Objective To explore the risk of new or recurrent cancer among patients with IBD and previous cancer, exposed or not to immunosuppressants.
Design Among the 17 047 patients of the CESAME prospective observational cohort who were enrolled from May 2004 to June 2005, and followed-up until December 2007, we identified 405 patients with cancer diagnosed previous to study entry. We calculated the rates of incident cancer in patients with or without previous cancer, and we assessed by survival analysis and nested case-control study the impact of immunosuppressants on the risk of incident new or recurrent cancer in patients with previous cancer.
Results The rate of incident cancer was 21.1/1000 patient-years (PY) and 6.1/1000 PY in patients with and without previous cancer, respectively. The multivariate-adjusted HR of incident cancer between patients with and without previous cancer was 1.9 (95% CI 1.2 to 3.0, p=0.003). Among patients with previous cancer, the rates of new and recurrent cancers were, respectively, 13.2/1000 PY and 6.0/1000 PY in the 312 patients who were not taking immunosuppressant at the time of study entry, and 23.1/1000 PY and 3.9/1000 PY in the 93 patients treated with immunosuppressants at study entry. There was no significant association between the exposure to immunosuppressants and the risk of new or recurrent cancer.
Conclusions Patients with IBD with a history of cancer are at increased risk of developing any (new or recurrent) cancer, with a predominant incidence of new cancers. Treatment with immunosuppressants has no overall major impact per se on this risk.
- Cancer Epidemiology
- Crohn's Disease
- Ulcerative Colitis
Statistics from Altmetric.com
Significance of this study
What is already known about this subject?
Immunosuppressants may promote various types of cancers.
In the post-transplant setting, immunosuppressants are associated with a high risk of cancer recurrence in patients with a recent history of cancer.
There are no data on the risk of new or recurrent cancer in patients using immunosuppressants who have IBD and a history of previous cancer.
What are the new findings?
In patients with IBD, a previous history of cancer is an independent risk factor for developing any cancer, in addition to age and exposure to immunosuppressants.
Patients with IBD with a history of cancer are more prone to develop a new cancer than to develop recurrence of the previous cancer.
In Patients with IBD with a history of cancer, exposure to immunosuppressants has no major overall impact on the risk of new cancer or cancer recurrence.
How might it impact on clinical practice in the foreseeable future?
Patients with IBD and previous cancer could be screened for another latent cancer, especially when these patients are older and are considering starting or resuming immunosuppressive therapy.
Reassuring data regarding the risk of cancer recurrence in patients undergoing immunosuppressive therapy suggest that gastroenterologists should not generally contraindicate the use of immunosuppressants in patients with a recent history of cancer and uncontrolled IBD. However, such decisions should always be made on a case-by-case basis together with the oncologist. This decision should take into account the characteristics of the previous cancer (organ, stage, histological type and prognosis), the time since completion of cancer treatment and the individual carcinogenic effects of immunosuppressants.
IBDs are lifelong diseases that primarily affect young patients. Sustained mucosal healing is becoming the standard objective of the long-term treatment of IBD.1 This objective can be achieved currently by maintenance treatment with immunosuppressants, including thiopurines, methotrexate and anti-tumor necrosis factor (TNF) agents.1–3 Unfortunately, there is no major trend towards the spontaneous extinction of IBD activity with time,4 ,5 and disease activity often recurs after withdrawal of immunosuppressants.6 ,7 Collectively, these factors lead to an extensive and prolonged use of immunosuppressants in patients with IBD.
De novo cancers and cancer recurrence may be promoted by immunosuppressants, due to various mechanisms that include decreased immunosurveillance,8 facilitated action of oncogenic viruses9 and direct alteration of DNA.10 These mechanisms differ considerably between immunosuppressants and cancer subtypes.
