Introduction

Post-transplant lymphoproliferative disorder (PTLD) is a severe complication of solid organ transplantation that develops due to uncontrolled proliferation of lymphocytes within the context of post-transplant immunosuppression. In the vast majority of cases, the proliferating cells are B lymphocytes of recipient origin and, most frequently, proliferation is induced by Epstein–Barr virus (EBV) infection [1].

The revised World Health organization (WHO) classification of PTLD considers PTLD histopathology independently of time of onset since transplantation and categorizes the continuum of disease into plasma cell hyperplasia/early lesion PTLD, polymorphic PTLD (polyclonal or monoclonal), and monomorphic PTLD [2, 3]. The term “early lesions PTLD” refers to localized early lesions that show the first morphological changes in the spectrum of PTLDs. A distinctive morphologic feature of early lesions PTLD is exuberant lymphoid proliferation with preservation of normal tissue architecture despite the presence of a mass effect, such as enlarged lymph nodes or tonsils and/or adenoid hypertrophy. This is in contrast with the other subtypes of non-early PTLDs (polymorphic, monomorphic) in which the normal architecture of the involved tissue is partially or completely destroyed [3, 4].

The incidence of the disease is higher in children who experience primary EBV infection post-transplantation, whereas the influence of the cumulative burden of immunosuppression is less evident [5]. Past studies have indicated that the incidence of PTLD after kidney transplantation is lower than that occurring after intestinal, heart, lung, and liver transplantation, but the increasing number of young renal transplant recipients may account for a higher rate of EBV-seronegative individuals receiving transplantation and, thus, for an increased prevalence of EBV-associated PTLD [6].

The aim of our study was to describe the natural history of EBV infection and its possible evolution toward PTLD in a large cohort of pediatric renal transplant recipients.

Methods

Patient population and data collection

This was a retrospective cohort study in which all patients who received a renal transplant in two large transplantation centers in Italy [Bambino Gesù Children’s Hospital (Center A); University of Padua, (Center B)] between January 2000 and December 2013 were eligible for entry. Patients were enrolled in the study if the results of at least two real-time PCR assays for EBV DNA per year were available following transplantation. Data on the results of the real-time EBV PCR assay and PTLD were retrospectively collected from clinical records.

EBV DNA was monitored in whole blood using the Artus real-time EBV TM PCR kit (Qiagen, Hamburg, Germany) (Center A) and an in house-developed real-time quantitative (q)PCR assay (Center B) [7] that targets the conserved EBV nuclear antigen 1 (EBNA1 gene) region and the highly conserved BLLF1 gene, respectively. Amplification cycles were run in a qPCR assay using ABI Prism 7900 HT Fast Real-time instruments (Applied Biosystems, Foster City, CA). The viral load was calculated from the slope and intercept of the standard curve, and results were expressed in copies per microliter. Good agreement between the two methods was demonstrated by inter-laboratory evaluations. Test proficiency was monitored by regular external quality assessments using Quality Control for Molecular Diagnostics (QCMD) programs. The analytical sensitivity (limit of detection) of both methods was 100 copies/ml.

The clinical cut-off for a positive PCR result for EBV DNA was arbitrarily set at 3000 copies/μl. A patient who tested positive for EBV DNA was subsequently evaluated at regular intervals for possible development of PTLD by means of abdominal scan (every 6 months) and chest X-ray (at least annually).

Patients were treated with induction treatment with basiliximab (two shots, on day 0 and day 4) or thymoglobulin (for 5–7 days post-transplantation, with the dosage adjusted to maintain a lymphocyte count of <200/mmc). As maintenance treatment, patients received calcineurin inhibitors (tacrolimus or cyclosporine) and steroids. Most patients were also treated with mycophenolate, or more rarely with azathioprine or mTOR inhibitors. Valganciclovir was provided as prophylaxis to all patients at high risk of developing cytomegalovirus (CMV) infection (CMV-positive donor vs. CMV-negative recipients) for the first 3 months after transplantation.

We categorized PTLDs according to the WHO subtype classification as localized early lesions (early lesions PTLD) or non-early lesions PTLD (polymorphic, monomorphic B-cell and other subtypes) that may disseminate independently from the time of occurrence since transplantation [2, 3].

