SEARCH

SEARCH BY CITATION

Keywords:

  • Brucella canis ;
  • diagnosis;
  • dog;
  • quantitative polymerase chain reaction;
  • zoonosis

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interests
  9. Funding
  10. References

Canine brucellosis is a reportable zoonotic disease that can lead to canine reproductive losses and human infection through contact with infected urine or other genitourinary secretions. Although many locations require testing and euthanasia of positive dogs, current diagnosis is limited by the time required for seroconversion, for example, presence of B. canis-specific antibodies. The goal of this study was to determine the diagnostic ability of Brucella canis-specific quantitative polymerase chain reaction (qPCR) assay to detect B. canis in field samples prior to serological positivity for faster diagnosis and prevention of transmission within kennels or in households. Two kennels, one of which was located in the owner's home, were sampled following observation of suggestive clinical signs and positive serology of at least one dog. Specimens obtained were comparatively analysed via serology and qPCR analysis. 107 dogs were analysed for B. canis infection via qPCR: 105 via whole-blood samples, 65 via vaginal swab, six via urine and seven via genitourinary tract tissue taken at necropsy. Forty-five dogs were found to be infected with canine brucellosis via qPCR, of which 22 (48.89%) were seropositive. A statistically significant number (= 0.0228) of qPCR-positive dogs, 5/25 (20.00%), seroconverted within a 30-day interval after initial serologic testing. As compared to serology, qPCR analysis of DNA from vaginal swabs had a sensitivity of 92.31% and specificity of 51.92%, and qPCR analysis of DNA from whole-blood samples had a sensitivity of 16.67% and specificity of 100%. B. canis outer membrane protein 25 DNA qPCR from non-invasive vaginal swab and urine samples provided early detection of B. canis infection in dogs prior to detection of antibodies. This assay provides a critical tool to decrease zoonotic spread of canine brucellosis, its associated clinical presentation(s), and emotional and economic repercussions.

Impacts
  • Brucella canis is a reportable zoonotic disease with legal mandates of canine euthanasia for disease control.
  • Diagnosis of brucellosis is difficult due to less-accurate diagnostics that require seroconversion.
  • Quantitative PCR of genitourinary secretions can determine infection prior to seroconversion and may be a better screening test for at-risk dogs prior to adoption into households or kennels.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interests
  9. Funding
  10. References

Brucella canis is a zoonotic coccobacillary bacterium that causes insidious reproductive disorders in dogs (Carmichael and Bruner, 1968; Greene and Carmichael, 2006; Hollett, 2006). The number of B. canis cases in dogs has recently increased throughout the United States and particularly the Midwest (Evans et al., 2005; Kauffman and Kinyon, 2006; Brower et al., 2007). The increase in number of cases corresponds to increased costs associated with reproductive losses from canine brucellosis and emotional losses due to euthanasia of infected dogs. Prevalence in Wisconsin alone has risen from 1.96% in 1995 to 26.8% in 2005 (Brower et al., 2007). Historically, culture of blood or reproductive tissues has served as the diagnostic gold standard (Evans et al., 2005; Greene and Carmichael, 2006; Hollett, 2006). Unfortunately, Brucella is difficult to culture (Greene and Carmichael, 2006; Hollett, 2006). Serological tests cannot adequately detect disease until 8–12 weeks post-infection when detectable levels of anti-Brucella antibody are present (Hollett, 2006).

Laws in many states list canine brucellosis as a reportable disease subject to quarantine (Iowa Code, 2011a,b). In multiple states, testing of all breeding animals and euthanasia of infected animals are instituted once a kennel is quarantined (Greene and Carmichael, 2006; Iowa Code, 2011b). There is no diagnostic test that can determine early infection. Earlier detection of fulminating disease in puppies or newly exposed breeding animals could promote shortening of the quarantine period and would provide better health information to people looking to acquire puppies and/or breeding stock. The purpose of this study was to determine the diagnostic ability of a Brucella canis-specific quantitative polymerase chain reaction (qPCR) assay to detect B. canis in field samples prior to detection via current serologic methods.

