Primary Epstein-Barr Virus Infection with Pneumonia Transmitted by Allogeneic Bone Marrow after Transplantation

Epstein-Barr virus (EBV) usually infects people who are <30 years old. Among immunocompetent patients, primary infection is usually asymptomatic or presents as infectious mononucleosis. After allogeneic hematopoietic stem cell transplantation (HSCT), reactivation of EBV infection most often induces post-transplant lymphoproliferative disorders (PTLDs). Reactivation of EBV infection after HSCT is almost always of donor origin [1]. Risk factors for PTLD that have been identified in retrospective studies include donor type (unrelated) and use of antithymoglogulin in the conditioning regimen [1]. To our knowledge, primary EBV infection presenting as pneumonia has not been reported for patients with recent HSCT. Here, we report a case of primary EBV infection, transmitted from the donor to an EBV-seronegative transplant recipient, that occurred several weeks after HSCT.

Case report. Our patient was a 9-year-old girl with normal development and severe homozygous sickle cell anemia. Allogeneic bone marrow transplantation was indicated because of multiple vaso-occlusive crises, recurrent chest syndromes, and elevated velocity of cerebral artery flow. The donor, who was HLA matched, was the recipient's healthy 17-year-old brother. The donor and the recipient had serological test results, obtained 1 month before transplantation, that were negative for IgM and IgG antibodies against cytomegalovirus, EBV, and toxoplasma. The conditioning regimen consisted of busulfan, cyclophosphamide, and rabbit antithymocyte globulin. Graft-versus-host disease (GVHD) prophylaxis consisted of the combination of methotrexate and ciclosporin. Prophylaxes used for infection were amoxicillin, ofloxacin, oral amphotericin B, fluconazole, and valacyclovir. Cotrimoxazole was added to the regimen when the patient's neutrophil count reached 0.5 × 10⁹ cells/L.

On day 21, fever (temperature, >39°C) and dry cough appeared. There was no sign of GVHD and no peripheral lymphadenopathy, chest auscultation was normal, and inflammatory markers in blood (e.g., fibrinogen and C-reactive protein level) were within normal ranges. The patient received broad-spectrum antibiotics without subsequent improvement. Cultures of blood, stool, sputum, and urine samples were negative for bacterial and fungal pathogens. Results of immunofluorescent staining of nasopharyngeal secretions were negative for respiratory syncytial virus; influenza virus A and B; parainfluenza virus 1, 2, and 3; and adenovirus. PCR analysis of blood samples did not show human herpesvirus 6, human herpesvirus 8, parvovirus B19, cytomegalovirus, adenovirus, or toxoplasma. CT scans of the lung (performed on days 23 and 29), head (day 29), and abdomen (day 29) had normal findings. Because of persistent fever associated with mononucleosis syndrome, quantitative EBV PCR was performed (as described in Sullivan et al. [2]) on day 25 for samples of blood and nasopharyngeal secretions, both of which had results that were highly positive (123,000 copies/mL and 1344 copies/mL, respectively) (figure 1). Quantitative PCR for EBV was retrospectively performed on stored blood samples and showed a progressive increase from day 11 (figure 1). Immunophenotyping of peripheral lymphocytes on day 26 showed that reconstitution was essentially a reconstitution of natural killer cells with activated T cells (94% of HLA-DR-positive cells) (table 1). Because a post-transplant EBV reactivation was suspected, rituximab (administered at a dose of 375 mg/m²) was introduced on day 26, with subsequent courses administered on day 29 and day 36. Fever disappeared by day 30, although cough persisted, and the results of PCR analysis of blood samples for EBV promptly decreased; the patient's results were negative by day 35 (figure 1). In contrast, hyperlymphocytosis of up to 11 × 10⁹ cells/L persisted in blood samples obtained after rituximab therapy was administered (figure 1), mainly because of an increase in CD8⁺ T cells. Furthermore, these CD8⁺ T cells shared features of activated (HLA-DR-positive cells, 98%) and effector memory cells (D45RA- CCR7-negative cells, 77%) (table 1). of note, B cells were absent, as was expected after administration of rituximab. Analysis of lymphocytes on day 24 confirmed 100% donor chimerism.

Figure 1

Clinical and biological correlation of primary Epstein-Barr virus (EBV) infection in an allogeneic stem cell transplant recipient. Febrile bronchitis occurred 3 weeks after allogeneic bone marrow transplantation and was suspected to be related to EBV infection. Indeed, we observed a high viral load at day 26, which promptly decreased after anti-CD20 monoclonal antibody treatment with rituximab on day 26 (D26), day 29 (D29), day 36 (D36), and day 49 (D49). An elevated lymphocyte count was detected on D26 and slowly decreased after the second course of rituximab. Fever (temperature, >39°C) definitively resolved after corticosteroid therapy. DC, Downey cells; Ly, lymphocyte count; MP, methylprednisolone.

Table 1

Lymphocyte and lymphocyte subtype counts for a patient with primary Epstein-Barr virus infection with lung involvement occurring 1 month after bone marrow transplantation.

On day 36, fever returned. Chest auscultation revealed bilateral crackles. A CT scan, performed on day 39, showed bibasal nodular and alveolar opacities. On day 40, bronchoalveolar lavage (BAL) was performed. Alveolar fluid contained 150,000 cells/mL with 69% lymphocytes, 28% macrophages, 2% neutrophils, and no RBCs. Immunophenotyping of BAL fluid showed 70% CD8⁺ T cells, 13% CD4⁺ T cells, 5% natural killer cells, and the absence of CD19⁺ B cells. An extensive search for bacteria, mycobacteria, fungal pathogens, toxoplasma, and pneumocystis had negative findings. PCR analysis was negative for mycoplasma, toxoplasma, pneumocystis, herpes simplex virus, human herpesvirus 6, cytomegalovirus, enterovirus, and adenovirus. Immunofluorescent staining did not show respiratory syncytial virus; influenza virus A or B; parainfluenza virus 1, 2, or 3; or adenovirus. Finally, EBV was identified by PCR (13,338 copies/mL) in alveolar fluid. The diagnosis of EBV pneumonia was retained despite PCR results that were negative for EBV in blood. This discrepancy could be explained by the rituximab treatment.

