Disseminated Infection With Varicella-Zoster Virus Vaccine Strain Presenting as Hepatitis in a Child With Adenosine Deaminase Deficiency

DISSEMINATED INFECTION WITH VARICELLA-ZOSTER VIRUS VACCINE STRAIN PRESENTING AS HEPATITIS IN A CHILD WITH ADENOSINE DEAMINASE DEFICIENCY The Oka strain of live attenuated varicella virus vaccine was licensed for use in healthy children in the United States in March, 1995. This vaccine is contraindicated in immunocompromised patients. We report a case in which varicella vaccination in a child resulted in the identification of his undiagnosed immunodeficiency. Case report. A 13-month-old white male child was hospitalized for suspected sepsis. His past medical history was significant for failure to thrive, chronic diarrhea, chronic thrush and multiple episodes of bronchiolitis. He had received varicella vaccine (Varivax, Oka/Merck strain) 2 weeks before admission. On admission he was in respiratory distress and required mechanical ventilation. He also had diarrhea, elevated serum transaminase values and abnormal coagulation studies. The liver was firm and palpable to the right iliac crest. No bacteria, fungi or parasites were isolated from blood, urine, cerebrospinal fluid or stool cultures. A viral culture of a bronchoalveolar lavage specimen yielded rhinovirus, enterovirus and parainfluenza virus type ΙΙΙ. He was treated with broad spectrum antibiotics . Chest computerized tomography revealed right upper and left lower lobe atelectasis. Computerized tomography of the abdomen showed ascites, hepatomegaly with no focal lesions and edematous bowel wall consistent with colitis. A liver biopsy demonstrated multifocal, well-demarcated areas of necrosis (Fig. 1). These necrotic foci varied in size and were associated with an inflammatory infiltrate composed of histiocytes with few neutrophils and lymphocytes. There were round, eosinophilic structures in the necrotic foci both within the parenchyma and within macrophages, which were interpreted to be most likely degenerating hepatocytes. Viral inclusions were not seen. Routine bacterial, fungal and viral cultures of the liver biopsy were negative. The supernatant of the liver homogenate prepared before culture was frozen and later submitted for viral DNA testing (see below).Fig. 1: A well-demarcated area of hepatic necrosis with associated inflammatory infiltrate and degenerating hepatocytes. Hematoxylin and eosin; original magnification, ×200.Initially the total leukocyte count was 1300/μl with an absolute neutrophil count of 585 cells/mm 3; then he developed leukocytosis of up to 60 000/μl with increased immature forms. The absolute lymphocyte count ranged from 270 to 455 cells/mm 3. Two weeks after admission (4 weeks after vaccination) he developed maculopapular and vesicular skin lesions on his trunk and extremities. A punch biopsy taken from one of the lesions demonstrated an intraepidermal vesicle with several multinucleated giant cells with characteristic inclusions consistent with either herpes simplex virus or varicella-zoster virus (VZV). Varicella-zoster virus was detected on three occasions by direct microscopic examination of vesicle material, which was stained using a fluorescein-labeled monoclonal antibody directed against antigenic determinants of the virus (Meridian Diagnostics, Inc., Cincinnati, OH). The presence of the virus in the skin vesicles was also confirmed by conventional viral culture. PCR amplification of the viral isolate from the skin lesion and the viral culture from the supernatant of the liver biopsy performed at Columbia University, New York, NY, by Dr. Philip La Russa revealed that both samples contained the OKA/Merck vaccine strain of VZV. 1 The patient was treated with intravenous acyclovir with complete resolution of his skin lesions. His past history of chronic infections, low lymphocyte count and positive PCR for the vaccine strain of VZV raised the issue of immunodeficiency. A PCR test for HIV-1 was negative. Serum immunoglobulins were in the low range of normal (IgG 345 mg/dl, IgM 62 mg/dl and IgA 52 mg/dl). Chemiluminescence assay was normal. Flow cytometric analysis revealed adequate expression of CD11b on neutrophils. The lymphocyte quantitative and qualitative analysis showed a marked decrease in the absolute and relative (percentage) numbers of B (CD19 34 cells/μl, 4.9%) and T cells (CD3 147 cells/μl, 21%; CD4 73 cells/μl, 11%; CD8 68 cells/μl, 10%). The functional assay demonstrated virtually no T cell proliferative activity to antigen and mitogen stimulation. Adenine deaminase blood concentration was 0.19 nmol/min/mg protein (normal range, 8 to 20), which is diagnostic for adenosine deaminase deficiency. Serum transaminase values returned to normal 2 weeks after