- Email: [email protected]
Mycophenolate Mofetil and Risk of Developing Malignancy After Orthotopic Heart Transplantation: Analysis of the Transplant Registry of the International Society for Heart and Lung Transplantation
CLINICAL HEART TRANSPLANTATION
Mycophenolate Mofetil and Risk of Developing Malignancy After Orthotopic Heart Transplantation: Analysis of the Transplant Registry of the International Society for Heart and Lung Transplantation James O. O’Neill, MB, BSc, Leah B. Edwards, PhD, and David O. Taylor, MD Background: Malignancy after organ transplantation has been described as the “price of immunotherapy.” Evolving strategies aimed at effective immunosuppression could have differing effects on the likelihood of developing malignancy. We analyzed data from the transplant registry of the International Society for Heart and Lung Transplantation (ISHLT) to ascertain which factors are associated with the development of malignancy after orthotopic heart transplantation (OHT). Methods: Multivariate modeling was performed to determine factors predictive of first post-transplant malignancy in patients taking standard immunosuppressive regimens, defined as cyclosporine or tacrolimus and azathioprine or mycophenolate mofetil (MMF), who underwent OHT between January 1, 1995 and December 31, 1997. Results: Of the 3,895 transplants described in the cohort, 703 (18%) developed post-transplant malignancy at any time during the follow-up period, and 549 (14%) developed malignancy within the first 5 years post-transplant. The breakdown of malignancy was as follows: skin: 47%; post-transplant lymphoproliferative disease: 10%; other malignancies: 24%; combination of types: 10%; and unreported: 10%. Multivariate modeling revealed that independent predictors of increased risk were prior malignancy and increased age, whereas the use of MMF as part of a standard immunosuppressive regimen was associated with an adjusted relative risk (RR) ⫽ 0.73 (95% confidence interval 0.56 to 0.95). Relative to a recipient age of 55 years, the risk of malignancy for 30, 45 and 60 years of age was 0.32, 0.46 and 1.37, respectively. Although the use of tacrolimus appeared protective in the univariate analysis, it was not significant according to multivariate analysis. Female gender appeared to be protective. Neither OKT3 nor antithymocyte globulin (ATG) use was associated with a significantly increased risk of malignancy. Conclusions: The choice of immunosuppressive regimen may affect the likelihood of developing malignancy after OHT. Induction immunosuppression does not appear to increase the risk of subsequent malignancy. The use of MMF in standard immunosuppressive regimens is associated with a significantly lower risk of developing malignancy. J Heart Lung Transplant 2006;25:1186 –91. Copyright © 2006 by the International Society for Heart and Lung Transplantation.
The earliest reports linking lymphoma to azathioprine use in renal transplant recipients date back to 1969.1 Subsequently, an increased risk of malignancy in transplant recipients has become recognized as “the price of
From the International Society for Heart and Lung Transplantation, Addison, Texas. Submitted January 22, 2006; revised May 22, 2006; accepted June 26, 2006. Supported by the Fulbright Commission (to J.O.). Reprint requests: David O. Taylor, MD, Department of Cardiovascular Medicine, Cleveland Clinic Foundation, Desk F25, 900 Euclid Avenue, Cleveland, OH 44195. Telephone: 216-42268. Fax: 216 444-3407. E-mail: [email protected] Copyright © 2006 by the International Society for Heart and Lung Transplantation. 1053-2498/06/$–see front matter. doi:10.1016/ j.healun.2006.06.010
1186
immunosuppression.”2 Data from the Cincinnati Transplant Tumor Registry indicate that the spectrum of malignancies in transplant recipients differs from that of the general population, with an increased incidence of skin and lymphoproliferative diseases.3 Other, de novo malignancies, including lung, prostate, colon breast and cervical cancers, are said not to occur more often in cardiac transplant recipients.4 However, there is evidence of a significant recurrence rate of prior malignancy after transplantation and rates of up to 22% have been described.5,6 The likelihood of developing recurrent malignancy after transplantation appears to be higher in cardiac than in renal transplant recipients, and may be related to the intensity of immunosuppression in the former. Current estimates are of a 2-fold increase in all neo-
The Journal of Heart and Lung Transplantation Volume 25, Number 10
O’Neill et al.
Table 1. Clinical Characteristics of Cohort (Continuous Variables) 5th 95th N Median percentile percentile Recipient age 3,895 54.0 29.0 66.0 Donor age 3,895 28.0 14.0 53.0 Recipient weight (kg) 3,701 77.0 54.0 102.0 Recipient body mass index 3,522 25.0 19.5 32.6 Donor weight (kg) 3,711 74.0 53.0 102.0 Donor BMI 3,249 24.2 18.9 32.9 Weight ratio 3,533 1.0 0.7 1.4 Body mass index ratio 2,937 1.0 0.7 1.4 Ischemia time (h) 3,722 2.8 1.4 4.5 Center volume (in prior year) 3,895 22.0 7.0 80.0
plasms and a 6-fold increase in visceral tumors.7 The increased recognition as malignancy is becoming the leading cause of death among long-term survivors of heart transplantation has prompted some authorities to suggest routine screening in these patients.8 Furthermore, in patients who develop colorectal cancer, 5-year survival is lower in transplant recipients compared with the general population, indicating the possibility of biologically more aggressive tumors or the effect of immunosuppression on the clinical progression of colorectal cancer in these patients.9 Summary presentation of the data from the International Society of Heart and Lung Transplantation (ISHLT) suggested that differing immunosuppressive regimens may have different carcinogenic potential.10 Differing immunosuppressive regimens may have varying promalignant potential, and there is significant evidence that mycophenolate mofetil may be associated with a lower risk when compared with other anti-metabolite preparations.11 Using data from the ISHLT registry, we sought to estimate the effect of various immunosuppressive drug
1187
therapy combinations on the risk of developing malignancy. METHODS Study Population All patients considered for inclusion were ⱖ18 years of age and underwent an orthotopic heart transplantation between January 1, 1995 and December 31, 1997 that was reported to the ISHLT registry. The cohort was limited to orthotopic transplant patients with at least 30-day survival and those in whom the malignancy status was known on the first follow-up form. Using these criteria resulted in 4,987 transplant recipients for analysis. As the focus of the analysis was the impact of immunosuppression on outcome, the study population was limited only to those patients receiving standard immunosuppressive regimens: cyclosporine (CyA) ⫹ azathioprine (Aza); CyA ⫹ mycophenolate mofetil (MMF); tacrolimus (Tac) ⫹ Aza; and Tac ⫹ MMF. In addition, patients who received non-standard induction therapy (e.g., combined OKT3 and anti-thymocyte globulin) were excluded. The final cohort consisted of 3,895 patients. Calculation of Event Status and Event Time Malignancy status was calculated based on all annual follow-up visits until the first unknown malignancy report or the first positive malignancy report, whichever came first. The date of malignancy diagnosis is not collected outside of the USA (and, in the USA, was not collected before 1999) so the date of the follow-up report was used as a surrogate of event time. This study includes data submitted to ISHLT through August 2003. The malignancy status was calculated based on annual follow-up visits. If all annual follow-ups during the 5-year window had a response of “yes” or “no” to the
Table 2. Clinical Characteristics of Cohort (Continuous Variables, With and Without Malignancy at 5 Years) Malignancy within 5 years No (N ⫽ 3,346)
Recipient age Donor age Recipient weight (kg) Recipient BMI Donor weight (kg) Donor BMI Weight ratio BMI ratio Ischemia time (h) Center volume (in prior year) BMI, body mass index.
