|Home | About | Journals | Submit | Contact Us | Français|
The diagnosis of bovine genital campylobacteriosis (BGC) presents significant challenges, as traditional methods lack sensitivity when prolonged transport of samples is required. Assays of preputial samples by means of real-time polymerase chain reaction (PCR) provide good sensitivity and high throughput capabilities. However, there is limited information on the acceptable duration of transport and temperature during transport of samples. In addition, the use of pooled samples has proven to be a valuable strategy for the diagnosis of other venereal diseases in cattle. The objectives of the present study were to determine the effect of sample pooling and of transport time and temperature on the clinical sensitivity of a real-time quantitative PCR (qPCR) assay for Campylobacter fetus subsp. venerealis in preputial samples from beef bulls. Eight infected bulls and 176 virgin yearling bulls were used as the source of samples. The qPCR sensitivity was comparable for unpooled samples and pools of 5 samples, whereas sensitivity was decreased for pools of 10 samples. Sensitivity for the various pool sizes improved with repeated sampling. For shorter-term transport (2 and 48 h), sensitivity was greatest when the samples were stored at 4°C and 30°C, whereas for longer-term transport (96 h) sensitivity was greatest when the samples were stored at −20°C. The creation of pools of 5 samples is therefore a good option to decrease costs when screening bulls for BGC with the qPCR assay of direct preputial samples. Ideally the samples should be stored at 4°C and arrive at the laboratory within 48 h of collection, but when that is not possible freezing at −20°C could minimize the loss of sensitivity.
Le diagnostic de la campylobactériose génitale bovine (CGB) présente des défis significatifs, étant donné que les méthodes traditionnelles manquent de sensibilité lorsqu’un transport prolongé des échantillons est requis. Les épreuves utilisant des échantillons prépuciaux dans des épreuves de réaction d’amplification en chaine par la polymérase en temps réel (PCR) ont une bonne sensibilité et une capacité de rendement élevée. Toutefois, il y a peu d’information sur la durée acceptable du transport et de la température durant le transport des échantillons. De plus, l’utilisation d’échantillons regroupés s’est avéré être une stratégie valable pour le diagnostic d’autres maladies vénériennes chez les bovins. Les objectifs de la présente étude étaient de déterminer l’effet du regroupement d’échantillons et du temps de transport et de la température sur la sensibilité clinique d’une épreuve PCR quantitative en temps réel (qPCR) pour Campylobacter fetus ssp. venerealis dans des échantillons prépuciaux provenant de taureaux. Huit taureaux infectés et 176 bouvillons vierges ont été utilisés comme source des échantillons. La sensibilité du qPCR était comparable pour des échantillons non-regroupés et des regroupements de 5 échantillons, mais diminuée pour des regroupements de 10 échantillons. La sensibilité pour les différentes tailles de regroupement s’améliorait suite à des échantillonnages répétés. Pour des transport de courte durée (2 et 48 h), la sensibilité était plus élevée lorsque les échantillons étaient entreposés à 4 °C et 30 °C, alors que pour le transport de longue durée (96 h) la sensibilité était plus élevée lorsque les échantillons étaient entreposés à −20 °C. La création de regroupement de 5 échantillons est une bonne option pour diminuer les coûts lors du tamisage de taureaux pour CGB avec le qPCR effectué directement sur des échantillons prépuciaux. Idéalement, les échantillons devraient être entreposés à 4 °C et arriver au laboratoire au plus tard 48 h après le prélèvement, si ce n’est pas possible, la congélation à −20 °C pourrait minimiser la perte de sensibilité.
(Traduit par Docteur Serge Messier)
Infection with Campylobacter fetus subsp. venerealis (Cfv) can be an important cause of reproductive loss in beef herds. Herds with Cfv infection often have pregnancy rates lower than expected, extended calving seasons, and abortions (1–5). Many important epidemiologic features of bovine genital campylobacteriosis (BGC) are shared with the features of trichomoniasis; however, diagnostic methods face the challenge of distinguishing Cfv from the other subspecies, C. fetus subsp. fetus (Cff). A diagnosis of herd infertility is often based on the screening of preputial samples from herd bulls for infectious agents, including Cfv (6). Although vaccines against BGC are available, detection and removal of infected bulls is important for the prevention and control of the disease in cow–calf herds.