Statistical analysis

Primary study outcomes were: (1) positive PCR test result for EBV following transplantation, defined as >3000 copies/μl, (2) onset of non-early lesions PLTD, and (3) onset of all PTLDs. Categorical variables were summarized using absolute frequencies and percentages, and continuous variables by the mean or median and range, as appropriate. To determine statistical differences between groups, we used the Chi-square test or Fischer’s exact test for categorical variables and the t test or Mann–Whitney test for continuous variables.

Three multivariate Cox proportional hazard models were developed to assess independent predictors of the three primary study outcomes. Variables for which the p value was ≤0.20 in univariate analysis were included in the multivariate models (likelihood ratio test p < 0.05). Proportional hazards assumptions were respected, and C-concordance statistics were used to measure the model performance.

All statistical analyses were performed using STATA, Statistical Software: Release 13 (StataCorp LP, College Station, TX).

Results

In the period considered, 318 patients underwent renal transplantation in the two participating centers. Of these, 14 patients were excluded from the analysis due to incomplete clinical data.

The characteristics of the 304 patients included in the study (201 from Center A, transplanted between January 2000 and December 2013, and 103 from Center B, transplanted between January 2004 and December 2013) are shown in Table 1. Of these 304 patients, 31 were older than 18 years at transplantation. Median duration of follow-up was significantly longer in Center A than in Center B, and significantly more patients were treated with antiviral prophylaxis in Center B than in Center A. Also, induction therapy with thymoglobulin and maintenance therapy with tacrolimus were more often used in Center B than Center A. All other patients’ characteristics were similar between the two participating centers (Table 1).

Table 1 Characteristics of patients included in the study, by center
Full size table

Data on donor EBV status before transplantation were available for 152 patients (50%), of whom 121 were positive (79.6%) and 31 were negative (20.4%) (Table 1).

At time of transplantation, 103 patients were seronegative for EBV (33.9%) (Fig. 1). Following the graft, 50 of these (48.5%) developed a first EBV infection after a median time of 8.4 months (range 7.0 days to 11.4 years). Upon detection of EBV infection, immunosuppression was reduced in all cases (withdrawal of mycophenolate in most cases) (data not shown). The PCR EBV assay results showed a reversion to <3000 copies/μl in 19 of these patients (38%) after a median time of 22.1 months (13 days to 84 months). Median maximum viral load was 337,443 (range 4214–99,470,000) copies/μl in those who maintained throughout the period of observation a positive viral load and 19,447 (range 1635–8,034,000) copies/μl in those who subsequently became negative (p < 0.001).

Fig. 1
figure 1

Distribution of patients included in the study by Epstein–Barr virus (EBV) status and onset of non-early lesions post-transplant lymphoproliferative disorder (PTLD)

Full size image

Among the 201 patients who were seropositive at time of transplantation, only 40 (19.9%) presented a reactivation of EBV, based on PCR assay values that exceeded the pre-determined threshold of 3000 copies/μl (Fig. 1), after a median time of 4.48 months (range 1 day to 9.3 years); in 19 (47.5%) of these patients EBV DNA decreased to <3000 copies/μl after a median time of 10 months. Median maximum viral load was 119,840 (range 4214–9,578,000) copies/μl in those patients who maintained a positive viral load and 6314 (range 1635–224,000) copies/μl in those who became negative (p = 0.001).

All EBV-seropositive patients (EBV ≥3000 copies/μL) were monitored over time using EBV PCR assays. determinations. The frequency of EBV seropositivity in the group of children who reached a high viral load (≥59,909.5 copies/μl; mean determinations per year upon detection of high viral load: 1.4) was similar to that the group who did not (<59,909.5 copies/μl; mean determinations per year: 1.6) (p = 0.166).

Very few patients (only 5) were treated with rituximab; these patients received one or two doses (375 mg/sq.m) with the intent to lower the EBV viral load in the blood. In all patients, the EBV viral load dropped very quickly after the rituximab treatment, becoming undetectable, but after a short time (2–6 months) it rose again in four patients to levels comparable to those detected before treatment, with only one patient having a persistent reduction of EBV viral load. None of these five patients developed PTLD after a follow-up of 3, 24, 28, 50 and 60 months, respectively. Given the very small number of patients in this group, no conclusions the effect of rituximab can be drawn.