n 3,346 3,346 3,171 3,016 3,197 2,802 3,035 2,526 3,190 3,346
Number missing 0 0 175 330 149 544 311 820 156 0
Median 53.0 28.0 77.0 25.1 73.0 24.1 1.0 1.0 2.8 22.0
5th percentile 27.0 14.0 53.5 19.4 52.2 18.8 0.7 0.7 1.4 7.0
Yes (N ⫽ 549) 95th percentile 66.0 53.0 102.6 32.8 102.0 32.8 1.4 1.3 4.5 80.0
n 549 549 530 506 514 447 498 411 532 549
Number missing 0 0 19 43 35 102 51 138 17 0
Median 58.0 28.0 77.0 24.9 75.8 24.7 1.0 1.0 2.8 20.0
5th percentile 43.0 14.0 56.0 19.7 55.0 19.5 0.7 0.7 1.5 6.0
95th percentile 67.0 53.0 99.1 31.2 104.3 32.9 1.4 1.4 4.3 74.0
1188
O’Neill et al.
The Journal of Heart and Lung Transplantation October 2006
Table 3. Clinical Characteristics of Cohort (Categoric Variables) Malignancy within 5 years No (N ⫽ 3,346) Year of transplant 1995 1996 1997 Diagnosis group Cardiomyopathy Congenital heart disease Coronary artery disease Re-transplant/graft failure Valvular heart disease Other Not reported Female recipient Female donor Recipient ABO A AB B O Donor ABO A AB B O Pre-transplant malignancy Maintenance immunosuppression Aza ⫹ CyA Aza ⫹ Tac MMF ⫹ CyA MMF ⫹ Tac Induction immunosuppression ALG/ATG OKT3 No induction
Yes (N ⫽ 549)
All (N ⫽ 3,895)
n
%
n
%
n
%
1,158 1,017 1,171
34.6 30.4 35.0
209 171 169
38.1 31.1 30.8
1,367 1,188 1,340
35.1 30.5 34.4
1,515 44 1,628 59 77 21 2 758 1,026
45.3 1.3 48.7 1.8 2.3 0.6 0.1 22.7 30.7
191
34.8
335 4 16 3
61.0 0.7 2.9 0.5
74 156
13.5 28.4
1,706 44 1,963 63 93 24 2 832 1,182
43.8 1.1 50.4 1.6 2.4 0.6 0.1 21.4 30.3
1,468 184 418 1,276
43.9 5.5 12.5 38.1
240 22 80 207
43.7 4.0 14.6 37.7
1,708 206 498 1,483
43.9 5.3 12.8 38.1
1,261 95 302 1,687 79
37.7 2.8 9.0 50.4 2.4
205 11 58 274 27
37.3 2.0 10.6 49.9 4.9
1,466 106 360 1,961 106
37.6 2.7 9.2 50.3 2.7
2,746 126 431 43
82.1 3.8 12.9 1.3
482 14 51 2
87.8 2.6 9.3 0.4
3,228 140 482 45
82.9 3.6 12.4 1.2
672 561 2,113
20.1 16.8 63.2
108 101 340
19.7 18.4 61.9
780 662 2,453
20.0 17.0 63.0
ALG, anti-lymphocyte globulin; ATG, anti-thymocyte globulin.
malignancy question, then the malignancy status was determined based on any malignancy report within the 5 years and the time was calculated based on the first report. However, if there was a response of “unknown” (with no positive malignancy report at a prior followup), then malignancy status was censored at the last follow-up with a known response. Statistical Methods Univariate freedom-from-malignancy rates were computed using the Kaplan–Meier method and compared using the log-rank test statistic. The proportional hazards method was used for multivariate analyses of post-transplant malignancy. The resulting relative risks (RRs) are calculated such that RR ⬎1 indicates there is an increased probability of developing a post-transplant
malignancy associated with the characteristic of interest when compared with the baseline; RR ⬍1 indicates a decreased probability. p-values were determined using a Wald chi-square statistic. To further confirm the strength of the associations, propensity analysis was performed. A multivariate logistic regression analysis was performed with MMF as the dependent variable. Based on the risk factors in the model, the probability of receiving MMF was estimated. The factors used in this model were center volume (⬍20 vs ⱖ20/year), year of transplant, recipient gender, recipient age, recipient weight, ABO blood group, diagnosis of coronary artery disease vs all others, cytomegalovirus seropositivity, hepatitis C seropositivity, hepatitis B surface antigen, serum creatinine, total bilirubin, diabetes, drug-treated chronic obstructive pulmonary
The Journal of Heart and Lung Transplantation Volume 25, Number 10
O’Neill et al.
