Canine Brucellosis: Insights Into the Epidemiologic Situation in Europe

Brucella canis is one of many responsible pathogens of discospondylitis in dogs and infections require specific management. Little is known about the epidemiologic situation in Europe. The purpose of the study was to get insights into the occurrence of brucellosis in dogs in Europe. The database of a European veterinary laboratory was screened for Brucella positive samples. Additionally, medical records of a veterinary hospital in Germany were screened for diagnosis of discospondylitis and brucellosis. The laboratory received samples from 20 European countries for Brucella testing in dogs: 3.7% of submitted samples were Brucella spp. PCR-positive (61/1,657), and Brucella canis antibodies were identified in 5.4% of submitted samples (150/2,764). Brucella spp. PCR-positive samples originated from Spain (11.1% of submitted samples), Poland (6.7% of submitted samples) and rarely from Italy and France. Samples with Brucella canis antibodies originated from 13 European countries (Sweden, Belgium, Austria, Switzerland, Italy, Finland, Germany, Denmark, Hungary, Norway, Poland, France, Netherlands). Young dogs (0–24 months) had a 5.4-fold increased risk of PCR positive samples. The supplementary medical records search identified four young female dogs (7–30 months) with Brucella canis discospondylitis in Germany. The four dogs had been imported to Germany from Eastern European countries (Moldavia, Romania, Macedonia). In conclusion, infection with Brucella canis needs to be considered in dogs in Europe and diagnostics for Brucella canis infection appear indicated in young dogs with discospondylitis.


INTRODUCTION
Veterinary neurologists are frequently confronted with spinal pain in dogs. One differential diagnosis of spinal pain is discospondylitis. The most common etiologies of discospondylitis in dogs are Staphylococcus species, and less frequently Streptococcus species, Escherichia coli, Aspergillus species and Brucella canis or Brucella suis, as well. Routine diagnosis relies on spinal radiographs, CT or MRI, blood cultures and needle aspirates of effected disc spaces using fluoroscopy or CT. A special approach is required for diagnosis of Brucella discospondylitis (1)(2)(3).
Brucella canis (B. canis) is a gram-negative, facultative intracellular coccobacillus which has been reported in many regions of the world and is considered endemic in Southern USA, in Central and South America and in Mexico (4)(5)(6)(7)(8). B. canis also occurs in Canada (9). Frequent reports of infections with B. canis also originate from Asia (China, Japan, India) and Africa (Nigeria, Zimbabwe) (5,(10)(11)(12)(13). It is exotic in Australia and does not occur in New Zealand (14)(15)(16). In Germany, B. canis was reported in 1976 in a colony of Beagle dogs (17) and 2003 in one male dog with epididymitis and orchitis (18). Rare cases originate also from other European countries, such as Sweden (19,20), the United Kingdom (21,22), Austria (23), Italy (24), and Hungary (25). It is generally assumed that countries with a large stray dog population have a higher prevalence of infection, since stray dogs can contribute to distribution and retention of this organism in dog populations (6,13,26,27). Complementary, breeding colonies are at increased risk of maintaining the infection, as well (28).
B. canis discospondylitis is the most frequently reported manifestation outside the genital tract (24,(29)(30)(31)(32)(33), others are generalized lymphadenopathy (19), intraocular inflammation (34,35), and rarely osteomyelitis (36) and meningoencephalitis (37) with similar appearance as neurobrucellosis in humans (38). Common consequences of B. canis infection in dogs are late abortion, stillbirth, and failure to conceive in female dogs (8,19,39,40) and epididymitis, orchitis, prostatitis and infertility in male dogs (24,29,41). Puppies can be born, which have very high risk of perinatal mortality (37). A major concern is that B. canis can cause a lifelong infection with intermittent shedding of bacteria (42,43). Awareness of canine brucellosis as a zoonosis increased in the last years (7,44,45), although symptomatic human infections are considered rare (46). Transmission of B. canis from dogs to humans is possible. Immunocompromized people with close contact to infected dogs (44,47,48), and laboratory workers handling infected specimen are considered at risk for infection (46,(49)(50)(51). Yet, surveillance for brucellosis as a zoonotic disease commonly focuses on B. melitensis, B. suis and B. abortus (52)(53)(54). Less attention has been paid to B. canis in Europe, although dogs usually live in close contact with their owners.
The purpose of the study was to get insights into the occurrence of Brucella canis in dogs in Europe.
We screened the database of a European laboratory for Brucella positive samples. Additionally, medical records of a German veterinary hospital were reviewed for dogs with Brucella canis discospondylitis and their geographic origin.

