Based on pathophysiology and known risk factors (Table 2), several strategies of prevention are described and aim to decrease incidence of PTLD. Current results, however, mainly allow earlier diagnosis, leading to decreased mortality and morbidity. Clear impact on incidence remains to be demonstrated by well-designed prospective studies, when ethically acceptable.

One corner stone of prevention relies on EBV load monitoring by quantitative polymerase chain reaction (PCR), as high viral load is clearly associated with PTLD. Numerous studies report the characteristics of EBV load after transplantation, measured in peripheral blood mononuclear cells (PBMC), plasma or whole blood (reviewed in Gulley and Tang 2010). Viral load is usually higher in patients under immunosuppression as compared to healthy controls, while the highest viral load is described after transplantation in subjects with acute primary infection, disease reactivation, and PTLD. In pediatric solid organ recipients, viral load >5,000/μg PBMC DNA was a diagnostic criterion for PTLD with a positive predictive value (PPV) of 89 % and negative predictive value (NPV) of 100 %, while both PPV and NPV were 100 % for an EBV load >1,000/100μl plasma (Wagner et al. 2001). High viral load was correlated with PTLD and normalized after treatment. Similar differences were described in other types of transplantation and led to preemptive assays. In Houston (Lee et al. 2005), after 2001, IS was decreased in liver-transplanted children after two EBV loads >4,000/μg PBMC DNA (previously described as giving PPV for PTLD of 56 % and NPV of 100 %). Before and after 2001, PTLD incidence was, respectively, 16 % (within 30 patients) and 2 % (out of 43 patients, 29 EBV-naïve patients before transplantation, p < 0.05). Preemptive IS alleviation (tacrolimus tapering, corticoids arrest) induced lowering of viral load in all and led to rejection in 1/11 patient. Another retrospective comparison of children after liver transplantation with or without EBV load monitoring (2001–2005 and 1993–1997), with preemptive IS tapering in case of high viral load, showed no difference in PTLD incidence (7.5 and 10.9 %, respectively) (Kerkar et al. 2010). PTLD occurring in the 2001–2005 era were, however, less advanced (polymorphic versus monomorphic, p = 0.03), with no mortality as compared to 3/10 deaths before. One problem is that each laboratory has to define its own cutoff for “high viral load” and risk assessment. The major drawback for using EBV load, as a predictive marker of PTLD, is the intermediate specificity and PPV that could lead to needless IS tapering and risk of graft rejection. The same group showed, always in liver-transplanted children, that reducing IS preemptively for EBV active infection led to graft rejection in 27.9 % of cases (with 1 graft loss), while IS tapering or cessation in case of PTLD was complicated by rejection in 83.3 % of patients, without graft loss (Weiner et al. 2012). Both the risk of PTLD and the risk of rejection after IS reduction were higher in case of past history of rejection. Finally, chronic high EBV load was also described as risk factor for late PTLD. EBV load >500/10⁵ PBMC was found in 12/18 EBV-seronegative children who underwent primary infection after liver transplantation, as compared to 0/13 seropositive children before transplant (D’Antiga et al. 2007). Sustained viral detection, meaning positive for more than 6 months, was only seen in those EBV-naïve patients (14/18), and three of them developed late PTLD (median time 47 months, range 15–121). In 71 cardiac-transplanted children, chronic high viral load was defined as EBV load >16,000/mL whole blood in >50 % of samples over at least 6 months (Bingler et al. 2008) and was found in 20 of them (8 with prior PTLD, 7 with prior symptomatic EBV). Late PTLD occurred in 9 of them (45 %) compared to 2 in the 51 controls (4 %, p < 0.001). Multivariate analysis showed independently increased odds ratio (ORs) to develop late PTLD for chronic high viral load (OR = 12.4, 95 % CI 2.1–74.4) and prior history of PTLD (OR = 10.7, 95 % CI 1.9–60.6). Therefore, experts recommend serial monitoring of EBV load in EBV-naïve patients undergoing primary EBV infection after transplantation, to assess the risk to develop early or late PTLD (Gulley and Tang 2010; KDIGO 2009).

