The natural history and clinical significance of posttransplant Epstein-Barr virus (EBV) infection remain largely unknown. The aims of this study are to describe the incidence, risk factors and consequences of EBV infection after kidney transplantation. A total of 383 consecutive patients having received a kidney transplant between January 2002 and December 2010 were included. EBV polymerase chain reaction (PCR) was performed every 2 weeks for 3 months, and every 4 weeks for the next 9 months. A total of 155 of the 383 patients (40%) had at least one positive viremia during the first year posttransplant. The median time to viremia was day 31 posttransplant (14–329). A total of 73 (47%) had EBV viremia > 103 log and 23 (15%) had positive viremia for more than 6 months. EBV D+/R− patients (12/18 (67%) versus 143/365 (39%), p = 0.02) and those having received antithymocyte globulins (ATG) (54% vs. 35%; p<0.001) were more likely to develop EBV infection. EBV infection (hazard ratio [HR], 3.03; 95% confidence interval [CI], 1.72–8.29; p = 0.01) was associated with the occurrence of opportunistic infections. A positive EBV PCR during the first 6 months posttransplant was associated with graft loss (HR, 3.04; 95% CI, 1.36–6.79; p = 0.014). EBV reactivation is frequent after transplantation and reflects overimmunosuppression. Prospective studies should examine the association between EBV and graft loss.
Measurement of the Epstein–Barr virus (EBV) load in the peripheral blood using polymerase chain reaction DNA amplification is currently used in patients in whom EBV infection could be associated with the subsequent development of lymphoma. Nevertheless, this strategy mainly concerns EBV D+/R− patients [1, 2]. Thus, contrary to other viruses, such as cytomegalovirus (CMV), the natural course of EBV infection is not accurately defined in adult solid organ transplant (SOT) recipients. Only one small study reported on the prospective monitoring of EBV infection in the first 3 months following kidney transplantation . Consequently, most evidence derives from pediatric organ transplant or bone marrow recipients and cannot be easily extrapolated to adult SOT recipients [4, 5]. Moreover, the available literature mainly concerns EBV infection in EBV D+/R− patients and little is known about the natural history of EBV reactivation. However, EBV reactivation occurs in adults in different other clinical situations associated with chronic immunosuppression and is likely to be underestimated in transplant patients [6, 7]. Similarly, a number of other questions remain without accurate answer. Stimuli that could trigger EBV reactivation are only partially defined. Different risk factors for posttransplant lymphoproliferative disease (PTLD) have been described including EBV sero-mismatch, degree of immunosuppression, age, type of organ transplanted and coinfection with CMV [8, 9]. Nevertheless, whether patients’ characteristics or immunosuppressive strategies influence posttransplant EBV infection is not known. Prolonged EBV infection is likely to reflect overimmunosuppression and may identify a population carrying an increased risk of serious infectious complications. Finally, whether EBV infection may influence graft survival is not clearly defined.
The aims of this study are to analyze the natural course of EBV infection in adult kidney transplant recipients, to define risk factors and to assess clinical consequences of EBV infection.
Patients and Methods
Study design and populations
We analyzed a prospective cohort of 383 consecutive patients having received a kidney transplant at the Transplant Unit of the University Hospital of Besançon between January 2002 and December 2010.
All the patients received a quadruple sequential immunosuppression. Induction consisted of either ATG (n = 107, 28%) (ATG Fresenius® [day 0: 9 mg/kg; days 1–4: 3 mg/kg/day]) or monoclonal anti-CD25 antibody (n = 276, 72%) (Simulect® [Novartis] [day 0: 20 mg, day 4: 20 mg]). The same maintenance immunosuppressive treatment was used including tacrolimus, mycophenolate mofetil and steroids. Target levels of tacrolimus were based on C0 of 0–3 months, 8–10 ng/mL, 3–6 months, 6–8 ng/mL and > 6 months, 5–7 ng/mL. Mycophenolate mofetil dosage was adapted to the MPA area under the curve (30–60).
