Aims: Management of patients with a single CHA2DS2-VASc score risk factor is controversial. We attempt to identify the "truly low risk" AF patients who will not benefit from oral anticoagulation (OAC) treatment.
Methods: Retrospective cohort analysis, all incident non-valvular AF (NVAF) cases between 2004 and 2015, and age 21 and older, with up to one thromboembolic risk factor besides sex (CHA2DS2-VASc score of up to 1 for men and up to 2 for women). A "low risk" score was created for these patients using a logistic regression model on the incidence of stroke within 30-2500 days following the NVAF diagnosis.
Results: We identified 15,621 patients. Average age was 53.7 ± 12.3 years, 56.6% male. Mean follow-up was 5.5 years. Significant predictors of ischemic stroke were age 65-74 and diabetes (2 points each), hypertension, vascular disease, and chronic kidney disease stage 2-3 (1 point each). Stroke incidence ranged from 0.8% for score 0 and up to 3.4% for scores ≤ 2. Odds ratio for stroke among patient group with a score ≤ 2 was 4.3 (2.9-6.6) compared with score 0. Our risk score's area-under-the-curve (AUC) for prediction of stroke was 0.68 (0.65-0.71), compared with 0.60 (0.57-0.62) for the CHAD2S2-VASc score, within this low-risk group.
Conclusion: Patients considered at low or intermediate risk using traditional risk stratification schemes, with ≥ 2 points using this proposed low-risk index (65-74 years old, diabetics or a combination of chronic renal failure and an additional risk factor), had an overall stroke risk that may justify anticoagulation therapy.
Keywords: Anticoagulants; Atrial fibrillation; Stroke; Thromboembolic risk.
© 2021. Royal Academy of Medicine in Ireland.