Abstract 336: Anticoagulation Prescribing Patterns and Rates of Bleeding for People With Cancer and Atrial Fibrillation
Introduction: Atrial fibrillation (AF) is a major cause of stroke and increased mortality among people with cancer. People with cancer have a higher risk of developing AF at time of cancer diagnosis, prior to treatment, and after therapy than patients without cancer. AF management includes anticoagulation (AC) for stroke prophylaxis. Risk stratification tools are used to assess patients’ stroke and bleeding risk. However, such tools have not been validated in people with cancer. Therapeutic AC poses challenges due to increased risk of bleeding, platelet dysfunction, and drug interaction. Given this, therapeutic AC is often deferred despite potentially higher thrombosis risk. We sought to better understand practice patterns of therapeutic AC among people with cancer and AF.
Hypothesis: We predicted reduced AC prescribing rates for AF among people with cancer despite high CHA2DS2-VASc score.
Methods: We conducted a retrospective cohort study of people with cancer and AF at a single institution from 2010 to 2018. Cancer type was ascertained from an institutional registry; AF and other characteristics were determined from the electronic medical record. Patients were classified as anticoagulated based on medication orders and lists. Bleeding event rates were assessed using ICD billing codes. The primary outcome was rate of anticoagulant use with a secondary outcome of frequency of bleeding events.
Results: We identified 2914 people with cancer and AF. 48% (1406 of 2914) were anticoagulated, and those prescribed AC were older (71.6 ± 9.7 vs. 69.8 ± 11, p<0.01), had higher CHA2DS2-VASc scores (3.7 ± 1.8 vs. 3.5 ± 1.9, p<0.01), and were more likely to have AF before cancer (49.7% vs. 39.7%p<0.01). People with breast (61%) and skin (66%) cancers had higher rates of AC, while people with hematologic (44%), brain (39%), lung (38%), and gastrointestinal (39%) cancers were less likely to be anticoagulated. Treated patients were more likely to experience a bleeding event (32% vs. 20%, p<0.01).
Conclusions: There is a large group of people with cancer and high CHA2DS2-VASc scores who are not anticoagulated. Further research is needed to determine if this lack of AC is due to bleeding risk, lack of prescribing, or invalid use of CHA2DS2-VASc stratification in this population. Apparent bleeding rates are higher than previously published.
