818
Brief Reports
Portal Vein and Bone Involvement in Disseminated
Cat-Scratch Disease: Report of 2 Cases
Cat-scratch disease (CSD) is a common cause of regional lymphadenopathy. We describe 2 children with an
unusual presentation of disseminated CSD, the first one
presenting with persistent fever, multilocular abscesses in
liver and spleen as well as osteolytic lesions in the lumbar
spine and the second one with portal vein thrombosis and
severe ascites.
Cat-scratch disease (CSD) caused by the fastidious growing
gram-negative rod Bartonella henselae is a common cause of
subacute regional lymphadenopathy in children and adults,
with typical clinical manifestations. It is usually a benign, selflimited disease. Three to 10 days after a scratch or bite by a
cat, the patient may develop pustules or papules at the primary
site of injury. After 2–3 weeks, regional lymph nodes enlarge
and persist for several weeks to months. Usually no antibiotic
therapy is required, and surgical resection of the lymph nodes
is rarely necessary. If CSD is clinically suspected by the typical
lymphadenopathy and a history of cat exposure, the diagnosis
is usually confirmed by serology [1–4]. In cases with atypical
clinical manifestations or low antibody levels against B. henselae, suspected CSD should be proved by serological followup, by molecular or histopathological investigations of the affected tissue, or both.
Disseminated CSD, in contrast, presents with a large variety
of clinical symptoms [5] and may be difficult to diagnose. It
has been described in immunocompetent and immunocompromised patients. Clinical manifestations include persistent fever, Parinaud’s oculoglandular syndrome, various skin lesions,
hepatic and splenic abscesses, arthritis, pneumonia, pleural effusion, thrombocytopenic purpura, and osteolytic lesions. Neurological complications can present as encephalitis, meningitis,
optic neuritis, convulsion, radiculitis, paraplegia, and cerebral
arteritis [6–22]. Here we report 2 cases of disseminated CSD in
immunocompetent children with hepatosplenic manifestations
and the rare symptoms of bone and portal vein involvement.
Three weeks after a bout of measles with an uneventful
course, a previously healthy 12-year-old girl developed fever
(387C–407C for 4 weeks), enlarged submandibular and axillary
lymph nodes 1 cm in diameter, severe diffuse abdominal pain
not related to meals or time of day, back pain, diffuse myalgia,
weight loss of 6 kg within 4 weeks, and hepatosplenomegaly.
There was no history of cat or dog contact. At admission, the
patient’s white blood cell count was 5900/mm3, C-reactive protein level was 7.7 mg/dL, and the erythrocyte sedimentation
CID 2000;31 (September)
rate was 57 mm/h. Liver function tests were slightly elevated,
with an alanine transferase of 46 U/L and an aspartate transferase of 28 U/L. Serological tests were performed that showed
negative results for cytomegalovirus, Epstein-Barr virus, hepatitis A and C virus, Salmonella typhi, Salmonella paratyphi,
Brucella abortus, and Entamoeba histolytica. Likewise, the antistreptolysin O titer, antinuclear antibodies, and antimitochondrial antibodies were all within normal limits. Rheumatoid factor was negative. Serological investigations for hepatitis B
revealed a chronic persistent hepatitis B infection (HBs antigen
and anti-HBc immunoglobulin [Ig] G positive and anti-HBc
IgM and HBe antigen negative). High titers of antibodies to
B. henselae (IgM 1:1024 and IgG 1:8000) were found by immunofluorescence assay 3 weeks after onset of the illness. Analysis of magnetic resonance images taken of the abdomen
showed multiple abscesses in the liver and spleen (figure 1). A
99m
Technetium bone scintigraph of the spine showed multiple
spots of hyperactivity in the lumbar vertebra (figure 2) corresponding to spondylitis, which was confirmed by magnetic resonance imaging scan as osteolytic lesions in the lumbar vertebra
(figure 1). After 3 weeks of therapy with erythromycin, the
patient clinically improved. Abdominal pain and back pain
decreased; the girl was afebrile and in good general condition.