In the post-transplant setting, prolonged use of immunosuppressants, including thiopurines and calcineurin inhibitors, is associated with an increased risk of de novo cancers.11 Cancer recurrence is also promoted by immunosuppressants.11 The excess risk of cancer recurrence depends primarily on the cancer subtype and on the time between completion of cancer treatment and the initiation of immunosuppressive therapy.12 ,13
In the general population, patients with a personal history of cancer have an overall small excess risk of developing a second cancer14 in addition to the risk of cancer recurrence. The risk of developing a second cancer is more pronounced in individuals who have experienced the first cancer during childhood or adolescence.14 In chronic inflammatory diseases, there are few data on cancer risk in patients with a history of cancer who were either exposed or not exposed to immunosuppressants. The data from two cohorts of patients with rheumatoid arthritis and a history of cancer suggested that the risk of new or recurrent cancer was not significantly different between patients treated with the combination of anti-TNF and methotrexate and patients treated with methotrexate alone.15 ,16
The risk of new or recurrent cancer under immunosuppressants in patients with IBD with a history of cancer is unknown. We addressed this question in the CESAME (Cancers Et Surrisque Associé aux Maladies inflammatoires intestinales En France) French nationwide observational cohort. This cohort was assembled in the early 2000s to assess the risks of any cancer or high-grade dysplasia in patients with IBD and to determine the impact of immunosuppressant use on these risks.17 ,18 In the present study, we identified patients who had been diagnosed with cancer prior to entry into the cohort. We calculated the incidence rates of any cancer in patients with or without previous cancer. Through a multivariate analysis, we assessed the independent roles of cancer history and immunosuppressant use on the risk of incident cancer. We then assessed, through survival analysis and a nested case-control study, the impact of immunosuppressants on the risk of new or recurrent cancer in patients with previous cancer.
CESAME cohort design
The CESAME cohort was designed to assess the risk of any malignancy or high-grade dysplasia in patients with IBD. Details on the design and management of the cohort are available elsewhere.17 Briefly, from May 2004 to June 2005, 680 French gastroenterologists (38.4% with a full-time hospital practice, 24.3% with a mixed public/private practice, and 37.3% with a full-time private practice) enrolled, on an unpaid basis, 19 486 consecutive patients with IBD from their individual practices (see the list of investigators in the online supplementary appendix). There were no exclusion criteria.
Data were collected on an electronic case-report form. The patients’ demographic characteristics, IBD type, date of diagnosis, cumulative disease location (small bowel, colon (more or less than 50% of the total mucosal area affected, estimated from macroscopic or microscopic findings), and anus), previous history of cancer, and exposure to immunosuppressants were recorded at the time of inclusion in the cohort. The gastroenterologists were asked to report all cases of cancer (including high-grade dysplasia) and death among their patients during the follow-up period, and to provide information on each surviving patient, obtained during a final visit between 1 January and 31 December 2007. They were also asked to notify all changes in immunosuppressants’ status at interim visits. The follow-up period ended on 31 December 2007. Follow-up was complete (ie, including the final visit) in 16 459 cases (84.5%), partial (interim visits but no final visit) in 588 cases (3.0%), and nonexistent in 2439 cases (12.5%). The current study was restricted to the 17 047 patients with complete or partial follow-up data. The median follow-up of the study population was 36 months (interquartile range, 32–43 months).
The protocol was approved by the institutional review boards of the French National Society of Gastroenterology, Association François Aupetit (the French IBD patient association), and Groupe d’Etude Thérapeutique des Affections Inflammatoires du Tube Digestif (GETAID). The study was authorised by the French Data Protection Agency (CNIL, registration number #04-1239, dated 6 July 2004). The patients’ specific written informed consent was not required for this observational study.
Characteristics of previous and incident cancers
Previous cancers were defined as those with histological confirmation previous to the day of entry into the CESAME observation period. Investigators provided the following details for all the previous cancers at the time of enrolment: cancer site, histological type and date of histological confirmation. Patients who were diagnosed with cancer within the 2 years before entry into the observational period were defined as patients with a 2-year maximal interval between the histological confirmation of cancer and the entry into the observational period. For patients with a previous history of two cancers, the interval between histological confirmation of the most recent cancer and the entry into the observational period was taken into account.
