Sex-specific FIND-AF risk thresholds aligned with CMR-defined atrial cardiomyopathy identify excess cardiovascular risk in heart failure across nationwide primary care data
K Kazi, C Hayward, A Helbitz, Y Nakao, P Swoboda, J Wu, C P Gale, R NadarajahAbstract
Background/Introduction
Atrial cardiomyopathy (AtCM) is common in heart failure (HF) and may predispose to adverse cardiovascular outcomes. Risk prediction tools have been developed to identify individuals at increased risk of AF, yet it is unclear the extent to which such tools capture AtCM and the prognostic implications.
Purpose
To derive sex-specific thresholds for a machine-learning AF risk score based on a CMR-defined AtCM phenotype in patients with HF and explore association with prognosis in a nationwide HF cohort.
Methods
In a CMR-characterised HF cohort1, predicted AF risk was calculated using the FIND-AF model2. Atrial and myocardial CMR characteristics were compared across risk deciles and against participants with known AF. Sex-specific high-risk thresholds were defined at the 90th percentile among participants without known AF. These thresholds were then applied to a nationwide primary care electronic health record cohort of adults with HF and no AF at baseline. The primary endpoint was a composite of HF hospitalisation, myocardial infarction, stroke, or cardiovascular death. Secondary endpoints were the individual components of the primary composite as well as all-cause death and incident AF. Associations were estimated using time-to-event models accounting for competing risk of death.
Results
The CMR cohort included 385 patients with HF (mean age 62±13 years; 39% female). Participants above the sex-specific thresholds exhibited left atrial characteristics comparable to those with established AF (Fig 1). In the nationwide HF cohort (n=657,277; median age 77.5 years), high-risk classification was associated with higher cumulative incidence of the primary composite endpoint at 5 years (684.5 vs 548.7 per 1000) (Fig 2). In adjusted analyses (at 5 years), high-risk status was independently associated with the primary composite outcome (HR 1.39, 95% CI 1.38–1.41). Stroke and incident AF showed marked separation (200.3 vs 118.1 per 1000; AF 136.8 vs 90.9 per 1000), with adjusted HRs of 1.73 (1.69–1.76) and 1.40 (1.36–1.44), respectively.
Conclusion(s)
Sex-specific FIND-AF thresholds aligned with a CMR-defined AtCM phenotype identify a heart failure subgroup with excess composite cardiovascular events. These findings demonstrate that high risk HF patients can be identified to consider for targeted rhythm surveillance and AtCM therapies.CMR traits by FIND-AF decile and AFPrimary composite by risk group