Abstract
Purpose:
Transplant recipients have an elevated risk of cancer because of organ rejection immunosuppressive medications, but no study has comprehensively examined associations between transplant status and mortality following a cancer diagnosis.
Methods:
For 16 different cancer types, we assessed cases in the US general population (N=7,147,476) ascertained from 11 cancer registries. Presence of a solid organ transplant prior to diagnosis (N=11,416 cancer cases) was identified through linkage with the national transplant registry (1987–2014). We used Cox models to examine the association between transplant status and cancer-specific mortality, adjusting for demographic characteristics and cancer stage.
Results:
For most cancers, cancer-specific mortality was higher in transplant recipients than for other cancer patients. The increase was particularly pronounced for melanoma (adjusted hazard ratio (aHR)=2.59, 95%CI 2.18–3.00) and cancers of the breast (1.88, 1.61–2.19), bladder (1.85, 1.58–2.17), and colorectum (1.77, 1.60–1.96), but it was also increased for cancers of the oral cavity/pharynx, stomach, pancreas, kidney, and lung, and diffuse large B-cell lymphoma (aHRs ranging from 1.21 to 1.47). Associations remained significant after adjustment for first-course cancer treatment and were generally stronger among local stage cancers for which potentially curative treatment was provided, e.g., for melanoma (aHR=3.82, 95%CI 2.94–4.97), and cancers of the colorectum (2.77, 2.07–3.70), breast (2.08, 1.50–2.88), and prostate (1.60, 1.12–2.29), despite lack of association for prostate cancer overall.
Conclusion:
For multiple cancer types, transplant recipients with cancer have an elevated risk of dying from their cancer, even after adjustment for stage and treatment, which may be due to impaired immunity.
Keywords: Population Based, Solid Organ Transplant Recipient, Immunosuppression, Melanoma Breast Colorectal Bladder, Neoplasms Survival
Precis:
For multiple cancer types, transplant recipients with cancer have an elevated risk of dying from their cancer, even after adjustment for stage and treatment, which may be due to impaired immunity.
Background
The number of solid organ recipients has increased in the last decade with almost 35,000 transplants occurring in the United States in 20171. Although solid organ transplant is life-saving, recipients have elevated risk for many cancer types2–4. Immunosuppression attributable to medications used to prevent organ rejection plays a large role in the increased risk, especially for cancers caused by viruses. Additionally, some medications given post-transplant may be inherently carcinogenic or promote tumor growth5–7.
If immunosuppression contributes to cancer outcomes, it could be reasoned that transplant recipients with cancer would also have higher cancer-specific mortality than cancer patients without a transplant, although the evidence to date is limited8,9. Several factors may contribute to mortality differences between transplant recipients and others who develop cancer. Because of frequent interaction with the medical system, transplant recipients may tend to be diagnosed at an early stage of cancer, which could favorably impact prognosis. Transplant recipients also have an elevated risk of dying from other transplant-related complications (e.g., organ failure, infections), so it is critical to accurately identify deaths attributable to cancer. Recipients of different organs can also have variable mortality. For example, differences for kidney recipients may result from a propensity for tumors to develop in native organs left in place at the time of transplantation
Because transplant recipients continue to live longer and the number of people living with a transplant has increased over time, characterizing outcomes following a cancer diagnosis in this population is important. The goal of this population-based study was to comprehensively examine the association between transplant status and mortality following a cancer diagnosis, focusing on cancer-specific mortality.
Methods
We used data from the Transplant Cancer Match (TCM) Study, which links the Scientific Registry of Transplant Recipients (SRTR) and 17 US regional and state cancer registries3. The SRTR contains information on recipient demographic and transplant characteristics. We used data from 11 participating cancer registries that provided vital status and cause of death (COD) information (Table 1). The TCM Study was approved by human subjects research review committees at the National Cancer Institute (NCI) and, as required, at participating cancer registries.
Table 1:
Characteristics of cancer patients according to transplant status*
| Characteristic | Cancer patients with prior transplant | Cancer patients without prior transplant | ||
|---|---|---|---|---|
| (n = 11,416) | (n=7,136,060) | |||
| n | % | n | % | |
| Sex | ||||
| Female | 3,596 | 31.5 | 3,336,109 | 46.8 |
| Male | 7,820 | 68.5 | 3,799,951 | 53.3 |
