Posttransplant Lymphoproliferative Disorder Following Pancreas Transplantation


* Corresponding author: Mark D. Stegall,


The incidence, risk factors and impact on patient and graft survival were evaluated for posttransplant lymphoproliferative disorder (PTLD) among 212 pancreas transplant recipients. Thirteen (6.1%) developed PTLD during 71 ± 27 months follow-up. Cumulative incidences of PTLD at 1, 3, 5 and 10 years posttransplant were 4.2%, 5.3%, 6.0% and 7.0%, respectively. Incidence of PTLD was lower for recipients of simultaneous pancreas kidney compared to pancreas after kidney transplant or pancreas transplant alone, though not significantly so. Recipient Epstein–Barr virus (EBV) seronegativity and number of doses of depleting antibody therapy administered at transplant were associated with increased risk of PTLD, while recipient age, gender, transplant type, cytomegalovirus mismatch maintenance immunosuppression type and treated acute rejection were not. All 13 cases underwent immunosuppression reduction, and 10 received anti-CD20 monoclonal antibody. During follow-up, 10/13 (77%) responded to treatment with complete remission, while 3 (23%) died as a result of PTLD. Patient and graft survivals did not differ for recipients with and without PTLD. The strong association of PTLD with EBV-seronegativity requires considering this risk factor when evaluating and monitoring pancreas transplant recipients. With reduction of immunosuppression and anti-CD20 therapy, survival for pancreas transplant recipients with PTLD was substantially better than previously reported.


Pancreas transplantation is a successful treatment for diabetic patients with severe progressive end-organ complications or life-threatening hypoglycemic unawareness. Compared to kidney transplantation, however, pancreas transplantation is associated with higher rates of surgical complication and immunological graft loss. Thus, given the availability of insulin and other pharmacotherapies for diabetes mellitus, the decision to pursue pancreas transplantation requires careful patient selection and risk assessment. The risk for posttransplant lymphoproliferative disorder (PTLD) among pancreas allograft recipient has not been extensively studied to date and much of the available information derives from case reports and series that combine recipients of various types of solid organ transplants. Single-center studies that have focused specifically on pancreas transplant recipients provide valuable clinical insights, but have described patient characteristics and outcomes over a wide time frame and have not included information on some of the important potential risk factors (1–3).

PTLD is a serious and potentially fatal complication of solid organ transplantation with an incidence of 1–13%, depending on the type of transplanted organ and the length of follow-up (4–12). PTLD represents a spectrum of tumors ranging from benign lymphoid proliferation to malignant monomorphic ß-cell lymphoma that is associated with high mortality (13–16). The incidence of PTLD typically peaks in the first year after transplantation and may be associated with primary or secondary Epstein–Barr virus (EBV) infection, intensity of immunosuppression and concurrent cytomegalovirus (CMV) disease (1,6,17). PTLD differs from other lymphoproliferative disorders that occur in the general population by its extranodal and allograft involvement. The tumors are most commonly ß-cell lymphomas, and less commonly originate from T cells (18). Depending on extent, histological characteristics and EBV positivity, treatment approaches for PTLD may include reduction or complete elimination of immunosuppressive medication, anti-CD20 monoclonal antibody therapy, conventional chemotherapy, surgical excision and radiotherapy (19–22).

During the past decade, PTLD has been recognized as a significant source of morbidity and as a transplant-related complication for which adequate risk assessment, preventive strategies and prognostic information are lacking. In the case of pancreas transplantation, a better understanding of PTLD risk could have a significant impact on the decision to pursue transplantation in the first place and on pre- and posttransplant management of individual patients. We hypothesized that analogous to other solid organ transplants, there may be particular factors that place pancreas transplant recipients at increased risk for the development of PTLD.

The purposes of this study were: (a) to identify the incidence of PTLD in a large, recent cohort of pancreas transplant recipients, (b) to describe the clinical course of PTLD in recent pancreas transplant recipients, (c) to delineate the risk factors associated with the development of PTLD and (d) to determine the overall prognosis and response to therapy among pancreas transplant recipients diagnosed with PTLD.

