A history of cancer is considered a relative contraindication for solid-organ transplantation; minimum cancer remission times are usually required before a transplant candidate with a pretransplant malignancy (PTM) can be listed for transplantation.1-4 The rationale supporting this recommendation is the increased risk of cancer recurrence attributed to immunosuppression.5,6 Historical data from the Israel Penn International Transplant Tumor Registry indicated an overall rate of posttransplant recurrence in patients with PTM of 21%.5 However, more contemporary population-based registry studies have reported lower recurrence rates.7,8
We have previously reported the results of a meta-analysis of 33 cohort studies showing that recipients with PTM have higher all-cause and cancer-specific mortality than those without PTM.9 However, it is unclear whether the increased risk of all-cause mortality in recipients with PTM is explained solely by the increased cancer mortality or if other factors, such as cardiovascular disease and graft failure, also contribute. Prolonging the wait time for transplantation for potential recipients with PTM to meet minimum cancer remission times means these patients live longer with end-stage organ disease, and this may result in worse outcomes after transplantation.
Currently, 7% of all solid organ transplant recipients (SOTR) in population-based studies have a PTM, and this number is expected to increase with the expansion of eligibility criteria to older patients.9 Improving our understanding of the causes of higher mortality in this population is important. We therefore designed this population-based study to compare the outcomes of SOTR with PTM in remission to recipients without prior history of cancer, to explore the risk of cancer-specific mortality in this population, and to identify factors associated with reduced survival.
MATERIALS AND METHODS
Study Design and Data Sources
We designed a retrospective, propensity score matched, cohort study including all patients who underwent solid organ transplantation from January 1, 1991, through December 31, 2010, in the province of Ontario, Canada. The cohort was assembled using the Canadian Organ Replacement Register, a national registry that contains information on approximately 98% all Canadian SOTR since 1981.10 Non-Ontarian residents were excluded from the study population (n = 1253) because these patients could not be linked to the provincial healthcare registries and databases in Ontario. Patients who died within 30 days of transplantation were also excluded (n = 311) because these were assumed to reflect perioperative mortality, unlikely to be malignancy-related. Transplants of the same organ occurring within 1 week of each other were combined into a single episode because they likely represent duplicate records. Patients were followed up from the date of transplantation until December 31, 2012, to allow a minimum of 2 years of follow-up for all cohort members.
The Ontario Health Insurance Plan database, the Canadian Institute for Health Information Discharge Abstract Database and National Ambulatory Care Reporting System were used to identify comorbidities, metastatic disease diagnosis, and billing codes for palliative care, radiotherapy, and chemotherapy. All data sources were deterministically linked using an encrypted unique patient identifier that allowed for longitudinal tracking of patients across all data sources. Direct personal identifiers were removed, and a unique identifier was applied using an algorithm. All data sets were held securely in linkable files without any direct personal identifiers and analyzed at the Institute for Clinical Evaluative Sciences, Toronto, Canada. The Research Ethics Board of St. Michael’s Hospital, Toronto, Canada approved the study (11-235).
Exposure and Matching
To determine if patients had a PTM, we linked members of our cohort to the Ontario Cancer Registry (OCR). This registry contains information on all incident cancers (other than nonmelanoma skin cancer) since 1964 in Ontario and has been estimated to be over 95% complete.11 At the time of the study, the OCR registered the first incident cancer for a cancer site and did not capture disease recurrence. Patients who were registered in OCR as having a cancer diagnosis at least 7 days before transplantation were considered to have a PTM. However, recipients identified by billing codes as having received chemotherapy within 6 months before transplantation were not considered in remission and were excluded. Recipients with incidentally diagnosed malignancies (diagnosed 7 days before 30 days after transplantation) were excluded. In addition, recipients whose indication for transplantation was a malignancy (ie, liver transplant recipients with prior history of a liver, biliary, or neuroendocrine malignancy, or lung transplant recipients whose indication for transplantation was a malignancy or with prior history of a lung malignancy) were also excluded. Malignancies occurring 30 days after transplantation were considered posttransplant de novo malignancies.