| Age at diagnosis (years) | ||||
| < 40 | 910 | 8.0 | 316,983 | 4.4 |
| 40–49 | 1,545 | 13.5 | 682,091 | 9.6 |
| 50–59 | 3,252 | 28.5 | 1,349,555 | 18.9 |
| 60–69 | 4,086 | 35.8 | 1,961,454 | 27.5 |
| 70–79 | 1,509 | 13.2 | 1,869,434 | 26.2 |
| 80+ | 114 | 1.0 | 956,543 | 13.4 |
| Race | ||||
| White | 8,286 | 72.6 | 5,856,905 | 82.1 |
| Black | 1,881 | 16.5 | 652,504 | 9.1 |
| Other | 1,249 | 10.9 | 626,651 | 8.8 |
| Cancer site | ||||
| Oral Cavity/pharynx | 696 | 6.1 | 213,384 | 3.0 |
| Colorectum | 942 | 8.3 | 1,011,114 | 14.2 |
| Esophagus | 140 | 1.2 | 86,465 | 1.2 |
| Stomach | 240 | 2.1 | 152,174 | 2.1 |
| Liver | 202 | 1.8 | 90,753 | 1.3 |
| Pancreas | 232 | 2.0 | 189,175 | 2.7 |
| Larynx | 170 | 1.5 | 82,874 | 1.2 |
| Lung | 1,910 | 16.7 | 1,131,970 | 15.9 |
| Melanoma | 552 | 4.8 | 333,779 | 4.7 |
| Breast | 858 | 7.5 | 1,473,944 | 20.7 |
| Prostate | 1,607 | 14.1 | 1,455,712 | 20.4 |
| Bladder | 350 | 3.1 | 236,360 | 3.3 |
| Kidney | 1,562 | 13.7 | 235,833 | 3.3 |
| Thyroid | 399 | 3.5 | 199,586 | 2.8 |
| DLBCL | 1,353 | 11.9 | 132,852 | 1.9 |
| Myeloma | 203 | 1.8 | 110,085 | 1.5 |
| Stage at diagnosis | ||||
| Local | 5,906 | 51.7 | 3,534,550 | 49.5 |
| Regional | 2,248 | 19.7 | 1,757,969 | 24.6 |
| Distant | 2,704 | 23.7 | 1,363,898 | 19.1 |
| Unknown | 558 | 4.9 | 479,643 | 6.7 |
| Year of diagnosis | ||||
| 1987–1996 | 1,061 | 9.3 | 2,148,823 | 30.1 |
| 1997–2001 | 2,204 | 19.3 | 1,551,233 | 21.7 |
| 2002–2005 | 2,533 | 22.2 | 1,280,815 | 18.0 |
| 2006–2009 | 3,304 | 28.9 | 1,360,411 | 19.1 |
| 2010–2014 | 2,314 | 20.3 | 794,778 | 11.1 |
| Surgical therapy† | ||||
| Yes | 6,055 | 57.8 | 3,654,689 | 61.3 |
| No | 4,158 | 39.7 | 2,140,106 | 35.9 |
| Unknown | 261 | 2.5 | 168,546 | 2.8 |
| Radiation therapy† | ||||
| Yes | 2,489 | 23.8 | 1,770,427 | 29.7 |
| No | 7,525 | 71.8 | 3,844,285 | 64.5 |
| Unknown | 460 | 4.4 | 348,629 | 5.9 |
| Chemotherapy† | ||||
| Yes | 2,663 | 25.4 | 1,514,814 | 25.4 |
| No | 7,296 | 69.7 | 4,114,024 | 69.0 |
| Unknown | 515 | 4.9 | 334,503 | 5.6 |
Abbreviations: DLBCL, diffuse large B-cell lymphoma
The cohort includes cancer cases from the following cancer registries that provided information on vital status and cause of death: California (years of cancer diagnosis and follow-up 1988–2012);Colorado (1988–2009);Connecticut (1987–2009);Georgia (1995–2010);Illinois (1987–2013);Iowa (1987–2009);Kentucky (1995–2011);New Jersey (1987–2010);Pennsylvania (1987–2013);Seattle (1987–2014);Texas (1995–2010).
Treatment information was available for all cancer registries for all calendar years of diagnosis, with the exception of Pennsylvania (data restricted to 1998–2013), Kentucky (2004–2011) and Illinois (2005–2013).
We selected cancer cases using cancer registry data and identified which cases were in individuals with a prior transplant through SRTR-linkage. We examined cancer types with at least 150 cases among transplant recipients in a preliminary tabulation (we included esophageal cancer in the study even though the final number of cases was 140). Using a modified version of the Surveillance, Epidemiology, and End Results (SEER) site recode10, we assessed cases of the following cancers: oral cavity/pharynx, colorectum (CRC), esophagus, stomach, liver, pancreas, larynx, lung, melanoma, breast, prostate, bladder, kidney, thyroid, and myeloma. We additionally included diffuse large B-cell lymphoma (DLBCL), the most common non-Hodgkin lymphoma among transplant recipients. We included only first cancer diagnoses, which could have been a first and only cancer (sequence 0) or first of multiple cancers (sequence 1).
The cohort of these cancer patients was followed from cancer diagnosis until the earlier of death or loss of follow-up or December 31, 2014. Individuals who received a transplant after cancer diagnosis initiated follow-up as non-recipients and were censored at the time of transplant. We excluded liver cancers diagnosed within 0–180 days after liver transplantation, because such cases are mostly cancers that were the indication for the liver transplant but which cancer registries record with a diagnosis date shortly after the transplant11.
Although we present results for overall mortality, the primary outcome was cancer-specific mortality, with death due to cancer defined as described by Howlader et al12. The algorithm uses the tumor sequence number, primary site, and COD to classify deaths as attributable to the cancer when the COD incorrectly specifies another cancer or related condition. We calculated overall mortality and cancer-specific mortality rates stratified by cancer site and transplant status. Cox regression was used to estimate the association (hazard ratio) between transplant status and mortality outcomes. Primary adjusted models included adjustment for sex, age, race, SEER summary cancer stage, and diagnosis year (Table 2). Because treatment of transplant recipients and cancer patients has changed over time, we performed secondary analyses restricted to cancers occurring in 2002 or later.
