A long-recognized complication of solid organ transplantation is an increased risk of cancer in the setting of immunosuppression. A high prevalence of de novo malignancies, both cutaneous and noncutaneous, has been reported, ranging from 3 to 17%,1–8 among which is the heterogeneous group of hematological malignancies that make up posttransplantation lymphoproliferative disorder (PTLD). First recognized in 1968 in transplant patients as a subgroup of lymphoid tumors that regressed after reduction in immunosuppression, PTLD was then referred to as the “pseudolymphomas.”9–11 Since then PTLD has been associated with Epstein-Barr virus (EBV) infection-induced polyclonal or monoclonal lymphoid proliferation of rapid onset, varied clinical presentation due to a predilection for extranodal sites, and high mortality. The prevalence of PTLD increased with the introduction of cyclosporin A (CYA), and varies according to the organ transplanted: liver 1 to 2%; kidney 1 to 3%; heart 2 to 6%; lung 2 to 9%9, 12–19; and intestine 11 to 33%.15, 20 In addition, PTLD prevalence varies with the patient's age at the time of transplantation, as the frequency is higher in children (≈10%) compared to adults (2 to 3%).21–23 The increased occurrence of PTLD in children is attributed to the higher rate of EBV-seronegativity at the time of their transplantation and the subsequent acquisition of primary EBV infection as a prelude to PTLD. Higher EBV viral loads occur in EBV-seronegative patients who subsequently develop primary EBV infection than in those who are already seropositive; the highest viral loads are seen in those who develop PTLD.22 These observations imply that EBV predisposes to PTLD, and indeed EBV infection has previously been reported to occur in up to 90% of PTLD cases.22, 24, 25 Notwithstanding, recently an increase in EBV-negative PTLD has been noted, ranging in prevalence from 23 to 36%.26–28
Other suggested predisposing factors for PTLD after liver transplantation (LTx) include the degree of immunosuppression particularly with respect to the use of antilymphocyte antibodies (ALA),9, 15, 22, 26, 27, 29 and the cause of the liver disease that necessitated LTx.8, 22 Yet the risk factors for PTLD, the prevalence of this disease, and its outcomes are actually not well defined, because there has been variability in patient populations among the different reported case series, inconsistencies in the histologic classification of PTLD, and evolution in EBV detection methods. Nonetheless, there are several well-held concepts concerning PTLD, namely that: 1) PTLD is usually associated with EBV infection22, 30; 2) EBV-negative PTLD occurs later after LTx and portends a worse prognosis than EBV-positive PTLD9, 22, 26, 30, 31; and 3) patients who develop PTLD have received high-dose immunosuppressive regimens. To appraise these concepts, we reviewed the LTx recipients with PTLD at our institution, to determine the prevalence of this disease, its association with traditional risk factors, and survival after specific therapy. In addition, this retrospective analysis allowed us to chronicle the various clinical and pathologic presentations of PTLD, which by their diversity could mimic other disorders and delay definitive diagnosis and initiation of appropriate therapy. Using this analysis, we aimed to determine whether the traditional dogmas about PTLD are still valid.
We reviewed retrospectively the liver transplant database at the Medical University of South Carolina between January 1, 1990 and May 30, 2005 for patients who were diagnosed with histologically-confirmed PTLD, and recorded their clinical presentations, laboratory results, pertinent imaging findings, and outcomes. All but 3 of the pathology specimens were reviewed in a masked fashion by 2 hematopathologists (J.L., L.C.) and a clinical oncologist (R.S.), who based the diagnoses on the available biopsy material and relevant histologic analysis. With these exceptions, the tumors were categorized according to the World Health Organization classification of PTLD.32
The patients' presenting symptoms, tumor locations, and histological classification were recorded, and were related to: the etiology of original liver disease, the patients' ages, the time that had elapsed from LTx to PTLD presentation, the rejection and immunosuppression history, and EBV status as determined by in situ hybridization (EBV-ISH) where possible. The ISH assays were performed by ARUP Laboratories (Salt Lake City, UT) using the INFORM EBER Probe system (Ventana Medical Systems, Tucson, AZ), which utilized fluorescence labeled oligonucleotide probes for the detection of ribonucleic acid transcripts (EBV early ribonucleic acid). The ISH assay was performed on archival tissue (paraffin blocks), and an ribonucleic acid–positive control was used to confirm the integrity of the specimen nucleic acids on all tissues submitted for analysis to avoid false-negative results. For the purpose of the study, a tumor was only classified as EBV-negative on the basis of EBV-ISH. Therapy administered and clinical outcomes were also retrieved to determine the response to treatment and mortality. The foci of this retrospective analysis therefore were: the spectrum of clinical presentations, the oncologic spectrum of the lymphomatous disorders, risk factors for developing PTLD, and responses to therapy including survival.