MATERIALS AND METHODS
The database of a veterinary diagnostic laboratory (IDEXX laboratories, Ludwigsburg, Germany), which received samples from dogs for B. canis testing from 20 European countries, was investigated (2011-2016). The laboratory had received 4,421 samples from dogs for testing for Brucella infection: 1,657 samples were submitted for detection of Brucella spp. with polymerase chain reaction (PCR), and 2,764 samples were submitted for detection of B. canis antibodies. Four samples were marked as originating from the same two dogs (Supplementary Tables 1-3). Samples were submitted by veterinarians or dog breeders. Polymerase chain reaction was a real-time PCR (IDEXX RealPCR TM ), which amplified the 76 bp-sequence of the internal transcribed spacer gene region of Brucella spp. The IDEXX RealPCR TM detects B. canis, B. microti, B. melitensis, B. abortus, B. suis, and B. ovis. Sequencing was not performed. Antibody test was routinely performed with an agglutination test (Institute of Hygiene and Infectious Diseases of Animals, JLU Gießen, Gießen, Germany). B. canis strain RM 6/66 was used as antigen. After growth on tryptone soy bean agar plates in a 5% CO 2 enriched atmosphere for 48 h at 37 • C bacteria were harvested and suspended in 0.15 M phosphate buffered saline (PBS) (pH 7.2). The suspension was filtered through four layers of gauze and subsequently heated in a water bath at 56 • C for 90 min to inactivate the microorganism. After washing the bacteria twice in PBS the last pellet was suspended in the same buffer to an about 10-fold higher concentration as required for the agglutination test. As a preservative Merthiolate was added (final concentration of 0.01%). For the agglutination test the concentrated antigen was diluted in 0.15 M NaCl giving a turbidity of McFarland no. 5. Each dilution (0.5 ml) of a log 2 dilution series in 0.15 M NaCl, beginning with 1:25, of the field sera was mixed with the same volume of the antigen suspension giving a final serum dilution of 1:50, and then incubated for 48 h at 37 • C in a humidified atmosphere. The reciprocal value of the last dilution which still revealed an at least 50% agglutination of the B. canis cells was recorded as titer. For controls known negative (field serum) and positive dog sera were used. The latter was from a dog experimentally infected with B. canis. According to Carmichael and Greene a titer of 100 was chosen as cut off value (43,55,56).
The following data were retrieved from the laboratory database: Submitted materials for diagnostic testing for B. The strength of association between positive PCR or antibody presence and sex and age was estimated by calculation of odds ratios (ORs). An odds ratio with a 95% CI excluding 1 was considered to indicate a significant association at the 5% level.
As a secondary supplementary study, the medical records of a neurology service in a German veterinary hospital were reviewed for dogs with a diagnosis of discospondylitis and Brucella canis infection. Data on diagnostics, traveling history and geographic origin of the dogs were extracted (Supplementary Table 4).
The ethics committee of the veterinary faculty LMU Munich approved the study (114-16-02-2018).

RESULTS
The European veterinary diagnostic laboratory had received 4,421 samples from 4,419 dogs for testing for canine brucellosis  Tables 2, 3. within a 5-year period ( Table 1). Samples submitted to the laboratory for Brucella spp. PCR originated from 15 European countries. PCR was positive in 3.7% (61/1,657) of all submitted samples. Brucella spp. PCR-positive samples originated from four European countries: Spain, Poland, Italy and France ( Figure 1A). Sample sizes for each country are described in Table 2. Young dogs (0-2 years old) had a 5.4-fold increased risk for being PCR-positive (OR 5.4; CI 95% 2.2-13.7) compared to dogs aged 3 years of age or older (OR 0.2; CI 95% 0.1-0.5). Statistical analysis failed to demonstrate an association between the sex of the dogs and positive Brucella spp. PCR. Samples submitted to the laboratory for B. canis antibody testing originated from 20 European countries. Presence of antibodies was documented in 5.4% (150/2,764) of all submitted samples. Samples with B. canis antibodies originated from 13 European countries: Sweden, Belgium, Austria, Switzerland, Italy, Finland, Germany, Denmark, Hungary, Norway, Poland, France and the Netherlands (Figure 1B). Sample sizes for each country are described in Table 3. Young dogs had a lower risk of being antibody-positive (OR 0.6; CI 95% 0.4-0.9). Conversely, an increased risk for B. canis antibody positivity was observed in dogs aged 3 years or older (OR 1.6; CI 95% 1.1-2.2). Statistical analysis failed to demonstrate an association between the sex of the dogs and the presence of Brucella canis antibodies.
The medical records search identified four dogs with B. canis infection and discospondylitis (Figures 2A,B). The four dogs (3 female-spayed, 1 female-intact; age 7-30 months) originated from Eastern European countries (1 Moldavia, 2 Romania, 1 North Macedonia) and had been brought to Germany between one and 23 months prior to presentation. None of the dogs had been used for breeding purposes. Diagnosis of brucellosis was based on growth on bacterial blood cultures and bacteria identification by MALDI-TOF mass spectrometry (3 dogs) or presence of B. canis antibodies [1 dog, tested twice, agglutination test and indirect fluorescent antibody test (IFAT 1:512)]. Two additional multiplex PCR assays (Bruce-ladder and New Bruceladder) confirmed growth of B. canis on blood cultures (3 dogs). Using these methods most of the known Brucella species can be identified by specific band patterns of amplificates of different sizes. It is even possible to differentiate between B. suis biovars and B. canis (57,58).