Table 2:
Association between transplant status and cancer-specific mortality
| Cancer site, and transplant status | Cancer-specific deaths | Cancer-specific mortality rate* | HR | 95%CI | aHR | 95%CI |
|---|---|---|---|---|---|---|
| Oral cavity/pharynx | ||||||
| Recipient | 207 | 77.1 | 0.87 | (0.76, 1.00) | 1.21 | (1.06, 1.39) |
| Non-recipient | 75,916 | 78.0 | 1 | referent | 1 | referent |
| Colorectum | ||||||
| Recipient | 369 | 132.4 | 1.38 | (1.25, 1.53) | 1.77 | (1.6, 1.96) |
| Non-recipient | 372,567 | 76.5 | 1 | referent | 1 | referent |
| Esophagus | ||||||
| Recipient | 87 | 416.3 | 0.81 | (0.66, 1.00) | 1.10 | (0.89, 1.36) |
| Non-recipient | 65,094 | 439.9 | 1 | referent | 1 | referent |
| Stomach | ||||||
| Recipient | 160 | 497.4 | 1.25 | (1.07, 1.46) | 1.47 | (1.26, 1.71) |
| Non-recipient | 102,930 | 301.0 | 1 | referent | 1 | Referent |
| Liver | ||||||
| Recipient | 118 | 392.1 | 0.74 | (0.62, 0.89) | 0.81 | (0.68, 0.97) |
| Non-recipient | 63,953 | 606.1 | 1 | referent | 1 | Referent |
| Pancreas | ||||||
| Recipient | 195 | 1165.9 | 1.24 | (1.08, 1.43) | 1.46 | (1.27, 1.68) |
| Non-recipient | 164,347 | 900.6 | 1 | referent | 1 | Referent |
| Larynx | ||||||
| Recipient | 51 | 101.5 | 1.14 | (0.87, 1.50) | 1.24 | (0.94, 1.63) |
| Non-recipient | 29,927 | 69.8 | 1 | referent | 1 | Referent |
| Lung | ||||||
| Recipient | 1,355 | 537.6 | 1.08 | (1.03, 1.14) | 1.35 | (1.29, 1.43) |
| Non-recipient | 870,934 | 424.9 | 1 | referent | 1 | Referent |
| Melanoma | ||||||
| Recipient | 152 | 73.3 | 2.54 | (2.17, 2.98) | 2.59 | (2.18, 3.00) |
| Non-recipient | 47,660 | 22.6 | 1 | referent | 1 | Referent |
| Breast | ||||||
| Recipient | 162 | 46.8 | 1.56 | (1.34, 1.82) | 1.88 | (1.61, 2.19) |
| Non-recipient | 265,229 | 27.0 | 1 | referent | 1 | Referent |
| Prostate | ||||||
| Recipient | 101 | 14.4 | 0.71 | (0.58, 0.86) | 1.07 | (0.88, 1.30) |
| Non-recipient | 188,303 | 20.4 | 1 | referent | 1 | Referent |
| Bladder | ||||||
| Recipient | 152 | 158.6 | 1.86 | (1.58, 2.18) | 1.85 | (1.58, 2.17) |
| Non-recipient | 75,981 | 61.7 | 1 | referent | 1 | Referent |
| Kidney | ||||||
| Recipient | 284 | 54.3 | 0.66 | (0.59, 0.75) | 1.23 | (1.09, 1.38) |
| Non-recipient | 71,838 | 68.4 | 1 | referent | 1 | Referent |
| Thyroid | ||||||
| Recipient | 14 | 7.8 | 0.85 | (0.50, 1.44) | 1.42 | (0.84, 2.39) |
| Non-recipient | 9,350 | 8.0 | 1 | referent | 1 | Referent |
| DLBCL | ||||||
| Recipient | 449 | 112.4 | 0.92 | (0.84, 1.01) | 1.31 | (1.20, 1.44) |
| Non-recipient | 56,983 | 108.8 | 1 | referent | 1 | Referent |
| Myeloma | ||||||
| Recipient | 74 | 129.8 | 0.73 | (0.58, 0.92) | 1.11 | (0.88, 1.39) |
| Non-recipient | 61,262 | 178.5 | 1 | referent | 1 | referent |
Abbreviations: HR, hazard ratio; CI, confidence interval; aHR, adjusted hazard ratio; DLBCL, diffuse large B-cell lymphoma
Mortality rate per 1000 person-years.
† Cox regression models were adjusted for sex, age (<40,40–84 in 5 year increments,85+), race (white,black,other), stage (local,regional,distant,unknown), and diagnosis year (1987–1991,1992–1996,1997–2001,2002–2005,2006–2009, 2010–2014).
We performed additional analyses of cancer-specific mortality accounting for appropriate (Table S2) first course cancer treatment using data provided by cancer registries, restricting these analyses to cancer registries and calendar years when treatment data were at least 90% complete (Table 1/Table S2). We also performed analyses in which we restricted to local stage cancers of the breast, colorectum, lung, prostate, and kidney, as well as melanoma and DLBCL, for which individuals were documented as having received treatment modalities appropriate for curative intent.
Additionally, we examined cancer-specific mortality for kidney, lung, and liver cancers among transplant recipients who received that organ vs. a different organ (e.g., for kidney cancers in kidney recipients vs. recipients of other organs). Finally, to examine whether other biological characteristics of tumors affected associations with mortality, we performed analyses for breast cancers stratified by estrogen receptor (ER) status (available 2004 or later), for CRC separately for colon and rectal cancers, and for lung cancers stratified by adenocarcinoma, squamous cell carcinoma, other non-small cell lung cancer (NSCLC), and small cell lung cancer (SCLC) histology.
Results
Our cohort contained 11,416 cancer patients with a prior transplant and 7,136,060 cancer patients without a transplant (34.6 million person-years of follow-up) (Table 1). Compared with cancer patients without a transplant, those with a prior transplant were younger at diagnosis (median age 59 versus 66 years) and more likely to be male (68.5% versus 53.3%) and of black race (16.5% versus 9.1%). The distribution of cancer diagnoses differed, with a larger fraction of patients with a prior transplant comprised of DLBCL and kidney cancer (Table 1). The timing of cancer diagnoses also differed by transplant status with non-transplant recipients more likely to be diagnosed in earlier calendar years than transplant recipients (median year 2000 versus 2005). Compared to patients without a transplant, those with a transplant were more likely to be diagnosed with distant-stage tumors (23.7% versus 19.1%). Cancer patients with a prior transplant were less likely to receive surgical treatment (57.8% versus 61.3%) and radiation treatment (23.8% versus 29.7%).