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interests
  9. Funding
  10. References

Description of animals

Specimens used for the study were from canines housed at two breeding kennels (Kennel A and Kennel B). These kennels both had a history of stillborn or aborted puppies. Animals at these two sites were identified after at least one dog from each kennel was found to be infected with B. canis via serology and the state veterinarian was notified. The 107 dogs tested in this study were primarily female (75%) and varied greatly in age (4 months to 11 years) and breed (Papillons, Lhasa Apsos, Miniature Australian Shepherds, Shih Tzus and Saint Bernards). At the time of sample collection, none of the dogs were showing overt clinical signs of B. canis infection. Whole-blood samples and tissues were obtained on-site and sent to Iowa State University College of Veterinary Medicine for analysis. Specimens analysed by qPCR included 105 whole-blood samples, 65 vaginal swabs, six urine samples, three testis samples, two uterine samples, one liver sample and one urinary bladder sample (Table 1). Samples were stored at −80°C until we were able to process them for PCR.

Table 1. Vaginal swabs have a higher likelihood of Brucella canis DNA detection via qPCR as compared to whole blood samples
 Dogs testedDogs seropositiveDogs qPCR positive
  1. IFA, Immunofluorescent Assay; RSAT, Rapid Slide Agglutination Test; AGID, Agar Gel Immunodiffusion; 2-ME-RSAT, 2-mercaptoethanol-RSAT.

  2. a

    all samples were taken at the time of euthanasia and were final samples.

  3. b

    20 of 21 dogs were euthanized at this time.

  4. c

    serology was reanalyzed on remaining 74 dogs, and remaining dogs in kennel euthanized based upon these final serology results.

  5. d

    PCR samples were taken at the same time as sampling for first serology, prior to euthanasia of dogs based on AGID results.

Total1073645
Kennel A1356
Assay+/DogsaSpecimen#/Specimens
IFA7/13Whole blood1/13
RSAT3/5Urine3/5
AGID3/5Testis1/3
Uterus2/2
Liver0/1
Bladder0/1
Kennel B943139
Assay+/DogsSpecimen+/Specimens
1st RSAT21/94dWhole blood3/92d
1st 2-ME-RSAT14/21Vaginal swab37/65
1st AGID19/21bUrine1/1
2nd RSAT34/74c
2nd 2-ME-RSAT14/34

DNA extraction

DNA purification from blood, urine, tissue and cotton swab samples was performed using the Qiagen ® QIAamp DNA isolation kit according to manufacturer's protocol (Qiagen, Valencia, CA, USA). Isolated DNA was analysed for concentration and quality via NanoDrop spectrophotometry (Thermo Scientific NanoDrop TM 1000, Vermont Hill, IL, USA).

Multiplex real time qPCR

DNA samples were assayed via qPCR in duplicate of three dilutions (straight, 1 : 10, 1 : 20) using the Mx3005P® qPCR System (Agilent Technologies, La Jolla, CA, USA), a 96-well format and Platinum qPCR SuperMix-UDG Master Mix (Invitrogen, Life technologies, Grand Island, NY, USA). Dilutions and duplicates of two negative controls (nuclease-free deionized water and non-infected canine blood DNA) and two positive controls (DNA isolated from B. canis 6/66 and B. abortus 2308) were included on each 96-well plate for quality control of the assay. Sample wells contained a reaction mixture of 19 μl of master mix and 6 μl of isolated DNA. The limit of detection of the qPCR in 1 ml of canine whole blood was determined via a standard curve to be 19.3 genome equivalents of B. canis DNA. The primer set used in the qPCR assay has been shown to be specific for B. canis and pan-Brucella spp., targeting the outer membrane protein (Omp) 25 gene (Foster et al., 2008). The probes used in the qPCR assay utilized two reporter fluors in one reaction (multiplex qPCR). The 6FAM probe specifically hybridized to B. canis Omp25 DNA, while the VIC probe hybridized to any Brucella species, using B. abortus to identify presence of other Brucella spp. Omp 25 DNA (Foster et al., 2008). Brucella-specific primers were as follows (Integrated DNA Technologies, Iowa City, IA, USA): F 5′-GGCTGGCGCCTTTGCT and R 5′-GGCCCAGGAATAACCTGCAT, TaqMan hydrolysis probes are as follows (Applied Biosystems, Carlsbad, CA, USA): 5′-6FAM-AACTTCCAGAAGGAC-MGBNFQ and 5′-VIC-ACTTCCAGCAGGACC-MGBNFQ. Primers were used at 775 nm and probes at 150 nm with thermocycling parameters of 2 min at 50°C and 2 min at 95°C followed by 50 cycles of 15 s at 95°C and 30 s at 60°C. Results were analysed via MxPro™ QPCR software version 4.01. For each sample, diluted 1 : 10 and 1 : 20 and run at full strength, a sample was considered positive if at least one of the six samples crossed the threshold w/in 40 cycles.

