Severe Bartonella henselae bone infection in a kidney transplanted young man
- Heidi Dahl Christensen 1,
- Ann Brinch Madelung 2,
- Anne Lerberg Nielsen 3 and
- Fredrikke Christie Knudtzen 4
- 1 Department of Nephrology, Odense University Hospital, Odense, Denmark
- 2 Department of Clinical Pathology, Odense University Hospital, Odense, Denmark
- 3 Department of Nuclear Medicine, Odense University Hospital, Odense, Denmark
- 4 Department of Infectious Diseases, Odense University Hospital, Odense, Denmark
- Correspondence to Dr Heidi Dahl Christensen; Heidi.dahl.christensen@rsyd.dk
Abstract
We present a case of a young kidney transplanted man. He was admitted with lymphadenopathy, fluctuating fever and night sweats 2 months after a cat bite. After admission, he developed severe pain around his right hip. An 18F-fluorodeoxyglucose (FDG)-positron emission tomography/CT revealed intense FDG-uptake in lymph nodes, spleen and bone, suggestive of lymphoma. An extracted lymph node showed confluent granulomas, microabscesses with neutrophils and scattered multinucleated giant cells histologically. The patient had history of latent tuberculosis and proteinase 3 -anti-neutrophil cytoplasmic antibodies associated (PR3-ANCA) vasculitis, making differential diagnostic considerations complicated. Bartonella henselae antibodies was detected in blood and B. henselae DNA in a lymph node. He was started on doxycycline and rifampicin. Due to severe drug interactions with both tacrolimus and increasing morphine doses, rifampicin was changed to azithromycin. He received 12 days of relevant antibiotic treatment and responded well. He was discharged after 16 days with close follow-up and was still in habitual condition 12 months later.
Background
Cat scratch disease (CSD) is caused by the bacteria Bartonella henselae. Transmission most commonly occurs by exogenous acquisition through bites or scratches from a cat.1 In immunocompetent individuals, an infection with B. henselae is often mild and self-limiting. Symptoms previously described in immunocompromised patients with CSD are fever, fatigue, headache, myalgia, arthralgia, night sweats among others, and findings of lymphadenopathy, skin lesions, splenomegaly, hepatomegaly and chorioretinitis.2
In this case, we present a young, immunocompromised cat owner with a rarely diagnosed manifestation of CSD. We will discuss how his medical history caused difficulties with the diagnosis, and how drug interactions complicated his treatment course.
Case presentation
A kidney transplanted man in his early 20s was admitted to a university hospital due to lymphadenopathy in the spring of 2020. He was diagnosed almost 10 years ago with proteinase 3-anti-neutrophil cytoplasmic antibodies (PR3-ANCA)-associated vasculitis with renal and extra renal manifestations, including upper respiratory tract and neurological manifestations. He was initially treated with plasmapheresis, haemodialysis, steroids and cyclophosphamide, later with rituximab. He regained his kidney function after 6 months of therapy and received maintenance therapy with mycophenolate mofetil and steroids. After another 11 months, the patient entered end-stage renal failure and returned to haemodialysis. During transplant investigation, he was found to have a positive interferon gamma release assay test and was treated with rifampicin and isoniazid according to Danish guidelines for latent tuberculosis (TB). Active TB was not suspected. He received a deceased donor kidney in 2015. There were donor-specific antibodies, and he received immunosuppressive therapy according to international guidelines.3
In 2020 he was admitted with inguinal lymphadenopathy and a 3-day history of night sweats and fluctuating fever (figure 1). He had no neurological, pulmonary, gastrointestinal, ocular or urogenital complaints at admission. He was living with a cat but could not remember recent scratches. However, the patient did recall being bitten on the hand 2 months prior. After the bite, there were no swelling, pain or signs of infection. He did not recall any bleeding after the bite. He did not receive any postbite healthcare or oral antibiotic prophylaxis. At admission, there were no longer any visible signs of the cat bite.
A timeline of events during hospital admission, presented chronologically.
Eight days after admission, increasingly severe muscle and bone pain occurred, primarily around the right hip and gluteal region. There was no trauma to the hip or visible or palpable changes. Twelve days into the admission, he developed cough and dyspnoea and required oxygen therapy for 2 days.
