Journal Logo

Original Clinical Science—General

Immunosuppression Is Associated With Clinical Features and Relapse Risk of B Cell Posttransplant Lymphoproliferative Disorder

A Retrospective Analysis Based on the Prospective, International, Multicenter PTLD-1 Trials

Zimmermann, Heiner MD1,2; Babel, Nina MD3; Dierickx, Daan MD4; Morschhauser, Franck MD5; Mollee, Peter MBBS6; Zaucha, Jan M. MD7,8; Dreyling, Martin H. MD9; Dührsen, Ulrich MD10; Reinke, Petra MD11; Verhoef, Gregor MD4; Subklewe, Marion MD9; Hüttmann, Andreas MD10; Tousseyn, Thomas MD12; Bachy, Emmanuel MD13; Hauser, Ingeborg A. MD14; Tarella, Corrado MD15; Van Den Neste, Eric MD16; Gheysens, Olivier MD17; Anagnostopoulos, Ioannis MD18; Leblond, Veronique MD19; Riess, Hanno MD20; Choquet, Sylvain MD19; Trappe, Ralf U. MD1,2,20

Author Information
doi: 10.1097/TP.0000000000002269

The incidence of posttransplant lymphoproliferative disorders (PTLD) in patients after solid organ transplantation is significantly increased compared to non-Hodgkin lymphoma in immunocompetent patients and varies with the transplanted organ and the intensity of immunosuppression.1-3

Drugs used for maintenance immunosuppression after solid organ transplantation can be grouped by their mechanism of action. They include calcineurin inhibitors (CNIs; ciclosporin and tacrolimus), antimetabolites (azathioprine and mycophenolate), inhibitors of mammalian target of rapamycin (mTOR inhibitors; rapamycin [syn. sirolimus] and everolimus), the selective T-cell costimulatory blocker betalacept and corticosteroids.4

The association between individual drug classes or drugs used as maintenance immunosuppression and the risk of development of PTLD is complex and the best-quality evidence stems from large transplant registries: The incidence of PTLD in patients treated with tacrolimus and an antimetabolite was increased twofold compared with those treated with ciclosporin and an antimetabolite in a retrospective analysis of 25 000 kidney transplant recipients.3 Immunosuppression with an mTOR inhibitor plus tacrolimus was associated with an increased risk of PTLD (hazard ratio [HR], 1.40; 95% confidence interval [CI], 1.03-1.902; P = 0.03) compared with mycophenolate plus tacrolimus in an analysis of 114 000 kidney transplant recipients.5 Conversely, Quinlan et al6 identified a significantly decreased risk of late-onset PTLD in patients on corticosteroid maintenance (HR, 0.64; 95% CI, 0.44-0.95; P = 0.005) using data from 150 000 kidney transplant recipients.

Immunosuppression reduction after diagnosis of PTLD has the goal of reestablishing host T-cell function sufficiently to control lymphoproliferation without compromising the grafted organ and has been reported to yield high response rates (45%) in a retrospective analysis.7 On the other hand, the only prospective trial conducted so far demonstrated a response to immunosuppression reduction in 1 (6%) of 16 cases.8 Current recommendations for immunosuppression reduction are based on guidelines originally formulated for renal transplant recipients and mirror the protocols used by both Reshef et al7 and Swinnen et al8: Stop antimetabolites (azathioprine and mycophenolate), reduce calcineurin inhibitors by 25% to 50% and maintain corticosteroids.9-11 This was also the recommendation for patients in the PTLD-1 trials.12,13 However, both Paya et al and Swinnen et al note that the ideal regime for immunosuppression reduction after PTLD is not known. The former describe their recommendation as a “potential algorithm” and the latter as based on “common clinical practice”.9,8

A succession of prospective clinical phase II trials has led to a successful, evidence-based standardized treatment protocol for the most common subtype of PTLD, CD20-positive B-cell PTLD, which accounts for 80% of cases. In 2007, the PTLD-1 trial (n = 70) demonstrated the safety and efficacy of sequential treatment with 4 cycles of weekly rituximab followed by 4 cycles of cyclophosphamide (CHOP-21 chemotherapy) 750 mg/m2 intravenously (IV) day (d) 1, doxorubicin 50 mg/m2 IV d1, vincristine 1.4 mg/m2 (max. 2 mg) IV d1, and prednisone 50 mg/m2 PO d1-5, every 21 days).12 The trial of risk-stratified sequential treatment (RSST, n = 152) introduced rituximab consolidation for patients with a complete response (CR) after rituximab induction (low-risk group) and rituximab and CHOP-21 (R-CHOP-21; high-risk group) for all other patients. It demonstrated in 2016 that a CR to rituximab induction identifies a group of patients with B-cell PTLD who do not need chemotherapy and that R-CHOP consolidation for all others is safe and effective.13

Our goal was to investigate the effect of immunosuppression on relapse risk in PTLD utilizing the unique opportunity provided by the large number of patients treated uniformly in the prospective PTLD-1 trials.

