• Open Access

Validation of a Commercially Available Immunoassay for the Measurement of Bovine Cardiac Troponin I

Authors


  • Presented in part at the annual Forum of the ACVIM, San Antonio, TX, June 4–7, 2008.

Corresponding author: Karsten Schober, DVM, The Ohio State University, 601 Vernon L. Tharp, 43210 Columbus, OH. e-mail: schober.4@osu.edu

Abstract

Background: Commercially available cardiac troponin I (cTnI) assays developed for use in humans have not yet been validated for use in cattle.

Hypotheses: The ADVIA Centaur TnI-Ultra immunoassay can be used for the detection of bovine cTnI. In healthy cattle, serum cTnI is undetectable or is present only in trace amounts.

Methods: Purified bovine cTnI and cTnI-free bovine serum were used for the evaluation of assay performance including intra- and inter-assay precision, sensitivity, interference, linearity, and recovery. Effects of storage at 23, 4, −20, and −80 °C for 2 days, and at −20 and −80 °C for 7 and 14 days and repeated freeze-thaw cycles on recovery of cTnI were analyzed. Serum cTnI concentrations in 30 healthy dairy cows were determined.

Results: Intra- and inter-assay precisions (mean ± SD) were 4.48 ± 2.26 and 13.36 ± 6.59%, respectively. The assay demonstrated linearity at 0.5, 2, 15, and 30 ng/mL cTnI. Mean recovery was 100.81, 85.26, 87.72, and 114.42%, respectively. Skeletal muscle homogenate added to serum of known cTnI concentration did not alter the concentration of the analyte (P > .05). Concentration of cTnI significantly decreased when samples were stored at 4 and 23 °C for 2 days (P < .05). Repeated freeze-thaw cycles and storage at −20 °C for 7 days had no significant influence on cTnI concentration (P > .05). Serum cTnI concentration in healthy cattle was ≤0.03 ng/mL.

Conclusion and Clinical Importance: ADVIA Centaur can be used reliably for the detection of serum cTnI concentration in cattle.

Analysis of cardiac troponin I (cTnI) is considered the “gold standard” for noninvasive diagnosis of myocardial injury in humans.1,2 It has replaced previously used cardiac biomarkers such as the MB isoenzyme of creatine kinase (CK-MB) because of its unique myocardial sensitivity and specificity. Increased concentrations of circulating cTnI were reported in dogs with acute myocardial damage,3 cardiac contusion,4 cardiomyopathy,5 babesiosis;6 in cats with hypertrophic cardiomyopathy;7,8 and in horses with myocardial disease.9,10

In cattle, increased blood concentrations of cTnI were reported with idiopathic pericarditis,11 traumatic reticuloperitonits,12,13 in a case of suspected foot-and-mouth disease,14 and in calves with experimentally induced endotoxemia.15 In such studies, various immunoassays, developed for use in humans, were used for the measurement of cTnI, but none had undergone prior validation for use in cattle. In the past, the diagnosis of myocardial injury in cattle has been made on the basis of results of physical examination; cardiac auscultation; and, more specifically, by radiography, electrocardiography, and echocardiography. However, diagnostic modalities such as radiography and cardiac ultrasound may not be readily available to the veterinarian. Therefore, easy-to-perform, reliable, and specific tests to detect myocardial injury in cattle are needed.

Bovine cTnI has high amino acid sequence homology (> 96%) with human cTnI.16 Therefore, it is anticipated that antibodies against human cTnI used in commercially available immunoassays will cross-react with bovine cTnI. The objective of the present study was to evaluate the analytical performance of the ADVIA Centaur TnI-Ultra immunoassaya for the detection of circulating bovine cTnI and to determine a reference range for serum cTnI concentration in healthy cows. We hypothesized that the human immunoassay would have sufficient test performance for use in cattle and that cTnI would not be detectable or occur only at very low concentrations in serum of healthy cows as previously demonstrated in healthy cows and calves by a different immunoassay.11,15

Materials and Methods

This study was approved by the Institutional Animal Care and Use Committee (IACUC) at the Ohio State University, Columbus, OH.

Because control bovine serum was not commercially available at the time of investigation, such serum was generated by the investigators. Two hundred milliliters of blood from a healthy steer (2 years old, 420 kg BW) was collected and serum was separated and used as control serum for all measurements during the study. Absence of detectable cTnI (<0.01 ng/mL) was assessed by analyzing a serum sample in duplicate.

