This figure shows the total number of patients eligible for study inclusion from the Brigham and Women’s Hospital and Massachusetts General Hospital, sirolimus-treated and non–sirolimus-treated patients after development of a first posttransplant cancer, and the number of patients who developed a second posttransplant cancer (of any type).
aBeginning of observation time.
bCensored at the time of skin cancer formation.
cCensored at the time of death, last follow-up, or the end of data collection.
Cumulative incidence rates for a second posttransplant cancer by sirolimus treatment. P value is calculated using the Gray test.23
eTable 1.ICD-9 Codes of Procedures and Diagnoses Used to Obtain Study Cohort
eTable 2. Results of Univariate Competing Risks Analysis for Postindex Cancer and Skin Cancer Formation
Customize your JAMA Network experience by selecting one or more topics from the list below.
Karia PS, Azzi JR, Heher EC, Hills VM, Schmults CD. Association of Sirolimus Use With Risk for Skin Cancer in a Mixed-Organ Cohort of Solid-Organ Transplant Recipients With a History of Cancer. JAMA Dermatol. 2016;152(5):533–540. doi:10.1001/jamadermatol.2015.5548
Solid-organ transplant recipients (OTRs) are at an increased risk for skin cancer. Prior studies have demonstrated a reduced incidence of skin cancer in renal OTRs treated with sirolimus. However, little information exists on the use of sirolimus for the prevention of skin cancer in nonrenal OTRs or those already diagnosed as having a posttransplant cancer.
To compare subsequent skin cancer formation in a mixed-organ cohort of OTRs who were or were not treated with sirolimus after developing a posttransplant index cancer of any type.
Design, Setting, and Participants
A 9-year retrospective cohort study at 2 academic tertiary care centers. Electronic medical records were reviewed for OTRs diagnosed as having a posttransplant cancer of any type to determine the type of organ transplanted, pretransplant and posttransplant cancer, and immunosuppressive medications. Patients underwent transplant from January 1, 2000, to December 31, 2008. Data were collected from July 30, 2011, to December 31, 2012, when follow-up was completed, and analyzed from April 28, 2013, to October 4, 2014.
Main Outcomes and Measures
Factors associated with subsequent skin cancer development were evaluated via multivariate Cox regression analysis.
Of 329 OTRs with an index posttransplant cancer (100 women and 229 men; mean [SD] age, 56  years), 177 (53.8%) underwent renal transplant; 58 (17.6%), heart transplant; 54 (16.4%), lung transplant; 34 (10.3%), liver transplant; and 6 (1.8%), mixed-organ transplant. Ninety-seven OTRs (29.5%) underwent conversion to sirolimus therapy after diagnosis. One hundred thirty OTRs (39.5%) developed second posttransplant cancers, of which 115 cases (88.5%) were skin cancers. An 11.6% reduction in skin cancer risk was observed in the sirolimus-treated vs non–sirolimus-treated groups overall (26 of 97 [26.8%] vs 89 of 232 [38.4%]; P = .045) and among nonrenal OTRs only (8 of 34 [23.5%] vs 44 of 112 [39.3%], respectively), although the latter difference was not significant (P = .09). Independent predictors of skin cancer formation after the index posttransplant cancer were history of pretransplant skin cancer (subhazard ratio, 2.1; 95% CI, 1.2-3.7), skin cancer as the index posttransplant cancer (subhazard ratio, 5.5; 95% CI, 2.5-6.4), and sirolimus treatment (subhazard ratio, 0.6; 95% CI, 0.4-0.9). These same risk factors were associated with skin cancer formation when the analysis was limited to nonrenal OTRs. No difference was found in allograft rejection or death between sirolimus-treated and non–sirolimus-treated groups.
Conclusions and Relevance
In this mixed-organ cohort of OTRs, patients taking sirolimus after developing posttransplant cancer had a lower risk of developing subsequent skin cancer, with no increased risk for overall mortality. Thus, conversion to sirolimus therapy may be considered in OTRs who develop cancer if the risk for skin cancer is of concern. Larger studies are needed to quantify sirolimus-associated risk reduction for other cancer types.
Abundant evidence suggests that solid-organ transplant recipients (OTRs) have an increased risk for cancer formation compared with the general population, with a 3- to 4-fold increased incidence reported.1 The most common cancer in OTRs is nonmelanoma skin cancer (predominantly cutaneous squamous cell carcinoma [CSCC]), followed by posttransplant lymphoproliferative disease, Kaposi’s sarcoma, and non-Hodgkin lymphoma.2-7 The incidence of CSCC is estimated to be 65 to 250 times more frequent in OTRs compared with the general population.8 Several studies9-11 have also shown a 28- to 49-fold increase in the incidence of non–Hodgkin lymphoma and a 400- to 500-fold increase in the incidence of Kaposi’s sarcoma in OTRs compared with the general population.