1189
Table 4. Reported Malignancy at Any Time During Follow-up After Orthotopic Heart Transplantation All (N ⫽ 703) Type of malignancy Skin only PTLD/lymphoma only Other only Type not reported Lymph ⫹ other Lymph ⫹ skin Other ⫹ skin Lymph ⫹ other ⫹ skin
n 329 71 166 70 12 18 29 8
%a 47 10 24 10 2 3 4 1
Aza ⫹ CyA (N ⫽ 621) n 291 65 141 62 12 18 25 7
Aza ⫹ Tac (N ⫽ 18)
% 47 10 28 10 2 3 4 1
n 6 1 7 3 0 0 1 0
MMF ⫹ CyA (N ⫽ 62)
% 33 6 39 17 0 0 6 0
n 32 5 17 5 0 0 2 1
MMF ⫹ Tac (N ⫽ 2)
% 52 8 27 8 0 0 3 2
n 0 0 1 0 0 0 1 0
% 0 0 50 0 0 0 50 0
PTLD, post-transplant lymphoproliferative disease; Lymph, lymphoma. a Because of rounding, total ⫽ 101%.
disease (COPD) and history of pulmonary embolism. This generated a C-statistic of 0.767, indicating a moderate predictive ability in discriminating between MMF and AZA use. A second proportional hazards regression model was generated for developing malignancy during follow-up, which included all of the risk factors as for the first model, but also included the propensity score. These results did not differ substantially from the initial model. All analyses were performed using SAS (v8.02, SAS Institute, Cary, NC). RESULTS The clinical characteristics of the patients are shown in Tables 1, 2 and 3. Of the 3,895 transplants described in the cohort, 703 (18%) developed post-transplant malignancy during the follow-up period, and 549 (14%) developed malignancy within the first 5 years after transplant. The median recipient age was 54 years and 21.4% were female. The etiology of heart failure was coronary artery disease in 50.4%, and 1.6% had undergone repeat transplantation. Increasing age was associated with increased risk of malignancy, and increasing center volume appeared to be protective. Although the former is intuitive, there is no clear explanation for the latter. Maintenance immunosuppression was Aza ⫹ CyA in 82.9%, MMF ⫹ CyA in 12.4%, Aza ⫹ Tac in 3.6% and MMF ⫹ Tac in 1.2%. Standard induction therapy was used in 37% of cases. Overall, 2.7% had a pre-transplant malignancy. The types and distributions of malignancy are shown in Table 4. Of the 703 transplants with a report of malignancy at any time during the follow-up period: 329 (47%) had only skin tumors; 71 (10%) had only posttransplant lymphoproliferative disease (PTLD)/lymphoma; and 166 (24%) had only other tumors. In 70 (10%) cases, although reported, the type was not specified. A com-
bination of two or more malignancies was reported in the remaining 10% of patients. The Kaplan–Meier plot showing freedom from malignancy for the four standard maintenance regimens is shown in Figure 1. Freedom from malignancy was greatest in those receiving MMF ⫹ Tac, followed by MMF ⫹ CyA, Aza ⫹ Tac and Aza ⫹ CyA ( p ⫽ 0.02, log-rank test). A comparison of Tac vs CyA indicated no significant difference in survival with malignancy ( p ⫽ 0.13). Freedom from malignancy was significantly higher with MMF compared with Aza ( p ⫽ 0.006, log-rank test; Figure 2). The use of induction therapy had no effect on the development of malignancy ( p ⫽ 0.39). Table 5 shows results from the multivariate proportional hazards analysis for the development of malignancy within the first 5 years after transplantation. MMF use was associated with a hazard ratio of 0.73 (95% confidence interval 0.56 to 0.95, p ⫽ 0.02) for the development of a malignancy during follow-up. A history of pre-transplant malignancy conferred an RR ⫽ 1.55 (95% confidence interval 1.1 to 2.19, p ⫽ 0.01) for Freedom from malignancy (%)
100 90 p = 0.02
80 70 60 50 0
6
12
18
24
30
36
42
48
54
60
Time (months)
AZA + CYC (N = 3228) MMF + CYC (N = 482)
AZA + Tac (N = 140) MMF + Tac (N = 45)
Figure 1. Kaplan–Meier plot of freedom from malignancy, stratified according to maintenance immunosuppression at discharge.
1190
O’Neill et al.
The Journal of Heart and Lung Transplantation October 2006
We sought to determine whether differing regimens affected the propensity to develop specific malignancies. In the case of skin malignancy, multivariate modeling indicated a trend toward a reduced incidence of this condition in patients taking MMF compared with Aza (adjusted RR ⫽ 0.73, 95% confidence interval 0.25 to 1.03, p ⫽ 0.09). For the sub-group with PTLD/ lymphoma, the risk was again found to be marginally significant (adjusted RR ⫽ 0.44, 95% confidence interval 0.19 to 1.01, p ⫽ 0.054). Maintenance steroid therapy was administered to all but 15 patients in this cohort, none of whom developed malignancy.
Freedom from malignancy (%)
100 90 p = 0.006
80 70 60 50 0
6
12
18
24
30
36
42
48
54
60
Time (months)
MMF (N = 527)
Azathioprine (N = 3368)
Figure 2. Kaplan–Meier plot of freedom from malignancy, stratified according to maintenance immunosuppression at discharge.
DISCUSSION
developing malignancy. Female gender was associated with a reduced risk (RR ⫽ 0.68, 95% confidence interval 0.54 to 0.86, p ⫽ 0.001). Blood group B was also associated with a reduced risk of malignancy (RR ⫽ 0.73, 95% confidence interval 0.56 to 0.95, p ⫽ 0.02). To confirm the results of the multivariate analysis, the calculated propensity score was included as a risk factor for the development of malignancy. Similar results were obtained, except that blood group no longer remained significant (Table 5).
Post-transplant malignancy, mainly skin cancer and PTLD/ lymphoma, remains a significant challenge after cardiac transplantation. In this series, it occurred in 18% of patients during a median follow-up of 2,175 days (the 10th percentile: 725 days; 25th percentile: 1,744 days). Thus, 75% of the cohort had at least 4 years and 9 months of follow-up. This is similar to the incidence of 27% at 7 years observed by the ISHLT registry.10 In almost 50% of the transplants initially considered for this study, malignancy status at 1 year was not reported. There may be a tendency for reporting bias, however.
