This article describes the cancer risks of commonly used inflammatory bowel disease (IBD) medications, with an emphasis on hematologic malignancy risks. The increasing use of immunosuppressant therapies in the treatment of IBD has raised this question to an even greater importance. Studies evaluating these medications are complicated due to varying disease severity and concomitant use of other immunosuppressant medication. The potential risks of all therapies must be weighed against the benefits these therapies can offer these patients.
An increasingly daunting challenge in the care of patients with inflammatory bowel disease (IBD) is managing the risks associated with medical therapies. One of the most worrisome issues for patients is the risk of cancer associated with their medical therapy. In this article, the cancer risks of commonly used IBD medications are described, with an emphasis on hematologic malignancy risks. Lymphoma, particularly non-Hodgkin lymphoma (NHL), has been associated with other immune-mediated diseases (eg, rheumatoid arthritis [RA] and lupus) and with states of immunosuppression, whether as a consequence of the underlying disease (eg, acquired immune deficiency syndrome [AIDS]) or the treatment (eg, organ transplantation). The increasing use of immunosuppressant therapies in the treatment of IBD has raised this question to an even higher level of importance for patients and physicians.
Does IBD increase the risk of lymphoma independent of medical therapies?
Before considering whether certain medical therapies increase the risk of malignancy, one must first determine whether the baseline risk of cancer among patients with IBD is elevated by virtue of their disease. Other immune-mediated diseases, such as RA, are associated with an increased risk of lymphoma that seems to be independent of the medical therapy. One could hypothesize that IBD patients have a similarly elevated risk of lymphoma due to chronic inflammation from their underlying IBD.
Several studies have sought to evaluate this underlying risk, with conflicting results ( Table 1 ). In evaluating these studies, 2 important features should be considered. First, it is important to assess the source of the IBD patients being studied. Population-based studies evaluate all IBD patients in a region. In contrast, tertiary-referral–based studies evaluate IBD patients referred to specialty centers. These patients may be under closer surveillance by their physicians, leading to over-reporting or earlier diagnosis of malignancies. They may also have other coexisting diseases in addition to their IBD, thus inflating any underlying association with lymphoma. Patients in tertiary referral centers may also have had a longer duration of disease, more severe disease, and may be more likely to have received immunosuppressive therapies. For example, in the study by Farrell and colleagues, which showed a significantly elevated risk of lymphoma in IBD, all 4 cases occurred in patients on immunosuppressive therapy, often on multiple immunosuppressants. In RA, it has been shown that the risk of lymphoma is greater in those with more severe arthritis, independent of the medical therapy. Such studies have not been performed in IBD. However, referral-based studies have tended to observe a higher relative risk of lymphoma among IBD patients than population-based studies (see Table 1 ). One interesting comparison came from a single study from Olmsted County and the Mayo clinic. In their population-based study of 454 IBD patients identified in Olmstead County between 1940 and 1993, they found only a single case of NHL in 6662 patient-years, resulting in a crude risk of 0.002 in IBD patients in Olmstead County and an incidence rate of 15 cases per 100,000 person-years. In an effort to clarify the clinical features and outcomes of IBD patients, all patients with IBD and lymphoma seen at the Mayo Clinic between 1976 and 1997 were also studied. Among approximately 15,000 patients seen at the Mayo Clinic during this time, 61 had a lymphoma, resulting in a crude risk of lymphoma in these tertiary-referral patients of 0.004, twice that seen in the population-based study. Although these are only rough calculations, these divergent findings from a single study illustrate two critical points. First, there was a higher risk of lymphoma in the tertiary-based study than in the population-based study. Second, although there was an elevated risk, the absolute risk remained low.