Materials and Methods

Study design and data collection

A retrospective review was performed of all adult pancreas transplant recipients at Mayo Clinic, Rochester, Minnesota from January 1, 1998 to December 31, 2006 using an electronic transplant database and medical records. Recipients of pancreas transplant alone (PTA), simultaneous pancreas–kidney transplants (SPK) and pancreas after kidney transplants (PAK) were included. Follow-up for vital status and occurrence of PTLD lasted through April 20, 2008. From a total of 242 pancreas transplant recipients, 30 were excluded from the analysis due to loss of the pancreas graft within 1 week of transplant (primarily due to early graft thrombosis), leaving 212 available patients for analysis. The study was approved by the Mayo Clinic Institutional Review Board.

Data collected on all pancreas transplant recipients included basic demographics, induction and maintenance immunosuppression, EBV and CMV recipient and donor pretransplant serological status, use of antiviral medication prophylaxis, biopsy-proven acute rejection (AR), treatment for AR and pancreas allograft and patient survival outcomes. Anti-EBNA and anti-VCA IgG and IgM were measured in serum by enzyme immunoassay (EIA). Recipients with negative result for both assays or with positive result for anti-VCA alone were considered to be EBV seronegative. Patients with a diagnosis of lymphoma, following pancreas transplantation were identified and detailed information was abstracted from the medical record of each. Clinical and laboratory information on histology, cell clonality, treatment modalities and survival outcomes of those recipients diagnosed with PTLD were reviewed. All patients with a diagnosis of PTLD were evaluated and followed closely by a hematologist and underwent staging with, at a minimum, whole-body CT scanning and bone marrow biopsy. Pathologic specimens were evaluated via routine ß-cell and T-cell immunophenotyping as well as cell clonality studies. EBV was detected in tumor cells by in situ hybridization utilizing probes for EBV-encoded RNA.

Immunosuppressive therapy and anti-viral prophylaxis

Rabbit antithymocyte globulin was given as induction therapy or therapy for biopsy-proven acute allograft rejection at 1.5 mg/kg/day intravenously. Induction and therapeutic courses consisted of daily doses for between 5 and 10 doses. Tacrolimus therapy was targeted to trough levels between 10 and 15 ng/mL during the first 4 months posttransplant and between 8 and 10 ng/mL, thereafter. Mycophenolate mofetil (MMF) was administered orally at a standard dose of 1 g twice daily. All recipients received methylprednisolone bolus therapy at the time of transplantation followed by oral prednisone tapered over 4 months to a baseline dose of 5–10 mg daily. Antiviral prophylaxis was prescribed for the first 3 months posttransplant on the basis of donor and recipient CMV serostatus. Donor CMV-positive (D+)/recipient CMV-positive (R+), D+/recipient CMV-negative (R−) and D−/R+ combinations received oral ganciclvir (1000 mg three times daily adjusted for GFR <60 mL/min) or oral valganciclovir (900 mg twice daily adjusted for GFR <60 mL/min) while D−/R− combination received oral acyclovir 200 mg three times daily.

Diagnosis and management of acute pancreas rejection

Acute pancreas transplant rejection was diagnosed by ultrasound-guided transcutaneous 18-guage needle biopsy. Biopsies were carried out for cause (persistent unexplained increase in serum pancreatic enzymes) or as part of a surveillance protocol. Surveillance biopsies were obtained 1–2 weeks after induction therapy, at 2–3 months and at 1 year posttransplant. Hematoxylin and eosin-stained sections were reviewed by a Mayo Clinic specialist gastrointestinal pathologist and were graded using the University of Maryland criteria (23). Antirejection treatment was administered for grade III rejection or greater, whether diagnosed by clinically-indicated or surveillance biopsy. During the time period studied, initial treatment consisted of corticosteroid boluses with either OKT3 (5 mg/day for 7–10 days) or rabbit antithymocyte globulin (1.5 mg/kg/day for 5–10 days). Grade II rejection was treated in some cases only based on clinical judgment and consisted of corticosteroid boluses with or without depleting antibody therapy. Treatment-resistant or recurrent AR was managed according to clinical judgment by additional courses of depleting antibody of variable duration, by corticosteroid boluses alone or, in some cases, with no additional therapy. In patients for whom additional immunosuppressive therapy was considered to be contraindicated (including those with a prior history of PTLD), treatment of AR was limited to corticosteroid bolus therapy alone.

