Cat scratch disease sepsis in an immunocompromised patient

Abstract

As a greater number of households have pets, the likelihood of zoonotic infections can rise too. Although in most healthy individuals these infections are self-limiting, they are more serious and can lead to adverse outcomes in the immunocompromised. There is minimal information available for the immunocompromised patient who are pet owners or on pet handling. We report a case of cat scratch disease-related sepsis in an immunocompromised patient. This case illustrates the need for a detailed history, including a pet history, in immunocompromised patients presenting with fever of unknown origin and the need for UK-based information for the immunocompromised on pet care and risks associated with having a pet.

Background

Pet ownership and contact is associated with many mental, physical and social benefits, but are a potential source of infection too.¹ The true incidence of zoonotic disease is difficult to identify, partly because these diseases are not notifiable and thus, there is a paucity of data at the local or national level.² Additionally, data are often under-reported, given confounding factors, such as non-pet exposure pathways, subclinical shedding by pets and pets being vectors to arthropod-borne diseases.^{2 3} Inasmuch as the risk of zoonotic disease transmission from pet to owner is very small,⁴ high-risk patients (those aged below 5 and over 65 years, pregnant women and immunocompromised) are at higher risk of zoonotic infections compared with the general population.³ In fact, high-risk patients have been observed to have more severe disease lasting longer and worsening secondary complications.^{2 3}

It can be assumed that high-risk patients are at a greater risk for zoonotic infections due to varying immune-related mechanisms and can be broadly categorised into primary (inherited/genetic) or secondary (acquired) immunodeficiency causes (figure 1).⁵ Secondary immunodeficiencies are more common and less life-threatening in itself.⁵ Acquired immunodeficiency can result from drugs (direct suppression of the immune cells or steroid-sparing agents), immunosuppressing infections and metabolic diseases physiological immunosuppression (figure 1).⁵ For the sake of brevity, we highlight the reviews by Travejo et al and Stull et al as good summaries on zoonotic infections in the immunocompromised.^{2 3 5}

Figure 1.

Broad categorisation of immunodeficiencies.

Cat scratch disease (CSD) is a zoonotic infection that adversely affects high-risk patients.⁶ As CSD is not on the list of Public Health England’s notifiable disease, the true epidemiology of this disease is not known. Recent studies from the USA, however, estimated that it affects 9.4 cases/100 000 of the children aged 5–6 years⁷ and up to 4.7/100 000 persons aged <65 years.⁶ Data on the incidence in immunocompromised patients, however, are lacking, and thus, there is no information available for cat-owning patients in the UK who are immunocompromised or have been initiated on immunosuppressive therapies.

We report the case of a 47-year-old cat owner who presented to the hospital with CSD-induced sepsis on mycophenolate mofetil (MMF).

Case presentation

A 47-year-old man with a background of scleritis and relapsing polychondritis, presented to the hospital in Feb 2020, with a 3-week history of fevers, night sweats and rigours. Besides this, he reported no other cardiac, respiratory or abdominal symptoms. There was no history of recent travel. He had first developed symptoms of relapsing polychondritis during the summer of 2017. He was started on a high dose of oral prednisolone towards the end of 2017. MMF was started in February 2018, at which point, steroid taper was commenced. His relapsing polychondritis was stable on 2.5 g daily, in divided doses, of MMF and 3 mg of prednisolone. On advice from his clinical nurse specialist, the patient had stopped taking his MMF 6 days prior to admission.

On examination, he looked unwell and was shivering. His chest auscultated clear, with normal heart sounds and his abdomen was palpated to be as soft and non-tender. His admission blood tests revealed a worsening of his chronic thrombocytopenia (platelets 49×10⁹/L), raised C reactive protein (100 mg/L) and normal neutrophils (3.25×10⁹/L). His radiograph and urinalysis were both described to be normal. His observations were suggestive of sepsis (respiratory rate 30 breaths/min, peripheral oxygen saturation of 97% on room air, heart rate 107 bpm, blood pressure 87/66 mm Hg and temperature 39.3°C) and as such, he was started on intravenous fluids and broad-spectrum antibiotics. The patient was commenced on amoxicillin 1 g three times a day, weight-based gentamicin 430 mg once daily and metronidazole 500 mg three times a day, which is based on our trust’s guidelines for sepsis of unclear source. After consultant review on day 2, intra-abdominal sepsis was thought unlikely, and thence, the regime was changed to co-amoxiclav 1.2 g three times a day (metronidazole and gentamicin were stopped).