Among these cancer patients, overall mortality was 209 vs. 118 per 1000 person-years in those with vs. without a prior transplant, respectively, and cancer-specific mortality was 114 vs. 73 in these two groups. However, mortality rates varied greatly across cancer types (Table S1, Table 2). Across cancer sites, overall mortality rates were generally higher in cancer patients with a prior transplant than in those without a prior transplant (Table S1), and associations with overall mortality strengthened and were all statistically significant after multivariate adjustment (adjusted HRs [aHRs] 1.37–5.19), except for liver cancer (aHR 0.93, 95%CI 0.79–1.08). The strongest associations between transplantation and overall mortality were for patients with thyroid cancer (aHR 5.19, 95%CI 4.35–6.19), melanoma (3.87, 3.47–4.31), and breast cancer (3.34, 3.04–3.67).
Associations for cancer-specific mortality were attenuated compared with associations for overall mortality; however, after adjustment for demographic factors and tumor stage, cancer-specific mortality remained significantly elevated in transplant recipients for all examined cancers except esophageal, liver, laryngeal, thyroid, prostate cancers and myeloma (Table 2). The strongest elevations in cancer-specific mortality associated with transplantation were for patients with melanoma (aHR 2.59, 95%CI 2.18–3.00), breast cancer (1.88, 1.61–2.19), bladder cancer (1.85, 1.58–2.17), and colorectal cancer (1.77, 1.60–1.96). Analyses restricted to cancers occurring in 2002–2014 yielded similar, and in some instances, qualitatively stronger results (Table S3).
Adjustment for first-course cancer treatment did not strongly affect associations between transplantation status and cancer-specific mortality (Table S2), and some associations were stronger, e.g., for melanoma (aHR 3.01, 95%CI 2.54–3.56) and breast cancer (2.23, 95% CI 1.89–2.62). Analyses restricted to patients with local stage cancers receiving treatment modalities appropriate for curative intent yielded similar or slightly stronger associations, compared with first-course treatment adjustment (Table 3). Specifically, prior transplantation was significantly associated with higher cancer-specific mortality for all examined cancers, with especially strong associations seen for melanoma (aHR 3.82, 2.94–4.97) and CRC (2.77, 2.07–3.70). Transplant status was significantly associated with cancer-specific mortality among individuals with local stage prostate cancers who received surgery or radiation treatment (aHR 1.60, 95%CI 1.12–2.29), despite the lack of association for prostate cancer overall (Table 2).
Table 3:
Association between transplant status and cancer-specific mortality among local-stage cancer patients receiving curative treatment *
| Cancer site, and transplant status | Cancer-specific deaths | Cancer-specific mortality rate† | aHR1‡ | 95%CI | aHR2§ | 95%CI |
|---|---|---|---|---|---|---|
| Colorectum | ||||||
| Recipient | 46 | 38.2 | 2.59 | (1.94, 3.47) | 2.77 | (2.07, 3.70) |
| Non-recipient | 32,305 | 19.3 | 1 | referent | 1 | referent |
| Lung | ||||||
| Recipient | 117 | 123.6 | 1.46 | (1.22, 1.75) | 1.66 | (1.38, 1.99) |
| Non-recipient | 47,405 | 85.7 | 1 | referent | 1 | referent |
| Melanoma | ||||||
| Recipient | 56 | 37.8 | 3.88 | (2.98, 5.04) | 3.82 | (2.94, 4.97) |
| Non-recipient | 16,595 | 10.5 | 1 | referent | 1 | referent |
| Breast | ||||||
| Recipient | 36 | 17.1 | 1.99 | (1.43, 2.76) | 2.08 | (1.50, 2.88) |
| Non-recipient | 50,563 | 9.9 | 1 | referent | 1 | referent |
| Prostate | ||||||
| Recipient | 30 | 7.3 | 1.64 | (1.14, 2.34) | 1.60 | (1.12, 2.29) |
| Non-recipient | 36,891 | 7.9 | 1 | referent | 1 | referent |
| Kidney | ||||||
| Recipient | 78 | 19.9 | 1.52 | (1.22, 1.90) | 1.56 | (1.25, 1.96) |
| Non-recipient | 10,396 | 17.2 | 1 | referent | 1 | referent |
| DLBCL | ||||||
| Recipient | 49 | 67.7 | 1.54 | (1.16, 2.04) | 1.44 | (1.09, 1.91) |
| Non-recipient | 6,120 | 53.0 | 1 | referent | 1 | referent |
Abbreviations: aHR, adjusted hazard ratio; CI, confidence interval; DLBCL, diffuse large B-cell lymphoma
Analyses were restricted to local stage cancers. Curative treatment is defined as receipt of surgery for colorectal, breast, and kidney cancers and melanoma; surgery or radiation therapy for lung and prostate cancers; and chemotherapy for DLBCL.
Mortality rate per 1000 person-years.
Cox regression models were adjusted for age (<40,40–84 in 5 year increments,85+), sex, race (white,black,other), stage (local,regional,distant,unknown), and diagnosis year (1987–1991, 1992–1996, 1997–2001, 2002–2005, 2006–2009, 2010–2014).
Cox regression models were adjusted for factors in HR1 plus additional treatment received including: surgery received (yes,no,unknown), radiotherapy received (yes,no,unknown), chemotherapy received (yes,no,other). Breast and prostate cancer models were also adjusted for hormone therapy (yes,no,unknown). DLBCL was not adjusted for surgery, but was adjusted for immune therapy (yes,no,unknown).
We also assessed cancer-specific mortality for cancers arising in an organ among individuals who had a prior transplant of that organ (Table 4). For kidney cancer, excess cancer-specific morality was observed among both patients with a prior kidney transplant (aHR 1.23, 95%CI 1.08, 1.40) and those who received other organs (aHR 1.21, 0.93–1.58). Similarly, for lung cancer, both lung recipients and recipients of other organs had elevated cancer-specific mortality (aHR 1.45, 95%CI 1.20–1.75, and 1.35, 1.27–1.42, respectively). In contrast, liver recipients had more favorable cancer-specific mortality than liver cancer patients without a transplant (aHR 0.59, 95%CI 0.44–0.80), but there was no difference between recipients of other organs and non-recipients (1.02, 0.81–1.27).