Statistical analysis

Data were summarized using means ± standard deviation for numeric variables, and counts and percents for categorical variables. Cumulative incidences of PTLD following pancreas transplantation accounting for the competing risk of death were derived using an extension of the Kaplan–Meier method (24). Cumulative incidences of death were derived using the Kaplan–Meier method. Cox regression analysis was used to assess risk factors for PTLD and risk factors for death and graft loss subsequent to PTLD. The impact of PTLD on graft loss and death was assessed using history of PTLD as a time dependent predictor in a Cox regression model. Because of the limited number of PTLD cases, the focus of the Cox regression analysis was on the univariate regression models with two-term models considered in an exploratory manner. Frequencies of treated AR among recipients with and without PTLD were compared by Fisher's exact test. Tests were conducted at the two-sided 5% level.


Patient characteristics

Overall patient characteristics are summarized in Table 1. The study population comprised 212 pancreas transplant recipients with a mean age of 43.0 ± 9.1 years (range 21–71 years) of which 109 (51%) were men. Relative to the last follow-up date, patients were followed up for 71 ± 27 months after transplantation. The induction immunosuppressive agent was rabbit antithymocyte globulin in 188 (88.7%), OKT3 in 23 (10.8%) and ATGAM in 1 (0.5%). Pretransplant recipient EBV serology was available for 209 recipients with 24 (11%) being EBV seronegative. Pretransplant CMV serology was available in 195 donor–recipient pairs, of whom 28 (14%) were CMV donor positive/recipient negative (CMV mismatch). Tacrolimus was the primary oral immunosuppressive agent in 208 (98%) patients while cyclosporine (CSA) was employed primarily in 4 (2%).

Table 1.  Characteristics of pancreas transplant recipients1
Pancreas recipients – (n)n = 212
  1. rATG = rabbit anti-thymoglobulin; MMF = mycophenolate mofetil.

  2. 1Data are expressed as mean ± SD (minimum, maximum) or n (%).

Age (years)43 ± 9 (21, 71) 
Female gender103 (49%)
Caucasian race207 (98%)
First pancreas transplantation203 (96%)
EBV negative 24 (11%)
CMV mismatch (donor +/recipient −) 28 (14%)
Induction immunosuppression
  rATG2 188 (88.7%)
  OKT3  23 (10.8%)
  ATGAM  1 (0.5%)
Maintenance immunosuppression
  Tacrolimus208 (98%)
  Cyclosporine 4 (2%)
  MMF200 (94%)
  Prednisone 212 (100%)
Antiviral prophylaxis 212 (100%)
Type of transplant
  PAK101 (48%)
  PTA 68 (32%)
  SPK 43 (20%)
PTLD  13 (6.1%)
Follow-up (months)70.7 ± 27.4

