Questions Remain About RA Treatment's Link to Lymphoma

NEW YORK — Data from large registries ultimately should provide answers about the long-term lymphoma risk associated with the tumor necrosis factor-blocking drugs, but for now, questions and contradictions remain.

Patients with rheumatoid arthritis (RA) have an elevated risk of lymphoma that has been estimated to be between two- and eightfold. How much of the risk relates to disease activity and how much relates to immunosuppressive treatment is not yet clear, Dr. Jeffrey Greenberg said in reviewing recent studies during a rheumatology meeting sponsored by New York University.

Helpful background evidence on the overall risk of lymphoma has emerged from a large Swedish registry that included 74,651 patients who received a diagnosis of RA between 1964 and 1995. Within this cohort, there were 378 cases of lymphoma; these patients were matched with 378 controls from the cohort who were cancer free when the lymphoma patients were diagnosed.

The cases were then analyzed for factors that might influence the development of lymphoma. The investigators also sought to determine whether RA-associated malignancy is disease- and inflammation-driven and could be associated with inadequate immunosuppres- sion, or whether it results directly from immunosuppression.

Level of disease activity strongly predicted risk in this study. Medium overall disease activity was associated with an 8-fold increase in lymphoma risk, while high disease activity was linked to a 70-fold increase. There also was a threefold increase in risk among patients with the highest erythrocyte sedimentation rates (Arthritis Rheum. 2006;54:692–701).

The investigators reported that having ever been treated with a traditional disease-modifying antirheumatic drug (DMARD) was not associated with an increased risk. Although increases in risk were seen with some individual drugs such as azathioprine, other drugs such as corticosteroids were associated with relative risk estimates of less than 1, reflecting a reduction in risk. Immunostimulation is the driving force in inflammation-associated lymphomas.

In a separate study, a different group of Swedish investigators looked at the risk of hematopoietic malignancies among RA patients being treated with TNF antagonists. The lymphoma risk was tripled in those being treated with these drugs compared with the general population, but the risk was no different from that in other RA patients (Ann. Rheum. Dis. 2005;64:1414–20).

Overall, in clinical trials thus far, the standardized incidence ratios for lymphoma for etanercept, infliximab, and adalimumab are 3.5, 7.0, and 5.5, respectively, said Dr. Greenberg, director of the Arthritis Translational Registry and Biorepository, New York University, New York. Those numbers do not account for background RA-related lymphoma risk, however, and therefore do not include the risk associated with treatment.

It's important to keep in mind that clinical trials are short in duration and are not powered to detect rare events such as malignancies, Dr. Greenberg said. To overcome this lack of power, another group of researchers undertook a meta-analysis of nine randomized trials of patients undergoing anti-TNF therapy, with 3,493 receiving active treatment and 1,512 receiving placebo. In this analysis, the pooled odds ratio for malignancy was 3.3, a finding that was “somewhat inconsistent” with the results seen in the Swedish studies (JAMA 2006;295:2275–85).

The meta-analysis had limitations, according to Dr. Greenberg. For example, it included only studies with infliximab and adalimumab; etanercept was omitted. The study also did not account for length of time on the drug. “They simply counted the number of people exposed to the drug versus placebo,” he said. This was in contrast to the method the Food and Drug Administration used in its analysis of the safety of all three TNF blockers, which found no elevations in malignancy rates in patients on these drugs (JAMA 2006;296:2201–2).

As to the method of analyzing drug exposure and risk, “There's an argument to be made on either side—it's not black and white—but the results of the meta-analysis have to be taken with a grain of salt,” Dr. Greenberg said.

The opposite approach, of analyzing risk according to time of exposure, had quite different results in another recent report. In an abstract presented at the 2006 meeting of the European League Against Rheumatism, a total of 124 lymphomas were seen during 109,884 patient-years of follow-up.

There were 79 observed cases of lymphoma, compared with 45 expected cases, giving a standardized incidence rate of 1.8, which does not differ from background risk in RA.

There also were no statistically significant increases in risk associated with any specific treatments, including DMARDs and biologics (Ann. Rheum. Dis. 2006;65[suppl. 2]:512–3).

So with regard to the effects of biologic treatment on lymphoma risk, there are still more questions than answers, Dr. Greenberg said. The way risk is determined also needs to change. “As we move toward personalized medicine, we need to move from population-based risk estimates, where we tell a patient, 'You have a 1.5-fold risk of developing a malignancy in the next 10 years,' to a position where we can say, 'Based on your genetic profile, you have a 10-fold risk of malignancy but an extremely high likelihood of benefiting from this drug.' Then we can balance risk and benefit on a personal level rather than relying on overall population-based estimates,” Dr. Greenberg said.

