Biomarker-Based ABC-AF Risk Scores for Personalized Treatment to Reduce Stroke or Death in Atrial Fibrillation: A Registry-Based, Multicenter, Randomized, Controlled Study

Abstract

Background: The clinical use of risk scores to guide treatment decisions and improve clinical outcomes has rarely been prospectively evaluated. This study aimed to evaluate whether a biomarker-based ABC-AF risk score-guided multidimensional treatment strategy improves long-term outcomes in patients with AF.

Methods: The multicenter, registry-based, randomized, controlled, open-label study enrolled adults with AF. In the ABC-AF strategy arm, the investigator was informed of each individual's ABC-AF score risks for stroke and bleeding, which were used as decision support to tailor treatment recommendations, including preference for type of direct oral anticoagulant treatment. In the standard of care arm, patient management was at the discretion of the investigator. Primary outcome was a composite of stroke or death. Secondary outcomes included stroke, death, major bleeding events, and their composite outcome.

Results: The intention-to-treat population comprised 3933 patients with a median age of 73.7 years; 33.6% were women, 51.3% had paroxysmal AF, 11.2% had a previous stroke or transient ischemic attack, and 85.7% had oral anticoagulant treatment. After randomization, 97.8% in the ABC-AF strategy arm and 92.6% in the standard of care arm received OACs (P<0.0001). Enrollment was prematurely terminated owing to safety concerns with a trend toward higher mortality in patients with CHA₂DS₂-VASc scores of ≥3, and the study was therefore underpowered for its primary objective. Over a median follow-up of 2.6 years, 175 primary events (3.18/100 patient-years [100PY]) occurred in the ABC-AF strategy and 148 (2.67/100PY) in the standard of care arm (hazard ratio [HR], 1.19 [95% CI, 0.96-1.48]; P=0.12). Major bleeding events were 152 (2.82/100PY) versus 141 (2.61/100PY; HR, 1.08 [95% CI, 0.86-1.36]; P=0.50), stroke 48 (0.87/100PY) versus 41 (0.74/100PY; HR, 1.18 [95% CI, 0.78-1.79]; P=0.44), death 136 (2.44/100PY) versus 113 (2.02/100PY; HR, 1.21 [95% CI, 0.94-1.55]; P=0.13), and rates of composite stroke, death, or major bleeding 277 (5.21/100PY) versus 244 (4.55/100PY; HR, 1.14 [95% CI, 0.96-1.36]; P=0.13). Primary outcome results were similar across ABC-AF score subgroups (interaction P=0.98).

Conclusions: The individually tailored multidimensional treatment strategy, based on ABC-AF risk scores, did not improve clinical outcomes compared with usual guideline-based care in patients with AF. The results emphasize the need for prospective testing of the use of risk stratification and precision medicine tools in different clinical settings before implementation in routine care.

Registration: URL: https://www.clinicaltrials.gov; Unique identifier: NCT03753490.

Keywords: atrial fibrillation; biomarkers; intracranial hemorrhages; risk factors; stroke.

Figure 1.

One-year persistence to key medical treatments. One-year persistence to key medical treatments in the ABC-AF strategy arm is shown in red and standard of care in blue. ACE indicates angiotensin-converting enzyme; ARB, angiotensin receptor blocker; and PPI, proton pump inhibitor.

Figure 2.

Primary outcome. The primary composite outcome was stroke or death. Shown is a Kaplan-Meier plot with the ABC-AF strategy arm in red and standard of care arm in blue. Shown is a Cox regression model with hazard ratios and 95% CI in parenthesis. HR indicates hazard ratio; and SoC, standard of care.

Figure 3.

Secondary outcomes. Secondary outcomes were a composite of stroke, death, or major bleeding. Shown are Kaplan-Meier plots with the ABC-AF strategy arm in red and standard of care arm in blue. Shown are Cox regression models with hazard ratios and 95% CI in parentheses. HR indicates hazard ratio.

Figure 4.

Primary outcome in prespecified subgroups. Shown is the primary composite outcome of stroke or death in prespecified subgroups: sex and age; previous stroke or transient ischemic attack; and CHA₂DS₂-VASc, with high defined as ≥2 points for men and ≥3 points for women. HR indicates hazard ratio; SoC, standard of care; and TIA, transient ischemic attack.

Figure 5.

Primary outcome in four risk subgroups by ABC-AF-stroke and ABC-AF-bleeding scores. Shown is the primary composite outcome of stroke or death in 4 risk subgroups by ABC-AF-stroke and ABC-AF-bleeding scores. A, High stroke risk, low bleeding risk. B, High stroke risk, elevated bleeding risk. C, Low or elevated stroke risk, low bleeding risk. D, Low or elevated stroke risk, elevated bleeding risk. Kaplan-Meier plots show the ABC-AF strategy arm in red and standard of care arm in blue, comprising 2840 (of the 3933) patients enrolled at study centers with biobanked blood samples for patients randomized to the standard of care arm. Because of the limited number (n=65) of patients categorized as having low bleeding risk in this analysis, patients with low stroke risk and low bleeding risk (Supplemental Material, category 5) were analyzed with those with elevated stroke risk and low bleeding risk (category 3), and patients with low stroke risk and elevated bleeding risk (category 6) were analyzed with those with elevated stroke risk and elevated bleeding risk (category 4). P=0.98 for interaction (risk group×randomization). HR indicates hazard ratio.