Investigations

Two blood cultures failed to grow anything after 5 days of incubation. Enteric PCR pathogen testing did not detect salmonella, shigella, campylobacter, Escherichia coli, cryptosporidium, giardia or Clostridium difficile. His hepatitis/HIV serology also tested negative. While he was an inpatient, he became more pyrexic and hypotensive, requiring further boluses of intravenous fluids and careful fluid monitoring. A CT abdomen and pelvis with contrast summarised that the patient had splenomegaly with significant right axillary lymphadenopathy. A further examination of the right axilla noted a tender 2×2 cm lymph node and blanching plaque-like rash on the posterior trunk. Further characterisation with CT of the thorax revealed enlarged right axillary lymph nodes, largest 3 cm in short axis. A transthoracic echocardiogram revealed no obvious vegetation.

An opinion from haematology was sought, given the possible differential of lymphoma. They suggested the need for an urgent biopsy of the lymph node and initiation of allopurinol and a cover of 1 unit of platelets to cover the surgical procedure. A discussion with microbiology suggested the addition of a hepatitis, Epstein-Barr virus and cytomegalovirus screen; all of which were negative. It transpired on further discussion with the patient that he had a pet cat, but could not recall any recent scratches. The patient underwent an ultrasound-guided core biopsy of the right axillary lymph node whereby multiple 14 g core specimens were obtained with no immediate complications and were sent for cytology, histopathology and 16S rDNA PCR.

Differential diagnosis

The initial provisional biopsy report suggested a necrotising inflammatory process and as such did not require further haematological input. Commencement of azithromycin resulted in clinical and biochemical improvement. Further biopsy results revealed morphological appearances suggestive of necrotising granulomatous lymphadenitis, which could be in keeping with the clinical impression of Bartonella/cat scratch lymphadenitis (figure 2). Other differential diagnoses would include lymphogranuloma venereum, tularaemia, tuberculosis and fungal infections. Systemic lupus erythematosus could also present with necrotising lymphadenitis and this would need to be excluded with the clinical and serological features. Necrosis can also be found in lymphomas and further excisional biopsy would be required to confirm this. The 16S rDNA PCR was detected using partial sequencing and, thus, the patient was managed as severe Bartonella infection and discharged with a further 2-week course of azithromycin.

Figure 2.

H&E staining of core biopsies of right axillary lymph nodes. (A) Magnified ×2: low power showing necrotic area. (B) Magnified ×4: necrotic areas with polymorphs. (C) Magnified ×10: higher magnification of necrosis admixed with inflammatory cells. (D) Magnified ×20: surrounding histiocytes and lymphocytes.

Treatment

Given the possibility of Bartonella resulting in CSD, the patient was commenced on azithromycin empirically. The treatment recommendations for isolated CSD lymphadenitis requiring antibiotics are azithromycin 500 mg on day 1 followed by 250 mg once a day for 4 days (total of 5 days). In patients with liver and splenic involvement, treatment is advised for 10–14 days. Our patient had deranged liver function tests and significant splenomegaly on CT scan and ultrasound; although no abscesses suggestive of peliosis hepatis were demonstrated. In view of liver and splenic involvement and features of severe sepsis at admission, a 2-week course of azithromycin was advised.

Outcome and follow-up

A clinic review 4 months later revealed that the patient was well and had no further fevers.

Discussion

CSD is a zoonotic Gram-negative bacterial infection caused by Bartonella henselae.⁶ It is usually transmitted to humans when cats, particularly kittens infected by fleas carrying Bartonella, scratch or bite humans.^8–10 The general availability of flea-control products means that stray cats are most likely vectors of Bartonella.^{8 10} Kittens are more likely to scratch or bite and, thus, are implicated to transmit the disease more than mature cats.¹⁰ Considering that CSD is not a reportable disease and often present with flu-like symptoms (fever, headache, poor appetite and lethargy), the true disease burden is unknown. Additionally, CSD classically presents with papules at the initial wound site and subsequent lymphadenopathy in the regional lymph nodes.¹⁰ About 5% of cases can present atypically with other systemic manifestations, including that of, but not limited to spleen, skin, eyes and nervous system (seizures and encephalitis).⁶

Epidemiological studies highlight that severe and atypical presentation of CSD is found among children aged 14 years and younger, with the lowest risk in those aged 50–64 years.⁶ This can be partly explained by immune-naivety of children, although further studies would be required to establish this. This also possibly explains why immunocompromised patients develop more severe and life-threatening disease associated with CSD.^{7 11} Similar to a previous case report,⁷ our immunocompromised patient was admitted with features of septic shock that could have been life-threatening.