Table 4:
Cancer-specific mortality associations related to transplanted organ or cancer-specific biological features.
| Cancer Deaths | Cancer-specific mortality rate | HR | 95%CI | aHR1* | 95%CI | ||
|---|---|---|---|---|---|---|---|
| Analyses for selected cancers, by transplanted organ | |||||||
| Kidney cancer | |||||||
| Kidney recipient | 230 | 52.4 | 0.65 | (0.57, 0.74) | 1.23 | (1.08, 1.40) | |
| Non-kidney recipient | 54 | 64.3 | 0.74 | (0.57, 0.97) | 1.21 | (0.93, 1.58) | |
| Non-recipient | 71,838 | 68.4 | 1 | 1 | |||
| Lung cancer | |||||||
| Lung recipient | 107 | 421.8 | 0.86 | (0.71, 1.04) | 1.45 | (1.20, 1.75) | |
| Non-lung recipient | 1,248 | 550.5 | 1.11 | (1.05, 1.17) | 1.35 | (1.27, 1.42) | |
| Non-recipient | 870,934 | 424.9 | 1 | 1 | |||
| Liver cancer | |||||||
| Liver recipient | 42 | 298.0 | 0.61 | (0.45, 0.82) | 0.59 | (0.44, 0.80) | |
| Non-liver recipient | 76 | 475.1 | 0.85 | (0.68, 1.06) | 1.02 | (0.81, 1.27) | |
| Non-recipient | 63,953 | 606.1 | 1 | 1 | |||
| Analyses for selected cancers, by biological features of the cancer | |||||||
| Breast cancer† | |||||||
| ER positive cases | |||||||
| Recipient | 41 | 41.7 | 1.95 | (1.43, 2.64) | 2.94 | (2.16, 4.00) | |
| Non-recipient | 28,141 | 21.6 | 1 | referent | 1 | referent | |
| ER negative cases | |||||||
| Recipient | 31 | 105.3 | 1.77 | (1.25, 2.52) | 2.21 | (1.56, 3.15) | |
| Non-recipient | 17,735 | 56.6 | 1 | 1 | referent | ||
| Colorectal cancer | |||||||
| Colon cancer | |||||||
| Recipient | 301 | 147.6 | 1.48 | (1.32, 1.66) | 1.81 | (1.62, 2.03) | |
| Non-recipient | 259,781 | 76.2 | 1 | referent | 1 | referent | |
| Rectal cancer | |||||||
| Recipient | 68 | 91.0 | 1.05 | (0.82, 1.33) | 1.59 | (1.26, 2.02) | |
| Non-recipient | 112,786 | 77.2 | 1 | referent | 1 | referent | |
| Lung cancer‡ | |||||||
| Adenocarcinoma | |||||||
| Recipient | 468 | 564.5 | 1.28 | (1.17, 1.40) | 1.56 | (1.42, 1.70) | |
| Non-recipient | 283,598 | 346.1 | 1 | referent | 1 | referent | |
| Squamous cell | |||||||
| Recipient | 399 | 363.6 | 0.87 | (0.79, 0.96) | 1.09 | (0.98, 1.20) | |
| Non-recipient | 190,938 | 383.7 | 1 | referent | 1 | referent | |
| Other NSCLC | |||||||
| Recipient | 329 | 752.6 | 1.17 | (1.05, 1.30) | 1.47 | (1.32, 1.64) | |
| Non-recipient | 238,823 | 561.1 | 1 | referent | 1 | referent | |
| SCLC | |||||||
| Recipient | 154 | 1445.6 | 1.63 | (1.39, 1.91) | 1.67 | (1.42, 1.95) | |
| Non-recipient | 152,187 | 719.3 | 1 | referent | 1 | referent | |
Abbreviations: HR,hazard ratio;aHR,adjusted hazard ratio;CI,confidence interval;ER,estrogen receptor;NSCLC,non-small cell lung cancer;SCLC,small cell lung cancer
Cox models were adjusted for age (<40,40–84 in 5 year increments,85+), sex, race (white,black,other), stage (local,regional,distant,unknown), and diagnosis year (1987–1991,1992–1996,1997–2001,2002–2005,2006–2009, 2010–2014).
Cases were restricted to diagnosis dates 2004–2014. ER status was missing for breast cancers in 15.3% of transplant recipients and 16.6% of non-recipients.
Adenocarcinoma was identified with histology codes: {8140,8141,8143–8145,8147,8190, 8250,8255,8260, 8262,8263,8290,8320,8323,8480,8481,8570–8574,8576}; squamous cell was identified with histology codes: 8052,8070–8076,8078; other NSCLC was identified with histology codes: {8010–8015,8020–8022,8030–8035,8040,8046,8050,8051,8082:8084,8146,8210,8230,8231,8244,8246,8251–8254,8280,8310, 8313,8315, 8330,8333,8341,8345,8350,8500,8510,8512,8520,8521,8525,8530,8550,8551,8560,8562,8575,9015}; SCLC was defined with histology codes: 8041–8045.
For breast cancer (Table 4), transplant recipients had elevated cancer-specific mortality regardless of ER status, but the association appeared qualitatively stronger for ER positive breast cancers (aHR 2.94, 95%CI 2.16–4.00) than ER negative breast cancers (2.21, 1.56–3.15). Among individuals with colorectal cancer (Table 4), patients with prior transplant had higher cancer-specific mortality than patients without a transplant, for both colon cancer (aHR 1.81, 95%CI 1.62–2.03) and rectal cancer (1.59, 1.26–2.02). With respect to lung cancer (Table 4), there was some heterogeneity by subtype in associations of cancer-specific mortality with transplant status, varying from no association for squamous cell cancer (aHR 1.09, 95%CI 0.98–1.20) to elevated mortality for adenocarcinoma, other NSCLC, and SCLC (aHRs 1.47–1.67).