Five patients had a diagnosis of early lesions PTLD and were treated with rituximab and a reduction of immunosuppression. These patients are still being closely monitored, but after a long follow-up (5–9 years) none of them has yet developed an aggressive lymphoma (Table 2).

Table 2 Description of early lesions post-transplant lymphoproliferative disorder occurring in patients enrolled in the study
Full size table

Non-early lesions PTLD was diagnosed in ten patients (3.3% of the whole cohort). In all cases, it occurred in patients in whom a positive EBV viral load was detected at some stage of their follow-up, even though one-half of the patients became negative at some stage of their follow-up and only one case occurred in this latter group of patients (Fig. 1).

Non-early lesions PTLD occurred in three of the 201 patients (1.5%) who were seropositive at transplantation and in seven of the 103 patients (6.8%) who were seronegative at transplantation (p = 0.035) (Table 3). In all ten cases, in situ hybridization demonstrated EBV positivity in tissues.

Table 3 Description of post-transplant lymphoproliferative disorder occurring in patients enrolled in the study
Full size table

Patients who were diagnosed with non-early lesions PTLD were referred to the hematology unit of the respective hospital. All of these patients were treated with a combination of chemotherapy and rituximab. The majority of patients recovered and kidney function was maintained. However, one patient died (10.0%) and two patients lost kidney function (20.0%).

The results of the multivariate Cox proportional hazard model investigating predictors of a positive PCR assay result for EBV following transplantation showed that seropositivity at transplantation and age of transplant recipient of >14 years were independently and significantly associated with a lower risk of EBV infection (Table 4).

Table 4 Predictors of positive PCR for Epstein–Barr virus. Multivariate Cox proportional hazard model
Full size table

Results of the multivariate models investigating predictors of non-early PLTDs and all PTLDs showed that having a maximum peak of EBV viral load above the median value observed in the whole cohort (59,909.5 copies/μl) was independently and significantly associated with higher risk of PTLD onset. In both models, center was also independently and significantly associated with the outcomes, even though this finding has no clear explanation and should be further investigated, while patient’s age >14 years was a risk factor for non-early PTLDs (Tables 5, 6).

Table 5 Predictors of non-early lesion post-transplant lymphoproliferative disorder onset. Multivariate Cox proportional hazard model
Full size table
Table 6 Predictors of all post-transplant lymphoproliferative disorder onset. Multivariate Cox proportional hazard model
Full size table

Discussion

Immunosuppressed patients may fail to develop an effective immune response against EBV. This may lead to a persistent infection, which may be responsible for the development of a PTLD, a well-known life-threatening complication of solid organ transplantation. Its prevalence depends on the type of transplanted organ, age of the patient, type and intensity of immunosuppression, and pre-transplantation EBV status [8, 9]. The highest incidence of PTLD occurs among pediatric recipients of heart, lung, and intestine (5–10%), while the incidence in pediatric liver and kidney recipients is 2.4 and 2–4%, respectively [1]. Intensity of immunosuppression may account for these differences [10]. The majority of cases of PTLD occur early, even within the first months after organ transplantation, i.e., during peak immunosuppression [10]. The risk of PTLD is higher in EBV-seronegative children who receive an organ from an EBV-seropositive donor [11].

Among the 304 patients enrolled in our study, the overall incidence of PTLD was consistent with that reported in the general literature [12] and, as expected, a naive status for EBV before transplantation was a high-risk factor for developing a steadily persistent high PCR EBV viral load. In total, 103 patients were seronegative at transplantation; of these 48% presented a first EBV infection, and in 62% of this latter group the viral load remained persistently detectable. These latter patients were therefore at high risk for PTLD. It is also likely that the other seronegative patients will experience a first infection at some stage of their life. Hopefully, it will occur when immunosuppression is less aggressive, which might provide them with some advantages in controlling the infection.

With the introduction of serial viral load monitoring, it has become possible to identify a population of children who are at major risk of developing PTLD [13].

