The prevalence of Pasteurella multocida, a cause of bovine respiratory disease, was studied in a random sample of beef suckler and dairy farms throughout Scotland, by means of a cross-sectional survey. A total of 637 calves from 68 farms from six geographical regions of Scotland were sampled between February and June 2008. Deep nasal swabs were taken, and samples that were culture-positive for P multocida were confirmed by PCR. Prevalence of P multocida was 17 per cent (105 of 616 calves); 47 per cent of farms had at least one positive animal. A higher prevalence was detected in dairy calves than beef calves (P=0.04). It was found that P multocida was associated with Mycoplasma-like organisms (P=0.06) and bovine parainfluenza type 3 virus (BPI-3) (P=0.04), detected by culture and quantitative PCR of nasal swabs, respectively. Detection of P multocida was not associated with bovine respiratory syncytial virus (BRSV), bovine herpesvirus type 1 (BoHV-1) or bovine viral diarrhoea virus (BVDV). Mycoplasma-like organisms, BPI-3, BRSV, BoHV-1 and BVDV were detected in 58, 17, four, 0 and eight calves, on 25, five, two, 0 and five of the 68 farms, respectively.
Statistics from Altmetric.com
IT HAS been estimated that annual losses to the UK cattle industry due to bovine respiratory disease amount to more than £80 million (Barrett 2000). Bovine respiratory disease is a syndrome with generally a multifactorial aetiology; the primary cause of an outbreak can be difficult to ascertain and is likely to include host, pathogen and environmental factors. Pathogens involved may include the bacteria Pasteurella multocida, Mannheimia haemolytica, Histophilus somni and Mycoplasma species, and the viruses bovine respiratory syncytial virus (BRSV), bovine parainfluenza type 3 virus (BPI-3) and bovine herpesvirus type 1 (BoHV-1). Bovine viral diarrhoea virus (BVDV) also has an established role in the calf respiratory disease complex (Potgieter 1997, Al-Haddawi and others 2007). In recent years, vaccines have been developed to aid farmers in the control of BRSV, BPI-3, BoHV-1, BVDV and M haemolytica. In Europe, there is currently no vaccine licensed for use against P multocida. Surveillance data, based on diagnostic submissions to regional veterinary investigation centres, show that P multocida accounts for approximately 15 per cent of bovine respiratory disease in England and Wales (Defra 2006), and for approximately 10 per cent in Scotland (Scottish Agricultural College [SAC] 2006), but may not reflect the situation within the general calf population.
Bacteria of the Pasteurellaceae family often exist as commensal organisms on mucosal surfaces of vertebrate animals, and it appears that they have the potential to be the primary or secondary cause of a range of infectious diseases (Biberstein 1990). Calves can be positive for upper respiratory tract carriage of P multocida within a few days of birth (Dowling and others 2004), and experimental challenge with P multocida alone has been shown to cause pneumonic pasteurellosis in calves (Dowling and others 2002). In addition, early stressors such as weaning or viral infection may precipitate disease (Weekley and others 1998, Dabo and others 2007); therefore, it is important to characterise pathogen carriage during the early weeks of life.
The aim of the present research was to determine, by means of a cross-sectional study, the prevalence of P multocida carriage within the upper respiratory tract of preweaned calves from randomly selected beef suckler and dairy herds located throughout Scotland. Quantitative PCR was used to detect viral RNA/DNA from a relatively non-invasive nasal swab method and to explore associations between bacterial and viral respiratory pathogens in Scottish calves.
Materials and methods
Six geographical areas of Scotland were delineated on the basis of administrative centres (Fig 1). Contact data for farms that had more than 20 beef suckler or 60 dairy cows or heifers in calf or in milk, according to the 2007 June Agricultural Census, were obtained from the Scottish Government. These criteria were used after analysis of a pilot study, indicating that they would generate data that were representative of the population in each of the six regions while giving the study a greater than 0.99 probability of contacting the target number of eligible farms in all regions. Random lists of contact data were generated, stratified by management system and region. The aim was to recruit six beef and six dairy farms from each of the six regions. Recruitment was conducted via telephonic interviews, and farmers were asked a number of questions designed to ascertain eligibility. Eligibility criteria included having a minimum of five calves aged less than 10 weeks at the time of sampling. Information was also collected from farmers who declined sampling (non-compliers) in order to assess potential selection bias.
