Preventive Veterinary Medicine 113 (2014) 620–624
Contents lists available at ScienceDirect
Preventive Veterinary Medicine journal homepage: www.elsevier.com/locate/prevetmed
Short Communication
A field study evaluation of PetrifilmTM plates as a 24-h rapid diagnostic test for clinical mastitis on a dairy farm Elisabeth Maria Mansion-de Vries a,b , Nicole Knorr a , Jan-Hendrik Paduch a , Claudia Zinke a , Martina Hoedemaker b , Volker Krömker a,∗ a b
University of Applied Sciences and Arts, Faculty II, Microbiology, Heisterbergallee 12, 30453 Hannover, Germany Clinic for Cattle, Hanover University of Veterinary Medicine, Foundation, Bischofsholer Damm 15, 30173 Hannover, Germany
a r t i c l e
i n f o
Article history: Received 11 May 2013 Received in revised form 24 November 2013 Accepted 27 November 2013
Keywords: Mastitis PetrifilmTM Microbiology Cultural diagnosis Test evaluation
a b s t r a c t Clinical mastitis is one of the most common and expensive diseases of dairy cattle. To make an informed treatment decision, it is important to know the causative pathogen. However, no detection of bacterial growth can be made in approximately 30% of all clinical cases of mastitis. Before selecting the treatment regimen, it is important to know whether the mastitis-causing pathogen (MCP) is Gram-positive or Gram-negative. The aim of this field study was to investigate whether using two 3M PetrifilmTM products on-farm (which conveys a higher degree of sample freshness but also bears a higher risk for contamination than working in a lab) as 24-h rapid diagnostic of clinical mastitis achieved results that were comparable to the conventional microbiological diagnostic method. AerobicCount (AC)-PetrifilmTM and ColiformCount (CC)-PetrifilmTM were used to identify the total bacterial counts and Gram-negative bacteria in samples from clinical mastitis cases, respectively. Missing growth on both plates was classified as no bacterial detection. Growth only on the AC-PetrifilmTM was assessed as Gram-positive, and growth on both PetrifilmTM plates was assessed as Gram-negative bacterial growth. Additionally, milk samples were analysed by conventional microbiological diagnostic method on aesculin blood agar as a reference method. Overall, 616 samples from clinical mastitis cases were analysed. Using the reference method, Gram-positive and Gram-negative bacteria, mixed bacterial growth, contaminated samples and yeast were determined in 32.6%, 20.0%, 2.5%, 14.1% and 1.1% of the samples, respectively. In 29.7% of the samples, microbiological growth could not be identified. Using the PetrifilmTM concept, bacterial growth was detected in 59% of the culture-negative samples. The sensitivity of the PetrifilmTM for Gram-positive and Gram-negative MCP was 85.2% and 89.9%, respectively. The specificity was 75.4% for Gram-positive and 88.4% for Gramnegative MCP. For the culture-negative samples, sensitivity was 41.0% and specificity was 91.0%. The results indicate that the PetrifilmTM concept is suitable for therapeutic decisionmaking at the farm level or in veterinary practice. As this concept does not allow any statement about the genus or species of microorganisms, relevant MCP should be assessed periodically at the herd level with conventional microbiological diagnostics. © 2014 Elsevier B.V. All rights reserved.
1. Introduction ∗ Corresponding author at: Microbiology, Faculty of Mechanical and Bioprocess Engineering, University of Applied Sciences and Arts, Heisterbergallee 12, 30453 Hannover, Germany. Tel.: +49 511 9296 2205; fax: +49 511 9296 2210. E-mail address: [email protected] (V. Krömker). 0167-5877/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.prevetmed.2013.11.019
Mastitis remains one of the most frequent disease and reason for antibiotic use in dairy cattle. In Germany the economic loss due to mastitis is approximately 1.4 billion D each year (German Veterinary Association
E.M. Mansion-de Vries et al. / Preventive Veterinary Medicine 113 (2014) 620–624
[GVA], 2012). In order to ensure a maximum treatment effect along with a minimum investment, an evidence-based diagnosis and a prudent use of antimicrobials are fundamental. Therefore, it is necessary to know the mastitis-causing pathogen (MCP). Antimicrobial treatment of clinical mastitis is only recommended when a pathogen is detected (Roberson, 2003). Gramnegative coliform bacteria could allow an antibiotic-free treatment of the infected animals, as long as they are afebrile (Guterbock et al., 1993). Conversely, with regard to antibiotic resistance, therapy with a narrow-spectrum antibiotic is usually recommended for infections with Gram-positive microorganisms (Wilson et al., 1999; Lago et al., 2011). Therefore, a milk sample from each quarter with clinical mastitis should be examined bacteriologically (Lago et al., 2011). The current predominant method used is the conventional microbiological culture (National Mastitis Council (NMC), 1999; GVA, 2009), which costs about 2.50 D per milk sample investigation. Despite reliable and reproducible results, the standard microbiological diagnostic method has several weaknesses, such as a high exposure time of at least 48 h and a need for a lot of diagnostic experience (Krömker et al., 2007). An alternative fast, easy and very sensitive diagnostic test is the commercial PathoProof® Mastitis PCR Assay, a real-time PCR for identifying eleven mastitis pathogens (Koskinen et al., 2009). The investigation by PathoProof® has to be performed in a special laboratory, so time is lost on sample transportation. Additionally, the price of 18 D per milk sample investigation is high. These limitations illustrate the need for an easy, fast and cost effective diagnostic test that can ideally even be performed on a farm. A simple diagnostic test system is offered by the 3M Company. They produce several PetrifilmTM plates, which can be used to detect bacteria or bacterial groups with different sample-ready selective culture systems. Each PetrifilmTM plate costs about 1.50 D. These films contain a cold water soluble gelling agent, nutrients and an indicator. Authors of several studies have shown that PetrifilmTM plates are suitable for identifying mastitis pathogens (McCarron et al., 2009; Neuling et al., 2011; Gitau et al., 2013) and can be used to diagnose infections due to MCP (Krömker et al., 2007). In a laboratory setting, McCarron et al. (2009) found that parallel investigation of mastitic milk samples by AerobicCount (AC)-PetrifilmTM and ColiformCount (CC)-PetrifilmTM allows a differentiation between Gram-positive and Gram-negative bacteria. The authors used conventional microbiological diagnoses as the reference method and obtained a sensitivity of 93.8% and specificity of 70.1%. The aim of this study was to evaluate whether using two PetrifilmTM products in an on-farm setting (which conveys a higher degree of sample freshness but also bears a higher risk for contamination than working in a lab) could achieve comparable results compared to the conventional microbiological diagnosis method. Achieving these goals would provide a rapid, cost effective and useful diagnostic method to significantly facilitate treatment decisions.