Incident cancers were defined as those with histological confirmation obtained within the observational prospective study period. The diagnostic histological report of incident cancers was obtained for all patients. Incident cancers that were diagnosed in patients without any history of previous cancer corresponded in all patients to the first incidence of cancer. Among the incident cancers that occurred in patients with previous cancer, new cancers were defined as those that developed in an organ different from the organ associated with the previous cancers, or that developed in the same organ but with an indisputably different histological type (eg, incident skin squamous-cell carcinoma and previous basal-cell carcinoma). Incident cancers were defined as recurrent cancers in the other cases.
We performed a multivariate analysis to determine the independent risk factors for incident cancers in the study population. We developed a Cox multivariate adjusted regression model in which each potential confounder (age, gender, IBD subtype, exposure to any immunosuppressant at the time of entry into the observational period), as well as a history of cancer, were entered. We used conventional methods to test a departure from the proportional hazards assumption that the effect exerted on the hazard by a non-time-dependent variable selected in the multivariate model was constant over time.19
We assessed the impact of exposure to any immunosuppressant at the time of study entry on the risk of new and recurrent cancer in patients with previous cancer. We first assessed the risk of new and recurrent cancer according to immunosuppressant status at the time of study entry. We created Kaplan–Meier survival curves, and we compared the survival free of new and recurrent cancer using a log-rank test between patients who were receiving any immunosuppressant at entry and patients who were receiving no immunosuppressant at entry.
We also assessed the role of the immunosuppressant status at entry using a nested case-control study. This study also examined the association between the shortest interval (less than or equal to 2 years) between the histological confirmation of previous cancer and entry into the study period and the risk of new and recurrent cancer. A conditional logistic regression model using an exact inferential method was estimated. The covariates were exposure to any immunosuppressant at study entry and the maximal 2-year interval between previous cancer incidence and study entry in the cases and their matched controls (four controls per case). Because non-melanoma skin cancers are often considered separately from other cancers in the literature, the analysis was repeated after exclusion of previous and incident cases of non-melanoma skin cancers.
For subgroup comparisons, proportions were compared using Fisher's exact test, and continuous outcomes were compared using the Kruskal–Wallis test. All tests were two-tailed at a 5% significance level.
Study population and follow-up time
Among the 19 486 patients enrolled into the CESAME cohort, 17 047 patients had undergone at least partial follow-up and constituted the study population. At the time of entry into the observational period, 405 (2.4%) patients had a personal history of cancer while 16 642 (97.6%) had no personal history of cancer. The patient characteristics, according to previous cancer status, are provided in table 1. Compared with patients without previous cancer, patients with previous cancer tended, at entry, to be older at study entry (p<0.001), to have ulcerative colitis or unclassified IBD (<0.001), to have a longer IBD duration (p<0.001), and less exposed to thiopurines (p<0.001), and anti-TNF therapy (p<0.001). There was no difference in the gender ratio (p=0.174), exposure to methotrexate (p=0.183) and other agents (p=0.530). The median follow-up time was 36 months (interquartile range: 32–43 months).
Characteristics of previous cancers
There were 405 patients with a personal history of cancer at study entry. There were 16 patients with a previous history of two subsequent cancers involving distinct organs, yielding 421 previous cancers. The details of cancer sites based on the time interval between the histological confirmation of the previous cancer and the entry into the observational period are provided in table 2. Among the patients with previous cancers, colorectal cancers and breast cancers accounted for 42.6% of the cases.
Incidence of cancers according to the previous cancer status
In the total study population, 316 patients developed an incident cancer during the observational period. The incidence rates of cancer according to the age at entry into the cohort and previous cancer status or exposure to any immunosuppressant at entry are shown in figure 1.
Among the 16 642 patients without previous cancer at the time of study entry, 293 developed an incident cancer (ie, the first personal history of cancer) during the observational period, corresponding to an incidence rate of 6.1/1000 patient-years.