Commercially available purified bovine cTnIb was used to spike cTnI-free serum and generate control serum of known cTnI concentration. Serum samples were immediately frozen and stored at −20 °C unless otherwise indicated. To assure expected concentrations of cTnI, the frozen samples were sent to a commercial laboratory for analysis within 48 hours. After arrival at the laboratory, the samples were thawed at room temperature and immediately analyzed. Calibration of the analyzer was performed with human cTnI standards.c The assaya used for all cTnI analyses is a 3-site sandwich immunoassay by direct chemiluminometry with 1 polyclonal goat antitroponin I antibody labeled with acridium ester and 2 biotinylated mouse monoclonal antitroponin I antibodies for detection of cTnI. The capture antibodies recognize amino acid sequences 41–49 and 87–91 located in the stable central region of the cTnI molecule.17 Magnetic latex particles conjugated with streptavidin are used as the solid phase reagent. The antibodies bind to cTnI in the sample and biotin as part of the immune complex and then bind to the streptavidin-labeled magnetic particles. A chemiluminescent reaction is initiated and a direct relationship exists between the amount of cTnI in the sample and the amount of relative light units detected by the assay.17 A sample volume of at least 100 μL is required.

Test performance of the assay has been evaluated previously in humans.14,18 Confirmation of similar assay characteristics with bovine cTnI was performed by evaluation of the assay's precision, sensitivity, interference with skeletal muscle, linearity, and recovery. Control samples of 4 different concentrations of cTnI (0.2, 1, 10, and 30 ng/mL) were used. Intra-assay precision (within-run) was evaluated by analyzing all samples 3 times in the same run within 1 day, and inter-assay precision (between-run) was evaluated by analyzing all samples twice each day for 3 consecutive days. For evaluation of the lower limit of detection of the immunoassay (ie, sensitivity of the assay) samples with concentrations of approximately 0.5, 0.1, 0.01, and 0.001 ng/mL cTnI were generated. Each concentration was assayed in 3 replicates to determine the lowest measurable cTnI concentration. Assay linearity was evaluated using serum with 4 different concentrations of cTnI (0.43, 2.33, 14.25, and 29.20 ng/mL). Serial dilutions to obtain 80, 60, 40, 20, and 10% of the original concentration were performed in each of the 5 samples of known cTnI concentration. Test recovery (in percent) was determined for each dilution as comparison of expected versus measured cTnI. Interference with skeletal muscle troponin I was determined by spiking a serum sample with known cTnI concentration with a skeletal muscle homogenate. To obtain the latter, a 50 g skeletal muscle block (Mm biceps femoris) was harvested from a cow humanely euthanized after a femoral fracture, and frozen at −70 °C within 2 hour of collection until further processing. Muscle tissue from the nonaffected limb was used. A 1.5 g sample of the frozen muscle block was removed and ultrafrozen in liquid nitrogen (−196 °C) for 2 minutes. Thereafter, the tissue was manually macerated and mixed with 6 mL of phosphate buffered saline.d A tissue tearore was used to break the skeletal muscle tissue down further over a time period of 2 minutes at 30,000 rpm. Finally, a Dounce homogenizatorf was used to homogenize the remainder of the muscle tissue. The resulting homogenate was centrifugedg at 2,800 rpm at 4 °C for 10 minutes, and the supernatant was collected for further analysis. Three different volumes of skeletal muscle homogenate (10, 100, and 200 μL) were added to 1 mL serum with a cTnI concentration of 0.89 ng/mL. Interference of the assay with skeletal muscle was evaluated by comparing baseline concentrations with the cTnI concentration of the mixed sample. Stability of the analyte in serum was determined by analysis of serum samples with low (0.5 ng/mL), medium (5 ng/mL), and high (20 ng/mL) cTnI concentrations stored at different temperatures over different time periods. The samples were stored for 48 hours at −80, −20, 4 °C, and room temperature (23 °C) as well as for an additional 7 and 14 days at −80 and −20 °C. Recovered cTnI was compared with cTnI concentration before storage, which were immediately analyzed after preparation. The effect of repeated freeze-thaw cycles on cTnI recovery (in percent) with samples with low (0.44 and 0.67 ng/mL) and medium (5.28 and 4.22 ng/mL) concentrations of cTnI stored at −80 and −20 °C, respectively, was evaluated. The serum samples were frozen for at least 24 hours before they were thawed over 30 minutes at room temperature. After completing the defrosting process, the samples underwent 2 additional freeze-thaw cycles on the same day. After each cycle, cTnI was analyzed and compared with baseline concentrations.