MATERIALS AND METHODS

Study Design and Patients

This retrospective analysis is based on the 222 patients treated in the sequential treatment and RSST PTLD-1 trials performed by the German PTLD study group and the European PTLD Network.12,13 One hundred fifty-nine patients with complete information on immunosuppression before and after diagnosis of PTLD were included in this analysis. Both trials were approved by the appropriate Ethics committees, and all patients gave informed consent according to the Declaration of Helsinki. Disease stage at enrolment was determined through a complete patient history, physical examination, laboratory investigations, bone marrow biopsy, and computed tomography (CT) scans of the head, chest and abdomen.12,13

Treatment

Treatment-naive adult solid organ transplant recipients diagnosed with CD20-positive PTLD who had failed to respond to upfront immunosuppression reduction had either received sequential treatment (38 patients) or RSST (121 patients) according to the PTLD-1 trial protocol and its 2007 amendment.12,13 All started with rituximab (375 mg/m2 IV) on days 1, 8, 15, and 22, followed by interim staging by CT (days 40-50). In the sequential treatment protocol, all patients received 4 cycles of CHOP (cyclophosphamide 750 mg/m2 IV day (d) 1, doxorubicin 50 mg/m2 IV d1, vincristine 1.4 mg/m2 (max. 2 mg) IV d1, and prednisone 50 mg/m2 PO d1-5) every 3 weeks starting from day 50.12 In RSST, patients with a CR at interim staging continued with 4 courses of rituximab monotherapy (375 mg/m2 IV) every 3 weeks, whereas all others received 4 cycles of R-CHOP-21 (rituximab 375 mg/m2 IV d1, cyclophosphamide 750 mg/m2 IV d1, doxorubicin 50 mg/m2 IV d1, vincristine 1.4 mg/m2 (max. 2 mg) IV d1, and prednisone 50 mg/m2 PO d1-5, every 21 days). Twenty-nine patients received rituximab consolidation according to protocol.13 The final response assessment was performed 1 month after the last cycle of therapy and subsequently patients completed follow-up examinations every 3 months for 2 years, every 6 months for years 3 to 5, and annually thereafter. Final response and follow-up assessment included a complete patient history, physical examination, laboratory investigations, and CT scans of the chest and abdomen.12,13 Follow-up data were evaluated up to June 2011 for the sequential treatment trial and July 2015 for the RSST trial. Median follow-up was 5.0 years.

Immunosuppression

Doses of immunosuppressive drugs were collected as part of the baseline patient documentation in the PTLD-1 trials. Immunosuppression before the diagnosis of PTLD was defined as immunosuppression at the last routine clinic visit before suspicion of PTLD; immunosuppression after diagnosis were the doses at the start of rituximab treatment, after reduction of immunosuppression. Median time from histopathological diagnosis of PTLD to start of trial treatment was 15 days. Failure to respond to upfront immunosuppression reduction was an inclusion criterion in the PTLD-1 trials. The recommendations for patients enrolled in these trials were to stop antimetabolites (azathioprine and mycophenolate) and reduce calcineurin inhibitors by 30% to 50% while maintaining corticosteroids, if feasible.12,13 Data on immunosuppression induction was not collected and is not analyzed here. Median time from transplantation to PTLD was 7.8 years.