621
2. Materials and methods 2.1. Samples Quarter milk samples of all animals with signs of clinical mastitis (i.e., clotting and/or discoloration of milk, udder swelling or redness, heat upon udder palpation) were obtained over an 11-month period at a single commercial dairy herd with 1000 lactating cows in Saxony-Anhalt (Germany). Immediately after detecting a clinical mastitis, a duplicate foremilk sample of the affected quarter was taken aseptically by the farmer. Samples were collected, as described by NMC (1999), and stored at a temperature between 2 ◦ C and 8 ◦ C. One of the duplicate samples (randomly allocated) was analysed on the farm within 12 h after collecting the sample using the AC- and CC-PetrifilmTM plates, whereas the second sample was transported to the microbiological Laboratory of the University of Applied Sciences and Arts Hannover (Germany) for conventional culture technique, twice a week. The laboratory samples were investigated within two hours after arrival by a trained blinded researcher through conventional microbiological diagnostic according to the GVA (2009). The data collection was planned before starting the study. 2.2. PetrifilmTM concept On the dairy farm, the examination of milk samples was performed with the AC- and CC-PetrifilmTM plates (3M, Neuss, Germany). The test was conducted following the manufacturer’s instructions. First, after vigorous shaking of the sample, a 1:10 dilution was prepared with sterile isotonic Ringer solution (Merck, Darmstadt, Germany), which was also thoroughly mixed. 1 ml of the dilution was added on each carrier film of both PetrifilmTM plates. After covering the sample with the top film it was spread out equally by the product-related spreader and, subsequently, incubated for 24 h at 37 ◦ C. After 24 h, colony growth was read with the aid of the PetrifilmTM integrated counting grid and the indicator dye. Enumeration and interpretation of the results were based on the recommendations of McCarron et al. (2009): A PetrifilmTM plate was categorised as positive when at least 20 colonies were counted on the CC-PetrifilmTM or five colonies were counted on the AC-PetrifilmTM . Based on these enumerations, a diagnosis was made (Table 1). The preparation and evaluation of the PetrifilmTM plates was conducted by a trained researcher at the farm. 2.3. Microbiological diagnosis For the microbiological diagnosis, 10 l of a wellmixed quarter foremilk sample were plated onto an aesculin blood agar plate (Oxoid, Wesel, Germany) and on a ChromoCult® Coliform Agar (Merck, Darmstadt Germany). The plates were incubated aerobically at 37 ◦ C and examined after 24 h and 48 h, according to the GVA recommendations (2009). The colonies were identified by Gram staining, cell morphology, aesculin hydrolysis and haemolysis patterns. Catalase and Gram-positive cocci were identified as staphylococci. Staphylococcus (S.) aureus
622
E.M. Mansion-de Vries et al. / Preventive Veterinary Medicine 113 (2014) 620–624
Table 1 Definition of the PetrifilmTM diagnosis (according to McCarron et al., 2009). CCa -PetrifilmTM
ACb -PetrifilmTM
Diagnosis
Number of diagnosis
Positive Negative Negative
Positive Positive Negative
Gram-negative pathogen Gram-positive pathogen No pathogen growth detected
218 368 144
a b
Coliform-count plate. Aerobic-count plate.
was defined as showing beta-haemolysis and having a positive clumping factor (Diamondial StaphPlus Kit, Diamondial, Sées, France) reaction. Other staphylococci were referred to as coagulase negative staphylococci (CNS), while Gram-positive, catalase negative cocci were identified as streptococci. Modified Rambach agar (Watts et al., 1993) was used to differentiate the aesculin hydrolysing cocci. -d-galactosidase-positive and aesculin hydrolysing cocci were identified as Streptococcus (Sc.) uberis, and aesculin hydrolysing, -d-galactosidase-negative cocci were identified as enterococci. Beta-haemolytic streptococci were further characterised by Lancefield serotyping (Diamondial Strep Kit, Diamondial, Sées, France). Streptococci from group C were referred to as Sc. dysgalactiae. Grampositive, asporogenic, beta-haemolytic, catalase-negative irregular rods with a V- or Y-shaped configurations were identified as Trueperella (T., formerly Arcanobacterium) pyogenes. Coliform bacteria were catalase negative, Gramnegative and cytochrome oxidase negative (Bactident oxidase, Merck, Darmstadt, Germany) rod-shaped bacteria that were capable of fermenting glucose (OF basal medium with addition of d [+]-glucose monohydrate, Merck, Darmstadt Germany). On ChromoCult® Coliform Agar (Merck, Darmstadt, Germany), Escherichia (E.) coli formed blue colonies, unlike other coliforms that formed pink-red colonies. Samples were categorised as positive when a minimum of three colonies grew. However S. aureus, Sc. dysgalactiae and T. pyogenes were identified if at least one colony was detected. If more than two different types of colonies were identified, a sample was considered to be contaminated. Even growth of S. aureus, Sc. dysgalactiae and T. pyogenes from contaminated samples were reported. Yeasts were differentiated microscopically after subculturing on YGC agar (Merck, Darmstadt, Germany). To compare the results with those of the PetrifilmTM method, they were summarised as Gram-negative MCP, Gram-positive MCP and no bacterial detection. Yeasts were numbered among Gram-positive MCP. If samples were contaminated, they were excluded from the evaluation of the PetrifilmTM . Samples with mixed growth which was not due to uniform Gram-positive or Gram-negative growth (not-uniformly mixed growth) were also excluded from evaluation.