Thirty healthy dairy cows (26 Holstein, 4 Jersey) were used for the generation of reference values of cTnI. Inclusion criteria were a normal history and physical examination including cardiac and pulmonary auscultation. Twenty Holstein cows were pregnant (between 40 and 215 days) and their average daily milk yield was 24.5 ± 6 kg. The remaining cows were in the dry-off period. The mean estimated body weight was 544 ± 71 kg based on body condition score and height of the animal assessed by 2 independent investigators and averaged. Twenty-five cows were <5 years old and 5 cows were >5 years old. A 10 mL blood sample was drawn from either the jugular or the coccygeal vein into a serum vacutainer.h Blood was left at room temperature for a maximum of 45 minutes before tubes were centrifugedg at 2,800 rpm for 20 minutes at 23 °C. Serum was then removed, separated into 2 aliquots, and frozen at −20 °C within 4 hours of collection. Samples were analyzed for cTnI within 2 days of collection.

Statistical Analysis

Statistical analyses were performed by commercially available software.i,j Precision was calculated as coefficient of variation (CV %= SD/average × 100%). Linearity of serial dilutions at different cTnI concentrations was determined by linear regression analysis. Linearity of the assay was assumed when the correlation coefficient was >0.95.19 Recovery (in percent) was calculated (obtained cTnI concentration/expected cTnI concentration × 100%). Kruskal-Wallis analysis of variance (ANOVA) was performed to determine interference with skeletal muscle homogenate as well as to determine differences induced by temperature, storage time, and cTnI concentration. A post hoc Mann-Whitney test was used to evaluate significant differences between all pairs of duration of storage on cTnI concentrations recovered. One-way ANOVA followed by a posthoc Tukey's test was used to determine significant differences in the recovery of cTnI caused by multiple freeze-thaw cycles. In the majority of healthy cows, the cTnI concentrations found were less than the detection limit of the assay. Therefore, LIFEREG procedure was used to handle such censored data and to estimate the 99th percentile as upper limit of the reference range. Cardiac troponin I concentrations reported as < 0.01 ng/mL were handled as interval-censored data for statistical purposes. Effect of age, body weight, lactation, and pregnancy status of the cows was evaluated by multiple regression analysis accounting for censored data. Results are expressed as mean ± standard deviation (SD) unless otherwise stated. Statistical significance was defined as P < .05.

Results

Intra-assay precision over the range 0.2–30 ng/mL cTnI was between 3.09 and 4.84% with a mean CV of 4.78 ± 2.26%. The interassay precision over the same range of cTnI concentrations was between 5.41 and 21.01% with a mean of 13.36 ± 6.59% (Table 1). Assessment of the lower limit of detection of the immunoassay revealed undetectable cTnI concentrations for all samples with an estimated concentration of < 0.01 ng/mL. The lowest concentration measured in this experiment was 0.13 ng/mL (Table 2). Regression analysis indicated good linearity for all sets of serial cTnI dilutions (Fig 1). The slope of the regression line ranged from 1.006 to 1.019 and the correlation coefficient from 0.98 to 0.994 for obtained versus expected cTnI concentrations. The average percent recovery for each dilution series was 100.81, 85.26, 87.72, and 114.42% with a mean percent recovery of 96.90 ± 16.59% for all dilutions. Recovered cTnI in serum samples containing 0.89 ng/mL of cTnI spiked with 10, 100, and 200 μL skeletal muscle homogenate was 0.98, 0.94 and 0.98 ng/mL, respectively (P= .188). There was no effect of the quantity of homogenate added on the cTnI concentration recovered. All troponin I concentrations decreased from baseline concentration when stored at 23 and 4 °C for 48 hours (P < .05). A significant decrease in recovery was observed when samples were stored at −80 °C (P < .001) for the same time period. However, cTnI concentrations were not affected when stored at −20 °C (P > .05). Storage temperature had a significant impact on the recovery of cTnI (P < .001; Table 3). Storage of samples at −20 °C for 7 days had no significant effect on the recovery of cTnI. Storage for 14 days at the same temperature resulted in a significant decrease in the recovered cTnI concentration (P= .001). When samples were stored at −80 °C for 7 and 14 days, no significant difference in recovery was observed (P > .05; Table 3).

