Visceral zoster as the presenting feature of

disseminated herpes zoster

Erik Stratman, MD Marshfield, Wisconsin

Visceral dissemination of herpes zoster may follow cutaneous dissemination in immunocompromised

patients. The skin is not necessarily the only organ affected and may not even be the presenting organ.
Immunohistochemical stains available for routine paraffin-embedded tissue biopsy specimens allow for
rapid diagnosis of varicella zoster virus. We describe a patient in whom gastric dissemination of herpes
zoster was proven by immunohistochemistry. Unexplained hepatitis, pancreatitis, gastritis, or complaints of
abdominal pain in immunocompromised patients with herpes zoster should prompt a high degree of
suspicion for visceral zoster and immediate treatment with intravenous acyclovir. (J Am Acad Dermatol
2002;46:771-4.)

I n the setting of immunocompromise, varicella

zoster virus (VZV) reactivation can take 4 forms.
These include local (classic) zoster, dermatomal
zoster with dissemination, atypical generalized
zoster with or without visceral involvement, and vis-
of the back. The pain was associated with anorexia,
insomnia, and emesis. This was the third time in 3
days that he had sought medical treatment for the
pain, having previously been given a diagnosis of
nonspecific musculoskeletal pain, supported by
ceral zoster without skin lesions.1-3 Disseminated unremarkable findings of laboratory and radiograph-
cutaneous zoster is defined as greater than 20 vesic- ic studies.
ular lesions outside the primary and immediately His history was significant for non-Hodgkin’s low-
adjacent dermatomes.4 Cutaneous dissemination is grade nodular lymphoma stage IIA diagnosed 8 years
uncommon, most often occurring in immunocom- earlier. During those 8 years, he received 8 cycles of
promised patients, especially those with lymphopro- cyclophosphamide, doxorubicin, vincristine, and
liferative malignancies.2-11 Visceral zoster has been prednisone (CHOP); 8 cycles of cyclophosphamide
classically defined as histologic or culture evidence and prednisone; 39 months of daily chlorambucil; 2
of VZV or clinical evidence of internal organ involve- months of fludarabine; and most recently, 4 cycles of
ment without other causes in the setting of cuta- rituximab ending 4 months before presentation.
neous zoster.1,10 Such visceral dissemination of her- Laboratory test results were unremarkable except
pes zoster is an extremely rare event, typically fol- for low white blood cell count (3300/µL; normal,
lowing cutaneous dissemination.1,11 We describe a 4100-10,900/µL), low platelet count (169,000/µL;
patient with an unusual presentation of pathologi- normal, 175,000-450,000/µL), elevated alkaline phos-
cally confirmed disseminated visceral zoster. phatase level (135 U/L; normal, 35-120 U/L), and min-
imally elevated levels of aspartate aminotransferase
CASE REPORT (46 U/L; normal, 8-41 U/L) and pancreatic amylase
A 77-year-old man was first seen in the emergency (51 U/L; normal, 12-50 U/L). The patient was admit-
department with a 6-day history of progressively ted for further evaluation of abdominal pain.
worsening, constant, dull, nonpleuritic, right-sided, Subsequent upper endoscopy to rule out peptic
localized thoracic wall pain radiating to the right side ulcer disease revealed erosive gastritis and 8 to 10
small, shallow ulcers in the body of the stomach (Fig
From the Department of Dermatology, Marshfield Clinic. 1), which were biopsied. Early in the hospital course,
Funding source: Marshfield Clinic Department of Medical abdominal pain intensified, and serial hepatic and
Education. pancreatic enzyme levels rose dramatically, consis-
Conflict of interest: None. tent with acute pancreatitis and hepatitis (aspartate
Reprint requests: John Melski, MD, Marshfield Clinic, 1000 N Oak
aminotransferase, 348 U/L; alanine aminotransferase,
Ave, Marshfield, WI 54449.
Copyright © 2002 by the American Academy of Dermatology, Inc. 259 U/L; normal, 6-47 U/L; alkaline phosphatase, 440
0190-9622/2002/$35.00 + 0 16/91/119091 U/L; pancreatic amylase, 347 U/L; lipase, 957 U/L;
doi:10.1067/mjd.2002.119091 normal, 5-80 U/L).

