• Open Access

Recurrent Osteomyelitis in a Cat due to Infection with Bartonella vinsonii subsp. berkhoffii Genotype II


  • M. Varanat,

    1. Intracellular Pathogens Research Laboratory, Center for Comparative Medicine and Translational Research, College of Veterinary Medicine, North Carolina State University (NCSU), Raleigh, NC
    Search for more papers by this author
  • A. Travis,

    1. River Road Animal Hospital, Southport, NC
    Search for more papers by this author
  • W. Lee,

    1. Intracellular Pathogens Research Laboratory, Center for Comparative Medicine and Translational Research, College of Veterinary Medicine, North Carolina State University (NCSU), Raleigh, NC
    Search for more papers by this author
  • R.G. Maggi,

    1. Intracellular Pathogens Research Laboratory, Center for Comparative Medicine and Translational Research, College of Veterinary Medicine, North Carolina State University (NCSU), Raleigh, NC
    Search for more papers by this author
  • S.A. Bissett,

    1. Intracellular Pathogens Research Laboratory, Center for Comparative Medicine and Translational Research, College of Veterinary Medicine, North Carolina State University (NCSU), Raleigh, NC
    Search for more papers by this author
  • K.E. Linder,

    1. Intracellular Pathogens Research Laboratory, Center for Comparative Medicine and Translational Research, College of Veterinary Medicine, North Carolina State University (NCSU), Raleigh, NC
    Search for more papers by this author
  • E.B. Breitschwerdt

    1. Intracellular Pathogens Research Laboratory, Center for Comparative Medicine and Translational Research, College of Veterinary Medicine, North Carolina State University (NCSU), Raleigh, NC
    Search for more papers by this author

Corresponding author: Dr Edward B. Breitschwerdt, Department of Clinical Sciences, College of Veterinary Medicine, North Carolina State University, 4700 Hillsborough Street, Raleigh, NC 27606; e-mail: ed_breitschwerdt@ncsu.edu.


Bartonella alpha proteobacteria growth medium

A 4-year-old spayed-female Domestic Shorthair cat, obtained from a shelter as a kitten, was examined in May 2007 because of an intermittent lameness. Rectal temperature was 102°F. The cat weighed 3.3 kg. There was a focal, nonpainful, 2.5-cm-diameter firm swelling at the medial aspect of the left metatarsal region. Radiographs of the left rear leg identified lysis of the distal aspect of the 1st metatarsal bone and adjacent soft tissue swelling. Thoracic radiographs were unremarkable. Neoplasia was suspected and digital amputation was elected. Preanesthetic blood tests identified thrombocytopenia (25,000/μL; reference range, 200,000–500,000) with an adequate estimated platelet count because of clumping, mild lymphopenia (1,071 lymphocytes/μL; reference range, 1,200–8,000), mild hyperglobulinemia (5.4 g/dL; reference range, 2.3–5.3), and negative FIV and FeLV ELISA test results. The neutrophil count was 4,725/μL (reference range, 2,500–8,500) with no bands or neutrophil toxicity. Urinalysis was unremarkable.

At surgery, the left medial metatarsophalangeal joint was disarticulated and the digit was removed. Histopathologic evaluation identified a mixed inflammatory response, characterized by a large number of well-differentiated plasma cells, macrophages with erythrophagocytosis and focal aggregates of neutrophils. Pyogranulomatous osteomyelitis was diagnosed. Plasma cell numbers were sufficiently high that plasma cell neoplasia was considered a plausible differential diagnosis. Fungal, acid-fast, and silver stains did not identify infectious agents. Culture of the amputated tissue was not performed. Two weeks postoperatively, an ulcer formed at the surgical site and was accompanied by moderate muscle atrophy. The distal limb became necrotic from the surgical site to the mid-tibia, which was accompanied by severe atrophy of the left rear leg to the level of the stifle and a loss of distal deep pain, which required mid femoral amputation. At no time had the cat traumatized the original incision. Despite necrosis, there was minimal to no purulent discharge and the cat never appeared overtly uncomfortable, despite rapid deterioration in the limb.