Matching was used to minimize the effect of confounding factors. A propensity score for PTM was created using a multivariable logistic regression. The following variables were selected a priori as possible confounders and included in the model: age at transplantation, sex, transplanted organ, year of transplantation, transplant centre, income quintile, and comorbidity using the Resource Utilization Bands (RUB) of the Johns Hopkins Adjusted Clinical Group (ACG) system. Information regarding the type of PTM was obtained from OCR. Malignancies were classified into high- and low-risk based on previously documented risk of recurrence in the transplant population.5,6,12,13 Cancers of the thyroid, prostate, bladder, kidney, testis, and oral cavity/pharynx were classified as low-risk malignancies. Lung, breast, melanoma, gastrointestinal, hematologic, and gynecologic cancers were classified as high-risk malignancies. Time between cancer diagnosis and transplantation was used as a measure of cancer remission time before transplantation and stratified into 0 to 5 or 5 years or longer. Each patient with a PTM was matched to 2 SOTR who did not have a PTM using 1:2 nearest-neighbor or “greedy” matching, without replacement, using a caliper of width 0.2 of the standard deviation of the logit propensity score distribution and exact matching for transplanted organ.14
Outcomes
The primary outcome was all-cause mortality. Secondary outcomes included cancer-specific mortality and cancer recurrence. Mortality was determined from death certificates using the Office of the Registrar General of Ontario death database and cancer-specific mortality was verified using OCR cause of death. A high level of agreement in the cause of death has been reported between the OCR and a prospective cohort of cancer patients with intensive clinical follow-up.15 Recurrence is not recorded in the OCR; therefore, we used a previously validated algorithm that considered diagnosis of metastatic disease, receipt of palliative care, or new chemotherapy/radiotherapy treatment found in administrative data to be evidence of recurrence.16-19 The recurrence date was defined as the earliest date of any palliative, radiotherapy, chemotherapy, or metastatic codes identified. For the ascertainment of recurrence, patients were censored if they developed a posttransplant de novo malignancy.
Transplant Recipient Covariates
Transplant characteristics were extracted from Canadian Organ Replacement Register. There are 6 adult transplant centers in the province and transplant center was also included as a covariate. Income quintile refers to the median annual income by neighborhood determined by linking the 2001 or 2006 census data (closest to the index procedure date) to forward sortation areas. RUB is a 6-level (low to high) simplified morbidity score formed by combining the ACG mutually exclusive cells that measure overall morbidity burden. The 6 levels are: 0, nonusers; 1, healthy users; 2, low morbidity; 3, moderate; 4, high; 5, very high.
Statistical Analysis
For the primary outcome, we performed a priori power calculations based on a previous study that found 4% of Ontario SOTR had a PTM, and SOTR who did not have a PTM had a median survival of 18.2 years.20 Using these data, we estimated we could detect a 10% or greater relative reduction in survival (α = 0.05, β = 0.8) for SOTR with PTM versus those without PTM.
Transplant recipients were censored when they were lost to follow-up (left the province or had no contact with the healthcare system for more than 6 months) or at the end of the follow-up period (December 31, 2012). We used Cox proportional hazard models to assess between-group differences in overall survival (OS) in the whole cohort and by cancer site, type of malignancy (low or high risk), and time between cancer diagnosis and transplantation (0 to 5 or ≥ 5 years). Because the association between type of malignancy and OS may vary based on time between cancer diagnosis and transplantation, interaction terms were created between type of malignancy and time between cancer diagnosis and transplantation.
All analyses of secondary outcomes were adjusted for competing risks given the nonabsorbing nature of the outcomes (ie, patients may experience events other than those of interest) and the high risk of death and diverse causes of death in this population. For the time to cancer death analysis, noncancer death was considered a competing risk. We modeled time to cancer death and time to noncancer death using cause-specific hazard models, and plotted the cumulative incidence function (CIF) of cancer death and noncancer death. Cumulative incidence functions were compared using the Gray test. These analyses were performed in the whole cohort and by type of malignancy (low or high risk), and time between cancer diagnosis and transplantation (0 to 5 or ≥5 years). Lastly, for time to cancer recurrence, death without cancer recurrence was considered a competing event. Analyses of time to cancer recurrence were also performed by type of malignancy (low or high risk), and time between cancer diagnosis and transplantation (0 to 5 or ≥5 years).
