Background/Purpose: Despite early concerns, evidence has suggested that use of TNFα inhibitors (TNFi) may not be associated with an increase in cancer risk in patients with rheumatoid arthritis (RA). Nevertheless, uncertainty remains about appropriate treatment decisions following a diagnosis of cancer in patients receiving TNFi. The aims of this analysis were to describe (i) treatment decisions and (ii) outcomes in the 5 years following a diagnosis of cancer in patients with RA receiving TNFi.

Methods: This descriptive analysis included participants from the British Society for Rheumatology Biologics Register for RA (BSRBR-RA) who develop a first cancer while receiving TNFi (+90 days) prior to 01/04/2011. Cancer and death data were captured through linkage with the UK National death and cancer registers held by the Health and Social Care Information Centre (HSCIC). Treatment data were captured via regular clinical follow-up forms. Initial treatment decisions in all patients following diagnosis and subsequent treatment decisions among patients who survived at least 6 months were examined. Decisions were examined in all cancers and separately for the five most common cancer sites. Overall and site-specific mortality are described using percent of deaths and Kaplan Meier survival function.

Results: During 118355 person years of follow-up in 12499 patients who had ever been exposed to TNFi, 404 first cancers occurred on TNFi (74 lung, 65 breast, 35 colorectal, 36 lymphoma, 18 prostate and 176 other). Within 6 months of diagnosis, 95 (24%) patients died (median time to death (TTD):52 days (IQR: 22-100)) with the highest death rate among patients with lung cancer. Of the 309 patients surviving the first 6 months, 67% stopped TNFi at cancer diagnosis.  Over the subsequent 5 years, 47% remained off all biologics, 26% continued on TNFi and 25% started an alternative class of biologic after a median of 510 days. Only 2 patients restarted a TNFi therapy and 5 discontinued their TNFi after 6 months. Treatment decisions varied among the 5 most common cancer sites, with a majority of patients (64%) with prostate cancer continuing their initial TNFi treatment over 5 years. Of patients with breast cancer and lymphoma 32% and 40% of patients, respectively started an alternative treatment; the majority switched to rituximab.

Conclusion: Treatment decisions following cancer diagnosis in patients receiving TNFi were highly variable and differed by the site of cancer diagnosis. Although around two thirds of patients discontinued their TNFi, only one third of patients with prostate cancer discontinued. Patients with cancer at sites associated with longer survival were also more likely to continue or restart a biologic over the following five year period.  If biologics were restarted, most patients started an alternative class of drug.