A previously healthy 11-year-old immunocompetent Caucasian male presented to the orthopedics department with a 2-week history of lower thoracic spine pain. There was no history of trauma or recent infection, and it was indicated that the complaint evolved progressively over a few days. The pain was originally considered to be of muscle origin and was offered pain relief and bed rest. Four days later, he re-presented with low-grade fever (up to 38.5°C), no rigors, but with worsening back pain that was now radiating to the anterior left hemithorax. The patient was admitted under the orthopedic team for evaluation and treatment.
On examination, the patient was in good general condition, alert with normal perfusion time and not tachycardic. There was only localized tenderness next to the lower thoracic vertebrae but not on the spine per se; this did not cause limitation on the range of movements. The rest of the physical examination was unremarkable. Initial laboratory evaluation showed a white blood cell count of 12600/μL (72.6% neutrophils, 16.0% lymphocytes), erythrocyte sedimentation rate of 90 mm/h and C-reactive protein of 34 mg/L. Blood cultures were taken for common bacterial pathogens. Cardiac ultrasound was normal, and chest radiogram was unremarkable. A bone scan was scheduled, and empiric therapy for possible vertebral osteomyelitis/discitis was initiated with intravenous clindamycin. A Tc99 hydroxymethylene diphosphonate (HDP) bone scan revealed an abnormal enhancement at the T7 vertebral body suggesting possible bacterial osteomyelitis. Treatment continued with the same antibiotic, but although the patient was clinically better on day 6 of therapy (less pain on palpation), fever continued, and advice was requested from the Infectious Diseases team.
On repeat questioning, it was revealed that the patient had frequent contact and abrasions from kittens, and he was occasionally consuming unpasteurized dairy products. No other risk factors were identified. Examination still revealed the localized tenderness on the paraspinal area.
Because of these findings, a magnetic resonance imaging (MRI) scan of the spine was requested, and serology for Bartonella species, Brucella melitensis, and a tuberculin skin test for tuberculosis was performed.
The contrast-enhanced MRI (Figure 1A) revealed abnormal signal at the T6–T7 vertebral bodies. In particular, there was an architectural distortion of the T7 vertebral body with left cortical defect and abnormal enhancement post contrast administration. Additionally, there was a large anterior/left paraspinal soft tissue mass extending from T5 down to T9 and through the T8–T9 neural foramen into the spinal canal without obvious infiltration of the spinal cord. Post contrast, the lesion demonstrates avid enhancement and areas of necrosis with no obvious abnormalities of the visualized intervertebral spaces. A paraspinal mass biopsy performed and showed evidence of chronic inflammation with granuloma formation. After these findings, antibiotic therapy was changed to doxycycline and gentamicin targeting possible Bartonella or Brucella infection. On day 2 of therapy with the new regimen, fever subsided, and erythrocyte sedimentation rate dropped from 70 to 24 mm/h. Clinical improvement continued especially on local back tenderness.
Five days into doxycycline and gentamicin, Bartonella henselae serology was strongly positive for IgM and IgG (1:512, normal value <1:64). Serology for Brucella was negative; blood cultures for common bacterial pathogens and the tuberculin skin test were also negative. Based on these findings, the diagnosis of B. henselae infection was established.
Our patient received 2 weeks of intravenous therapy and was discharged home on oral antibiotics (doxycycline and rifampicin) to continue until a repeat MRI scan was scheduled 4 weeks after the first imaging. Serology for Bartonella was performed again on week 4 of therapy. IgM antibodies were borderline positive and IgG remained strongly positive. Blood inflammatory markers were improving (C-reactive protein: 4 mg/L and erythrocyte sedimentation rate: 20 mm/h). Repeat MRI (Figure 1B) indicated significant improvement. There was a major decrease on the size of the paraspinal soft tissue mass, which was now expanding from T6 down to T8 while there was only a minor architectural distortion of T7. The intervertebral discs were normal.