| Study | Type of Study | N | Cases of Lymphoma CD | Cases of Lymphoma UC | SIR, RR or OR in CD Patients (95% CI) | SIR, RR or OR in UC Patients (95% CI) |
|---|---|---|---|---|---|---|
| Greenstein et al | H | 1227 CD; 734 UC | 3 | 3 | 4.69 ( P <.05) | 8.82 ( P <.005) |
| Ekbom et al | P | 1655 CD; 3121 UC | 1 | 8 | 0.4 (0.0–2.4) | 1.2 (0.5–2.4) |
| Persson et al | P | 1251 CD | a | N/A | 1.4 (0.4–3.5) | — |
| Kerlen et al | P | 1573 UC | N/A | 3 | N/A | 1.2 (0.3–3.5) |
| Mellemkjaer et al | P | 2645 CD | 4 | N/A | 1.5 (0.4–3.7) | N/A |
| Loftus et al | P | 216 CD; 238 UC | 1 | 0 | 2.4 (0.1–13.1) | 0 (0–6.4) |
| Palli et al | P | 231 CD; 689 UC | 2 a |
| 2.5 (0.28–9.0) a |
|
| Farrell et al | H | 267 CD; 515 UC | 1 | 3 | 31.2 (2.0–85) b | 31.2 (2.0–85) b |
| Bernstein et al | P | 2857 CD; 2672 UC |
|
| 2.40 (1.17–4.97) | 1.03 (0.47–2.24) |
| Lewis et al | P | 7988 CD; 12,185 UC | 7 | 11 | 1.59 (0.63–3.29) | 1.20 (0.59–2.15) |
| Winther et al | P | 1160 UC | N/A | 2 | — | 0.51 (0.06–1.82) |
| Askling et al | P | 20,120 CD; 27,559 UC | 65 | 87 | 1.3 (1.0–1.6) | 1.0 (0.8–1.3) |
| Goldacre et al | P | 5127 CD; 6990 UC | 6 | 14 | 1.01 (0.37–2.20) | 1.19 (0.64–2.01) |
| Hemminki et al | P | 27,606 UC | — | 75 | — | 1.52 (1.20–1.91) |
| von Roon et al | M | 36,576 CD | — | — | 1.42 (1.16–1.73) | — |
a Included 1 case of HD, 1 case of myeloma; no cases of NHL.
Another important aspect in evaluating these studies is the ability to control for concomitant medication use. Because patients with the most severe disease are also more likely to be treated with more potent immunosuppression, it is difficult to separate these factors. Furthermore, most population-based studies do not have the ability to control for medication use. In a meta-analysis of cancer risk in Crohn disease (CD) patients, the pooled relative risk of developing lymphoma was 1.42 (95% CI 1.16–1.73). However, when evaluating the two studies within the meta-analysis in which patients were specifically only on corticosteroids, the pooled relative risk of developing lymphoma in 9462 patients was 2.01 (95% 1.17–3.46).
Although the estimated relative risk of lymphoma in ulcerative colitis (UC) and CD may differ from study to study, most population-based studies suggest minimal, if any, increased risk of lymphoma compared with the general population. Furthermore, in the general population, lymphoma is rare. For example, in the United States, the annual incidence is approximately 22 per 100,000 people per year. As a result, in the IBD population as a whole, the absolute risk of lymphoma is low.
Immunosuppression and lymphoma risk
Defective host immune function increases the risk of lymphoma, as shown by the increased risk of lymphoma in patients with human immunodeficiency virus (HIV) and AIDS populations. Elevated risks for lymphoma have been found in HIV-positive patients with even modest levels of immunosuppression, before the onset of AIDS. Furthermore, the risk of lymphoma has been found to be inversely related to the CD4 count in several studies, lending additional credence to the concept that a deficient immune system enhances oncogenesis.
The post-transplant population has broadened the link between immunosuppression and malignancy to incorporate medications as causal mechanisms. By reducing components of the host immune system, and therefore undermining the immunologic surveillance of tumor cells, it is believed that chronic use of immunosuppressive medications can prevent early elimination of malignant cells. This has specifically been implicated in the prevalence of Epstein-Barr virus (EBV)–positive lymphomas in immunocompromised patients. EBV is a human herpesvirus that infects more than 90% of the population, with the primary infection usually occurring in childhood. After this primary infection, which can range from asymptomatic forms to overt mononucleosis, EBV-infected B-lymphocytes exist in a life-long asymptomatic latent infection that is typically controlled by cell-mediated host immunosurveillance including natural killer cells, and CD4+ and CD8+ T cells. Immunocompromised states, including congenital syndromes and medically induced immunosuppression, remove the host’s control and allow the emergence of lymphoproliferative disorders associated with EBV infections. Data from the transplant population also support the hypothesis that the risk of lymphoma is directly correlated with the degree of immunosuppression, and that lymphoma can develop shortly after the onset of immunosuppressive therapy. Many of these tumors completely regress when the immunosuppressive therapy is reduced or stopped.