Higher viral load is also described in patients developing PTLD after bone marrow/stem cell transplantation, and same recommendations are applied. Additional factors are important to determine high-risk patient as EBV status before transplantation does not influence the occurrence of PTLD in this type of graft. In the large series from Landgren et al. (26.901 patients after allogeneic hematopoietic cell transplantation, 127 PTLD, 105 in the first-year post-transplant), the cumulative incidence of PTLD was 0.2 % in patients with no major risk factors (T-cell-depleted graft, HLA-mismatched graft, ATG, age >50 years), 1.1 % if 1 risk factor, 3.6 % if 2, and 8.1 % if 3 or more (Landgren et al. 2009). They underlined the theoretical interest to prospectively monitor EBV infection in this sub-group of high-risk patients, aiming to early treatment intervention. Additionally, preemptive strategy by low-dose rituximab in the conditioning regimen was proposed (Savani et al. 2009; Bacigalupo et al. 2010). This was based on one study showing no PTLD in 27 adult recipients of intestinal and multivisceral transplantation whom induction IS included one single dose of rituximab 150 mg/m² (Vianna et al. 2008). Moreover, 14 patients having developed EBV reactivation after allogeneic stem cell transplantation received preemptive rituximab therapy (standard doses) and did not develop PTLD (Kindwall-Keller et al. 2009). Finally, PTLD did not occur in 38 patients after rituximab-containing conditioning regimen or prior rituximab, even in 8 of them with 3 high risk factors (Savani et al. 2009). These preliminary data were confirmed by a retrospective study comparing 55 patients preemptively treated with 200 mg/m² rituximab at day+5 of conditioning to 68 who did not receive prophylaxis (Dominietto et al. 2012). Both groups received ATG. EBV viremia was found in 56 % of rituximab-treated patients as compared to 85 % of the controls (p = 0.0004). The maximal median viral load was also lower (91 vs 1321/10⁵ cells, p = 0.003), and the risk to have EBV >1,000/10⁵ cells, which is a known risk factor for PTLD, was found in 14 versus 49 % of patients (p = 0.0001). No PTLD were observed in the treated group as compared to 2 in the controls. No side effects were reported although leukocyte and lymphocyte counts were lower at day+50 and +100 after rituximab. Acute GVHD was reduced after preemptive treatment (20 vs 38 %, p = 0.02). Finally, in 70 pediatric stem cell transplant recipients, preemptive strategy was prospectively applied in case of EBV high load (>40,000/mL whole blood) (Worth et al. 2011). When CD3 count was <300 × 10⁶/L, classical doses rituximab was given (13/70), if not, IS was decreased (6/70). Incidence of PTLD was decreased as compared to historical controls (1.4 vs 21.7 %, p = 0.003), even more in patients with high viral load (2.7 vs 62.5 %, p < 0.0001). There were few side effects of rituximab, the main being longer time for B-cell recovery (27.6 vs 8.9 months). All PTLD occurred in patients with low CD3 count. The possibility to further improve this prevention in high-risk patients by using sirolimus as GVHD prophylaxis was proposed but should be demonstrated (Reddy et al. 2011; Cen and Longnecker 2011).

