| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Department of Large Animal Sciences, Faculty of Life Sciences, University of Copenhagen, Grønnegårdsvej 8, DK-1870 Frederiksberg C, Denmark
Danish Meat Association, Vinkelvej 11, 8620 Kjellerup, Denmark
1 Corresponding author: ssn{at}life.ku.dk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Key Words: paratuberculosis • robust regression • somatic cell count
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Younger animals are more susceptible to infection with MAP (Doyle, 1953), although clinical signs generally do not appear until adulthood or may never develop (Chiodini, 1996). During the incubation period, asymptomatic animals can shed MAP into the environment, resulting in further transmission of infection in the herd. Therefore, for each clinically diseased animal detected by the producer, there might be 15 to 20 animals infected within the herd (Whitlock and Buergelt, 1996). For an efficient MAP control strategy, repeat testing and establishment of management practices for the protection of susceptible calves are needed to break transmission routes (Groenendaal et al., 2002).
An increased susceptibility to other infections, such as mastitis, after MAP infections has been suggested (Merkal et al., 1975). Cows can become infected with mastitis pathogens and develop subclinical or clinical mastitis, which depends on several factors, such as management practices, environmental factors, and the condition of the immune system (Smith et al., 1985). Somatic cell count is a measure of the number of leukocytes and epithelial cells per milliliter of milk. Increased SCC is indicative of mammary gland inflammation, and if associated with isolation of mastitis pathogens, indicates mastitis (Harmon, 1994). Somatic cell count is used for milk pricing as one of the indicators of milk quality and is a widely accepted criterion for representing the udder health status of a dairy herd.
A legal limit of 400,000 cells/mL was established by the European Union as the maximum SCC in bulk milk collected at the farm (Council Directive 92/46/EEC, European Economic Community, 1992), although various quality programs reward farmers additionally for lower SCC in bulk milk. Milk from high-SCC individuals increases bulk tank SCC results. Therefore, this price differentiation may lead to farmers withholding milk from cows with high SCC and feeding it to calves.
The presence of antibodies to MAP is correlated with shedding of MAP (Nielsen and Ersbøll, 2006; Nielsen and Toft, 2006). Because the ingestion of contaminated milk or colostrum can be a primary route for calf infection (Taylor et al., 1981), showing an association between a high SCC and the occurrence of antibodies to MAP can be crucial for effective control in dairy herds. During the course of a MAP infection, the immune response varies, first presenting a cell-mediated immune response in early subclinical stages and later a humoral immune response with production of antibodies (Stabel, 2000). Kreeger and Snider (1992) described a reduced immune system cell response as a consequence of chronic MAP infection, which may be the basis for the increased risk of secondary diseases. If an association between MAP infection and the occurrence of mammary infections is present, evaluation of the time between the occurrence of MAP antibodies and a high SCC may provide evidence of the underlying immune mechanisms. At least 3 different hypotheses of this interrelationship can be established: 1) progression of a MAP infection may lead to progression of mammary infections, resulting in a high SCC; 2) mastitis may be the primary cause for reduced immune competence, leading to the progression of a MAP infection and the occurrence of antibodies; and 3) an unknown factor unrelated to the MAP infection and mastitis may lead to progression of both MAP and mammary infections.
McNab et al. (1991) showed an association between MAP (defined by a lipoarabinomannan ELISA-positive result) and higher SCC. Nordlund et al. (1996) and Hendrick et al. (2005) found no significant differences between SCC in MAP-positive and MAP-negative cows. Nonetheless, all these studies were cross-sectional, and if there is a time lag between the 2 events, an association may be detected only in a longitudinal study. The objective of the present study was to determine and characterize the association between occurrence of MAP antibodies and high SCC in dairy cattle in a longitudinal study with follow-up.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Data from repeated ELISA sampling, milk yield (kg), SCC records, and DIM were obtained through the Danish milk recording system. Information on calving identification and herd number, culling, parity, breed, and date of birth was extracted from the Danish Cattle Database. Table 1
summarizes information regarding age at first sampling for ELISA and SCC and the time at risk in each herd. The time at risk by herd was expressed as the sum of individual animal time units (cow years) from the beginning of the study to withdrawal or study end. Mycobacterium avium ssp. paratuberculosis was isolated on at least one occasion in all herds but one (herd 3), yet the antibody profile indicated MAP presence.
|
Sampling was carried out longitudinally as a cross-sectional survey with follow-up. Therefore, the number of cows across the study periods, and consequently the number of samples per cow, varied as follows: ELISA: minimum = 1, median = 5, and maximum = 40; SCC: minimum = 1, median = 6, and maximum = 45.
Laboratory Analyses
Somatic cell counts of milk samples were determined according to the International Dairy Federation/International Organization for Standardization standard 148A (International Dairy Federation, 1995) by using a Fossomatic 5000 instrument (Foss, Hillerød, Denmark).