Serology

Serum samples were collected from all animals and sent to diagnostic laboratories by the referring veterinarian for serological testing of canine brucellosis. Serum samples collected from 13 dogs from Kennel A were submitted for Brucella Canine Test, an indirect immunofluorescent assay (IFA) (Antech, Irvine, CA, USA). Sera from animals that tested positive were retested via rapid slide agglutination test (RSAT) and agar gel immunodiffusion (AGID) (Cornell Animal Health Diagnostic Center, Ithaca, NY, USA). Serum samples collected from 94 dogs from Kennel B were first tested via RSAT and 2-mercaptoethanol-RSAT (2-ME-RSAT) (Iowa State University Diagnostic Laboratory, Ames, IA, USA), and when animals were still available for resampling, positive samples confirmed via AGID. As serological confirmation via AGID was not possible for each dog, where AGID serological results were not available, 2-ME-RSAT serology was used as the ‘best’ available result and used to compare qPCR to the combined AGID or 2-ME-RSAT serology.

Statistical analysis

Statistical significance of the comparison of qPCR-positive seronegative samples which a month later were seropositive was determined via a nonparametric Wilcoxon matched-pairs signed-rank test, GraphPad Prism 4 (GraphPad Software Inc, La Jolla, CA, USA) on vaginal swab qPCR-positive dogs with RSAT/2-ME-RSAT results from two time points. P-values below 0.05 were considered significantly different. Values for qPCR sensitivity, specificity, predictive value positive and predictive value negative were calculated according to 2 × 2 contingency table analysis of qPCR versus serological results (RSAT, 2-ME-RSAT and AGID).

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interests
  9. Funding
  10. References

B. canis clinical signs

All dogs enrolled in this study were initially identified via the presence of positive B. canis serological cases in the two kennels reported to the Iowa State Veterinarian. Breeders willing to participate in this study were contacted and informed consent was received to test affected kennels. In both kennels, breeders had been observing late-term abortion and infertility in bitches prior to initial testing.

Serological diagnosis of B. canis leads to conflicting results

Additional serology was performed based on the recommendations of the State Veterinarian after initial testing and index case identification in each kennel. Initially, thirteen dogs from Kennel A were tested via IFA, of which seven dogs were found to be positive for canine brucellosis. After euthanasia of two clinically affected animals, the remaining five seropositive animals were tested via RSAT and AGID, of which three dogs were found to be positive by both of these assays (Table 1). Due to discrepancies between multiple serological tests performed on these thirteen dogs, five animals were determined to be truly seropositive based on a positive AGID result or suggestive clinical signs combined with a positive IFA result. A total of 94 dogs from Kennel B were screened by RSAT, of which 21 dogs tested positive. The 21 RSAT-positive samples were then tested by 2-ME-RSAT, of which fourteen were positive, and by AGID, of which nineteen were positive. After euthanizing twenty dogs, Kennel B re-tested 74 dogs, including two dogs who had initially tested positive by AGID, just over a month (34 days) later. Of the 74 dogs, 34 animals tested positive on RSAT. These 34 animals were retested by 2-ME-RSAT, of which fourteen were determined positive including the two dogs previously determined to be seropositive via AGID. Based on positive AGID results from nineteen dogs on the first date of testing and positive 2-ME-RSAT results from an additional twelve dogs on the second date of testing, 31 animals in total were determined seropositive of 94 dogs from Kennel B. Due to the state of Iowa's current guidelines regarding B. canis seropositive dogs, confirmed positive cases were euthanized and long-term follow-up within these populations was not possible. Kennel B was depopulated after learning the results of the second round of RSAT and 2-ME-RSAT serologic testing, making a second round of PCR testing impossible.