2.4. Statistical analyses The data were recorded in Microsoft Access (Microsoft, Redmond, USA). The microbiological results of the PetrifilmTM method were compared with the results of the reference method. The PetrifilmTM results (Grampositive MCP, Gram-negative MCP or no bacterial detection) were considered false positives or false negative
Table 2 Numbers and percentages of mastitis causing pathogens isolated with the reference method from milk samples from 730 clinical mastitis cases recorded on one German farm with app. 1000 cows during an 11-month study period. Finding/pathogen
Number
Percentage (%)
No growth E. coli Other Gram-negative pathogens Sc. uberis S. aureus Sc. dysgalactiae T. pyogenes CNSa Other Gram-positive pathogens Mixed growthb Yeast Contaminated Total
217 129 17 107 35 28 19 19 30 18 8 103 730
29.7 17.7 2.3 14.7 4.8 3.8 2.6 2.6 4.1 2.5 1.1 14.1 100
a b
Coagulase negative staphylococci. Two pathogens were detected.
when the reference methods did not result in confirmation. The epidemiological software Win Episcope 2.0 (http://www.clive.ed.ac.uk/winepiscope/) was used to estimate sensitivity, specificity, positive and negative predictive value, Youden’s index, and the true and apparent prevalence, referring to the standard epidemiological methods of Kreienbrock and Schach (2000). 3. Results A total of 763 double quarter foremilk samples from German Holstein Friesian dairy cows with clinical mastitis were taken from 01 February 2012 to 31 December 2012. Because of either the reference method or the PetrifilmTM method results were lacking in 33 samples, these samples were excluded from the data analysis. Additionally, 114 samples were not used in the evaluation of the PetrifilmTM , because of contamination (n = 103) or nonuniformly mixed growth (n = 11). As a result, 616 double foremilk samples were included in the analysis. With the reference method, microbiological growth could not be identified in 217 samples (Table 2), whereas in 128 (59.0%) of these samples, the PetrifilmTM concept was able to detect bacterial growth. In 70.7% of the samples, the findings of the PetrifilmTM method were consistent with the results of the reference method (Table 3). The results of the statistical analyses are illustrated in Table 4. 4. Discussion In the present study, the diagnostic accuracy of the PetrifilmTM method was examined on the division of
E.M. Mansion-de Vries et al. / Preventive Veterinary Medicine 113 (2014) 620–624
623
Table 3 Contingency table for the comparison of the results of the reference method with the results of the PetrifilmTM concept. Milk samples from 616 clinical mastitis cases from one German farm with app. 1000 cows were included into the 11-month study. PetrifilmTM method
Reference method
a
Gram-negative MCP Gram-positive MCP No MCP detected Total
Gram-negative MCPa
Gram-positive MCP
No MCP detected
Total
134 8 46 188
8 213 82 303
7 29 89 125
149 250 217 616
Mastitis causing pathogen.
mastitis milk samples into no growth, Gram-negative and Gram-positive bacterial growth compared to the standard diagnostic method to clarify whether a treatment decision could be made based on results using the PetrifilmTM method. Consistently high sensitivities and specificities and both positive and negative predictive values were determined. Furthermore the results of the PetrifilmTM plates were obtained in less than 24 h after collection, much faster than the results from the conventional microbiological diagnostic method. Low sensitivity and a correspondingly moderate Youden’s index only occur when identifying no bacterial growth because the PetrifilmTM concept detected bacterial growth in 59% of the culture-negative samples. This result illustrates that the PetrifilmTM concept identified more MCPs compared to the standard microbiological diagnostic method, which could be explained by the 10-fold greater inoculum of 100 l, which increased the probability of pathogen detection (Sears et al., 1990; Hogan and Smith, 2003). In this way lacking diagnostic sensitivity of the reference method negatively affects the outcome of the PetrifilmTM method. It is also possible to increase the detection of MCP’s in the conventional microbial diagnostic by increasing the inoculum (Krömker et al., 2010). However, in Germany the official required test method is the conventional microbial diagnostic with 10 l inoculum, so the PetrifilmTM method had to be compete with these standards. Furthermore, faster sample processing could have influenced the rate of negative samples. The PetrifilmTM plates were prepared immediately after collecting the samples, with no more than a 12-h delay after the collection. In contrast, because of the need for only portage of the reference method samples, at least one 24-h period is expected from the collection time to arrival at the laboratory. The negative predictive value verifies whether a result is truly negative when the test has declared it negative. The negative predictive value of the PetrifilmTM method were between 73.9% (in the
identification of no pathogen growth) and 96.5% (in the detection of Gram-negative MCP), which indicates the benefits of the method. The sensitivity and specificity values that were calculated in the present study were comparable to the results of McCarron et al. (2009) and Gitau et al. (2013). In both studies, the preparation of the PetrifilmTM plates was performed in the laboratory. McCarron et al. (2009) investigated also the interreader agreement on interpreting the PetrifilmTM results. They state that the agreement between readers is very good (kappa around 0.85) in spite of different microbiological training. Still, tests were performed by one and the same technician as a courtesy for the farmers’ will to cooperate. Despite high consensus when comparing the results of the conventional microbiological diagnostic method and the PetrifilmTM concept and many benefits of the PetrifilmTM plates, even a few difficulties in the implementation process must be noted. The authors of the present study estimate that the greatest weakness is the lack of information about the quality of the milk samples. A statement of whether the milk sample was collected aseptically or was contaminated is impossible when using the PetrifilmTM technique. Identification of mixed infections is also excluded because of the growth of Gram-negative MCP on both the AC and the CC-PetrifilmTM . The clinical impact of the inability of the PetrifilmTM method to identify mixed infections is low, since only 2–3% of the samples are usually mixed infections (Olde Riekerink et al., 2008; Gitau et al., 2013), as was the case in the present study. Yeast does only grow on the AC PetrifilmTM . In most cases they are not detectable within 24 h of incubation because they grow very slowly. In our study 6 out of 8 milk samples with yeast results in the reference method had no pathogen detection with the PetrifilmTM method (data not shown). The clinical impact of the determination yeast as Grampositive MCP in the PetrifilmTM method is also low, since only 1–2% of the samples are usually yeast infections (Olde
Table 4 Test characteristics of the PetrifilmTM concept based on its ability to correctly classify the bacterial growth (classes: Gram-positive, Gram-negative, no growth) from milk samples from 616 clinical mastitis cases recorded on one German farm with app. 1000 cows during an 11-month study period. Test characteristic
Gram-negative MCPa (%)
Gram-positive MCP (%)
No bacterial growth (%)
Sensitivity (95% CI) Specificity (95% CI) Positive predictive value (95% CI) Negative predictive value (95% CI) Youden’s index (95% CI) True prevalence (95% CI) Apparent prevalence (95% CI)
89.9 (85.1–94.8) 88.4 (85.5–91.3) 71.3 (64.8–77.7) 96.5 (94.8–98.2) 0.78 (0.73–0.84) 24.2 (20.8–27.6) 29.9 (26.9–34.2)
85.2 (80.8–89.6) 75.4 (71.0–79.8) 70.3 (65.2–75.4) 88.2 (84.6–91.8) 0.61 (0.54–0.67) 40.6 (36.7–44.5) 49.2 (45.2–53.1)
41.0 (34.5–47.6) 91.0 (88.2–93.8) 71.2 (63.3–79.1) 73.9 (70.0–77.8) 0.32 (0.25–0.39) 35.2 (31.5–39.0) 20.3 (17.1–23.5)
a
Mastitis causing pathogen.
