Cancer risk with tocilizumab/sarilumab, abatacept and rituximab treatment in patients with rheumatoid arthritis: a Danish cohort study.

Abstract

To investigate cancer risk in RA patients treated with tocilizumab/sarilumab, abatacept or rituximab compared with those who received TNF inhibitors (TNFi) and compared with biological DMARDs (bDMARD)-naïve RA patients. Nationwide registry-based cohort study of RA patients who initiated bDMARD treatment with tocilizumab/sarilumab, abatacept, rituximab, and TNFi, as well as bDMARD-naive patients who initiated their second type of conventional synthetic DMARD. Patients were identified in the Danish Rheumatology Quality Register (DANBIO) and followed for cancer from 2006 to 2020. Patients could contribute multiple treatments, with person years, deaths and cancers allocated to each treatment group in a 'latest type of treatment' manner. Inverse probability of treatment weighting and weighted cause-specific Cox models were used to calculate hazard ratios (HRs) for cancer in each tocilizumab/sarilumab, abatacept and rituximab group compared with TNFi-treated and bDMARD-naïve groups, respectively. In total, 21 982 treatment initiations, 96 475 person years and 1423 cancers were identified. There were no statistically significant increased HRs for overall cancer in tocilizumab/sarilumab, abatacept or rituximab treatment groups (HRs ranged from 0.7 to 1.1). More than 5 years of abatacept exposure showed a non-significantly increased HR compared with TNFi (HR 1.41, 95% CI 0.74-2.71). For haematological cancers, rituximab treatment showed non-significantly reduced HRs: vs TNFi-treated (HR 0.09; 95% CI 0.00-2.06) and bDMARD-naïve (HR 0.13; 95% CI 0.00-1.89). Treatment with tocilizumab/sarilumab, abatacept or rituximab in RA patients was not associated with increased risks of cancer compared with TNFi-treated and with bDMARD-naïve RA patients in a real-world setting.