Key take-aways
- Novel biomarker-based risk scores, such as the ABC-AF scores, have been developed to estimate stroke and bleeding risk in patients with atrial fibrillation (AF).
- The ABC-AF trial evaluated whether tailoring of treatment recommendations based on patients' ABC-AF risk scores could improve clinical outcomes as compared with usual guideline-based care.
- The individually tailored risk-based treatment strategy was not associated with reduced death, stroke or major bleeding compared with usual care.
- These findings emphasise the need for prospective testing of the utility of risk stratification tools in different clinical settings before implementation in routine care.
Madrid, Spain – 30 August 2025: An individually tailored multidimensional risk-based treatment strategy was not associated with improvements in clinical outcomes compared with usual guideline-based care in patients with atrial fibrillation (AF), according to late-breaking research presented in a Hot Line session today at ESC Congress 2025.1
Patients with AF are at increased risk of stroke and are often treated with oral anticoagulant (OAC) therapy, mainly direct OACs or warfarin. While OACs substantially reduce the risk of stroke, they may also increase the risk of bleeding events. ESC Guidelines recommend that a risk-based approach is used to make decisions on OACs and other treatments for stroke prevention in patients with AF.2
Several biomarker-based risk scores for stroke and bleeding have been developed. These include the biomarker-based ABC-AF-stroke score (Age, Biomarkers [NT-proBNP and hs-troponin T] and Clinical history of stroke/transient ischaemic attack) and the ABC-AF-bleeding score (Age, Biomarkers [growth differentiation factor 15, haemoglobin and hs-troponin T] and Clinical history of bleeding).
Principal Investigator, Professor Jonas Oldgren from Uppsala University, Uppsala, Sweden, explained why and how implementation of these risk scores was assessed in the ABC-AF trial: "While novel biomarker-based risk scores have been validated in different populations, the clinical utility of risk scores to guide treatment decisions and improve clinical outcomes has rarely been prospectively evaluated. We performed a pragmatic registry-based trial to evaluate whether the tailoring of treatment recommendations based on patients' ABC-AF risk scores improves clinical outcomes as compared with usual guideline-based care in patients with AF."
This open-label registry-based, randomised controlled study enrolled patients from 39 Swedish sites. Eligible patients were adults with a diagnosis of AF, including newly or previously diagnosed AF, with or without current OAC treatment. Patients were randomised 1:1 to either an ABC-AF risk score-guided treatment strategy or to standard of care. Plasma samples were obtained at randomisation and ABC-AF risk scores were automatically calculated based on the ABC-AF-stroke variables and the ABC-AF-bleeding variables. The investigator received a visual presentation of the ABC-AF risk scores along with specific treatment recommendations tailored to each individual patient. Based on this information, the investigator decided on medical treatments and other interventions. Patients in the control arm were managed in accordance with usual practices at the discretion of the investigator, without measurements of any ABC-AF risk score biomarkers and without individual treatment recommendations. The primary outcome was a composite of stroke or death. Data for study outcomes were retrieved from mandatory national registers with complete coverage of all in-patient care at Swedish hospitals and vital status for all Swedish residents. Enrolment was prematurely terminated owing to safety concerns with a trend towards higher mortality in patients with CHA2DS2-VASc scores of 3 or above.
The intention-to-treat population comprised 3,933 patients who had a median age of 73.9 years and 33.6% were women. In total, 51.3% had paroxysmal AF, 11.2% had prior stroke or transient ischaemic attack and 85.7% had OAC treatment.
After randomisation, the proportion of patients receiving any OAC increased to 97.8% in the active arm and 92.6% in the control arm (p<0.0001). In the active group, there were changes in the use of some direct OACs – increased use of apixaban and dabigatran, and reductions in rivaroxaban – and reductions in the use of warfarin. In the control group, there was increased use of apixaban, edoxaban and rivaroxaban, and reductions in the use of warfarin. In both study groups, the proportion of patients on antiplatelet treatment was halved and statin treatment increased.
Over a median follow-up of 2.6 years, the primary outcome occurred at a similar rate between the groups: 3.18/100 patient-years (100PY) in the active group and 2.67/100PY in the control group (hazard ratio [HR] 1.19; 95% confidence interval [CI] 0.96 to 1.48; p=0.12).
There were no significant differences in the active vs. control groups in the rate of stroke (0.87/100PY vs. 0.74/100PY; HR 1.18; 0.78 to 1.79; p=0.44) or death (2.44/100PY vs. 2.02/100PY; HR 1.21; 0.94 to 1.55; p=0.13).
Major bleeding events occurred at a similar rate in the active vs. control group (2.82/100PY vs. 2.61/100PY; HR 1.08; 95% CI 0.86 to 1.36; p=0.50).
Professor Oldgren concluded: "We found no benefit of individually tailored, multidimensional treatment recommendations based on ABC-AF-stroke and ABC-AF-bleeding risk scores compared with usual guideline-based care in this study population who had lower-than-expected event rates. Due to premature termination of recruitment, the study was underpowered for its primary objective, but we will continue to follow-up randomised patients to assess any long-term effects. Overall, the results emphasise the need for prospective testing of the utility of risk stratification and precision medicine tools in different clinical settings before being implemented for tailoring of treatment in routine clinical care."