Monitoring of EBV-specific immunity in correlation with EBV load is proposed to better differentiate patients with asymptomatic high EBV load from those at risk of PTLD. This strategy was efficient in children after liver transplantation for which an inadequate ratio between high viral load and low specific EBV T lymphocytes (EBV–CTL) at the time of EBV primary infection was 100 % predictive of following PTLD (Smets et al. 2002). EBV–CTL were detected by the enzyme-linked immunospot assay (Elispot) through their specific IFNγ secretion after stimulation with an autologous EBV + cell line. No PTLD occurred when EBV–CTL were >2/mm³ of blood, independently of the viral load. The same profile was described in 14 healthy adults followed prospectively (Vogl et al. 2012). EBV reactivation was related to lower Elispot-detected EBV-TL in 8/12 subjects. However, the Elispot monitoring is time-consuming, requires the production of autologous EBV cell line for all the patients, and is therefore difficult to implement for routine follow-up. An alternative would be the tetramer technique that allows easy detection of EBV-peptide-specific HLA-restricted lymphocytes by flow cytometry. No correlation was found with viral reactivation in healthy carriers (Vogl et al. 2012). In 31 prospectively followed liver-transplanted children, level of tetramer EBV-specific CD8 T cells were the same in 11 patients with chronic high viral load and 20 controls (Gotoh et al. 2010). Genetic profile of EBV in these 11 children showed type 0 latency (EBER, BART, and LMP2), suggesting that high viral load can be related to viral immune escaping. No early or late PTLD occurred during the study (follow-up 0.8–7.4 years). In a small series, tetramer EBV-specific CD8 T cells were lower in 10 PTLD patients and 6 healthy controls as compared to the 2 transplanted patients without complication, with only 30 % of them being functional (Sebelin-Wulf et al. 2007). Moreover, the absolute CD4 T-cell count was also lower (336 ± 161/μL vs 1008 ± 424/μL, p = 0.00001). CD4 T cells <230/μL were associated with high viral load >1,000/μg DNA. In 307 patients after allogeneic hematopoietic cell transplant, EBV–CTL at D28 post-transplantation, assessed by flow cytometry after specific peptide stimulation, did not allow us to predict the 25 PTLD (Hoegh-Petersen et al. 2011) although they tend to be higher in patients without complication. In pediatric liver-transplanted children, the PPV of high viral load for PTLD was increased from 29 to 57 % if associated with low IFNγ (A/A) polymorphism (Lee et al. 2006b). No differences were found for polymorphism in TGF-β, TNF-α, and interleukin 2, 6, and 10. Children had 100 % risk of having high viral load in case of low release of ATP by CD3+ cells, stimulated for 24 h with phytohemagglutinin (<125 ng/mL). Following IS tapering, increase in ATP release and decreased viral load were observed (Lee 2006a). Finally, the potential implication of EBV-specific natural killer (NK) cells was recently proposed. Important NK response was described in adults followed prospectively after IM, with particular expansion of a CD56 highly positive sub-population (CD56^bright) (Williams et al. 2005). Higher NK cells were correlated with lower viral load. In a retrospective review of adult and pediatric solid organ transplant recipients, polymorphism of Fc-γ receptor 3A (FCGR3A) and killer cell immunoglobulin-like receptor (KIR) genes, involved in NK cell function, was shown to influence significantly survival after PTLD diagnosis but were not risk factors for PTLD (Stem et al. 2010). More recently, both NK cell phenotype and function were different in pediatric patients developing PTLD after thoracic transplantation and those without complication (Wiesmayr et al. 2012). Patients with chronic high viral load but no PTLD displayed intermediate values. The interest to monitor NK cells in transplanted patients should, however, be confirmed.

Antiviral drugs were also proposed as preemptive therapy. Although they are not active on latent EBV cycle involved in PTLD, they can impair or reduce lytic cycle, limiting intercellular EBV transmission. Use of ganciclovir in adults and children after renal transplantation was associated, especially within the first-year post-transplant, with up to 83 % reduction in the risk of PTLD (Funch et al. 2005). One month of treatment in this period gave a 38 % reduced risk. Effect of acyclovir was less marked. Forty-seven children with positive PCR for EBV after liver transplant were treated with valganciclovir but no IS reduction (Hierro et al. 2008). Symptoms were present in 28 % of them. After a short treatment (1 month), negative PCR was observed in 34 % of patients, with 82 % relapse after valganciclovir arrest. After longer therapy (8 months), PCR negativity was found in 47.6 % children and was maintained in 60 % of them off treatment. PTLD was suspected but not confirmed in one child (2.1 %).