Investigations
At admission blood samples showed near normal inflammatory markers with a C reactive protein of 10 mg/L (normal range <6 mg/L) and a normal white cell count (5.1 × 109/L, normal range 3.5–8.8 × 109/L). His creatinine was 170 µmol/L, only slightly elevated compared with his baseline around 150 µmol/L. An enlarged inguinal lymph node (3.5 × 5 cm) was surgically removed the following day. Histopathological findings showed typical CSD histology with confluent granulomas, microabscesses with neutrophils and scattered multinucleated giant cells (figure 2).4
Histopathology of affected lymph node, H&E stain. (A) At low magnification, the lymph node architecture is distorted by areas of confluent granulomas. (B) High magnification showing a multinucleated giant cell (arrow) and (C) microabscess with neutrophils (arrow) surrounded by a rim of epithelioid histiocytes.
On the third day of admission, an 18F-fluorodeoxyglucose (FDG)-positron emission tomography (PET)/CT showed intense FDG-uptake in the hilum of the liver and in enlarged axillary, iliac and inguinal lymph nodes on the left side (figure 3). The largest lymph node was measured to 1.4 cm at the hepatic hilus. The spleen was enlarged with several small areas of intense FDG-uptake with no morphological correlation on non-contrast CT. In the skeleton, focal intense FDG-uptake was revealed in the fifth lumbar vertebra, in both sides of the pelvis and in the right femoral head correlating with a small osteolytic lesion in one of the sites (left iliac bone). The pattern of lesions was suggestive of lymphoma.
18F-fluorodeoxyglucose (FDG)-positron emission tomography (PET), maximal intensity projection (A), axial PET/CT through upper abdomen (B), lower abdomen (C) and inguinal area (D) show intense FDG-uptake in lymph nodes (red arrows), in small areas in the spleen (blue arrows) and in the skeleton (yellow arrows). Dotted green line shows kidney transplant in the lower left side of the abdomen.
Ten days after admission, we received the results of the immunofluorescence assay based bartonella serology, with a detected B. henselae IgG with a titre of 1:2048, IgM not detected. All other microbiological tests performed were negative (table 1). The diagnosis was later confirmed by PCR using microbiome sequencing of the 16S rRNA gene and a species-specific B. henselae PCR on the extracted inguinal lymph node.5
Microbiological results
| Serum | Results |
| Blood cultures | No growth (at six different times) |
| Epstein-Barr virus, PCR | Not detected |
| Cytomegalovirus, PCR | Not detected |
| Varicella zoster virus, serology | IgM not detected, IgG detected |
| Parvovirus, serology | IgM not detected, IgG detected |
| BK polyomavirus, PCR | Not detected |
| HIV Ag/Ab | Not detected |
| Viral hepatitis B+C, serology | Not detected |
| Herpes simplex virus 1+2, serology | Not detected |
| Bartonella henselae, serology | IgG detected 1:2048 titre IgM not detected <1:20 titre |
| Bartonella quintana, serology | IgG +IgM not detected |
| Francisella tularensis, serology | <1:10 titre |
| Brucella abortus, serology | <1:25 titre |
| Brucella melitensis, serology | <1:25 titre |
| Toxoplasma gondii, serology | IgM +IgG not detected |
| Sputum | |
| SARS-CoV-2, PCR | Not detected |
| Influenza A+B, PCR | Not detected |
| Legionella pneumoniae, PCR | Not detected |
| Mycoplasma pneumoniae, PCR | Not detected |
| Chlamydia pneumoniae, PCR | Not detected |
| Chlamydia psittaci, PCR | Not detected |
| Urine | |
| Bacterial cultures | No growth (at two different times) |
| Lymph node | |
| Microbiome bacteria 16S DNA | B. henselae, detected |
| Microbiome parasite 18S DNA | Not detected |
| Microbiome fungi 18S DNA | Not detected |
| PCR B. henselae DNA | Detected |
| PCR B. quintana DNA | Not detected |
| PCR F. tularensis DNA | Not detected |
ANCA-associated vasculitis was excluded by no elevation of myeloperoxidase (MPO) or PR3 titers. Mycobacterial infection was ruled out based on microscopy of the lymph node with Ziehl-Nielsen staining without acid-fast bacilli, PCR non-detectable for Mycobacterium tuberculosis complex and cultivation for 56 days without mycobacterial growth.
A transesophageal echocardiography showed biventricular good contractility and normal valve conditions with no evidence of infectious endocarditis. An MRI scan of the left hip confirmed the results from the PET-CT; In the iliac bone, there was a focal area of 2.5 cm in diameter with oedema. No malignancy was suspected and there was no joint effusion. A chest X-ray, at time of oxygen need, showed discrete bilateral pneumonia in the lower lobes, without pleural effusion. Pulmonary findings were not seen on the PET-CT performed 10 days prior, and he only had minor symptoms of pneumonia. A ventilation/perfusion scintigraphy excluded pulmonary embolisms.