624
E.M. Mansion-de Vries et al. / Preventive Veterinary Medicine 113 (2014) 620–624
Riekerink et al., 2008), as was the case in the present study. Furthermore, incubated PetrifilmTM plates cannot be easily disposed of through the municipal waste. They must be autoclaved prior to properly disposing of them. With the implementation of PetrifilmTM plates on a dairy farm, the use of a steam cooker would be conceivable. Despite these weaknesses the present study showed that the PetrifilmTM achieved good results. The sensitivity and NPV of Gram-negative detection was as excellent as established by Leslie et al. (2005). If treatment of Gramnegative MCP will be done without antibiotics the in-field PetrifilmTM method could help to reduce antibiotic application by approximately 20%. 5. Conclusion The study results lead to the assessment that the present PetrifilmTM method can be used for classifying samples from clinical mastitis cases with Gram-positive and Gramnegative MCPs or samples without growth; therefore, it is suitable for therapeutic decision making at the farm level or in veterinary practice. However, this method only allows the determination of Gram-positive or Gramnegative pathogens, without providing any information about the species; therefore, standard diagnostics may still be needed. To monitor the prevalence of MCP’s at the herd level, conventional microbiological diagnostics should be performed because of its greater discriminative potential. Because changes in the prevalence of MCP are possible, this method should be periodically evaluated. Furthermore, note that reliable results can be achieved only if the sampling is performed aseptically. Conflict of interest statement None declared. Acknowledgment The authors thank Ohreland KG for providing their cows and Vrieswoud KG to enable us to carry out the study. Furthermore we are grateful to 3M for supplying the PetrifilmTM plates. References German Veterinary Association (GVA), 2009. Leitlinien zur Entnahme von Milchproben unter antiseptischen Bedingungen und Leitlinien
zur Isolierung und Identifizierung von Mastitiserregern. GVA, Gießen, Germany. German Veterinary Association (GVA), 2012. Leitlinien zur Bekämpfung der Mastitis des Rindes als Herdenproblem. GVA, Gießen, Germany. Gitau, G.K., Bundi, R.M., Vanleeuwen, J., Mulei, C.M., 2013. Evaluation of PetrifilmsTM as a diagnostic test to detect bovine mastitis organisms in Kenya. Trop. Anim. Health Prod. 45, 883–886. Guterbock, W.M., Van Eeenennaam, A.L., Anderson, R.J., Gardner, I.A., Cullor, J.S., Holmberg, C.A., 1993. Efficacy of intramammary antibiotic therapy for treatment of clinical mastitis caused by environmental pathogens. J. Dairy Sci. 76, 3437–3444. Hogan, J., Smith, K.L., 2003. Coliform mastitis. Vet. Res. 34, 507–519. Kreienbrock, L., Schach, S. (Eds.), 2000. Epidemiologische Methoden. Spektrum Akademischer Verlag, Heidelberg, Germany. Koskinen, M., Holopainen, J., Pyörälä, S., Bredbacka, P., Pitkälä, A., Barkema, H., Bexiga, R., Roberson, J., Sølverød, L., Piccinini, R., Kelton, D., Lehmusto, H., Niskala, S., Salmikivi, L., 2009. Analytical specificity and sensitivity of a real-time polymerase chain reaction assay for identification of bovine mastitis pathogens. J. Dairy Sci. 92, 952–959. Krömker, V., Hauptmann, T., Bormann, A., 2007. Bacteriological diagnosis of mastitis pathogens in quarter milk samples and bulk milk samples with petrifilm. Dtsch. Tierärztl. Wschr. 114, 378–383. Krömker, V., Paduch, J.-H., Klocke, D., Zinke, C., 2010. Microbiological investigation of culture negative milk samples of clinical mastitis cows. Milchwissenschaft 65, 123–126. Lago, A., Godden, S.M., Bey, R., Ruegg, P.L., Leslie, K., 2011. The selective treatment of clinical mastitis based on on-farm culture results: I. Effects on antibiotic use, milk withholding time, and short-term clinical and bacteriological outcomes. J. Dairy Sci. 94, 4441–4456. Leslie, K., Walker, M., Vernooy, E., Bashiri, A., Dingwell, R., 2005. Evaluation of the PetrifilmTM culture system for the identification of mastitis bacteria as compared to standard bacteriological methods. In: Hogeveen, H. (Ed.), Mastitis in Dairy Production: Current Knowledge and Future Solutions, Mastitis in Dairy Production: Current Knowledge and Future Solutions. Wageningen Academic Publishers, Wageningen, pp. 416–421. McCarron, J.L., Keefe, G.P., McKenna, S.L., Dohoo, I.R., Poole, D.E., 2009. Laboratory evaluation of 3M Petrifilms and University of Minnesota Bi-plates as potential on-farm tests for clinical mastitis. J. Dairy Sci. 92, 2297–2305. National Mastitis Council (NMC), 1999. Laboratory Handbook on Bovine Mastitis. NMC, Madison, WI. Neuling, R., Paduch, J.-H., Bormann, A., Zinke, C., Krömker, V., 2011. Evaluation of bacteriological diagnosis in samples of clinical mastitis cases with PetrifilmTM in veterinary practice. Tierarztl. Prax. Ausg. G Grosstiere Nutztiere 39, 77–80. Olde Riekerink, R.G.M., Barkema, H.W., Kelton, D.F., Scholl, D.T., 2008. Incidence rate of clinical mastitis on Canadian dairy farms. J. Dairy Sci. 91, 1366–1377. Roberson, J.R., 2003. Establishing treatment protocols for clinical mastitis. Vet. Clin. Food Anim. 19, 223–234. Sears, P.M., Smith, B.S., English, P.B., Herer, P.S., Gonzalez, R.N., 1990. Shedding pattern of Staphylococcus aureus from bovine intramammary infections. J. Dairy Sci. 73, 2785–2789. Watts, J.L., Salmon, S.A., Yancey Jr., R.J., 1993. Use of modified Rambach agar to differentiate Streptococcus uberis from other mastitis streptococci. J. Dairy Sci. 76, 1740–1743. Wilson, D.J., Gonzalez, R.N., Case, K.L., Garrison, L.L., Gröhn, Y.T., 1999. Comparison of seven antibiotic treatments with no treatment for bacteriological efficacy against bovine mastitis pathogens. J. Dairy Sci. 82, 1664–1670.
