Study population
Seventy-seven dairy herds from Thuringia, Hesse and Saxony, all federal states of Germany, with different within herd prevalence of paraTB were selected for this study. Nineteen herds were certified as ‘paraTB-unsuspicious’ that is comparable to the ‘herd classification no. 6’ of the uniform program standards of the VBJDCP [4]. The other 58 herds were classified as MAP positive based on recent results of FC of individual faecal samples obtained for whole herd testing. All herds were housed in free stalls.
BS and LM sampling
The BS samples were collected as described by Eisenberg et al. [6] by a veterinarian of the animal health service from each federal state. Briefly, three BS were taken in parallel mainly in localizations with high cow traffic involving manure concentration areas which had been proven to be suitable for environmental sampling [31].
Additionally, samples from LM storage areas (tanks, lagoons, pits or pre-flooders) were obtained as described recently by Donat et al. [7].
Three laboratories working under quality management standards were involved in the study (LI – LIII). After sampling, BS and LM were immediately transported to either LI or LII. LM was divided into three aliquots of equal volume. All samples were frozen and stored at −20 °C for 1 to 2 months except LM for qPCR B. For technical reasons these latter samples were stored for 8 to 10 months at −20 °C. One BS and one aliquot of LM of each herd were shipped frozen to the two other laboratories.
FC of all samples was performed in the three laboratories. Furthermore, the samples were tested by two different qPCR methods, qPCR A (LII) and qPCR B (LIII).
Boot swab handling
BS were placed in 125 mL plastic cups with screw cap and cut into small pieces with sterile scissors. In LI and LII, 100 mL of sterile physiologic salt solution (Merck, Darmstadt, Germany) were added and faecal material was rinsed off the boot swab pieces by automatically shaking at 200 rpm for 30 min. The eluate was transferred into a sterile 50 mL centrifuge tube (Sarstedt, Nümbrecht-Elsenroth, Germany). The supernatant was discarded after centrifugation for 15 min at 2000 x g.
In LIII, 3 g of cut BS were directly processed for bacterial culture and qPCR as described below.
Bacterial culture
FC of processed BS as well as LM was done according to the official manual of diagnostic procedures published by the Friedrich-Loeffler-Institut, the German Federal Research Institute of Animal Health [32].
First, 3 g of faecal material or 5 g of cut BS (LIII) were mixed with 30 mL of a 0.75% hexadecylpyridinium chloride solution (Acros Organics, Fisher Scientific GmbH, Nidderau, Germany) for decontamination. After sedimentation of coarse material for 5 min, 20 mL of the supernatant were decanted, horizontally shaken for 30 min at 200 rpm and subsequently incubated for 48 h at room temperature in the dark. Then, the supernatant was discarded and 0.2 mL of the sediment was inoculated onto each of 3 tubes of commercial Herrold’s Egg Yolk Medium (HEYM) with mycobactin J, amphotericin B, nalidixic acid and vancomycin (ANV; Becton Dickinson, Heidelberg, Germany). After cultivation under aerobic conditions (37 °C ± 2 °C) for 7 days in a slanting position, the tubes were locked airtight and incubated vertically for another 11 weeks. Differentiation of suspected colonies was carried out by IS900 PCR [33]. Samples were characterized as MAP-positive (MAP+), MAP-negative (MAP-) or non-assessable (n.a.) due to microbial contamination. The number of cultivable bacteria was recorded semi-quantitatively as a score; score 1: up to 10 colonies/tube, score 2: 11–50 colonies/tube, score 3: 51–100 colonies/tube, score 4: >100 colonies/tube. The average score of the three tubes per sample was calculated.
DNA preparation and PCR
qPCR A, utilizing nucleic acid purification with magnetic beads
Nucleic acids were extracted from faeces using the MagMax™ Total NucleicAcid isolation kit (Life Technologies, Darmstadt, Germany) according to the instruction of the manufacturer, which uses zirconia beads for the mechanical disruption and magnetic beads for nucleic acid purification.