In the absence of background data for formal sample size calculation, the number of animals to be sampled per farm was estimated using a rule of thumb that 1/expected prevalence × 3 provided an adequate sample size for prevalence estimation. Carriage rate of P multocida in young calves on supply farms used for experimental studies is 25 to 30 per cent (J. C. Hodgson, unpublished data), implying that a sample size of nine to 12 animals for the present study would be adequate. At sampling, if 10 or more eligible calves were present, 10 were sampled, ensuring where possible that proportional numbers of calves from all groups or units of the farm were included as an informal stratification. If there were between five and 10 calves, all were sampled. All sampled calves were examined clinically and records were made of demeanour, any recent treatment, cough, auscultation findings, nasal discharge, diarrhoea and rectal temperature. No calves were assessed as clinically ill. A deep nasal swab was taken from each calf as follows. Any frank discharge or detritus was first cleaned from the nares using a dry paper towel. For bacteriology, a swab (BBL CultureSwab Plus; Becton Dickinson) was inserted full length (12 cm) into the ventral meatus, taking care not to touch the nares. The swab was rotated against the mucous membrane before withdrawing and by repeating the procedure with the same swab on the contralateral side. The swab was then inserted into Amies transport medium without charcoal. Swabbing was repeated for viral sampling, using a plain swab, which was then placed into virus transport medium (E&O Laboratories) containing Hanks balanced salt solution, 1 per cent bovine serum albumin, sodium bicarbonate, phenol red, 1 per cent of antibiotics benzyl penicillin, streptomycin sulphate and polymixin B, and 0.5 per cent nystatin. A jugular venous blood sample was collected into a plain Vacutainer (Becton Dickinson). Samples were either returned that day or, where necessary, sent by overnight post at ambient temperature to the laboratory for immediate processing. All experimental protocols were approved by the Moredun Research Institute Experiments and Ethical Review Committee and authorised under the UK Animals (Scientific Procedures) Act 1986.
Detection of P multocida
Swabs were plated directly on to selective sheep blood agar (5 per cent citrated sheep blood in agar number 2 base with 10 mg/l vancomycin and 10 mg/l amphotericin B; E&O Laboratories). Plates were cultured aerobically overnight at 37°C before assessment. One colony phenotypically resembling P multocida per plate was subcultured to ensure purity of colonies and stored at −70°C in brain-heart infusion broth with glycerol. These isolates were confirmed as P multocida by PCR, using primers specific for the kmt1 gene encoding a protein involved in the synthesis of capsular polysaccharide, as described by Townsend and others (1998). Briefly, frozen material was recultured, and loopfuls of bacterial growth were suspended in 500 μl of nuclease-free water, incubated at 95°C for 10 minutes and the resultant lysates stored at −20°C. Template (20 μl) was added to a reaction mixture (total volume 50 μl) containing 1 × PCR buffer, 1.5mM MgCl2 and 1 unit DNA polymerase (Platinum Taq DNA polymerase; Invitrogen), in addition to 0.2mM of each dNTP (dNTP mix; Invitrogen) and 50 pmol of each primer KMT1SP6 and KMT1T7. Samples were run on a thermal cycler (G-Storm GS1; Gene Technologies), with initial denaturation at 95°C for two minutes, followed by 35 cycles at 95°C for 30 seconds, 55°C for 30 seconds and 72°C for 30 seconds, and a final extension at 72°C for five minutes. PCR products underwent agarose gel electrophoresis on 1 per cent gels in 1 × tris-acetate-EDTA buffer before UV transillumination.
Detection of Mycoplasma species
After plating on to selective sheep blood agar, swabs were smeared directly on to a basic agar (CM401; Oxoid) without phenol red, enriched with G-selective supplement (SR59; E&O Laboratories) to support the growth of Mycoplasma species. Plates were incubated at 37°C in 5 per cent CO2 between three and seven days, a period used to produce typical cultures of M bovis (Nicholas and Ayling 2003), and examined daily for signs of microbial growth. Identification of cultures was restricted to allocating those with a typical ‘fried-egg’ appearance as Mycoplasma-like organisms.