Three months later, the abdominal pain was completely resolved. After a follow-up period of 15 months, the girl was
doing clinically well, and sonography indicated that the hepatic
and splenic abscesses had disappeared. Titers to B. henselae
decreased in the follow-up, as shown in table 1.
A previously healthy 2.5-year-old boy with a history of cat
contact presented for routine physical examination. He was in
good general condition and afebrile, with mild abdominal tenderness in the paraumbilical area. Splenomegaly 10 cm in dia-
Reprints or correspondence: Dr. Michael Ruess, University Children’s
Hospital, Mathildenstraße 1, D-79106 Freiburg, Germany (ruess@kkl200
.ukl.uni-freiburg.de).
Clinical Infectious Diseases 2000; 31:818–21
q 2000 by the Infectious Diseases Society of America. All rights reserved.
1058-4838/2000/3103-0036$03.00
Figure 1.
Magnetic resonance image scan showing multiple abscesses in liver and spleen and an osteolytic lesion in the lumbar vertebra (arrow) in patient 1.
CID 2000;31 (September)
Brief Reports
Figure 2. Bone scan showing multiple spots of hyperactivity in the
lumbar spine of patient 1.
meter was confirmed by ultrasound. At admission, his white
blood cell count was 10,200 cells/mm3, platelet count was
155,000/mm3, hemoglobin was 100.1 g/L, and lactate dehydrogenase was slightly elevated, at 380 U/L. C-reactive protein
was !6 mg/L. To exclude the possibility of a neoplastic disorder,
a bone marrow puncture was performed; results indicated a
normal cell distribution. Liver function tests were within normal limits. Serological tests for cytomegalovirus, Epstein-Barr
virus, hepatitis A, B, and C virus, parvovirus B19, Toxoplasma
gondii, and human immunodeficiency virus were all negative.
Standardized tuberculin test was negative. Analysis by ultrasound and computed tomography revealed splenomegaly,
thrombosis of the extrahepatic portal vein with cavernous
transformation (figure 3), and several enlarged intra-abdominal
lymph nodes.
Three weeks later, the boy developed massive ascites but
showed no signs of jaundice. A serum sample obtained 6 weeks
after onset of illness showed high titers of antibodies to B.
henselae (table 1). The abdominal lymph nodes were resected
by laparotomy, and B. henselae DNA was detected by polymerase chain reaction (PCR). PCR that was simultaneously
performed on mycobacterial DNA remained negative. At the
same time, an EDTA blood sample was analyzed by PCR and
was also found to be positive for B. henselae. Histological investigation of the liver biopsy showed no pathological findings.
The ascites were treated with spironolactone and decreased
slowly within several weeks. Because of the rather good clinical
condition and the absence of fever, the boy received no antibiotic therapy. After 5 months, the ascites and splenomegaly
had disappeared, and, clinically, the patient had recovered com-
819
pletely. Antibodies against B. henselae decreased, as shown in
table 1. Two and a half years after the onset of the disease, the
patient is in good health.
Fever and abdominal pain may be leading symptoms in disseminated CSD. In a recently published study, Arisoy et al. [9]
described 19 children with hepatosplenic CSD. Fever (100%)
and abdominal pain (68%) were the 2 most frequent symptoms.
The initial diagnosis of fever of unknown origin was made in
79% of the children. Dunn et al. [6] reported a series of 11
children with hepatosplenic CSD. Fever for 12 weeks was found
in 72% of the patients, and 64% of the children complained of
abdominal pain. In typical CSD, regional lymphadenopathy is
a characteristic feature and one of the major diagnostic criteria.
In disseminated CSD, however, lymphadenopathy can be absent [6, 8, 15]. In the study of Dunn et al. [6], approximately
half the children showed lymphadenopathy, and in a recently
reported series of 9 cases of disseminated CSD, only 4 patients
had peripheral lymphadenopathy [15].
The first patient of our report showed the most frequent
symptoms of disseminated CSD: fever and abdominal pain.