Incidence of PTLD after pancreas transplantation

Thirteen of 212 pancreas transplant recipients (6.1%) developed PTLD during 1046 person years of follow-up. PTLD was diagnosed between 1.3 months and 6.1 years posttransplant. The cumulative incidence of PTLD was 4.2%, 5.3%, 6.0%, 7.0% and 7.0% at 12, 36, 60, 84 and 120 months, respectively (Figure 1A). At 6.1 years, the latest time of PTLD diagnosis, 73 patients remained at risk. Overall, there was an accelerated incidence of PTLD in the first year after transplantation with a lower and relatively constant rate thereafter. Considering these two periods, the rate for the first year was 9/205 or 44 PTLDs per 1000 person years, and thereafter 4/841 or 4.8 PTLDs per 1000 years, for a rate ratio of 9.3 (95% CI 3.0–34.2, p = 0.0001). As shown in Figure 1B the cumulative incidence of PTLD was significantly higher in EBV seronegative recipients than EBV seropositive recipients (p = 0.008) with the difference occurring exclusively during the first posttransplant year. There was only one PTLD case diagnosed in the SPK group. The cumulative incidences of PTLD at 12, 24, 60 and 120 months for SPK were 0%, 0%, 0% and 3.9%, respectively, and those for PAK were 4.0%, 4.0%, 6.9% and 6.9%, respectively, and for PTA 7.4%, 7.4%, 8.9% and 8.9%, respectively. Given the case numbers involved, however, these cumulative incidences did not differ significantly between the three groups (Log-rank p = 0.37, Figure 2). For the PAK cohort, the median interval between kidney and pancreas transplants was 15.3 (interquartile range 7.6–45.2) months. The intervals between transplants for PAK recipients that did and did not develop PTLD following pancreas transplantation were 61.1 (9.9–100.1) and 13.8 (7.1–33.7), respectively (Mann–Whitney U-test p = 0.09).

Figure 1.

(A) The cumulative incidence of PTLD accounting for competing risk of death is shown for the entire study cohort following pancreas transplantation. The estimate is depicted as a solid line with 95% confidence interval shown as dashed lines. (B) Cumulative incidences of PTLD following transplantation are shown for pancreas transplant recipients with negative (dashed line) and positive (solid line) baseline EBV serologic testing.

Figure 2.

Cumulative incidences of PTLD are shown for recipients of pancreas transplant alone (PTA, short dashes), pancreas after kidney transplant (PAK, long dashes) and simultaneous pancreas kidney transplant (SPK, solid line).

Risk factors for PTLD

In univariate analyses EBV seronegativity prior to transplantation was associated with a higher risk of PTLD (HR = 5.33, 95% CI 1.74–16.3, p = 0.008). The number of doses of depleting antilymphocyte antibody given as induction therapy was also associated with an increased rate of PTLD (HR = 2.13 per dose, 95% CI 1.33–3.41, p = 0.0007). Recipient age, gender, transplant type (SPK, PAK or PAT), CMV mismatch, maintenance immunosuppressant and treatment for AR (grade ≥ II or grade ≥ III) were not associated with risk of PTLD. Considered in an exploratory manner, both EBV seronegativity and number of doses of depleting antibody therapy remained as significant risk associations with PTLD when entered together in a Cox model (HR = 4.12 95% CI 1.34–12.6, p = 0.013; and HR = 2.14 95% CI 1.28–3.58, p = 0.004, respectively).

Relationship between PTLD and acute pancreas transplant rejection

As shown in Table 2, Cox regression analysis demonstrated no significant association between treatment for AR ≥ grade II or ≥ grade III and diagnosis of PTLD although hazard ratios were 3.0 and 2.4, respectively. In order to more clearly delineate the relationship between treatment for biopsy-proven AR and development of PTLD, the frequency of treated AR among recipients without PTLD was compared with that of the PTLD cohort either limited to rejection occurring prior to PTLD diagnosis or including rejection episodes that occurred either before and after diagnosis. As shown in Table 3, treated AR occurred in 27.8% of all non-PTLD recipients compared to 30.8% of all PTLD recipients prior to diagnosis (p = NS). In contrast, when AR occurring after PTLD diagnosis was included, the frequency increased to 61.5% in this cohort (p = 0.04 compared to non-PTLD). Similar trends were observed for PAK, PTA and SPK subgroups, although it was notable that treated AR of the pancreas was less frequent among SPK recipients.