We used a robust variance estimator to account for the matching of PTM and non-PTM recipients by the propensity score.21,22 We tested the assumption of proportional hazards in our Cox models by plotting log[−log(survival)] versus log(time) and assessing parallelism. No important departures from proportionality were detected. All tests were 2-sided with a P value of less than 0.05 considered statistically significant. Analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC).
RESULTS
Between January 1, 1991, and December 31, 2010, 11 434 patients received a solid organ transplant in Ontario. Of these, 1161 had a PTM, and 10 273 had no prior history of cancer. Of the 1161 recipients with PTM, we excluded 450 who underwent transplantation for a malignancy, and 258 who were incidentally diagnosed with cancer around the time of transplantation, leaving 443 recipients with a PTM in remission (Figure 1). We derived 1-to-2 matched subcohorts: 443 patients with PTM and 886 patients without PTM, giving a total of 1329 SOTR in our analysis. The cohorts were well matched for key confounders (Table 1).
;)
FIGURE 1: Cohort selection flow diagram.
TABLE 1: Baseline characteristics transplant recipients with a preexisting malignancy in remission and those without preexisting malignancy antecedent after matching
All-Cause Mortality
There were 559 deaths among the 1329 SOTR in our study. A total of 344 (38.8%) of the recipients in the no-PTM cohort died compared with 215 (48.5%) in the PTM group. Transplant recipients with PTM had a worse OS, with a median 10.3 years compared with 14.5 years for those without PTM (Figure 2). The hazard ratio (HR) for all-cause mortality in recipients with PTM compared with those without PTM was 1.38 (95% confidence interval [CI], 1.16-1.64).
FIGURE 2: OS by history of PTM.
Risk of all-cause mortality by PTM cancer site is shown in Figure 3 and Figure 4. Compared with recipients without PTM, those with PTM classified as low risk had similar all-cause mortality (HR, 1.06; 95% CI, 0.86-1.31; Figure 3A), whereas recipients with high-risk PTM were at increased risk of mortality (HR, 1.81; 95% CI, 1.47-2.23; Figure 3A). Recipients with PTM transplanted within 5 years of cancer diagnosis had similar survival compared with recipients without PTM (HR, 1.34; 95% CI, 0.73-2.47; Figure 3B). In contrast, those with intervals between cancer diagnosis and transplantation longer than 5 years had worse OS (HR, 2.27; 95% CI, 1.37-3.75; Figure 3B). Because clinical practice guidelines for the selection of transplant candidates recommend minimum cancer remission intervals before transplantation according to the risk of cancer recurrence, we explored the interactions between the malignancy risk classification and the time from cancer diagnosis to transplantation (TCT) (Figure 3C). Of the 249 recipients transplanted 5 years or longer after cancer diagnosis, 121 (48.6%) had a high-risk malignancy, compared with 77 recipients with high-risk malignancies among 194 recipients transplanted less than 5 years after cancer diagnosis (39.7%). High-risk malignancies were associated with worse OS compared with no PTM for both cancer diagnosis to transplantation intervals (<5 years: HR, 1.47; 95% CI, 1.14-1.90; and ≥ 5 years: HR, 1.79; 95% CI, 1.40-2.67). In contrast, the OS for patients with low-risk malignancies was better than that of patients without PTM for patients transplanted less than 5 years from cancer diagnosis (HR, 0.72; 95% CI, 0.52-0.99). Recipients with low-risk malignancies and transplanted 5 years or longer after cancer diagnosis had worse OS (HR, 1.76; 95% CI, 1.27-2.49).
FIGURE 3: OS by (A) pretransplant cancer site, (B) time between cancer diagnosis and transplantation, and (C) interaction terms of time between cancer diagnosis and pretransplant cancer site.
FIGURE 4: Risk of all-cause mortality by PTM cancer site. HRs for OS were generated using 2 Cox proportional hazard models: (1) by type of malignancy (low- or high-risk) compared to no pretransplant cancer (grey squares), and (2) by cancer site compared with no pretransplant cancer (black squares). Cancer sites listed under low- or high-risk malignancy correspond to each category.
