Suspected Hepatosplenic Cat Scratch Disease with No Major Symptoms and Negative Serology: A Case Report

Introduction

Cat scratch disease (CSD), first reported in 1950 by Debré et al, can affect immunocompetent patients of all ages [1]. It is almost exclusively caused by Bartonella henselae, with rare reports implicating B. clarridgeiae [2]. With 80% of patients reportedly being younger than 21 years of age, the condition is considered the most common cause of chronic, benign adenopathy in children and young adults [1,3]. Typical CSD presents with fever and an erythematous papule at the cat scratch or bite site, followed by self-limiting lymphadenopathy in the regional lymph nodes [4]. Atypical presentations are seen in up to 25% of cases [5] and can involve the eyes, nervous system, heart, liver, spleen, skin, or musculoskeletal system [4]. These include Parinaud oculoglandular syndrome, neuroretinitis, encephalopathy, hepatosplenic involvement, osteomyelitis, endocarditis, and erythema nodosum [2,5]. Disease severity is generally higher in children younger than 14 years of age, whereas adults aged 50 to 64 years reportedly have the lowest risk for developing atypical CSD [4].

Case Report

A 39-year-old woman with no significant past medical history presented to the Emergency Department with peri-umbilical and right lower quadrant pain lasting for 12 h. She was afebrile and had stable vital signs. Physical examination revealed tenderness in the right iliac fossa. There was no rebound tenderness or guarding, and her abdomen was soft. She had no palpable cervical or supraclavicular lymphadenopathy, hepatosplenomegaly, or rash. The laboratory test results showed leukocytosis, with a white blood cell count of 14 400/μL and neutrophil predominance. The C-reactive protein level was elevated at 2.69 mg/dL. CT of the abdomen with contrast revealed no apparent enlargement of the appendix. Incidentally, it showed multiple low-enhancing nodules in the liver and spleen, as well as hepatic hilar lymphadenopathy (Figure 1). Although findings of appendicitis were not apparent on the CT image, appendicitis was clinically suspected. Treatment was initiated with intravenous cefmetazole 1 g, followed by levofloxacin 500 mg orally once daily, plus metronidazole 500 mg orally once daily for 7 days.

Although her abdominal symptoms disappeared, follow-up CT showed remaining hepatosplenic nodules along with enlarged hepatic hilar lymph nodes, indicating lymphomas, metastasis of malignant tumor, sarcoidosis, or tuberculosis. Laboratory test results showed anemia (hemoglobin 11.6 g/dL) and a slightly elevated C-reactive protein level (0.37 U/mL) but were otherwise unremarkable. Serology results for hepatitis A virus, hepatitis B virus, HIV, and syphilis were negative, as was the tuberculosis Interferon-Gamma Release Assay. Serum angiotensin-converting enzyme and soluble interleukin-2 receptors were within the reference range. Chest CT demonstrated absence of pulmonary hilar lymphadenopathy, and the electrocardiogram findings were within normal limits, with no evidence of atrioventricular block – features atypical for sarcoidosis. Magnetic resonance imaging (MRI) of the spleen with contrast revealed multiple low-intensity nodules on T2-weighted images, relative to the splenic parenchyma (Figure 2). Subsequently, an interventional radiology-guided percutaneous splenic biopsy was performed to evaluate for malignancy and expedite the diagnostic process. Histopathology of the specimen showed non-necrotizing epithelioid granulomas with no evidence of malignancy (Figure 3). Ziehl-Neelsen staining did not reveal any organism.

The presence of epithelioid cell granulomas in the biopsy specimen suggested that the nodular lesion identified on MRI was accurately targeted. Given the absence of malignant features, caseous necrosis, and the negative result on acid-fast staining within the biopsy tissue, malignancy, lymphoma, and tuberculosis were considered unlikely. The patient was subsequently referred to the Infectious Disease Department for further evaluation of possible infectious causes. Upon further history-taking, she reported having frequent contact with more than 30 cats, including many kittens, and multiple scratches over the years. However, it is worth noting that no visible scars or bite marks were found on physical examination. An indirect fluorescent antibody assay was performed to detect anti-Bartonella henselae and B. quintana immunoglobulin M (IgM) and immunoglobulin G (IgG). Based on assay criteria, a titer of greater than or equal to 1: 64 was defined as positive for IgG, and a titer of greater than or equal to 1: 20 was defined as positive for IgM [6,7]. All titers remained below these thresholds, indicating negative serology. Despite negative serologic results, her strong exposure history raised clinical suspicion for CSD. An MRI of the brain with contrast and an ophthalmologic examination confirmed the absence of central nervous system lesions. Subsequently, she was started on rifampicin 300 mg orally twice daily, azithromycin 500 mg once daily on the first day, followed by 250 mg once daily thereafter for 2 weeks in total. Abdominal echography performed 1 month after treatment showed a decrease in the size of the splenic lesions. Abdominal CT with contrast taken 2 months after treatment showed a resolution of the hepatosplenic nodules, revealing calcifications instead (Figure 4).