Contents lists available at ScienceDirect
Preventive Veterinary Medicine journal homepage: www.elsevier.com/locate/prevetmed
Short Communication
A field study evaluation of PetrifilmTM plates as a 24-h rapid diagnostic test for clinical mastitis on a dairy farm Elisabeth Maria Mansion-de Vries a,b , Nicole Knorr a , Jan-Hendrik Paduch a , Claudia Zinke a , Martina Hoedemaker b , Volker Krömker a,∗ a b
University of Applied Sciences and Arts, Faculty II, Microbiology, Heisterbergallee 12, 30453 Hannover, Germany Clinic for Cattle, Hanover University of Veterinary Medicine, Foundation, Bischofsholer Damm 15, 30173 Hannover, Germany
a r t i c l e
i n f o
Article history: Received 11 May 2013 Received in revised form 24 November 2013 Accepted 27 November 2013
Keywords: Mastitis PetrifilmTM Microbiology Cultural diagnosis Test evaluation
a b s t r a c t Clinical mastitis is one of the most common and expensive diseases of dairy cattle. To make an informed treatment decision, it is important to know the causative pathogen. However, no detection of bacterial growth can be made in approximately 30% of all clinical cases of mastitis. Before selecting the treatment regimen, it is important to know whether the mastitis-causing pathogen (MCP) is Gram-positive or Gram-negative. The aim of this field study was to investigate whether using two 3M PetrifilmTM products on-farm (which conveys a higher degree of sample freshness but also bears a higher risk for contamination than working in a lab) as 24-h rapid diagnostic of clinical mastitis achieved results that were comparable to the conventional microbiological diagnostic method. AerobicCount (AC)-PetrifilmTM and ColiformCount (CC)-PetrifilmTM were used to identify the total bacterial counts and Gram-negative bacteria in samples from clinical mastitis cases, respectively. Missing growth on both plates was classified as no bacterial detection. Growth only on the AC-PetrifilmTM was assessed as Gram-positive, and growth on both PetrifilmTM plates was assessed as Gram-negative bacterial growth. Additionally, milk samples were analysed by conventional microbiological diagnostic method on aesculin blood agar as a reference method. Overall, 616 samples from clinical mastitis cases were analysed. Using the reference method, Gram-positive and Gram-negative bacteria, mixed bacterial growth, contaminated samples and yeast were determined in 32.6%, 20.0%, 2.5%, 14.1% and 1.1% of the samples, respectively. In 29.7% of the samples, microbiological growth could not be identified. Using the PetrifilmTM concept, bacterial growth was detected in 59% of the culture-negative samples. The sensitivity of the PetrifilmTM for Gram-positive and Gram-negative MCP was 85.2% and 89.9%, respectively. The specificity was 75.4% for Gram-positive and 88.4% for Gramnegative MCP. For the culture-negative samples, sensitivity was 41.0% and specificity was 91.0%. The results indicate that the PetrifilmTM concept is suitable for therapeutic decisionmaking at the farm level or in veterinary practice. As this concept does not allow any statement about the genus or species of microorganisms, relevant MCP should be assessed periodically at the herd level with conventional microbiological diagnostics. © 2014 Elsevier B.V. All rights reserved.
1. Introduction ∗ Corresponding author at: Microbiology, Faculty of Mechanical and Bioprocess Engineering, University of Applied Sciences and Arts, Heisterbergallee 12, 30453 Hannover, Germany. Tel.: +49 511 9296 2205; fax: +49 511 9296 2210. E-mail address: [email protected] (V. Krömker). 0167-5877/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.prevetmed.2013.11.019
Mastitis remains one of the most frequent disease and reason for antibiotic use in dairy cattle. In Germany the economic loss due to mastitis is approximately 1.4 billion D each year (German Veterinary Association
E.M. Mansion-de Vries et al. / Preventive Veterinary Medicine 113 (2014) 620–624
[GVA], 2012). In order to ensure a maximum treatment effect along with a minimum investment, an evidence-based diagnosis and a prudent use of antimicrobials are fundamental. Therefore, it is necessary to know the mastitis-causing pathogen (MCP). Antimicrobial treatment of clinical mastitis is only recommended when a pathogen is detected (Roberson, 2003). Gramnegative coliform bacteria could allow an antibiotic-free treatment of the infected animals, as long as they are afebrile (Guterbock et al., 1993). Conversely, with regard to antibiotic resistance, therapy with a narrow-spectrum antibiotic is usually recommended for infections with Gram-positive microorganisms (Wilson et al., 1999; Lago et al., 2011). Therefore, a milk sample from each quarter with clinical mastitis should be examined bacteriologically (Lago et al., 2011). The current predominant method used is the conventional microbiological culture (National Mastitis Council (NMC), 1999; GVA, 2009), which costs about 2.50 D per milk sample investigation. Despite reliable and reproducible results, the standard microbiological diagnostic method has several weaknesses, such as a high exposure time of at least 48 h and a need for a lot of diagnostic experience (Krömker et al., 2007). An alternative fast, easy and very sensitive diagnostic test is the commercial PathoProof® Mastitis PCR Assay, a real-time PCR for identifying eleven mastitis pathogens (Koskinen et al., 2009). The investigation by PathoProof® has to be performed in a special laboratory, so time is lost on sample transportation. Additionally, the price of 18 D per milk sample investigation is high. These limitations illustrate the need for an easy, fast and cost effective diagnostic test that can ideally even be performed on a farm. A simple diagnostic test system is offered by the 3M Company. They produce several PetrifilmTM plates, which can be used to detect bacteria or bacterial groups with different sample-ready selective culture systems. Each PetrifilmTM plate costs about 1.50 D. These films contain a cold water soluble gelling agent, nutrients and an indicator. Authors of several studies have shown that PetrifilmTM plates are suitable for identifying mastitis pathogens (McCarron et al., 2009; Neuling et al., 2011; Gitau et al., 2013) and can be used to diagnose infections due to MCP (Krömker et al., 2007). In a laboratory setting, McCarron et al. (2009) found that parallel investigation of mastitic milk samples by AerobicCount (AC)-PetrifilmTM and ColiformCount (CC)-PetrifilmTM allows a differentiation between Gram-positive and Gram-negative bacteria. The authors used conventional microbiological diagnoses as the reference method and obtained a sensitivity of 93.8% and specificity of 70.1%. The aim of this study was to evaluate whether using two PetrifilmTM products in an on-farm setting (which conveys a higher degree of sample freshness but also bears a higher risk for contamination than working in a lab) could achieve comparable results compared to the conventional microbiological diagnosis method. Achieving these goals would provide a rapid, cost effective and useful diagnostic method to significantly facilitate treatment decisions.