In November 2007, the cat was examined because of lethargy. The cat was thrombocytopenic (97,000/μL; reference range, 175,000–600,000) and neutropenic (1,064/μL; reference range, 2,500–8,500). Urinalysis identified proteinuria and neutrophils suggestive of a urinary tract infection and the cat was treated with amoxicillin-clavulanate 62.5 mg PO q12h for 7 days. Urine culture was not performed.

In February 2008, diffuse distal swelling and lameness developed in the left front leg. Radiographs identified mild soft tissue swelling and mild degenerative joint disease. The cat was treated with meloxicam 0.45 mg PO q24h for 3 days, then q72h to alleviate pain and swelling. Despite this treatment, the cat remained intermittently lame. Swelling and lameness of the left front leg again were documented by physical examination in May 2008, at which time the swelling had become firmer and more localized over the left carpus. There also was thickening of the right carpal region, which was firmer and less pliable than the left carpal region. Radiographs identified osteolysis of the left radial carpal bone with concurrent periosteal reaction. There was also cortical expansion of the right 5th metacarpal bone. Although serum globulin concentration was within reference range, serum protein electrophoresis documented polyclonal gammopathy associated with increased gamma globulins (2.54 g/dL; reference range, 0.5–1.90).

By June 2008, the cat had persistent decreased weight bearing on the left thoracic limb. The left carpus was swollen and painful, and the distal limb had developed a valgus deformity. The right thoracic leg was also painful during carpal palpation and range of motion manipulations, but the owner reported no lameness or pain at home. The lateral phalanges of the right thoracic limb were swollen and firm. The cat was sedated with 0.15 mg dexmeditomidine IM. Radiographs of the left and right thoracic limbs, chest, and right pelvic limb were obtained. Cytology samples were also obtained from the lesions in the left and right thoracic limbs by fine needle aspiration. Radiographs documented an ovoid expansile lesion of the right 5th metacarpal bone, characterized by extensive and irregular osteoproliferation, marked cortical bone destruction, and moderate thickening of the soft tissues lateral to the metacarpal lesion (Fig 1). Severe osteolysis of left radial carpal bone was also present, as well as irregular periosteal proliferation involving all left carpal bones and the proximal aspects of all left metacarpal bones. Moderate left carpal joint effusion was also noted. The radiographic findings of polyostotic osteolysis and irregular osteoproliferation in a patient of this age were consistent with multifocal osteomyelitis of either bacterial or fungal etiology. Given the presence of left carpal joint effusion and involvement of the left carpal bones, septic arthritis of the left carpal joint was considered likely. Soft tissue swelling adjacent to the right 5th metacarpal bone was thought to represent cellulitis. Other aggressive bone lesions (ie, primary or metastatic osseous neoplasia) were considered less likely. Right pelvic limb and thoracic radiographs were within normal limits.

Figure 1.

 (A) Dorsopalmar radiograph of the right pes. Note the aggressive expansile lesion of the right 5th metacarpal bone and adjacent soft tissue swelling (white arrow). (B) Dorsopalmar radiograph of the left pes. Note the severe osteolysis of the radial carpal bone (large white arrow), osteoproliferation of the carpal and metacarpal bones (small white arrow), and carpal joint effusion (white arrowhead).

Cytology from the left carpal lesion identified erythrophagocytosis, nontoxic neutrophils, epithelioid, and multinucleated cells thought to be osteoclasts, scattered plasmacytoid osteoblasts, and numerous benign plasma cells. Organisms were not seen, including lack of viral, anaplasma, or ehrlichial inclusions. The cytology results and clinical information supported chronic neutrophilic and plasmacytic osteomyelitis. Based on cellular appearance and the duration of the disease process, plasma cells were reported to be reactive rather than neoplastic, but myeloma could not be excluded from the differential diagnosis. The cat was treated with tramadol hydrochloride 12.5 mg PO q12h while the owner decided whether to pursue additional diagnostic testing.