DISCUSSION
This study provides insights into the regional occurrence of B. canis in dogs in Europe. The evaluation of the European laboratory's database showed that PCR-and antibody-positive samples originated from dogs from a variety of European countries, suggesting the widespread presence of B. canis in Europe. Brucella spp. DNA was present in 3.7% of all submitted samples and B. canis antibodies in 5.4% of samples submitted for diagnostic testing for canine Brucella infection from 20 European countries. Yet, the true occurrence of B. canis infection in dogs in Europe remains unknown. Data presented here reveal only occurrence in samples in which testing for B. canis was specifically requested by the veterinarian or originate from populations at risk and, thus, are not representative of the exact occurrence in the dog population of the respective country. Limitations of the present study are that the results refer to samples from preselected dogs and do not reflect countrywide occurrence. The samples were collected from a heterogeneous group of dogs, so the proportions of positive tested samples must be interpreted critically and cannot measure a real country prevalence of canine brucellosis. Another limitation in the interpretation of the data is the non-species-specific PCR used by the laboratory. Dogs are susceptible to infections with B.  abortus, B. suis and B. melitensis. Infections with other Brucella spp. than B. canis should be considered in countries in which brucellosis is not yet eradicated, if the dog is fed raw pig meat or is a hunting dog or is housed on a farm (2,16,59). In addition, number of samples from several countries was insufficient for estimating an accurate frequency of positivity as reflected by the confidence intervals (Tables 2 and 3). In the present study, the laboratory database only specified the country of origin, but otherwise the precise origin of the samples remained unknown.  (19,20), the United Kingdom (21,22), Austria (23), Italy (24), Hungary (25) or Switzerland (61). Bacterial isolation has been considered the only proof of B. canis infection (39,60,62). More recently, PCR of clinical samples has been suggested as a diagnostic test for detection of subclinical and early infections prior to antibody development (63)(64)(65)(66)(67). PCR detects Brucella spp. DNA in samples, and positive results indicate infection with Brucella spp. Thus, results of samples from Spain and Poland suggest that B. canis occurs in these countries (20). In samples from Spain, 11.1% had positive PCR results, and in samples from Poland, 6.7% showed positive PCR results and 3.7% revealed the presence of B. canis antibodies. So far, there is a lack of reports on B. canis infections in Polish dogs. Iwaniak et al. (68) even stated that B. canis had never been confirmed in Poland (68). Yet, a case report from Sweden in 2012 established a connection between a bitch with B. canis infection that aborted repeatedly, and a stud dog from Poland (19). Antibody testing was only available from few Spanish samples, which are not representative, but previously Mateu de Antonia et al. (69) indicated B. canis antibody prevalence for Spanish stray dogs as high as 6.5% (69). There was also a report on a brucellosis outbreak in a Swedish kennel, in which two diseased bitches had been mated with stud dogs from Spain (20). It may also be of interest that the present study revealed B. canis antibodies in 6.9% of samples from Finland, which is in contrast to data from Dalhbom et al. (70), who had shown that 388 dogs from 94 Finnish kennels had no B. canis antibodies (70). An important limitation of epidemiologic studies of Brucella canis based on antibodies is the fact that positive samples were not tested twice. In order to exclude false positive results, confirmation with a second serological test would be advantageous. Thus, further investigations are warranted to investigate regional differences in B. canis occurrence. The results of this study have also implications for the diagnosis of discospondylitis. The gold standard for diagnosis of the specific underlying etiology of discospondylitis is bacterial isolation with blood and urine cultures, or in case of negative blood/urine cultures with cultures of image-guided disc biopsies (1,3). However, advantages of PCR are to detect acute or chronic infections and the possibility to examine blood, urine and secretions. Especially, vaginal swabs or semen are suitable materials for Brucella diagnosis with PCR (63,71). In our study, swabs and aspirates of unknown origin were the most frequently