Does the use of immunosuppressive therapy in inflammatory bowel disease increase the risk of developing lymphoma?

Abstract

Recent case reports have raised concerns regarding the risks of non-Hodgkin’s lymphoma in patients with inflammatory bowel disease treated with immunosuppressive agents. This evidence-based review examines this issue from data derived from the use of immunosuppression in other conditions (and inflammatory bowel disease).

We conclude that, in transplant (cardiac and renal) recipients, immunosuppression increases the risk of non-Hodgkin’s lymphoma. For non-transplant patients (with psoriasis and rheumatoid arthritis), debate remains as to whether the observed increase in the incidence of non-Hodgkin’s lymphoma is due to drug or disease. For inflammatory bowel disease per se, population studies show no significant increase in the risk of non-Hodgkin’s lymphoma, with a relative risk of 1.3 (95% confidence interval, 0.9–1.7) compared to expected rates, and several studies of immuno- suppression in inflammatory bowel disease do not appear to confirm a significant rate of lymphoma incidence. Reported cases of lymphoma from single centres should be viewed with caution as evidence of increased risk.

If any association exists, it is likely to be of minimal clinical significance compared to the established and more frequent risks of myelosuppression and infection, and is unlikely to outweigh the benefit of immunosuppression in inflammatory bowel disease.

INTRODUCTION

Immunosuppressive agents, such as azathioprine and 6-mercaptopurine, have been used in the management of inflammatory bowel disease since the 1960s.¹ They are now established effective treatments for both ulcerative colitis and Crohn’s disease,^2–4 particularly as steroid-sparing agents, given the increased concern over the potential long-term side-effects of the use of corticosteroids. Ever since transplant patients on immunosuppressive therapy were shown to have an increased risk of non-Hodgkin’s lymphoma, there has been concern that immunosuppressive use in inflammatory bowel disease might also result in an increased risk of this condition. Two recent reports of non-Hodgkin’s lymphoma and Hodgkin’s disease in patients with inflammatory bowel disease^5, 6 have resurrected this issue, and this article reviews the available data to assess whether immunosuppressive therapy in inflammatory bowel disease really does increase the risk of non-Hodgkin’s lymphoma.

To address this issue, we attempt to answer four questions.

1 What is the risk of developing non-Hodgkin’s lymphoma in immunosuppressed transplant patients?

2 What is the risk of developing non-Hodgkin’s lymphoma in non-transplant (non-inflammatory bowel disease) patients receiving immunosuppression?

3 What is the incidence of non-Hodgkin’s lymphoma and Hodgkin’s disease in inflammatory bowel disease?

4 Is immunosuppression in inflammatory bowel disease associated with an increased risk of non-Hodgkin’s lymphoma?

The aim of this article is not to balance the potential benefits against the calculated risk as this has recently been investigated elsewhere.⁷

Our search strategy used Medline 1966 to the present with the keywords lymphoma and inflammatory bowel disease (combined). Specific searches were also carried out in a similar manner for azathioprine, 6-mercaptopurine, methotrexate and ciclosporin. Limited searches were carried out for rheumatoid arthritis and psoriasis. Further articles not found on Medline, but in lists of references of articles, were used, as well as abstracts presented to the American Gastroenterological Association in the last 2 years.

WHAT IS THE RISK OF DEVELOPING NON- HODGKIN’S LYMPHOMA IN IMMUNOSUPPRESSED TRANSPLANT PATIENTS?

The relative risk of developing non-Hodgkin’s lymphoma in organ transplant recipients is extremely high, although the absolute risk is low. A study in Europe and North America examined the incidence of non-Hodgkin’s lymphoma in over 50 000 transplant recipients between 1983 and 1991 (45 141 renal, 7634 heart).⁸ Patients had received a variety of immunosuppressive regimes, but most contained ciclosporin and/or azathioprine and steroids. One hundred and ninety-four cases of non-Hodgkin’s lymphoma occurred during the first year after transplantation (relative risk, 58.8 against age- and sex-matched population), and 123 cases in the subsequent 5 years (relative risk, 13.7). The relative risk after heart transplantation was three times higher than after renal transplantation, and the incidence of non-Hodgkin’s lymphoma in North America was twice that found in Europe. The incidence of non-Hodgkin’s lymphoma also varied with the type of immunosuppression given (i.e. the combination of ciclosporin and azathioprine increased the risk by 50% against other treatments). It should be noted that higher doses of ciclosporin and azathioprine were used in heart transplant recipients than in renal patients, and that the doses of immunosuppression used in transplant recipients are generally higher than those used in inflammatory bowel disease. In addition, the predilection of the lymphoma for the transplanted organ itself and the confounding effect of the presence of a transplant organ need to be considered in comparisons with inflammatory bowel disease. A recent US study has indicated that the latest immunosuppressive regimes (using tacrolimus, mycophenolate and antibody induction regimes) are also associated with an increased incidence of lymphoma in renal transplant recipients, with an incidence rate of 2–4% in the first 2 years after transplantation.⁹

WHAT IS THE RISK OF DEVELOPING NON- HODGKIN’S LYMPHOMA IN NON-TRANSPLANT PATIENTS RECEIVING IMMUNOSUPPRESSION?

