Background/Purpose:  Biologic disease-modifying anti-rheumatic drugs (DMARDs) have been widely used for treatment of rheumatoid arthritis (RA) in United States. However, it is not clear whether biologics are associated with elevated cancer risk. We estimated cancer incidence rates (IRs) among anti-TNF users and evaluated biologic use after cancer diagnosis among RA patients, using validated cancer outcomes from the U.S. SEER program.

Methods: Using 2006-2013 SEER-Medicare linked data and a 5% non-cancer sample, we identified new anti-TNF users (etanercept, adalimumab, certolizumab, golimumab, infliximab) among RA patients, identified using algorithms that required ≥ 1 rheumatologist diagnosis code for RA at any time before specific drug use. New users were defined specific to each drug as no use of that therapy in the ‘baseline’ period. Eligible subjects were continuously enrolled in Medicare Parts A, B and D in baseline and throughout follow up. Patients with history of cancer were excluded. We identified all cancers reported to SEER nationally, as confirmed nationally by trained tumor registrars. Follow up started from the drug initiation date and ended at the earliest date of: malignancy, a 90-day gap in current exposure, death, switch to another biologic, or loss of Medicare coverage. We calculated the number of cancer events in both SEER and 5% non-cancer sample, and then calculated the weighted national IRs of cancer for each drug, along with 95% confidence interval, by multiplying the observed person years for non-cancer patients with 20. Subgroup analyses based on the concurrent use of methotrexate (MTX) were conducted. We also evaluated patients’ subsequent biologic use after their cancer diagnosis.

Results: We identified 1,374 new anti-TNF users in SEER-Medicare linked cancer registry and 5% non-cancer sample. Of these, 36.5% used infliximab, 25.9% adalimumab, 20.3% etanercept, 11.0% certolizumab, and 6.1% golimumab. During follow-up, we identified 231 cancers yielding a weighted IR from a low of 0.8 (golimumab) to a high of 1.9 (certolizumab) per 1000 person years across different anti-TNFs. IRs with concurrent MTX use were greater than those without MTX (table). After cancer diagnosis, 60-70% of patients resumed the same anti-TNF, 15-25% discontinued all biologic use, and the remainder switched to non-biologic DMARDs or other biologics. This pattern was consistent across different anti-TNFs.  

Conclusion: Crude incidence rates of cancer among RA patients were comparable between users of different anti-TNF users. Concurrent MTX use was associated with higher cancer risk, although ongoing work is evaluating potential confounding.