She also had peripheral lymphadenopathy and bone involvement, a symptom that was first described in 1954 [16] and that
is rarely seen in disseminated CSD. A recently published review
[17] of the literature reports ∼15 cases of CSD associated with
osteomyelitic lesions. Bone involvement as a consequence of
B. henselae infection usually presents with localized pain without erythema, swelling, or tenderness and appears only in !1%
of CSD cases [18]. An even more atypical course of the disease
was observed in the second patient. Splenomegaly found on
routine physical examination and slight abdominal tenderness
were the leading signs. The delayed onset of ascites 3 weeks
after onset of the illness and the extrahepatic portal vein thrombosis are very rare complications. To our knowledge, this is
only the second report of portal vein involvement in disseminated CSD [19]. These sequels can be explained by the pressure
of the enlarged abdominal lymph nodes to the portal vein and
subsequent deceleration of the blood flow. A thrombophilic
disorder was excluded in this patient. In 1996, Zinzindohoue
et al. [19] described portal vein involvement due to disseminated
CSD in a 16-year-old boy. This patient had symptoms of portal
hypertension, which were caused by an abscess behind the porTable 1. Bartonella henselae antibody titers and serological followup in patients 1 and 2.
Time of serum collection
after onset of illness
Patient 1
3 Weeks
3 Months
6 Months
25 Months
Patient 2
6 Weeks
4 Months
B. henselae titer
IgG
IgM
8000
4000
2000
256
1024
128
!64
!64
1024
512
!64
128
820
Brief Reports
CID 2000;31 (September)
showed a faster decrease of peripheral lymph node enlargement
during azithromycin therapy during the first 30 days [22]. But
after 1 month, there was no difference in the outcome of antibiotic treated and untreated patients. There are many case reports in the literature with empirical antibiotic treatment for
disseminated CSD, including erythromycin, tetracycline, gentamicin, rifampin, ciprofloxacin, and other drugs [23]. But there
are no prospective controlled clinical trials to prove the effectiveness of an antimicrobial therapy for disseminated CSD.
Therefore, the use of antibiotics in a patient with disseminated
CSD has to be decided on a case-by-case basis. A general
recommendation of a specific antibiotic as drug of choice at
this time remains questionable.
Diagnosis and treatment of disseminated CSD remains difficult. The broad range of clinical symptoms and the unspecific
course of the disease may cause problems in recognizing CSD
and lead to misdiagnosis. Therefore, disseminated CSD has to
be taken into account in children presenting with a systemic
illness, and B. henselae infection should be considered in the
initial evaluation of prolonged fever, especially in children with
a history of exposure to cats.
Figure 3. Computed tomography scan showing extrahepatic portal
vein thrombosis with cavernous transformation (arrow) in patient 2.
tal triad. In the analysis of the resected clinical specimen, B.
henselae DNA was detected by PCR.
Immunofluorescence assay showed that both patients had
high IgG and less elevated IgM. The titers dropped when the
children were improving, but they still had persisting, high levels
of IgG. Approximately 15 months after onset of the disease,
the first patient had an IgG titer of 1:256. In several cases in
the literature, very high IgG and IgM titers are reported at the
time of diagnosis [6, 9, 11, 13–15, 20], but serological followup after recovery is not well known. Waldvogel et al. [20] described the titer course of a patient with disseminated CSD,
but within a period of 5 months, no significant decrease could
be observed. The sensitivity of the immunofluorescence assay
in CSD ranges from 83% to 100% [1–4] but has less satisfying
specificity. Especially high titers to B. henselae are diagnostic
key findings, and serological testing is a method to confirm
clinically suspected CSD within 48 h [2, 4]. PCR analysis is a
helpful tool to confirm the diagnosis. Sensitivity of PCR assays
varies depending on the primers used for amplification and the
preparation of the specimen [21]. Formalin-fixed, paraffinembedded specimens seem to be less suitable for PCR analysis
than fresh or frozen specimens. However, obtaining clinical
specimens is always combined with invasive procedures.
Antimicrobial treatment differed in our patients. The first
patient received erythromycin for 3 weeks and clinically improved after treatment and was completely recovered after 3
months. The second patient received no antibiotic treatment.