Immunosuppression and lymphoma risk
Defective host immune function increases the risk of lymphoma, as shown by the increased risk of lymphoma in patients with human immunodeficiency virus (HIV) and AIDS populations. Elevated risks for lymphoma have been found in HIV-positive patients with even modest levels of immunosuppression, before the onset of AIDS. Furthermore, the risk of lymphoma has been found to be inversely related to the CD4 count in several studies, lending additional credence to the concept that a deficient immune system enhances oncogenesis.
The post-transplant population has broadened the link between immunosuppression and malignancy to incorporate medications as causal mechanisms. By reducing components of the host immune system, and therefore undermining the immunologic surveillance of tumor cells, it is believed that chronic use of immunosuppressive medications can prevent early elimination of malignant cells. This has specifically been implicated in the prevalence of Epstein-Barr virus (EBV)–positive lymphomas in immunocompromised patients. EBV is a human herpesvirus that infects more than 90% of the population, with the primary infection usually occurring in childhood. After this primary infection, which can range from asymptomatic forms to overt mononucleosis, EBV-infected B-lymphocytes exist in a life-long asymptomatic latent infection that is typically controlled by cell-mediated host immunosurveillance including natural killer cells, and CD4+ and CD8+ T cells. Immunocompromised states, including congenital syndromes and medically induced immunosuppression, remove the host’s control and allow the emergence of lymphoproliferative disorders associated with EBV infections. Data from the transplant population also support the hypothesis that the risk of lymphoma is directly correlated with the degree of immunosuppression, and that lymphoma can develop shortly after the onset of immunosuppressive therapy. Many of these tumors completely regress when the immunosuppressive therapy is reduced or stopped.
Classes of medications in IBD and their lymphoma risk
5-Aminosalicylate Medications
To date, there has been no association between the 5-aminosalicylate (5-ASA) class of medications and lymphoma in IBD. 5-ASA is a major metabolite of salicylazosulfapyridine (SASP). In rodent models and in vitro studies, SASP, 5-ASA, and sulfapyridine (SP) have not been demonstrated to exhibit mutagenicity or DNA reactivity, nor were they found to be genotoxic.
Corticosteroids
Corticosteroids, specifically oral prednisone, have been a mainstay in the therapy for IBD. Despite having numerous well-known side effects, corticosteroids have not been associated with an increased risk of lymphoma. Corticosteroids have long been a standard medication in first-line and salvage therapy for NHL. Of theoretical interest, however, is the possibility of masking evolving malignancies by the use of corticosteroids in IBD. In addition, because of the widespread immunosuppressant effects of corticosteroids, combination therapy with other immunosuppressive medications could theoretically potentiate the risk of immunosuppression-related lymphoma.
Cyclosporine
Cyclosporine (CsA) is a calcineurin inhibitor that blocks production of interleukin 2 by activated CD4-positive T cells. Although not directly mutagenic, it may impair the hosts’ immunosurveillance for transformed cells and has been implicated in EBV-positive lymphoproliferative disorders. The number of IBD patients treated with CsA is too small to precisely estimate the risk of rare events such as malignancies associated with this medication. However, there is greater experience in the fields of dermatology and organ transplantation. A worldwide study of more than 5000 organ transplant patients found that 0.5% of men (10 men in total) and fewer than 0.1% of women (1 female patient) developed lymphoma, with differing latency periods from time of drug therapy to malignancy diagnosis for different transplant indications. This translated into an approximately 28-times higher risk of lymphoma in patients treated with CsA. A 5-year prospective cohort study of 1252 patients with psoriasis treated with CsA found no trend toward an increased incidence of malignancies over time within the cohort. Two patients developed lymphoma in the study, translating into a standardized incidence ratio (SIR) of 2.0 (0.2–7.2), similar to a previously reported incidence in moderate to severe psoriasis patients not exposed to CsA. However, the investigators conceded that this cohort was not large enough to adequately assess this risk.
In the IBD literature, there is a paucity of data on lymphomas associated with CsA therapy. A hospital-based study followed 782 patients for 8 years at a single institution. Of those patients, 238 received immunosuppressive therapy. Four cases of NHL were reported, all of which occurred in patients on immunosuppression. One of these patients had received 5 months of methotrexate (MTX) therapy followed by 12 months of CsA and ultimately underwent a colectomy. At surgery, the patient was found to have a lymph node with pathology consistent with large cell diffuse B-cell type lymphoma. No other evidence of lymphoma was found and the patient required no treatment. Two additional case reports in IBD patients with CsA exposure have been published. The first was a case report of rectal lymphoma reported in a patient with a 13-year history of UC who had been treated with CsA for 4 years. Therapy with CsA was discontinued and the patient was treated medically with good response. Another case report involved a UC patient who was found to have B-cell lymphoma after ileal pouch-anal anastomosis (IPAA) surgery in the pouch and diffusely in the lymph nodes. The patient had only a 4-year duration of IBD before surgery, and had been treated with a combination of 4 immunosuppressant therapies including CsA, 6-MP, and infliximab.