Early diagnosis and treatment are mandatory to avoid PTLD-related morbidity and mortality, the latter having been often reported between 30 and 60 % (Bakker et al. 2007). High index of suspicion is necessary due to non-specificity of initial symptoms. Unexplained fever, adenopathies, hepatosplenomegaly, sepsis like syndrome, and extra-nodal involvement (graft, gastrointestinal tract, and sinonasal cavity) are common presentations. Graft localization is more often reported in early PTLD. The interest of fluorodeoxyglucose positron emission tomography for disease staging and follow-up after treatment is described, but its role in precocious diagnosis remains to be demonstrated. Unrecognized PTLD can rapidly lead to graft and multiorgan severe dysfunction and death.

In early PTLD after solid organ transplantation, reducing IS is the first step of treatment, aiming to allow restoration of specific immunity and control of the EBV-infected proliferative cells, with a response rate reported up to 75 % (Heslop 2009). While it could be an option in late PTLD after bone marrow/stem cell transplantation, IS tapering is usually needless in early cases because of the profound immunosuppression of the patient at that time. For those, unmanipulated donor cell injection or in vitro expanded EBV–CTL adoptive transfer would be a better treatment. The sustained response rate to such therapy was demonstrated to be >70 % in 49 PTLD patients after hematopoietic stem cell transplantation (Doubroniva et al. 2012; Heslop 2012), with same results observed for donor unselected lymphocytes, donor EBV–CTL, or third-party EBV–CTL. Symptoms improved after 5–15 days, with radiological and complete response described after 3 weeks and 3 months, respectively. Cell therapy was also efficient in central nervous system localization of PTLD. Authors showed that treatment failure was associated with mismatched viral antigen or HLA presentation between EBV–CTL and tumor, explaining why the target proliferative cells were not recognized and eliminated. In a PTLD SCID mice model, injection of EBV–CTL resulted in delayed tumor development and tumors were prevented in 40 % of cases (p = 0.001). Results were improved by preselection of CD8+ cells or conditioning of CTLS with a combination of IL 2, 7, and 15 (Johannessen et al. 2011). Daily injection of IL2 in addition to EBV–CTL infusion led to PTLD prevention in 78 % of animals. Cell therapy was also shown to be an efficient preemptive strategy in 13 high-risk children receiving partially matched stem cell transplant (Leen et al. 2009). One single dose of EBV–CTL was administered 30 days after transplantation, and neither PTLD nor GVHD was observed. Limitations for cell therapy are GVHD risk (mostly with unmanipulated donor T cells) and need for special facilities to expand EBV–CTL and have them readily available in rapid progressing diseases, as their expansion requires 2–3 months (Heslop 2009; Bollard et al. 2012).

Targeting the proliferative cells with anti-CD20 antibodies (rituximab, 4-weekly injections of 375 mg/m²) is well recognized as part of initial treatment in case of CD20+ tumor, with response rates around 50 %. In a review of 80 PTLD in adults after solid organ transplantation, 74 % received first-line rituximab treatment in addition to reduced IS, with or without following chemotherapy (Evens et al. 2010). The 3-year progression-free survival and overall survival were 70 and 73 % in patients having received rituximab as compared to 21 and 33 % in others (p < 0.0001 and = 0.0001). Multivariate analysis showed 3 risk factors associated with disease progression and lower survival: central nervous system or bone marrow involvement and hypoalbuminemia. Overall survival in case of 0, 1, or more than 2 risk factors was, respectively, 93, 68, and 11 %. In case PTLD do not respond to IS reduction and rituximab within 8 weeks, it is now recommended to rapidly start chemotherapy (Montserrat et al. 2012). Seventy adult patients developed B-cell PTLD after solid organ transplant and were prospectively included in a phase 2 international study (Trappe et al. 2012). Subjects that did not respond to IS reduction received 4 courses of rituximab followed after 4 weeks by chemotherapy (cyclophosphamide, doxorubicin, vincristine, and prednisolone = CHOP). PTLD was of late type in 76 %, with monomorphic classification in 96 % and associated with EBV in only 44 % of patients. Fifty-nine patients received the complete treatment and showed partial and complete response in 90 and 68 % of cases. The median survival was 6.6 years. The main adverse events were severe leucopenia (68 %) or infections (41 %), with CHOP-related mortality in 11 % (5/7 being non-rituximab responders). EBV-associated PTLD occurred earlier and in younger patients, involved more often the graft and the kidney, and less frequently lymph nodes, were associated with lower performance status and developed more severe infections. Of the 40 patients with complete response, 8 relapsed and died from PTLD (5 within 6 months). Globally, lower risk of disease progression was observed in EBV-associated PTLD (HR 0.007) and higher risk was seen in patients in partial remission after sequential treatment (HR 20.83). The interest of low-dose chemotherapy was also described in children after liver transplant (Gross et al. 2005). Patients with EBV-related PTLD received six cycles of low-dose cyclophosphamide and prednisone (one cycle every 3 weeks with cyclophosphamide 600 mg/m² on day 1, prednisone 1 mg/kg orally twice a day on days 1–5) after a trial of IS reduction. Two-year disease-free survival and overall survival were 67 and 73 %. The same group reported results with rituximab added to the low-dose chemotherapy (375 mg/m² on days 1, 8, and 15 of the 2 first cycles) (Gross et al. 2012). Two-year disease-free survival and overall survival were 71 and 83 %. Side effects were similar to the previous study. Of the 55 children, 10 died, 7 of PTLD, and 3 of infections.