In August 2008 the cat was sedated with a combination of 80 mg ketamine, 0.16 mg dexmedetomidine, and 0.27 mg buprenorphine IM. Radiographs showed moderate progression of the osseous lesions identified in June. Blood was obtained for a CBC and serum biochemical profile. With the exception of mild thrombocytopenia (193,000/μL; reference range, 200,000–500,000) potentially associated with platelet clumping, CBC findings were within reference ranges and the neutrophil count was 5,355/μL (reference range, 2,500–8,500). Serum biochemical abnormalities included hypercalcemia (11.6 mg/dL; reference range, 8.2–10.8) and hyperglobulinemia (6.0 g/dL; reference range, 2.3–5.3). Fungal serology for aspergillosis, blastomycosis, coccidiomycosis, and histoplasmosis was negative and repeat FIV and FeLV test results were negative. A bone marrow aspiration sample, obtained to further assess the possibility of a plasma cell myeloma, was unremarkable. Aerobic culture of the bone marrow aspirate yielded mild growth of Salmonella enterica subsp. enterica, which was considered a potential sample collection contaminant. Mycoplasma and fungal cultures of the bone marrow aspirate were negative. Aerobic, anaerobic, fungal, and Mycoplasma cultures of joint fluid, aspirated from the left carpal-metacarpal joint were negative for growth. An aseptically obtained jugular blood sample, submitted to the Vector Borne Diseases Diagnostic Laboratory, NCSU-CVM, was cultured in BAPGM (Bartonella alpha proteobacteria growth medium).1 Although awaiting test results, the cat was treated with 20 mg PO q12h and azithromycin 20 mg PO q48h for presumptive intracellular bacterial infection. Bartonella spp. PCR following direct extraction from the blood sample was negative. After a 7-day incubation period, Bartonella vinsonii subsp. berkhoffii genotype II DNA was PCR amplified and sequenced from the BAPGM enrichment blood culture.1Bartonella henselae DNA and DNA from a healthy dog served as positive and negative controls, respectively. Subculture onto an agar plate did not result in bacterial growth. By IFA testing, B. vinsonii subsp. berkhoffii and B. henselae antibody titers were 1 : 128 and 1 : 64, respectively.

The cat was treated for bartonellosis for 3 months with azithromycin 36 mg (10 mg/kg) PO q48h and concurrently with amoxicillin-clavulanate 62.5 mg PO q12h for 2 months for possible infection with S. enterica. During this time period, the owner reported a progressive increase in weight bearing on the left forelimb and decreased pain when walking, and meloxicam and tramadol were discontinued. In February 2009, repeated BAPGM blood culture failed to detect B. vinsonii subsp. berkhoffii DNA or growth of bacteria. B. vinsonii subsp. berkhoffii and B. henselae antibody titers were negative. The cat continued to be less painful, but both the left and right carpi were grossly larger than in November 2008. Radiographs showed continued progression of the previously described thoracic limb lesions. Radiographs of the right pelvic limb and thorax remained within normal limits. Azithromycin was discontinued once blood culture results became available. Reculture was recommended, but not performed 1 month after cessation of antibiotics. As of June 2009, the cat was healthy, was walking without pain, and had not received additional antibiotics or pain medications.

After successful blood culture detection of B. vinsonii subsp. berkhoffii in August 2008, the original formalin-fixed, paraffin-embedded left medial metatarsal osteomyelitis lesion, was sent to the Intracellular Pathogens Research Laboratory to determine if the cat was infected with B. vinsonii subsp. berkhoffii at the time of digital amputation. B. vinsonii subsp. berkhoffii genotype II DNA again was amplified and sequenced from the original lesion.