Table 2.  Risk factors for PTLD: results from univariable Cox regressions
 Hazard ratio (95% CI)   p-Value
Age at transplant0.96 (0.91–1.03)0.3
Male gender1.14 (0.38–1.14)0.8
EBV seronegativity5.33 (1.74–16.3)  0.008
CMV mismatch0.19 (0.08–5.10) 0.65
Induction-related depleting antibody therapy (per dose)2.13 (1.33–3.41)   0.0007
Tacrolimus therapy0.27 (0.35–2.07)0.3
Treated AR grade II or greater  3.0 (0.87–10.48)0.2
Treated AR grade III or greater 2.4 (0.61–9.39)0.1
Type of transplant
 PTA4.04 (0.49)     
 PAK2.84 (0.34)    0.3
Table 3.  Treated acute pancreas rejection within the study cohorts
 AllNo PTLDPTLD (prediagnosis)p1PTLD (pre- and postdiagnosis)p
  1. 1PTLD versus no PTLD.

  2. Fisher's Exact test.

Recipients with 1 or more treated AR n (%)63 (29.7)55 (27.8)4 (30.8)NS8 (61.5)0.02
 PAK33 (32.7)30 (31.6)2 (33.3)NS3 (50) 0.3 
 PTA27 (39.7)23 (37.1)2 (33.3)NS 4 (66.7)0.15
 SPK3 (7.0)2 (4.9)0 (0)    NS1 (100)0.07

Characteristics and treatment of the PTLD cases

The patient characteristics, histological features, tumor localizations and treatment modalities for all 13 PTLD cases are depicted in Table 4. The most common sites of occurrence were lymph node (5/13, 39%), CNS and liver (4/13, 31% each). Of note, only 1/13 (8%) lymphomas involved the allograft, this rate being comparable to the 10% graft involvement rate reported by Opelz and Döhler for kidney recipients and lower than rates reported in the same large study for liver (22%) and lung/heart-lung (53%) (7). In total, nine (69%) patients exhibited extranodal involvement. From a histological perspective, one lymphoma was shown to consist of polyclonal ß- and T cells while the remaining 12 were monoclonal ß-cell type. All 13 cases were CD20 positive and 10/13 (77%) were EBV positive by in situ hybridization. In general, the histological characteristics were not notably different to those reported for cases of series consisting of kidney, liver, heart and lung recipients (4). Treatment of PTLD was tailored according to histology (Table 4). Immunosuppression reduction was carried out in all cases and consisted either of complete withdrawal of both mycophenolate mofetil and tacrolimus with continuation of prednisone (8/13 [62%]) or withdrawal of mycophenolate mofetil and continuation of prednisone combined with low dose tacrolimus (target trough level 4 to 6 ng/mL; 5/13 (38%)). Monoclonal anti-CD20 antibody therapy (rituximab) was administered to 10/13 (77%) cases. Chemotherapy with R-CHOP (rituximab, cyclophosphamide, hydroxydaunorubicin, vincristine and prednisone) followed by a standard relapse regimen, ICE (ifosfamide, carboplatin and etoposide) was used in one patient.

Table 4.  PTLD in pancreas transplant recipients at the Mayo Clinic, Rochester, MN 1998–2006: patient and disease characteristics
PatientTransplant typeSexAge (yrs)EBV statusOnset of PTLD (mos)Cell typeCD20In situ EBVClonalityOrgan involvementTreatmentFollow-up (mos)Graft survivalPatient survival
  1. 1PTA = pancreas transplant alone; SPK = simultaneous pancreas-kidney transplant; PAK = pancreas after kidney transplant; Pos = positive; Neg = negative; CNS = central nervous system; LN = lymph nodes.

  2. 2‘Immunosuppression reduction’ refers to withdrawal of all immunosuppression except for prednisone.

  3. 3‘Low TAC’ refers to continuation of low-level tacrolimus in combination with prednisone.

  4. 4Recipient died of non-PTLD-related cause following prolonged complete remission.

