TNF-alpha Inhibitors and Lymphoma: More Data Needed to Assess Risk

Do the new biologics for rheumatoid arthritis (RA), all of which block the action of tumor necrosis factor (TNF), increase the risk of lymphoma? This question was considered at length at the March 2003 meeting of the Arthritis Advisory Committee (AAC) of the U.S. Food and Drug Administration (transcript available at www.fda.gov). The Committee concluded that a connection between the drugs and the disease, although possible, could not be established with certainty and recommended long-term monitoring of patients with RA who are taking these agents. What changes, if any, the agency will require in the package inserts of the three products - adalimumab (Humira), etanercept (Enbrel), and infliximab (Remicade) - are as yet unknown.

Last year, investigators from the Center for Biologics Evaluation and Research at the FDA published a discussion of 26 cases of lymphoma in RA patients who were taking TNF-alpha inhibitors. These cases had been reported to the post-market adverse event surveillance system MedWatch[1]. Most had non-Hodgkin's lymphoma, and in these patients the disease tended to appear within a short time (median of 8 weeks) after treatment started. In two cases, discontinuation of the drug was reportedly followed by disease regression. Because these are spontaneous reports, as was pointed out at the meeting, there is no way to determine what percentage of the total lymphoma cases they represent nor what percentage of the number of patients exposed to the drugs.

At the March meeting, data from the premarketing clinical trials were updated by the sponsors (Table 1). The histologic subtypes observed in patients receiving TNF-alpha inhibitors were representative of what is seen in the general population, and no predominant histologic subtype was noted. Although the data on the incidence of lymphomas in clinical trials suggest that TNF-alpha inhibitors may predispose patients to lymphoma, this conclusion is by no means certain due to a number of factors.

Table 1. Lymphoma Occurrence in Patients Receiving TNF-Alpha Inhibitors in Clinical Trials and Open-Label Extensions

*The number of lymphomas expected is derived from the SEER database (i.e., not the general RA population).

** Standard incidence ratio (number seen/number expected)

First, the number of patients in the clinical trials was relatively small, and the placebo arms were too small to provide an adequate comparator arm for such rare events.

Second, and even more important, RA itself is a risk factor for lymphoma, particularly the B-cell types[2],[3],[4],[5]. This well-documented association may be the result of chronic antigenic stimulation and selection[6],[7],[8],[9],[10],[11], although iatrogenic immunosuppression may contribute to the development of some lymphomas (analogous to the Epstein-Barr virus-linked lymphoproliferative disorders, including lymphomas, seen in recipients of organ and bone marrow transplants and in patients in the later stages of HIV infection)[12],[13],[14],[15],[16],[17].

Third, the risk of lymphoma parallels the severity of RA - patients with severe disease have a 25.8-fold higher risk than those with less active disease[18]. This finding is relevant because most of the clinical trials enrolled only patients with moderate to severe RA, although different patient populations were studied in different trials and with different agents. The potential impact of disease characteristics was highlighted by the observation that in one study restricted to subjects with early RA (mean disease duration <1 year), no lymphomas were seen.

Fourth, most of the patients had taken or were taking other disease-modifying antirheumatic drugs, which also have immunosuppressive properties.

Fifth, although the uniform absence of lymphomas in the placebo-treated patients is suggestive, it must be borne in mind that most of these patients crossed over into the active-drug arm after a short period, so their follow-up is much shorter than that of patients given the active drug. Consequently, they had less opportunity to develop lymphoma while receiving the test agent.

Sixth, the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) data used to calculate the standardized incidence ratios (SIRs) were collected between 1992 and 1999, but the incidence of lymphoma in the general population is increasing, making the "expected" figure too low.

The bottom line is that although there are hints that TNF-alpha inhibitors increase the risk of lymphoma, this conclusion cannot be stated with certainty. As was pointed out by Elaine S. Jaffe, MD, an FDA consultant from another advisory committee, "You have to sort out what is due to disease, what is due to treatment, and what is due to background noise." At the present time, this is not possible.

One attempt to circumvent the problem of the predisposition of RA patients to lymphoma was an analysis of the data from the infliximab trials in Crohn's disease. Unfortunately, some evidence suggests that this disease also increases the risk of lymphoma[19],[20], although other studies have reached the opposite conclusion[21]. The impact of disease severity on lymphoma risk, however, has not been studied in patients with Crohn's disease.

Two lymphomas appeared after a single dose of infliximab, and 20 cases have been identified in patients since the drug received marketing approval for use in Crohn's disease. Here, the problem of crossing over was even greater than in the RA trials, as there are few options for the treatment of serious Crohn's disease.

In another attempt to define the risk, the latest information from the National Databank for Rheumatic Diseases was examined at the AAC meeting. This database, maintained by Fred Wolfe, MD, of the Arthritis Research Center Foundation, includes information on 18,557 RA patients followed by more than 900 rheumatologists, who were queried twice a year from 1998 to mid-2002 about adverse drug events (Table 2).

Table 2. Lymphoma Occurrence in Patients Receiving TNF-alpha Inhibitors (National Databank)

Although these data help establish the "expected" number of lymphomas in the RA population, there could be differences in the severity of disease in the different groups. All three of the manufacturers are doing long-term follow-up, and in the UK all physicians prescribing TNF inhibitors must register, so the required data should eventually be forthcoming. The FDA asked the Advisory Committee to determine what if any changes should be made in the labeling for the three products. The Committee decided that a boxed warning about lymphoma was not warranted, given the deficiencies in the available data, but that a standardized statement on the need for vigilance would be appropriate. Standard incidence ratios were disparaged as potentially misleading since the observed rates were low, confidence intervals were wide, and study populations lacked uniformity.

"If I'm a patient, I don't want to know that I am twice or ten times as likely to get something as somebody else is if I don't know the likelihood that somebody else is going to get it," stressed Hal Paulus, MD, Professor of Medicine, UCLA. "What I want to know is, if I start this drug, what's the chance I am going to get lymphoma?" That, of course, is the question to which, at present, no one has an answer. Another comment came from Wendy W. McBrair, RN, who was the consumer representative on the Committee. "It is important not to scare patients. People with rheumatoid arthritis will be forever grateful for these medications, and I don't think anything I have heard here today is going to keep patients away from them. Physicians need to be vigilant, and patients need to be educated on how to be vigilant. That's the most important message."