B. vinsonii subsp. berkhoffii genotype I was isolated for the 1st time in 1993 from a dog with shifting leg lameness, epistaxis, recent-onset seizures, and endocarditis.2 Subsequently, 3 additional genotypes (designated II–IV), all of which have been implicated as a cause of endocarditis in dogs, were described based upon sequence differences in the Bartonella 16S-23S intergenic spacer region and the pap31 gene.3,4 In pet dogs, both seroprevalence studies and blood culture isolation results indicate infrequent exposure to or active infection with any of the 4 B. vinsonii subsp. berkhoffii genotypes, whereas exposure is more frequent in rural and working dogs, and in coyotes and feral dog populations.1,3,5,6 On a comparative medical basis, dogs and humans infected with Bartonella spp. can develop similar disease manifestations, including endocarditis, granulomatous lymphadenitis, granulomatous hepatitis, bacillary angiomatosis, peliosis hepatis, seizures, and arthritis.5,6 Although seemingly well-adapted on an evolutionary basis to induce persistent infection in canine reservoir hosts (eg, coyotes), B. vinsonii subsp. berkhoffii has only rarely been isolated from pet dogs, foxes, or humans.2,3 To our knowledge, this case report describes the 1st isolation of B. vinsonii subsp. berkhoffii from a cat and the 1st association of a Bartonella spp. with osteomyelitis in any animal other than a human being.

B. henselae and Bartonella clarridgeiae are transmitted among cats by Ctenocephalides felis.5,7 Healthy cats are considered reservoir hosts for these 2 Bartonella spp. and are generally nonclinical carriers of these intravascular bacteria.5,7 In various study populations throughout the world, B. henselae bacteremia has been reported in 8–56% of healthy cats whereas prevalence of B. clarridgeiae bacteremia generally is much lower.7 Although disease associated with B. henselae and B. clarridgeiae is not commonly recognized in cats, other non–reservoir-adapted Bartonella spp., such as B. vinsonii subsp. berkhoffii, for which dogs are the only known reservoir hosts,3–5 may be pathogenic when transmitted to cats.

In humans, musculoskeletal manifestations, including osteomyelitis, have been reported as a complication of cat scratch disease (CSD), caused by B. henselae. In a study by Maman et al,8 about 10% of CSD patients developed chronic musculoskeletal complications including myalgia, arthritis, osteomyelitis, and neuralgia. There are also case reports of B. henselae-associated osteomyelitis, most often in children, at times, not accompanied by fever or lymphadenopathy.9,10 Non–host-adapted Bartonella spp. are more likely to be associated with development of pathology, for example B. henselae osteomyelitis in people and B. vinsonii subsp. berkhoffii osteomyelitis in this cat. Because B. vinsonii subsp. berkhoffii genotype II was PCR amplified and sequenced from an enrichment blood culture obtained 15 months after digital amputation and then was retrospectively amplified and sequenced from the paraffin block produced at the time of initial surgery, it is likely that this organism caused persistent bacteremia, recurrent osteomyelitis, and arthritis in this cat. Recently, we described Bartonella spp. DNA carryover when processing necropsy and biopsy tissue samples, which represents a potentially unique problem for the molecular diagnosis of bartonellosis in veterinary medicine.11 In this cat, recurrent disease accompanied by repeated sequencing of an infrequently detected subspecies and defined genotype makes DNA carryover less likely. In addition, special precautions were taken in our laboratory when sampling paraffin blocks to minimize this possibility. Whether infection with B. vinsonii subsp. berkhoffii after digital amputation contributed to postoperative necrosis and what appeared to be ischemic atrophy of the left rear leg could not be determined because tissues from the amputated leg were not submitted for histopathology.

Current recommendations for the treatment of Bartonella musculoskeletal infections are based predominantly on empirical data. Although reinfection cannot be ruled out, it seems likely that B. vinsonii subsp. berkhoffii was not immunologically or therapeutically eliminated after digital and rear leg amputations in this cat. BAPGM enrichment cultures of blood and joint fluid recently were used to document failure of azithromycin or marbofloxacin to eliminate B. henselae and B. vinsonii subsp. berkhoffii from a dog that progressed from nonerosive to erosive polyarthritis.12 Resistance to macrolide antibiotics has been reported for B. henselae,13 but additional studies are required to determine if B. vinsonii subsp. berkhoffii can also develop macrolide resistance. In addition to osteomyelitis, clinical and radiographic findings in this cat suggested chronic, progressive polyarthritis, which also was reported in 3 dogs that were seroreactive to B. vinsonii subsp. berkhoffii antigens.14 As reported in this cat, clinical improvement occurred in these dogs in conjunction with antimicrobial therapy and was accompanied by a rapid decrease in B. vinsonii subsp. berkhoffii antibodies.14 In humans with B. henselae osteomyelitis, prognosis is good with patients being treated for a median duration of 32 days with mono, dual, or triple antibiotic therapy.9