621
2. Materials and methods 2.1. Samples Quarter milk samples of all animals with signs of clinical mastitis (i.e., clotting and/or discoloration of milk, udder swelling or redness, heat upon udder palpation) were obtained over an 11-month period at a single commercial dairy herd with 1000 lactating cows in Saxony-Anhalt (Germany). Immediately after detecting a clinical mastitis, a duplicate foremilk sample of the affected quarter was taken aseptically by the farmer. Samples were collected, as described by NMC (1999), and stored at a temperature between 2 ◦ C and 8 ◦ C. One of the duplicate samples (randomly allocated) was analysed on the farm within 12 h after collecting the sample using the AC- and CC-PetrifilmTM plates, whereas the second sample was transported to the microbiological Laboratory of the University of Applied Sciences and Arts Hannover (Germany) for conventional culture technique, twice a week. The laboratory samples were investigated within two hours after arrival by a trained blinded researcher through conventional microbiological diagnostic according to the GVA (2009). The data collection was planned before starting the study. 2.2. PetrifilmTM concept On the dairy farm, the examination of milk samples was performed with the AC- and CC-PetrifilmTM plates (3M, Neuss, Germany). The test was conducted following the manufacturer’s instructions. First, after vigorous shaking of the sample, a 1:10 dilution was prepared with sterile isotonic Ringer solution (Merck, Darmstadt, Germany), which was also thoroughly mixed. 1 ml of the dilution was added on each carrier film of both PetrifilmTM plates. After covering the sample with the top film it was spread out equally by the product-related spreader and, subsequently, incubated for 24 h at 37 ◦ C. After 24 h, colony growth was read with the aid of the PetrifilmTM integrated counting grid and the indicator dye. Enumeration and interpretation of the results were based on the recommendations of McCarron et al. (2009): A PetrifilmTM plate was categorised as positive when at least 20 colonies were counted on the CC-PetrifilmTM or five colonies were counted on the AC-PetrifilmTM . Based on these enumerations, a diagnosis was made (Table 1). The preparation and evaluation of the PetrifilmTM plates was conducted by a trained researcher at the farm. 2.3. Microbiological diagnosis For the microbiological diagnosis, 10 l of a wellmixed quarter foremilk sample were plated onto an aesculin blood agar plate (Oxoid, Wesel, Germany) and on a ChromoCult® Coliform Agar (Merck, Darmstadt Germany). The plates were incubated aerobically at 37 ◦ C and examined after 24 h and 48 h, according to the GVA recommendations (2009). The colonies were identified by Gram staining, cell morphology, aesculin hydrolysis and haemolysis patterns. Catalase and Gram-positive cocci were identified as staphylococci. Staphylococcus (S.) aureus
622
E.M. Mansion-de Vries et al. / Preventive Veterinary Medicine 113 (2014) 620–624
Table 1 Definition of the PetrifilmTM diagnosis (according to McCarron et al., 2009). CCa -PetrifilmTM
ACb -PetrifilmTM
Diagnosis
Number of diagnosis
Positive Negative Negative
Positive Positive Negative
Gram-negative pathogen Gram-positive pathogen No pathogen growth detected
218 368 144
a b
Coliform-count plate. Aerobic-count plate.
was defined as showing beta-haemolysis and having a positive clumping factor (Diamondial StaphPlus Kit, Diamondial, Sées, France) reaction. Other staphylococci were referred to as coagulase negative staphylococci (CNS), while Gram-positive, catalase negative cocci were identified as streptococci. Modified Rambach agar (Watts et al., 1993) was used to differentiate the aesculin hydrolysing cocci. -d-galactosidase-positive and aesculin hydrolysing cocci were identified as Streptococcus (Sc.) uberis, and aesculin hydrolysing, -d-galactosidase-negative cocci were identified as enterococci. Beta-haemolytic streptococci were further characterised by Lancefield serotyping (Diamondial Strep Kit, Diamondial, Sées, France). Streptococci from group C were referred to as Sc. dysgalactiae. Grampositive, asporogenic, beta-haemolytic, catalase-negative irregular rods with a V- or Y-shaped configurations were identified as Trueperella (T., formerly Arcanobacterium) pyogenes. Coliform bacteria were catalase negative, Gramnegative and cytochrome oxidase negative (Bactident oxidase, Merck, Darmstadt, Germany) rod-shaped bacteria that were capable of fermenting glucose (OF basal medium with addition of d [+]-glucose monohydrate, Merck, Darmstadt Germany). On ChromoCult® Coliform Agar (Merck, Darmstadt, Germany), Escherichia (E.) coli formed blue colonies, unlike other coliforms that formed pink-red colonies. Samples were categorised as positive when a minimum of three colonies grew. However S. aureus, Sc. dysgalactiae and T. pyogenes were identified if at least one colony was detected. If more than two different types of colonies were identified, a sample was considered to be contaminated. Even growth of S. aureus, Sc. dysgalactiae and T. pyogenes from contaminated samples were reported. Yeasts were differentiated microscopically after subculturing on YGC agar (Merck, Darmstadt, Germany). To compare the results with those of the PetrifilmTM method, they were summarised as Gram-negative MCP, Gram-positive MCP and no bacterial detection. Yeasts were numbered among Gram-positive MCP. If samples were contaminated, they were excluded from the evaluation of the PetrifilmTM . Samples with mixed growth which was not due to uniform Gram-positive or Gram-negative growth (not-uniformly mixed growth) were also excluded from evaluation.