771
772 Stratman J AM ACAD DERMATOL
MAY 2002

One-month follow-up evaluation of lymphoma

showed enlarged but stable abdominal lymph nodes.

DISCUSSION
After initial infection with primary VZV, the virus
spreads from the epidermis to sensory nerve end-
ings, then centripetally to the dorsal root ganglion
where it enters a latent stage.4 There are conflicting
reports in the literature regarding reinfection as a
significant cause of herpes zoster in patients with
cancer, but the best evidence suggests that reactiva-
tion is most likely.1,2,3,5,6,11,12 The mechanisms of
reactivation are poorly understood, but aging, trau-
ma, stress, radiation, and especially immunosup-
pression are reported triggers.3,4
The patient in this report had documented VZV in
the stomach and probable infection in the pancreas
and liver as well. Because visceral zoster usually pre-
sents in the setting of disseminated cutaneous
zoster, hematogenous spread of the virus to viscera
Fig 1. Endoscopic view of shallow ulcers in body of
stomach. cannot be ruled out. However, the level of thoracic
dermatomal skin involvement (right T7) shares the
same dorsal ganglion as the afferent sympathetic
fibers that supply these abdominal organs.13,14 Once
During this time, a band-like cluster of burning, triggered to reactivate, VZV may travel from the dor-
edematous, erythematous plaques appeared in the sal root ganglion centrifugally toward the skin or
right T7 distribution along with simultaneous scat- centripetally toward a viscus, because posterior sen-
tered vesicles over the body, particularly concentrat- sory nerve roots contain fibers from both skin and
ed on the head. Intravenous administration of viscera.15,16 Centripetal spread of the virus toward a
acyclovir, 800 mg every 8 hours, was started for pre- viscus is a rare event for unclear reasons and appears
sumed disseminated cutaneous zoster, thought to to occur primarily in the setting of immunocompro-
have been triggered by the acute pancreatitis. Direct mise. This centripetal spread would explain a mech-
fluorescence antibody testing from the base of an anism for visceral dissemination of herpes zoster.
unroofed skin vesicle was positive for VZV. Results of That a majority of patients with visceral dissemina-
radiographic biliary system evaluation, including an tion have or eventually have thoracic zoster supports
abdominal sonogram and computed tomographic this theory.1
scan, were unremarkable. A diagnosis of acalculous Visceral zoster should be defined as laboratory-
cholecystitis was entertained. Findings on endoscop- proven VZV infection of visceral organs. Proof can be
ic retrograde cholangiopancreatography were equiv- attained through immunohistochemical, direct
ocal for biliary sludge. immunofluorescence, genetic, or culture positivity
Pathologic examination of a gastric biopsy speci- for VZV. Clinical or serologic criteria alone are inade-
men demonstrated viral inclusions (Fig 2). quate in proving infection, especially when the
Immunoperoxidase staining was negative for herpes comorbid conditions and the co-pathogens that
simplex virus and cytomegalovirus. At the urging of often accompany VZV in immunocompromised
the dermatology resident, immunoperoxidase stain- patients are considered.1,17 Although there are sev-
ing for VZV was performed on gastric tissue, and eral reports of visceral involvement with VZV in the
results were positive (Fig 3). literature, only those with positive test results as list-
After the initiation of treatment with acyclovir, ed previously offer direct proof.5,8,18-20
hepatic and pancreatic enzyme levels began to Visceral zoster presents in the setting of immuno-
decline. The patient improved considerably, as did compromise, occurring in 3% to 15% of immuno-
his herpetic skin lesions. All laboratory abnormali- compromised patients with zoster.1,2,4,10,11 It mani-
ties, with the exception of the alkaline phosphatase fests in 10% of those with cutaneous dissemination
level, returned to normal within 12 days of acyclovir and almost always follows the rash.5 It has not been
therapy. The patient was then switched to twice-daily reported in immunocompetent patients. The lungs
oral valacyclovir, 1000 mg, for an additional 5 days. are the most frequent noncutaneous organ involved
J AM ACAD DERMATOL Stratman 773
VOLUME 46, NUMBER 5

Fig 2. Photomicrograph of gastric biopsy specimen demonstrates viral inclusions (arrow).