Detection of BRSV, BPI-3, BoHV-1 and BVDV by real-time RT-PCR
Swabs were collected in virus transport medium, mixed in a vortex, and the resulting suspension was stored at −70°C until processing. Samples were thawed and aliquots of each sample were pooled on a per-farm basis (maximum of 10 samples per pool). At the same time, individual aliquots were made and stored at −70°C for individual sample testing in cases where a positive pool result was given. Individual sera were similarly pooled on a per-farm basis, and individual aliquots were retained for individual testing. RNA was extracted from both virus transport medium and sera sample pools (and from individual samples as required) using a commercial kit (QIAamp Viral RNA Mini-kit; QIAGEN). DNA was not removed from the preparation. Samples were stored at −70°C until testing. To investigate the use and sensitivity of the Viral RNA kit for DNA virus detection, a comparison was made between two nucleic acid extraction methods, using serial 10-fold dilutions of BoHV-1 (strain 6660)-infected cell culture supernatants. These dilutions were extracted using both the QIAamp Viral RNA Mini kit (QIAGEN) and the QIAamp Minelute virus spin kit (QIAGEN) following the manufacturer's protocol. Samples were stored at −70°C until testing.
Respiratory virus assays were carried out on RNA extracted from pooled virus transport medium. The Pestivirus (BVDV) assays were carried out on both RNA extracted from pooled virus transport medium and RNA extracted from pooled sera. Where any assay gave a positive result for any pool, the individual samples constituting that pool were tested to identify individual infected animals. For BRSV RNA detection, real-time RT-PCR was carried out as described previously (Willoughby and others 2008b), using SuperScript III Platinum One-Step Quantitative RT-PCR System with ROX (Invitrogen) on an ABI Prism 7000 real-time PCR platform (Applied Biosystems). Each run included no-template extraction and positive controls (Rispoval RS; Pfizer). For BPI-3 RNA detection, the real-time RT-PCR protocol was applied (Willoughby and others 2008a), using the same reagents and platform as described previously. Each run included no-template extraction and two previously characterised UK virus isolates as positive controls (L3047/1 and A2084). For BoHV-1 genome detection, a previously published real-time PCR targeting the gB gene was used (Wang and others 2008), with the Platinum Taq Quantitative PCR System with ROX (Invitrogen) on the ABI Prism 7000. Each run included no-template extraction and a previously characterised UK virus isolate as a positive control (IBR 6660). For pestivirus detection, a multiplex real-time RT-PCR to detect BVDV type 1 (BVDV1), BVDV2 and Border disease viral RNA was used as previously described (Willoughby and others 2006). Each run included no-template extraction and reaction controls: BVDV1 (strain NADL [National Animal Disease Laboratory]), BVDV2 (strain 890) and Border disease virus (Moredun strain) were used as positive controls. All real-time PCR assays included an endogenous internal control targeting β-actin RNA and DNA as required (K. Willoughby, unpublished data) to validate negative results.
Determination of BVDV antibody status
Serum was tested for antibody to BVDV following a standard operating procedure used by Fenton and others (1991). The results were expressed as a ratio matrix of optical density at 450 nm. The cut-off ranges for interpretation (routinely used in diagnosis of BVDV antibodies in serum of the calves older than 12 weeks of age) were: negative, less than 0.100, equivocal, 0.101 to 0.150, and positive, more than 0.150.