NEW YORK — Data from large registries ultimately should provide answers about the long-term lymphoma risk associated with the tumor necrosis factor-blocking drugs, but for now, questions and contradictions remain.

Patients with rheumatoid arthritis (RA) have an elevated risk of lymphoma that has been estimated to be between two- and eightfold. How much of the risk relates to disease activity and how much relates to immunosuppressive treatment is not yet clear, Dr. Jeffrey Greenberg said in reviewing recent studies during a rheumatology meeting sponsored by New York University.

Helpful background evidence on the overall risk of lymphoma has emerged from a large Swedish registry that included 74,651 patients who received a diagnosis of RA between 1964 and 1995. Within this cohort, there were 378 cases of lymphoma; these patients were matched with 378 controls from the cohort who were cancer free when the lymphoma patients were diagnosed.

The cases were then analyzed for factors that might influence the development of lymphoma. The investigators also sought to determine whether RA-associated malignancy is disease- and inflammation-driven and could be associated with inadequate immunosuppres- sion, or whether it results directly from immunosuppression.

Level of disease activity strongly predicted risk in this study. Medium overall disease activity was associated with an 8-fold increase in lymphoma risk, while high disease activity was linked to a 70-fold increase. There also was a threefold increase in risk among patients with the highest erythrocyte sedimentation rates (Arthritis Rheum. 2006;54:692–701).

The investigators reported that having ever been treated with a traditional disease-modifying antirheumatic drug (DMARD) was not associated with an increased risk. Although increases in risk were seen with some individual drugs such as azathioprine, other drugs such as corticosteroids were associated with relative risk estimates of less than 1, reflecting a reduction in risk. Immunostimulation is the driving force in inflammation-associated lymphomas.

In a separate study, a different group of Swedish investigators looked at the risk of hematopoietic malignancies among RA patients being treated with TNF antagonists. The lymphoma risk was tripled in those being treated with these drugs compared with the general population, but the risk was no different from that in other RA patients (Ann. Rheum. Dis. 2005;64:1414–20).

Overall, in clinical trials thus far, the standardized incidence ratios for lymphoma for etanercept, infliximab, and adalimumab are 3.5, 7.0, and 5.5, respectively, said Dr. Greenberg, director of the Arthritis Translational Registry and Biorepository, New York University, New York. Those numbers do not account for background RA-related lymphoma risk, however, and therefore do not include the risk associated with treatment.

It's important to keep in mind that clinical trials are short in duration and are not powered to detect rare events such as malignancies, Dr. Greenberg said. To overcome this lack of power, another group of researchers undertook a meta-analysis of nine randomized trials of patients undergoing anti-TNF therapy, with 3,493 receiving active treatment and 1,512 receiving placebo. In this analysis, the pooled odds ratio for malignancy was 3.3, a finding that was “somewhat inconsistent” with the results seen in the Swedish studies (JAMA 2006;295:2275–85).

The meta-analysis had limitations, according to Dr. Greenberg. For example, it included only studies with infliximab and adalimumab; etanercept was omitted. The study also did not account for length of time on the drug. “They simply counted the number of people exposed to the drug versus placebo,” he said. This was in contrast to the method the Food and Drug Administration used in its analysis of the safety of all three TNF blockers, which found no elevations in malignancy rates in patients on these drugs (JAMA 2006;296:2201–2).

As to the method of analyzing drug exposure and risk, “There's an argument to be made on either side—it's not black and white—but the results of the meta-analysis have to be taken with a grain of salt,” Dr. Greenberg said.

The opposite approach, of analyzing risk according to time of exposure, had quite different results in another recent report. In an abstract presented at the 2006 meeting of the European League Against Rheumatism, a total of 124 lymphomas were seen during 109,884 patient-years of follow-up.

There were 79 observed cases of lymphoma, compared with 45 expected cases, giving a standardized incidence rate of 1.8, which does not differ from background risk in RA.

There also were no statistically significant increases in risk associated with any specific treatments, including DMARDs and biologics (Ann. Rheum. Dis. 2006;65[suppl. 2]:512–3).

So with regard to the effects of biologic treatment on lymphoma risk, there are still more questions than answers, Dr. Greenberg said. The way risk is determined also needs to change. “As we move toward personalized medicine, we need to move from population-based risk estimates, where we tell a patient, 'You have a 1.5-fold risk of developing a malignancy in the next 10 years,' to a position where we can say, 'Based on your genetic profile, you have a 10-fold risk of malignancy but an extremely high likelihood of benefiting from this drug.' Then we can balance risk and benefit on a personal level rather than relying on overall population-based estimates,” Dr. Greenberg said.