Statistical Analysis

Response to treatment and disease progression were classified according to World Health Organization criteria based on CT imaging. Time to progression (TTP) was defined from start of treatment to disease progression (all patients). To evaluate the effect of immunosuppression after diagnosis of PTLD, we performed a landmark TTP analysis with time zero defined as 1 year after start of treatment to minimize the effect of PTLD treatment and immunosuppression prior to diagnosis of PTLD. Landmark TTP was defined from 365 days after start of treatment to disease progression. The landmark analysis included all 112 patients without evidence of progression at 365 days after start of treatment. Time-to-event outcomes were described using Kaplan-Meier statistics. Exploratory analyses were performed using 2-sided stratified log-rank tests as well as the χ2 test for categorical variables and the Independent-Samples Kruskal-Wallis Test for continuous variables. For paired observations, the McNemar test for proportions and the Wilcoxon matched pairs test for continuous variables were used. Multivariable analyses were performed with Cox regression models (log-rank ratio test, step-wise forwards). The 2-sided significance level was set at 0.05 and statistical tests were performed using IBM SPSS 22.0.0.0.

RESULTS

Patients

The baseline characteristics of the 159 patients included in this analysis are summarized in Table 1. Median age was 56.4 years (range, 18-82). Seventy-six of 159 patients were kidney, 39 liver, 17 lung, 17 heart, 6 kidney/pancreas, and 4 heart/kidney transplant recipients. Median time from transplantation to PTLD was 7.8 years. Most cases (118/159, 74%) were of the diffuse large B cell type, 65 (43%) of 152 PTLD were Epstein-Barr virus (EBV)-associated, and 112 (70%) of 159 patients were Ann Arbor stage III or IV. Ninety-seven (61%) of 158 patients had an elevated serum lactate dehydrogenase (LDH) activity at diagnosis and 64 (41%) of 158 patients an international prognostic index score of 3 or higher (risk factors are age > 60 years, Ann Arbor stage ≥ III, Eastern Oncology Group (ECOG) performance status of 2 or greater, elevated LDH, and more than 1 extranodal disease manifestation).14 Four cases were reclassified with a diagnosis other than CD20-positive PTLD on pathology review.13

TABLE 1
TABLE 1:
Baseline patient characteristics (n = 159)

Immunosuppression Before and After Diagnosis of PTLD

Maintenance immunosuppression before and after diagnosis of PTLD is listed by drug class and individual drug in Table 2, including median doses and dose ranges. Before diagnosis of PTLD, 105 (66%) of 159 patients received an antimetabolite (70 mycophenolate, 35 azathioprine), 141 (89%) of 159 patients received a CNI (50 ciclosporin, 91 tacrolimus), 14 (9%) of 159 patients received an mTOR inhibitor (5 everolimus, 9 rapamycin), and 85 (54%) of 159, a corticosteroid. After diagnosis of PTLD, significantly fewer patients received an antimetabolite (42/159 [26%] patients, P < 0.001) or a CNI (121/159 [76%] patients, P < 0.001). No patient switched from either azathioprine or mycophenolate to the other antimetabolite or from either ciclosporin or tacrolimus to the other CNI. In those patients who continued a CNI, the doses were significantly lower (P = 0.003 for ciclosporin and P < 0.001 for tacrolimus). Median mycophenolate doses were also lower (P = 0.002). In contrast, we observed a significant increase in the use of both mTOR inhibitors (23/159 [14%] patients, P = 0.049) and corticosteroids (94/159 [59%] patients, P = 0.035), with higher median doses of corticosteroids. The combinations of immunosuppressive drugs used before and after diagnosis of PTLD are listed in Table 3. Before diagnosis of PTLD, 54 (34%) of 159 patients received triple-agent immunosuppression, 78 (49%) of 159 double-agent immunosuppression, and only 27 (17%) of 159 single-agent immunosuppression. The majority of patients on single-agent immunosuppression (23/27) received CNI monotherapy. The most common dual-agent combination was a CNI and an antimetabolite (42/159 [26%] patients). The most common triple-agent combination and most common overall drug combination (47/159 [30%] patients) was a CNI, an antimetabolite, and a corticosteroid. After diagnosis of PTLD, the use of triple-agent immunosuppression decreased to 12 (8%) of 159 patients, whereas the use of double-agent (98/159 [62%]) and single-agent (48/159 [30%]) immunosuppressions increased. The use of CNI and mTOR inhibitors monotherapy increased, and 8 patients started corticosteroid monotherapy. The use of CNI/antimetabolite double-agent immunosuppression decreased (18/159, 11%, P < 0.001). The combination of CNI and corticosteroid increased in frequency (58/159, 37%, P < 0.001) and became the most common drug combination over all.