2.4. Statistical analyses The data were recorded in Microsoft Access (Microsoft, Redmond, USA). The microbiological results of the PetrifilmTM method were compared with the results of the reference method. The PetrifilmTM results (Grampositive MCP, Gram-negative MCP or no bacterial detection) were considered false positives or false negative
Table 2 Numbers and percentages of mastitis causing pathogens isolated with the reference method from milk samples from 730 clinical mastitis cases recorded on one German farm with app. 1000 cows during an 11-month study period. Finding/pathogen
Number
Percentage (%)
No growth E. coli Other Gram-negative pathogens Sc. uberis S. aureus Sc. dysgalactiae T. pyogenes CNSa Other Gram-positive pathogens Mixed growthb Yeast Contaminated Total
217 129 17 107 35 28 19 19 30 18 8 103 730
29.7 17.7 2.3 14.7 4.8 3.8 2.6 2.6 4.1 2.5 1.1 14.1 100
a b
Coagulase negative staphylococci. Two pathogens were detected.
when the reference methods did not result in confirmation. The epidemiological software Win Episcope 2.0 (http://www.clive.ed.ac.uk/winepiscope/) was used to estimate sensitivity, specificity, positive and negative predictive value, Youden’s index, and the true and apparent prevalence, referring to the standard epidemiological methods of Kreienbrock and Schach (2000). 3. Results A total of 763 double quarter foremilk samples from German Holstein Friesian dairy cows with clinical mastitis were taken from 01 February 2012 to 31 December 2012. Because of either the reference method or the PetrifilmTM method results were lacking in 33 samples, these samples were excluded from the data analysis. Additionally, 114 samples were not used in the evaluation of the PetrifilmTM , because of contamination (n = 103) or nonuniformly mixed growth (n = 11). As a result, 616 double foremilk samples were included in the analysis. With the reference method, microbiological growth could not be identified in 217 samples (Table 2), whereas in 128 (59.0%) of these samples, the PetrifilmTM concept was able to detect bacterial growth. In 70.7% of the samples, the findings of the PetrifilmTM method were consistent with the results of the reference method (Table 3). The results of the statistical analyses are illustrated in Table 4. 4. Discussion In the present study, the diagnostic accuracy of the PetrifilmTM method was examined on the division of
E.M. Mansion-de Vries et al. / Preventive Veterinary Medicine 113 (2014) 620–624
623
Table 3 Contingency table for the comparison of the results of the reference method with the results of the PetrifilmTM concept. Milk samples from 616 clinical mastitis cases from one German farm with app. 1000 cows were included into the 11-month study. PetrifilmTM method
Reference method
a
Gram-negative MCP Gram-positive MCP No MCP detected Total
Gram-negative MCPa
Gram-positive MCP
No MCP detected
Total
134 8 46 188
8 213 82 303
7 29 89 125
149 250 217 616
Mastitis causing pathogen.
mastitis milk samples into no growth, Gram-negative and Gram-positive bacterial growth compared to the standard diagnostic method to clarify whether a treatment decision could be made based on results using the PetrifilmTM method. Consistently high sensitivities and specificities and both positive and negative predictive values were determined. Furthermore the results of the PetrifilmTM plates were obtained in less than 24 h after collection, much faster than the results from the conventional microbiological diagnostic method. Low sensitivity and a correspondingly moderate Youden’s index only occur when identifying no bacterial growth because the PetrifilmTM concept detected bacterial growth in 59% of the culture-negative samples. This result illustrates that the PetrifilmTM concept identified more MCPs compared to the standard microbiological diagnostic method, which could be explained by the 10-fold greater inoculum of 100 l, which increased the probability of pathogen detection (Sears et al., 1990; Hogan and Smith, 2003). In this way lacking diagnostic sensitivity of the reference method negatively affects the outcome of the PetrifilmTM method. It is also possible to increase the detection of MCP’s in the conventional microbial diagnostic by increasing the inoculum (Krömker et al., 2010). However, in Germany the official required test method is the conventional microbial diagnostic with 10 l inoculum, so the PetrifilmTM method had to be compete with these standards. Furthermore, faster sample processing could have influenced the rate of negative samples. The PetrifilmTM plates were prepared immediately after collecting the samples, with no more than a 12-h delay after the collection. In contrast, because of the need for only portage of the reference method samples, at least one 24-h period is expected from the collection time to arrival at the laboratory. The negative predictive value verifies whether a result is truly negative when the test has declared it negative. The negative predictive value of the PetrifilmTM method were between 73.9% (in the
identification of no pathogen growth) and 96.5% (in the detection of Gram-negative MCP), which indicates the benefits of the method. The sensitivity and specificity values that were calculated in the present study were comparable to the results of McCarron et al. (2009) and Gitau et al. (2013). In both studies, the preparation of the PetrifilmTM plates was performed in the laboratory. McCarron et al. (2009) investigated also the interreader agreement on interpreting the PetrifilmTM results. They state that the agreement between readers is very good (kappa around 0.85) in spite of different microbiological training. Still, tests were performed by one and the same technician as a courtesy for the farmers’ will to cooperate. Despite high consensus when comparing the results of the conventional microbiological diagnostic method and the PetrifilmTM concept and many benefits of the PetrifilmTM plates, even a few difficulties in the implementation process must be noted. The authors of the present study estimate that the greatest weakness is the lack of information about the quality of the milk samples. A statement of whether the milk sample was collected aseptically or was contaminated is impossible when using the PetrifilmTM technique. Identification of mixed infections is also excluded because of the growth of Gram-negative MCP on both the AC and the CC-PetrifilmTM . The clinical impact of the inability of the PetrifilmTM method to identify mixed infections is low, since only 2–3% of the samples are usually mixed infections (Olde Riekerink et al., 2008; Gitau et al., 2013), as was the case in the present study. Yeast does only grow on the AC PetrifilmTM . In most cases they are not detectable within 24 h of incubation because they grow very slowly. In our study 6 out of 8 milk samples with yeast results in the reference method had no pathogen detection with the PetrifilmTM method (data not shown). The clinical impact of the determination yeast as Grampositive MCP in the PetrifilmTM method is also low, since only 1–2% of the samples are usually yeast infections (Olde
Table 4 Test characteristics of the PetrifilmTM concept based on its ability to correctly classify the bacterial growth (classes: Gram-positive, Gram-negative, no growth) from milk samples from 616 clinical mastitis cases recorded on one German farm with app. 1000 cows during an 11-month study period. Test characteristic
Gram-negative MCPa (%)
Gram-positive MCP (%)
No bacterial growth (%)
Sensitivity (95% CI) Specificity (95% CI) Positive predictive value (95% CI) Negative predictive value (95% CI) Youden’s index (95% CI) True prevalence (95% CI) Apparent prevalence (95% CI)
89.9 (85.1–94.8) 88.4 (85.5–91.3) 71.3 (64.8–77.7) 96.5 (94.8–98.2) 0.78 (0.73–0.84) 24.2 (20.8–27.6) 29.9 (26.9–34.2)
85.2 (80.8–89.6) 75.4 (71.0–79.8) 70.3 (65.2–75.4) 88.2 (84.6–91.8) 0.61 (0.54–0.67) 40.6 (36.7–44.5) 49.2 (45.2–53.1)
41.0 (34.5–47.6) 91.0 (88.2–93.8) 71.2 (63.3–79.1) 73.9 (70.0–77.8) 0.32 (0.25–0.39) 35.2 (31.5–39.0) 20.3 (17.1–23.5)
a
Mastitis causing pathogen.