Categorical data were summarized as the proportion of patients with a given attribute. Comparisons for categorical data were made using chi-squared analysis; quantitative comparisons of continuous data were made by parametric and nonparametric analysis as appropriate. For all comparisons, statistical significance was defined as a P value of less than 0.05.
A total of 621 LTx were performed during the study period, of whom 359 (58%) were male, 262 (42%) were female, and 67 (11%) were children. The patients' ages ranged from 2 months to 75 yr. A total of 22 cases of PTLD were diagnosed in 21 patients, giving an overall disease prevalence of 3.5% (see Table 1). One child had Hodgkin lymphoma followed 2 yr later by non-Hodgkin lymphoma. For 3 cases that had undergone histologic evaluation at the time of original diagnosis but for whom pathologic material was no longer available for histologic review in this study, classification according to the current World Health Organization criteria was not possible since their original diagnoses were made prior to its introduction.32 Therefore, these 3 cases are not included in the World Health Organization histologic analysis and EBV status, but they do appear in the demographics (Table 1) and clinical presentation (Table 2) sections below.
Table 1 summarizes the demographic information for the patients in the study. The median age of the cohort was 53 yr (range 7 months to 73 yr); 52% were male and 5 (24%) were children (as defined by age less than 18 yr).
The median age of the 5 children was 5.5 yr (range, 7 months to 17 yr); 3 were boys and 2 were girls. The prevalence of PTLD in our pediatric transplant population was 7.5%. Biliary atresia was the cause of the native liver disease in 3 patients; the others consisted of one case each of fulminant liver failure, Wilson's disease, and cholestatic cryptogenic liver disease. The immunosuppression regimen was tacrolimus (TAC)-based in 3 cases and CYA-based in 1 case; the patient with 2 sequential tumors had a CYA-based regimen at the time of the first PTLD and was switched to TAC prior to acquiring the second tumor. Four patients also required high-dose steroids for acute rejection (mean, 1.5 ± 1.3 standard deviation rejection episodes per patient), 3 of whom were each treated twice; and 1 patient had 4 rejection episodes, 1 of which occurred greater than 1 yr after LTx. Two of the 5 patients had steroid-resistant rejection and were treated with ALA. EBV infection was confirmed in 3 cases (out of 4 tested) by EBV-ISH, and the median time of onset of PTLD after LTx was 18 months (range, 2.5-60 months). Extranodal disease occurred at presentation in 4 cases.
The median age of the 16 adults diagnosed with PTLD was 54.5 yr (range, 36 to 73 yr); 9 were men and 7 were women. The prevalence of PTLD in the adult population was 3%. Alcohol was the cause of liver disease in 5 (31%) of the patients, while 3 (19%) of the transplants were done for hepatitis C. The etiologies of liver disease in the remaining patients transplanted are listed in Table 1. The immunosuppression regimen for the adult cohort was CYA-based in 12 patients (75%) and TAC-based in 4 (25%). Additionally, mycophenolate mofetil was included in the regimen of 5 patients (31%) and 6 (28%) received azathioprine. A total of 12 patients were treated for acute rejection (mean 1.1 ± 1.0 episodes per patient); 11 episodes occurred within the first year after LTx, while 2 patients had acute rejection greater than a year after LTx. Nine of the 12 had a single episode of rejection, 1 patient had 2 occurrences, and an additional 2 patients each had 3 episodes. The prevalence of steroid-resistant rejection requiring ALA was 25% (4/16 patients). Underlying EBV infection was confirmed by EBV-ISH in 2 cases (out of 9 tested); the median time of onset of PTLD after LTx was 18 months (range, 1-132 months). Extranodal disease was present at the time of diagnosis in 13 cases (81%).