Omp25 DNA qPCR analysis of B. canis infection

In comparison with the serology data, B. canis DNA was detected in six dogs from Kennel A by qPCR analysis. To determine optimal specimens for detection of B. canis DNA, multiple sample types were tested (Fig. 1). Of the different specimens tested, vaginal swabs (56.92%) provided a higher likelihood of detection for B. canis Omp 25 DNA compared with whole blood (3.26%) (Fig. 1). A total of 45 dogs from both kennels were positive for B. canis by qPCR, of which 22 (48.89%) were found to be seropositive (Fig. 2). Of the total dogs tested in the study, 32 vaginal swab samples with positive qPCR results were compared with initial and final 2-ME-RSAT serological results. On the second testing date, 1 month later, five of 25 qPCR-positive dogs (20.00%) with previously undetectable antibody levels now had detectable levels. This is a statistically significant population (= 0.0228) of dogs that were detected as positive for B. canis Omp25 DNA prior to detection of antibody (Fig. 3).

image

Figure 1. Vaginal swabs have a higher likelihood of Brucella canis detection compared with whole-blood samples.

Download figure to PowerPoint

image

Figure 2. Venn diagram indicating overlap between seropositive and qPCR-positive tests of all serology-positive samples.

Download figure to PowerPoint

image

Figure 3. Brucella canis qPCR is able to detect infected dogs prior to detection of antibody.

Download figure to PowerPoint

Sensitivity and specificity of B. canis Omp25 DNA qPCR

The diagnostic value of this B. canis Omp 25 DNA qPCR assay was determined via 2 × 2 contingency tables comparing qPCR results to serology (Table 2). Separate tables were compiled for whole-blood, vaginal swabs and total samples. As compared to combined serology (2-ME-RSAT and AGID assays), whole-blood samples were not sensitive for perceived B. canis infection, with a sensitivity of 16.67%, but were very specific for B. canis infection with 100% specificity. Comparatively, samples taken from the genitourinary tract, vaginal swabs, had much higher sensitivity, 92.31%, but lower test specificity, 51.92%. It should be noted that the gold standard serology was unlikely to detect infection in recently infected dogs. This phenomenon was reflected in a change in relative sensitivity after a second round of serologic testing (Fig. 3).

Table 2. Tabular comparison of serological results to quantitative PCR for diagnosis of brucellosis
 Initial sample/dogSensitivitySpecificityPositive predictive valueNegative predictive value
  1. Combined serologic results included sequential RSAT, 2-ME-RSAT, and AGID analysis as detailed in Table 1.

  2. Cases were considered to be serologically positive by a positive AGID result or a positive 2-ME-RSAT result in cases where AGID testing was not performed.

  3. RSAT, rapid slide agglutination test; 2-ME-RSAT, 2-mercaptoethanol-RSAT; AGID, agar gel immunodiffusion; qPCR, quantitative polymerase chain reaction.

Whole bloodCombined serology    
qPCRPositiveNegativeTotal 
Positive30316.67%100.00%100.00%83.15%
Negative157489
Total187492
Vaginal swab
qPCR
Positive12253792.31%51.92%32.43%96.43%
Negative12728
Total135265
Total
qPCR
Positive14253973.68%66.67%35.90%90.91%
Negative55055
Total197594

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interests
  9. Funding
  10. References

At present, diagnosis of B. canis infection in dogs is limited by the time it takes to observe a rise in B. canis-specific antibodies and cross-reactivity of other antibodies found in canine serum to B. canis antigen (Carmichael et al., 1984). To limit exposure to this zoonotic disease, we wanted to determine whether qPCR analysis could detect B. canis Omp25 DNA in a variety of field samples (blood, urine, vaginal swab) prior to seroconversion. We compared qPCR results to the current serological detection methods for B. canis infection (RSAT, IFA, AGID). qPCR analysis identified the presence of B. canis Omp25 DNA in multiple dogs prior to antibody detection (Fig. 3). Non-invasive samples from the genitourinary tract, including vaginal swabs and urine, were found to be the most sensitive for detection of B. canis Omp25 DNA via qPCR. Use of these samples would make collection of diagnostic samples within the ability of some dog owners and breeders, highlighting the ability of this diagnostic to potentially streamline Brucella canis screening to decrease canine transmission and human exposure.

Stringent quality control methods are necessary when performing qPCR as a diagnostic tool (Queipo-Ortuno et al., 2008). The qPCR assay discussed herein utilized multiple positive and negative controls on each 96-well test plate. These included DNA isolated directly from purified B. canis and B. abortus in multiple dilutions to provide an on-plate positive control standard curve without inhibition of amplification. This in-well multiplex reaction was additionally able to distinguish B. canis infections from infection with other Brucella spp. by using two differently labelled probes, FAM and VIC, to distinguish the possibility of non-B. canis infection of dogs, as was detected in one canine sample, and may be important in at-risk owners or other host species.