Among the 405 patients with previous cancer at study entry, 23 developed an incident cancer (ie, a new or recurrent cancer) during the observational period. This result corresponds to an incidence rate of 21.1/1000 patient-years. Table 3 shows the clinical data of the 23 patients with previous cancer who developed new or recurrent cancer. Most of the patients (n=16) developed new cancers affecting a different organ, or affecting the same organ but with a different histological type (previous basal-cell skin carcinoma and incident squamous-cell carcinoma (n=2)). Among the patients exposed to any immunosuppressant at study entry, one patient had a recurrent meningioma. Among patients not exposed to immunosuppressants at study entry, five patients developed recurrent cancer. There were cases of local breast cancer recurrence, non-Hodgkin lymphoma recurrence, peritoneal regional recurrence of small bowel adenocarcinoma, metastatic progression of a previously controlled prostate cancer, and peritoneal regional recurrence of a surgically treated (nephrectomy) renal cancer with extension to the small bowel. All the patients with previous cancer who were exposed to any immunosuppressant at study entry and who developed new or recurrent cancer were exposed to thiopurines alone at study entry. None of the ten, seven and four patients with previous cancer who were exposed to methotrexate, anti-TNF agents or other immunosuppressants, respectively, developed a new or recurrent cancer.
Independent risk factors of incident cancer in the total study population
The multivariate HR analysis adjusted for different risk factors showed an independent association between a previous history of cancer and the risk of incident cancer (ie, the first cancer among patients without previous cancer, and a new or recurrent cancer among patients with previous cancer) during the observational period (table 4). The multivariate-adjusted HR of incident cancer between patients with and without previous cancer was 1.9 (95% CI 1.2 to 3.0, p=0.003). Other statistically significant risk factors for incident cancers were increased age and exposure to any immunosuppressant at entry into the observational period.
Risk of new or recurrent cancer by immunosuppressant status at entry in patients with previous cancer
Among the 405 patients with previous cancer, 93 (23.0%) were exposed to any immunosuppressant at the time of entry into the observational period, and 312 (77.0%) were not exposed. The mean (SD) age of patients exposed to any immunosuppressant at entry was 51.6 (12.7) years but was 57.7 (14.6) years in patients not exposed to any immunosuppressant. Among the 19 patients with a previous history of non-Hodgkin lymphoma, two were exposed to an immunosuppressant (thiopurines in both cases) at study entry. The first patient had had a Helicobacter pylori-related gastric lymphoma, successfully treated by H pylori eradication. The second patient had a rectal lymphoma, successfully treated by radiation therapy. Among the 21 patients with previous non-melanoma skin cancer, 7 patients were exposed to any immunosuppressant at entry (5 to thiopurines, 1 to anti-TNF therapy and 1 to cyclosporin). All 7 of these patients had a previous history of basal cell carcinoma.
The rates of incident cancer (ie, new or recurrent cancer) were 27.0/1000 patient-years and 19.2/1000 patient-years in patients exposed or not exposed to any immunosuppressant at the time of study entry, respectively. The distribution of the new or recurrent cancer incidences by immunosuppressant status is shown in figure 2.
The probability of remaining free of new or recurrent cancer in patients exposed or not exposed to any immunosuppressant at entry into the observational period is shown in figure 3. There was no significant difference in the incidence of new or recurrent cancers between patients exposed or not exposed to any immunosuppressant.
The results of the nested case-control study are provided in table 5. Being exposed to any immunosuppressant at the time of entry, or having a short (<2 years) interval between histological confirmation and entry into the observational period were not associated with a significant excess risk of new or recurrent cancer in patients with previous cancer. These results were not substantially altered when non-melanoma skin cancers were excluded from the analysis.
This is the first prospective study on the risk of new or recurrent cancer in patients with IBD with previous cancer. The results show that patients with IBD with previous cancer are at increased risk of developing any (new or recurrent) cancer, with a predominant incidence of new cancers. Another important finding is that immunosuppressants have no overall major quantitative impact per se on this risk.