Table 5. Multivariate Analysis for Development of Malignancy Over 5 Years, With and Without Incorporation of Propensity Score Outcome: all malignancies combined Multivariate analysis Risk factor Mycophenolate mofetil Tacrolimus Propensity score Pre-transplant malignancy Female recipient Female donor OKT3 ALG/ATG Tx year 1996 Tx year 1997 Congenital or other diagnosis Cardiomyopathy Re-transplant Valvular heart disease ABO ⫽ O ABO ⫽ B ABO ⫽ AB Age Center volume BMI Donor BMI
Multivariate analysis with propensity score
p-value 0.02 0.16
RR 0.73 0.73
95% confidence limits for RR 0.56 0.95 0.46 1.14
0.01 0.001 0.49 0.78 0.39 0.15 0.31 0.72 0.47 0.57 0.40 0.28 0.02 0.16 ⬍0.0001 0.027 0.17 0.12
1.55 0.68 0.93 1.03 0.92 0.43 0.92 0.86 1.17 1.05 1.11 0.81 0.73 0.73
1.1 0.54 0.77 0.86 0.75 0.14 0.78 0.38 0.76 0.89 0.88 0.55 0.56 0.46
2.19 0.86 1.13 1.23 1.12 1.36 1.08 1.95 1.82 1.24 1.4 1.19 0.95 1.14
(continuous)
RR, relative risk; ALG, anti-lymphocyte globulin; ATG, anti-thymocyte globulin; BMI, body mass index.
p-value 0.02 0.19 0.24 0.009 0.0003 0.34 0.96 0.54 0.13 0.09 0.12 0.12 0.62 0.71 0.77 0.60 0.28 ⬍0.0001 0.017 0.11 0.13
RR 0.72 0.74 1.58 0.64 0.92 1.01 0.94 0.62 0.43 0.41 0.87 0.81 1.09 1.03 1.07 0.81
95% confidence limits for RR 0.55 0.95 0.47 1.16 (continuous) 1.12 2.23 0.50 0.82 0.78 1.09 0.83 1.23 0.78 1.14 0.33 1.15 0.16 1.13 0.13 1.28 0.73 1.04 0.36 1.84 0.70 1.70 0.87 1.21 0.84 1.35 0.55 1.19 (continuous)
The Journal of Heart and Lung Transplantation Volume 25, Number 10
This factor may help to explain the apparently higher risk of malignancy seen here compared with other studies, a finding which varies widely. The cited risk was found to vary from 3.5% at 3 years to a 7% incidence of lymphoproliferative disease at 7 years.12,13 In the clinical trial setting, 2 of 46 patients in an angiographic study of rapamycin developed solid-organ tumors, during a follow-up period of ⬍2 years.14 In a large, multicenter, randomized, controlled trial comparing MMF and Aza over 36 months, with 289 patients in each group, the combined end-point of death or repeat transplantation was significantly higher with Aza (18.3% vs 11.8%, p ⬍ 0.01).15 Post-transplant malignancy was reported in 15.6% of the Aza group and 12.5% of MMF-treated patients (p ⫽ NS). Although it did not reach statistical significance, that study supports our findings. We identified increasing age, male gender and a prior malignancy as important and powerful indicators of risk for post-transplant malignancy. These factors are, however, unmodifiable. On the other hand, MMF was associated with a 27% adjusted risk reduction for the development of any malignancy during follow-up. The choice of immunosuppressive regimen is therefore an important modifiable risk factor, and the present data appear to favor MMF over azathioprine. There are several limitations of our study. This was not a randomized, controlled trial. Despite the use of multivariate analysis, there may have been significant, unrecorded differences between patients receiving MMF and those receiving Aza. In an attempt to overcome this, a propensity score was generated to predict the use of MMF in patients. Incorporation of the propensity score in the multivariate analysis did not materially affect our results. Another concern is the broad classification of malignancy as “other” in 24% and “not reported” in 3%. As malignancy is being increasingly recognized as a limiting factor in the survival of transplant recipients, modification of the future data set to include a more accurate breakdown of malignancy types would appear necessary. Other co-factors, specifically nicotine use and exposure to ultraviolet radiation, also merit consideration for inclusion, although the latter would be particularly difficult to quantify. In addition, we excluded a large number of transplants—those not receiving conventional drug regimens and those in whom the first annual report did not report on the presence or absence of malignancy. This could theoretically have limited our ability to generalize to the larger transplant population. Against this, however, is the fact that the clinical characteristics of the
O’Neill et al.
1191
patients excluded in our analysis were similar to those included. Finally, the period of our study, 1995 to 1997 (inclusive), was at a relatively early stage in the use of MMF in cardiac transplant recipients. It is possible that the profile of patients receiving MMF may now be different. In addition, we considered malignancy only, and did not analyze other potential complications of MMF, including viral infections, leukopenia and severe gastrointestinal side effects, which can limit tolerability of this drug. Despite these limitations, the data presented herein indicate that MMF is associated with a significantly lower risk of developing post-transplant malignancy and is the preferred agent compared with azathioprine in this setting. REFERENCES 1. Penn I, Hammond W, Bretschneider L, Starzl TE. Malignant lymphomas in transplantation patients. Transplant Proc 1969;1: 106 –12. 2. Penn I. The price of immunotherapy. Curr Probl Surg 1981;18: 681–751. 3. Penn I. Posttransplant malignancies. Transplant Proc 1999;31: 1260 –2. 4. Hunt SA. Malignancy in organ transplantation: heart. Transplant Proc 2002;34:1874 – 6. 5. Penn I. The effect of immunosuppression on pre-existing cancers. Transplantation 1993;55:742–7. 6. Penn I. Effect of immunosuppression on preexisting cancers. Transplant Proc 1993;25:1380 –2. 7. Ippoliti G, Rinaldi M, Pellegrini C, Vigano M. Incidence of cancer after immunosuppressive treatment for heart transplantation. Crit Rev Oncol/Hematol 2005;56:101–13. 8. Haupton P, Medra MR. It is time to stop ignoring malignancy in heart transplantation: a call to arms. J Heart Lung Transplant 2005;24:1111–3. 9. Buell JF, Papaconstantinou HT, Skalow B, Hanaway MJ, Alloway RR, Woodle ES. De novo colorectal cancer: five-year survival is markedly lower in transplant recipients compared with the general population. Transplant Proc 2005;37:960 –1. 10. Taylor DO, Edwards LB, Mohacsi PJ, et al. The registry of the International Society for Heart and Lung Transplantation: twentieth official adult heart transplant report—2003. J Heart Lung Transplant 2003;22:616 –24. 11. Keogh A. Long-term benefits of mycophenolate mofetil after heart transplantation. Transplantation 2005;79(suppl):S45– 6. 12. deSalvo TG, Naftel DC, Kasper EK, et al. The differing hazard of lymphoma vs. other malignancies in the current era. J Heart Lung Transplant 1998;17:70. 13. Opelz G. Are post-transplant lymphomas inevitable? Nephrol Dial Transplant 1996;11:1952–5. 14. Mancini D, Pinney S, Burkhoff D, et al. Use of rapamycin slows progression of cardiac transplantation vasculopathy. Circulation 2003;108:48 –53. 15. Eisen HJ, Kowabashiga J, Keogh A, et al. Three-year results of a randomized, double-blind, controlled trial of mycophenolate mofetil versus azathioprine in cardiac transplant recipients. J Heart Lung Transplant 2005;24:517–25.