On the basis of the organ transplant experience, we have examined other conditions in which immunosuppressive therapy is often used to investigate whether an increased incidence of non-Hodgkin’s lymphoma also occurs. A UK-based prospective study of 1634 immuno suppressed non-transplant recipients (underlying disorders were a mixture of inflammatory conditions, the commonest diagnoses being rheumatoid arthritis (n=643), inflammatory bowel disease (n=321) and chronic glomerulonephritis (n=253)) was carried out to examine this question.¹⁰ Most of these patients (68%) had received azathioprine, and the follow-up was between 1 and 12 years. Overall, the incidence of any cancer was slightly increased (relative risk, 1.5), but, when categorized for subtype, there was an excess of non-Hodgkin’s lymphoma (relative risk, 10.9) and squamous type skin cancer (relative risk, 5). Four of the six lymphomas were in patients with rheumatoid arthritis, of whom three had received azathioprine and one cyclophosphamide. In this study, there was a 13-fold increased risk of developing lymphoma if suffering from rheumatoid arthritis compared with the general population. A possible confounding factor was the contribution of the underlying disease (rheumatoid arthritis) to the risk of lymphoma. Several large prospective studies have reported an increased risk of non-Hodgkin’s lymphoma in patients with rheumatoid arthritis (see Table 1), even in the absence of immunosuppression.^10–14 The larger Scandinavian population-based studies suggest an approximately two- to three-fold increase in the risk of non-Hodgkin’s lymphoma or Hodgkin’s disease in rheumatoid arthritis patients. More recent data have raised concerns over the risk of developing lymphoma in rheumatoid arthritis patients specifically treated with methotrexate,¹⁵ especially since spontaneous remission of lymphoma has been observed on cessation of treatment in some of these patients.^15, 16 However, the literature on methotrexate is inconsistent. A retrospective study from the Mayo Clinic between 1976 and 1992 cross-indexed all patients with rheumatoid arthritis (n=16 263) and haematological malignancy (n=21 270).¹⁷ The authors found no differences between the patients who had received methotrexate and those who had received other disease-modifying drugs, and no relationship between the peak or cumulative dose and duration of methotrexate therapy and the subsequent development of haematological malignancy.

Table 1.   Studies of non-Hodgkin’s lymphoma (NHL)/Hodgkin’s disease (HD) incidence in rheumatoid arthritis (RA)/psoriasis

The data for the risk of lymphoma in patients with psoriasis per se are more consistent than for rheumatoid arthritis, with two large European studies^18, 19 suggesting small increases in risk for both non-Hodgkin’s lymphoma and Hodgkin’s disease (although one used hospital in-patients as controls¹⁸). Immunotherapy-related reports of lymphoma in psoriasis are sparse and restricted to case reports, mainly for ciclosporin²⁰ and methotrexate;²¹ however, a retrospective review of 224 patients with psoriasis treated with methotrexate found no increased incidence of lymphoma or indeed any internal malignancy.²²

There is also evidence in other inflammatory/autoimmune conditions, such as systemic lupus erythematosus,²³ Sjögren’s syndrome²⁴ and myasthenia gravis,²⁵ of an increased risk of developing non-Hodgkin’s lymphoma, suggesting that immune modulation (primary or secondary), sometimes with subsequent activation of Epstein–Barr virus infection²⁶ (evidence of which is found in many of the described lymphomas), is important in the aetiology of these cases.

So, how much of this is relevant to the gastroenterologist faced with treating an inflammatory bowel disease patient with immunotherapy? In rheumatoid arthritis and psoriasis, the data suggest that the disease itself is associated with an increased risk of developing non-Hodgkin’s lymphoma, irrespective of treatment. However, given the evidence of the spontaneous regression of Epstein–Barr virus-driven lymphomas on cessation of treatment, there may be an additional aetiological role for immunosuppressive drugs.

WHAT IS THE INCIDENCE OF NON-HODGKIN’S LYMPHOMA AND HODGKIN’S DISEASE IN INFLAMMATORY BOWEL DISEASE?

Lymphoma has long been recognized in association with inflammatory bowel disease in the absence of immunotherapy.²⁷ The first major series suggesting an increased incidence came from Greenstein et al. in 1985.²⁸ In their retrospective analysis of 1961 inflammatory bowel disease patients at the Mount Sinai Hospital, they found an excess of lymphomas in patients with inflammatory bowel disease (n=8; relative risk, 4.8). In 1992, they described nine cases of lymphoma (eight extraintestinal) occurring in patients with ulcerative colitis (n=5) and Crohn’s disease (n=4) in a longer follow-up of this cohort.²⁹ Four of the nine had previously received steroids or sulfasalazine, but none had received antimetabolites. Studies from other tertiary referral centres support the observation of an increased risk,^5, 30 with estimates as high as 31.2 for non-Hodgkin’s lymphoma in patients with inflammatory bowel disease,⁵ although others have found no significantly increased risk.³¹ One of the considerations that should be made when interpreting these data on the incidence of lymphoma in inflammatory bowel disease is the question of ascertainment bias, and the publication of case series from single centres. To avoid this potential bias, it is best to examine prospective population-based data (Table 2), where incidence rates (usually derived from National Cancer Registries) are compared with age- and sex-matched population controls, thereby overcoming potential problems of using hospital-based controls. Five of the eight studies (Table 2) come from Scandinavia,^32–36 with the others from Italy,⁶ the UK³⁷ and the USA.³⁸ Most of these studies have identified cases of cancer through National Cancer Registries, thereby avoiding ascertainment bias.