All the patients except CMV seronegative recipients of a CMV seronegative donor received CMV prophylaxis with Valaciclovir (2002–2007) or Valganciclovir (2008–2010) in the first 3 months following transplantation. Antiviral prophylaxis dose was adapted to renal function (maximal dose for Valaciclovir and Valganciclovir were 8 g and 900 mg/day, respectively). All patients received Pneumocystis antimicrobial prophylaxis with trimethoprim-sulfamethoxazole.
EBV D+/R− patients received intravenous immunoglobulins (0.4 g/kg/day given on 5 consecutive days) until month 6 posttransplant and Valaciclovir independently of CMV serologic status. ATG was proscribed in EBV D+/R− patients.
Characteristics of the study population are described in Table 1.
Table 1. Characteristics and outcomes of patients
No EBV viremia (n = 228)
EBV viremia (n = 155)
ATG = antithymocyte globulins; HLA = human leukocyte antigen.
48 ± 14
49 ± 14
Gender male (%)
HD duration (months)
18 ± 16
16 ± 15
3.7 ± 1
3.6 ± 1.1
ATG use (%)
Delayed graft function (%)
Acute rejection (%)
All the patients were prospectively monitored for viremia eomN EBV polymerase chain reaction (PCR) every 2 weeks for 3 months, followed by every 4 weeks for the next 9 months.
After the first year posttransplant, all the patients had EBV viremia once a year. If positive, viremia was repeated on each visit.
Persistent viremia was defined as being unchanged or increasing viral loads over two separate measurements.
A decrease in immunosuppression was performed in patients with EBV PCR > 104 log/mL on two consecutive separate measurements. Mycophenolate mofetil (MMF) was first tapered and eventually withdrawn if PCR remained above 104 log/mL.
The samples were whole-blood Ethylenediaminetetraacetic acid-treated, homogenized for 10 min at room temperature by gentle shaking and stored in 1 mL aliquots in 1 mL at −20°C until use. The routine EBV diagnosis was performed twice a week. Prior amplification, samples were thawed and DNA was extracted with a QIAamp DNA Blood kit (Qiagen, Courtaboeuf, France). Two hundred microliters of whole blood was mixed with proteolysis solution (20 μL of proteinase K, 200 μL of AL buffer) homogenized for 30 s, briefly centrifuged and incubated at 70°C for 10 min. DNA was immobilized on magnetic beads in a King fisher device (Thermolab system) using “BS15 DNA Blood 200” program. This procedure yields 100 μL of DNA extract. A 10-fold serial dilutions in water (100 000 to 1 000 copies of EBV DNA) were prepared from Namalwa diploid cell line (ATCC CRL 1432) for quantification curve. Each cell harbors two integrated copies of the EBV genome . Negative and positive controls were water and a previous positive clinical sample chosen for an EBV viral load comprised between 5 000 and 25 000 copies, respectively. Stock solutions of these controls were frozen at −20°C. A 76 bp region of the BALF4 gene (coding for the surface glycoprotein, 110 kDa) is targeted for amplification using primers selected by BLAST database analysis. The upstream and downstream primer sequences were 5′-CTCTTTTTGCTCCTGGTTTTGC-3′ (position 157920 of the sequence) and 5′-GGCATTAATCCCATTAGTAAGACAGAA-3′ (position 157996), respectively. The sequence of the taqman probe is 5′ FAMCATGCAGCGCTAACATGATGGCTTG -TAMRA3′.
The β globin gene as DNA control is amplified with the following upstream primer sequence: 5′-TCCCCTCCTACCCCTACTTTCTA-3′, downstream primer 5′-TGCCTGGACTAATC TGCAAGAG-3′ and the taqman probe 5′VIC-TCACAGAGGCTTTTTGTTCCCCCAGAC-TAMRA3′. Primers and probe were purchased from Eurogentec. β globin gene amplification is carried out simultaneously with BALF4 target in order to control the presence or absence of DNA polymerase inhibitors. The PCR mixture consist in 5 μL of DNA extract, 100 nM of each β globin primer, 250 nM of each BALF4 primer, 100 nM of each taqman probe in a 50 μL final volume with a TaqMan® Universal PCR Master Mix (N° 4318157 Applied Biosystems Foster City, CA, USA).