Because the patient remained febrile and required oxygen therapy, on day 36, we began a course of corticosteroid treatment with methylprednisolone, administered at a dosage of 1 mg/kg per day. Fever disappeared definitively by day 40. To avoid reoccurrence of EBV infection, a fourth injection of rituximab was administered on day 47.

Thus, our patient was considered to have developed primary EBV infection with pulmonary involvement. Retrospectively, a cryopreserved sample of the donor's bone marrow was found to be positive for EBV by PCR (2568 copies/mL), and the donor described a history of afebrile upper respiratory symptoms just before stem cell donation. Moreover, by day 90, the recipient's brother showed EBV seroconversion (both IgM and IgG antibodies were detected in serum samples).

Corticosteroid therapy was progressively decreased, and it was changed to therapy with mycophenolate mofetil for unique GVHD prophylaxis by day 105, because cyclosporine therapy was discontinued on day 36 as the result of renal toxicity. Cough and bilateral crackles and squeaks persisted until 8 months after HSCT. CT scans of the lungs, performed at 4 months and 6 months after transplantation, indicated the development of bronchectasis and hypoattenuated lung areas, evocating small airways disease. At this time, pulmonary function tests found a mild decrease of forced expiratory volume in 1 second (to 70% of the predicted value) and an increase in airflow resistance (to 156% of the predicted value) and residual volume (to 143% of the predicted value) with normal total lung capacity. Diagnosis of post-transplantation bronchiolitis probably triggered by EBV infection was made 4 months after HSCT. The patient had no other signs of chronic GVHD.

Discussion. Primary EBV infection after HSCT is currently not well defined, and the differentiation between a primary infection and reactivation of infection is not always easy. Primary EBV infection concomitant with HSCT was reported >2 decades ago to cause multiorgan infiltration mimicking GVHD [2]. In our case, EBV reactivation was unexpected, because the recipient and the donor were seronegative for EBV 1 month before transplantation. The only risk factor for the patient was the use of antithymoglogulin in the conditioning regimen [1]. Because the patient had a fever that was resistant to the usual antibiotic therapy, PCR testing of blood samples for EBV was performed, and very high levels of EBV were found. In investigating the origin of this EBV transmission, clinical signs compatible with primary infection in the donor were retrospectively found; in addition, EBV DNA was found in a cryopreserved bone marrow sample obtained from the donor. These findings suggest transmission by the donor's bone marrow. The donor acquired EBV immunity 1 month after donation of stem cells. With respect to the immune response to EBV in the recipient, this response could not be attributed to naive T cells, which are never detectable before the third month after transplantation [3]. Thus, although the bone marrow graft was infected with EBV, it possibly also contained EBV-specific T lymphocytes, preformed in the donor, who was undergoing primary infection. The early proliferation of natural killer cells may be a nonspecific innate response consistent with either a response to EBV or a typical post-transplantation lymphocyte reconstitution [4]. In contrast, proliferation of activated effector memory CD8⁺ T cells 40 days after transplantation is unusual and may suggest an immune response to EBV infection. In our patient, we had no evidence for a PTLD, which can be localized in the lung [5]; no tumoral cells were detected in the BAL fluid, and no monoclonal band was disclosed by protein electrophoresis.

EBV pneumonitis was suspected in our patient, because a high EBV load was found in samples of blood and BAL fluid. The respiratory tract is known to be a reservoir for EBV [6], although pneumonia remains a rare manifestation of EBV infection. Among immunocompetent patients, pulmonary involvement is uncommon but possible [7], notably during childhood. Recently, case reports of 3 infants with clinical features similar to those of our patient have been published [8, 9]. Respiratory symptoms progressively resolved with corticosteroid and antiviral therapy (acyclovir and gancyclovir). Analysis of BAL fluid showed lymphocytic alveolitis for 2 patients, similar to the findings of BAL fluid analysis in our case.

Reasons for lung involvement in patients with EBV infection are not clear, because the virus has been found in cells that do not express CD21, the EBV receptor. Indeed, EBV may infect not only B lymphocytes in the lung but also macrophages in immunocompetent adults [10], type 2 alveolar cells in adults with cryptogenic fibrosing alveolitis [11] or with AIDS [12], and cuboidal epithelial cells in adults with idiopathic pulmonary fibrosis [13]. In our case, the persistence of EBV in cells other than B lymphocytes after rituximab therapy could explain the persistence of infection. Moreover, rituximab is not the recommended treatment for severe primary EBV infection, but this patient was treated for a post-transplantation PTLD [14]. On the other hand, control of EBV load in blood by rituximab therapy is not necessarily related to a decrease in tumor burden for patients with PTLD [15].

Conclusion. Primary EBV infection transmitted by a donor's bone marrow can occur after HSCT. Physicians became aware that signs of benign respiratory infection in a donor seronegative for EBV infection can reveal an EBV infection. Rituximab allowed a rapid decrease of EBV load in blood but was not sufficient to control the symptoms. A reactive T lymphocyte proliferation and a local inflammatory response in the lung probably resulted in the persistence of respiratory symptoms. The patient's clinical response to corticosteroid therapy supports this hypothesis.

• Received March 3, 2006.
• Accepted May 16, 2006.

• © 2006 by the Infectious Diseases Society of America