All cases in this study either had IFA or RSAT testing performed to initially diagnose B. canis infection or were kennel mates of such dogs. All positive RSAT cases were automatically followed up by 2-ME-RSAT and/or AGID testing for final diagnosis as is common procedure in many diagnostic laboratories. Each of these serologic tests differs in sensitivity and specificity. RSAT is found to have a sensitivity ranging from 50 to 62.5% and a specificity of 95 to 99.7% (Brown et al., 2007; Pretzer, 2008). 2-ME-RSAT has been shown to have a sensitivity of 50–70% and a specificity of nearly 100% provided the dog was not infected in the two to 3 weeks required for seroconversion and had measurable antibodies at the time of sampling (Carmichael and Shin, 1996). Based on the less than 100% specificities of these assays, particularly in early infection, qPCR may have some diagnostic advantages. Culture was not used as a diagnostic gold standard for these studies, as we did not have the budget for these additional tests. AGID performed by Cornell Animal Health Diagnostic Center is often considered to be the current best diagnostic gold standard and was used for RSAT confirmation whenever possible (Carmichael et al., 1984).

During determination of a ‘best’ diagnostic sample for B. canis Omp25 DNA detection, blood was found to be very specific for B. canis infection (100%, Table 2), but not very sensitive (16.67%) as a diagnostic screening tool. Vaginal swabs were a much more reliable sample for detection of B. canis infection in female dogs as analysed by qPCR, with excellent sensitivity (92.31%) and relatively good specificity (51.92%) (Table 2). These results are similar to previous work using PCR to detect B. canis infection (Keid et al., 2007a,b). While the means of PCR product detection were different in these previous publications, B. canis DNA was detected in blood, vaginal tissue and semen. The previous studies were performed using both symptomatic and asymptomatic dogs at the time of sample collection, while the work presented here focused on transmission from chronically infected dogs to naïve dogs in the kennel, all of whom were asymptomatic at the time of sample collection. These case differences may have influenced the location of the organism, for example blood versus tissue, and thus influenced qPCR results from these samples. qPCR will detect B. canis DNA in blood if the organism is there to detect. As presented here, asymptomatic cases may not be bacteremic at the time of sampling: B. canis was more frequently found in vaginal swabs because the organism has a tropism for reproductive tissue. Hemoculture was not performed as part of this study. We hypothesized that B. canis bacteremia does not last very long after initial infection after which time B. canis Omp25 DNA is found in tissues/samples from the genitourinary tract. The findings in this manuscript, particularly qPCR results compared to serology, are notable when considering that qPCR ‘false positives’ may be truly infected animals that have not yet seroconverted. Alternatively, these ‘false positives’ may have had very low levels of bacterial DNA present in tissue samples but had a robust innate immune response, which cleared systemic infection without developing an adaptive immune response and therefore have not seroconverted. The most novel finding with strong implications for this assay as an improved diagnostic test for B. canis is that qPCR detection of Omp25 DNA detects B. canis infection in dogs prior to detection of antibody (Fig. 3).

The results of this study are very encouraging for the use of B. canis Omp25-specific qPCR as a diagnostic tool for detection of B. canis. Non-invasive diagnostic samples such as vaginal swabs and urine had valuable diagnostic use and could be safely and easily obtained by breeders and veterinarians alike. Further studies including semen and additional urine samples are necessary to fully define the best diagnostic samples for male dogs via B. canis Omp25 qPCR. The potential of this qPCR assay for early detection could be very valuable for elimination of B. canis from kennels and replace the need to demonstrate detectable levels of antibody. Additionally, Omp25 qPCR could be a valuable screening tool for B. canis in newly purchased dogs prior to adding these dogs into a new home or kennel and risking exposure to other household dogs or people. Based on the potential of this assay, several private laboratories and two university-based diagnostic laboratories are currently offering B. canis PCR as a diagnostic test or plan to have it available in the near future. B. canis is a re-emerging infectious disease in the canine breeding industry (Evans et al., 2005; Kauffman and Kinyon, 2006; Brower et al., 2007). Better screening and detection methods will be very useful to prevent further canine spread of this insidious disease and reduce the risk of human exposure. B. canis Omp25 qPCR may be a critical diagnostic component to decrease both economic and emotional effects of canine brucellosis and to reduce the risk of human transmission.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interests
  9. Funding
  10. References

The authors would like to thank Kyle Metz, Clara Haydée Quevedo Salazar, Marie Bockenstedt, Katherine Gibson-Corley and Avanti Sinha for their technical assistance. We thank our collaborating kennels for their support and contribution.