Mycophenolate Mofetil and Risk of Developing Malignancy After Orthotopic Heart Transplantation: Analysis of the Transplant Registry of the International Society for Heart and Lung Transplantation James O. O’Neill, MB, BSc, Leah B. Edwards, PhD, and David O. Taylor, MD Background: Malignancy after organ transplantation has been described as the “price of immunotherapy.” Evolving strategies aimed at effective immunosuppression could have differing effects on the likelihood of developing malignancy. We analyzed data from the transplant registry of the International Society for Heart and Lung Transplantation (ISHLT) to ascertain which factors are associated with the development of malignancy after orthotopic heart transplantation (OHT). Methods: Multivariate modeling was performed to determine factors predictive of first post-transplant malignancy in patients taking standard immunosuppressive regimens, defined as cyclosporine or tacrolimus and azathioprine or mycophenolate mofetil (MMF), who underwent OHT between January 1, 1995 and December 31, 1997. Results: Of the 3,895 transplants described in the cohort, 703 (18%) developed post-transplant malignancy at any time during the follow-up period, and 549 (14%) developed malignancy within the first 5 years post-transplant. The breakdown of malignancy was as follows: skin: 47%; post-transplant lymphoproliferative disease: 10%; other malignancies: 24%; combination of types: 10%; and unreported: 10%. Multivariate modeling revealed that independent predictors of increased risk were prior malignancy and increased age, whereas the use of MMF as part of a standard immunosuppressive regimen was associated with an adjusted relative risk (RR) ⫽ 0.73 (95% confidence interval 0.56 to 0.95). Relative to a recipient age of 55 years, the risk of malignancy for 30, 45 and 60 years of age was 0.32, 0.46 and 1.37, respectively. Although the use of tacrolimus appeared protective in the univariate analysis, it was not significant according to multivariate analysis. Female gender appeared to be protective. Neither OKT3 nor antithymocyte globulin (ATG) use was associated with a significantly increased risk of malignancy. Conclusions: The choice of immunosuppressive regimen may affect the likelihood of developing malignancy after OHT. Induction immunosuppression does not appear to increase the risk of subsequent malignancy. The use of MMF in standard immunosuppressive regimens is associated with a significantly lower risk of developing malignancy. J Heart Lung Transplant 2006;25:1186 –91. Copyright © 2006 by the International Society for Heart and Lung Transplantation.
The earliest reports linking lymphoma to azathioprine use in renal transplant recipients date back to 1969.1 Subsequently, an increased risk of malignancy in transplant recipients has become recognized as “the price of
From the International Society for Heart and Lung Transplantation, Addison, Texas. Submitted January 22, 2006; revised May 22, 2006; accepted June 26, 2006. Supported by the Fulbright Commission (to J.O.). Reprint requests: David O. Taylor, MD, Department of Cardiovascular Medicine, Cleveland Clinic Foundation, Desk F25, 900 Euclid Avenue, Cleveland, OH 44195. Telephone: 216-42268. Fax: 216 444-3407. E-mail: [email protected] Copyright © 2006 by the International Society for Heart and Lung Transplantation. 1053-2498/06/$–see front matter. doi:10.1016/ j.healun.2006.06.010
1186
immunosuppression.”2 Data from the Cincinnati Transplant Tumor Registry indicate that the spectrum of malignancies in transplant recipients differs from that of the general population, with an increased incidence of skin and lymphoproliferative diseases.3 Other, de novo malignancies, including lung, prostate, colon breast and cervical cancers, are said not to occur more often in cardiac transplant recipients.4 However, there is evidence of a significant recurrence rate of prior malignancy after transplantation and rates of up to 22% have been described.5,6 The likelihood of developing recurrent malignancy after transplantation appears to be higher in cardiac than in renal transplant recipients, and may be related to the intensity of immunosuppression in the former. Current estimates are of a 2-fold increase in all neo-
The Journal of Heart and Lung Transplantation Volume 25, Number 10
O’Neill et al.
Table 1. Clinical Characteristics of Cohort (Continuous Variables) 5th 95th N Median percentile percentile Recipient age 3,895 54.0 29.0 66.0 Donor age 3,895 28.0 14.0 53.0 Recipient weight (kg) 3,701 77.0 54.0 102.0 Recipient body mass index 3,522 25.0 19.5 32.6 Donor weight (kg) 3,711 74.0 53.0 102.0 Donor BMI 3,249 24.2 18.9 32.9 Weight ratio 3,533 1.0 0.7 1.4 Body mass index ratio 2,937 1.0 0.7 1.4 Ischemia time (h) 3,722 2.8 1.4 4.5 Center volume (in prior year) 3,895 22.0 7.0 80.0
plasms and a 6-fold increase in visceral tumors.7 The increased recognition as malignancy is becoming the leading cause of death among long-term survivors of heart transplantation has prompted some authorities to suggest routine screening in these patients.8 Furthermore, in patients who develop colorectal cancer, 5-year survival is lower in transplant recipients compared with the general population, indicating the possibility of biologically more aggressive tumors or the effect of immunosuppression on the clinical progression of colorectal cancer in these patients.9 Summary presentation of the data from the International Society of Heart and Lung Transplantation (ISHLT) suggested that differing immunosuppressive regimens may have different carcinogenic potential.10 Differing immunosuppressive regimens may have varying promalignant potential, and there is significant evidence that mycophenolate mofetil may be associated with a lower risk when compared with other anti-metabolite preparations.11 Using data from the ISHLT registry, we sought to estimate the effect of various immunosuppressive drug
1187
therapy combinations on the risk of developing malignancy. METHODS Study Population All patients considered for inclusion were ⱖ18 years of age and underwent an orthotopic heart transplantation between January 1, 1995 and December 31, 1997 that was reported to the ISHLT registry. The cohort was limited to orthotopic transplant patients with at least 30-day survival and those in whom the malignancy status was known on the first follow-up form. Using these criteria resulted in 4,987 transplant recipients for analysis. As the focus of the analysis was the impact of immunosuppression on outcome, the study population was limited only to those patients receiving standard immunosuppressive regimens: cyclosporine (CyA) ⫹ azathioprine (Aza); CyA ⫹ mycophenolate mofetil (MMF); tacrolimus (Tac) ⫹ Aza; and Tac ⫹ MMF. In addition, patients who received non-standard induction therapy (e.g., combined OKT3 and anti-thymocyte globulin) were excluded. The final cohort consisted of 3,895 patients. Calculation of Event Status and Event Time Malignancy status was calculated based on all annual follow-up visits until the first unknown malignancy report or the first positive malignancy report, whichever came first. The date of malignancy diagnosis is not collected outside of the USA (and, in the USA, was not collected before 1999) so the date of the follow-up report was used as a surrogate of event time. This study includes data submitted to ISHLT through August 2003. The malignancy status was calculated based on annual follow-up visits. If all annual follow-ups during the 5-year window had a response of “yes” or “no” to the
Table 2. Clinical Characteristics of Cohort (Continuous Variables, With and Without Malignancy at 5 Years) Malignancy within 5 years No (N ⫽ 3,346)
Recipient age Donor age Recipient weight (kg) Recipient BMI Donor weight (kg) Donor BMI Weight ratio BMI ratio Ischemia time (h) Center volume (in prior year) BMI, body mass index.