TABLE 2
TABLE 2:
Comparison of immunosuppression before and after diagnosis of PTLD by drug class and individual drug
TABLE 3
TABLE 3:
Maintenance immunosuppression before and after diagnosis of PTLD by drug combination (n = 159), P values are shown for McNemar tests. Drug combinations received by if less than 10% of patients both before and after diagnosis are excludeda

Clinical Baseline Patient Characteristics by Immunosuppression Before Diagnosis

We observed significant differences in clinical baseline factors depending on the use of different classes of immunosuppressive drugs before diagnosis of PTLD. Among the 105 patients on antimetabolites, the transplanted organs were differently distributed (P = 0.001) compared with the remaining 54 patients, with kidney transplant recipients overrepresented and liver transplant recipients underrepresented. Furthermore, EBV-associated PTLD was more common (50% vs 29%, P = 0.012), while advanced Ann Arbor stage was less common (65% vs 81%, P = 0.029). One hundred forty-one patients received CNIs, and their baseline characteristics were very similar to the overall cohort. Nonetheless, we observed highly significant differences compared to the 18 patients not on CNIs at diagnosis of PTLD: a shorter median time from transplantation (7.2 years vs 11.9 years, P = 0.005), more common advanced Ann Arbor stage (75% vs 33%, P < 0.001) and more frequent nodal disease (77% vs 39%, P = 0.001). In the 14 patients on mTOR inhibitors at diagnosis, early PTLD was significantly more common than in the remaining 145 patients (43% vs 19%, P = 0.040). In addition, nodal disease manifestations were less frequent (43% vs 76%, P = 0.008). We noted most significant differences in the comparison of the 85 patients on corticosteroid-containing immunosuppression before diagnosis of PTLD with the 74 other patients. These included a lower age at diagnosis (median, 51.7 years vs 59.0 years, P = 0.024), a shorter time from transplantation to PTLD (median, 4.0 years vs 9.8 years, P < 0.001), a higher proportion of patients with early PTLD (35% vs 6%, P < 0.001), a different distribution of transplanted organs (overrepresentation of kidney and lung transplant recipients, P < 0.001), more polymorphic and early lesion PTLD (17% vs 5%, P = 0.026), a higher proportion of EBV-positive disease (55% vs 29%, P = 0.001) as well as a higher proportion of patients with elevated serum LDH at diagnosis (69% vs 53%, P = 0.035). Furthermore, we compared the baseline characteristics of patients receiving the 2 most commonly used classes of immunosuppressive drugs: antimetabolites and CNIs. The only significant differences between the 2 antimetabolites, azathioprine and mycophenolate, were the distribution of transplanted organs (P = 0.022) and the median year of transplantation (1998 for azathioprine, 2002 for mycophenolate, P = 0.026). The comparison of the 2 CNIs ciclosporin and tacrolimus showed that patients on tacrolimus were younger at diagnosis of PTLD (median, 51.8 years vs 60.7 years, P = 0.003), had a shorter time from transplant to PTLD (median, 5.2 years vs 9.3 years, P < 0.001), a later median year of transplantation (2004 vs 1997, P < 0.001), a different distribution of transplanted organs (liver and lung overrepresented, P = 0.029), and a higher proportion of patients with an ECOG performance status of 2 or higher (36% vs 16%, P = 0.015) compared with patients on ciclosporin.

Immunosuppression Before Diagnosis and TTP

We investigated the effect of immunosuppression before the diagnosis of PTLD on PTLD outcome measured by TTP. Overall, there were 46 progression events. There were no significant TTP differences comparing single-, double-. and triple-agent immunosuppression (P = 0.754). Neither could we identify TTP differences between patients who received a given class of immunosuppressant compared to those who did not. This was the case for antimetabolites (3-year TTP, 72.0% [95% CI, 63%-81%] vs 62.7% [95% CI, 48%-77%], P = 0.181), CNIs (3-year TTP, 68.2% [95% CI, 60%-77%] vs 74.3% [95% CI, 52%-97%], P = 0.652), mTOR inhibitors (3-year TTP 64.1% [95% CI, 34%-94%] vs 69.5% [95% CI, 61%-78%], P = 0.581), and corticosteroids (3-year TTP 62.8% [95% CI, 51%-74%] vs 76.4% [95% CI, 66%-87%], P = 0.107).