Milk samples were tested for antibodies to MAP by using an adapted ELISA based on a commercially available antigen (Allied Monitor, Fayette, MO), whose performance was reported previously (Nielsen and Toft, 2006). Each sample was tested in duplicate. Polysorb microtiter plates (Nunc, Roskilde, Denmark) were coated with the antigen (2 µg/mL in 0.1 M carbonate buffer, pH 9.6) and left for 3 d at 5°C. The samples were diluted 1:2 in Mycobacterium phlei (Allied Monitor) to absorb possible cross-reacting antibodies to non-Mycobacterium avium strains. The absorbed samples were left overnight at 5°C and washed 5 times in PBS (pH 6.8) with 0.05% Tween20 (PBST20). After adding the samples to the coated plates in duplicate (100 µL/well), they were incubated at 5°C overnight. Two sets of positive and negative controls were included on each plate. After 5 repeated washings with PBST20, 100 µL of a 1:2,000 dilution of peroxidase-labeled goat antibovine IgG (H + L; Kirkegaard and Perry Laboratories, Gaithersburg, MD) was added to each well. The plates were incubated at 20°C and washed.
A solution of ortho-phenylene diamine (Kem-En-Tec, Copenhagen, Denmark) was diluted in citrate buffer (pH 5.0) to a concentration of 300 µg/mL. One hundred milliliters of this solution was supplemented with 80 µL of H2O2 and added to each well. The chemical reaction was stopped by adding 100 µL of 0.5 M H2SO4 to each well. Optical density (OD) values were obtained by using an ELISA reader (492 nm) and a 620-nm filter as reference. For each sample, a corrected OD (ODc) was calculated by subtracting the mean OD value of negative controls from the mean OD value of a test sample. The ODc were used for statistical analysis.
Statistical Methods
Descriptive Statistics.
The continuous SCC was dichotomized into high and low by using a cutoff of 300,000 cells/mL. An animal with at least 1 ODc >0.3 was defined as ELISA positive, based on the cutoff used in the Danish Paratuberculosis program (Nielsen et al., 2007a). All cows were categorized into parity groups (1, 2, and 
3) and lactation stages (0 to 100, 101 to 200, and >200 DIM).
Descriptive statistics were carried out, providing the distributions of ELISA and SCC groups stratified by herd, parity group, and lactation stage, respectively. Age at first sampling, high SCC, and ELISA positivity were computed by herd, and the distributions of age at first positive detection were calculated.
Association Between High SCC and ELISA-Positive Results.
Fisher’s exact test was used to test the association between the ELISA-positive or ELISA-negative groups and the high or low SCC groups. The results were statistically compared by herd and breed.
Robust Regression Analysis.
Robust regression analysis was used to assess the association between the age at first ELISA-positive sample and age at first high-SCC sample. Robust regression is a method that is more resistant to the presence of outliers than linear regression (Staudte and Sheather, 1990). The term outlier refers to data that appear unusual relative to the majority of the sample and may represent extraordinary (or unlikely) values of the dependent variable or data that are a result of errors in data collection. Robust regression can protect from high leverage points in the data (i.e., extreme values of the independent variable). Model fitting and parameter estimation were performed by using MM-estimation (Yohai, 1987) in the SAS procedure ROBUSTREG (Chen, 2002). The MM-estimation is a 3-step process in which the initial estimates are calculated by using least trimmed squares (which adjusts for outliers in the dependent variable) and subsequently adjusts to also account for outliers in the independent variable.
The general robust regression model was:
 |
where log (age at first positive ELISA) is the log-transformed value of age at the first positive ELISA; a is the intercept; β is the slope for log (age at first high SCC); log (age at first high SCC) is the log-transformed value of age at first high SCC; breed is the fixed effect of breed on the log (age at first positive ELISA); and 
is the random error. The log transformation was applied to improve the assumption of normally distributed observations.
Plots of residuals and predicted values were carried out to evaluate heteroscedasticity and the normal distribution of residuals. Residuals were tested for normality by using the Kolmogorov-Smirnov test. Leverage and outlier points were identified.
Additional Exploratory Analysis.
Additional exploratory analysis was used to further explore patterns in the data, specifically the relationship between ELISA and SCC results and age.
The distribution of the 4 outcomes (ELISA positive/ high SCC; ELISA positive/low SCC; ELISA negative/ high SCC; and ELISA negative/low SCC) was evaluated by plotting the predicted values vs. age. The SAS procedure GAM was used to visualize the pattern among the variables, representing the data in a smoothed format. A generalized additive model was fitted by using a B-spline smoother with 3 degrees of freedom. All of the above analyses were performed in the software SAS 9.1.3 (SAS Inst. Inc., Cary, NC).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
. The distribution of the high-SCC and low-SCC cows stratified by ELISA group, herd, parity, and lactation stage is given in Table 3. The number of positive animals increased from parity group 1 to 3 for both SCC and ELISA. The distributions of SCC values were first quartile (Q1): 59,000 cells/mL; median (Q2): 131,000 cells/mL; third quartile (Q3): 318,000 cells/mL; and maximum (Q4): 9999,000 cells/ mL (detection limit).
|
|
, respectively. Outliers for each herd are shown. For each of the 26 herds, a similar age distribution was found for both the ELISA-positive and high-SCC first sample.
|
The "age at first positive ELISA" plotted against "age at first high SCC" is shown in Figure 2 (2,509 animals were found both negative to ELISA and with low SCC and consequently were not plotted). No cows under 2 yr of age were either ELISA positive or had high SCC, although the average age at first sampling was 1.9 yr. The linear appearance of the graph supports the view that a MAP-positive classification is strongly associated with a high SCC. This association was further examined by robust regression.
|
 |
The results imply that when cows have high SCC and are ELISA positive, the former is expected to occur first, but the expected time between the 2 events increases with an increase in age at first high SCC.
