343–352. nation protocols. Can J Anim Sci 2008;88:439–447.

318 The Canadian Journal of Veterinary Research 2000;64:0–00 FOR PERSONAL USE ONLY

2000;64:0–00 The Canadian Journal of Veterinary Research 319 FOR PERSONAL USE ONLY

“Best Medicine Practices – Timely Topics” ST. JOHN’S, NEWFOUNDLAND & LAbrADOr JULY 9–12, 2014 Join us for top notch scientific sessions in: Sign up for Workshops: Companion Animals, Equine, Ruminant Level 2 Dental Wet Lab and Other Species/Topics Tracks - Difficult extractions made easier - Dental radiographic positioning CVMA’s Business management sessions and interpretation will discuss “Not What You Say But How!” Orthopaedic Lab - Basic fracture fixation principles New for 2014 - CVMA 1st Annual using plastinated bones Veterinarian Student Poster Competition! A Social Evening to look forward to: “Havin’ a Time in Newfoundland!” An evening to experience our colourful culture and world renowned hospitality.

In Collaboration with: The Canadian Association A New Discovery Down Every Lane of Animal Health Technologists and Technicians www.canadianveterinarians.net FOR PERSONAL USE ONLY

Produits et services Supplies and services

Reaching Canada’s Veterinarians Canadian Journal Get your message into of Veterinary Research available online now! The Canadian Veterinary Journal Articles are now easily accessible For more information contact: over the Internet at Laima Laffitte Advertising Manager Tel.: (613) 673-2659 PubMed Central Fax: (613) 673-2462 www pubmedcentral. .com e-mail: [email protected]

Four issues of the RENEW YOUR SUBSCRIPTION Canadian Journal of Veterinary Research for 2013 NOW! (online) Quatre numéros de la RENOUVELEZ VOTRE ABONNEMENT Revue canadienne de recherche vétérinaire pour 2013 DÈS MAINTENANT! $150 $ ( applicable GST/HST; Canadian/Canadien TPS/TVH en vigueur) $165 $ US/É.-U. Foreign/Étranger $50 $ ( applicable GST/HST) Back issues/Anciens numéros each/chacun

Contact/Communiquez avec : Beverley Kelly Editorial Coordinator/Coordonnatrice de la rédaction Canadian Journal of Veterinary Research Revue canadienne de recherche vétérinaire 339 rue Booth Street Ottawa, Ontario K1R 7K1 Canada Tel./Tél. : (613) 236-1162, x117 Fax/Télécopieur : (613) 236-9681 E-mail/Courriel : [email protected]

www.canadianveterinarians.net

Tables of contents Tables des matières Coming events Événements à venir Article abstracts Résumés d’articles Classified advertising Annonces classées Subscription information Information sur l’abonnement Instructions for authors Directives à l’intention des auteurs

www.veterinairesaucanada.net Evaluation of cerebrospinal fluid lactate and plasma lactate concentrations in anesthetized dogs with and without intracranial disease Deanne Caines, Melissa Sinclair, Darren Wood, Alexander Valverde, Doris Dyson, Luis Gaitero, Stephanie Nykamp ...... 297

Sandwich-dot enzyme-linked immunosorbent assay for the detection of canine distemper virus Zhi Li, Yanlong Zhang, Huiguo Wang, Jinhua Jin, Wenzhe Li ...... 303

Short Communications/ Communications brèves Needle-free delivery of an inactivated avian influenza H5N3 virus vaccine elicits potent antibody responses in chickens Oladele Ogunremi, John Pasick, Yohannes Berhane ...... 309

The effects of dietary omega fatty acids on pregnancy rate, plasma prostaglandin metabolite levels, serum progesterone levels, and milk fatty-acid profile in beef cows Gavin F . Richardson, Mary A . McNiven, Hélène V . Petit, John L . Duynisveld . . . . . 314