In our series all patients who developed a PTLD had a steady increase of EBV viremia. In our statistical model, a maximum peak of EBV viral load above the median value observed in the whole cohort (59,909.5 copies/μl) was significantly and independently associated with the development of a PTLD. This association remains an important point of debate in the literature: correlation with the maximum peak of EBV viral load has been demonstrated for various conditions, but it is not as evident for kidney transplant recipients [14,15,16,17].

Exposure to EBV begins early in life, with approximately 50% of children in developed countries becoming seropositive by 5 years of age [18]. The age of children at transplantation is therefore an indirect risk factor due to this higher likelihood of being seronegative in early infancy. Consequently, transplanting young children carries a higher risk of PTLD development.

In our series, patient’s age above 14 years also correlated with the development of non-early lesions PTLD. This finding deserves comment. Our study population was a pediatric cohort, and all patients had been placed on the waiting list for a transplant before the age of 18 years. However, transplantation may have occurred after 18 years of age and, as expected, almost all of them (97.8%) received a transplant before 22 years of age. Younger patients carry a higher probability of developing EBV infection and thus are at high risk for PTLD, but it is not clear why older patients should be at higher risk for non-early lesions PTLD. One possible explanation may be related to the worse adherence to treatment of adolescents. It has been demonstrated that patients in this age group are more at risk of rejection due to poor compliance [19,20,21], and it is possible that some of these patients may have received steroid pulses for rejection treatment which may have added to overall immunosuppression, inducing a higher risk for the development of non-early lesions PTLD.

Some authors have suggested a possible beneficial effect of valganciclovir in the treatment of EBV infection [22, 23], although the benefit of this drug has not been clearly demonstrated, as it acts only on replicating viruses, while in PTLD most of the viral genome is in a non-replicative phase [1]. In our series, CMV prophylaxis was provided for 3 months only to high-risk patients (positive donor/negative recipient). However, prophylaxis did not modify the risk for PTLD development in any of the statistical modeling.

Recent data from the literature [24] suggest that a possible alternative treatment of EV infection may be EBV-specific cytotoxic T lymphocytes. In our patient series, however, this approach was used only in exceptional cases, and there is therefore insufficient data to draw any conclusion on this efficacy of this treatment.

We must acknowledge several limitations to this study. It is retrospective, and we have included patients with different durations of follow-up. Some patients have a short follow-up and might have achieved a spontaneous recovery of EBV infection or recovered from the infection after immunosuppression reduction. Detection of a positive EBV PCR assay result often prompted physicians to increase the frequency of tests, thus EBV-positive patients may have had more determinations than the others.

There were a number of differences between the two centers: Center B used induction therapy with thymoglobulin and maintenance therapy with tacrolimus more often than Center A. Both induction and maintenance treatments were included in the multivariate model which investigated predictors of non-early lesions PTLD onset. Although thymoglobulin induction was significantly associated with EBV viremia, none of these covariates were significantly and independently associated with PTLD development. Thus, these variables do not explain the center effect in this model. This is an important issue that deserves further investigation in the future.

The most difficult issue is to predict which patients will actually develop PTLD, as not all patients with a persistent infection will develop lymphoma [25]. In our cohort, EBV maximum viral load significantly correlated with PTLD development, and we should question if, in such a context, a sharp reduction of immunosuppression would have been appropriate [26], even at the risk of losing the kidney. In patients in whom PTLD occurred early after the detection of EBV viremia, any preventative measures would likely have been ineffective, while in those in whom the diagnosis of PTLD occurred several months or years after EBV detection, we would have had time initiate more effective treatment. In all cases immunosuppression was moderately reduced with little or no effect. However, it is important to consider that even among our high-risk patients (those with a persistently high EBV viral load) only a minority developed PTLD. Moreover, following the development of non-early lesions PTLD, the majority of patients recovered, although one of the ten patients who developed lymphoma died and two additional patients, who had recovered from PTLD, lost their graft. Stopping immunosuppression in all patients at risk is therefore inappropriate.

The definition of a reliable, more specific, marker of risk is essential to personalize the clinical approach to the single patient [27, 28].