n 3,346 3,346 3,171 3,016 3,197 2,802 3,035 2,526 3,190 3,346
Number missing 0 0 175 330 149 544 311 820 156 0
Median 53.0 28.0 77.0 25.1 73.0 24.1 1.0 1.0 2.8 22.0
5th percentile 27.0 14.0 53.5 19.4 52.2 18.8 0.7 0.7 1.4 7.0
Yes (N ⫽ 549) 95th percentile 66.0 53.0 102.6 32.8 102.0 32.8 1.4 1.3 4.5 80.0
n 549 549 530 506 514 447 498 411 532 549
Number missing 0 0 19 43 35 102 51 138 17 0
Median 58.0 28.0 77.0 24.9 75.8 24.7 1.0 1.0 2.8 20.0
5th percentile 43.0 14.0 56.0 19.7 55.0 19.5 0.7 0.7 1.5 6.0
95th percentile 67.0 53.0 99.1 31.2 104.3 32.9 1.4 1.4 4.3 74.0
1188
O’Neill et al.
The Journal of Heart and Lung Transplantation October 2006
Table 3. Clinical Characteristics of Cohort (Categoric Variables) Malignancy within 5 years No (N ⫽ 3,346) Year of transplant 1995 1996 1997 Diagnosis group Cardiomyopathy Congenital heart disease Coronary artery disease Re-transplant/graft failure Valvular heart disease Other Not reported Female recipient Female donor Recipient ABO A AB B O Donor ABO A AB B O Pre-transplant malignancy Maintenance immunosuppression Aza ⫹ CyA Aza ⫹ Tac MMF ⫹ CyA MMF ⫹ Tac Induction immunosuppression ALG/ATG OKT3 No induction
Yes (N ⫽ 549)
All (N ⫽ 3,895)
n
%
n
%
n
%
1,158 1,017 1,171
34.6 30.4 35.0
209 171 169
38.1 31.1 30.8
1,367 1,188 1,340
35.1 30.5 34.4
1,515 44 1,628 59 77 21 2 758 1,026
45.3 1.3 48.7 1.8 2.3 0.6 0.1 22.7 30.7
191
34.8
335 4 16 3
61.0 0.7 2.9 0.5
74 156
13.5 28.4
1,706 44 1,963 63 93 24 2 832 1,182
43.8 1.1 50.4 1.6 2.4 0.6 0.1 21.4 30.3
1,468 184 418 1,276
43.9 5.5 12.5 38.1
240 22 80 207
43.7 4.0 14.6 37.7
1,708 206 498 1,483
43.9 5.3 12.8 38.1
1,261 95 302 1,687 79
37.7 2.8 9.0 50.4 2.4
205 11 58 274 27
37.3 2.0 10.6 49.9 4.9
1,466 106 360 1,961 106
37.6 2.7 9.2 50.3 2.7
2,746 126 431 43
82.1 3.8 12.9 1.3
482 14 51 2
87.8 2.6 9.3 0.4
3,228 140 482 45
82.9 3.6 12.4 1.2
672 561 2,113
20.1 16.8 63.2
108 101 340
19.7 18.4 61.9
780 662 2,453
20.0 17.0 63.0
ALG, anti-lymphocyte globulin; ATG, anti-thymocyte globulin.
malignancy question, then the malignancy status was determined based on any malignancy report within the 5 years and the time was calculated based on the first report. However, if there was a response of “unknown” (with no positive malignancy report at a prior followup), then malignancy status was censored at the last follow-up with a known response. Statistical Methods Univariate freedom-from-malignancy rates were computed using the Kaplan–Meier method and compared using the log-rank test statistic. The proportional hazards method was used for multivariate analyses of post-transplant malignancy. The resulting relative risks (RRs) are calculated such that RR ⬎1 indicates there is an increased probability of developing a post-transplant
malignancy associated with the characteristic of interest when compared with the baseline; RR ⬍1 indicates a decreased probability. p-values were determined using a Wald chi-square statistic. To further confirm the strength of the associations, propensity analysis was performed. A multivariate logistic regression analysis was performed with MMF as the dependent variable. Based on the risk factors in the model, the probability of receiving MMF was estimated. The factors used in this model were center volume (⬍20 vs ⱖ20/year), year of transplant, recipient gender, recipient age, recipient weight, ABO blood group, diagnosis of coronary artery disease vs all others, cytomegalovirus seropositivity, hepatitis C seropositivity, hepatitis B surface antigen, serum creatinine, total bilirubin, diabetes, drug-treated chronic obstructive pulmonary
The Journal of Heart and Lung Transplantation Volume 25, Number 10
O’Neill et al.