Table 2.   Population studies of lymphoma incidence in inflammatory bowel disease

Of 31 873 patients with either Crohn’s disease or ulcerative colitis, 49 had lymphoma (six Hodgkin’s disease; 43 unclassified), giving (assuming a Poisson distribution) a relative risk of non-Hodgkin’s lymphoma in inflammatory bowel disease of 1.3 (95% confidence interval (CI), 0.9–1.7). When Hodgkin’s disease is specified as the type of lymphoma, we have calculated a relative risk of 4.0 (95% CI, 1.5–8.7); however, there was a positive association between Hodgkin’s disease and inflammatory bowel disease in only one of these studies (four cases, none of which incidentally had received immunosuppressive therapy).⁶

Although these studies did not all consider the effect of treatment, they originate from different geographical settings over a period of more than 50 years, allowing us to conclude that the overall incidence of non-Hodgkin’s lymphoma in inflammatory bowel disease does not appear to be significantly increased. The opposite may apply for Hodgkin’s disease, but this is primarily based on the results of one study.

IS IMMUNOSUPPRESSION IN INFLAMMATORY BOWEL DISEASE ASSOCIATED WITH AN INCREASED RISK OF NON-HODGKIN’S LYMPHOMA?

There are several studies that have specifically examined the question of whether immunosuppression leads to an increased risk of non-Hodgkin’s lymphoma in inflammatory bowel disease (Table 3).^37, 39–43 These studies are heterogeneous in that the doses and type of immunosuppression, as well as the total dosage and follow-up, vary widely.

Table 3.   Previous studies of inflammatory bowel disease, use of immunosuppression and incidence of lymphoma

In the first report from the UK, in 1985, no cases of lymphoma were found in 327 prospectively followed inflammatory bowel disease patients (mean follow-up is not given nor details of the exact immunosuppression, although many of the patients in the study had received azathioprine at some point).¹⁰ A New York study initially published in 1989,³⁹ and 10 years later with continued follow-up of patients treated with 6-mercaptopurine,⁴⁰ found only two cases of lymphoma, and only one new case was diagnosed in the 7471 patient-years of follow-up between the two publications. In a study from the UK examining the safety of azathioprine, no cases of lymphoma were found in almost 7000 years of patient follow-up,⁴¹ but data from Oxford looking at the long-term risk of malignancy in inflammatory bowel disease identified seven cases of lymphoma, three of whom had received azathioprine;⁴³ one of these had rheumatoid arthritis and ulcerative colitis. In a retrospective study from the Mayo clinic, 61 cases of lymphoma were found in 15 000 inflammatory bowel disease patients, only three of whom had received immunotherapy (others had also been immunosuppressed but had received liver transplantation).³⁸ The total number of patients who had received immunosuppression from the original 15 000 is not stated (hence the omission of this study from Table 3), but the fact that only 5% of the identified lymphomas could be linked with the use of immunosuppression is reassuring.

In total (Table 3), only 11 cases of lymphoma are described in over 17 000 years of follow-up after a variety of immunosuppressive regimens. This sizeable experience contrasts with a recent paper reporting four cases in 238 patients treated with immunosuppressive therapy between 1990 and 1999.⁵ It is of note that the inflammatory bowel disease register used in this study was set up in mid-1996, by which time three of the four cases had already been diagnosed.

Therefore, we conclude that immunosuppression in patients with inflammatory bowel disease does not appear to be associated with a clinically significant increased rate of lymphoma.

CONCLUSIONS

In organ transplant recipients, immunosuppression increases the risk of non-Hodgkin’s lymphoma (especially in the transplanted organ). For patients with rheumatoid arthritis and psoriasis, immunosuppressive therapy may be associated with a small increase in risk of non-Hodgkin’s lymphoma, but debate remains as to whether the increased risk is due to drug or disease.

For patients with inflammatory bowel disease, neither the background disease nor immunotherapy appears to be associated with an increased risk of non-Hodgkin’s lymphoma.

Reported increases of lymphoma in inflammatory bowel disease from single centres should be viewed with caution as evidence of increased risk. If any such association does indeed exist, it is likely to be of minor importance compared to the established and more frequent risks of immunosuppression, such as myelosuppression and infection.

ACKNOWLEDGEMENT

The authors would like to thank Professor Richard Logan for his useful comments and contributions to the preparation of this review.