Table 1.   Intra- and interassay precision at different cTnI Concentrations.
cTnI Concentration
(ng/mL)
Intra-assay
CV (%)
Interassay
CV (%)
  1. cTnI, serum cardiac troponin I concentration; CV (%), coefficient of variation in percent.

0.23.095.42
1.03.2511.44
10.07.9521.01
30.04.8415.56
Table 2.   Serial dilution of serum cTnI for assessment of the lower limit of detection of the immunoassay (analyzed in triplicates).
cTnI (ng/mL)Replicate 1Replicate 2Replicate 3
  1. ND, not detectable; cTnI, serum cardiac troponin I concentration.

0.50.450.470.46
0.10.130.140.13
0.01NDNDND
0.001NDNDND
Figure 1.

 Plots of observed versus expected concentrations of cardiac troponin I (cTnI) after serial dilution. (A) baseline cTnI concentration 0.43 ng/mL; (B) baseline cTnI concentration 2.33 ng/ml; (C) baseline cTnI concentration 14.25 ng/mL; (D) baseline cTnI concentration 29.20 ng/mL; Samples were run in duplicates. P<.0001 at all concentrations.

Table 3.   Effect of storage on recovery of cTnI.
StorageAverage Recovery (%)
Temperature (°C)Time (days)cTnI (ng/mL)
0.5520
  1. cTnI, serum cardiac troponin I concentration.

22–23262.9251.5443.22
4280.8979.0377.64
−202100.097.25101.0
−80797.7696.5997.39
1484.3385.3290.58
289.8986.8582.92
7105.60104.093.30
14107.90102.3092.13

Repeated freeze-thaw cycles had no significant influence on the cTnI recovery (P > .05). After 1, 2, and 3 freeze-thaw cycles, the recoveries at low cTnI concentration (0.5 ng/mL) at −20 °C were 97.76, 96.27, and 96.21%, respectively, and at −80 °C were 105.6, 110, and 116.9%, respectively. At medium concentration (5 ng/mL), the recovery of cTnI at −20 °C was 96.88, 94.22, and 96.97% and at −80 °C were 102.8, 105.7, and 113.6%, respectively.

In healthy cows, all serum cTnI concentrations were ≤ 0.03 ng/mL. Twenty-three cows had cTnI concentrations ≤ 0.01 ng/mL (Fig 2). Assuming the measured cTnI concentration followed an exponential distribution, the mean cTnI concentration in healthy cows was 0.02 ng/mL and the 99th percentile was 0.07 ng/mL. Variables such as age, body weight, lactation, and pregnancy status had no effect on baseline cTnI concentration (P > .05).

Figure 2.

 Point plot of serum cardiac troponin I (cTnI) concentrations in healthy dairy cows (n = 30). The line in the center represents the mean value.

Discussion

To the author's knowledge, this is the first report on the validation of a commercially available cTnI assay for the detection of bovine cTnI. Our results indicate that the ADVIA Centaur TnI-Ultra immunoassay can be used for analysis of bovine cTnI. The immunoassay had similar analytical performance to that observed in previous studies of small animals and horses3,20–22 and performs with adequate precision, linearity, and recovery across different concentrations of purified bovine cTnI. The intra-assay (within-run) precision was <10%, which is similar to that reported in studies in humans with the same assay.23 The interassay precision (between-runs) at lower cTnI concentration (0.2 ng/mL) also was <10%. However, at higher cTnI concentrations (1.0, 10.0, and 30 ng/mL), precision decreased. For clinical acceptance of assay performance, the International Federation of Clinical Chemistry and Laboratory Medicine recommends a CV<10%.24,25 This recommendation was made to obtain correct risk classification in humans with myocardial infarction.26 High assay precision is of particular importance if low cutoff values (<0.01 ng/mL) are used for the detection of myocardial injury thus providing an earlier identification of patients at risk. Although we observed an imprecision above such recommendations at higher cTnI concentrations, its clinical significance for cattle may not be as important. To the author's knowledge, the relationship between the magnitude of myocardial injury and concentration of circulating cTnI in cattle has not yet been reported.