Discussion
In this large population-based study, we examined how the presence of a prior organ transplant among cancer patients affected overall and cancer-specific mortality. Overall mortality was higher in transplant recipients, which was expected because this population is at risk of dying from complications of end-stage organ disease and transplantation. For most of the evaluated cancer sites, transplant-recipients also had elevated cancer-specific mortality. In particular, cancer-specific mortality was strongly elevated for melanoma, breast cancer, and bladder cancer.
Our results are consistent with previous studies examining cancer mortality in immunosuppressed populations. Miao et al.8 examined several cancer sites using data from the Israel Penn International Transplant Tumor Registry (IPITTR). They generally reported similar if slightly stronger associations to those we observed, with the exception of prostate cancer (for which they documented a very strong elevation in cancer-specific mortality). However, IPITTR is not a population-based study and had substantially fewer cases than our study, and Miao et al. excluded a large fraction of cases that lacked stage information. Our results showing elevated cancer-specific mortality in transplant recipients with melanoma are similar to those reported by Robbins et al.13 using an earlier version of TCM data and Vajdic et al.14 for Australian patients.
HIV infection causes immunosuppression similar to that observed in transplant recipients, through depletion of CD4-positive T-cells. A comprehensive study reported that HIV-infected individuals with cancer generally experienced higher cancer-specific mortality than HIV-uninfected cancer patients15. Those findings are largely mirrored by our results, although the associations were generally weaker in HIV-infected individuals than in our study. No elevation in cancer-specific mortality was reported for HIV-infected individuals with DLBCL15, which may reflect the misclassification of some DLBCL deaths as due to acquired immunodeficiency syndrome.
As a result of their close follow-up for post-transplant medical care, transplant recipients are likely to receive timely work-up and diagnosis of cancer. In turn, this early diagnosis would be expected to result in a relatively early stage at cancer diagnosis. Overall, however, we observed a slight shift to more advanced stage at diagnosis for all cancers as a group (Table 1). Transplant recipients with cancer in our study were also less likely to receive surgery and radiation therapy than other cancer patients. Differences in stage and treatment were partly driven by the distribution of cancers in the two groups, as individual cancer sites varied considerably with respect to stage and therapy (data not shown). Shiels et al. previously observed a shift towards earlier stage at diagnosis among cancer patients with a prior transplant for a number of individual cancer sites16. Nonetheless, the associations with elevated cancer-specific mortality that we demonstrate were present for individual cancer sites with adjustment for stage and cancer treatment, and certain estimates appeared stronger when we restricted to patients with local stage cancers who received treatment appropriate for curative intent as well as for cancer cases treated recently (2002–2014).
Our results support a model in which immunosuppression, attributable to medications given to prevent organ rejection, increases cancer-specific mortality. Immunity is increasingly recognized as critical to cancer control. In particular, heightened immune function as reflected by the presence of tumor infiltrating lymphocytes (TILs) is associated with lower mortality in patients with melanoma17, colorectal cancer18,19, bladder cancer20, and breast cancer21–23, and with favorable tumor features, such absence of lymph node metastases in melanoma24 and chemotherapeutic response in breast cancer23,25,26. However, we are not aware of any data regarding the presence of TILs in tumors among transplant recipients. Additionally, immunotherapy, in particular monoclonal antibodies that target programmed death-ligand 1 (PD-L1) and its receptor PD-1 on T-cells, is increasingly used to manage advanced stage cancers including melanoma, bladder cancer, and lung cancer27. Finally, immunosuppressive drugs may directly contribute to cancer-specific mortality among transplant recipients by promoting tumor invasiveness, angiogenesis, and metastasis5–7.
The emerging understanding of how TILs and immunotherapy influence prognosis for the aforementioned cancers make the strong associations for melanoma, bladder cancer, and breast cancer more intriguing. Moreover, in the study by Shiels et al.16, melanoma and bladder cancer presented at more distant stages among both transplant recipients and HIV-infected individuals than among cancer patients in the general population.
With the exception of prostate cancer, we observed stronger associations with cancer-specific mortality for cancers with typically better prognoses, such as local stage cancers and ER-positive breast cancers, compared with associations for more lethal cancers like pancreatic cancer. This pattern may be attributable to the small relative impact immunosuppression has on mortality in very aggressive cancers. Additionally, apparently curable cancers may be differentially susceptible to micro-metastases in immunosuppressed individuals, whereby a seemingly good prognosis cancer is more serious than staging would indicate28. Alternatively, the algorithm may have differentially misclassified some deaths as cancer-specific deaths because of cause of death errors. Although this algorithm has been validated for a range of cancer sites and demographic groups in the general population12, its performance in a transplant population is unknown. Differentially misclassifying more deaths as cancer-attributable in transplant recipients could have had a relatively large impact on analyses of less aggressive cancers. The reason for the observed heterogeneity across the lung cancer subtypes was unclear.
The 40% reduction in cancer-specific mortality observed for liver cancers that developed in liver recipients was surprising. One possible explanation is that liver recipients were under close surveillance, and any liver cancers developing after transplantation were detected when they were small and amenable to treatment. Additionally, the transplanted liver in which liver cancers arose was from a healthy donor. In contrast, lung cancers in lung recipients and kidney cancers in kidney recipients generally arise in a native, damaged organ left in place at transplantation29–32, which may contribute to the poor outcomes.
Our study has several strengths. It is the largest and most comprehensive examination of the association between transplant status and cancer mortality, and our sample of cancer patients was population-based, incorporating all cases reported to central cancer registries in 11 US areas. Additionally, we utilized an algorithm previously validated for the general population to more accurately classify deaths attributable to cancer12.
