Ventura, Alessandro*; Massei, Francesco; Not, Tarcisio*; Massimetti, Mauro; Bussani, Rossana†; Maggiore, Giuseppe
Cat-scratch disease (CSD) is a benign disease manifested in most cases by self-limiting lymphadenitis after an inoculation, usually inflicted by a cat. Bartonella henselae is the causative agent of CSD, and laboratory confirmation of the diagnosis is now possible by testing the specific antibody response (1). Systemic B. henselae infection has been reported in small series of immunocompetent (2) and immunocompromised children (3) and generally has initial symptoms of prolonged fever and multiple granulomatous lesions in liver and spleen (4).
The purpose of our study was to report the clinical features of hepatic and/or splenic involvement in children with serologic evidence of B. henselae infection. For this reason, we retrospectively reviewed the records of all patients in whom hypoechogenic lesions in liver and/or spleen was diagnosed from 1990 through 1996, in three pediatric clinics in northern Italy. The reports of nine cases involving serologic evidence of B. henselae infection are the basis of this article.
The immune response to B. henselae was evaluated by an enzyme-linked immunosorbent assay (ELISA) and by an indirect fluorescent antibody (IFA) test.
This procedure was performed as previously reported (5). Briefly, B. henselae (strain Huston-T-isolate; American Type Culture Collection 49882) was grown at 37°C in 5% carbon dioxide for 5 days on Columbia blood agar. Bacteria were harvested in sterile distilled water and maintained frozen at -80°C. The bacterial cells were sonicated and centrifuged, and the pellet was resuspended in distilled water and stored at -20°C until ready for use. After dilution in 0.1 M carbonate buffer (pH 9.6) to yield the protein concentration of 5 µg/ml, a 0.1-ml aliquot of this solution was placed in each well of a flat-bottomed plate. Serum dilutions were 1:200 for both immunoglobulin (Ig)G and IgM. Phosphatase conjugates of goat anti-human IgM and IgG were diluted 1:4000. The immune reaction was developed by adding substrate solution. The absorbance was read in a microplate reader at 405 nm, and the positive control reached an optical density value of 1. Normal values were taken as less than 0.5 optical density for both IgM and IgG, which represented a value of more than 2 standard deviations above the mean of normal healthy people. The cut-off value was established from the sera of 32 patients with clinically proven CSD and from 55 healthy control subjects. The serologic test has a sensitivity of 87.5% and a specificity of 96%.
Indirect Fluorescent Antibody Test
Determination of IgG anti-B. henselae by indirect immunofluorescence was performed according to the manufacturer's instructions (M.R.L. Diagnostics, U.S.A.). IFA results were expressed as the highest reciprocal IFA serum titer at which bacillus-specific fluorescence was noted. Positive values had been determined to be greater than or equal to 128. In our hands, this serologic test has a sensitivity of 90.6% and a specificity of 92%.
We retrospectively reviewed the records of all patients with diagnoses of hypoechogenic lesions in liver and/or spleen, from 1990 through 1996 in three pediatric clinics in northern Italy. Patients reviewed were 11 boys and 2 girls (median age, 8 years 9 months; range, 2 years 6 months to 14 years). All had serologic evaluation for B. henselae infection (five retrospectively on frozen sera), and eight had liver biopsy (four percutaneous and four by laparotomy). All had single or multiple hypoechogenic lesions in liver and/or spleen. Serologic evidence of B. henselae infection was present in nine. Three of the remaining four patients received diagnoses of Crohn's disease, primary hepatic actinomycosis, and chronic granulomatous disease, and one 5-year-old girl who had fever of unknown origin and lymphadenomegaly that remained undiagnosed, showed a progressive amelioration of the clinical course on broad-spectrum antibiotic treatment.
Seven of the nine patients with B. henselae infection had prolonged unexplained fever lasting 6 weeks to 4 months, and 3 had acute lymphadenitis and fever. All had increased sedimentation rate (range, 30-110 mm/hr) and normal serum aminotransferase activity. Hypoechogenic lesions on ultrasound and/or hypodense areas on computed tomographic (CT)-scan involved the liver and the spleen in seven, only the liver in one and only the spleen in two. Liver histology performed in five showed necrotizing granulomatous lesions in all. The main clinical and laboratory features of the nine cases are summarized in Table 1. All patients were treated with broad-spectrum antibiotics. Fever lasted from 3 weeks to 4 months, and hepatic and splenic images normalized in all with residual calcifications in one.