Cycling conditions are as follows: one cycle at 50°C for 2 min followed by 1 cycle at 95°C for 10 min followed by 45 cycles of 95°C for 30 s and 60°C for 90 s.
Comparison between calculated threshold cycle (Ct) of each sample with Ct of quantification curve points allowed the determination of viral loads in copies/mL and transformed in log/mL.
Age, gender, weight, size, hemodialysis duration before transplantation, diabetes, pretransplant history of cancer, panel reactive antibody, rank of transplantation (first vs. iterative), human leukocyte antigen (HLA) matching, donor type, delayed graft function, acute rejection, posttransplant lymphoproliferative disease (PTLD) and immunosuppressive treatment (type of induction) were assessed.
Acute rejection was considered in the presence of serum creatinine elevation. Only biopsy-proven acute rejections were considered. Acute rejection was defined according to the Banff classification .
Severe bacterial infections
Diagnosis of severe bacterial infections required bacterial infection-related hospitalization.
Diagnosis of CMV disease required the presence of viral replication and clinical symptoms.
Death-censored graft losses occurring after 6 months were recorded.
These analyses were performed without knowledge of baseline characteristics.
Two physicians independent of the study were responsible for diagnostic ascertainment.
Arithmetic mean was calculated and expressed as ± SD.
The patients were divided into two groups. Patients with at least one positive EBV PCR during the first year of follow-up were designed as EBV(+) whereas the others were designed as EBV(−).
Logistic regression was used to determine which factors were associated with EBV viremia.
Using log-rank tests on Kaplan–Meier nonparametric estimates of the survival without different infectious outcomes and graft loss distribution, we selected variables with a p-value lower than, or equal to, 0.10. The selected variables were included into a Cox proportional hazards model, and a backward stepwise selection process was performed, this time at a classical α = 0.05. Results were expressed as hazard ratio (HR) and 95 % confidence interval (CI), with a p-value testing the null hypothesis: HR = 1. Therefore when p value is less than 0.05, HR is significantly different from 1, either greater than 1 (i.e. risk of event is increased) or less than 1 (i.e. risk of event is decreased). Assumptions of Cox models (log-linearity, proportionality of risk in time) were met in this analysis.
Clinical characteristics of the study population are depicted in Tables 1 and 2.
Table 2. Characteristics and outcomes of the study patients according to EBV viral load
Persistent < 6 months
Persistent > 6 months
<0.05 *persistent EBV > 6 months vs. other groups.
A total of 357 patients were EBV seropositive before transplantation. Most of them (311, 87.1%) received an EBV seropositive kidney. Of 26 EBV seronegative patients, 18 (4.7%) received a seropositive kidney.
Incidence of viremia
A total of 6371 samples were expected during the first year and 6198 were collected (97.3%).
A total of 155 of the 383 patients (40%) had at least one positive viremia during the first year posttransplant. The median value of EBV viremia was 0.95 × 103/mL (range: 0.87 × 102–2.2 × 106). The median time to viremia was day 31 posttransplant (14–329). Most of first positive viremia occurred in the first 3 months posttransplant (< 1 month, 49%, months 1–3, 35%, months 3–6, 5%, > 6 months, 9%) (Figure 1A).
Seventy patients (18.3%) had persistent viremia and 23 (6%) had positive viremia for more than 6 months. Eighty-one patients (52%) had EBV viremia < 103/mL, 59 (38%) had EBV viremia between 103 and 104/mL and 15 (10%) had EBV viremia > 104/mL. Viral load was highly predictive of persistent viremia after the first year posttransplant (19%, 38% and 92%, respectively; p<0.0001).
Factors associated with EBV infection
Patients with and without viremia did not differ for age, gender, underlying renal disease, type of transplant and CMV serostatus.
BKV viremia was not more frequent in patients with and without EBV infection (12.2% vs. 11.4%; p = 0.926).
EBV D+/R− patients were more likely to develop EBV infection (12/18 (67%) versus 143/365 (39%), p = 0.02).