Advances in immunosuppressive medications have been responsible for considerable improvements in acute and chronic organ rejection and the life expectancy of OTRs. However, the risk for cancer in OTRs is related to the level of immunosuppression. A high risk for malignant neoplasms has been observed for heart and lung OTRs who receive high levels of immunosuppressive therapy, whereas a lower risk has been observed in kidney and especially liver OTRs who require lower levels of immunosuppressive medications.12-14 Multiple-drug regimens appear to be associated with the highest risk.15 Conversely, mammalian target of rapamycin (mTOR) inhibitors have been shown to reduce the growth and proliferation of tumor cells.16,17 Recent clinical studies have demonstrated a reduced incidence of skin cancer in kidney OTRs treated with sirolimus as first-time therapy18,19 and those who switch therapy from calcineurin inhibitors to sirolimus.20-22 Studies examining the effect of mTOR use on the risk for subsequent cancer formation in OTRs who already have been diagnosed as having a posttransplant cancer are limited. Because cancer formation is a major reason for conversion to mTOR therapy, data on the reduction of cancer risk in patients with a history of cancer have high clinical relevance. In addition, data regarding the impact of conversion to mTOR inhibitor therapy in nonrenal OTRs are limited. The present study was undertaken to assess the risk for subsequent cancer in a mixed-organ cohort of OTRs diagnosed as having various posttransplant cancers and to compare subsequent cancer risk by use vs nonuse of mTOR inhibitors.
We searched the Research Patient Data Registry for all OTRs at the Brigham and Women’s Hospital and Massachusetts General Hospital from January 1, 2000, to December 31, 2008, who were diagnosed as having a pathologically confirmed posttransplant cancer of any type. The year 2008 was chosen as the cutoff so that a 4-year follow-up was possible, allowing time for cancer formation, conversion to sirolimus treatment, and second cancer formation. The database was searched using the organ transplant category under inpatient procedure codes to identify OTRs and was stratified using the malignant neoplasm diagnosis codes to exclude patients without posttransplant cancer (eTable 1 in the Supplement). Inpatient procedures and cancer diagnoses were based on codes from the International Classification of Diseases, Ninth Revision. Patients 18 years or older with a confirmed heart, lung, liver, or kidney transplant (based on operative notes), a posttransplant cancer (confirmed via pathology reports), and at least 2 visits at Brigham and Women’s Hospital or Massachusetts General Hospital were included in the final cohort. The study was approved by the Partners Human Research Office, who waived the need for informed consent. Patient data were deidentified.
Patient data were collected from July 30, 2011, to December 31, 2012, when follow-up was completed. We reviewed electronic medical records for all OTRs who met the inclusion criteria. The following information was extracted for each patient: age; sex; race; transplant history, including the date and type of transplant, transplant rejection, and posttransplant infections; immunosuppressive therapy history, including induction therapy, name of the medication, dose of the medication, and start and end dates of treatment; and pretransplant and posttransplant cancer history. Additional information recorded from the electronic medical records included the history of induction therapy (if any), smoking, alcohol use, and use of tanning beds and any family history of cancer.
Data were analyzed from April 28, 2013, to October 4, 2014. We analyzed baseline demographic variables and medication data using descriptive statistics and frequency tabulation. The primary comparison group consisted of patients who were prescribed an mTOR inhibitor between their first posttransplant cancer and their second posttransplant cancer or censorship vs those who did not use an mTOR inhibitor between their first posttransplant cancer and their second posttransplant cancer or censorship. The follow-up period for statistical modeling began at the time of the first posttransplant diagnosis of a malignant neoplasm (the index cancer). Patients were censored (follow-up stopped) when they developed a second posttransplant cancer if such a tumor developed. Because skin cancer was the predominant cancer, patients who developed nonskin cancers were excluded from multivariate analyses. If no skin cancer developed after the index cancer, patients were censored at the time of death or on December 31, 2012 (end of data collection). Figure 1 provides a schematic of these methods.
We used 2-tailed t tests for continuous variables and χ2 tests for categorical variables. Competing-risks survival regression performed with the method of Fine and Gray23 was used to determine univariate and multivariate associations of risk factors with the development of subsequent skin cancers after the index cancer in those treated with mTOR inhibitors vs those who were not. All models were adjusted for the competing risk for all-cause mortality.