1PAK1F43Neg 1.3T + BPosPosPolyclonalCNSImmunosuppression reduction2113 YesYes
2PAKM39Neg 1.7BPosPosMonoclonalLiverImmunosuppression reduction, rituximab80NoYes
3PTAM37Pos 3.5BPosPosMonoclonalCNSImmunosuppression reduction, rituximab81YesYes
4PAKM39Pos 4.0BPosPosMonoclonalLiverLow TAC3, rituximab74YesYes
5PTAF23Neg 5.4BPosPosMonoclonalLNLow TAC, rituximab16NoYes
6PTAM27Neg 7.1BPosPosMonoclonalLNImmunosuppression reduction, rituximab83YesYes
7PTAF43Neg 7.3BPosPosMonoclonalGraft, colonImmunosuppression reducation 0NoNo
8PAKM49Pos 9.4BPosNegMonoclonalLN (mesenteric, retroperitoneal)Immunosuppression reduction, rituximab + chemotherapy: (R-CHOP, ICE)14YesNo
9PTAM46Pos11.3BPosPosMonoclonalLiver, spleen, skeletonLow TAC, rituximab14YesYes
10PTAF43Pos30.4BPosPosMonoclonalLN, tongueLow TAC75NoYes
11PAKM42Pos32.8BPosPosMonoclonalCNSImmunosuppression reduction, rituximab 1YesNo
12PAKF51Pos56.7BPosNegMonoclonalLN (intraperitoneal, mesenteric)Immunosuppression reduction, rituximab34NoNo4
13SPKF44Pos73.8BPosPosMonoclonalCNSImmunosuppression reduction, rituximab32YesYes

Graft and patient survival and prognostic factors

Of the entire cohort, 39 (18.4%) recipients died during follow-up. Four (31%) of the 13 patients diagnosed with PTLD died compared to 35/199 (18%) of those without PTLD. This difference in mortality during follow-up for patients with or without PTLD was not significantly different (HR = 1.9, 95%CI 0.59–6.4, p = 0.32). Overall patient survival at 12, 36 and 96 months posttransplant was 96%, 87% and 77%, respectively compared to 84%, 65% and 65%, respectively for those diagnosed with PTLD. Pancreas allograft survival was lower but not significantly so in patients with PTLD compared to those without (HR = 1.9, 95% CI 0.60–6.5, p = 0.31). Death-censored pancreas allograft survival for the entire cohort was 96%, 89%, 80% and 73% at 12, 36, 60 and 96 months, respectively posttransplant. In comparison, death-censored allograft survival subsequent to PTLD diagnosis was 90%, 77%, 77% and 38%, respectively at the same time points.

Following treatment for PTLD, 10 (77%) recipients became disease free during a mean period of observation of 47 months (range 0–110 months). Of these responders, 1/10 (10%) were managed by immunosuppression reduction alone and the remaining 9 (90%) were treated with rituximab. Of the four deaths that occurred among pancreas transplant recipients with PTLD, three (75%) were directly related to lymphoma progression. The fourth recipient died as a result of severe cognitive difficulties and multiple neurologic deficits following prolonged hypoglycemia due to exogenous insulin and status epilepticus that occurred 34 months following PTLD diagnosis and during prolonged disease-free remission.


This study builds upon a relatively limited existing literature regarding the incidence, risk factors, treatment and prognosis of PTLD among pancreas transplant recipients. Importantly, the study is confined to the most recent era of pancreas transplantation and includes a comprehensive amount of clinical, laboratory and histological data for our entire cohort of transplant recipients during this time period. Furthermore, we have been able to determine the impact of factors such as EBV serostatus on PTLD risk. Finally, we provide an updated report of PTLD remission rates and patient and graft survival following the introduction of anti-CD20 monoclonal therapy for lymphoma. The results suggest that assessment of the risk for PTLD should be incorporated into patient evaluation for pancreas transplantation and that this group of transplant recipients should be considered for PTLD surveillance and prevention strategies.