Statistical analysis was carried out after exclusion of ineligible samples and animals. Prevalence of P multocida at the animal level was calculated as the number of animals confirmed as positive for P multocida by PCR divided by the total number of animals sampled. Farm prevalence was calculated as the number of farms with at least one positive animal divided by the total number of farms sampled. The within-herd prevalence was calculated as the number of positive animals identified on each farm divided by the total number sampled on that farm. Exact 95 per cent confidence intervals were calculated for prevalences using Minitab version 13 (Armitage and others 2002). Farm prevalences in different regions were compared using variants of Fishers Exact test using StatXact v 6. The housing strategy of the beef farms sampled was also compared with that of those that declined to be sampled using these methods. Farm-level prevalences in beef and dairy farms/animals were compared using the Cochran-Mantel-Haenszel test in StatXact v 6, stratified by region. Animal-level prevalences on positive farms were compared using generalised linear mixed models (GLMMs) (Genstat version 11), thus allowing for clustering at farm stratum. GLMMs were used to quantify the risk that fewer animals were sampled per farm in some regions. National estimates of prevalences (adjusting for the size of the animal population in different regions) were calculated using parametric and non-parametric bootstrapping in Excel 2003 and Genstat v 11, respectively. The univariate association between detection of P multocida and BPI-3 or Mycoplasma species on nasal swabs was assessed using generalised linear models with logistic link functions and binomial responses, allowing for overdispersion associated with clustering at the farm level (Intercooled Stata 8). BVDV status was modelled using a GLMM with logistic link functions and Bernoulli responses, fitting farm as a random effect.
Farmer compliance was 68 per cent of those eligible to take part. Beef farms with calves born outside or turned out soon after calving show weak evidence of under-representation. Of the beef suckler farms sampled, 15 of 33 calved outside or turned calves out within a week of birth. Of farms that declined to be sampled, from which information was gained, 19 of 27 calved outside (P=0.05).
Sampling took place between February and June 2008, during which time 33 beef suckler and 35 dairy farms were sampled. The geographical distribution of farms sampled approximately reflects the distribution of cattle density (Volkova and others 2008).
A total of 637 calves were included in the study. Bacterial samples from 14 animals (2 per cent) were overgrown at culture with Proteus species. These animals were eliminated from analysis, as it could not be ascertained whether P multocida was present. Of the remaining 623 animals, seven calves were older than 10 weeks of age and were also excluded from analysis. The median age of calves included in the analysis (616 animals) was 22 days.
Before any animals were excluded from the study, the majority of farms (58 of 68) had 10 animals sampled (mean 9.37, median 10, range 4 to 10). After exclusions were implemented, dairy farms in the Northern Isles region had appreciably fewer animals included in the analysis per farm (mean 6.1), but this difference was not statistically significant (P=0.17).
The majority of farms (55 of 68) were sampled by staff from the Moredun Research Institute, and the remainder were sampled by staff from the Scottish Agricultural College (SAC) Epidemiology Research Unit.
Detection of P multocida
A total of 105 isolates were identified phenotypically as P multocida and stored at −70°C. All isolates were confirmed by PCR; therefore, the mean animal level prevalence was 17 per cent (105 of 616 animals) (Table 1); on average, 9 per cent of beef and 26 per cent of dairy animals were positive (P=0.04, stratifying by region). However, appreciably higher prevalences in dairy animals were seen only in the south east and Northern Isles regions. The median number of positive animals on positive farms was two animals, with a range from one to 10. With one exception, all beef farms had a low animal-level prevalence of P multocida on nasal swabs; that is, three or fewer calves per farm tested positive. In contrast, on 40 per cent of dairy farms, P multocida was detected in nasal swabs from more than half of the healthy calves that were sampled (Fig 2). Mean farm prevalence was 47 per cent (32 of 68 farms), with 36 per cent and 57 per cent of beef suckler and dairy farms, respectively, having at least one positive animal (P=0.15). There was no statistically significant evidence of differences in prevalence by region for either dairy of beef farms.
National prevalence figures
Allowing for differences in estimated prevalence and in the size of the eligible cattle population in different regions, it was estimated that nationally, 37 per cent (95 per cent confidence interval 22 to 55 per cent) of beef farms and 57 per cent (35 to 78 per cent) of dairy farms were positive for P multocida. Nationally, 8 per cent (3 to 12 per cent) of eligible animals on beef farms and 16 per cent (9 to 25 per cent) of animals on dairy farms were positive for P multocida. On farms observed as positive, 21 per cent (12 to 31 per cent) of animals from beef herds and 29 per cent (17 to 45 per cent) from dairy herds were assessed as positive. The national prevalence estimates for dairy herds are very different from those that might be inferred from simple calculations based on the results shown in Table 1. This is because of the highly heterogeneous distribution of dairy animals across the regions of Scotland, where 63 per cent of eligible animals from the sampling frame were found in the south west region, but only 1 per cent of animals were found in the Highlands and the Northern Isles regions, respectively.
