Association of CHADS-VASc and HAS-BLED to frailty and frail outcomes: From the TREAT-AF study.

Background: Morbidity and mortality associated with high CHADS-VASc and HAS-BLED scores is not specific to atrial fibrillation (AF). Frailty could be an important contributor to this morbidity and mortality while being mechanistically independent from AF. We sought to evaluate the association of stroke and bleeding risk to noncardiovascular frail events and the association of stroke prevention therapy to outcomes in frail patients with AF.

Methods: Using the TREAT-AF (The Retrospective Evaluation and Assessment of Therapies in AF) study from the Veterans Health Administration, we identified patients with newly diagnosed AF from 2004 to 2014. Baseline frailty was identified using a previously validated claims-based index requiring ≥2 of 12 ICD-9 diagnoses. Logistic regressions modeled the association between CHADS-VASc and modified HAS-BLED and frailty. Cox proportional hazard regressions were used to evaluate the association between CHADS-VASc and modified HAS-BLED and a composite of noncardiovascular frail events (fractures, urinary tract infections, bacterial pneumonia, or dehydration). We also evaluated the association of oral anticoagulant (OAC) use with stroke, bleeding, and 1-year mortality in frail patients and non-frail patients.

Results: In 213,435 patients (age 70 ± 11; 98% male; CHADS-VASc 2.4 ± 1.7) with AF, 8,498 (4%) were frail. CHADS-VASc > 0 and HAS-BLED > 0 were strongly associated with frailty (odds ratio [OR] 13.3 (95% CI: 11.6-15.2) for CHADS-VASc 4+ and OR 13.4 (10.2-17.5) for HAS-BLED 3+). After adjusting for covariates, CHADS-VASc, and HAS-BLED > 0 were associated with higher risk of non-cardiovascular frail events (hazard ratio [HR] 2.1 (95% CI: 2.0-2.2) for CHADS-VASc 4+ and HR 1.4 (95% CI: 1.3-1.5) for HAS-BLED 3+). In frail patients, OAC use was associated with significantly lower risk of 1-year mortality (HR 0.82; 95% CI 0.72-0.94, P = .0031) but did not reach significance for risk of stroke (HR 0.80; 95% CI 0.55-1.18, P = .26) or major bleeding (HR 1.08; 95% CI 0.93-1.25, P = .34).

Conclusions: High CHADS-VASc and HAS-BLED scores are strongly associated with frailty. However, in frail patients, OAC use was associated with reduction in 1-year mortality. For this challenging clinical population with competing risks of frailty and frail events, focused prospective studies are needed to support clinical decision-making. Until then, careful evaluation of frailty should inform shared decision-making.