When explored by pretransplant cancer sites, most of the malignancies types classified as high risk were significantly associated with increased mortality (Figure 4). Gastrointestinal PTM were associated with the highest risk of mortality (HR, 2.54; 95% CI, 1.85-3.45), followed by pretransplant melanoma (HR, 1.76; 95% CI, 1.12-2.77), hematologic PTM (HR, 1.68; 95% CI, 1.15-2.46), and breast PTM (HR, 1.55; 95% CI, 1.01-2.20). Although the mortality HR for lung PTM was notably increased (HR, 1.81; 95% CI, 0.84-3.91), it was not statistically significant.
Cause-specific Mortality
Competing risk methodology was used to compare cancer-specific mortality between the 2 groups given the high rate of noncancer deaths as a competing event (Figures 5A and B). Recipients with PTM had a higher rate of both cancer-specific mortality (cause-specific hazard ratio [CSHR], 1.85; 95% CI, 1.20-2.86) and noncancer death (CSHR, 1.29; 95% CI, 1.08-1.54). When stratified by type of PTM, only those with high-risk PTM were at increased risk of cancer death (CSHR, 3.16; 95% CI, 1.98-5.06; Figure 5C). Time from cancer diagnosis to transplantation 5 years or longer was not only associated with an increased risk of cancer death (CSHR, 2.32; 95% CI, 1.44-3.73; Figure 5D) but also with an increased risk of noncancer death (CSHR, 1.53; 95% CI, 1.22-1.90).
FIGURE 5: Cumulative incidence of (A) cancer death and (B) noncancer death comparing recipients with PTM to those without prior cancer history. Cancer death by (C) pretransplant cancer site and (D) TCT. CIFs were compared using the Gray test.
Cancer Recurrence and Cause of Cancer Death
There were 97 cancer recurrences in 443 transplant recipients with PTM, with 137 recipients with PTM dying without experiencing a cancer recurrence during follow-up. At 5 years, the cumulative incidence of recurrence was 14.4%, and death without recurrence was 15.2%. Recurrences by PTM site are presented in Table 2. The cumulative incidence of PTM recurrence was higher for recipients with high-risk versus low-risk malignancies (incidence of recurrence at 5 years: 21.1% vs 9.2%; P = 0.001) and those with prolonged (≥5 year) versus shorter (<5 year) intervals between cancer diagnosis and transplantation (incidence of recurrence at 5 years: 16.8% vs 11.3%, P = 0.03, Figure 6).
TABLE 2: Recurrence of PTM
FIGURE 6: Cancer recurrence after transplantation. Cumulative incidence of (A) cancer recurrence and (B) death without recurrence by time between cancer diagnosis and transplantation (upper panels) and type of PTM (lower panels). CIFs were compared using the Gray test.
DISCUSSION
This large population-based study demonstrates that recipients with PTM are at increased risk of all-cause death compared with recipients without history of cancer. Importantly, recipients with PTM were not only at increased risk of cancer death but also at increased risk of noncancer death. Outcomes of recipients with low-risk PTM were similar to those of recipients with no previous malignancy. In contrast, recipients with high-risk PTM had worse outcomes, independent of the time between cancer diagnosis and transplantation. Recipients with prolonged intervals between cancer diagnosis and transplantation (≥5 years) were at increased risk of noncancer death. Cancer recurrence occurred in 21% of recipients with PTM. Recipients with high-risk PTM had higher rates of recurrence, but notably, time between cancer diagnosis and transplantation was not associated with rate of recurrence.