submitted material with positive test results. Combination of different materials for PCR-testing can increase the diagnostic success (71). Consequently, routine testing with PCR from several specimens could enable rapid diagnosis of canine brucellosis in practice and may also be employed for future epidemiologic investigations. The laboratory data shows an increased risk (5.4-fold) of PCR-positive samples in young dogs up to 2 years of age and confirms the overrepresentation of young dogs in previous case reports (19,22,29,35). Another study strengthens this consideration, here, 21% of 200 canine neonates had a positive B. canis-PCR (72). Adjacent to vertical transmissions between bitches and their puppies via placenta, during parturition or weaning, potential other modes of infection are venereal transmission in sexually mature dogs and infection via mucosal membranes of oropharynx or conjunctiva (urine, vaginal secretions) (37,73). B. canis is well-known to cause abortions and stillbirth in infected bitches. However, puppies can be born and may either die shortly after birth or appear seemingly healthy and develop the disease later (19,37). Thus, chronically infected surviving puppies could present a serious source for harbourage and spread of B. canis (25).
In general, treatment of B. canis diseased dogs is controversially discussed because of the risk of harboring this agent lifelong and its zoonotic potential (19,39). Veterinary literature considers treatment of B. canis infection in dogs with combined use of tetracyclines and aminoglycosides (26,29,43). WHO guidelines recommend combined treatment with tetracycline or doxycycline and an aminoglycoside antibiotic or rifampicin in humans (74,75). Combination of doxycycline and rifampicin seems to be effective in dogs infected with B. suis, as well (2). Treatment with enrofloxacin eliminated clinical signs and frequent abortions in a kennel with twelve infected breeding dogs (76). A new approach was the additional administration of hydroxychloroquine, which improved clinical signs at an early stage, and reduced treatment failures and relapses in human patients with brucellosis (77). This drug has already been used as an immunomodulatory drug in dogs with lupus erythematosus or lymphoma (78,79). Indeed, no antibiotic therapy appears to eliminate B. canis completely from affected dogs, and treatment failures and relapses are frequently reported (19,26,39). Literature considers the zoonotic potential of B. canis is low compared to B. melitensis, B. suis and B. abortus, which are more frequently reported as underlying cause of human brucellosis (52,(80)(81)(82)(83). Less attention has been paid to B. canis in Europe, although dogs usually live in close contact with their owners and there is an increasing incidence of dog trafficking and import of puppies from breeding kennels with poor state of health (19,84). An increasing number of case reports describes B. canis infections in people, especially in immunocompromized adults or children (7,9,47,48,50,51,(85)(86)(87)(88)(89). Clinically conspicuous humans exhibit unspecific signs like undulant fever, fatigue, weakness, lymphadenopathy, liver and spleen enlargement (7,44,47,49,50,90).
In summary, reasons for the occurrence of B. canis in Europe could be poor health in selected breeding kennels, the rising international dog trade of breeding animals (5,19) and presence of stray dog populations (6) in some Southern and Eastern European countries (91,92). Breeding for commercial purposes in very poor housing conditions without veterinary care may constitute additional risks. Import of those puppies could promote transboundary carry-over of B. canis infections in dog populations (84,(93)(94)(95)(96).

CONCLUSION
In conclusion, infection with Brucella canis needs to be considered in dogs in Europe and diagnostics for Brucella canis infection appear indicated in young dogs with discospondylitis. PCR for rapid diagnosis of Brucella infection may be included in the routine work-up of dogs with discospondylitis.

ETHICS STATEMENT
The ethics committee of the Faculty of Veterinary Medicine at LMU Munich approved the study (114-16-02-2018).

AUTHOR CONTRIBUTIONS
GB, AF, and KH designed and coordinated the study. FP provided the laboratory data. GB, AF, and KH determined the evaluation of the clinical and laboratory data. WH designed the antibody testing. GW designed the microbiological culture of clinical cases. FM designed the PCR examinations of clinical cases. GB, AF, KH, and WH wrote the manuscript. All authors read and approved the final manuscript.