Detection of Mycoplasma species
In total, 58 animals were culture-positive for Mycoplasma-like organisms; 18 of these were also positive for P multocida (Table 2). There was weak evidence for a positive association between the detection of P multocida and Mycoplasma-like organisms in the nasal tract (P=0.06).
Detection of BRSV, BPI-3, BoHV-1 and BVDV in nasal swabs
Using quantitative PCR, BRSV and BPI-3 were detected in the study population (Table 2). For BRSV, four animals from two farms were positive. Both types of BPI-3 (BPI-3a and BPI-3b) were detected (Horwood and others 2008, Willoughby and others 2008a); the latter was only detected on one farm, with half of the sampled animals on that farm (five of 10) giving a positive result. Four other farms gave positive results for BPI-3a. In total, 17 of 39 calves were positive for BPI-3a from these four farms. On one farm, both BPI-3 and BRSV were detected. There was statistically significant evidence of a positive association between the detection of P multocida and BPI-3 in the nasal tract (P=0.04). BVDV1 RNA was detected in five nasal swab pools and in eight individual animal sera from these five farms. Where BVDV RNA was not detected in nasal swab pools, the sera pools were also negative.
Determination of BVDV status based on serology
The proportion of animals seropositive for BVDV on farms ranged from 0 to 100 per cent (median 70 per cent). On four farms, all sampled calves were seronegative; two of these were on Shetland, which is officially BVDV-free. There was a statistically significant difference in the mean proportion of seropositive calves on farms that reported vaccinating cows against BVDV compared with those that did not (P<0.001) (Fig 3). The probability of an animal being seropositive was negatively associated with age of calf (P=0.02), with no evidence that this relationship was different on farms that did or did not vaccinate their animals (P=0.89).
The aim of this cross-sectional survey was to determine the prevalence of carriage of P multocida in the upper respiratory tract of calves on Scottish beef suckler and dairy farms assessed from samples obtained using short (15 cm) nasal swabs. Unpublished observations in the authors' laboratory have shown agreement between isolates obtained from clinically normal calves, using short nasal and long (30 cm), guarded, nasopharyngeal swabs, with no cases of P multocida culture recorded from a long nasopharyngeal swab, which were not detected using a short nasal swab. In addition, analysis by PFGE of isolates obtained by short nasal swab or retrieved from diseased lung tissue from the same calf showed the same banding pattern. These observations indicated that detection of P multocida by nasal swab would be a good indicator of the P multocida status in the nasopharynx and lungs. Others, using alternative sampling methods, reported good agreement comparing upper and lower respiratory tract samples (Allen and others 1991, DeRosa and others 2000). In the authors' opinion, the use of swabs for surveys is preferable to bronchoalveolar lavage (BAL) because BAL is fairly invasive and, in studies of clinically normal populations, may not be justified.
Almost half of the farms sampled had at least one positive animal. The study provided evidence of clustering of colonisation on dairy farms but did not show any such effect in the case of beef animals. There was a lower mean prevalence of carriage of P multocida in beef suckler calves, compared with dairy calves of the same age. This was not unexpected and probably reflected less close contact between calves, less stress, better maternal care and passive transfer of immunity experienced by outdoor beef suckler compared with indoor-housed dairy calves. Indeed, it has been shown that confinement and stress when beef calves are weaned at approximately six months of age coincide with an increase in bovine respiratory disease (Gibbs 2001).
To minimise any variation due to seasonal effects, the field work was carried out within a defined temporal window that included the spring calving period for Scottish beef suckler farms. Pilot studies indicated that few farms calve solely in autumn months; autumn-born calves were not included in the study. Management of neonates in autumn may vary significantly from spring-born calves, particularly on beef farms, and results from the current study may not be representative of this population.