Discussion

A case-series study conducted by Rodriguez et al is the largest cohort of adult patients with CSD to date [8]. It included 30 adult patients with CSD and reported peripheral lymphadenopathy as the most common clinical presentation, observed in 90% of cases. Hepatosplenic CSD in immunocompetent adults is rare, and much of the current knowledge has been obtained from previous reports on cases in children [9]. Its presentations can resemble those of potentially lethal diseases, such as lymphoma, tuberculosis, or malignancies [9]. Nevertheless, a timely and accurate diagnosis is vital to prevent prolonged hospitalization [10], potentially invasive diagnostic testing, or surgical interventions required when testing for such etiologies [9].

There are no standard diagnostic criteria for CSD [11]. Traditionally, the diagnosis of CSD has been confirmed with 3 of the 4 criteria shown in Table 1 [2]. Serology, especially indirect fluorescent antibody assay for anti-B. henselae IgG, is currently considered a first-line diagnostic test [11]; however, in the literature, its sensitivity and specificity are variable. Lack of standardization on the definition of CSD, variability between testing kits, or differences in study populations may have contributed to this range in the performance of the test [11,12]. One factor that can limit IgG specificity is the high seroprevalence of B. henselae IgG, as many asymptomatic patients are seropositive due to prior animal contact [11]. In a study by Sanders et al, a seroprevalence of B. henselae antibody was found to be as high as 30% in healthy German individuals, irrespective of cat ownership [12]. B. henselae IgG can be detected in the blood for up to 22 to 28 weeks after exposure, with 25% of patients remaining IgG seropositive after 1 year [11]. A positive IgM is a hallmark of acute Bartonella infection, but IgM production is brief [13]. Cross-reactivity to Coxiella burnetti, Brucella sp, Chlamydophila sp, as well as non-henselae Bartonella is also known [9]. Furthermore, Epstein-Barr virus infection reportedly reduces the specificity of Bartonella serology testing [10]. For these reasons, a negative serology at one point does not exclude CSD, making CSD a challenging diagnosis.

Polymerase chain reaction (PCR) assays enable rapid detection of Bartonella species in blood samples, biopsy specimens, and swabs of lesions [9]. Although PCR is high in specificity, sensitivity is shown to vary depending on specimen type, ranging between 45% and 95% [3]. In one study, sensitivity was highest in lymph node pus aspirates, followed by primary lesions, lymph node fine needle aspirations, lymph node biopsy specimens, and paraffin-embedded lymph nodes [3]. The sensitivity of blood, as well as that of serum samples, although readily available, is reported to be low [3,14]. Another factor that can alter the sensitivity of PCR assays is the timing of the biopsy [15]. In a previous study, PCR was more likely positive if performed during the first 6 weeks of infection [15]. Thus, PCR can be helpful in certain situations, but is considered to play solely a supplementary role in the diagnosis of CSD. Another point to note is that PCR is not readily available in all institutions and is not cost-effective [2]. The unavailability of Bartonella PCR in our facility was one factor that deterred us from prioritizing PCR over serology.

Histopathology of lymph nodes is usually nonspecific and can show stellate caseating granulomas (with acellular, necrotic center), microabscesses, or follicular hyperplasia, depending on the stage of disease [15]. In the presence of a strong clinical suspicion of CSD, Warthin-Starry staining of the specimen should be considered, although in many cases, it is not performed, owing to the low prevalence of CSD in adults [9].

In our case, diagnostic testing for hepatosplenic CSD was not performed until the diagnosis of sarcoidosis, lymphoma, and malignancies was excluded. At the time of biopsy, Warthin–Starry staining and microbiological culture were not performed, as a history of cat exposure had not yet been elicited, and the initial differential diagnosis favored sarcoidosis or malignancy. Absence of major symptoms, such as fever and lymphadenopathy, may have contributed to delayed recognition of CSD in our patient. However, once a detailed history revealed possible cat exposure, empiric antibiotic therapy was initiated despite negative serologic tests, resulting in successful clinical resolution. Although hepatosplenic CSD in adults is uncommon, early recognition can be essential to avoiding extensive testing and procedures. Additionally, our case suggests that, in the presence of a strong clinical suspicion for CSD, timely consideration of an empiric treatment can be beneficial.

Conclusions

Our case depicts a suspected case of hepatosplenic CSD with minimal symptoms and a negative Bartonella serology. A 2-week course of rifampicin and azithromycin successfully led to clinical improvement. Despite the rarity of CSD, clinicians should consider hepatosplenic CSD in patients with multiple nodular lesions in the liver and spleen, since its presentations can overlap with potentially lethal conditions, such as malignancies or tuberculosis. Given the low sensitivity of Bartonella serology, we conclude that a negative serologic result would not rule out CSD, and that initiation of treatment with antibiotics could be beneficial in a case with strong clinical suspicion.