Skip Navigation

A Fatal Case of West Nile Virus Infection in a Bone Marrow Transplant Recipient

  1. Bradley Hiatt1,
  2. Lucy DesJardin2,
  3. Thomas Carter1,
  4. Roger Gingrich1,
  5. Curt Thompson2, and
  6. Margarida de Magalhaes-Silverman1
  1. 1Division of Hematology, Oncology, Blood and Marrow Transplantation, Department of Internal Medicine, College of Medicine, Iowa City
  2. 2Hygienic Laboratory, University of Iowa, Iowa City
  1. Reprints or correspondence: Dr. Margarida de Magalhaes-Silverman, Div. of Hematology, Oncology, Blood and Marrow Transplantation, Dept. of Internal Medicine, University of Iowa Hospital and Clinics, 200 Hawkins Dr., GH C-32, Iowa City, IA 52242 (margarida-silverman{at}uiowa.edu).

Abstract

West Nile virus (WNV) can cause severe, potentially fatal neurological illnesses, which include encephalitis, meningitis, Guillain-Barré syndrome, and anterior myelitis. Because of the short viremic phase, WNV infection is most commonly diagnosed by detection of immunoglobulin M antibody to WNV in serum or cerebrospinal fluid (CSF). We describe a patient with T cell lymphoma who had undergone a T cell-depleted bone marrow transplantation and developed fatal WNV infection. The results of serological tests of blood samples and of CSF tests were negative. Diagnosis was made postmortem by a positive result of reverse-transcriptase polymerase chain reaction (ABI 7700; TaqMan) for WNV in stored CSF and serum samples.

Case report. In September 2002, a 41-year-old man was admitted to the hospital with a 48-h history of fever, nausea, vomiting, and fatigue. A routine outpatient assessment performed 1 week earlier had revealed no abnormalities. The patient had undergone T cell-depleted, allogeneic, HLA, minor-mismatched-sibling bone marrow transplantation for relapsed T cell lymphoma 118 days before hospital admission. He had no recollection of any contact with ill persons. He was a resident of Illinois and did not have any history of travel. On 2 occasions during the month before admission, he underwent transfusion with packed RBCs; on another 2 occasions, he underwent transfusions with platelets. Medications being taken at the time of admission included cyclosporin (for graft-versus-host disease [GVHD] prophylaxis), prednisone (for a recent history of diffuse alveolar hemorrhage), voriconazole (for antifungal prophylaxis), and trimethoprim-sulfamethoxazole (for Pneumocystis carinii prophylaxis). At the time of admission, the patient did not have active GVHD.

The findings of a physical examination were remarkable only for a temperature of 38.5°C, with no signs of meningeal inflammation or skin rash. Blood and urine samples were obtained for culture. CT of the chest was nondiagnostic. The patient's hemoglobin concentration was 8.9 g/dL, his leukocyte count was 2.8 × 109 leukocytes/L (with 1.5 × 109 neutrophils/L and 0.2 × 109 lymphocytes/L), and his platelet count was 32 × 109 platelets/L. His biochemical profile was unremarkable. Empirical therapy with broad-spectrum intravenous antibiotics was started.

A lumbar puncture was performed on the second day of hospitalization because of persistent fever, and examination of the specimen revealed clear CSF at normal pressure. CSF protein and glucose levels were normal, and the leukocyte count was 3 leukocytes/mm3. The results of a Gram stain were negative, as were a cryptococcal antigen test and bacterial, fungal, herpesvirus, cytomegalovirus, and enterovirus cultures. The findings of an MRI of the brain, which was performed the following day to detect progressive hypersomnolence, were normal. Over the subsequent 4 days, the patient developed rapid, progressive, ascending weakness. The patient had no nuchal rigidity or sensory loss. However, he did have bilateral ptosis and areflexic weakness in both arms (proximal more than distal), bilateral leg weakness with hypoactive reflexes, neck weakness, and dysarthria. A tensilon test was performed, and the results were equivocal. Electromyography and nerve conduction studies were indicative of an acute demyelinating polyneuropathy, with associated axonal loss compatible with the diagnosis of Guillain-Barré syndrome. On the fifth day of hospitalization, the patient's mental status further deteriorated, and he underwent intubation and received mechanical ventilatory support for airway protection. An electroencephalogram (EEG) performed on day 6 of hospitalization showed a poorly defined, 5–7-Hz background, with diffuse frontal dominant θ slow waves without epileptiform activity most consistent with diffuse cerebral dysfunction. The patient's mental status progressively deteriorated to a vegetative state, and treatment with intravenous immunoglobulin was started.