624
E.M. Mansion-de Vries et al. / Preventive Veterinary Medicine 113 (2014) 620–624
Riekerink et al., 2008), as was the case in the present study. Furthermore, incubated PetrifilmTM plates cannot be easily disposed of through the municipal waste. They must be autoclaved prior to properly disposing of them. With the implementation of PetrifilmTM plates on a dairy farm, the use of a steam cooker would be conceivable. Despite these weaknesses the present study showed that the PetrifilmTM achieved good results. The sensitivity and NPV of Gram-negative detection was as excellent as established by Leslie et al. (2005). If treatment of Gramnegative MCP will be done without antibiotics the in-field PetrifilmTM method could help to reduce antibiotic application by approximately 20%. 5. Conclusion The study results lead to the assessment that the present PetrifilmTM method can be used for classifying samples from clinical mastitis cases with Gram-positive and Gramnegative MCPs or samples without growth; therefore, it is suitable for therapeutic decision making at the farm level or in veterinary practice. However, this method only allows the determination of Gram-positive or Gramnegative pathogens, without providing any information about the species; therefore, standard diagnostics may still be needed. To monitor the prevalence of MCP’s at the herd level, conventional microbiological diagnostics should be performed because of its greater discriminative potential. Because changes in the prevalence of MCP are possible, this method should be periodically evaluated. Furthermore, note that reliable results can be achieved only if the sampling is performed aseptically. Conflict of interest statement None declared. Acknowledgment The authors thank Ohreland KG for providing their cows and Vrieswoud KG to enable us to carry out the study. Furthermore we are grateful to 3M for supplying the PetrifilmTM plates. References German Veterinary Association (GVA), 2009. Leitlinien zur Entnahme von Milchproben unter antiseptischen Bedingungen und Leitlinien
zur Isolierung und Identifizierung von Mastitiserregern. GVA, Gießen, Germany. German Veterinary Association (GVA), 2012. Leitlinien zur Bekämpfung der Mastitis des Rindes als Herdenproblem. GVA, Gießen, Germany. Gitau, G.K., Bundi, R.M., Vanleeuwen, J., Mulei, C.M., 2013. Evaluation of PetrifilmsTM as a diagnostic test to detect bovine mastitis organisms in Kenya. Trop. Anim. Health Prod. 45, 883–886. Guterbock, W.M., Van Eeenennaam, A.L., Anderson, R.J., Gardner, I.A., Cullor, J.S., Holmberg, C.A., 1993. Efficacy of intramammary antibiotic therapy for treatment of clinical mastitis caused by environmental pathogens. J. Dairy Sci. 76, 3437–3444. Hogan, J., Smith, K.L., 2003. Coliform mastitis. Vet. Res. 34, 507–519. Kreienbrock, L., Schach, S. (Eds.), 2000. Epidemiologische Methoden. Spektrum Akademischer Verlag, Heidelberg, Germany. Koskinen, M., Holopainen, J., Pyörälä, S., Bredbacka, P., Pitkälä, A., Barkema, H., Bexiga, R., Roberson, J., Sølverød, L., Piccinini, R., Kelton, D., Lehmusto, H., Niskala, S., Salmikivi, L., 2009. Analytical specificity and sensitivity of a real-time polymerase chain reaction assay for identification of bovine mastitis pathogens. J. Dairy Sci. 92, 952–959. Krömker, V., Hauptmann, T., Bormann, A., 2007. Bacteriological diagnosis of mastitis pathogens in quarter milk samples and bulk milk samples with petrifilm. Dtsch. Tierärztl. Wschr. 114, 378–383. Krömker, V., Paduch, J.-H., Klocke, D., Zinke, C., 2010. Microbiological investigation of culture negative milk samples of clinical mastitis cows. Milchwissenschaft 65, 123–126. Lago, A., Godden, S.M., Bey, R., Ruegg, P.L., Leslie, K., 2011. The selective treatment of clinical mastitis based on on-farm culture results: I. Effects on antibiotic use, milk withholding time, and short-term clinical and bacteriological outcomes. J. Dairy Sci. 94, 4441–4456. Leslie, K., Walker, M., Vernooy, E., Bashiri, A., Dingwell, R., 2005. Evaluation of the PetrifilmTM culture system for the identification of mastitis bacteria as compared to standard bacteriological methods. In: Hogeveen, H. (Ed.), Mastitis in Dairy Production: Current Knowledge and Future Solutions, Mastitis in Dairy Production: Current Knowledge and Future Solutions. Wageningen Academic Publishers, Wageningen, pp. 416–421. McCarron, J.L., Keefe, G.P., McKenna, S.L., Dohoo, I.R., Poole, D.E., 2009. Laboratory evaluation of 3M Petrifilms and University of Minnesota Bi-plates as potential on-farm tests for clinical mastitis. J. Dairy Sci. 92, 2297–2305. National Mastitis Council (NMC), 1999. Laboratory Handbook on Bovine Mastitis. NMC, Madison, WI. Neuling, R., Paduch, J.-H., Bormann, A., Zinke, C., Krömker, V., 2011. Evaluation of bacteriological diagnosis in samples of clinical mastitis cases with PetrifilmTM in veterinary practice. Tierarztl. Prax. Ausg. G Grosstiere Nutztiere 39, 77–80. Olde Riekerink, R.G.M., Barkema, H.W., Kelton, D.F., Scholl, D.T., 2008. Incidence rate of clinical mastitis on Canadian dairy farms. J. Dairy Sci. 91, 1366–1377. Roberson, J.R., 2003. Establishing treatment protocols for clinical mastitis. Vet. Clin. Food Anim. 19, 223–234. Sears, P.M., Smith, B.S., English, P.B., Herer, P.S., Gonzalez, R.N., 1990. Shedding pattern of Staphylococcus aureus from bovine intramammary infections. J. Dairy Sci. 73, 2785–2789. Watts, J.L., Salmon, S.A., Yancey Jr., R.J., 1993. Use of modified Rambach agar to differentiate Streptococcus uberis from other mastitis streptococci. J. Dairy Sci. 76, 1740–1743. Wilson, D.J., Gonzalez, R.N., Case, K.L., Garrison, L.L., Gröhn, Y.T., 1999. Comparison of seven antibiotic treatments with no treatment for bacteriological efficacy against bovine mastitis pathogens. J. Dairy Sci. 82, 1664–1670.