There are also several limitations. Information on some tumor characteristics, such as grade and molecular features, was unavailable or missing for some cases, which prevented us from assessing their impact. For example, there was a suggestion of more ER negative breast cancers among transplant recipients, but ER status was missing for ~15% of individuals, and there were too few data on human epidermal growth factor receptor 2 status of breast cancer cases to evaluate. Additionally, we only had data from cancer registries on the first course of cancer treatment, and data on some treatment modalities (especially chemotherapy) were likely incomplete. Registry treatment data also lack granularity. For example, we have no information on treatment tolerance, length of treatment, or specific medications or procedures.
In conclusion, we provide evidence that transplant recipients who develop cancer generally have higher mortality due to their cancer than other cancer patients. As transplant recipients continue to live longer with improved outcomes, cancer will likely increase as a cause of morbidity and mortality in this population. More research is needed to understand whether tumors arising in this population are affected by the patients’ immunosuppression. Finally, additional work is needed to identify optimal treatment regimens in cancer patients with a prior transplant.
Supplementary Material
Acknowledgements
This research was supported in part by the Intramural Research Program of the NCI. The authors gratefully acknowledge the support and assistance provided by individuals at the Health Resources and Services Administration (Monica Lin), the SRTR (Ajay Israni, Bertram Kasiske, Paul Newkirk, Jon Snyder), and the following cancer registries: the states of California (Tina Clarke), Colorado (Jack Finch), Connecticut (Lou Gonsalves), Florida (Brad Wohler), Georgia (Rana Bayakly), Hawaii (Brenda Hernandez), Illinois, Iowa, Kentucky (Jaclyn Nee), Michigan (Glenn Copeland), New Jersey (Xiaoling Niu), New York (Amy Kahn), North Carolina, Texas (Leticia Nogueria), and Utah (Janna Harrell), and the Seattle-Puget Sound area of Washington (Margaret Madeleine). We also thank Kelly Yu at the NCI for study management, and analysts at Information Management Services for programming support (David Castenson, Matthew Chaloux, Michael Curry, Ruth Parsons).
The views expressed in this paper are those of the authors and should not be interpreted to reflect the views or policies of the NCI, Health Resources and Services Administration, SRTR, cancer registries, or their contractors.
The SRTR is currently operated under contract number HHSH250201500009C (Health Resources and Services Administration) by the Minneapolis Medical Research Foundation, Minneapolis, MN. Previously the SRTR was managed under contracts HHSH250201000018C and HHSH234200537009C. The following cancer registries were supported by the SEER Program of the NCI: California (contracts HHSN261201000036C, HHSN261201000035C, HHSN261201000034C), Connecticut (HHSN261201300019I), Hawaii (HHSN261201000037C, N01-PC-35137, N01-PC-35139), Iowa (HSN261201000032C, N01-PC-35143), New Jersey (HHSN261201300021I, N01-PC-2013-00021), Seattle-Puget Sound (N01-PC-35142), and Utah (HHSN2612013000171). The following cancer registries were supported by the National Program of Cancer Registries of the Centers for Disease Control and Prevention: California (agreement 1U58 DP000807-01), Colorado (U58 DP000848-04), Georgia (5U58DP003875-01), Illinois (5U58DP003883-03), Maryland (U58DP12-1205 3919-03), Michigan (5U58DP003921-03), New Jersey (NU58DP003931-05-00), New York (U58DP003879), North Carolina (U58DP003933) and Texas (5U58DP000824-04). Additional support was provided by California, Colorado, Connecticut, Illinois, Iowa, Massachusetts (Massachusetts Cancer Prevention and Control Cooperative Agreement 5458DP003920), New Jersey, New York (including the Cancer Surveillance Improvement Initiative), Texas, Utah, and Washington, as well as the University of Utah and Fred Hutchinson Cancer Research Center in Seattle, WA.
Funding:
This research was supported the National Cancer Institute, Health Resources and Services Administration, and Centers for Disease Control and Prevention. Additional support was provided by California, Colorado, Connecticut, Illinois, Iowa, Massachusetts, New Jersey, New York, Texas, Utah, and Washington; the University of Utah; and Fred Hutchinson Cancer Research Center.
Footnotes
Conflicts of Interest: None
References
- 1.OPTN. Organ Procurement and Transplantation Network. 2018; http://optn.transplant.hrsa.gov. Accessed Feb 4, 2018, 2018.
