Major bleeding and thromboembolic complications associated with antithrombotic treatment in patients with atrial fibrillation/flutter and incident cancer

Abstract

Background

Anticoagulant management of patients with atrial fibrillation with active cancer is complex because cancer increases the risk of thrombosis as well as bleeding. Previous studies have investigated the impact of any type of cancer, while outcomes may differ per specific type. We performed the present study to provide more insight into the impact of specific types of cancer on clinical outcomes.

Objectives

We examined major bleeding (MB) and thromboembolism (TE) rates associated with antithrombotic treatment in patients with atrial fibrillation/flutter (AF) who develop cancer and examined whether cancer type affected MB and TE risks.

Methods

This Danish population-based cohort study included all patients aged ≥ 50 years discharged with incident AF between January 1, 1995, and December 31, 2016, and identified those who subsequently developed cancer. Data on cancer type, outcomes, and antithrombotic exposure were obtained from hospital and drug prescription databases. Follow-up continued from the time of cancer diagnosis until the occurrence of an outcome or the end of the 2-year follow-up. Incidence rates (IRs) per 100 patient-years and adjusted hazard ratios with corresponding 95% CIs were calculated using Cox regression.

Results

A total of 22,996 patients with AF with subsequent incident cancer were identified. These patients had higher MB (IR, 5.36 [95% CI, 5.09-5.64] vs 2.27 [95% CI, 2.22-2.32]) and TE (IR, 3.91 [95% CI, 3.68-4.15] vs 2.71 [95% CI, 2.66-2.76]) rates than those without cancer. The higher MB rate was observed across all antithrombotic exposure categories. Urogenital (IR, 6.43 [95% CI, 5.94-6.95]) and intracranial cancer (IR, 6.36 [95% CI, 3.85-9.76]) demonstrated the highest MB rates; hematologic (IR, 4.92 [95% CI, 4.12-5.82]) and gastrointestinal cancer (IR, 4.82 [95% CI, 4.31-5.36]) had the highest TE rates. A particularly high MB rate was observed in patients with AF with gastrointestinal cancer and triple antithrombotic therapy (IR, 39.0 [95% CI, 15.5-79.1]).

Conclusion

Patients with AF with certain incident cancer types experienced higher rates of MB and TE than those without cancer. Dual/triple antithrombotic therapy in patients with AF with incident cancer was associated with high bleeding rates, particularly with gastrointestinal cancer.