hospitalization. During the third week of hospitalization he was weaned from the ventilator. He was inadvertently given dexamethasone (0.3 mg/kg every 6 h for a total of 10 doses) in preparation for extubation. Five days later he developed multiple vesicular lesions at new sites on the trunk and extremities. A viral antigenic determinant from the lesions was again positive for VZV. Treatment with intravenous acyclovir was restarted with prompt response. He was extubated 1 week later. He was discharged home after 5 weeks of hospitalization. Enzyme replacement therapy with bovine ADA covalently modified with polyethylene glycol was initiated (Adagen; Enzon, Inc., Piscataway, NJ). Discussion. Varicella vaccine (Varivax) was licensed in the United States in 1995 for routine use in healthy children and adults. Clinical studies have demonstrated its effectiveness in preventing severe disease and its favorable safety profile. 2 The vaccine is contraindicated in immunocompromised individuals, such as those with congenital immunodeficiency, blood dyscrasias, leukemia, lymphoma, symptomatic HIV-1 infection and malignancy for which they are receiving immunosuppressive therapy. 3 An exception includes children with acute lymphocytic leukemia to whom vaccine may be given under a special compassionate use program. Children receiving prednisone for allergic conditions in dosages of <2 mg/kg/day are not considered immunocompromised and may be vaccinated. Immunodeficiency should be excluded before immunization in children with a family history of hereditary immunodeficiency. In our patient, who had a history of failure to thrive and chronic infections, varicella vaccine should not have been administered until immunodeficiency has been excluded. Hepatic involvement has been described in childhood varicella but is most often subclinical. 4 A severe form of the disease, progressive varicella, can develop in the immunocompromised patient, with a mortality of 20%. These patients have a high incidence of visceral involvement, including pneumonitis, meningoencephalitis and hepatitis. 5 Adenosine deaminase (ADA) deficiency is inherited as an autosomal recessive trait. Complete deficiency of ADA results in severe combined immunodeficiency (SCID). Most patients present as infants with SCID, whereas others are diagnosed in childhood (delayed onset) or as adults (late onset). It accounts for 10 to 20% of all cases of SCID. Compared with other types of SCID, ADA deficiency causes the most profound lymphopenia. 6 Moderately elevated serum hepatic transaminases, the cause and importance of which are uncertain, are found in some patients with ADA deficiency. In a small series of autopsy findings in eight patients with ADA deficiency, severe bridging portal fibrosis was found in two of them. 7 The histologic changes in the liver are the first description of the pathologic alterations of hepatitis caused by the vaccine strain of VZV. Varicella hepatitis is histologically distinctive, with patchy coagulative hepatocytic necrosis that incites only a mild inflammatory response. Viral inclusions characteristic of varicella may be seen at the periphery of the lesions. The absence of viral inclusions in this patient seemed inconsistent with the clinical suspicion of immunodeficiency, in which one would suspect large numbers of inclusions. However, the vaccine strain of the virus is an attenuated strain, which might account for the lack of viral inclusions in the liver. Hepatitis secondary to either enterovirus or parainfluenza type ΙΙΙ was a clinical consideration, but these viruses do not typically produce the pathologic alterations seen in this biopsy. 8 In addition viral culture of the liver was negative, but the vaccine strain was detected in viral culture supernatant of the liver biopsy using PCR analysis. Our patient had undiagnosed SCID at the time of vaccine administration, was unable to mount a humoral and cellular immune response and thus developed a severe case of varicella. VZV IgM was <1:10 and IgG was 1:20, 2 weeks after the onset of the rash. He had a relapse of varicella after he inadvertently received intravenous dexamethasone (dose equivalent to 7-mg/kg/day dose of prednisone). He responded well to acyclovir therapy on both occasions. 9 This case represents the unusual presentation of disseminated varicella virus infection from a vaccine strain. Whereas disseminated infection with the vaccine strain has been reported in immunosuppressed patients, the primary presentation of hepatitis and the histopathologic changes have not been previously described. This case underscores the importance of early diagnosis of immunodeficiency in children with repeated infections before administration of live virus vaccines.