As noted previously, the hypothesis that immunosuppression directly increases the risk of lymphoma is supported by observations of lymphoma regression with discontinuation of immunosuppression therapy. In the setting of organ transplantation and treatment of dermatologic disorders, complete remission of lymphomas associated with CsA treatment has occurred after dose reduction or discontinuation of the drug.
Methotrexate
MTX is a structural analogue of folic acid that inhibits the activity of dihydrofolate reductase, the enzyme responsible for converting folic acid to folate cofactors. This inhibition, in turn, inhibits DNA synthesis. MTX is cytotoxic and can induce immunosuppression and leucopenia. It has been implicated in impairing cellular immune control of tumor proliferation, and specifically immune control of EBV-induced B-cell proliferation. Animal studies examining this primary carcinogenic effect of MTX have shown a trend toward increased rates of malignant lymphomas in Swiss mice and Syrian hamsters, although numbers were small and no significant difference was detected. In a second animal study, buccal pouches of hamsters injected with MTX showed increased rates of chemically induced carcinomas.
MTX has been used for several decades to treat psoriasis and RA, and most studies of MTX-associated lymphoproliferative disease come from these specialties. In examining these rates, it is important to appreciate that there seems to be an increased risk of lymphoma at baseline in patients with connective-tissue disorders as a result of the disease itself. The theory is that prolonged stimulation of lymphoid tissue by the underlying disease inflammation induces lymphoma development, and it is believed that this explains the 2- to 3-fold increased risk of lymphoma in RA patients who are not receiving any immunosuppressive therapy. However, a similar increased rate of malignancy at baseline in psoriasis patients has been questioned.
Numerous case reports of lymphoproliferative disorders occurring in patients with RA and psoriasis treated with MTX have been published, including several cases of EBV-positive lymphomas. However, several cohort and case-control studies have not shown an elevated risk of lymphoma in patients with RA and psoriasis treated with MTX compared with those not treated with MTX. In psoriasis, in a multicenter prospective cohort study that included 28,554 person-years of follow-up, the incidence of lymphoma in patients with psoriasis who were not exposed to MTX was comparable to the general population (incidence rate ratio [IRR] 0.85, 95% CI 0.27–1.67). However, the investigators of this study did find an elevated risk of lymphoma in patients with psoriasis who were treated with MTX for 3 or more years (IRR 3.74, 95% CI 1.61–7.36). In contrast, a matched case-control study of 26 patients with severe psoriasis and noncutaneous cancer and 104 matched psoriasis controls nested within a cohort of 1380 patients found a relative risk of noncutaneous cancer of 0.95 (95% CI 0.4–2.2). The only 2 cases of leukemia or lymphoma occurred in patients with psoriasis who were not treated with MTX. The RA literature also supports no increased risk of lymphoma associated with MTX. A large study of 18,572 patients with RA compared overall rates and MTX-associated rates of lymphoma to those expected from the survey, epidemiology and end results (SEER) cancer database. The rate of lymphoma for patients with RA who were treated with MTX specifically was nonsignificantly increased compared with expected rates in the United States (SIR 1.7, 95% CI 0.9–3.2). In this same study, patients who had not been exposed to MTX or anti-TNF therapies had no increased incidence of lymphoma (SIR 1.0, 95% CI 0.4–2.5). Two studies compared MTX-treated RA with their nonexposed RA counterparts. In a comparison of patients associated with MTX, those not associated with MTX, and those with sporadic lymphoproliferative disease compiled from several studies, RA patients exposed to MTX had a shorter time between their RA and lymphoproliferative disorder diagnosis, but no other significant differences were found. In a nested case-control study of 23,810 patients with RA from Canada, a nonsignificant adjusted relative risk of developing lymphoma was found for patients with exposure to MTX (risk ratio [RR] = 1.23, 95% CI 0.99–1.40). These studies suggest that among patients with RA and psoriasis, if there is any increased risk of lymphoma associated with MTX therapy, the increase is likely small, although long-term therapy may be associated with a higher risk.