The samples were thawed, homogenized and 0.3 g were transferred into 1 mL phosphate buffered saline. After homogenization by vortexing for 3 min and centrifugation at 100 x g for 60 s, 175 μl of the supernatant were added to the zirconia bead tubes previously filled with 235 μl lysis/binding solution. The samples were homogenized three times for 30 s at 6800 rpm (Precellys® 24 homogenizer, Bertin Technologies, Montigny-le-Bretonneux, France) and centrifuged at 16,000 x g for 3 min.
To clarify the lysate, 300 μl of the sample were transferred into a new 1.5 mL tube and centrifuged for a second time at 16,000 x g for 6 min.
The MagMax™ Express 96 instrument (Life Technologies) was applied for nucleic acid purification. After four washing steps the purified DNA was directly used for the qPCR.
For the detection of MAP DNA, the VetMAX™ MAP Real-Time PCR screening kit (Life Technologies), performed on a 7500 Fast qPCR cycler (Life Technologies), was used according to the instruction of the manufacturer.
A sample was detected to be positive at CT-values ≤37.0, inconclusive at CT-values >37.0 and <40.0. Samples with atypical amplification curves or with undetermined CT-values for target and internal control (PCR inhibition) were considered n.a.
qPCR B, utilizing sample enrichment and silica membrane-based nucleic acid purification
5.0 g of cut BS or 3.0 g of LM, respectively, were suspended in 20 mL sterile distilled water and left for 10 to 20 min at room temperature for sedimentation of coarse material. To concentrate bacteria and remove PCR inhibiting substances, 10 mL of the supernatant were centrifiltrated using Adiafilter (ADIFIL 100, Adiagene, Saint Brieux, France) and the pellet was re-dissolved in 500 μL sterile distilled water. Bacterial cells were mechanically disrupted by bead-beating with 300 mg glas beads (Peqlab, Erlangen, Germany) for 10 min at 30 Hz using the TissueLyser (Qiagen, Hilden, Germany). After an initial centrifugation step (5 min, 15,000 x g) the nucleic acid extraction was carried out using the QIAamp® DNA Mini kit (Qiagen) according to the instruction of the manufacturer. The qPCR was performed with the Adiavet™ Paratb Real time kit (Adiagene) using the 7500 Real Time PCR system (Applied Biosystems, Foster City, CA, USA) as recommended. Samples were tested in duplicate and the mean CT-value of both replicates was calculated. A sample was detected to be positive at mean CT-values ≤38.0, inconclusive at mean CT-values >38.0 and ≤40.0 and n.a. when one replicate yielded a measurable and the other an undetermined CT-value for the target.
Statistical data analysis
Data recording and descriptive statistics were performed using a Microsoft Excel® spreadsheet (Microsoft Office 2007, Microsoft Corporation, Redmond, USA). Further statistical examinations were done by means of the statistical programme packages BMDP [34] and StatXact 9.0 [35]. The figures were created with the statistical software package PASW Statistics 17 (SPSS, Quarry Bay, Hongkong).
For each matrix, BS and LM samples, the Spearman’s rank correlation coefficient (programme BMDP3D) was calculated in order to analyse the relationship between the semi-quantitative colony growth score of FC and the CT-values of qPCR.
Kappa coefficients were calculated with the programme BMDP4F to describe the degree of agreement between the dichotomized outcomes (negative or positive) of the FC and the qPCR test systems. The following levels of agreement were considered for the interpretation of the kappa coefficient (ĸ) [36]: < 0.20: slight; 0.21–0.40: low; 0.41–0.60: moderate; 0.61–0.80: substantial; 0.81–1.00: excellent. Furthermore, both calculations were performed for the description of the accordance between qPCR A and B for each matrix, and for the accordance between BS and LM using either PCR A or B. Herd level Se and Sp for both qPCR test systems applied either on BS or LM samples were estimated by a frequency-based approach classifying herds according to their certification status (MAP+, MAP-) using the cut-off CT-values as recommended by the manufacturers for faecal samples. The corresponding exact 95% confidence intervals of the binomial distribution were calculated according to Clopper and Pearson [37]. Cut-offs for both qPCR tests were estimated by TG-ROC-analysis [38, 39]. Herd level Se and the cut-offs of the qPCRs were re-calculated after adjusting the values for herd level Sp to 100%. These calculations were performed using R software [40], version 2.15.2 (2012–10-26), with package DiagnosisMed [41]. In general, the statistical significance level was set at α = 0.05.