Fig 3. Photomicrograph of gastric biopsy specimen shows positive staining for varicella zoster.
(Immunoperoxidase stain; original magnification ×100.)

in VZV infections including varicella, and the liver is la (55%).1 It is important to note that some pub-
the most frequent abdominal organ involved.1 lished mortality rates predate antiviral treatment.2,6
Reports of stomach and pancreas involvement are Diagnosis can be difficult when visceral symptoms
rare, although both appear to be involved in this precede skin lesions. Because a lack of skin lesions
patient.1,10,12,17,21 When visceral dissemination often leads to a delay in diagnosis and treatment,
occurs, mortality rates are high, approximating the mortality rates are very high (in some reports nearly
mortality rates associated with disseminated varicel- 67%).1,5 As in our case, the presentation can be clin-
774 Stratman J AM ACAD DERMATOL
MAY 2002

ically confused with acalculous cholecystitis.8 3. Schimpff S, Serpick A, Stoler B, Rumack B, Mellin H, Joseph J, et
Abdominal symptoms can precede the rash by 1 to al. Varicella-zoster infection in patients with cancer. Ann Intern
Med 1972;76:241-54.
10 days, and in some reports, visceral reactivation
4. McCrary M, Severson J,Tyring S.Varicella zoster virus. J Am Acad
occurred in the absence of skin lesions.10,11,15,22 If Dermatol 1999;41:1-14.
this diagnosis is suspected, intravenous acyclovir 5. Rogers S, Irving W, Harris A, Russell N. Visceral varicella zoster
treatment should be started, given the high mortali- infection after bone marrow transplantation without skin
ty rates with delayed treatment and the relative safe- involvement and the use of PCR for diagnosis. BMT 1995;
15:805-7.
ty and superiority of treatment.6,8,22,23
6. Meyers J. Infection in bone marrow transplant recipients. Am J
Patients with lymphoma, particularly advanced Med 1986;81(Suppl 1A):27-38.
stage Hodgkin’s lymphoma, are at significantly 7. Goffinet D, Glatstein E, Merigan T. Herpes zoster-varicella infec-
increased risk for cutaneous and visceral dissemina- tions and lymphoma. Ann Intern Med 1972;76:235-40.
tion of zoster.3,7 This likely reflects the severity of the 8. Stemmer S, Kinsman K, Tellschow S, Jones R. Fatal noncuta-
neous visceral infection with varicella-zoster virus in a patient
underlying immune dysfunction as compared with
with lymphoma after autologous bone marrow transplanta-
other internal or pharmacologic causes of immuno- tion. Clin Infect Dis 1993;16:497-9.
suppression. A majority of cases of visceral zoster are 9. Mazur M, Dolin R. Herpes zoster at the NIH: a 20 year experi-
found in severely immunosuppressed patients after ence. Am J Med 1978;65:738-44.
bone marrow transplantation for an underlying lym- 10. Schiller S, Nimer S, Gajewski J, Golde D. Abdominal presentation
of varicella-zoster infection in recipients of allogeneic bone
phoma.1,2,5,6,8,10,12,22,24
marrow transplantation. BMT 1991;7:489-91.
In addition to underlying malignancy type, risk 11. Balfour H. Varicella zoster virus infections in immunocompro-
factors for VZV reactivation include chemotherapy mised hosts: a review of the natural history and management.
(30% of patients with lymphoma are actively receiv- Am J Med 1988;85(Suppl 2A):68-73.
ing chemotherapy at the time of VZV reactivation), 12. McCluggage W, Fox J, Baillie K, Coyle P, Jones F, O’Hara M.