NEW YORK — Data from large registries ultimately should provide answers about the long-term lymphoma risk associated with the tumor necrosis factor-blocking drugs, but for now, questions and contradictions remain.

Patients with rheumatoid arthritis (RA) have an elevated risk of lymphoma that has been estimated to be between two- and eightfold. How much of the risk relates to disease activity and how much relates to immunosuppressive treatment is not yet clear, Dr. Jeffrey Greenberg said in reviewing recent studies during a rheumatology meeting sponsored by New York University.

Helpful background evidence on the overall risk of lymphoma has emerged from a large Swedish registry that included 74,651 patients who received a diagnosis of RA between 1964 and 1995. Within this cohort, there were 378 cases of lymphoma; these patients were matched with 378 controls from the cohort who were cancer free when the lymphoma patients were diagnosed.

The cases were then analyzed for factors that might influence the development of lymphoma. The investigators also sought to determine whether RA-associated malignancy is disease- and inflammation-driven and could be associated with inadequate immunosuppres- sion, or whether it results directly from immunosuppression.

Level of disease activity strongly predicted risk in this study. Medium overall disease activity was associated with an 8-fold increase in lymphoma risk, while high disease activity was linked to a 70-fold increase. There also was a threefold increase in risk among patients with the highest erythrocyte sedimentation rates (Arthritis Rheum. 2006;54:692–701).

The investigators reported that having ever been treated with a traditional disease-modifying antirheumatic drug (DMARD) was not associated with an increased risk. Although increases in risk were seen with some individual drugs such as azathioprine, other drugs such as corticosteroids were associated with relative risk estimates of less than 1, reflecting a reduction in risk. Immunostimulation is the driving force in inflammation-associated lymphomas.

In a separate study, a different group of Swedish investigators looked at the risk of hematopoietic malignancies among RA patients being treated with TNF antagonists. The lymphoma risk was tripled in those being treated with these drugs compared with the general population, but the risk was no different from that in other RA patients (Ann. Rheum. Dis. 2005;64:1414–20).

Overall, in clinical trials thus far, the standardized incidence ratios for lymphoma for etanercept, infliximab, and adalimumab are 3.5, 7.0, and 5.5, respectively, said Dr. Greenberg, director of the Arthritis Translational Registry and Biorepository, New York University, New York. Those numbers do not account for background RA-related lymphoma risk, however, and therefore do not include the risk associated with treatment.

It's important to keep in mind that clinical trials are short in duration and are not powered to detect rare events such as malignancies, Dr. Greenberg said. To overcome this lack of power, another group of researchers undertook a meta-analysis of nine randomized trials of patients undergoing anti-TNF therapy, with 3,493 receiving active treatment and 1,512 receiving placebo. In this analysis, the pooled odds ratio for malignancy was 3.3, a finding that was “somewhat inconsistent” with the results seen in the Swedish studies (JAMA 2006;295:2275–85).

The meta-analysis had limitations, according to Dr. Greenberg. For example, it included only studies with infliximab and adalimumab; etanercept was omitted. The study also did not account for length of time on the drug. “They simply counted the number of people exposed to the drug versus placebo,” he said. This was in contrast to the method the Food and Drug Administration used in its analysis of the safety of all three TNF blockers, which found no elevations in malignancy rates in patients on these drugs (JAMA 2006;296:2201–2).

As to the method of analyzing drug exposure and risk, “There's an argument to be made on either side—it's not black and white—but the results of the meta-analysis have to be taken with a grain of salt,” Dr. Greenberg said.

The opposite approach, of analyzing risk according to time of exposure, had quite different results in another recent report. In an abstract presented at the 2006 meeting of the European League Against Rheumatism, a total of 124 lymphomas were seen during 109,884 patient-years of follow-up.

There were 79 observed cases of lymphoma, compared with 45 expected cases, giving a standardized incidence rate of 1.8, which does not differ from background risk in RA.

There also were no statistically significant increases in risk associated with any specific treatments, including DMARDs and biologics (Ann. Rheum. Dis. 2006;65[suppl. 2]:512–3).

So with regard to the effects of biologic treatment on lymphoma risk, there are still more questions than answers, Dr. Greenberg said. The way risk is determined also needs to change. “As we move toward personalized medicine, we need to move from population-based risk estimates, where we tell a patient, 'You have a 1.5-fold risk of developing a malignancy in the next 10 years,' to a position where we can say, 'Based on your genetic profile, you have a 10-fold risk of malignancy but an extremely high likelihood of benefiting from this drug.' Then we can balance risk and benefit on a personal level rather than relying on overall population-based estimates,” Dr. Greenberg said.