Patients having received ATG were more likely to develop EBV viremia (54% vs. 35%; p<0.001). Viral load was higher (4.54 ± 5.33 vs. 3.67 ± 4.4 log; p<0.0001) in ATG-treated patients. Time to first positive viremia was not different, but patients who had received ATG were more likely to have persistent viremia (27 vs. 13%, p = 0.001).
By contrast, both Tacrolimus and MMF dosages were similar as well as mean Tacrolimus though levels and MPA AUC. The incidence of acute rejection did not differ in patients without and without viremia.
EBV PCR and changes in immunosuppression
Changes in immunosuppressive drugs were performed in 15 patients having the previously described criteria. Six of them (40%) were EBV seronegative before transplant and had received a seropositive organ.
MMF was withdrawn in nine patients. MMF dose and Tacrolimus dose reduction were effective in reducing EBV PCR under the predefined thresholds in 10 patients. Tacrolimus was converted to Sirolimus in two other patients because of concomitant skin cancer. The switch had no effect on EBV viral load.
Three additional patients received Rituximab (375 mg/m2) because of persistent high viral load. Clearance of the virus was obtained but recurrence occurred in all cases as CD19+ B cells reappeared. One patient received a second infusion of Rituximab with the same evolution.
No case of PTLD occurred during the study follow-up.
EBV reactivation and infections
A total of 74 (19.3%) patients experienced at least one episode of severe acute bacterial infection. Severe acute bacterial infections were not significantly more frequent in EBV(+)-patients (21.3% vs. 18%; p = 0.421).
A total of 31 (8.1%) patients experienced at least one episode of CMV disease. CMV disease was slightly more frequent in subjects with EBV reactivation (10.3% vs. 6.6%; p = 0.187).
Opportunistic infections (other than CMV)
A total of 27 (7%) patients experienced at least one opportunistic infection (Zoster infections, 9; Pneumocystis Jirovicii, 6; Aspergillosis, 1; Herpes infections, 5; Legionellosis, 3; Tuberculosis, 1; Nocardiosis, 2). Opportunistic infections were more frequent in EBV(+)-patients (11.6% vs. 3.9%; p = 0.008). In multivariate analysis, age (HR, 2.24; 95% CI, 1.41–4.27) and EBV reactivation (HR, 3.03; 95% CI, 1.72–8.29; p = 0.01) increased the risk of opportunistic infections.
Because most of EBV viremia (91%) occurred before 6 months posttransplant, we analyzed whether EBV replication in the first 6 months after transplant may also affect graft survival (Figure 1B). Fourteen patients lost their graft during the first 6 months posttransplant. Mean follow-up after the first 6 months posttransplant was 67 ± 24 months. Among 369 patients, graft loss occurred in 27/133 (20.3%) with at least one positive EBV PCR and in 25/236 (10.6%) without EBV viremia. Causes of graft loss are depicted in Table 3.
Table 3. Causes of graft loss in patients with and without EBV reactivation
A positive EBV PCR during the first year posttransplant increased the risk of graft loss (HR, 1.84; 95% CI, 1.11–6.22; p = 0.033). Acute rejection (p = 0.014), delayed graft function (p = 0.009), 1-year creatinine clearance < 50 mL/min (p<0.001), pretransplant diabetes (p = 0.045) and preformed HLA antibodies (p = 0.09) were also associated with graft loss.
The effect of EBV infection remained significant even after adjustment for confounding factors (HR, 1.54; 95% CI, 1.05–5.44; p = 0.043). Acute rejection (HR, 1.94; 95% CI, 1.16–3.59; p = 0.018), delayed graft function (HR, 2.61; 95% CI, 1.011–5.98; p = 0.021) and 6 months creatinine clearance < 50 mL/min (HR, 4.77; 95% CI, 1.76–12.91; p = 0.001) were also associated with graft loss (Table 4).
Table 4. Cox model: hazard ratio (HR) of death-censored graft loss and 95% confidence intervals (CI)
Delayed graft function
≥ 50 mL/min
< 50 mL/min
EBV infection was not associated with death.
Data are very scarce regarding the natural course of EBV infection after solid-organ transplantation. Our study demonstrates that EBV reactivation is a frequent event in the first year posttransplant affecting up to 40% of the patients. Even when most EBV reactivation occurs at low levels and are transient, some patients exhibit asymptomatic persistent EBV viremia with high viral load. However, EBV infection was associated with the occurrence of opportunistic infections and a higher risk of graft loss.