Differential diagnosis
Based on the symptoms of fever in this immunocompromised patient, we wanted to rule out multiple viral, bacterial and fungal infections. There was a risk of reactivation of previously known infection with Epstein-Barr virus and cytomegalovirus. To rule out reactivation of Epstein-Barr virus was especially important due to the association with post-transplant lymphoproliferative disorder (PTLD). With lymphadenopathy and the patient’s history of latent tuberculosis and ANCA-associated vasculitis, examinations for these diseases were performed. Based on the PET-CT, lymphoma was suspected and ruled out through histopathological examination of the extracted lymph node.
Treatment
On the fifth day of admission, before serology and PCR results were available, the patient was started on tablet doxycycline 100 mg and tablet rifampicin 300 mg, both two times per day on suspicion of CSD, based on clinical findings and pathology results. Initially azithromycin was recommended in addition to doxycycline, but due to the severity of the findings on PET-CT it was changed to rifampicin.
The patient’s usual dose of tacrolimus was 18 mg daily. We measured blood levels of tacrolimus 3–4 times a week during rifampicin treatment because of the known interaction between rifampicin and tacrolimus. Tacrolimus was gradually increased to 60 mg daily but was still at subtherapeutic levels (<5 ng/mL). The severe musculoskeletal pain required increasing morphine dosing from initially 20 mg/day to a maximum of 167.5 mg daily. On this dose, the patient was barely pain free. After 7 days of antibiotic treatment the patient developed a generalised rash. Therefore, and because of the severe drug interactions, rifampicin was discontinued and replaced with azithromycin 500 mg daily for the remainder of the treatment duration. Hereafter tacrolimus and morphine doses were slowly reduced.
Outcome and follow-up
The patient was discharged after 16 days of admittance, without fever and no longer requiring morphine. Antibiotics were discontinued at discharge, after a total treatment duration of 12 days. At consultation 1 week later, the patient was pain-free and in a habitual state. At follow-up 12 months later there were no symptoms or complications related to CSD. A PET-CT performed 20 months after admittance showed no FDG-uptake in lymph nodes or bones. He was advised not to keep the cat.
Discussion
This case describes a severe presentation of B. henselae infection with skeletal involvement. The clinical manifestations of fever and lymphadenopathy are not uncommon in solid organ transplanted patients and requires thorough and extensive investigations including testing for CSD. Of relevant differential diagnoses, PTLD is one of the more common malignancies found in solid organ transplanted patients and is a result of immunosuppression.6 The cumulative incidence of PTLD 20 years post-transplant has been found to be 3.5 per 100 person-years. PTLD was our primary diagnostic suspicion at admission. Investigations to rule out ANCA-associated vasculitis, as well as tuberculosis, were also performed as the symptoms were compatible with these conditions, and was already known to the patient. These differential diagnoses were especially relevant to consider due to the necrotising granulomatous inflammation found in the extracted lymph node.
A case series and literature review from 2021 found that Bartonella can induce vasculitis-like symptoms and induce ANCA positivity, particularly PR3.7 However, these patients most often had Bartonella endocarditis. In previous studies up to 60% of patients with Bartonella endocarditis have been found to be PR3 ANCA positive.8 These patients were most often without classic vasculitis ear-nose-throat granulomatous symptoms. In the case of our patient, CSD was not suspected when he was diagnosed with ANCA-associated vasculitis in 2014.