Conflict of Interests

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interests
  9. Funding
  10. References

The authors declared no conflict of interests with respect to the research, authorship and/or publication of this article.

Funding

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interests
  9. Funding
  10. References

This work was supported by the American Kennel Club (AKC) CHF ACORN grants 1159-A and 1352-A.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interests
  9. Funding
  10. References
  • Brower, A., O. Okwumabua, C. Massengill, Q. Meunks, P. Vanderloo, M. Duster, K. Homb and K. Kurth, 2007: Investigation of the spread of Brucella canis via the U.S. interstate dog trade. Int. J. Infect. Dis. 11, 454458.
  • Brown, J., J. L. Blue, R. E. Wooley, D. W. Dreesen and L. E. Carmichael, 1976: A serologic survey of a population of Georgia dogs for Brucella canis and an evaluation of the slide agglutination test. J. Am. Vet. Med. Assoc. 169, 12141216.
  • Carmichael, L. E. and D. W. Bruner, 1968: Characteristics of a newly-recognized species of Brucella responsible for infectious canine abortions. Cornell Vet. 48, 579592.
  • Carmichael, L. E. and S. J. Shin, 1996: Canine brucellosis: a diagnostician's dilemma. Semin. Vet. Med. Surg. (Small Anim). 11, 161165.
  • Carmichael, L. E., S. J. Zoha and R. Flores-Castro, 1984: Problems in the serodiagnosis of canine brucellosis: dog responses to cell wall and internal antigens of Brucella canis. Dev. Biol. Stand. 56, 371383.
  • Evans, L., L. K. Kauffman and R. Cox, 2005: Brucellosis on the rise in canine breeding kennels. Iowa Veterinary Medical Association Update Newsletter. 13, 10.
  • Foster, J. T., R. T. Okinaka, R. Svensson, K. Shaw, B. K. De, R. A. Robison, W. S. Probert, L. J. Kenefic, W. D. Brown and P. Keim, 2008: Real-time PCR assays of single-nucleotide polymorphisms defining the major Brucella clades. J. Clin. Microbiol. 46, 296301.
  • Greene, C. E., and L. E. Carmichael, 2006: Canine brucellosis In: Greene, C. E. (ed.), Infectious Diseases of the Cat and Dog, 3rd ed. pp. 369381. Elsevier, St. Louis.
  • Hollett, R. B., 2006: Canine brucellosis: outbreaks and compliance. Theriogenology 66, 575587.
  • Iowa Code § 163.1(2011a)
  • Iowa Code § 163.2 (2011b)
  • Kauffman, L. and J. Kinyon, 2006: Canine Brucellosis on the rise in Iowa breeding kennels. IVMA Update News. 14, 3.
  • Keid, L. B., R. M. Soares, S. A. Vasconcellos, D. P. Chiebao, J. Megid, V. R. Salgado and L. J. Richtzenhain, 2007a: A polymerase chain reaction for the detection of Brucella canis in semen of naturally infected dogs. Theriogenology 67, 12031210.
  • Keid, L. B., R. M. Soares, S. A. Vasconcellos, D. P. Chiebao, V. R. Salgado, J. Megid and L. J. Richtzenhain, 2007b: A polymerase chain reaction for detection of Brucella canis in vaginal swabs of naturally infected bitches. Theriogenology 68, 12601270.
  • Pretzer, S. D., 2008: Bacterial and protozoal causes of pregnancy loss in the bitch and queen. Theriogenology 70, 320326.
  • Queipo-Ortuno, M. I., F. Tena, J. D. Colmenero and P. Morata, 2008: Comparison of seven commercial DNA extraction kits for the recovery of Brucella DNA from spiked human serum samples using real-time PCR. Eur. J. Clin. Microbiol. Infect. Dis. 27, 109114.