The clinical spectrum of disease presentation was broad, as illustrated in Table 2. Lymphadenopathy was the presenting finding in 5 patients (23%), while 17 (77%) had extranodal disease. Of the 17 patients with extranodal disease, 6 (36%) had lesions in the liver allograft; 5 were mass lesions and 1 patient had lymphoproliferation on liver biopsy associated with monoclonal gammopathy and elevated liver enzymes. In the remaining 11 patients with extranodal disease, many different organ systems were involved, i.e., skin, breast, gastrointestinal tract, urinary tract, tonsils, and red blood cells. In addition, 1 patient presented with infiltration of the spleen and lymph nodes, multiorgan failure, and rapid death.
Tissue was available for review from 19 cases (see Table 3) but no specimen was available from 3 cases that had been diagnosed originally as a “B-cell lymphoma.” The reviewed material comprised 5 polymorphous PTLDs, 12 monomorphic PTLDs, and 2 Hodgkin lymphomas. The monomorphic PTLDs consisted of 5 diffuse large B-cell lymphomas, 2 cases of Burkitt or Burkitt-like lymphoma, 2 cases of follicular lymphoma, and 1 each of anaplastic myeloma, peripheral T-cell lymphoma, and lymphoplasmacytoid lymphoma. No cases were classified as so-called “early lesions”.32 The morphologic classification of the PTLD and the corresponding results of EBV-ISH in each case are shown in Table 3. The histologic heterogeneity of the cases examined is illustrated in the composite Figure 1.
Table 3. Tissue Source, Morphologic Diagnoses, and EBV Status of Cases
NOTE: Tissue specimens were available for microscopic review from 19 of 22 cases.
Abbreviations: WHO, World Health Organization; nd, not done; DLBCL, diffuse large B-cell lymphoma.
Small cleaved lymphoma
Peripheral T-cell lymphoma
EBV-ISH was performed on archived tissue from 13 of the 19 cases that were available for microscopic review; 5 were positive, while 8 were negative. In the other 6 cases, there was no remaining tissue for EBV-ISH testing. The characteristics of the 13 patients stratified according to EBV-ISH status are shown in Table 4.
Table 4. EBV-ISH Results With Outcome of PTLD Patients
EBV positive by ISH (n = 5)
EBV negative by ISH (n = 8)
NOTE: Tissue was available from 13 of the 22 cases for EBV-ISH examination.
The median time of onset of PTLD after transplantation for the 5 EBV-positive patients was 48 months (range, 6-132 months); 2 were adults and 3 were children, median age was 9 yr (range, 5-73 yr). Extranodal disease was present in 4 cases, 2 of whom had involvement of the liver allograft. Three patients had had acute rejection, with a mean of 1.4 ± 1.7 episodes of rejection for the cohort; 1 patient received ALA for steroid-resistant rejection. Immunosuppression was reduced in all patients, and, in addition, 4 patients received chemotherapy. Two of the 5 EBV-ISH positive patients died of PTLD.
EBV-ISH–Negative Patients (n = 8)
In the EBV-ISH–negative group, the median age was 52 yr (range, 17-56 yr); 7 were adults and 1 was a child. The median time of onset after LTx was 18 months (range, 2-108 months). Five patients had extranodal disease, but only 1 case involved the liver allograft. Seven patients were treated for acute rejection with steroids; the mean number of episodes of rejection for the EBV-ISH–negative group was 1.1± 0.6. None of the patients had steroid-resistant rejection requiring ALA use. Chemotherapy was utilized in addition to a reduction in immunosuppression in 7 cases, and the overall mortality for the group was 3 out of 8, but only 1 patient died from PTLD. The other 2 patients died several years later of causes unrelated to PTLD.