Antiviral drugs (ganciclovir, acyclovir) are not described as an efficient treatment for overt PTLD because they are inactive on latent EBV cycle present in these diseases. Surgery or radiotherapy can be useful in localized disease or central nervous system involvement, but often require additional treatment (Heslop 2009).

Outcome of the different treatments for EBV–PTLD in hematopoietic stem cell transplant recipients was extensively reviewed (Styczynski et al. 2009a), showing that preemptive therapy with rituximab of EBV–CTL reduced most significantly the risk of death (survival rates of 89.7 and 94.1 %). The same treatments used for overt PTLD also improved survival rates (63 and 88.2 %, respectively). Reduction in IS approximately allowed a survival rate of 56.6 %. Chemotherapy and antiviral agents did not influence survival of the patients.

According to all the results described, guidelines for PTLD treatment after solid organ or stem cell transplant were defined (Styczynski et al. 2009b; Parker et al. 2010). After solid organ transplantation, recommendations are the following. Reduction in IS should be initiated in all patients (limited disease, 25 % reduction; extensive disease, 50 % reduction in calcineurin inhibitors, stop azathioprine/mycophenolate, keep prednisone at maximum 10 mg/day; extensive disease and critically ill patient, stop all drugs except prednisone). Graft function should be carefully monitored. For patients with low-risk PTLD (age >60 years, normal LDH, performance status ECOG 0–1) and disease progression despite IS reduction, rituximab monotherapy is recommended. Addition of chemotherapy (CHOP-based regimen) should not be delayed in patients with disease progression despite this treatment or immediately in clinically aggressive PTLD. When chemotherapy is given, prophylactic granulocyte colony-stimulating factor and anti-infectious agents should be used. In case of central nervous system involvement, reduced IS should be followed by local radiotherapy and steroids. High-dose methotrexate can be considered in young patients. Surgery or radiotherapy may be adequate for localized diseases. Treatment with EBV–CTL infusions, antiviral drugs, and/or arginine butyrate, interferon or intravenous immunoglobulin should be currently limited to clinical trials. In cases where the graft failed or has been removed, retransplantation can be considered, if possible at least one year after PTLD healing. After bone marrow/stem cell transplantation, recommendations are the following. Regarding prevention of EBV reactivation, recipients of allogeneic grafts and graft donors should be tested for EBV serology, and antiviral agents and immunoglobulin have no impact and are not recommended. For PTLD preemptive therapy, high-risk patients after allogeneic transplant should be closely monitored for EBV load and PTLD symptoms, and for preemptive treatment with rituximab (1–2 doses), IS reduction or EBV–CTL should be considered in patients with high viral load. Monitoring is not necessary in HLA identical without T-cell depletion or auto transplantation. For PTLD treatment, rituximab (4–8 doses), IS reduction, or EBV–CTL are recommended as first-line therapy. Chemotherapy is a second-line treatment, while antiviral drugs are not recommended.