- 2.Collett D, Mumford L, Banner NR, Neuberger J, Watson C. Comparison of the incidence of malignancy in recipients of different types of organ: a UK Registry audit. Am J Transplant. 2010;10(8):1889–1896. [DOI] [PubMed] [Google Scholar]
- 3.Engels EA, Pfeiffer RM, Fraumeni JF Jr., et al. Spectrum of cancer risk among US solid organ transplant recipients. Jama. 2011;306(17):1891–1901. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Grulich AE, van Leeuwen MT, Falster MO, Vajdic CM. Incidence of cancers in people with HIV/AIDS compared with immunosuppressed transplant recipients: a meta-analysis. Lancet. 2007;370(9581):59–67. [DOI] [PubMed] [Google Scholar]
- 5.Hojo M, Morimoto T, Maluccio M, et al. Cyclosporine induces cancer progression by a cell-autonomous mechanism. Nature. 1999;397(6719):530–534. [DOI] [PubMed] [Google Scholar]
- 6.Zhou AY, Ryeom S. Cyclosporin A promotes tumor angiogenesis in a calcineurin-independent manner by increasing mitochondrial reactive oxygen species. Mol Cancer Res. 2014;12(11):1663–1676. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Basu A, Contreras AG, Datta D, et al. Overexpression of vascular endothelial growth factor and the development of post-transplantation cancer. Cancer Res. 2008;68(14):5689–5698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Miao Y, Everly JJ, Gross TG, et al. De novo cancers arising in organ transplant recipients are associated with adverse outcomes compared with the general population. Transplantation. 2009;87(9):1347–1359. [DOI] [PubMed] [Google Scholar]
- 9.Robbins HA, Clarke CA, Arron ST, et al. Melanoma Risk and Survival among Organ Transplant Recipients. J Invest Dermatol. 2015;135(11):2657–2665. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Engels EA, Pfeiffer RM, Ricker W, Wheeler W, Parsons R, Warren JL. Use of surveillance, epidemiology, and end results-medicare data to conduct case-control studies of cancer among the US elderly. American Journal of Epidemiology. 2011;174(7):860–870. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Koshiol J, Pawlish K, Goodman MT, McGlynn KA, Engels EA. Risk of hepatobiliary cancer after solid organ transplant in the United States. Clin Gastroenterol Hepatol. 2014;12(9):1541–1549 e1543. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Howlader N, Ries LA, Mariotto AB, Reichman ME, Ruhl J, Cronin KA. Improved estimates of cancer-specific survival rates from population-based data. Journal of the National Cancer Institute. 2010;102(20):1584–1598. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Rodrigues LK, Klencke BJ, Vin-Christian K, et al. Altered clinical course of malignant melanoma in HIV-positive patients. Arch Dermatol. 2002;138(6):765–770. [DOI] [PubMed] [Google Scholar]
- 14.Vajdic CM, Chong AH, Kelly PJ, et al. Survival after cutaneous melanoma in kidney transplant recipients: a population-based matched cohort study. Am J Transplant. 2014;14(6):1368–1375. [DOI] [PubMed] [Google Scholar]
- 15.Coghill AE, Shiels MS, Suneja G, Engels EA. Elevated Cancer-Specific Mortality Among HIV-Infected Patients in the United States. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2015;33(21):2376–2383. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Shiels MS, Copeland G, Goodman MT, et al. Cancer stage at diagnosis in patients infected with the human immunodeficiency virus and transplant recipients. Cancer. 2015;121(12):2063–2071. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Azimi F, Scolyer RA, Rumcheva P, et al. Tumor-infiltrating lymphocyte grade is an independent predictor of sentinel lymph node status and survival in patients with cutaneous melanoma. J Clin Oncol. 2012;30(21):2678–2683. [DOI] [PubMed] [Google Scholar]
- 18.Naito Y, Saito K, Shiiba K, et al. CD8+ T cells infiltrated within cancer cell nests as a prognostic factor in human colorectal cancer. Cancer Res. 1998;58(16):3491–3494. [PubMed] [Google Scholar]
- 19.Pages F, Berger A, Camus M, et al. Effector memory T cells, early metastasis, and survival in colorectal cancer. N Engl J Med. 2005;353(25):2654–2666. [DOI] [PubMed] [Google Scholar]
- 20.Liakou CI, Narayanan S, Ng Tang D, Logothetis CJ, Sharma P. Focus on TILs: Prognostic significance of tumor infiltrating lymphocytes in human bladder cancer. Cancer Immun. 2007;7:10. [PMC free article] [PubMed] [Google Scholar]
- 21.Adams S, Gray RJ, Demaria S, et al. Prognostic value of tumor-infiltrating lymphocytes in triple-negative breast cancers from two phase III randomized adjuvant breast cancer trials: ECOG 2197 and ECOG 1199. J Clin Oncol. 2014;32(27):2959–2966. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Ladoire S, Mignot G, Dabakuyo S, et al. In situ immune response after neoadjuvant chemotherapy for breast cancer predicts survival. J Pathol. 2011;224(3):389–400. [DOI] [PubMed] [Google Scholar]
- 23.Loi S, Michiels S, Salgado R, et al. Tumor infiltrating lymphocytes are prognostic in triple negative breast cancer and predictive for trastuzumab benefit in early breast cancer: results from the FinHER trial. Ann Oncol. 2014;25(8):1544–1550. [DOI] [PubMed] [Google Scholar]
- 24.Taylor RC, Patel A, Panageas KS, Busam KJ, Brady MS. Tumor-infiltrating lymphocytes predict sentinel lymph node positivity in patients with cutaneous melanoma. J Clin Oncol. 2007;25(7):869–875. [DOI] [PubMed] [Google Scholar]
- 25.Issa-Nummer Y, Loibl S, von Minckwitz G, Denkert C. Tumor-infiltrating lymphocytes in breast cancer: A new predictor for responses to therapy. Oncoimmunology. 2014;3:e27926. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Kochi M, Iwamoto T, Niikura N, et al. Tumour-infiltrating lymphocytes (TILs)-related genomic signature predicts chemotherapy response in breast cancer. Breast Cancer Res Treat. 2018;167(1):39–47. [DOI] [PubMed] [Google Scholar]
- 27.Rotte A, Jin JY, Lemaire V. Mechanistic overview of immune checkpoints to support the rational design of their combinations in cancer immunotherapy. Ann Oncol. 2017. [DOI] [PubMed] [Google Scholar]
- 28.Strauss DC, Thomas JM. Transmission of donor melanoma by organ transplantation. Lancet Oncol. 2010;11(8):790–796. [DOI] [PubMed] [Google Scholar]
- 29.Dickson RP, Davis RD, Rea JB, Palmer SM. High frequency of bronchogenic carcinoma after single-lung transplantation. J Heart Lung Transplant. 2006;25(11):1297–1301. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Minai OA, Shah S, Mazzone P, et al. Bronchogenic carcinoma after lung transplantation: characteristics and outcomes. J Thorac Oncol. 2008;3(12):1404–1409. [DOI] [PubMed] [Google Scholar]
- 31.Leveridge M, Musquera M, Evans A, et al. Renal cell carcinoma in the native and allograft kidneys of renal transplant recipients. J Urol. 2011;186(1):219–223. [DOI] [PubMed] [Google Scholar]
- 32.Tsaur I, Obermuller N, Jonas D, et al. De novo renal cell carcinoma of native and graft kidneys in renal transplant recipients. BJU Int. 2011;108(2):229–234. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.