Varicella zoster gastritis in a bone marrow transplant recipient.
radiation therapy (particularly to the area of subse-
J Clin Pathol 1994;47:1054-6.
quent zoster), and even neoplastic involvement of 13. Williams P. Gray’s anatomy. 38th ed. New York: Churchill
the involved dermatome.4,7 That 66% of patients Livingstone; 1995. p. 1306.
with lymphoma are clinically free of lymphoma at 14. Haubrich W, Schaffner F, Berk J. Bockus gastroenterology. Vol 2.
the time of zoster diagnosis suggests that therapy- 5th ed. Philadelphia: Saunders; 1995. p. 568, 2816, 1589.
15. Gais E, Abrahamson R. Herpes zoster and its visceral manifesta-
induced immunosuppression is a significant factor in
tions. Am J Med Sci 1939;197:817-25.
VZV reactivation. 3,7 16. Wisloff F, Bull-Berg J, Myren J. Gastric involvement in herpes
The significance of zoster as it relates to the zoster. Endoscopy 1980;12:134-5.
underlying lymphoma is not entirely clear. Local 17. Parenti D, Steinberg W, Kang P. Infectious causes of acute pan-
zoster does not appear to affect survival rates in creatitis. Pancreas 1996;13:356-71.
18. Gilden D, Dueland A, Devlin M, Mahalingam R, Cohrs R.Varicella-
patients with lymphoma.7 In bone marrow trans-
zoster virus reactivation without rash. J Infect Dis 1992:
plant recipients with zoster, there is no significantly 166(Suppl 1):S30-S34.
increased malignancy relapse rate.1 However, in 19. Nikkels A, Delvenne P, Sadzot-Delvaux C, Debrus S, Piette J,
other patients with lymphoma, 72% had recurrence Rentier B, et al. Distribution of varicella zoster virus and herpes
or progression of their lymphoma after zoster.7 This simplex virus in disseminated fatal infections. J Clin Pathol
1996;49:243-8.
suggests that zoster represents a poor prognostic
20. Dueland A, Devlin B, Martin J, Mahalingam R, Cohrs R, Manz H,
sign in patients with lymphoma and should indicate et al. Fatal varicella-zoster virus meningoradiculitis without skin
the need for careful surveillance for lymphoma involvement. Ann Neurol 1991;29:569-72.
recurrence or progression.3,7 21. Coehlo J, Wiederkehr J, Campos A, Zeni Neto C, Oliva V. Acute
pancreatitis caused by varicella-zoster virus after liver trans-
I thank Al Martin, MD, Don Schreiber, MD, Terri plantation. J Chir (Paris) 1994;131:96-8.
22. Munoz L, Balmana J, Martino R, Sureda A, Rabella N, Brunet S.
Perrone, MD, David Grekin, MD, John Melski, MD, Laura
Abdominal pain as the initial symptom of visceral varicella
Bliven, HT, and the Marshfield Clinic medical library staff
zoster infection in hematopoietic stem cell transplant recipi-
for their help in the completion of this manuscript. ents. Med Clin (Barc) 1998;111:19-22.
23. Pulik M,Teillet F,Teillet-Thiebaud F, Lionnet F, Genet P, Petitdidier
REFERENCES C. Varicella-zoster virus pancreatitis in hematologic diseases.
1. Locksley R, Flournoy N, Sullivan K, Meyers J. Infection with vari- Ann Med Interne (Paris) 1995;146:292-4.
cella-zoster virus after marrow transplantation. J Infect Disease 24. Perez-Oteyza J, Pascual C, Garcia-Larana J, Odriozola J,
1985;152:1172-81. Rocamora A, Navarro J. Abdominal presentation of varicella
2. Schuchter L, Wingard J, Piantadosi S, Burns W, Santos G, Saral R. zoster infection after bone marrow transplantation. BMT
Herpes zoster infection after autologous bone marrow trans- 1992;9:217-20.
plantation. Blood 1989;74:1424-7.