Effect of Immunosuppression After Diagnosis on the Risk of Relapse—Landmark Analysis

To examine the effect of immunosuppression after the diagnosis of PTLD on relapse risk, we performed a landmark analysis of TTP and defined time zero as 1 year after start of treatment. We thus included all 112 patients who were without evidence of PTLD progression at this time point. 17 progression events occurred in this group. In this way, we intended to minimize the effect of PTLD treatment and immunosuppression prior to diagnosis of PTLD. The 34 patients on immunosuppression containing an antimetabolite had very similar landmark TTP with a 3-year estimate of 86.2% (95% CI, 73%-99%) compared with the other 78 patients (85.2%; 95% CI, 77%-94%; P = 0.912; Figure 1A). This was also the case in a subgroup analysis limited to patients with monomorphic PTLD (n = 99, P = 0.954). There was no significant difference in landmark TTP between the 6 patients on azathioprine and the 28 patients on mycophenolate (P = 0.844). Similarly, there was no significant landmark TTP difference between the 43 patients initially on antimetabolite-containing immunosuppression who had stopped the antimetabolite and the 34 who had not (P = 0.490). Neither was there a significant difference in landmark TTP between the 81 patients receiving a CNI (3-year estimate, 87.0%; 95% CI, 79%-95%) and the 31 patients not receiving a CNI (3-year TTP, 82.0%; 95% CI, 67%-97%,; P = 0.186, Figure 1B). In a direct comparison of patients receiving a CNI after diagnosis of PTLD, the 51 patients on tacrolimus had significantly better landmark TTP than the 30 patients on ciclosporin (3-year estimate 97.0% [95% CI, 91%-100%] vs 69.8% [95% CI, 52%-87%], P = 0.002, Figure 1C). Landmark TTP in the 15 patients who had stopped CNI-containing immunosuppression was not significantly different from landmark TTP in the 81 patients who continued CNI-containing immunosuppression (P = 0.147). In addition, we found no significant differences in landmark TTP comparing grouped dose changes in patients on tacrolimus and ciclosporin, respectively (Figure S1, SDC,http://links.lww.com/TP/B579). For the 16 patients on immunosuppression containing an mTOR inhibitor, there was no significant difference in landmark TTP versus all others (3-year TTP 93.3% [95% CI, 81%-100%] vs 84.7% [95% CI, 77%-92%], P = 0.379, Figure 1D). Finally, the 65 patients on immunosuppression containing a corticosteroid after diagnosis of PTLD had a significantly poorer landmark TTP compared with those on steroid-free immunosuppression (3-year TTP 78.7% [95% CI, 68%-90%] vs 95.4% [95% CI, 89%-100%], P = 0.010, Figure 1E). To exclude effects caused by either remaining tumor mass, high corticosteroid doses, or PTLD subtype, we repeated this analysis in 3 specific subgroups: first, those with a complete or partial remission 1 year after start of treatment (n = 110; P = 0.010); second, those with a corticosteroid dose of ≤ 20 mg prednisolone equivalent per day (n = 100, P = 0.010); and third, those with monomorphic PTLD (n = 99, P = 0.004). All subgroup analyses had similar results to the complete patient cohort. The groups of patients who started or stopped either mTOR inhibitors or corticosteroids were too small (n ≤ 8) for meaningful landmark TTP analysis.

FIGURE 1
FIGURE 1:
Landmark analysis of TTP in 112 patients without evidence of progression 1 year after start of treatment by immunosuppression after diagnosis of PTLD. A, Landmark TTP by antimetabolite-containing immunosuppression (P = 0.912). Patients on antimetabolites (n = 34) are denoted by a solid line, all others (n = 78) by a dashed line. B, Landmark TTP by CNI-containing immunnosuppression (P = 0.186). Patients on CNIs (n = 81) are denoted by a solid line, all others (n = 31) by a dashed line. C, Landmark TTP by immunosuppression containing tacrolimus vs containing ciclosporin (P = 0.002). Patients on tacrolimus (n = 51) are denoted by a solid line, those on ciclosporin (n = 30) by a dashed line. D, Landmark TTP by immunosuppression containing an mTOR inhibitor (P = 0.379). Patients on mTOR inhibitors (n = 16) are denoted by a solid line, all others (n = 96) by a dashed line. E, Landmark TTP by corticosteroid-containing immunosuppression (P = 0.010). Patients on corticosteroids (n = 65) are denoted by a solid line, all others (n = 47) by a dashed line.