The development and clinical validation of molecular diagnostic techniques for direct testing for Cfv in bovine preputial samples have facilitated field investigation of poor reproductive performance under extreme environmental conditions in remote beef herds (4,7). Among the many such techniques, the real-time polymerase chain reaction (PCR) assay designed by Hum et al (8) is probably the most extensively evaluated. This assay is based on the primers VenSF and VenSR, which amplify a region 142 base pairs long in a genomic island of the parA gene of Cfv (9). These primers were recently adapted to the SYBR Green quantitative PCR (qPCR) platform, and sample processing was optimized for direct preputial sample testing (10). The resulting analytic sensitivity of the assay was the equivalent of a single cell per reaction (10), and the average clinical sensitivity was 85.4% (7). However, these parameters were determined under nearly ideal collection and transport conditions (samples were stored at 25°C after collection and processed within 2 to 4 h).
Because testing for BGC is often done at the same time as testing for trichomoniasis, there was interest in the potential for pooling samples to reduce the costs associated with testing. Recent studies have shown that the sensitivity of Tritrichomonas foetus testing is not significantly diminished for pool sizes of 5 or 10 samples when the samples are collected into either InPouch TF enrichment medium or phosphate-buffered saline (PBS) (11,12). This strategy allows for the screening of large number of bulls more affordably.
Preputial samples collected in the field are subject to long transport times and variable temperatures before arrival at the laboratory. Conditions in North America range from high temperatures in the summer to extremely low temperatures during the winter (13,14). Since Cfv is temperature-sensitive, temperature extremes during transport can result in false-negative results, particularly when viable organisms are needed (15). In addition, preputial samples have a wide variety of normal microbiota in addition to fecal and other contaminants. Control of factors affecting the growth of contaminants is critical for culture-based diagnosis (15,16). However, although transport conditions have been extensively studied with regard to culture techniques, the impact of variable transport times and temperatures on direct molecular techniques is unknown.
The overall goal of this project was to optimize a real-time PCR test for use by veterinarians under field conditions. The first objective was to evaluate whether the laboratory-related costs for screening bulls could be decreased by pooling samples for PCR testing. The second objective was to determine the effect of variations in time of travel to the laboratory and transport temperature on the efficiency of this testing procedure.
The animal procedures were done in accordance with the guidelines of the Canadian Council on Animal Care and were approved by the University of Saskatchewan’s Animal Research Ethics Board. Bulls with and without a Cfv infection were sampled. Preputial samples were collected in May and June 2013 from 176 virgin yearling bulls at the Agri-Environmental Services Branch Spring Creek Bull Station, Broderick, Saskatchewan, to serve as a source of negative samples for the pooling study. This group included 4 breeds: Black Angus, 47% (83/176); Red Angus, 30% (52/176); Simmental, 15% (26/176); and Charolais, 8% (15/176).
Eight Cfv-infected bulls were housed and sampled from June to August 2013 at the Western College of Veterinary Medicine, Saskatoon, Saskatchewan. Three of the bulls were naturally infected with Cfv and originated from an infected herd as previously reported (4). The remaining 5 bulls were artificially infected with isolates from the naturally infected bulls 1 y before the start of this project; the procedure and results of the infection study have been reported elsewhere (17). This group included 6 Black Angus and 2 Limousin bulls aged 4 to 8 y. A preputial sample was collected weekly for 9 wk from the in-house bulls and divided into aliquots.
Both qPCR and culture were used to test all the samples from both virgin and known-positive bulls, as the sensitivity of any currently available technique is less than 100% even under ideal conditions (7,16). Also, because the sensitivity of culture, the gold standard for diagnosis of Cfv, is limited under field conditions (16), we considered the in-house bulls to be infected if Cfv was cultured from the prepuce at the beginning and end of the testing period. For example, if a bull was determined to be culture-positive in week 1, that bull was sampled every week thereafter, and if week 9 was the last sampling point for that bull and that sample was culture-positive, all the samples from week 1 to week 9 were considered to be from a Cfv-positive bull. As a result, the sensitivity of a particular test would be calculated as the proportion of all samples testing positive during that period.