The difficulty in defining a disorder such as PTLD that is heterogeneous by nature and infrequent in its occurrence has understandably resulted in discrepant findings among the studies that described its associated risk factors, pathophysiology, clinical presentation, and outcome. A variable and evolving classification of PTLD and increasingly sensitive and specific methods for determining the EBV status of tumors have also affected the reporting of PTLD cases over the years. In addition, variability occurs in the disorder itself, with respect to the organ transplanted and the age of the recipient. Finally, the evolution of posttransplantation care over time may have also contributed to variations in the disorder, as suggested by Nelson et al.30 who showed a significant time-dependent increase in EBV-negative PTLD at their institution after 1991. As a result, the discrepancies among the observational studies of PTLD—depending on the organ transplanted, population studied, and the era in which the transplant was performed—allowed preconceptions to be formed. This has occurred regarding the role of EBV infection, specifically with respect to its role in the pathogenesis of disease, the time of onset after transplantation, and prognosis.
Historically, the prevalence of EBV infection in PTLD has been reported to be as high as 90 to 95%.15, 22, 26, 29, 33 However, the 62% prevalence of EBV-negative lesions among our 13 cases tested by ISH agrees well with the 21 to 34% prevalence reported in other recent case series.9, 28, 30 Even if the 6 cases in which ISH was not performed had been EBV-positive, the prevalence of EBV-negative PTLD in our patients would still be 42%. Thus, our data contradict the conventional wisdom that PTLD is primarily an EBV-induced disorder in which the virus escapes detection and eradication by the suppressed immune system. Our data also suggests that alternative measures to detect the presence of EBV infection (which were performed in a few patients in the cohort) are not as reliable as EBV-ISH. Two confirmed EBV-negative cases had evidence of rising serum EBV antibody titers, falsely suggesting the presence of EBV infection, while immunohistochemical staining for EBV was negative in 2 cases that were confirmed positive by ISH. Had we used immunohistochemical staining to determine EBV status, this would have falsely elevated our prevalence of EBV-negative tumors. Finally, the results of other studies have demonstrated a failure of serum polymerase chain reaction for EBV to determine accurately the EBV status of PTLD cases.33 This apparent discrepancy between diagnostic modalities for determining EBV status underscores the need to standardize better the manner in which the presence of EBV is determined in future studies of PTLD. Given our experience and the opinion of experts in the field concerning the high sensitivity of EBV-ISH,30 we feel ISH should be that standard.
Our cohort also deviated from the current conventional pattern of EBV-negative PTLD with respect to both the time of onset from transplant and mortality. While generally reported to occur later after transplantation than EBV-positive tumors,9, 15, 26, 28, 30 the median time of onset from transplantation in our EBV-negative patients was only 18 months vs. 45 months in those patients with confirmed EBV-positive PTLD. As shown in Figure 1, EBV-negative tumors occurred as early as 2 months after transplant, and 4 of the 8 cases (50%) were diagnosed within the first postoperative year. In addition, EBV-positive PTLD occurred any time from 2 months up to 11 yr after organ grafting. The dictum that EBV-negative PTLD portends a worse prognosis than does EBV-positive PTLD, due to higher mortality,22, 26, 28, 30 is contradicted in our study, as our EBV-positive cohort had a higher mortality rate than the EBV-negative cohort (57% vs. 13%, respectively).