Biochemical diagnosis is often difficult because of poor serological sensitivity and cross-reaction between other species, including Coxiella and Chlamydia.¹² PCR, therefore, would be the most specific and sensitive test.¹³ Given this uncertainty, an accurate diagnosis should be based cumulatively on a detailed history, clinical phenomenology/symptomology, histopathological and biochemical results.¹⁴ CSD treatment in an otherwise immunocompetent patient is self-limiting; however, given the systemic inflammatory response noted in immunocompromised patients, antibiotics are key in the overall survival of the patient.¹⁵

Our case illustrates the importance of a detailed history, including a pet history, which would have helped aid in the early diagnosis and treatment of this patient. More importantly, our case explicates the need for patients on long-term immunosuppressive therapy to be made aware of pet care. To our knowledge, there was minimal information available in the UK on CSD, and for a matter of fact, on any zoonotic disease in the immunocompromised patient. In fact, we advised our patient to follow guidance from the US National Library of Medicine—MedlinePlus—patient information website, for pets in immunocompromised patients, which has the relevant information presented clearly.¹⁶ Our learning points, thus, would be threefold: (1) the conscious need to elucidate a detailed history in an immunocompromised patient presenting with fever of unknown origin; (2) awareness about the association of certain zoonoses with different types of pets (figure 3); and (3) the need for more information to be cascaded, especially to the immunocompromised patients on pet care and the potential risks associated with having a pet.

Figure 3.

Infographic about common zoonotic pathogens of concern in immunocompromised patients.

Conclusion

Immunocompromised persons are at risk of zoonotic infections and can present as atypically and severely ill, given their immunological status. A thorough history and clinical examination remain integral to their diagnosis and subsequent management. Immunocompromised patients should be advised on pet care and liaising with veterinarians early in case of concerns regarding the well-being of their pets.

Patient’s perspective.

I have been taking mycophenolate and prednisolone to manage relapsing polychondritis and scleritis since 2018. Before contracting and being hospitalised with this infection, other than mild common colds, I had never really been taken ill in that period.

I have had many consultations with consultants, my general practitioner and clinical nurse specialists as an outpatient, and I am sure that I have never seen or heard any advice or guidance in relation to pets/animals and immunocompromised patients. I have also read through the information leaflets for my medication, and I don’t recall seeing any reference to it in there.

In 2019, there were a couple of occasions, once at a zoological park and once at a family children’s party, where I had the opportunity to handle a variety of animals, including mammals, reptiles and birds, but was advised by the animal handlers that, as an immunocompromised person, I should avoid doing so. It actually never occurred to me that this guidance should apply to domestic animal/pets as well!

We do have a mature cat at home and I had always handled him closely, petting and feeding him, and cleaning his food bowl. He often slept or sat on the same furniture as us, and would sometimes use the kitchen counter as an exit route to the garden.

Because he is an outdoor cat, we have always kept his worming and flea medication up to date.

We do maintain a good standard of hygiene at home, but I wouldn’t say that we had ever paid any extra/special attention to cleaning any spaces or surfaces that the cat has come into contact with.

I do not recall ever being bitten or scratched by our cat, but we did sometimes find fleas in the house, and I did have a small puncture mark on my right hand that I caused myself at the time that it became inflamed.

I began to fall ill approximately 4–6 weeks prior to admission to hospital. I found myself becoming very fatigued early in the evening and would fall asleep before I went to bed, but would wake up in the early hours feeling feverish and sweaty. The fatigue and fever symptoms eventually became constant, and I found it difficult to concentrate on simple tasks at home and work. I also found myself becoming breathless from even the slightest exertion.

During my illness, we had also noticed a change in the cat’s behaviour, in that he had become quite lethargic—after I left hospital we had him checked over by a vet who noticed that he had quite serious tooth decay. He was given dental surgery, followed by a 28-day course of antibiotics.

Since recovering, I have changed my behaviour towards the cat. I always use gloves whenever I feed him or clean his food bowl. I avoid close contact with him, and do not let him sit or sleep or sit on the furniture, as much as possible.

Learning points.

• The conscious need to elucidate a detailed history in an immunocompromised patient presenting with fever of unknown origin.

• Awareness and education on the association of common zoonoses with different pets.

• The need for more information to be cascaded, especially to the immunocompromised patients on pet care and the potential risks associated with having a pet.

• Advice to pet owners to seek early advice from veterinarians regarding their pets before initiation of immunocompromised therapy.

Ethics statements

Patient consent for publication

Obtained.