Seronegative recipients of a seropositive kidney carry a higher risk of viremia . In this study, 60% of D+/R− patients developed EBV infection. Of note, no PTLD was observed in this high-risk population. Systematic use of antiviral pharmacologic prevention, passive immunization with intravenous immunoglobulins, close monitoring and prompt intervention to reduce immunosuppression in patients with persistent positive EBV viremia may have contribute to this result. Few data support the relevance of passive immunization to prevent EBV-induced proliferation. Nadal et al.  found that the infusion of human immunoglobulin. Further studies should examine whether intravenous immunoglobulins may reduce the risk of PTLD in EBV D+/R− patients. Besides EBV sero-mismatch, the use of ATG was also clearly associated with EBV infection. Viral load was higher and the duration of viremia was longer in patients having received ATG. All together these data suggest that ATG should be used with caution in EBV D+/R− patients.
Reduction in immunosuppression was effective to reduce EBV viral load in the majority of patients with persistent viremia. However, there is no consensus regarding the way to minimize immunosuppressive drugs. Although MMF has been reported to reduce the incidence of EBV infection, MMF dose reduction was effective in most patients to reduce EBV viral load . Preemptive administration of Rituximab is widely used in the setting of HSCT and may have some benefits [13, 14]. Nevertheless, we observed EBV reactivation in all Rituximab-treated patients as CD19+ B cells reappeared. We believe that Rituximab could not be recommended in solid-organ transplant patients with persistent high EBV load.
EBV reactivation is likely to reflect overimmunosuppression. Indeed, EBV infection is more frequent in ATG-treated patients and is reversible after a significant reduction in immunosuppression. Moreover, our data suggest that persistent EBV infection is associated with the subsequent occurrence of opportunistic infections. There are few markers of overimmunosuppression, but EBV monitoring may help to recognize such patients and could allow prompt adaptation of immunosuppressive drugs.
We observed a higher rate of graft loss in patients having had a positive EBV PCR during the first 6 months posttransplant. Li et al.  also reported worse renal function in pediatric transplant recipients with subclinical viremia. Nevertheless, CMV and EBV viremia were analyzed together and a specific relationship between EBV infection and graft loss was not described. More recently, Smith et al.  reported a significantly greater decline in GFR in transplant patients with subclinical EBV viremia. Moreover, EBV viremia was associated with histologic evidence of chronic allograft injury . The mechanisms of EBV-related allograft injury are not well established. Whereas a direct cytopathogenic effect is unlikely, indirect inflammatory effects may be involved . Specific anti-EBV exhausted CD8+ CD28− T cells may play a role in these mechanisms . Nevertheless, some bias may also explain this association. EBV reactivation may identify a high-risk population carrying a higher risk of graft loss and receiving a greater immunosuppression. Indeed EBV infection was found to be more frequent in ATG-treated patients. However, the association between EBV viremia and graft loss seems to independent of major pretransplant confounding factors. Furthermore, the relationship between EBV and graft loss is likely influenced by multiple sequential events that are not attributable directly to the virus, but rather to the myriad of clinical decisions that cascade from the detection of the virus. Especially, reduction in immunosuppressive drugs in patients with high EBV viremia may have influence graft loss. Nevertheless, we did not observe any acute rejection after drug withdrawal and/or minimization in this population. Finally, BK virus nephropathy was very rare in the study population and cannot explain the association between EBV reactivation and graft loss. Nevertheless, we acknowledge that the clinical scenario leading to EBV viremia, likely confounds this relationship. Further prospective studies are needed to conclude on a possible deleterious effect of EBV reactivation on graft survival.
Our study shows that EBV infection is a frequent infectious complication following kidney transplantation. EBV reactivation is associated with the occurrence of opportunistic infections. Reduction in immunosuppression is effective in patients with persistent high viral load. EBV viremia seems to be associated with graft loss. Nevertheless, this relationship could depend on numerous confounding factors and clearly deserves prospective analysis.
The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.