Our patient received a total of 3-month therapy. On regular follow-up, he was completely asymptomatic, fully mobile, although it was recommended to avoid intense physical activity for at least 6 months.
Frequency of osteomyelitis in children varies in different parts of the world with a recent review estimating its incidence between 1.94 and 13/100,000 in developed countries and 43 and 200/100,000 in developing countries.1 The predominant causative organism is Staphylococcus aureus. Vertebral involvement is uncommon and was found in 3.8% of pediatric osteomyelitis cases.1 The main causes of pediatric discitis and spondylodiscitis are Staphylococcus aureus, Kingella kingae, Coagulase-negative Staphylococcus, Streptococcus pneumoniae, Escherichia coli, Salmonella spp, Mycobacterium tuberculosis, Brucella spp and fungi.2
Cat scratch disease (CSD) in childhood is typically benign, commonly presenting with regional lymphadenopathy and is usually caused by B. henselae. B. henselae infection, however, can manifest with less common clinical forms including prolonged fever, Parinaud oculoglandular syndrome, endocarditis, encephalopathy, neuroretinitis and hepatosplenic disease.
Bone involvement in CSD is rare with reports describing its incidence up to 0.2%.3,4 Usually, the disease is not multifocal, and most commonly described in the vertebrae.5 Diagnosis of osteomyelitis caused by B. henselae can be very challenging because culture isolation requires 2- to 6-week incubation, and polymerase chain reaction lacks in sensitivity.6 Serology can be a considerable tool in the quest for identifying B. henselae as the causative agent.
The literature does not provide robust evidence on the treatment of the nontypical CSD as the existing data derive small case studies and case reports. Similarly, there is no consensus on the choice of agent and duration of treatment of bone infection caused by B. henselae. The 2004 recommendation by Rolain et al7 for complicated CSD is doxycycline and rifampin but with no recommendation on the duration of treatment. The 2007 review by Hajjaji et al5 is proof of the variability in treatment both in the choice of agents and in the duration; agents selected for treatment were beta-lactams, macrolides, aminoglycosides, fluoroquinolones, rifampin and co-trimoxazole either as monotherapy or as combined treatment with a median duration of 32 days (interquartile range: 21–63 days). Similarly, a review of 13 cases of B. henselae osteomyelitis from 2007 to 2013 reveals a wide variability in use of treating agents, such as macrolides or beta-lactams monotherapy to combinations of clarithromycin, clindamycin and rifampin. Of note, only aminoglycosides have in vitro bactericidal activity.7
Bartonella osteomyelitis should be considered in patients with unusual presentation. Diagnosis, antibiotic choice and duration of therapy remain challenging.
1. Dartnell J, Ramachandran M, Katchburian M. Haematogenous acute and subacute paediatric osteomyelitis: a systematic review of the literature. J Bone Joint Surg Br. 2012;94:584–595.
2. Principi N, Esposito S. Infectious discitis and spondylodiscitis in children. Int J Mol Sci. 2016;17:539.
3. Carithers HA. Cat-scratch disease. An overview based on a study of 1,200 patients. Am J Dis Child. 1985;139:1124–1133.
4. Maman E, Bickels J, Ephros M, et al. Musculoskeletal manifestations of cat scratch disease. Clin Infect Dis. 2007;45:1535–1540.
5. Hajjaji N, Hocqueloux L, Kerdraon R, et al. Bone infection in cat-scratch disease: a review of the literature. J Infect. 2007;54:417–421.
6. Florin TA, Zaoutis TE, Zaoutis LB. Beyond cat scratch disease: widening spectrum of Bartonella henselae
infection. Pediatrics. 2008;121:e1413–e1425.
7. Rolain JM, Brouqui P, Koehler JE, et al. Recommendations for treatment of human infections caused by Bartonella
species. Antimicrob Agents Chemother. 2004;48:1921–1933.