CSF analysis was repeated 10 days after hospital admission, and the results were again unremarkable, except for a leukocyte count of 6 leukocytes/mm3 (5 lymphocytes). The results of PCR for human herpesvirus 6 and enterovirus were negative. Serum and CSF specimens were sent to the University of Iowa Hygienic Laboratory (Iowa City) for detection of IgM antibodies against WNV using an IgM antibody capture (MAC) ELISA. Both the serum and the CSF samples tested negative for WNV IgM antibody. EEG was repeated on hospital day 10, and the findings remained consistent with diffuse cerebral dysfunction. Repeated brain MRI showed increased signal within the superior posterior medulla at the pontomedullary junction. The patient died 17 days after hospital admission. An autopsy was not granted by the family. However, premortem serum and CSF samples were available for additional testing.

Because of the clinical picture and the fact that the patient was immunocompromised after his bone marrow transplantation and, consequently, that antibody response could have been blunted, 2 serum samples and 1 CSF specimen were tested for the presence of WNV RNA using an ABI 7700 assay (TaqMan) [1]. The CSF sample from day 10 of hospitalization tested positive by RT-PCR, thus meeting the national case definition of a confirmed case of WNV infection [2]. A serum sample obtained on day 3 of hospitalization tested negative by RT-PCR. However, a second serum sample, which had been obtained 4 days later, tested positive using the TaqMan assay. All RT-PCR results were confirmed by TaqMan RT-PCR using primers that target a different region of the WNV genome [1]. There was no evidence of contamination in any of the control samples. An additional 48 CSF samples that were submitted to the Hygiene Laboratory were also tested, and only 1 sample tested positive by TaqMan RT-PCR. This sample was correspondingly positive for IgM using the MAC ELISA.

Discussion. WNV was first identified in the United States in 1999 during an outbreak among birds and horses and an epidemic of meningitis and encephalitis in the New York City area [3]. By 2002, WNV had caused the largest documented arboviral outbreak of meningitis/encephalitis in the Western hemisphere. The patient we describe resided in Illinois, which had the greatest number of cases of WNV infection in the United States in 2002: there were 884 cases of WNV infection and 64 associated fatalities [4]. Although our patient lived in an area with high levels of WNV activity, he had received several transfusions during the month before symptom development. Recent investigations provided evidence that WNV can be transmitted through blood transfusions, through organ transplantation, and in utero [5]. For that reason, we cannot exclude blood transfusion as the route of transmission. Investigations are currently in progress to further evaluate that possibility.

Neurological manifestations of WNV infection include aseptic meningitis, encephalitis, and an acute flaccid paralysis attributed to a peripheral demyelinating process (Guillain-Barré syndrome) or to a polio-like syndrome involving the anterior horn cells of the spinal cord and motor axons [69]. The pathogenesis of severe infection with WNV is not well understood, but the increased risk of neurological infection and more-aggressive disease in elderly persons suggests that immune incompetence is a risk factor. Experiments involving mice support the role of B cells and antibody in defense against disseminated WNV infection [10]. The described patient had undergone T cell-depleted bone marrow transplantation ∼3 months before development of his illness, and he was severely immunosuppressed (and was also receiving immunosuppressive drugs) as a consequence. Immune reconstitution has been carefully evaluated in recipients of allogeneic bone marrow transplants. Quantitatively, cellular components of the immune system return to normal levels in 4–6 months after conventional bone marrow transplantation, with the exception of CD4+ T cells and B cells, for which a return to normal levels is often delayed. T cell proliferation and immunoglobulin production remain impaired usually until the second half of the first year after transplantation. T cell depletion of the graft further compromises T cell engraftment and function and is associated with an additional delay in immune recovery [9, 11]. The diagnosis of WNV infection rests on a high index of clinical suspicion and on results of specific laboratory tests, such as the MAC ELISA, viral culture, and RT-PCR. Viral cultures of specimens of CSF or human brain tissue have had very low yield; results of nucleic acid amplification have been positive in up to 55% of CSF samples [2].

This patient had results similar to those for a similar patient who was receiving immunosuppressive therapy [12], for whom the results of serological tests were negative for antibodies. The most likely explanation is that our patient was unable to mount an antibody response. However, the sensitivity of MAC ELISA is not 100%, and thus it is possible that antibody was present but at levels less than the limit of detection. The presence of WNV RNA was detected in both serum and CSF specimens by RT-PCR, which indicates infection with WNV.

Recently, a patient who developed encephalitis while undergoing treatment with cyclophosphamide, hydroxydoxubicin, vincristine, and prednisone (CHOP) and rituximab for non-Hodgkin B cell lymphoma has been described elsewhere [12]. In that patient, the diagnosis of WNV infection was established by TaqMan PCR testing of the CSF. Several CSF and serum samples tested negative for WNV antibody, as occurred for the patient we describe here. As with the patient in our study, this patient was immunosuppressed as a consequence of her disease and likely because of its treatment [13].

In summary, we report a fatal case of WNV infection in an immunocompromised patient who had undergone allogeneic, T cell-depleted bone marrow transplantation. The patient's serological test results were negative, and the diagnosis was made after death by RT-PCR of both CSF and serum specimens.

  • Received April 4, 2003.
  • Accepted July 8, 2003.

References

| Table of Contents

Navigate This Article