In contrast to the unconventional behavior of our PTLD cases with respect to EBV status, we did find, however, that higher dose immunosuppression is still a risk factor for PTLD. The rates of acute allograft rejection after LTx have been reported to range from 40 to 50% for CYA-based regimens and 30 to 40% for TAC-based regimens.34, 35 The prevalence of acute rejection in our population was 77%, of whom 7 patients (32%) had several episodes of rejection. A high rate of steroid-resistant rejection requiring ALA (6/22, 27%) was also observed in our cases, which is greater than the generally reported rate of 10%.24, 25, 33 With respect to EBV status, there was no demonstrable difference between the EBV-positive and EBV-negative groups regarding the number of patients with acute rejection (60% vs. 75%, respectively). Neither could we tell if there was a difference between the groups concerning ALA use since only 1 EBV-positive patient and none of the EBV-negative patients received such therapy. The remaining 5 cases of ALA use occurred in patients in whom EBV-ISH was not done.
The inclusion of children in the study does allow for a few observations to be made with respect to the nature of PTLD in adults (see Table 1); however, the small number of children precludes statistical comparisons between the groups. As has been previously described, the children had a higher prevalence of disease (7.5% vs. 3%) and more EBV-positive PTLD than in adults (75% vs. 22%, see Table 4). Additionally, it appeared that a greater proportion of children received TAC-based immunosuppression as opposed to CYA-based, as was the case for the adults. Nonetheless, there was no difference between the groups with respect to occurrence of acute rejection, ALA use, time of presentation after LTx, and prevalence of extranodal disease.
A striking observation in our cohort was the diversity of both pathologic (Table 3) and clinical (Table 2) spectra of disease. Histologically, the most common tumors were polymorphic and diffuse large B-cell lymphoma, as would be expected. However, many other morphologic patterns were seen that are not prototypical for PTLD, including: lymphocytoplasmic, follicular, Burkitt, and anaplastic myeloma, and peripheral T-cell and Hodgkin lymphomas. As varied as the tumor histologies were, so too were the clinical presentations. The high prevalence of extranodal disease resulted in many organ systems being involved, including the gastrointestinal tract, red blood cells, hepatobiliary system, reticuloendothelial tissue, kidneys, lungs, and urinary system. That a mass lesion need not be present at all was evident in 4 patients, 2 of whom had lymphomatous ascites, 1 with acute onset of fever and renal failure secondary to peripheral T cell lymphoma, and another with multisystem organ failure. The presence of extranodal disease occurred irrespective of tumor morphology or EBV status, as illustrated in Figure 1. Moreover, extranodal PTLD is not necessarily a late occurrence, but can actually present within weeks of transplantation. In fact, only 1 of 9 of our cases of PTLD diagnosed within 6 months of transplantation presented with classical lymphadenopathy and an infectious mononucleosis-like syndrome. Indeed, the spectrum of clinical presentations of PTLD in liver transplant recipients is truly varied, including sore throat, hoarseness, breast lumps, subcutaneous nodules, biloma and biliary stricture, hemolysis, and pelvic mass with ureteral obstruction. Taking all of these observations into account, PTLD in LTx often masquerades as other disorders that are not obviously lymphoproliferative at first sight. PTLD in LTx can be considered a “great mimic,” like syphilis of old and other more modern protean disorders.36
PTLD is a conglomeration of hematologic malignancies whose heterogeneity is rivaled only by the diverse attempts to define this infrequent and challenging diagnosis. Whereas the current World Health Organization classification provides a consistent method for defining the tumors, further clarification is still required with respect to determination of EBV status. It is our conviction that this should be based on EBV-ISH as practiced by experts in the field37 since serum EBV antibody titers, EBV polymerase chain reaction, and tissue EBV immunohistochemical staining are unreliable. EBV-related PTLD appears to be less prevalent nowadays than previously. Additionally, the conventional notion that EBV-negative PTLD occurs later after transplantation and is more frequently fatal than EBV-negative tumors may no longer be true. We draw attention to the wide clinical spectrum of PTLD because of the high prevalence of extranodal disease, which may result in difficulty and allow delay in diagnosis. High vigilance and a low threshold for diagnosis are necessary to initiate therapy promptly in this disorder that still caries a high mortality rate.