The analysis of this cohort demonstrates that the incidence of PTLD in pancreas transplant recipients appears to be greater than that of kidney recipients and comparable to that of liver and heart transplant recipients (1,5,13). Reasons for the increased risk level compared to kidney recipients may include higher rates of EBV seronegativity and higher overall level of immunosuppressive therapy among pancreas recipients as well as organ-specific or immunologic factors that cannot be addressed through review of clinical data alone. Based on the striking effect of EBV seronegativity on early PTLD risk in this cohort, practice changes for EBV-negative pancreas transplant candidates could be considered. Although donor EBV serostatus was not available to us for the pancreas allografts described here, all PTLD cases among EBV-negative recipients occurred between 1 and 8 months posttransplant and were positive for EBV DNA by in situ hybridization. Thus, it is highly likely that the donors were EBV positive in these cases and that donor seropositivity is the primary determinant of PTLD risk among EBV-naïve recipients. Therefore, confirmation of donor EBV seronegative status prior to transplantation or even delaying transplantation until EBV seroconversion occurs in potential recipients without life-threatening hypoglycemic unawareness should be considered. Consideration of more formal practice guidelines to limit overall exposure to depleting antibody therapy in pancreas transplant recipients may also merit consideration on the basis of the results reported here although such measures must be balanced against the recognized value of such therapy to prevent and treat acute pancreas allograft rejection. It should be noted that the relationship between AR and development of PTLD among pancreas recipients is a complex one as rejection risk may be increased following reduction of immunosuppressive medications. Our analyses indicated that treated AR was not a significant risk factor for PTLD and that treatment for AR did not occur more frequently in pancreas allograft recipients who subsequently developed PTLD (although AR occurring after diagnosis resulted in an eventual increased rejection risk in this cohort). Nonetheless, dose number of depleting antibody therapy administered during induction was clearly associated with PTLD risk and pancreas recipients requiring prolonged or repeated courses of depleting antibody therapy should be closely monitored for clinical or laboratory evidence of EBV reactivation and development of lymphoma.

Also of interest, while PAK recipients had obviously received additional immunosuppression exposure (including depleting antibody therapy in some cases) related to kidney transplantation, there was no evidence of increased PTLD incidence in this cohort compared to PTA and SPK recipients. Importantly, the initial and subsequent immunosuppressive protocols for PAK and PTA recipients did not differ with regard to the pancreas transplants. The median interval between kidney and pancreas transplants was, in fact, longer for PAK recipients who developed PTLD compared to those who did not. These observations, which provide some reassurance regarding the safety of sequential transplantation in general, may indicate that the risk relationship between immunosuppression exposure and EBV-associated PTLD applies predominantly during the first posttransplant year (for both kidney and pancreas transplants) and that the typical separation in time between the two transplants in the PAK cohort results in an avoidance of additive risk. For patients with perceived high risk for PTLD prior to pancreas transplantation, consideration of alternative induction strategies may be merited (25). Additional long-term outcome studies of pancreas transplant recipients receiving diverse immunosuppressive regimens would be of value in this regard. Prolonged use of antiviral prophylaxis with acyclovir, ganciclovir, valganciclovir or anti-CMV immunoglobulin may also be considered although the relative efficacy of these agents to prevent PTLD occurrence remains debatable (26,27). An additional practice strategy that may be specifically applicable to EBV-naïve pancreas transplant recipients is viral surveillance of blood using DNA-based techniques to detect primary EBV infection at the earliest possible time point posttransplant. Sequential quantitative PCR (qPCR) assays have been well validated as a sensitive means to identify increasing EBV replication in transplant recipients and, coupled with immunosuppression reduction and/or initiation of antiviral therapy, may have the potential to prevent or halt the development of EBV-associated lymphoproliferation (28–30). A blood qPCR-based viral surveillance protocol among EBV-seronegative recipients has recently been introduced into our practice for kidney and pancreas recipients but we do not have data regarding viremia and viral load for the large majority of patients reported in the current study. Preliminary results for EBV-negative kidney recipients at our center indicate that PTLD is associated with detectable viremia and higher viral loads at the time of diagnosis (MD Griffin, TS Larson, TM Habermann and MD Stegall unpublished data), however, further prospective studies of organ transplant recipients will be necessary to determine whether EBV screening with preemptive intervention can be successfully applied to reduce the occurrence of PTLD with an acceptable rejection risk.