Our findings are largely consistent with the existing literature.9 We recently conducted a meta-analysis and found that recipients with PTM were at increased risk of all-cause and cancer-specific mortality compared with recipients without PTM.9 However, our results indicate that recipients with PTM are also at increased risk of noncancer mortality. A previous study by Brattstrom et al12 reported that PTM were not associated with increased risk of noncancer deaths, yet, PTM were associated with cardiovascular mortality. Although the underlying cause of increased noncancer mortality is uncertain, recipients with PTM are more likely to have prolonged time with end-stage organ disease and dialysis before transplantation. These factors are associated with increased risks of cardiovascular disease and graft failure, both of which are associated with increased mortality.23,24 In our cohort, we observed that only recipients with longer intervals between cancer diagnosis and transplantation (≥5 years) were at increased risk of noncancer death, which supports this hypothesis. Differences in immunosuppressant management (ie, recipients with PTM may receive mammalian target of rapamycin inhibitors to reduce the risk of recurrence) may result in higher risks of both organ/graft failure and mortality and could also explain the increased risk of noncancer mortality in recipients with PTM.25-28
In addition, in our cohort we did not find an association between prolonged remission times before transplantation and reduced risk of recurrence or death from malignancy. A similar finding was recently reported by Dahle et al29 in a population-based cohort of kidney transplant recipients in Norway. However, previous studies had reported that shorter intervals between cancer diagnosis and transplantation were associated with higher risk of cancer recurrence and cancer-specific mortality.12,13 Nevertheless, these studies did not use competing risk methodology. Failure to account for competing events (eg, noncancer deaths), particularly in a group of patients with high rates of mortality, can lead to incorrect or biased inferences (22). Therefore, for many recipients with PTM, particularly those with malignancies categorized as low-risk, prolonging wait times before listing for transplantation might not reduce the risk of PTM recurrence and death from PTM, but the resulting prolonged time the patient is exposed to organ failure may increase the risk of noncancer death.
The major strengths of this study are its population-based nature, the large number of recipients with PTM, and the comprehensive availability of the cause of death for all SOTRs. The use of competing risk methodology to evaluate cancer-specific mortality and cancer recurrence is particularly important in this population at high risk of death. However, our study has some limitations. Although adjustment was made for multiple confounding factors through propensity score matching, lack of information on induction and maintenance immunosuppression could confound the association between PTM and outcomes. Since cancer remission dates or waitlisting dates were not available, the use of time between cancer diagnosis and transplantation as surrogate likely overestimated the waiting time applied to patients in this study. Recipients with PTM were more likely to have pretransplant comorbidities. However, comorbidities were considered as a confounder and adjusted for in the propensity score matching. The higher prevalence of comorbidities in recipients with PTM would likely increase the difference in all-cause and noncancer mortality observer between both groups. Lastly, the lack of information for cancer stage and did not allow us to explore the effect of pretransplant cancer stage on outcomes. Patients transplanted with shorter remission times were likely a selected group that may have had a lower risk of recurrence based on tumour characteristics. Still this suggests that some recipients with PTM can be clinically identified as having a low risk of recurrence.
In conclusion, our findings indicate that SOTR with PTM experience an increased risk of death when compared with those without a history of cancer. Recipients with PTM are not only at increased risk of cancer-specific mortality but also at increased risk of noncancer death. Recipients with low-risk PTM had good outcomes and low rates of cancer death even when transplanted early, therefore consequences of delaying transplantation should be weighed against the risk of cancer recurrence in this subcohort of patients. Recipients with high-risk PTM were at increased risk of all-cause and cancer-specific mortality and cancer recurrence.
REFERENCES
1. Knoll G, Cockfield S, Blydt-Hansen T, et al. Canadian Society of Transplantation: consensus guidelines on eligibility for kidney transplantation.
CMAJ. 2005;173:S1–S25.
2. Kasiske BL, Cangro CB, Hariharan S, et al. The evaluation of renal transplantation candidates: clinical practice guidelines.
Am J Transplant. 2001;1(S2):3–95.
3. Mehra MR, Kobashigawa J, Starling R, et al. Listing criteria for heart transplantation: International Society for Heart and Lung Transplantation guidelines for the care of cardiac transplant candidates—2006.
J Heart Lung Transplant. 2006;25:1024–1042.
4. Murray KF, Carithers RL Jr. AASLD. AASLD practice guidelines: evaluation of the patient for liver transplantation.
Hepatology. 2005;41:1407–1432.
5. Penn I. Evaluation of transplant candidates with pre-existing malignancies.
Ann Transplant. 1997;2:14–17.
6. Penn I. The effect of immunosuppression on pre-existing cancers.
Transplantation. 1993;55:742–747.
7. Kauffman HM, Cherikh WS, McBride MA, et al. Transplant recipients with a history of a malignancy: risk of recurrent and de novo cancers.
Transplant Rev. 2005;19:55–64.
8. Chapman JR, Sheil AG, Disney AP. Recurrence of cancer after renal transplantation.
Transplant Proc. 2001;33:1830–1831.
