Background Clinical guidelines caution against the use of TNFi in individuals with a recent (5–10 years) history of cancer, but evidence of an increased risk of cancer recurrence is limited (1–2).
Objectives We investigated the risk of recurrent solid non-skin cancer in patients with rheumatoid arthritis and TNFi-treatment, and if time between index-cancer and TNFi-start influence this risk.
Methods 61.950 individuals with RA were identified in the Swedish national outpatient-care register Jan 2001-Dec 2014. Among these, 446 Individuals with at least one diagnosis of cancer (index cancer) prior to the start of TNFi-treatment were identified through linkage to the national cancer register and the ARTIS register of biologic treatment. Individuals (n=1.278) with a history of equally recent cancer of the same type and stage (invasive/in situ) were matched 3:1 to each patient starting TNFi against a background of solid cancer. Study participants were required to be in cancer remission during a period of 6 months prior to start of follow-up. The primary outcome was first recurrence or second primary of the same cancer type, identified through register-linkages until Dec 2014. Hazard ratios (HR) for recurrence or second primary were calculated using a Cox regression model with TNFi-treatment start (and a corresponding date among the matched biologics-naïve individuals) as start of follow-up. The final models were stratified for the matching variables sex, birth year, year of diagnosis of the index cancer and index cancer type and stage (invasive vs in situ), and adjusted for education level and comorbidities.
Results The mean time from index cancer diagnosis until TNFi-treatment/start of follow-up was 9.9 and 9.5 years among the TNFi treated and their matched biologics-naïve controls, respectively. The mean follow-up (SD) from TNFi start was 4.9 (3.5) and 4.1 (3.1) years, respectively. The cancer stage distribution was similar between the two groups, apart from stage IV (0.6% among the TNFi-treated and 1.6% among the biologics-naïve). Thirty individuals (7%) among the 446 TNFi-treated developed a cancer recurrence (crude incidence rate 14/1000 person-years), compared with 89 (7%) among the 1.278 matched biologics-naive (crude incidence rate 17/1000 person-years). This corresponded to an adjusted HR for recurrent cancer of 0.69 (95% CI 0.42–1.12) in the matched analysis (table 1) comparing the TNFi treated to the biologics-naïve individuals. Stratified analyses indicated no increased risk associated with any specific cancer type with the possible exception of uterine cancer where HR for recurrence was 14.8 (95% CI 1.17–187.5), based on only 1 event among the TNFi-treated. HR for recurrence among individuals starting TNFi treatment within 5 year from index cancer was 0.67 (0.31–1.44).
Conclusions Among patients with RA and a history of cancer, those selected to receive TNFi-treatment in clinical practice did not experience more cancer recurrences than patients with RA treated otherwise. We detected no differential risk depending on the timing of TNFi-start in relation to the index cancer. The generalizability of our findings to individuals with a very recent cancer, or a poor prognosis, remains unknown.
Raaschou P, et al. Ann Rheum Dis 2015;74:2137–2143.
Mamtani R, et al. Arthritis Rheumatol 2016; 68: 2403–2411.
Disclosure of Interest P. Raaschou: None declared, J. Söderling Grant/research support from: previous research projects fully or partly funded by Novo Nordisk and Combine Sweden, and has served as an external consultant to AbbVie, Merck and Novartis., J. Askling Grant/research support from: Abbvie, Pfizer, UCB, MSD, Roche, Lilly.