Preputial samples were collected by the aspiration method (18,19). The material collected from each virgin bull was transferred into a screw-capped tube containing 2 mL of PBS (20 mM phosphate, 150 mM NaCl). The first sample was packed in an insulated container with bags of water at approximately 25°C and immediately transported to the laboratory for culturing. The second sample was transported to the laboratory in a Styrofoam box containing ice packs and immediately frozen at −80°C. An aliquot from the first sample was cultured as previously described (16). Another aliquot from the first sample was screened by qPCR with the VenSF/VenSR primers (8) in a preoptimized SYBR Green assay using a heat-lysis preparation method (10,20).
Samples from the Cfv-positive bulls were collected and transported in a similar manner except that the collected preputial material was transferred into 2.5 mL of PBS. Immediately after collection, aliquots for culture and qPCR were removed, and the remainder of the first sample was frozen at −80°C for use in the pooling study (weeks 1 to 5). The second sample from each of the Cfv-positive bulls was split into 200-μL aliquots, which were dispensed into microcentrifuge tubes 12 × 1.5 mL for use in the transport-conditions study (weeks 2 to 10).
Culture for Cfv was conducted with a filter-based method using nonselective medium optimized for Campylobacter isolation (16,21). All fresh preputial samples were transported to the laboratory and processed within 3 h of collection.
For qPCR, DNA was released from the preputial samples by the heat-lysis method as previously described (10,20); all samples were tested in duplicate. The PCR reactions were carried out as reported with the primers VenSF and VenSR (10). The data obtained were analyzed with the use of computer software (iQ5 Optical System; Bio-Rad, Mississauga, Ontario); end-point analysis with this software calculates the average fluorescence, measured in relative fluorescence units (RFU), over the last 5 cycles (7). Samples were considered to have a positive result if they had a mean RFU greater than the RFU of samples known to be negative (no template control and extraction negatives) plus a tolerance value of 2.5%. The optimal tolerance value had previously been determined by use of a receiver operating characteristics curve (7). In addition, all positive samples were evaluated to determine if the correct melting-peak signal was generated at 78.5 ± 0.5°C (10). To ensure that nontarget products such as primer dimers were not recorded, the RFU reading was taken after each PCR cycle at 76°C between the PCR extension and denaturing steps (7).
Pools were made by combining three 200-μL aliquots from each of 35 samples from the 8 known-positive bulls collected during the first 5 wk of the study period and pooling these aliquots with aliquots from negative samples. One positive sample was pooled with negative samples to form sample pools described as number positive/total number of samples in a pool. The pools evaluated were 1/3 (1 positive + 2 negative), 1/5 (1 positive + 4 negative), and 1/10 (1 positive + 9 negative). The resulting volumes of the pooled samples were 0.6 mL (1/3), 1.0 mL (1/5), and 2.0 mL (1/10). Samples that were not pooled, including undiluted samples from the positive bulls, were reported as 1/1. Thirty-five pools were made for each ratio and were tested by real-time PCR as described below.
All pools of 1/3, 1/5, and 1/10 and the corresponding unpooled samples for each positive bull that were tested in consecutive 1-, 2-, and 3-wk periods were identified. The evaluation of sequentially tested pools from each positive bull was structured to assess the utility of repeated sampling. Repeated sampling is intended to increase the probability of identifying positive bulls and is routinely recommended for T. foetus (11). The results for each positive bull were then summarized as a cumulative positive or negative result for each series of 1, 2, and 3 sequentially tested pools. For example, if at least 1 pool out of a series of 3 sequentially collected and tested pools of a given size containing material from a given positive bull was qPCR-positive, then the testing strategy was successful in identifying the positive bull.
Crude sensitivity was estimated for each pool size by calculating the proportion of positive samples among all samples tested. Generalized linear mixed models (GLIMMIX procedure, version 9.2; SAS Institute, Cary, North Carolina, USA) with a logit link function and binomial distribution were used to compare the sensitivity estimates between pools of different sizes. The repeated use of samples from the same bull was accounted for with a random intercept for the sample and a random intercept for the bull. A Tukey adjustment was used to account for post-hoc multiple comparisons among unpooled samples and pools of 3, 5, and 10 samples.