Cox Regression Analysis of Landmark TTP

To examine our significant findings on landmark TTP in the context of potentially confounding clinical baseline characteristics, we performed multivariable Cox regression analyses. Patient characteristics of the 112 patients included in the landmark analysis according to antimetabolite- or corticosteroid-containing immunosuppression after diagnosis of PTLD are listed in Table 4. Significant differences between patients on antimetabolite-containing immunosuppression after diagnosis of PTLD (n = 34) and the remaining patients (n = 78) were the transplanted organ (kidney transplant recipients overrepresented and liver, lung and heart transplant recipients underrepresented), a longer time from transplant to PTLD as well as a lower proportion of early PTLD and a lower proportion of patients with Ann Arbor Stage III/IV and elevated LDH at diagnosis. A confirmatory Cox regression analysis of landmark TTP including the risk factors antimetabolite-containing immunosuppression after diagnosis of PTLD as well as these significantly different baseline parameters is provided in Table S1 (SDC,http://links.lww.com/TP/B579). Significant differences between patients on corticosteroid-containing immunosuppression after diagnosis of PTLD (n = 65) and the remaining patients (n = 47) were a lower age, a lower proportion of male patients, the transplanted organ (kidney and lung transplant recipients overrepresented and liver transplant recipients underrepresented), a shorter time from transplant to PTLD, a higher proportion of early PTLD and a higher rate of EBV association. A confirmatory Cox regression analysis of Landmark TTP including the risk factors corticosteroid-containing immunosuppression after diagnosis of PTLD as well as these significantly different baseline parameters is provided in Table S2 (SDC,http://links.lww.com/TP/B579). Here, we included the baseline factors age at diagnosis, transplanted organ, year of transplantation, time from transplant to PTLD, histology (polymorphic/early lesion PTLD vs monomorphic PTLD), EBV association, Ann Arbor stage of disease (III/IV vs I/II), serum LDH activity at diagnosis (elevated vs normal), presence of extranodal disease, and ECOG performance status (≥2 vs <2) as well as corticosteroid-containing and ciclosporin-containing immunosuppression after diagnosis of PTLD in a multivariable log-rank, forwards step-wise Cox regression analysis of landmark TTP, including 101 patients with available data for all baseline characteristics (see Table 5A). Only age at diagnosis (P = 0.001, HR, 1.076/year; 95% CI, 1.030-1.124) and corticosteroid-containing immunosuppression after diagnosis of PTLD (P = 0.002, HR 11.195, 95% CI, 2.441-51.346) were identified as independent, significant risk factors. We repeated this analysis replacing steroid-containing immunosuppression after diagnosis of PTLD with the corticosteroid dose after diagnosis of PTLD in percent of the dose before diagnosis of PTLD (Table 5B). As only half of the patients had received corticosteroids prior to transplantation and thus had available data on change in corticosteroid dose, this analysis was limited to 48 patients. Age (P = 0.007, HR 1.080/year, 95% CI, 1.022-1.143) and corticosteroid dose after diagnosis of PTLD (P = 0.008, HR 1.002/percent dose change, 95% CI, 1.000-1.003) were identified as independent, significant risk factors. Thus, an additional year of age was associated with an 8% increase in relapse risk, whereas a 5% increase in steroid dose was associated with a 1% increase in relapse risk.

TABLE 4
TABLE 4:
Characteristics of 112 patients included in the Landmark analysis by use of immunosuppression after diagnosis of PTLD
TABLE 5
TABLE 5:
Multivariable Cox-regression analyses (log-rank ratio test, forward stepwise) for landmark TTP

DISCUSSION

This is the largest analysis to date of the role of immunosuppression in PTLD characteristics and relapse risk. Most importantly, all patients included were enrolled, treated and followed up uniformly in prospective clinical trials. By design, any analysis based on immunosuppression at the time of PTLD is potentially confounded by different induction regimes as well as changes to immunosuppression over the follow-up period, due to factors, such as the development of side effects, episodes of rejection, and retransplantation. Measureable confounding factors, such as immunosuppression before PTLD and patient baseline characteristics, have been provided in the Results section. Cox regression analyses have been used to analyze the impact of patient baseline characteristics on time-to-event outcomes. Age (P = 0.001; HR, 1.076/year) was the only baseline factor identified as an independent risk factor for PTLD relapse after successful PTLD treatment with immuno(chemo)therapy. However, as the choice of immunosuppression may be determined by patient characteristics not recorded in the PTLD trials, unmeasured confounding cannot be excluded. Additional limitations include the retrospective character of this analysis and multiple testing. Therefore, the findings presented here require prospective confirmation.