In the IBD literature there is limited data evaluating the risk of lymphoproliferative disease associated with MTX use. In a single-center study of 782 IBD patients, 238 of whom were receiving immunosuppression, 4 cases of NHL were seen, all of which occurred in immunosuppressed IBD patients. Two of the 4 cases had exposure to MTX: 1 patient had been on MTX for 4 years, the other had received MTX for 5 months followed by 12 months of CsA. These data translated into a statistically significant overall risk of NHL in the population (SIR 31.2, 95% CI 2.0–85) and substantially increased odds of NHL in immunosuppressed IBD patients. However, issues of power and selection bias, given the hospital base of this study, confound interpretation of these results.
6-Mercaptopurine/Azathioprine
Azathioprine (AZA) is a prodrug that yields 6-mercaptopurine (6-MP). The final active metabolite of these compounds is believed to be 6-thioguanine (6TG), which becomes incorporated into ribonucleotides and causes an antiproliferative effect on mitotically active lymphocytes. It is also believed that 6-MP and AZA may directly inhibit cytotoxic T cell and natural killer cell function, yielding defective cell-mediated immune surveillance and allowing the development of lymphoproliferative disorders, including EBV-infected lymphocytes, to proliferate out of check.
Transplant and RA literature have pointed to an increased incidence of lymphoma in patients treated with the thiopurine analogues. An early large collaborative study of 3823 renal transplant patients reported a nearly 60-fold increased risk of NHL in patients exposed to AZA. Fifteen of the NHLs affected the brain and presented as space-occupying lesions, an observation noted in prior transplantation literature. A more recent multicenter study of 6500 transplant patients found a lower, but still significantly elevated, risk of post-transplant lymphoproliferative disorder (PTLD) in those with AZA exposure. However, in many of these studies, patients were concomitantly receiving multiple immunosuppressive medications or receiving dosages much higher than those given in nontransplant populations. Several studies of patients with RA have also shown increased risks of NHL in patients exposed to AZA or 6-MP with estimates ranging from a 3- to 13-fold increased risk of reticuloendothelial cancers. A large nested case-control study found no increased risk of lymphoma associated with AZA after adjustment for concomitant medications. When evaluating the rheumatologic studies, issues of baseline risk of cancer in this patient population and increased dosages of AZA used in some of these studies must be taken into account.
Several cohort studies have also been performed examining the risk of lymphoma in IBD patients treated with thiopurine analogues ( Table 2 ). A study from Mount Sinai in the late 1980s evaluated their experience with 6-MP in their IBD patients in the previous 18-year period. Twelve neoplasms were seen during this period, including 1 diffuse lymphoma of the brain in a 53-year-old patient with long-standing CD who had been treated with 6-MP for approximately 9 months. Connell and colleagues conducted a prospective study of 755 patients with IBD treated with 2 mg/kg AZA for a median of 12.5 months between the years of 1962 and 1991. They found no excess cancer in the patients treated with AZA and no cases of NHL. A study from France surveyed 157 CD patients treated with AZA/6-MP for at least 6 months. A large percentage (27%) of these patients stopped taking their medication during the 20-year follow-up period. Four malignancies were seen in the follow-up period including 1 case of CNS lymphoma that occurred in a 37-year-old patient who had been on AZA for 17 months. A study of 6-MP in 90 patients with chronic refractory UC found 6 malignancies (including 3 colonic neoplasms) in 368 cumulative patient-years, but no cases of NHL. In a single-center study of 550 patients taking 6-MP for an average of 5 years, 2 cases of NHL developed, including 1 cerebral and 1 abdominal.