Nonparametric tests were used to compare mean cycle threshold (Ct) values across pool sizes because the Ct values for test-negative samples were right-censored (≥40), and exact values were not available. A Kruskal–Wallis test was used to detect any difference among pool sizes, and then post-hoc pairwise Mann–Whitney U-tests were used and adjusted for multiple comparisons by means of IBM SPSS Statistics, version 22.0 (IBM Corporation, Armonk, New York, USA).
A similar logistic regression model, with a random intercept to account for use of the same bull in multiple pools, was constructed to evaluate the cumulative sensitivity of serial testing of pooled samples across 1, 2, and 3 subsequent collection weeks. Fixed effects for the serial testing model included pool size, number of cumulative samples tested for each bull, and interaction between pool size and number of cumulative samples tested. For both models, plots of standardized residuals compared with predicted values were used to assess the models for extreme outliers. The level of significance used was 5%.
One aliquot from each sample was allocated to each specific time (2, 48, and 96 h) and temperature (−20°C, 4°C, and 30°C) immediately after collection. Each week, 1 set of three 200-μL aliquots was put in the refrigerator and held at 4°C for various times. The first aliquot was held for 2 h; the second and third aliquots were held for 48 and 96 h, respectively. In weeks 2, 4, 6, and 8 of the study the second set of 3 aliquots from each bull was held at −20°C for the same periods. In weeks 5, 7, 9, and 10 of the study the second set of 3 aliquots was held at 30°C for the same periods. The samples collected in week 3 were not moved at the required time and were discarded from the study. All samples were moved to long-term storage at −80°C immediately at the end of their period. For all temperatures and transport times the conditions were simulated in a controlled environment to avoid fluctuations within the temperatures evaluated.
Potential differences in test sensitivity among time and temperature combinations were examined with the use of generalized linear mixed models by means of the SAS GLIMMIX procedure with a logit link function and binomial distribution. Random intercepts were used to account for repeated measures among bulls and repeated analysis of individual samples from bulls within weeks. The model included fixed effects for holding time, temperature, and an interaction between time and temperature. Plots of standardized residuals were compared with predicted values to assess the models for extreme outliers. The level of significance used was 5%.
Nonparametric tests were used to compare mean Ct values across time and temperature combinations as described for pool sizes.
All 8 Cfv-infected bulls were confirmed to be culture-positive and qPCR-positive at the first sample collection. Seven of the bulls were culture-positive at or after the end of the trial (Table I). The final bull was culture-positive at week 9 and qPCR-positive at week 10. The sensitivity of the direct qPCR for all baseline samples collected from the 8 positive bulls each week was 91% [95% confidence interval (CI) 80% to 96%]. From the 176 virgin yearling bulls tested, 163 samples were confirmed to be negative by culture and qPCR and were included as negative samples in the pooling study. The resulting specificity for the direct qPCR was 93% (95% CI 88% to 96%).
Of the 35 samples from the 8 Cfv-infected bulls (5 samples from 3 bulls and 4 samples from 5 bulls) collected during the first 5 wk of the study and used as positive samples for the pooling study, 88% (31/35) were culture-positive and 94% (33/35) were qPCR-positive at the time of sample collection. The first and last samples from each bull used in the pooling study were culture-positive and qPCR-positive.
The crude sensitivity of the qPCR test for each pool size was as follows: unpooled, 94% (33/35); 1/3, 77% (27/35); 1/5, 88% (31/35); and 1/10, 57% (20/35). After accounting for repeated use of samples in multiple pool sizes and repeated sampling of bulls (Figure 1), we detected no significant differences in sensitivity between unpooled samples and pools of 1/3 (P = 0.10) or 1/5 (P = 0.73) or between pools of 1/3 and 1/5 (P = 0.38). However, the sensitivity of pools of 1/10 was lower than the sensitivity of unpooled samples (P = 0.001) and pools of 1/5 (P = 0.015); the sensitivity of pools of 1/3 did not differ from that of pools of 1/10 (P = 0.14). The variance accounted for by individual samples [σ2 = 5.5, standard error (SE) 4.3] and repeated analysis of individual bulls (σ2 = 1.5, SE 2.0) was not significant.
The mean qPCR assay Ct values also differed significantly between pool sizes (P < 0.0001). Unpooled samples had Ct values [median 30.0, interquartile range (IQR) 26.6 to 32.2] that were significantly lower than the Ct values of pools of 3 (median 34.2, IQR 31.9 to 35.3), 5 (median 34.1, IQR 31.6 to 35.1), and 10 (median 34.4, IQR 32.1 to ≥ 40) samples. After accounting for multiple comparisons, we found no significant differences (P = 0.99) among the pools of 3, 5, and 10 samples.