The most common immunosuppressive drug regimens before diagnosis of PTLD in the PTLD-1 trials were the triple combination of CNI, an antimetabolite and a corticosteroid (30%), and the doublet of CNI and antimetabolite (26%). In keeping with available guidelines, the recommendations for patients enrolled in these trials were to stop antimetabolites and reduce CNIs by 30% to 50% while maintaining corticosteroids.9-11 These recommendations were followed: after diagnosis of PTLD, significantly fewer patients received an antimetabolite or a CNI (both P < 0.001) and in those patients who continued a CNI or mycophenolate, the doses were significantly lower. Failure to respond to immunosuppression reduction was an inclusion criterion in the PTLD-1 trials. Therefore, the effect of immunosuppression reduction on treatment response could not be evaluated. The focus of this analysis was thus the impact of different classes of maintenance immunosuppression on relapse risk after successful PTLD therapy, measured as landmark TTP starting 1 year after start of PTLD treatment.

Antimetabolites have been singled out for discontinuation in guidelines on the management of immunosuppression reduction in PTLD.9-11 To our knowledge, these “potential algorithms” (Paya et al) were based on expert consensus and current clinical practice rather than prospective clinical data.9 However, the 34 patients on immunosuppression containing an antimetabolite after diagnosis had very similar landmark TTP compared with the other 78 patients (P = 0.912). Neither could we identify a difference in landmark TTP between patients who had stopped the antimetabolite and those who had not (P = 0.490). Our data suggest that the specific step of stopping the antimetabolite does not change the risk of PTLD relapse after successful immuno(chemo)therapy.

Compared with those on steroid-free immunosuppression, patients on corticosteroids after diagnosis of PTLD were significantly younger, diagnosed earlier after transplantation, and more commonly had EBV-associated disease. However, the patients on immunosuppression containing a corticosteroid after diagnosis of PTLD had a significantly poorer landmark TTP compared with those on steroid-free immunosuppression (P = 0.010). In a Cox regression analysis of landmark TTP, corticosteroid-containing immunosuppression after diagnosis of PTLD (P = 0.002, HR 11.195, 95% CI, 2.441-51.346) was identified as an independent, significant risk factor for PTLD relapse. In addition, the corticosteroid dose after diagnosis of PTLD in percent of the dose pre-PTLD was correlated with relapse risk. The number of patients who had started or stopped corticosteroids was too small for landmark TTP analysis so that a potential benefit of stopping corticosteroids could not be analyzed. Taking into account the reported beneficial outcomes of steroid-free immunosuppression in kidney, liver, and heart transplantation, our observations can be interpreted as a reason to critically appraise corticosteroid dosing after PTLD in suitable patients.15,16

The 30 patients on ciclosporin after diagnosis of PTLD had significantly poorer landmark TTP than the 51 patients on tacrolimus (P = 0.002). However, ciclosporin use after diagnosis of PTLD was not a significant risk factor in a Cox regression analysis of landmark TTP (P = 0.053). The observed significantly lower age and shorter time to PTLD of patients on tacrolimus fits with the reported higher incidence of PTLD in patients treated with tacrolimus and an antimetabolite compared to those treated with ciclosporin and an antimetabolite but might be confounded by the shift from ciclosporin to tacrolimus as the preferred CNI over time.3 We conclude that tacrolimus can safely be continued after diagnosis of PTLD. As better renal graft function has been reported in patients maintained on reduced doses of CNIs after PTLD compared with those on corticosteroid monotherapy or CNI-free immunosuppression, maintaining CNIs after PTLD appear safe both from the perspectives of graft survival and PTLD relapse risk.17,18 Because no patient started a CNI after diagnosis of PTLD and none switched over from tacrolimus to ciclosporin or vice versa, our data neither support nor argue against a switch from ciclosporin to tacrolimus after PTLD.

For patients receiving mTOR inhibitors as maintenance immunosuppression after diagnosis of PTLD, there was no significant difference in landmark TTP compared with all other patients (P = 0.379). Numbers starting or stopping mTOR inhibitors were too small for meaningful landmark TTP analysis. Our data cannot answer the question if starting an mTOR inhibitor after PTLD provides a landmark TTP benefit.

In summary, we identified corticosteroid-containing immunosuppression after diagnosis as a risk factor associated with PTLD relapse in the PTLD-1 trial cohort. In contrast, antimetabolite-containing immunosuppression was not associated with PTLD relapse. Our findings from a retrospective analysis require prospective validation.