Multivariate models were built through forward stepwise variable addition followed by backward elimination. In this form of model building, modeling begins with the variable with the largest effect estimate on univariate modeling. Other variables were added based on the next variable with the largest effect estimate and retained in the model if the Wald test comparing the smaller model with the larger model was significant at P ≤ .05 or if the P value ranged from greater than .05 to greater than .99 (nonsignificant). Cumulative incidence function curves were generated to illustrate the incidence of a second skin cancer formation. All analysis was performed using STATA statistical software (version 12.0; StataCorp).
We reviewed the electronic medical records of 329 patients (100 women and 229 men; mean [SD] age, 56  years) with a first posttransplant cancer of any type. Most of the transplants were renal (177 [53.8%]), followed by heart (58 [17.6%]), lung (54 [16.4%]), and liver (34 [10.3%]). Only 6 patients (1.8%) underwent mixed-organ transplants, including 1 patient with a heart-lung transplant, 3 patients with liver-kidney transplants, and 2 patients with heart-kidney transplants.
Sirolimus was the only mTOR inhibitor drug used in the cohort because other mTOR inhibitors were not yet in use during the study period. Median follow-up time was 40 (range, 4-143) months in the sirolimus-treated group (n = 97) and 36 (range, 2-245) months in the non–sirolimus-treated group (n = 232). The median time to the development of the first posttransplant cancer (index cancer) was 42 (range, 3-341) months. The most common posttransplant index cancers by frequency included CSCC (105 [31.9%]), basal cell carcinoma (BCC) (74 [22.5%]), lung cancer (29 [8.8%]), B-cell lymphoma (26 [7.9%]), prostate cancer (10 [3.0%]), renal cancer (9 [2.7%]), melanoma (9 [2.7%]), thyroid cancer (8 [2.4%]), bladder cancer (7 [2.1%]), breast cancer (6 [1.8%]), colon cancer (6 [1.8%]), and other (40 [12.2%]).
Patients were stratified by those prescribed sirolimus between their first posttransplant index cancer and censorship (97 [29.5%]) and those not prescribed sirolimus between their first posttransplant index cancer and censorship (232 [70.5%]) (Table 1). The reasons for sirolimus therapy included posttransplant nonskin cancer (37 patients [38.1%]), improvement of allograft function and /or prevention of calcineurin inhibitor toxicity (30 patients [30.9%]), posttransplant skin cancer (28 patients [28.9%]), and unknown (2 patients [2.1%]).
No difference in age (P = .55), sex (P = .51), race (P = .51), history of pretransplant skin cancer (P = .21), type of index posttransplant cancer (P = .83), history of posttransplant infections (P = .06), or induction therapy (P = .78) was observed between the sirolimus-treated and non–sirolimus-treated patients. A higher proportion of kidney (62 of 97 [63.9%] vs 115 of 232 [49.6%]; P = .02) and liver (16 of 97 [16.5%] vs 18 of 232 [7.8%]; P = .03) OTRs and a lower proportion of lung OTRs (5 of 97 [5.2%] vs 49 of 232 [21.1%]; P < .001) were prescribed sirolimus. In addition, a higher proportion of patients prescribed sirolimus had a history of a pretransplant nonskin cancer compared with those not prescribed sirolimus (20 of 97 [20.6%] vs 27 of 232 [11.6%]; P = .03). We found no significant difference in the proportion of heart OTRs (P = .19) and multiorgan OTRs (P = .68) in the sirolimus-treated vs non–sirolimus-treated groups. No significant difference in azathioprine (P = .83), prednisone (P = .11), or cyclosporine (P = .62) use during the study period was observed between the sirolimus-treated vs non–sirolimus-treated groups. However, compared with the non–sirolimus-treated group, a higher proportion of the sirolimus-treated group had used mycophenolate mofetil (90 of 97 [92.8%] vs 194 of 232 [83.6%]; P = .03) and tacrolimus (88 of 97 [90.7%] vs 187 of 232 [80.6%]; P = .02) (Table 1).
Table 2 shows the risk for subsequent cancer formation stratified by sirolimus use. In the overall cohort of 329 patients, 130 (39.5%) developed a second posttransplant cancer of any type. The median time to the development of a second posttransplant cancer after an index posttransplant cancer was 14 (range, 1-182) months. A 12.2% reduction in the risk for a second cancer (of any type) was observed in the sirolimus-treated group compared with the non–sirolimus-treated group (30 of 97 [30.9%] vs 100 of 232 [43.1%]; P = .04). This reduction in risk for a second cancer was largely attributed to a reduction in skin cancer formation (26 of 97 [26.8%] vs 89 of 232 [38.4%]; P = .045). Of the 329 patients with a postindex cancer, 58 (17.6%) developed CSCC; 53 (16.1%), BCC; 3 (0.9%), melanoma; and 1 (0.4%), sebaceous carcinoma. Only 15 of 130 patients (11.5%) with a second posttransplant cancer developed a nonskin cancer. A significant difference by sirolimus use in the proportion of patients with poor outcomes was not observed.