There were also significant differences in the sensitivity of the qPCR test between unpooled samples and pools of 3 to 10 samples (P = 0.0003) when differences among the cumulative results of 1, 2, or 3 sequential samplings (P = 0.0008) were considered after we accounted for repeated sampling of individual bulls (Figure 2). Regardless of pool size, the sensitivity of the qPCR test was higher when bulls were considered positive if they tested positive in any of 2 (P = 0.013) or 3 (P = 0.002) repeated samples as compared with a single sample. After considering the improvement in sensitivity associated with repeated sampling to detect positive bulls, we determined that pools of 10 samples had lower sensitivity than unpooled samples (P = 0.004) or pools of 1/5 (P = 0.003). However, there were no significant differences in sensitivity for repeated assessment between pools of 10 and 3 (P = 0.07) or pools of 5 and 3 (P = 0.06).
From the 8 Cfv-positive bulls 29 samples were divided into aliquots for testing and then held at 4°C and −20°C for 2, 48, and 96 h. A second set of 29 paired samples from the same bulls were divided into aliquots for testing and held at 4°C and 30°C for 2, 48, and 96 h. This process resulted in 348 tested aliquots (2 × 29 × 2 × 3) from the 8 bulls. Complete sets of 4 replicates of each of the 12 time and temperature combinations were available for 6 bulls (48 aliquots per bull), 3 repeats were available for 1 bull (36 aliquots), and 2 repeats were available for the final bull (24 aliquots).
Of the 58 samples used for analysis, 86% (50/58) were culture-positive and 84% (49/58) were qPCR-positive at the time of collection. The variance accounted for by differences across bulls (σ2 = 3.8, SE 1.4) was significant; however, the variance accounted for by repeated analysis of samples from individual bulls within the week (σ2 = 6.5, SE 4.5) was not.
Overall, there was no significant difference among temperature and time combinations (P = 0.13) (Table II). However, pairwise comparison of sample storage at either 4°C or 30°C for 96 h suggested lower sensitivity of the qPCR as compared with storage at 30°C for either 2 or 48 h (P = 0.04). Short-term storage at −20°C for 2 h also appeared to decrease the sensitivity of the qPCR relative to storage at 30°C for either 2 h (P = 0.04) or 48 h (P = 0.04).
After adjustment for multiple comparisons, there was only 1 significant difference in mean Ct values between time and temperature combinations (Table III): samples stored at 30°C for 48 h had lower mean Ct values than those stored at −20°C for 2 h (P = 0.04).
Diagnostic evaluation of herds for BGC presents several challenges. The development of PCR-based techniques has minimized some issues associated with sample handling and transport. However, the laboratory cost of testing individual samples is still an important deterrent for producers, often limiting Cfv testing unless there is evidence of reproductive failure. The ability to pool samples for Cfv testing provides a practical strategy to reduce cost to facilitate continual herd monitoring in commercial operations and enable large epidemiologic studies to evaluate potential control strategies. One objective of this study was to evaluate the effect of pooling preputial samples collected by aspiration directly into PBS on the sensitivity of qPCR testing for BGC. Pools of 10 samples had significantly lower sensitivity for detection of a single positive bull than pools of 5 samples or unpooled samples, potentially owing to dilution. Previous research has shown that pooling samples for PCR testing increases the efficiency of testing for trichomoniasis (11,12,22) without resulting in a significant difference between qPCR sensitivity for pools of 5 or 10 samples and unpooled samples (12).
Together with the previous work (12), the present study suggests that pools of samples from 5 bulls can be cost-effectively screened for both BGC and trichomoniasis when the samples are directly collected into PBS. The resulting sensitivity is comparable to that for unpooled samples, particularly when the bulls are sampled on 2 or 3 dates. In addition, PBS is an inexpensive and readily available collection medium as compared with, for example, the commercial medium pouches traditionally recommended for trichomoniasis testing. The field and laboratory steps in this BGC screening protocol were designed as a cost-effective method for assessing a large number of bulls. However, the PCR protocol for the pooled samples might be enhanced in future studies with the addition of an internal PCR control to evaluate the presence of amplifiable DNA in the sample. Although this could increase the cost of testing, the issue is especially relevant when there is inhibition of the PCR reaction, as has previously been recognized with preputial samples (12).