Patients with previous cancer may develop a metastatic lesion or local recurrence in the originally affected organ. They may also develop a new cancer that may be a new primary tumour of the same organ or a primary tumour of another organ. It is difficult in some contexts, such as the diagnosis of ipsilateral subsequent breast cancers, to make the distinction between tumour recurrence and the development of a new tumour.20 Our study design pooled all these forms of tumours under the classification of incident cancer in patients with previous cancer. We found that patients with IBD with previous cancer had a doubled risk of any form of incident cancer (ie, new or recurrent cancer), compared with the risk of an incident first cancer in patients without cancer. Qualitatively, most of the patients with previous cancer developed new cancers rather than cancer recurrence. Similar findings were reported in two separate cohorts of patients with rheumatoid arthritis.15 ,16 Our result is also in accordance with cancer epidemiology studies of the general population. Data from the SEER Cancer Registries show that cancer survivors are at increased risk of new malignancy, partly because of environmental or genetic susceptibilities that are common to the original and new cancers. The overall relative risk is 1.14 in the general population, but it is much higher in patients who developed the initial cancer at a young age.14 This latter feature is attributed to the cancer-promoting role of initial chemotherapy or radiotherapy, and the over-representation of genetic susceptibility in young patients who develop cancer.14
Most of the CESAME patients who were exposed to any immunosuppressant were treated with thiopurine monotherapy. Thiopurines have been shown to promote EBV-related lymphomas and non-melanoma skin cancers in patients with IBD.17 ,18 The main objective of our study was to investigate whether patients with previous cancer are at increased risk of developing of a new cancer or cancer recurrence when exposed to thiopurines. In the post-transplant state, a combination of steroids and azathioprine was used as the standard immunosuppressive post-transplant regimen until the 1980s, when calcineurin inhibitors became available. In this context, the role of thiopurines in promoting first cases of lymphomas and non-melanoma skin cancers was established.11 It was also demonstrated in studies by Penn, that post-transplant immunosuppressive therapy was associated with a high risk of cancer recurrence in patients with previous cancer. This risk exceeded 20% for some cancer subtypes (such as melanomas and non-melanoma skin cancers), and the greatest incidence occurred in the 2 years following completion of the treatment for the initial cancer.12 ,13 In our study, exposure to any immunosuppressant, particularly thiopurines, was an independent risk factor for the development of any incident cancer, together with increased age and a previous history of cancer. By contrast, we did not demonstrate a high risk of cancer recurrence in the subset of patients with previous cancer who were exposed to immunosuppressants. This association held true in the subset of patients with recent cancer (≤2 years), exposed to any immunosuppressant (mainly thiopurines).
The main limitation of our study is the limited statistical power of the analyses when restricted to the subset of patients with previous cancer. We showed a strong statistical trend toward a carcinogenic role of immunosuppressants in the total cohort population but no significant impact of immunosuppressant use on the risk of new or recurrent cancer in the limited population of patients with previous cancer. One could speculate that this apparent paradox is not real and can just be attributed to the reduced statistical power of the latter analysis. However, first, the level of the HR of incident cancer in the total population exposed to immunosuppressants was 1.7, whereas this figure was 1.3 and 0.3 for new and recurrent cancer, respectively, in patients with previous cancer who were exposed to immunosuppressants. Second, patients with previous cancer, either exposed or not to immunosuppressants, had a similar mean age, and thus, the actuarial comparison of the cancer risks according to immunosuppressant use was spontaneously roughly adjusted for the major potential confounder. The actuarial curves were not distinct between the two groups and were crossed-over, suggesting no real major impact of immunosuppressants, which contrasts with the observations in the post-transplant population in the literature. Third, when examining the details of the use of immunosuppressants in patients with previous cancer, we reported indices towards a reasoned and prudent use of immunosuppressants in our study patients. In patients with previous lymphomas and non-melanoma skin cancers, which are considered the most important immunosuppressant-promoted cancers, the resumption of immunosuppressants was restricted to the less severe cases of cancer, that is, localised lymphomas and basal-cell carcinomas. This prudent attitude could constitute one element of the explanation for the absence of obvious deleterious consequences of immunosuppressant use in patients with previous cancer.