1189
Table 4. Reported Malignancy at Any Time During Follow-up After Orthotopic Heart Transplantation All (N ⫽ 703) Type of malignancy Skin only PTLD/lymphoma only Other only Type not reported Lymph ⫹ other Lymph ⫹ skin Other ⫹ skin Lymph ⫹ other ⫹ skin
n 329 71 166 70 12 18 29 8
%a 47 10 24 10 2 3 4 1
Aza ⫹ CyA (N ⫽ 621) n 291 65 141 62 12 18 25 7
Aza ⫹ Tac (N ⫽ 18)
% 47 10 28 10 2 3 4 1
n 6 1 7 3 0 0 1 0
MMF ⫹ CyA (N ⫽ 62)
% 33 6 39 17 0 0 6 0
n 32 5 17 5 0 0 2 1
MMF ⫹ Tac (N ⫽ 2)
% 52 8 27 8 0 0 3 2
n 0 0 1 0 0 0 1 0
% 0 0 50 0 0 0 50 0
PTLD, post-transplant lymphoproliferative disease; Lymph, lymphoma. a Because of rounding, total ⫽ 101%.
disease (COPD) and history of pulmonary embolism. This generated a C-statistic of 0.767, indicating a moderate predictive ability in discriminating between MMF and AZA use. A second proportional hazards regression model was generated for developing malignancy during follow-up, which included all of the risk factors as for the first model, but also included the propensity score. These results did not differ substantially from the initial model. All analyses were performed using SAS (v8.02, SAS Institute, Cary, NC). RESULTS The clinical characteristics of the patients are shown in Tables 1, 2 and 3. Of the 3,895 transplants described in the cohort, 703 (18%) developed post-transplant malignancy during the follow-up period, and 549 (14%) developed malignancy within the first 5 years after transplant. The median recipient age was 54 years and 21.4% were female. The etiology of heart failure was coronary artery disease in 50.4%, and 1.6% had undergone repeat transplantation. Increasing age was associated with increased risk of malignancy, and increasing center volume appeared to be protective. Although the former is intuitive, there is no clear explanation for the latter. Maintenance immunosuppression was Aza ⫹ CyA in 82.9%, MMF ⫹ CyA in 12.4%, Aza ⫹ Tac in 3.6% and MMF ⫹ Tac in 1.2%. Standard induction therapy was used in 37% of cases. Overall, 2.7% had a pre-transplant malignancy. The types and distributions of malignancy are shown in Table 4. Of the 703 transplants with a report of malignancy at any time during the follow-up period: 329 (47%) had only skin tumors; 71 (10%) had only posttransplant lymphoproliferative disease (PTLD)/lymphoma; and 166 (24%) had only other tumors. In 70 (10%) cases, although reported, the type was not specified. A com-
bination of two or more malignancies was reported in the remaining 10% of patients. The Kaplan–Meier plot showing freedom from malignancy for the four standard maintenance regimens is shown in Figure 1. Freedom from malignancy was greatest in those receiving MMF ⫹ Tac, followed by MMF ⫹ CyA, Aza ⫹ Tac and Aza ⫹ CyA ( p ⫽ 0.02, log-rank test). A comparison of Tac vs CyA indicated no significant difference in survival with malignancy ( p ⫽ 0.13). Freedom from malignancy was significantly higher with MMF compared with Aza ( p ⫽ 0.006, log-rank test; Figure 2). The use of induction therapy had no effect on the development of malignancy ( p ⫽ 0.39). Table 5 shows results from the multivariate proportional hazards analysis for the development of malignancy within the first 5 years after transplantation. MMF use was associated with a hazard ratio of 0.73 (95% confidence interval 0.56 to 0.95, p ⫽ 0.02) for the development of a malignancy during follow-up. A history of pre-transplant malignancy conferred an RR ⫽ 1.55 (95% confidence interval 1.1 to 2.19, p ⫽ 0.01) for Freedom from malignancy (%)
100 90 p = 0.02
80 70 60 50 0
6
12
18
24
30
36
42
48
54
60
Time (months)
AZA + CYC (N = 3228) MMF + CYC (N = 482)
AZA + Tac (N = 140) MMF + Tac (N = 45)
Figure 1. Kaplan–Meier plot of freedom from malignancy, stratified according to maintenance immunosuppression at discharge.
1190
O’Neill et al.
The Journal of Heart and Lung Transplantation October 2006
We sought to determine whether differing regimens affected the propensity to develop specific malignancies. In the case of skin malignancy, multivariate modeling indicated a trend toward a reduced incidence of this condition in patients taking MMF compared with Aza (adjusted RR ⫽ 0.73, 95% confidence interval 0.25 to 1.03, p ⫽ 0.09). For the sub-group with PTLD/ lymphoma, the risk was again found to be marginally significant (adjusted RR ⫽ 0.44, 95% confidence interval 0.19 to 1.01, p ⫽ 0.054). Maintenance steroid therapy was administered to all but 15 patients in this cohort, none of whom developed malignancy.
Freedom from malignancy (%)
100 90 p = 0.006
80 70 60 50 0
6
12
18
24
30
36
42
48
54
60
Time (months)
MMF (N = 527)
Azathioprine (N = 3368)
Figure 2. Kaplan–Meier plot of freedom from malignancy, stratified according to maintenance immunosuppression at discharge.
DISCUSSION
developing malignancy. Female gender was associated with a reduced risk (RR ⫽ 0.68, 95% confidence interval 0.54 to 0.86, p ⫽ 0.001). Blood group B was also associated with a reduced risk of malignancy (RR ⫽ 0.73, 95% confidence interval 0.56 to 0.95, p ⫽ 0.02). To confirm the results of the multivariate analysis, the calculated propensity score was included as a risk factor for the development of malignancy. Similar results were obtained, except that blood group no longer remained significant (Table 5).
Post-transplant malignancy, mainly skin cancer and PTLD/ lymphoma, remains a significant challenge after cardiac transplantation. In this series, it occurred in 18% of patients during a median follow-up of 2,175 days (the 10th percentile: 725 days; 25th percentile: 1,744 days). Thus, 75% of the cohort had at least 4 years and 9 months of follow-up. This is similar to the incidence of 27% at 7 years observed by the ISHLT registry.10 In almost 50% of the transplants initially considered for this study, malignancy status at 1 year was not reported. There may be a tendency for reporting bias, however.
Table 5. Multivariate Analysis for Development of Malignancy Over 5 Years, With and Without Incorporation of Propensity Score Outcome: all malignancies combined Multivariate analysis Risk factor Mycophenolate mofetil Tacrolimus Propensity score Pre-transplant malignancy Female recipient Female donor OKT3 ALG/ATG Tx year 1996 Tx year 1997 Congenital or other diagnosis Cardiomyopathy Re-transplant Valvular heart disease ABO ⫽ O ABO ⫽ B ABO ⫽ AB Age Center volume BMI Donor BMI
Multivariate analysis with propensity score
p-value 0.02 0.16
RR 0.73 0.73
95% confidence limits for RR 0.56 0.95 0.46 1.14
0.01 0.001 0.49 0.78 0.39 0.15 0.31 0.72 0.47 0.57 0.40 0.28 0.02 0.16 ⬍0.0001 0.027 0.17 0.12
1.55 0.68 0.93 1.03 0.92 0.43 0.92 0.86 1.17 1.05 1.11 0.81 0.73 0.73
1.1 0.54 0.77 0.86 0.75 0.14 0.78 0.38 0.76 0.89 0.88 0.55 0.56 0.46
2.19 0.86 1.13 1.23 1.12 1.36 1.08 1.95 1.82 1.24 1.4 1.19 0.95 1.14
(continuous)
RR, relative risk; ALG, anti-lymphocyte globulin; ATG, anti-thymocyte globulin; BMI, body mass index.