Collecting an individual sample from each bull and having the diagnostic laboratory pool the samples with an equal volume from each bull will result in pools that contain consistent volumes for each bull and also provide the opportunity to test individuals when a positive pool is identified. This recommendation is similar to that reported for pooling samples for trichomoniasis testing. Just as with trichomoniasis testing in bulls (12), serial sampling when testing pools of samples from bulls for BGC will increase the sensitivity for detection of infection.
Previous research supports the recommendation for serial sampling when testing individual animals (7). In that study, as in the present one, bulls known to be Cfv-positive did not test positive at every sampling, and there was significant variability among samplings in the number of Cfv organisms detected. Others have also reported variations in the numbers of cultured Cfv organisms (6) and have supported the use of serial testing with 3 to 6 samples at weekly intervals to confirm negative status (23–25). In the present study, the cumulative sensitivity of the pooled qPCR assay for both 2 and 3 consecutive preputial samples was significantly higher than the sensitivity for a single sample. However, no significant improvement was found when 3 consecutive samples were tested as compared with 2. The industry is already familiar with this practice since repeated sampling is commonly recommended for T. foetus testing (18).
The second objective of this study was to determine whether variations in the travel time to the laboratory and the temperature during transport would have an impact on the sensitivity of this testing procedure. It is not unusual for samples collected before the weekend to take 3 days or more to arrive at the laboratory and to be exposed to a range of temperatures during transport, even when practitioners have attempted to insulate the samples for shipment. Information regarding transport times and conditions for samples tested through real-time PCR-based methods is limited in the literature. One study of the use of conventional PCR testing for preputial samples collected and transported in modified Weybridge transport enrichment medium indicated that transport exceeding 24 h at room temperature had a detrimental effect on assay sensitivity when there was contamination and overgrowth with other organisms (26). The presence of Pseudomonas spp. in preputial samples is relatively common (16), and these organisms, in addition to others present in the prepuce, are capable of producing DNase, which might explain the observed reduction in sensitivity. McMillen et al (20) reported a decrease in the number of Cfv organisms detected by qPCR as storage at ambient temperature increased to 5 d but not when storage was for only 2 d. The same authors showed a similar effect when testing preputial samples for T. foetus (27).
In the present study, transport times of up to 48 h at 30°C did not reduce sensitivity significantly. However, sensitivity decreased from 99% at 48 h to 84% at 96 h, indicating that prolonged transport at higher temperatures could decrease sensitivity, in agreement with previous research. In temperate climates it is also important to consider exposure to ambient temperatures higher than 30°C during transport, particularly when courier services are used. Moderate increases in temperature increase the growth rate of many bacteria, including pseudomonads that produce DNases (28). The exposure of preputial samples used for trichomoniasis diagnosis to higher ambient temperatures (46°C and 54°C) (14) resulted in higher Ct values, indicating lower concentrations of organisms. Although temperatures above 30°C were not evaluated herein, one could expect similar effects for Cfv as reported for T. foetus.
Samples are commonly stored in the refrigerator at 4°C before being shipped to the laboratory. Although we found no reports of previous studies evaluating the effect of storage or transport at 4°C on the detection of Cfv in preputial samples by real-time PCR, there was information for preputial samples used for trichomoniasis diagnosis. Two studies concluded that storage at 4°C for up to 24 h had no detrimental effect on the sensitivity of a conventional PCR assay (19,29). However, in one of the studies storage of samples for 5 d reduced sensitivity (29). Our results also suggest that storage and transport times beyond 48 h at 4°C could reduce the sensitivity of the qPCR assay.