REFERENCES

1. Penn I, Hammond W, Brettschneider L, et al. Malignant lymphomas in transplantation patients. Transplant Proc. 1969;1:106–112.
2. Opelz G, Henderson R. Incidence of non-Hodgkin lymphoma in kidney and heart transplant recipients. Lancet. 1993;342:1514–1516.
3. Opelz G, Döhler B. Lymphomas after solid organ transplantation: a collaborative transplant study report. Am J Transplant. 2004;4:222–230.
4. Krisl JC, Doan VP. Chemotherapy and transplantation: the role of immunosuppression in malignancy and a review of antineoplastic agents in solid organ transplant recipients. Am J Transplant. 2017;17:1974–1991.
5. Sampaio MS, Cho YW, Shah T, et al. Association of immunosuppressive maintenance regimens with posttransplant lymphoproliferative disorder in kidney transplant recipients. Transplantation. 2012;93:73–81.
6. Quinlan SC, Pfeiffer RM, Morton LM, et al. Risk factors for early-onset and late-onset post-transplant lymphoproliferative disorder in kidney recipients in the United States. Am J Hematol. 2011;86:206–209.
7. Reshef R, Vardhanabhuti S, Luskin MR, et al. Reduction of immunosuppression as initial therapy for posttransplantation lymphoproliferative disorder. Am J Transplant. 2011;11:336–347.
8. Swinnen LJ, LeBlanc M, Grogan TM, et al. Prospective study of sequential reduction in immunosuppression, interferon alpha-2B, and chemotherapy for posttransplantation lymphoproliferative disorder. Transplantation. 2008;86:215–222.
9. Paya CV, Fung JJ, Nalesnik MA, et al. Epstein-Barr virus-induced posttransplant lymphoproliferative disorders. ASTS/ASTP EBV-PTLD Task Force and The Mayo Clinic Organized International Consensus Development Meeting. Transplantation. 1999;68:1517–1525.
10. EBPG Expert Group on Renal Transplantation. European best practice guidelines for renal transplantation. Section IV: Long-term management of the transplant recipient. IV.6.1. Cancer risk after renal transplantation. Post-transplant lymphoproliferative disease (PTLD): prevention and treatment. Nephrol Dial Transplant. 2002;17(Suppl 4):31–33, 35–36.
11. Parker A, Bowles K, Bradley JA, et al. Management of post-transplant lymphoproliferative disorder in adult solid organ transplant recipients—BCSH and BTS guidelines. Br J Haematol. 2010;149:693–705.
12. Trappe R, Oertel S, Leblond V, et al. Sequential treatment with rituximab followed by CHOP chemotherapy in adult B-cell post-transplant lymphoproliferative disorder (PTLD): the prospective international multicentre phase 2 PTLD-1 trial. Lancet Oncol. 2012;13:196–206.
13. Trappe RU, Dierickx D, Zimmermann H, et al. Response to rituximab induction is a predictive marker in B-cell post-transplant lymphoproliferative disorder and allows successful stratification into rituximab or R-CHOP consolidation in an international, prospective, multicenter phase II trial. J Clin Oncol. 2017;35:536–543.
14. International Non-Hodgkin's Lymphoma Prognostic Factors Project. A predictive model for aggressive non-Hodgkin's lymphoma. N Engl J Med. 1993;329:987–994.
15. Opelz G, Döhler B, Laux G; Collaborative Transplant Study. Long-term prospective study of steroid withdrawal in kidney and heart transplant recipients. Am J Transplant. 2005;5:720–728.
16. Lerut JP, Pinheiro RS, Lai Q, et al. Is minimal, [almost] steroid-free immunosuppression a safe approach in adult liver transplantation? Long-term outcome of a prospective, double blind, placebo-controlled, randomized, investigator-driven study. Ann Surg. 2014;260:886–891.
17. Serre J-E, Michonneau D, Bachy E, et al. Maintaining calcineurin inhibition after the diagnosis of post-transplant lymphoproliferative disorder improves renal graft survival. Kidney Int. 2014;85:182–190.
18. Rabot N, Büchler M, Foucher Y, et al. CNI withdrawal for post-transplant lymphoproliferative disorders in kidney transplant is an independent risk factor for graft failure and mortality. Transpl Int. 2014;27:956–965.

Supplemental Digital Content

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.