A granulomatous fungal infection, though rare, especially in northern Europe, was also considered among differential diagnosis. They primarily present with skin lesions, but lymphadenopathy and slowly developing generalised symptoms have been described in the literature.9 Invasive fungal infections such as candidiasis or aspergillosis are more common fungal infections in immunocompromised patients.10 To rule out these infections, microbiome 18S DNA testing was performed on the extracted lymph node and blood cultures were cultivated for fungal infection. This case and other cases from the literature highlight the importance of rapid pathological investigations of lymph nodes in immunocompromised patients with lymphadenopathy.11
The majority of cases with CSD are found in paediatric patients or young adults.12 A review found 47 cases of CSD with bone infection, where the included patients had a median age of 9 years. Only two patients were immunocompromised.13 The vertebral and pelvic regions were the most common areas involved, which concurs with the findings in our case. Previous studies have shown that the sites of lymphadenopathy and osteomyelitis do not always correlate, indicating that the bacterial spread can happen either via lymphatic vessels or haematogenously.14
Due to the relatively prompt treatment after symptom onset and the clinical response to treatment, the antibiotic treatment was discontinued after 12 days. Other similar paediatric cases have received antibiotic treatment for an average of 3 months (5 days to 12 months).11 Doxycycline or azithromycin is the recommended treatment options in invasive CSD, but there is no conclusive evidence on the optimal duration of treatment, as there are no high-quality studies available that have investigated treatment duration in disseminated CSD. In a systematic review of 51 immunocompetent children with CSD with bone involvement, the median antibiotic treatment duration was 23 days (IQR 20–42 days).15 In this study, 12.5% of the patients did not receive antibiotic treatment and still had favourable outcomes, only 1 of 51 children had a relapse. As our patient was immunocompromised and had bone involvement, one could argue that there would have been indication for a longer duration of antibiotic treatment. However, the close monitoring of the patient after discharge without suspicion of relapse led us to conclude that 12 days of combination antibiotic therapy was sufficient in this case. At follow-up 12 months after treatment, there was still no sign of recurrent disease. More studies are needed on optimal treatment duration of the different manifestations of CSD in immunocompromised patients.
The patient developed symptoms and chest X-ray findings compatible with pneumonia during his hospitalisation. Pulmonary findings are rarely described in CSD.16 Examinations from sputum did not reveal the aetiology of the pulmonary infection, which resolved spontaneously within a few days. We cannot know if this represented a hospital-associated pneumonia, a viral infection or pulmonary infection with B. henselae. As the patient had been on relevant antibiotic treatment targeting B. henselae for 7 days when the pneumonia occurred, CSD seems a less likely cause of the pulmonary symptoms.
During the time of rifampicin treatment, our patient experienced severe drug interactions. Rifampicin induces CYP3A metabolism, whereas tacrolimus inhibits CYP3A. Therefore, this combination results in lower concentrations of tacrolimus. A study performed on healthy young men showed an increase in clearance of tacrolimus up to 47% when simultaneously treated with rifampicin.17 Likewise, rifampicin and morphine interact via CYP3A4, where the area under the serum concentration-time curve for morphine is reduced by approximately 30% when given concomitantly with rifampicin.18 These interactions became clinically very evident in our case and resulted in painful days for the patient. The interactions could have had consequences for the transplanted kidney if there had been less focus on frequent blood tacrolimus measurements, leading to more insufficient immunosuppression than was already the case.
The American Society of Transplantation recommends that patients after solid organ transplantations pay extra attention to hygiene when handling cats, especially kittens, to limit the risk of contracting B. henselae.19 If they plan to acquire new pets, it is recommended to wait 6–12 months post-transplantation, so that the patient’s immunosuppressive drugs are at a lower, stable dose.
With this case, we want to draw attention to the importance of differential diagnostics in infected immunocompromised patients with fever, lymphadenopathy and suspected bone infection. In addition, this case serves as a reminder of the risk of drug interactions with immunosuppressive drugs and the complications that follows.
Learning points
-
A swollen lymph node in an immunocompromised patient should lead to a positron emission tomography-CT.
-
Bartonella henselae can present with severe, invasive disease in immunocompromised patients.
-
Cat scratch disease is an important differential diagnosis in bone infections in immunocompromised patients.
-
Always consider the many drug interactions between tacrolimus and antimicrobial drugs in transplant patients with infections.
Ethics statements
Patient consent for publication
Footnotes
-
Contributors HDC: Main author of this paper and corresponding author. Clinical investigation of this case. Planning, conduct, reporting, conception and design, acquisition of data or analysis and interpretation of data. ABM: Creator of pathology images as well as primary interpreter of these. In addition, participation in the overall process of reporting and analysing data. ALN: Creator of PET-CT images as well as primary interpreter of these. In addition, participation in the overall process of reporting and analysing data. FCK: Planning, conduct, reporting, conception and design, acquisition of data or analysis and interpretation of data.
-
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
-
Case reports provide a valuable learning resource for the scientific community and can indicate areas of interest for future research. They should not be used in isolation to guide treatment choices or public health policy.
-
Competing interests None declared.
-
Provenance and peer review Not commissioned; externally peer reviewed.
- © BMJ Publishing Group Limited 2022. No commercial re-use. See rights and permissions. Published by BMJ.
References
Use of this content is subject to our disclaimer