Freezing of samples has traditionally been recommended for long-term storage. We are unaware of any studies of the effect on test performance of freezing preputial samples. Freezing and thawing have been shown to decrease DNA yield by as much as 25% in blood samples (30). Improved sensitivity of PCR assays for other pathogens has, however, been reported when samples are frozen, the hypothesis being that the freeze–thaw cycle may aid in DNA release from cells (31). Studies conducted on stool samples indicated that storage of frozen samples was suitable for downstream DNA applications; however, partial thawing before definitive storage had a detrimental effect, increasing DNA fragmentation (32). Our results also support decreased sensitivity due to short-term freezing at −20°C for 2 h before transfer to long-term storage at −80°C. The very short time at −20°C could have prevented complete freezing and thus exposed the sample to a partial freeze and thaw before storage. Sensitivity, although not significantly different, appeared to be higher for samples stored at −20°C for 48 or 96 h that had sufficient time to completely freeze.
It is important to consider that in this study the transport conditions were simulated in the laboratory, temperatures being held constant. Field conditions and shipping of samples by courier services will likely result in fluctuations in temperature, and the effect of this variation remains unknown. In summary, while there was no overall significant difference across time and temperature combinations, it appears that for short periods of transport, samples can be safely stored in the refrigerator. There was also no evidence of sample damage in the first 48 h if the sample was maintained at 30°C. However, for samples that might be delayed up to 96 h before reaching the laboratory, freezing the sample at −20°C is suggested before shipping on ice, as this option appeared to provide the best sensitivity for storage up to 96 h.
Factors to consider in interpreting these results include the statistical model used to determine sensitivity. We considered a bull to be infected if Cfv was cultured from the prepuce at the beginning and end of the testing period. As previously reported, the potential limitation of this approach is the assumption that Cfv was present and could be detected in each sample (7). Some samples obtained between the times of collection of the samples confirmed to be positive by culture could have had numbers of Cfv undetectable by any testing strategy. However, any differences among proportions of positive bulls for different pool sizes or time and temperature combinations should not have been affected by this limitation, given the random allocation of bulls and samples to study treatments.
The ability to differentiate clearly between the subspecies of C. fetus presents significant challenges. Although certain molecular-based tools such as multilocus sequence typing and amplified fragment length polymorphism analyses (33) are able to distinguish between the 2 subspecies, these technique are laborious and require an isolate. The parA gene, used in the present study, has been questioned as a suitable target for Cfv identification since it has been detected in non-Cfv organisms, including Cff, in certain geographic regions (34–37). However, it is the most thoroughly described Cfv-specific target found to date (7), and both conventional PCR and real-time PCR formats are currently available. In the present study and a previous study by our research group (7), we reported a relatively high percentage of false-positive results in virgin bulls. Acquisition of the parA gene by other non-Cfv organisms could explain this, but further research is needed to evaluate this possibility. A recent study conducted in New Zealand (37) found no association between the percentage of PCR-positive bulls and pregnancy rates when the researchers used the parA gene assay described by McMillen et al (20). Although the assay used in the New Zealand study and the assay evaluated in the present study rely on detection of the parA gene, they use different PCR primers. It remains to be determined whether the difference in primers explains the lack of association with pregnancy outcomes or whether the presence of the parA gene in non-Cfv organisms is limited to certain geographic regions. Attempts to develop new PCR assays based on Cfv-novel targets have failed to increase the specificity beyond that reported for the VenSF-VenSR primers (38), and no additional studies (7) assessing clinical utility with field preputial samples have been reported.
The cow–calf industry requires the ability to efficiently screen large numbers of bulls for Cfv under field conditions. Culture-based methods, although highly specific, lack sensitivity, particularly when prolonged transport is required (16). The real-time PCR assay evaluated in the present study could be used as a preliminary screening method; however, its reported clinical sensitivity and specificity should be taken into account when the results are interpreted. Until more specific diagnostic targets are developed and optimized to high-throughput techniques, this assay could be followed by more specific tests in animals whose results are positive, particularly when no clinical signs of BGC have been observed. Previous reports of the clinical sensitivity and specificity of the direct qPCR assay and the present report of how sensitivity is affected by pooling as well as transport time and temperature provide baseline information needed for applying this tool to BGC control in field situations.
Evaluation of diagnostic strategies for bovine venereal infections is essential to minimizing economically significant reproductive losses in beef herds through identification and removal of infected bulls and to aid in the differential diagnosis of the underlying causes of reproductive failure in beef herds. Further research is needed to optimize the timing of sample collection as well as the amount and frequency of sampling needed to safely establish a negative status.
The authors thank Champika Fernando for technical assistance in the laboratory. This study was supported in part by the Saskatchewan Agriculture Development Fund.