In contrast to the finding of higher PTLD incidence compared to previous studies, our data also suggest that the prognosis for PTLD in pancreas transplant recipients may be substantially better during the past decade than the outcomes reported by Paraskevas et al. (1) over a longer period of observation. It is likely that the predominance of early, EBV-associated, CD20+ lymphomas among our patient cohort resulted in a high rate of response to initial therapeutic interventions. Our experience also indicates that a combined approach of immunosuppression reduction and rituximab therapy is associated with prolonged disease-free remission in the majority of pancreas transplant recipients with EBV-associated, CD20+ PTLD (31–34). Nonetheless, three deaths occurred that were directly attributable to PTLD, reaffirming the potential severity of this complication. As regards graft outcome, the results reported here also provide some reassurance. We did not observe a lower rate of death-censored graft survival among pancreas recipients diagnosed with PTLD compared to those without. Indeed, the 5-year graft survival of 77% among those with PTLD, who did not succumb to the disease indicates that irreversible AR remains relatively uncommon despite prolonged substantial reduction in immunosuppression. Although the potential to determine longer-term allograft survival among PTLD survivors in this cohort was limited due to the number of cases followed beyond 5 years, the observed 8-year death-censored graft survival of 38% (compared to 73% for those without PTLD), while not statistically significant, raises a concern regarding eventual chronic graft deterioration.

Some general limitations to the study should be acknowledged. As a single-center retrospective analysis, practice-specific and demographic influences on PTLD incidence and outcomes may restrict applicability of the results to other transplant centers — particularly those with differing approaches to immunosuppressive management. Furthermore, the relatively low number of cases and lack of available diagnostic testing for some facets of antiviral immunity make it likely that additional risk factors for PTLD remain undetected or unaccounted for. For instance, while we observed that PTLD occurred at a higher rate among PTA and PAK recipients compared to SPK recipients, this observation did not reach statistical significance likely as a result of available numbers for the three groups. Similarly, while we demonstrate an association between PTLD risk and induction-related dose number of depleting antibody therapy, case numbers were not sufficient to determine whether a ‘threshold dose’ for depleting antibody exists above which PTLD risk increases to an unacceptable level or to examine the role of depleting antibody given as therapy for AR. Examples of other factors unaccounted for in the study include EBV exposure through blood transfusions, variations in cellular immunity to EBV or in viral strain, and modifying influences of other viral infections.

Nonetheless, the series brings into focus the significance of primary EBV infection and overall burden of immunosuppressive therapy on the risk for early development of PTLD among diabetic patients receiving pancreas transplantation over the past decade. During this time period many centers, including our own, have successfully applied newly available immunosupppressive combinations to reduce clinical and subclinical rejection rates among pancreas recipients (35–37). Given the unique and complex dynamics involved in counseling diabetic patients about the potential benefits and risks of pancreas transplantation, we believe that the PTLD risk factors and outcomes we report here should prompt pancreas transplant programs to reevaluate practices related to pretransplant PTLD risk stratification. While it appears that outcomes for PTLD among pancreas transplant recipients have improved compared to prior reports, it is also clear that prevention of this complication should remain a priority in the field.


This study was presented in part at the 2008 American Transplant Congress in Toronto, Canada. The authors would like to acknowledge the contribution of the kidney–pancreas transplant coordinators Mayo Clinic, Rochester for their thorough follow-up and dedication to the care of the patients described in the study.

Conflict of Interest Statement

There is no conflict of interest from any of the authors.