| Study | Type of Study | Drug Therapy | Number of Patients Studied | Mean Duration of Therapy (Months) | Number of Cases of Lymphoma | SIR, RR or OR (95% CI) |
|---|---|---|---|---|---|---|
| Kinlen | H | AZA, 6-MP | 321 IBD | N/R | 2 | 12.5 (1.2–46.0) |
| Present et al | H | 6-MP |
|
|
| N/R |
| Connell et al | H | AZA |
| 12.5 (median) | 0 | 0 |
| Bouhnik et al | H | AZA, 6-MP | 157 CD | 25 | 1 a | N/R |
| George et al | H | 6-MP | 90 UC | 42 | 0 | 0 |
| Korelitz et al | H | 6-MP |
| 60 |
| 4.9 (0.9–14.5) |
| Farrell et al | H | AZA | 238 IBD | 21.6 | 2 | 37.5 (3.5–138) |
| Lewis et al | P | AZA/6-MP |
| 24 |
| 1.6 (0.0006–9.0) |
| Dayharsh et al | P | AZA/6-MP | 1200 IBD | 42 (median) | 6 b | N/R |
| Fraser et al | H | AZA |
| 27 |
| 4.6 (0.9–13.7) |
| Glazier et al | H | 6-MP |
| 27 |
| N/R |
| Kandiel et al | M | Pooled analysis | 3891 IBD | N/R | 11 | 4.18 (2.07–7.51) |
a Included 1 case of CNS lymphoma.
b Reported 18 total lymphomas, 6 of which occurred in patients with AZA/6-MP exposure.
Farrell and colleagues reported a dramatically increased incidence of lymphoma in IBD patients treated with immunosuppressive medications. In their tertiary care center–based retrospective study of 238 IBD patients, they reported 4 cases of NHL, 2 of which were on AZA. Their overall reported SIR for NHL in patients receiving AZA was 37.5 (95% CI 3.5–138). However, in a population-based study using the General Practice Research Database (GPRD) from the United Kingdom evaluating 1465 IBD patients treated with the thiopurine analogues, only 1 case of Hodgkin lymphoma was found in a patient with UC who had received 1 prescription for AZA 10 months earlier. This resulted in a nonsignificant SIR of 1.57 (95% CI 0.04–8.75) that remained nonsignificant even after sensitivity analyses were performed. A 15-year review from the Mayo Clinic identified all IBD patients who developed lymphoma and then divided the patient population into 2 8-year intervals (1985–1992, 1993–2000) corresponding to the introduction of thiopurine analogue therapy in 1993. They identified 18 patients with lymphoma, 6 within the first interval and 12 occurring in the second interval. Fifty percent of the lymphomas occurring between 1993 and 2000 occurred in patients treated with thiopurine analogues; and of these 6 patients, 5 were EBV-positive. This figure was in comparison to only one EBV-positive lymphoma occurring during the first period. The investigators also noted that 5 of the 6 lymphomas arising in the AZA/6-MP–treated patients were extranodal lymphoma, in comparison with 5 of the 12 lymphomas in the unexposed patients. In addition, 5 of the 6 lymphomas in the AZA/6-MP–treated patients were diffuse large B-cell type in comparison with only 4 of the 12 lymphomas in the unexposed patients. Although the investigators did not evaluate dosages or duration of treatment during their study, they did estimate that from 1993 to 2000 they treated approximately 1200 patients with AZA or 6-MP, giving an estimated risk of developing EBV-positive lymphoma of 0.5%. Two retrospective chart reviews from 2 different hospital-based studies observed no significantly increased risk of NHL in AZA/6-MP treated patients. Fraser and colleagues evaluated 626 patients and found 3 cases of NHL, a rate that was not significantly different from IBD patients not receiving thiopurine analogue therapy. A 10-year single-center experience of 285 IBD patients found a single case of large B-cell colonic lymphoma in a 44-year-old patient with a history of Crohn colitis.
A meta-analysis was performed using 6 studies evaluating the risk of NHL in IBD patients treated with AZA or 6-MP. When pooling the data from the 6 studies, the investigators observed a significant increased risk of lymphoma in IBD patients treated with thiopurine analogues compared with the expected rates in the general population (SIR 4.18 95% CI 2.07–7.51). There was significant heterogeneity among the studies but excluding any 1 study did not appreciably affect their results, with statistically significant SIRs ranging from 2.90 to 5.21. When evaluating studies that directly compared IBD patients treated with thiopurine analogues with those who were not treated with thiopurine analogues, the investigators found a lower, but still significantly elevated, relative risk of NHL in IBD patients treated with thiopurine analogues.
Most of these studies have been single-center or hospital-based, raising the issues of ascertainment and referral bias. In addition, there is potential bias in studies from tertiary care centers if the cases of lymphoma that prompted the investigators to undertake the study were included in the study cohort. It is also unclear whether the increased incidence of lymphoma in patients receiving AZA or 6-MP is due to the immunomodulators or is a consequence of the more aggressive or refractory nature of disease in patients who tend to receive these medications.