Bartonella sp. are highly fastidious organisms that can be difficult or impossible to isolate by conventional blood culture or to detect by PCR amplification after direct extraction of DNA from patient samples.1,15 These diagnostic limitations are especially applicable when attempting to document infection with a Bartonella spp. in a nonreservoir host.16,17 In conjunction with efforts to enhance the sensitivity of PCR for detection of Bartonella-specific DNA sequences, we have recently incorporated pre-enrichment culture of aseptically obtained diagnostic samples (blood, cerebrospinal, aqueous, and joint fluids and effusions) using a liquid insect cell culture-based medium (BAPGM) before PCR testing.8 Combining pre-enrichment culture with PCR amplification has substantially improved diagnostic sensitivity when testing samples from dogs and humans infected with novel Bartonella species.18 Historically, because B. henselae and B. clarridgeiae could be readily isolated from cat blood after an incubation period of a few weeks in a high CO2 incubator, our laboratory did not initially recommend the use of BAPGM when culturing cat blood samples. As illustrated by this cat, direct PCR from blood was negative, suggesting that an enrichment culture approach may be necessary to diagnose infection with non–reservoir-adapted Bartonella spp. in cats, as has been shown for dogs and human patients.1,4,16,17 In addition, PCR amplification of B. vinsonii subsp. berkhoffii DNA only after enrichment culture supports the presence of viable bacteria in the blood sample, because the initial PCR-negative blood sample was diluted 1 part blood to 10 parts BAPGM before incubation for 7 days.

Various Bartonella spp. are transmitted among reservoir hosts by fleas, lice, sand flies, keds, and possibly biting flies and ticks.19 Transmission to non–reservoir-adapted hosts can occur via an arthropod bite or a scratch or bite by a carrier animal. The mode of B. vinsonii subsp. berkhoffii transmission to this cat is unknown. B. vinsonii subsp. berkhoffii DNA has been amplified and sequenced from saliva obtained from healthy dogs, and dogs have been implicated in the direct transmission of B. henselae to humans.20 Therefore, dog bite transmission could have been a source of infection for this cat. Although there is clinical and epidemiological evidence to support B. vinsonii subsp. berkhoffii transmission by Rhipicephalus sanguineus, tick transmission has not been proven.19

In recent years, Bartonella spp. have been associated with a wide spectrum of inflammatory lesions in dogs and human patients, including endocarditis, encephalitis, meningitis, granulomatous hepatitis, splenitis, and necrotizing granulomatous lymphadenitis.6,21,22 Classically these facultative, intracellular bacteria induce a B-cell–associated granulomatous reaction, characterized by numerous histiocytes, plasmacytoid monocytes, small lymphocytes, multinucleated giant cells, and plasma cells intermingled with monocytoid B-cells.23 Granulomatous inflammation, accompanied by an unusually large number of plasma cells and hyperglobulinemia, was most likely related to B. vinsonii subsp. berkhoffii infection in this cat. Importantly, 2 pathologists thought plasma cell tumor was a diagnostic consideration.

Fever and leukocytosis were not associated with polyarthritis and osteomyelitis in this cat. Thrombocytopenia and neutropenia have been reported in dogs and humans infected with Bartonella spp.24 Although osteomyelitis may have caused hypercalcemia, endogenous production of active vitamin D may have contributed, as reported in twins with CSD.25 Additional studies are needed to determine the frequency of B. vinsonii subsp. berkhoffii infection in cats. Based on this case report, Bartonella spp. infection should be considered in the differential diagnosis of osteomyelitis, hypercalcemia, hyperglobulinemia, and thrombocytopenia in cats.


This study was supported by the State of North Carolina and in part through graduate student stipend or salary support provided by Bayer Animal Health and IDEXX Laboratories.