In 1994, a boy aged 2 years 8 months had symptoms of prolonged fever and hepatosplenomegaly. Laboratory evaluation showed microcytic (69/fl) anemia (hemoglobin, 8.6 g/dl), thrombocytosis (500,000 mm3), leukocytosis (18,350 mm3), high sedimentation rate, and high C reactive protein level (60 mg/dl; normal value, <5 mg/dl). Multiple serologies for bacterial, mycotic, and viral infections were negative. Results of a Mantoux test were negative. Abdominal ultrasound showed multiple hypoechogenic hepatic and splenic lesions. The patient was treated with 10 days of imipenem intravenously with resolution of fever and normalization of biochemical abnormalities. One year later an ultrasound examination showed normalization of liver images and multiple hyperechogenic calcified lesions in the spleen. Retrospective analysis of a convalescent serum for B. henselae demonstrated specific IgG anti-B. henselae in absence of IgM. The family had a kitten in the home.
In 1993, a 13-year-old boy had high fever, fatigue, anorexia, and diarrhea. He was treated at home with clotrimazole and cefuroxime. He was in good general condition, the liver was not enlarged, and the spleen was palpable. Laboratory tests showed high sedimentation rate and high C reactive protein levels (150 mg/dl), hypergammaglobulinemia (2.5 g/dl), and hyper-IgG (3082 mg/dl). Abdominal ultrasound and CT-scan (Fig. 1) showed multiple hepatic and splenic lesions with abdominal lymph node enlargement. The child underwent open liver and lymph node biopsies. Histology showed granulomatous lymphadenitis and necrotizing granuloma in the liver with central microabscess and palisading epithelioid cells. Fever lasted 3 weeks after surgery. He was treated with rifampin and ceftriaxone with resolution of fever and normalization of image abnormalities with 3 months. Retrospective analysis of a serum at admission was positive for B. henselae by ELISA determination of IgM and IgG and by IFA. The boy had been scratched by a kitten before onset of acute disease.
In 1996, a boy aged 13 years 10 months had high fever lasting 3 weeks and axillary lymphadenitis with evidence of multiple cat scratches. He had been treated with amoxicillin and clotrimazole. The liver and spleen were normal at clinical examination. Sedimentation rate was 36 mm/hr, C reactive protein 66 mg/dl, and IgG 2035 mg/dl. Abdominal ultrasound showed multiple hepatic and splenic hypoechogenic lesions. B. henselae IgG and IgM ELISA and IFA results were positive. He was treated with rifampin with resolution of fever and normalization of biochemical abnormalities within 1 month and of hepatic images within 10 months.
A 10-year-old boy was scratched by a kitten in 1996 1 month before admission. He had high fever lasting 1 week and axillary lymphadenitis with evidence of an inoculation lesion. Liver and spleen were normal at clinical examination. Sedimentation rate was 30 mm/hr, and C reactive protein level was 19 mg/dl. Abdominal ultrasound showed a unique hypoechogenic lesion in the spleen. B. henselae IFA serology was positive. Fever and biochemical abnormalities resolved without treatment within 2 weeks, and the splenic lesion resolved within 2 months.
A boy aged 2 years 6 months was admitted in 1990 for evaluation of presumed metastatic neuroblastoma. He had septic-like fever and multiple hepatic and splenic hypoechogenic lesions on abdominal ultrasound. General conditions were not compromised, but the liver and spleen were enlarged. An open liver biopsy was performed, and analysis showed necrotizing granulomata with multiple abscesses. He was treated with different antibiotics which included rifampin, vancomycin, and netilmicin with resolution of fever and normalization of biochemical and image abnormalities. Diagnosis of B. henselae infection was made retrospectively by the analysis of Warthin-Starry staining of the liver biopsy specimen which showed bacillary forms (Fig. 2) and by ELISA and IFA serology. The patient had a kitten at home.