The manufacturer reports a low (<0.007%) cross-reactivity for the ADVIA Centaur immunoassay when skeletal muscle troponin I is added to a solution with known concentration of cTnI in humans. Capture antibodies used in commercially available cTnI assays have negligible cross reactivity with skeletal muscle troponin.19,27,28 We did not observe a clinically relevant increase in cTnI concentration when various amounts of skeletal muscle homogenate were added. The increase in cTnI concentration was not proportionate to the volume of skeletal muscle homogenate added. The mild increase in cTnI concentration observed may be because of interference with other substrates such as myoglobin, muscle proteins, or lipids. Based on our results, we conclude that concurrent skeletal muscle disease will not limit the use of the ADVIA Centaur immunoassay in cattle with myocardial injury.

The assay demonstrated excellent linearity and recovery for all sets of serial dilutions for the detection of bovine cTnI. Cardiac TnI can occur in circulation after myocardial injury as free, binary (complex of cTnI with cTnC), as well as ternary forms (complex of cTnI with cTnC and cTnT).29,30 The troponin complex released after injury undergoes rapid posttranslational modification. The type of myocardial insult determines which forms of circulating cTnI dominate.31 Cardiac troponin I is susceptible to proteolysis by serum proteases as well as phosphorylation and oxidation after it is released into the bloodstream. This leads to conformational changes and a variety of peptides with different stabilities. The rate of degradation of cardiac troponin depends on different factors such as size of the cTnI fragments released and their complex formation with other troponin subunits.29,30 In humans, free cTnI has a lower stability than the binary or ternary forms. The N-terminal as well as C-terminal regions of troponin I are rapidly cleaved during proteolysis. Therefore, using an immunoassay that employs antibodies against the stable core region of the protein will increase analytical performance in comparison with assays using antibodies recognizing either the C- or N-terminus of cTnI. The stable core region lies between amino acid residues 30 and 110.32 The ADVIA Centaur immunoassay uses antibodies that recognize immunologic epitopes located within this stable region.

In humans, the majority (>97%) of cTnI released is complexed with TnC after myocardial infarction.33,34 Free cTnI has a very short half life after release (approximately 5 min) and occurs only in small quantities in circulation.35 Other minor fractions of released troponins are the TnI-TnT binary complex and the TnI-TnC-TnT ternary complex.36 Shi et al36 reported that some assaysk,l,m preferentially recognized cTnI in complex form over free cTnI in humans. Considerable differences in cTnI recovery among various commercially available immunoassay methods may occur34 mainly because of the presence of various forms of cTnI in the sample and the use of capture antibodies with different affinities for free and complexed forms of cTnI.37 In cattle, it is unknown which of the several forms of cTnI is released with regard to a particular myocardial insult. In our study, we only validated the assay with free bovine cTnI. Therefore, our results are limited because conclusions about the recovery of complex isoforms of cTnI by the immunoassay cannot be drawn.

We detected a significant decrease in cTnI recovery when cTnI samples were stored at room temperature and at 4 °C for 2 days. These findings are similar to previous studies in laboratory animals and humans.19,38–40 Storage at −20 °C for 7 days had no significant effect on cTnI recovery. In contrast, a significant, although possibly clinically irrelevant, decrease in cTnI recovery was noticed after storage for 2 days at −80 °C. However, the same batch did not experience a significant decrease in cTnI when stored at the same temperature for 7 and 14 days. Interassay variance or possible laboratory error are the likely reasons for the latter finding. Further investigation by means of a larger sample size is needed to fully understand the effect of storage on cTnI recovery. Our results suggest that cTnI measurements are best performed on the day of sampling or with samples stored at −20 °C for a brief period of time.

All cTnI concentrations determined in clinically healthy dairy cows were ≤0.03 ng/mL, resembling reported concentrations by others in cattle.11,13,15 The majority of cows had a serum cTnI concentration at or below the lower limit of detection of the assay used. Therefore, we conclude that cTnI in healthy dairy cows should not be detectable or occur only in trace amounts. Because of the lack of standardization of numerous commercially available cTnI assays, results of this study are only valid for cTnI measurements by the ADVIA Centaur TnI-Ultra.41 Different manufacturers utilize different capture and proprietary antibodies with differing abilities to detect free or complexed forms of cTnI, resulting in significant and clinically relevant differences among cTnI analyzers with the same control serum.42 Therefore, results of studies with different cTnI assays cannot be directly compared without bias. Lactation and pregnancy status, age, and weight of the cows did not influence serum cTnI. However, additional studies involving a larger number of cows and groups of different breeds, sex, and body weights are required to determine potential confounding effects on the concentrations of circulating cTnI concentration.

