Abstract
OBJECTIVE: We aim to evaluate the malignancy risk between Janus kinase inhibitors (JAKi) users and tumor necrosis factor inhibitors (TNFi) users in rheumatoid arthritis (RA) patients, using a large real-world electronic health record database (TriNetX). METHODS: In this retrospective cohort study, we identified adult RA patients initiating JAKi or TNFi therapy between January 1, 2018, and December 31, 2022, within the TriNetX global federated network. The hazard ratio (HR) and confidence intervals (CI) of incident-specific cancers, overall cancer incidence, and all-cause mortality, were calculated between the propensity score matched JAKi and TNFi cohorts. The probability of the outcome of interest was estimated using the Kaplan-Meier analysis. RESULTS: After propensity score matching, there were 4045 each in JAKi or TNFi user cohorts. The mean (standard deviation) age was 57.7 (13.3) and 57.6 (13.9) years, 81.4% and 80.8% were female, and median (interquartile range) follow-up time 3.69 (2.61) years vs 3.69 (2.68) years, in the JAKi and the TNFi cohorts, respectively. No significant differences were observed in risks of overall cancer incidence and all-cause mortality between the two cohorts. JAKi users had a reduced risk of incident digestive organ cancers compared with TNFi users (adjusted HR: 0.599, 95% CI: 0.439-0.817), mainly observed among females. The risks of incident respiratory and intrathoracic organs cancers were increased in JAKi users compared to TNFi users among females (adjusted HR: 2.582, 95% CI: 1.109-6.011) but not in males. CONCLUSION: In a large real-world database, we did not find an increased risk of overall cancer incidence in JAKi compared to TNFi users among RA patients. JAKi users had a lower risk of incident digestive organs cancers, and a higher risk of respiratory cancer in females, when compared with TNFi users.