Overall, treatment with thiopurines seems likely to be associated with an increased risk of lymphoma, particularly EBV-associated lymphoproliferative disorder. The magnitude of this risk is likely on the order of a 3- to 5-fold relative risk. Several unanswered questions remain, such as whether the risk persists after the medication is discontinued, and whether combination therapy with other immunosuppressants potentiates the risk.
Anti-TNF Therapy
Three anti-TNF agents, infliximab, adalimumab, and certolizumab, have been approved for the treatment of IBD. In addition, etanercept is approved for the treatment of RA and psoriasis in the United States. Shortly after the approval of the first three anti-TNF agents, questions were raised regarding whether these agents may increase the risk of lymphoma.
Infliximab
Infliximab, previously known as cA2, is a chimeric monoclonal antibody directed against human tumor necrosis factor (TNF). It consists of a linkage of variable regions of mouse antihuman TNF monoclonal antibody to human IgG1 with k light chains. Mouse studies have shown that infliximab binds to and neutralizes soluble TNF in vitro and protects against the effects of TNF in vivo. Infliximab has also been shown to bind to the transmembrane form of TNF with a high specificity, affecting down-regulation of several inflammatory mediators. Like many of the immunosuppressive medications, infliximab was initially used extensively for RA. Initial clinical trials did not observe an increased incidence of lymphoma in RA patients treated with infliximab, although these studies were not designed to specifically address this question and were underpowered for this outcome. More recent population-based prospective observational studies have suggested a possible increased risk of lymphoma among RA patients treated with anti-TNF agents. Investigators from the National Data Bank for Rheumatologic Disease compared outcomes in 18,572 RA patients to the SEER cancer database matched for age and gender. The investigators found an overall SIR for lymphoma with biologic use of 2.9 (95% CI 1.7–4.9). For infliximab specifically (with or without etanercept), the SIR for lymphoma was 2.6 (1.4–4.5); and for infliximab alone, the SIR was 2.2 (1.0–4.9). This must be viewed against the expected increased risk of lymphoma among RA patients in general, albeit such an increased risk was not observed in this study population among patients who had not received MTX or TNF therapy. In a large multicenter registry of 757 RA patients treated with either etanercept or infliximab and 800 RA patents who were treated conventionally, the investigators identified a nearly 5-fold increased rate of lymphoma in the anti-TNF–treated patients compared with the patients treated conventionally (RR = 4.9, 95% CI 0.9–26.2). A recent meta-analysis of randomized controlled trials of anti-TNF therapy in RA patients calculated a pooled odds ratio (OR) for total malignancies of 3.3 (95% CI 1.2–9.1). Although a separate, specific analysis for lymphoma risk was not made, a total of 10 lymphomas were observed in the anti-TNF–treated patients, whereas no lymphomas were observed in the control arms. Six of these 10 lymphomas were not included in the analysis because they occurred in follow-up after the trials ended. Even after updating their estimates to include 2 additional RA studies, the OR for overall malignancy remained significant (OR 2.4, 95% CI 1.2–4.8). The results of these studies must be interpreted with caution. First, the limited number of observations and, in some cases, short follow-up, limit interpretation, as demonstrated by the wide and often overlapping confidence intervals. Studies evaluating lymphoma risk in RA are also complicated by the uncertainty of the true cause. Given the known association between RA and lymphoma, and RA disease severity and lymphoma, it is impossible to determine whether these elevated risks are due to the anti-TNF therapy, the underlying disease severity, or a combination of both.
Infliximab received US Food and Drug Administration (FDA) approval for CD in October 1998 and licensing in the European Union in September 1999; it received FDA approval for UC in September 2005. At the time of FDA approval, there were 2 small open-labeled studies and another 2 randomized placebo-controlled trials ( Table 3 ). Neither of the randomized controlled trials at that time had reported any lymphomas; however, given the small number of patients, the FDA felt there was not even evidence to exclude a relationship between infliximab and lymphoma development. Subsequently, an extension of 1 of the studies did find a single case of lymphoma in a patient with a 6-year history of AZA and a single infusion of infliximab.
Related posts:
Treatment of Fistulizing Inflammatory Bowel Disease
Pediatric Inflammatory Bowel Disease: Highlighting Pediatric Differences in IBD
Pouchitis and Pouch Dysfunction
Pregnancy and Inflammatory Bowel Disease
Safety Profile of IBD Therapeutics: Infectious Risks
Clostridium Difficileand Inflammatory Bowel Disease
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