He also clinically improved, but the length of recovery—5
months—was significantly longer. For typical CSD, one study
Michael Ruess, 1 Anna Sander, 2 Matthias Brandis, 1
and Reinhard Berner 1
1
University Children’s Hospital and 2Institute for Medical
Microbiology and Hygiene, University of Freiburg, Freiburg, Germany
References
1. Regnery RL, Olson JG, Perkins BA, Bibb W. Serological response to
Rochalimaea henselae antigen in suspected cat-scratch disease. Lancet
1992; 339:1443–5.
2. 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
3. Szelc-Kelly CM, Goral S, Perez-Perez GI, Perkins BA, Regnery RL,
Edwards KM. Serologic response to Bartonella and Afipia antigens
in patients with cat-scratch disease. Pediatrics 1995; 96:1137–42.
4. Sander A, Posselt M, Oberle K, Bredt W. Seroprevalence of antibodies
to Bartonella henselae in patients with cat scratch disease and in
healthy controls: evaluation and comparison of two commercial serological tests. Clin Diagn Lab Immunol 1998; 5:486–90.
5. Bass JW, Vincent JM, Person DA. The expanding spectrum of Bartonella
infections: II. Cat-scratch disease. Pediatr Infect Dis J 1997; 16:
163–79.
6. Dunn MW, Berkowitz FE, Miller JJ, Snitzer JA. Hepatosplenic catscratch disease and abdominal pain. Pediatr Infect Dis J 1997; 16:
269–72.
7. Margileth AM, Wear DJ, English CK. Systemic cat-scratch disease: report of 23 patients with prolonged or recurrent severe bacterial infection. J Infect Dis 1987; 155:390–402.
8. Apalsch AM, Nour B, 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.
9. Arisoy ES, Correa AG, Wagner ML, Kaplan SL. Hepatosplenic catscratch disease in children: selected clinical features and treatment.
Clin Infect Dis 1999; 28:778–84.
10. Hayem F, Chacar S, Hayem G. Bartonella henselae infection mimicking
CID 2000;31 (September)
11.
12.
13.
14.
15.
16.
17.
Brief Reports
systemic onset of juvenile chronic arthritis in a 2½-year-old girl. J
Rheumatol 1996; 23:1263–5.
Jacobs RF, Schutze GE. Bartonella henselae as a cause of prolonged
fever and fever of unknown origin in children. Clin Infect Dis 1998;26:
80–4.
Golden SE. Hepatosplenic cat-scratch disease associated with elevated
anti-Rochalimaea antibody titers. Pediatr Infect Dis J 1993; 12:
868–71.
Tan TQ, Wagner ML, Kaplan SL. Bartonella (Rochalimaea) henselae
hepatosplenic infection occurring simultaneously in two siblings. Clin
Infect Dis 1996; 22:721–2.
Caniza MA, Granger DL, Wilson KH, et al. Bartonella henselae: etiology of pulmonary nodules in a patient with depressed cell-mediated
immunity. Clin Infect Dis 1995; 20:1505–10.
Ventura A, Massei F, Not T, Massimetti M, Bussani R, Maggiore G.
Systemic Bartonella henselae infection with hepatosplenic involvement. J Pediatr Gastroenterol Nutr 1999; 29:52–6.
Adams WC, Hindman SM. Cat-scratch disease associated with osteolytic
lesions. J Pediatr 1954; 44:665–9.
Hulzebos CV, Koetse HA, Kimpen JLL, Wolfs TFW. Vertebral osteo-
Thiabendazole for the Treatment of Strongyloidiasis
in Patients with Hematologic Malignancies
A total of 21 patients with hematologic malignancies
were given thiabendazole for treatment of strongyloidiasis. Fifteen patients were cured. Since there were no relapses, it is unlikely that maintenance therapy has a role
in the management of strongyloidiasis in this population
of patients.