The incidence of detection of P multocida in nasal swabs was lower (approximate range 30 to 60 per cent) than that observed in other published work relating to healthy calves (Allen and others 1992, Catry and others 2006). This may be due to a number of factors, including differences in management systems that influence calf-to-calf contact. For example, Allen and others (1992) studied beef calves in feedlots, which were not included in the present study. Another reason may have been a low survival rate of P multocida during transport in the present study, reducing in turn the sensitivity of its detection by culture. However, P multocida can survive for 10 days at 21°C in Amies transport medium (Tefera and Smola 2002), and it is noteworthy that the highest percentage (100) of swabs taken during the current survey that were positive for P multocida originated from Orkney and, therefore, had spent the longest time in transit at ambient temperatures. Furthermore, an average incidence of 25 per cent in dairy calves aged up to one week has been observed routinely in swabs processed immediately after sampling in the author's laboratory, from calves on Scottish supply farms selected for challenge studies (J. C. Hodgson, unpublished data). A further possibility is that bacteria of interest could be outcompeted for nutrients by more prolific bacterial species or those overgrown by contaminating species. However, only 2 per cent of samples were overgrown by contaminants (Proteus species) – a small and acceptable rate of contamination, given that samples were collected under field conditions – and unlikely to have biased the prevalence estimates. These observations indicate that the rates of colonisation of young calves by P multocida reported in the present study are not low due to any sampling, transport or analytical considerations. Specificity of P multocida detection was likely to be high, as isolates were confirmed using a validated P multocida-specific PCR; primers were confirmed as specific by a database search (Primer-BLAST; NCBI).
M haemolytica was not detected on selective sheep blood agar plates in the course of the present study. M haemolytica is considered to be the main cause of pneumonic pasteurellosis in beef cattle; therefore, this result was surprising. Reasons for this may include the widespread use of vaccines effective against M haemolytica. In dairy cattle, the findings are consistent with the view that P multocida is more frequently isolated from cases of enzootic calf pneumonia (Maheswaran and others 2002). The co-occurrence of P multocida with other pathogens is frequently reported (Autio and others 2007, Dabo and others 2007), and evidence was found for relationships between the presence of P multocida and BPI-3 or Mycoplasma-like organisms in the nasal tract. Given the detection of Mycoplasma-like organisms in a considerable proportion of calves, future studies should investigate the prevalence and relevance of such organisms in more detail following the use of molecular methods such as PCR and sequencing to identify definitively the involvement of Mycoplasma species.
The present study detected respiratory viruses from single nasal swabs using quantitative PCR. The technique may also be useful in clinically affected animals, for which many studies have used paired serology to identify active infection of respiratory viruses. This has some disadvantages, as results are retrospective and can be difficult to interpret, especially in young animals with maternally derived antibodies. In addition, two visits are necessary, which may decrease farmer compliance, particularly in the more extensive beef suckler systems. There may also be loss to follow-up due to male dairy and dairy-cross calves being sold before the second visit.
Taking into account the age of calves sampled, seropositivity to BVDV was probably due to maternally derived antibodies, as indicated by a negative relationship with the age of the calf, and the higher prevalence on farms where dams were vaccinated against BVDV. The detection of BVDV RNA may represent a persistent or transient infection; serial testing is required to determine this with certainty.
Further work will include exploring the associations between detection of P multocida by nasal swab and any subsequent development of clinical signs of bovine respiratory disease, as well as investigating how the contemporaneous presence of bacteria and viruses in the nose may relate to disease. Studies of differences between isolates at the molecular level are also being carried out to explore the molecular epidemiology of P multocida within this calf population, which will add to the current knowledge of disease pathogenesis and transmission.
The authors thank the farmers and veterinarians for their participation, members of the Pasteurella group and Epidemiology and Population Biology division (Moredun Research Institute) and Andrew Brownlow and members of the Epidemiology Research Unit (SAC Inverness) for help with sampling, and Janice Gilray and Madeleine Maley for help processing the viral samples. The study was part-funded by the Scottish Government via its Centre of Excellence in Epidemiology, Population Health and Infectious Disease Control (EPIC) and part-funded by the Pasteurella group at the Moredun Research Institute.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.