We observed no new or recurrent cancer in patients exposed to immunosuppressants other than thiopurines, including methotrexate and anti-TNF agents. However, the number of patients in this situation was too small to enable any conclusions to be drawn.
The overall trend in this study regarding the risk of cancer recurrence under immunosuppressants, mainly thiopurines, in patients with IBD with previous cancer, should be confirmed in a larger study. Other studies should also specifically assess the role of anti-TNF agent, before definitely concluding that gastroenterologists should no longer generally contraindicate the use of immunosuppressants in patients with recently controlled cancer and uncontrolled IBD. However, of major importance in such patients, individual decisions of immunosuppressant use in clinical practice should always be made now on a case-by-case basis with the oncologist. The decision should take into account the characteristics of the previous cancer (organ, stage, histological type and prognosis), the time since completion of cancer treatment, and the individual carcinogenic effects of each class of immunosuppressants.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online appendix
Collaborators See online supplementary appendix for the investigators of the CESAME Study Group.
Contributors LB, FC and LP-B were jointly responsible for the study concept, design and implementation, as well as for the writing of the first draft of the report. FC was responsible for all statistical and sensitivity analyses. A-MB, J-FC, AG and DH participated in the writing of the first draft. JBC and HS participated in histological report collection. LB, FC, J-FC, J-LF, DL and LP-B were study investigators. As an oncologist, AdG provided expertise in the interpretation of the results. All authors took part in the revision of the manuscript.
Funding The CESAME project was supported by grants from Programme Hospitalier de Recherche Clinique National (AOM05157), Association François Aupetit, Délégation Inter-régionale de la Recherche clinique Ile de France-AP-HP, Ligue contre le Cancer, and Fonds de Recherche de la Société Nationale Française de Gastro-entérologie.
Competing interests LB has received consulting fees from Abbvie, lecture fees from Merck and Abbvie, and unconditional research grants form Biocodex, Ferring Pharmaceuticals and Abbvie. J-FC has received consulting fees from Abbvie, Amgen, Biogen Idec, Boehringer-Ingelheim, Bristol–Myers Squibb, Cellerix SL, Celltrion, Chemocentryx Inc, Centocor, Cosmo Technologies, Elan Pharmaceuticals, Genentech, Giuliani, Given Imaging, GlaxoSmithKline, Hospira, Immune Pharmaceuticals, Merck Sharp & Dohme, Millenium Pharmaceuticals Inc., Neovacs, Ocerra Therapeutics, Pfizer, Prometheus, Sanofi, Shire Pharma, Synta Pharma, Takeda, Teva Pharma, Therakos, UCB Pharma, TXCell, Vertex, Wyeth, Dr. August Wolff GmbH & Co; lecture fees from Abbvie, Centocor, Ferring, Given Imaging, Merck Sharp & Dohme, UCB Pharma, Tillotts; grant support from Ferring, Merck Sharp & Dohme, Abbvie; payment for the development of educational presentations from Abbvie, Centocor, Merck Sharp & Dohme; stock options from Intestinal Biotech Development. DL has received consulting fees from Abbvie, Ferring, MSD and Norgine, and act as speaker for Abbvie, Ferring, MSD and Norgine. LP-B has received lecture and consulting fees from Abbott, Merck, UCB-pharma, Tillots, Ferring, Norgine, and Shire FaC, A-MB, J-LF, AdeG, HS and TS disclose no competing interests in relation with this study.
Ethics approval The protocol was approved by the institutional review boards of the French National Society of Gastroenterology, Association François Aupetit (the French IBD patient association), and Groupe d'Etude Therapeutique des Affections Inflammatoires du Tube Digestif (GETAID). The study was authorised by the French Data Protection Agency (CNIL, registration number #04-1239, dated 6 July 2004).
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.