Thiabendazole has been considered the mainstay of treatment for strongyloidiasis for decades, with a clinical efficacy
of ∼90% in the immunocompetent host [1]. Nevertheless, information regarding its use in immunosuppressed patients is
limited to anecdotal reports of patients treated for the disseminated syndrome [2, 3], with little being known about the best
therapeutic regimen and its efficacy. In this study, we report
the treatment response to thiabendazole in 21 cases of strongyloidiasis diagnosed in a cohort of 163 patients with hematologic malignancies. All patients were screened for strongyloidiasis by examination of 3 stool samples with use of the direct
and Baermann-Moraes methods [4]. Patients with strongyloidiasis were treated with thiabendazole, 25 mg/kg, given twice
daily for 3 days (maximum dose, 3000 mg/day), and they were
subsequently monitored with monthly stool examinations. PaPresented in part: 10th International Symposium on Infections in the
Immunocompromised Host, Davos, Switzerland, 21–24 June 1998 (abstract
105).
Financial support: Conselho Nacional de Pesquisa, Brazil (grant CNPq
300235/93-3).
Reprints or correspondence: Dr. Marcio Nucci, Serviço de Hematologia,
Hospital Universitário Clementino Fraga Filho, Av. Brigadeiro Trompovsky,
s/n CEP 21941-590 Rio de Janeiro, Brazil ([email protected]).
Clinical Infectious Diseases 2000; 31:821–2
q 2000 by the Infectious Diseases Society of America. All rights reserved.
1058-4838/2000/3103-0037$03.00
18.
19.
20.
21.
22.
23.
821
myelitis associated with cat-scratch disease. Clin Infect Dis 1999; 28:
1310–2.
Margileth AM. Cat-scratch disease: etiology, diagnosis and therapy. Infect Med 1993; 10:38–43.
Zinzindohoue F, Guiard-Schmid JB, La Scola B, Frottier J, Parc R.
Portal triad involvement in cat-scratch disease. Lancet 1996; 348:
1178–9.
Waldvogel K, Regnery RL, 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.
Sander A, Posselt M, Böhm N, Ruess M, Altwegg M. Detection of
Bartonella henselae DNA by two different PCR assays and determination of the genotypes of strains involved in histologically defined
cat scratch disease. J Clin Microbiol 1999; 37:993–7.
Bass JW, Freitas BC, 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.
Margileth AM. Antibiotic therapy for cat-scratch disease: clinical study
of therapeutic outcome in 268 patients and a review of the literature.
Pediatr Infect Dis J 1992; 11:474–8.
tients with >3 negative stool samples after treatment were considered to be cured. The definition of disseminated strongyloidiasis was the same as that in our previous report [5].
From April 1995 through December 1997, 163 consecutive
patients with hematologic malignancies were evaluated, and 22
(13%) had strongyloidiasis. One patient was excluded because
he died of uncontrolled acute leukemia before the beginning of
treatment. The median age of the 21 patients was 46 years
(range, 12–76 years), and the male:female ratio was 16:5. Six
patients had acute leukemia, 7 had chronic leukemia, 6 had
lymphoma, and 2 had multiple myeloma. Corticosteroids were
being used by 17 (81%) of the patients. Symptoms attributable
to strongyloidiasis were present in 17 patients (81%), and cough
was the most frequent symptom (7 patients), followed by diarrhea and abdominal pain (6 patients each). Four patients
(19%) were asymptomatic. The median number of samples examined was 2 (range, 1–6), and the median number of positive
stool samples was 1 (range, 1–3).
One patient complained of dizziness, but all patients took
the full dose of thiabendazole. The median follow-up was 14
months (range, 1–29 months). Eighteen patients (86%) were
shown to be free of infection by all subsequent stool examinations (median, 4 examinations; range, 1–12 examinations).
Fifteen patients (71%) had >3 negative examination results and
were considered to be cured. Of the 3 patients for whom !3
examinations were performed, 2 died of acute leukemia in relapse and 1 died of sepsis of cutaneous origin. Three patients
remained infected after 1 course of thiabendazole (cure rate in
15 of 18 patients, 82%). Of the 3 patients for whom therapy
failed, 1 was cured after a 10-day course of treatment with
thiabendazole, whereas the other 2 died during relapse of their
underlying diseases, before the results of a second course of
treatment could be evaluated. Ten patients are still alive. One
patient was lost to follow-up and was censored after the last