In 1992, a 6-year-old boy was admitted for evaluation of an unexplained fever lasting for 1 month. He had been treated with different antibiotics including ceftriaxone and netilmicin. General conditions were not compromised; the liver was moderately enlarged. Abdominal ultrasound showed multiple hypoechogenic liver lesions with a unique lesion in the spleen. The boy had a kitten at home, and he had been scratched several times. B. henselae ELISA and IFA serology were positive. He was treated with rifampin, and fever and biochemical and image abnormalities normalized.
In 1995, a 10-year-old boy had fever, inguinal lymphadenitis, and evidence of a scratch from a kitten at home. Liver and spleen were not palpable. Abdominal ultrasound showed two hypoechogenic splenic lesions. B. henselae ELISA and IFA serology were positive. Fever lasted 2 weeks after admission, and he had been treated with rifampin with resolution of splenic lesions.
In 1992, a girl aged 7 years 9 months, living at a farm with kittens had fever, abdominal pain, and diarrhea. She was treated with amoxicillin, erythromycin, and trimethoprim-sulfamethoxazole. Her general condition was good, there was no lymphadenopathy, and liver and spleen were not palpable. Laboratory test results showed high sedimentation rate and high C reactive protein levels (180 mg/dl); fibrinogen was 650 mg/dl, albumin 3 g/dl, and total γ-globulins 1.5 g/dl. Abdominal ultrasound showed multiple hepatic and splenic lesions. Fever lasted 4 weeks, during which she was treated with teicoplanin with resolution of image abnormalities. Retrospective analysis of a convalescent serum by IFA was positive for B. henselae.
In 1995, a boy aged 11 years 2 months had high fever. He was treated at home with ceftriaxone for submandibular lymphadenitis. There was no adenomegaly, and the liver and spleen were not palpable. Laboratory tests showed high sedimentation rate and high C reactive protein level (130 mg/dl); fibrinogen was 500 mg/dl. γ-Glutamyl transpeptidase activity was 1.5 times the upper normal value. Abdominal CT scan showed multiple hypodense hepatic lesions. The child underwent an open liver biopsy. The liver had a "nutmeg" appearance, and histology showed stellate necrotizing granulomata surrounded by palisading epithelioid cells. The boy had had several contacts with a kitten. Warthin-Starry staining was negative, but B. henselae serology was positive by IFA.
Systemic B. henselae infection with hepatosplenic involvement has been reported with increasing frequency in immunocompetent children (2,4). Prolonged fever, with or without peripheral lymphadenitis associated with such systemic symptoms as malaise, fatigue, myalgia, and weight loss are the most frequent clinical manifestations. All our patients had long-lasting fever of unknown origin, and peripheral lymphadenitis was present in approximately half of the patients, a rate similar to that reported by others (4,6). Peripheral lymphadenitis may help in determining the diagnosis, even if systemic CSD may develop in absence of peripheral lymphadenitis. Two of our patients (patients 4 and 7 in Table 1) who had lymphadenitis and fever with no systemic symptoms showed evidence of systemic involvement demonstrated by the presence of isolated hypoechogenic lesions of the spleen on ultrasound. This suggests that systemic involvement may be more frequent than currently thought and may not be associated with severe symptoms.
Visceral involvement was clinically latent in most of our patients; only two had hepatosplenomegaly and/or reported mild abdominal pain. Abdominal pain has recently been reported to be an important clinical symptom suggesting visceral involvement in CSD (6), but our and others' experiences (2) suggest that systemic B. henselae infection should be considered even in the absence of this symptom.
Abdominal ultrasound was helpful in the diagnosis of visceral B. henselae infection. It should always be performed in cases of prolonged fever of unknown origin, even in absence of clinical and biochemical evidence of liver involvement. The finding of single or multiple hypoechogenic lesions in liver and/or spleen in a child with prolonged fever should suggest systemic B. henselae infection.
In our experience, B. henselae infection was a frequent cause of hypoechogenic hepatic and/or splenic lesions. Histopathologic examination, when performed, showed granulomatous lesions. It is worth noting that B. henselae is not included in the list of causative agents in granulomatous lesion in the liver in a distinguished pediatric hepatology textbook (7). This could be related to the peculiar epidemiology of CSD. In fact, so far, systemic B. henselae infection has been reported mostly in people living in North America (2,4,6). Only recently have there been reports in Europe (8).