In the nonrenal transplant cohort, 57 of 146 OTRs (39.0%) developed a second posttransplant cancer of any type. A 16.4% reduction in second cancer risk (of any type) was observed in sirolimus-treated patients compared with the non–sirolimus-treated patients; however, the difference was not significant (9 of 34 [26.5%] vs 48 of 112 [42.9%]; P = .09). Similar to the overall cohort, the reduction in risk for a second cancer was attributed to a reduction in skin cancer formation (8 of 34 [23.5%] vs 44 of 112 [39.3%]; P = .09), although the difference was not significant. Of these 146 OTRs, 24 (16.4%) developed CSCC; 25 (17.1%), BCC; 2 (1.4%), melanoma; and 1 (0.7%), sebaceous carcinoma. Only 5 of the 146 nonrenal OTRs (3.4%) developed a nonskin cancer as their second posttransplant cancer.
No significant difference in death from posttransplant infection (P = .25) or rejection (P = .33) was observed among sirolimus-treated vs non–sirolimus-treated OTRs in the overall cohort. A 9% reduced risk for overall death was observed among sirolimus-treated patients in the overall cohort; however, the difference did not reach statistical significance (27.1% vs 36.6%; P = .10).
The cumulative incidence rates of skin cancer at 1, 3, and 5 years after the index posttransplant cancer were 9.3%, 20.6%, and 24.7%, respectively, in the sirolimus-treated group vs 17.7%, 31.0%, and 35.8%, respectively, in the non–sirolimus-treated group, thus demonstrating a lower risk for skin cancer with sirolimus treatment (Gray test, 5.97; P = .02) (Figure 2). Because skin cancer was the major driver of second posttransplant cancers in this cohort, competing risks regression analysis was focused on identifying independent risk factors associated with second skin cancer formation.
Results of the univariate analysis of risk factors associated with postindex skin cancer formation cancer are shown in eTable 2 in the Supplement. Multivariate analysis to identify predictors of postindex skin cancer formation demonstrated that history of pretransplant skin cancer (subhazard ratio [SHR], 2.1; 95% CI, 1.2-3.7; P = .02), history of the index posttransplant cancer being skin cancer (SHR, 5.5; 95% CI, 2.5-6.4; P < .001), and sirolimus treatment (SHR, 0.6; 95% CI, 0.4-0.9; P = .03) were significantly associated with formation of postindex skin cancer. The same factors were associated with postindex skin cancer formation when multivariate analysis was limited to nonrenal OTRs (history of pretransplant skin cancer [SHR, 2.7; 95% CI, 1.5-5.0; P = .001], history of the posttransplant index cancer being skin cancer [SHR, 3.9; 95% CI, 1.9-7.9; P < .001], and sirolimus treatment [SHR, 0.5; 95% CI, 0.3-1.1; P = .08]), although the association with sirolimus treatment did not reach statistical significance.
In this 2-center mixed-organ cohort of OTRs diagnosed as having posttransplant cancer, 39.5% of patients developed a second posttransplant cancer. Most of the subsequent cancers (88.5%) were skin cancers; therefore, only factors associated with subsequent skin cancer formation could be determined. A history of skin cancer (before or after the transplant) was associated with a higher risk for subsequent skin cancer formation, and the use of sirolimus was protective, with a 40% adjusted reduction in risk (SHR, 0.6; 95% CI, 0.4-0.9) in the overall cohort. When multivariate analysis was limited to the 146 nonrenal OTRs (44.4% of the cohort), the same independent factors were associated with subsequent skin cancer development with sirolimus exposure (P = .08), although not significantly so owing to reduced statistical power. No increase in rejection or mortality was associated with sirolimus use in the overall cohort. Patients with a history of skin cancer (before or after transplant) had a 2- to 5-fold higher risk for subsequent skin cancer formation consistent with the results of prior studies.24,25
Prior studies20,26,27 have shown a reduction in skin cancer risk in OTRs treated with sirolimus. However, these studies have been limited to patients undergoing renal transplant. To our knowledge, this study is the first to show a reduction of skin cancer risk in a cohort of patients with various types of organ transplants and a history of posttransplant cancer. This risk reduction in skin cancer formation was present even after adjusting for prior skin cancer, which predisposes patients to subsequent skin cancer formation. Thus, even for patients who have already had difficulty with skin cancer formation, mTOR inhibition appears to be of benefit. No difference in cancer outcomes was observable between sirolimus-treated and non–sirolimus-treated groups because poor outcomes were rare.