9. Acuna SA, Huang JW, Daly C, et al. Outcomes of solid organ transplant recipients with preexisting malignancies in remission: a systematic review and meta-analysis.
Transplantation. 2017;101:471–481.
10. Canadian Institute for Health Information. Canadian Organ Replacement Register Annual Report: Treatment of End-Stage Organ Failure in Canada, 2001 to 2010. Ottawa, Ont.: Canadian Institute for Health Information; 2011.
https://secure.cihi.ca/free_products/2011_CORR_Annua_Report_EN.pdf.
11. McLaughlin JR, Kreiger N, Marrett LD, et al. Cancer incidence registration and trends in Ontario.
Eur J Cancer. 1991;27:1520–1524.
12. Brattstrom C, Granath F, Edgren G, et al. Overall and cause-specific mortality in transplant recipients with a pretransplantation cancer history.
Transplantation. 2013;96:297–305.
13. Sigurdardottir V, Bjortuft O, Eiskjaer H, et al. Long-term follow-up of lung and heart transplant recipients with pre-transplant malignancies.
J Heart Lung Transplant. 2012;31:1276–1280.
14. Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies.
Multivariate Behav Res. 2011;46:399–424.
15. Brenner DR, Tammemägi MC, Bull SB, et al. Using cancer registry data: agreement in cause-of-death data between the Ontario Cancer Registry and a longitudinal study of breast cancer patients.
Chronic Dis Can. 2009;30:16–19.
16. Tan J.
The Processes of Care after Colorectal Cancer Surgery in Ontario [thesis]. University of Toronto; 2008.
17. Forbes SS, Sutradhar R, Paszat LF, et al. Long-term survival in young adults with colorectal cancer: a population-based study.
Dis Colon Rectum. 2010;53:973–978.
18. Richardson DP, Daly C, Sutradhar R, et al. Hospitalization rates among survivors of young adult malignancies.
J Clin Oncol. 2015;33:2655–2659.
19. Daly C, Urbach DR, Stukel TA, et al. Patterns of diagnostic imaging and associated radiation exposure among long-term survivors of young adult cancer: a population-based cohort study.
BMC Cancer. 2015;15:1.
20. Acuna SA, Fernandes KA, Daly C, et al. Cancer mortality among recipients of solid-organ transplantation in Ontario, Canada.
JAMA Oncol. 2016;2:463–469.
21. Lin DY, Wei L-J. The robust inference for the Cox proportional hazards model.
J Am Stat Assoc. 1989;84:1074–1078.
22. Austin PC. The use of propensity score methods with survival or time‐to‐event outcomes: reporting measures of effect similar to those used in randomized experiments.
Stat Med. 2014;33:1242–1258.
23. Keith DS, Cantarovich M, Paraskevas S, et al. Duration of dialysis pretransplantation is an important risk factor for delayed recovery of renal function following deceased donor kidney transplantation.
Transpl Int. 2008;21:126–132.
24. Israni AK, Snyder JJ, Skeans MA, et al. Predicting coronary heart disease after kidney transplantation: Patient Outcomes in Renal Transplantation (PORT) Study.
Am J Transplant. 2010;10:338–353.
25. Delgado JF, Alonso‐Pulpón L, Mirabet S, et al. Cancer incidence in heart transplant recipients with previous neoplasia history.
Am J Transplant. 2016;16:1569–1578.
26. Cortazar F, Molnar MZ, Isakova T, et al. Clinical outcomes in kidney transplant recipients receiving long-term therapy with inhibitors of the mammalian target of rapamycin.
Am J Transplant. 2012;12:379–387.
27. Hoogendijk-van den Akker JM, Harden PN, Hoitsma AJ, et al. Two-year randomized controlled prospective trial converting treatment of stable renal transplant recipients with cutaneous invasive squamous cell carcinomas to sirolimus.
J Clin Oncol. 2013;31:1317–1323.
28. Knoll GA, Kokolo MB, Mallick R, et al. Effect of sirolimus on malignancy and survival after kidney transplantation: systematic review and meta-analysis of individual patient data.
BMJ. 2014;349:g6679.
29. Dahle DO, Grotmol T, Leivestad T, et al. Association between pretransplant cancer and survival in kidney transplant recipients.
Transplantation. 2017;101:2599–2605.