Diagnosis of disseminated B. henselae infection was difficult until recently, and most of our patients had many invasive investigations, including surgical liver biopsy. Warthin-Starry staining of affected tissue can suggest diagnosis but is of low diagnostic sensitivity during later stages of the disease (9). Serologic test for B. henselae showed in a different experience a sensitivity and a specificity close to 100% for indirect fluorescence antibody test and for titers more than 1:256 (1,5,10). The presence of high IgM titers to B. henselae is suggested to indicate a recent or ongoing infection (5). Serology for B. henselae thus provides a rapid diagnosis, avoids invasive investigations, and excludes potentially life-threatening disease associated with lymphadenopathy. It should always be performed in children with hypoechogenic hepatosplenic lesions.
All our patients were treated with different types of antibiotics. There is limited evidence of a beneficial effect of antibiotic treatment in disseminated B. henselae infection in the immunocompetent host (9). Treatment with oral azithromycin was recently shown in a controlled study to produce significant clinical benefit in typical CDS (11). In our experience, in patients with hepatosplenic involvement, lesions resolved in all within 6 months of onset, with residual calcification of a splenic lesion in one. In light of recent experience, azithromycin could be proposed in patients with systemic B. henselae infection to accelerate healing of visceral lesions.
In conclusion, systemic B. henselae infection must always be considered in children with fever of unknown origin and represents a frequent cause of inflammatory hepatosplenic lesions in children. Serologic tests allow a rapid and noninvasive diagnosis. Systemic B. henselae infection in a self-limiting disease in immunocompetent children, and diagnostic procedures should be limited to serology. Hepatosplenic involvement can be found even in children with classic CSD with peripheral lymphadenitis, even in absence of apparent systemic manifestations. The frequency of systemic involvement in typical CSD is probably underestimated.
1. Nadal D, Zbinden R. Serology to Bartonella (Rochalimaea) henselae may replace traditional diagnostic criteria for cat-scratch disease. Eur J Pediatr 1995;154:906-8.
2. Estrada B, Silio M, Begue RE, Van Dyke RB. Unsuspected hepatosplenic involvement in patients hospitalized with cat-scratch disease. Pediatr Infect Dis J 1996;15:720-1.
3. Apalasch AM, Bakr N, Jaffe R. Systemic cat-scratch disease in a pediatric liver transplant recipient and review of the literature. Pediatr Infect Dis J 1993;12:769-74.
4. Dangman BC, Albanese BA, Kacica MA, Lepow ML, Wallach MT. Cat-scratch disease in two children presenting with fever of unknown origin: imaging features and association with a new causative agent, Rochalimaea henselae. Pediatrics 1995;95:767-71.
5. Not T, Canciani M, Buratti E, et al. Serologic response to Bartonella henselae in patients with cat scratch disease and in sick and healthy children. Acta Paediatr 1999;88:1-6.
6. Dunn MW, Berkowitz FE, Miller JJ, Snitzer JA. Hepatosplenic cat scratch disease and abdominal pain. Pediatr Infect Dis J 1997;16:269-72.
7. Mowat AP. Liver disorders in childhood. Oxford, UK: Butterworth-Heinemann, 1994.
8. Waldvogel K, Regnery R, Anderson BE, Caduff R, Caduff J, Nadal D. Disseminated cat-scratch disease: detection of Rochalimaea henselae in affected tissue. Eur J Pediatr 1994;153:23-7.
9. Bass J, Vincent JM, Person DA. The expanding spectrum of Bartonella infections: II. Cat scratch disease. Pediatr Infect Dis J 1997;16:163-79.
10. Zangwill KM, Hamilton DH, Perkins BA, et al. Cat scratch disease in Connecticut. Epidemiology, risk factors and evaluation of a new diagnostic test. N Engl J Med 1993;329:8-13.
11. Bass JW, Cary Freitas B, Freitas AD, et al. Prospective randomized double blind placebo-controlled evaluation of azithromycin for treatment of cat scratch disease. Pediatr Infect Dis J 1998;17:447-52.
© 1999 Lippincott Williams & Wilkins, Inc.