Breed was not significant (P = 0.89) and was excluded from the final model. Results from fitting the robust regression model via MM-estimation are summarized in Table 4. The coefficient of determination (R2) was 0.51. For comparison, ordinary least squares regression yielded a R2 = 0.72. After performing the log-log age transformation, the distribution of residuals was found suitable for the model.
|
.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The cutoff value of 300,000 cells/mL was chosen arbitrarily. A sensitivity analysis, in which cutoff values of 200,000 and 100,000 cells/mL were used, suggested that the selected cutoff value did not influence the final outcome (data not shown). Robust regression estimates for each of the selected values were not significantly different, illustrating the robustness and reliability of the results.
The exploratory analysis of the 4 outcomes (ELISA positive/high SCC; ELISA positive/low SCC; ELISA negative/high SCC; ELISA negative/low SCC) indicates that MAP-positive animals have higher values of SCC. In previous cross-sectional studies (Nordlund et al., 1996; Hendrick et al., 2005), no significant association was found between MAP and SCC, but they did not include a longitudinal component, and the lack of association could be due to the 2 events not occurring simultaneously. Our study included repeated sampling over time from each cow, which improved the detection of MAP-infected animals and highlighted the time association between ELISA positivity and having high SCC.
The Three Hypotheses
Three different scenarios were evaluated through the analysis of time of occurrence to MAP antibody positivity and having high SCC. A considerable proportion (40%) of cows first tested positive by ELISA. For example, based on the regression coefficients, an ELISA-positive animal at 4 yr of age may have had a high SCC 1.8 yr earlier; a 6-yr-old ELISA-positive animal may be expected to have experienced a high SCC 2.4 yr earlier. Although the events leading to mastitis are often of a multifactorial nature, the resistance of the host to mastitis pathogens may be influenced by the immune status of the animal. Therefore, a compromised immune response induced by progression of a MAP infection may predispose cows to mastitis, leading to increased SCC, as described by hypothesis 1. It is generally considered that antibodies are produced as the infection progresses from cell-mediated to humoral immune responses (Stabel, 2000), although antibodies occur during both immune phases. The duration of each process is not known, but is probably highly variable according to the results of Nielsen and Ersbøll (2006). A variable duration of the cell-mediated immune responses may explain the variation in whether an animal first tested positive in ELISA or in SCC, because a small amount of antibodies could be present during cell-mediated immune responses and elicit a positive ELISA. If that is the situation, hypothesis 2 or 3 is more likely. Only 12% of the ELISA-positive animals had low SCC. A large proportion (46%) of the animals had high SCC prior to a positive ELISA (hypothesis 2), suggesting that the process of mastitis infection may cause alterations in immune function that may lead to impaired cellular immunity to MAP, with progression of the infection and consequent antibody production. Because of the long incubation period of the infection, it can be hypothesized that a proportion of high-SCC animals never become ELISA positive because of withdrawal or incomplete observation at the end of the follow-up study. Depending on management practices and the condition of the immune system, cows can become infected with a mastitis pathogen and never become MAP positive if they have never been exposed to MAP.
A small proportion (14%) of the animals tested positive to MAP and high SCC simultaneously. Although the association between MAP and SCC was present, a third factor is probably affecting general immunity, resulting in both high SCC and antibodies to MAP (hypothesis 3).
The large variation in the distribution of first occurrence of MAP-positive assessment and high SCC indicates that none of the 3 hypotheses could be firmly excluded. Nevertheless, the variation in the results may indicate that the effect of an unknown mechanism negatively affecting the general immune competences may lead to both SCC and MAP deterioration, thereby favoring hypothesis 3.
Implications for Use of Test Results
The study results indicate that there is an association between ELISA-positive measures and high SCC. Therefore, an increased risk that cows with high SCC become infectious with MAP should be considered by farmers and advisors. Test results can be used for targeting potentially infected and infectious animals. The management of MAP-infected animals should include procedures to avoid the use of milk from high-SCC cows.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Received for publication July 5, 2007. Accepted for publication September 18, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. S. Nielsen, H. Bjerre, and N. Toft Colostrum and Milk as Risk Factors for Infection with Mycobacterium avium subspecies paratuberculosis in Dairy Cattle J Dairy Sci, December 1, 2008; 91(12): 4610 - 4615. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
3), and those with high SCC are plotted on the horizontal axis (y 