Anticoagulation therapy in patients with atrial fibrillation and active cancer

Abstract

Abstract Background Atrial fibrillation (AF) is common among cancer patients. Oral anticoagulation therapy (OAT) is important to prevent stroke in AF patients, but active cancer may also predispose to bleeding. Current knowledge on the use of OAT among AF patients with active cancer is scarce. Purpose To assess the influence of active cancer on the initiation of OAT in patients with AF and to explore cancer features predicting withholding OAT. Methods The nationwide FinACAF registry linked data from several registries covering all levels of care as well as drug purchases and included all 123 618 patients diagnosed with incident AF in Finland during 2011- 2018 with an indication for OAT (≥1 non-sex CHA2DS2-VASc point). Patients were considered to have active cancer if they had received a new cancer diagnosis or treatment for cancer during the preceding year from AF diagnosis. The primary outcome was the initiation of OAT (including non-vitamin K oral anticoagulants (NOAC) and warfarin) and the secondary outcome the initiation of any anticoagulation therapy (AAT; including OAT and low molecular-weight heparin) during the first year from AF diagnosis. Results The mean age of the patients was 73.9 (standard deviation 11.5) years and 62 955 (50.9%) of them were women. Altogether, 7015 (5.7%) patients had active cancer at AF diagnosis. The median CHA2DS2-VASc (3 [interquartile range 2-4] for both groups) and HAS-BLED (2 [interquartile range 2-3] for both groups) scores were similar in patients with active cancer and those without cancer history. Patients with active cancer, compared to those without cancer history, less often initiated OAT (47.7% vs. 70.5 %) or AAT (61.8% vs. 71.8%) (Figure). In Fine-Gray regression analyses with all-cause mortality as a competing risk, after adjusting for confounding factors, active cancer remained associated with lower likelihood of OAT (adjusted subdistribution hazard ratio (aSHR 0.501) (95% confidence interval (CI): 0.485-0.518), p<0.001) as well as AAT (aSHR 0.745 (95%-CI 0.723-0.769, p<0.001) initiation. Among cancer patients, metastases, gastrointestinal cancer and radiotherapy were associated with lower likelihood of OAT initiation, while only metastases predicted lower likelihood of AAT initiation (Table). Among patients initiating OAT, active cancer was associated with lesser likelihood of choosing a NOAC over warfarin compared to those without cancer (36.0% vs. 39.2%; adjusted hazard ratio 0.712 (95%-CI: 0.640-0.792), p<0.001). Conclusion AF patients with active cancer are less likely to initiate anticoagulation therapy, and those starting OAT are less likely to initiate with NOACs, compared to patients without cancer history. Metastases are the most important predictor of withholding anticoagulation therapy among AF patients with cancer.