p-value 0.02 0.19 0.24 0.009 0.0003 0.34 0.96 0.54 0.13 0.09 0.12 0.12 0.62 0.71 0.77 0.60 0.28 ⬍0.0001 0.017 0.11 0.13
RR 0.72 0.74 1.58 0.64 0.92 1.01 0.94 0.62 0.43 0.41 0.87 0.81 1.09 1.03 1.07 0.81
95% confidence limits for RR 0.55 0.95 0.47 1.16 (continuous) 1.12 2.23 0.50 0.82 0.78 1.09 0.83 1.23 0.78 1.14 0.33 1.15 0.16 1.13 0.13 1.28 0.73 1.04 0.36 1.84 0.70 1.70 0.87 1.21 0.84 1.35 0.55 1.19 (continuous)
The Journal of Heart and Lung Transplantation Volume 25, Number 10
This factor may help to explain the apparently higher risk of malignancy seen here compared with other studies, a finding which varies widely. The cited risk was found to vary from 3.5% at 3 years to a 7% incidence of lymphoproliferative disease at 7 years.12,13 In the clinical trial setting, 2 of 46 patients in an angiographic study of rapamycin developed solid-organ tumors, during a follow-up period of ⬍2 years.14 In a large, multicenter, randomized, controlled trial comparing MMF and Aza over 36 months, with 289 patients in each group, the combined end-point of death or repeat transplantation was significantly higher with Aza (18.3% vs 11.8%, p ⬍ 0.01).15 Post-transplant malignancy was reported in 15.6% of the Aza group and 12.5% of MMF-treated patients (p ⫽ NS). Although it did not reach statistical significance, that study supports our findings. We identified increasing age, male gender and a prior malignancy as important and powerful indicators of risk for post-transplant malignancy. These factors are, however, unmodifiable. On the other hand, MMF was associated with a 27% adjusted risk reduction for the development of any malignancy during follow-up. The choice of immunosuppressive regimen is therefore an important modifiable risk factor, and the present data appear to favor MMF over azathioprine. There are several limitations of our study. This was not a randomized, controlled trial. Despite the use of multivariate analysis, there may have been significant, unrecorded differences between patients receiving MMF and those receiving Aza. In an attempt to overcome this, a propensity score was generated to predict the use of MMF in patients. Incorporation of the propensity score in the multivariate analysis did not materially affect our results. Another concern is the broad classification of malignancy as “other” in 24% and “not reported” in 3%. As malignancy is being increasingly recognized as a limiting factor in the survival of transplant recipients, modification of the future data set to include a more accurate breakdown of malignancy types would appear necessary. Other co-factors, specifically nicotine use and exposure to ultraviolet radiation, also merit consideration for inclusion, although the latter would be particularly difficult to quantify. In addition, we excluded a large number of transplants—those not receiving conventional drug regimens and those in whom the first annual report did not report on the presence or absence of malignancy. This could theoretically have limited our ability to generalize to the larger transplant population. Against this, however, is the fact that the clinical characteristics of the
O’Neill et al.
1191
patients excluded in our analysis were similar to those included. Finally, the period of our study, 1995 to 1997 (inclusive), was at a relatively early stage in the use of MMF in cardiac transplant recipients. It is possible that the profile of patients receiving MMF may now be different. In addition, we considered malignancy only, and did not analyze other potential complications of MMF, including viral infections, leukopenia and severe gastrointestinal side effects, which can limit tolerability of this drug. Despite these limitations, the data presented herein indicate that MMF is associated with a significantly lower risk of developing post-transplant malignancy and is the preferred agent compared with azathioprine in this setting. REFERENCES 1. Penn I, Hammond W, Bretschneider L, Starzl TE. Malignant lymphomas in transplantation patients. Transplant Proc 1969;1: 106 –12. 2. Penn I. The price of immunotherapy. Curr Probl Surg 1981;18: 681–751. 3. Penn I. Posttransplant malignancies. Transplant Proc 1999;31: 1260 –2. 4. Hunt SA. Malignancy in organ transplantation: heart. Transplant Proc 2002;34:1874 – 6. 5. Penn I. The effect of immunosuppression on pre-existing cancers. Transplantation 1993;55:742–7. 6. Penn I. Effect of immunosuppression on preexisting cancers. Transplant Proc 1993;25:1380 –2. 7. Ippoliti G, Rinaldi M, Pellegrini C, Vigano M. Incidence of cancer after immunosuppressive treatment for heart transplantation. Crit Rev Oncol/Hematol 2005;56:101–13. 8. Haupton P, Medra MR. It is time to stop ignoring malignancy in heart transplantation: a call to arms. J Heart Lung Transplant 2005;24:1111–3. 9. Buell JF, Papaconstantinou HT, Skalow B, Hanaway MJ, Alloway RR, Woodle ES. De novo colorectal cancer: five-year survival is markedly lower in transplant recipients compared with the general population. Transplant Proc 2005;37:960 –1. 10. Taylor DO, Edwards LB, Mohacsi PJ, et al. The registry of the International Society for Heart and Lung Transplantation: twentieth official adult heart transplant report—2003. J Heart Lung Transplant 2003;22:616 –24. 11. Keogh A. Long-term benefits of mycophenolate mofetil after heart transplantation. Transplantation 2005;79(suppl):S45– 6. 12. deSalvo TG, Naftel DC, Kasper EK, et al. The differing hazard of lymphoma vs. other malignancies in the current era. J Heart Lung Transplant 1998;17:70. 13. Opelz G. Are post-transplant lymphomas inevitable? Nephrol Dial Transplant 1996;11:1952–5. 14. Mancini D, Pinney S, Burkhoff D, et al. Use of rapamycin slows progression of cardiac transplantation vasculopathy. Circulation 2003;108:48 –53. 15. Eisen HJ, Kowabashiga J, Keogh A, et al. Three-year results of a randomized, double-blind, controlled trial of mycophenolate mofetil versus azathioprine in cardiac transplant recipients. J Heart Lung Transplant 2005;24:517–25.