The clinical use of the immunoassay in cattle with naturally occurring myocardial disease was not determined. Mellanby et al13 reported a significant difference in circulating cTnI concentrations in 4 of 5 cows with reticulopericarditis when compared with a healthy control group (n = 34) by the Immulite troponin I immunometric chemiluminescent assay system.n Unfortunately, this study did not correlate histopathologic evidence of myocardial injury to serum cTnI concentrations. Studies in humans43 and rats44 indicated good correlation between the severity of morphologically detectable myocardial cell damage and blood troponin I concentrations. Additional investigations are needed to confirm similar findings in cattle.

Circulating cTnI is substrate (myocardium) specific, but lacks specificity for the particular type of cardiac and noncardiac disease. Various studies in humans45,46 and small animals47 reported on increased cTnI concentrations in patients with renal failure, sepsis, blunt chest trauma, hyperthyroidism, gastric dilatation-volvulus, and diabetic ketoacidosis. Also, strenuous exercise may be associated with increased cTnI.45,46 However, the effect of extracardiac disease on circulating concentrations of cTnI is not well documented in cattle although endotoxemia was found to be related to plasma cTnI concentration in calves.15 Additional studies are needed to investigate the relationship between circulating cTnI and systemic disease.

Certain limitations of this study need emphasis. One limitation was that only free bovine cTnI was used for the assessment of analytical performance of the ADVIA Centaur immunoassay. Information on assay performance with complex forms of cTnI was not obtained, and, therefore, no information with regard to analytical differences between free and complexed bovine cTnI could be obtained. Another limitation relates to the assessment of the lower limit of detection. The magnitude of dilutional steps used was possibly too large to detect minor differences in cTnI concentrations. Therefore, the lower limit of detection obtained may be an underestimation of the true limit. Sample size was small leading to low statistical power to detect significant differences. Moreover, based on statistical recommendations, reference values should be established from at least 120 independent observations.47 We only used 30 cows in our study. Furthermore, we did not test the upper limit of detection of the assay. Moreover, we only used custom-made standards for the measurement of cTnI because standards for cattle were not commercially available.

In summary, the ADVIA Centaur TnI-Ultra immunoassay has sufficient analytical performance for the measurement of bovine cTnI. The assay reliably detects free cTnI, but additional studies are needed to evaluate the ability of the immunoassay to detect complex cTnI isoforms. Storage of serum affects recovery of the analyte. Storage at −20 °C is recommended if serum cannot be analyzed within a few hours of sampling. Results of this study are similar to those of previous studies in other species.16,19–21 but cannot necessarily be extrapolated to results of studies with different cTnI assays for the detection of bovine cTnI. Additional research on the evaluation of circulating cTnI in cows with naturally occurring myocardial injury secondary to viral or bacterial infections, traumatic reticuloperitonitis, glycoside and ionophore toxicity, and nutritional deficiencies affecting myocardial integrity and function as well as the effect of noncardiac diseases on circulating cTnI concentrations are needed.

Footnotes

aADVIA Centaur TnI-Ultra, Siemens Medical Solutions Diagnostics, NY

bBiosPacific, Emeryville, CA

cADVIA Centaur Calibrator UL, Siemens Medical Solutions Diagnostics

dPhosphate Buffer Solution, Fisher Scientific, Pittsburgh, PA

eTissue-tearor, Fisher Scientific

fDounce homogenizer, Fisher Scientific

gSorvall Legend T/RT, Fisher Scientific

hBD, Franklin Lakes, NJ

iMinitab 15.1, Minitab Inc, State College, PA

jSAS 9.1, SAS Institute Inc, Cary, NC

kStratus II, Dade International, Miami, FL

lOpus, Behring Diagnostics Systems, Westwood, MA

mAccess, Beckman Coulter Inc, Brea, CA

nImmulite, Siemens Medical Solutions Diagnostic, Los Angeles, CA

Acknowledgment

The study was supported by United States Department of Agriculture (USDA) Animal Health Formula Funds provided through the Ohio State University.

Ancillary