A recent study25 found that sirolimus exposure was not associated with a reduction in incident CSCC after transplant in a mixed-organ cohort of OTRs. Our study is different in that it examines the risk for skin cancer formation in a mixed-organ cohort of OTRs who had already developed some form of posttransplant malignant neoplasm. This patient population had already demonstrated a propensity toward cancer formation and was at higher risk for developing skin cancer than the general OTR population.28,29 Our data support consideration of this subset of OTRs who develop posttransplant cancer for sirolimus chemoprevention. Our results indicate a 40% reduction in the risk for a second skin cancer even after adjusting for a history of skin cancer. In addition, a recent meta-analysis24 showed an increase in overall mortality with sirolimus use, but mortality was not elevated in low-dose sirolimus studies.24 The present study found a 9% reduction in mortality in the sirolimus group (although it did not reach statistical significance, at P = .10). In many early trials,24 patients underwent immediate conversion to mTOR inhibition therapy, and high doses were used. At our institution, patients undergo gradual conversion from traditional regimens to mTOR-based treatment, and maintenance therapy tends to consist of relatively low doses of sirolimus (mean [SD] dose, 2.2 [1.4] mg), with a potential for fewer mTOR-related adverse effects.
Our study has limitations. The study was underpowered to assess the effect of mTOR inhibitor use on cancers other than skin cancer because other cancers were rare. The study was nonrandomized; thus, other factors that were associated with cancer formation may have driven the choice of mTOR use or nonuse, although this seems unlikely. Conversion to mTOR inhibition therapy is inevitably accompanied by discontinuation of therapy with or dose reduction of other immunosuppressive agents; therefore, the differential effect on cancer formation of mTOR treatment initiation vs withdrawal of other agents cannot be ascertained. However, this effect reflects real-world management choices, so the overall impact of mTOR conversion, including the effects of sparing other drugs, has clinical relevance. Finally, the study data are derived from a retrospective electronic medical record review, and adherence to mTOR inhibition and other immunosuppressive therapies could not be assessed. However, adherence should not differ substantially between mTOR and non-mTOR groups; therefore, this lack should have little effect on our results and conclusions.
The present study provides further evidence of the antineoplastic effects of mTOR inhibition on the development of skin cancers in a mixed-organ cohort of OTRs with a history of cancer. This study is among the first to report the effect of mTOR conversion on cancer formation in a mixed-organ cohort of OTRs, and the findings in the nonrenal OTRs paralleled those of the cohort as a whole.
The findings that a history of skin cancer is associated with subsequent skin cancer formation and that mTOR conversion affects this risk highlight the necessity for dermatologists and transplant physicians to be aware of skin cancer history, coordinate regular posttransplant surveillance of skin cancers in OTRs (particularly patients with a history of skin cancer), and have close communication as skin cancers form to consider reduction in immunosuppressive therapy or conversion to an mTOR-based regimen if skin cancer formation is of concern. Conversion to mTOR therapy was not associated with an increased risk for rejection and mortality in this study. Gradual conversion to a low-dose mTOR-based regimen may be considered in patients who develop multiple or high-risk skin cancers to decrease their skin cancer burden. Further studies are needed to define optimal conversion regimens and dosing in such scenarios.
Corresponding Author: Chrysalyne D. Schmults, MD, MSCE, Department of Dermatology, Brigham and Women’s Hospital, Harvard Medical School, 1153 Centre St, Ste 4349, Boston, MA 02130 (email@example.com).
Accepted for Publication: November 11, 2015.
Published Online: January 20, 2016. doi:10.1001/jamadermatol.2015.5548.
Author Contributions: Mr Karia and Dr Schmults had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Karia, Azzi, Heher, Schmults.
Acquisition, analysis, or interpretation of data: Karia, Hills, Schmults.
Drafting of the manuscript: Karia, Hills.
Critical revision of the manuscript for important intellectual content: Karia, Azzi, Heher, Schmults.
Statistical analysis: Karia.
Obtained funding: Karia, Schmults.
Administrative, technical, or material support: Karia, Hills.
Study supervision: Azzi, Heher, Schmults.
Conflict of Interest Disclosure: None reported.
Funding/Support: This study was supported by Novartis Pharmaceuticals.
Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.