Epicardial adipose tissue attenuation as an independent marker of cardiovascular outcomes in de novo heart failure (RECOVER-EAT).

BACKGROUND: Epicardial adipose tissue (EAT) is a visceral fat depot implicated in cardiovascular disease. While EAT volume has been associated with risk, it provides limited information regarding adipose tissue composition. Computed tomography (CT)-derived EAT attenuation is a marker of adipose tissue composition, but its prognostic significance in heart failure (HF) remains uncertain. We evaluated the association between EAT attenuation and adverse cardiovascular outcomes in HF. METHODS: We studied patients with HF undergoing clinically indicated cardiac or non-cardiac CT at a tertiary referral centre. Patients were included if CT was performed within ±6 months of echocardiography. EAT volume and attenuation, defined as the mean Hounsfield unit (HU) value of voxels within the segmented EAT volume, were quantified. The primary endpoint was a composite of cardiovascular death, myocardial infarction, stroke/transient ischaemic attack, and HF hospitalisation. Associations were assessed using multivariable Cox proportional hazards models, and incremental prognostic value was evaluated using likelihood ratio testing. RESULTS: Among 529 patients with de novo HF (mean age 58.0 ± 14.5 years; 29.7% female), 522 had sufficient longitudinal follow-up for time-to-event analysis. Of these, 87 experienced the primary composite outcome during a median follow-up of 24 months. In the fully adjusted model incorporating EAT attenuation and volume simultaneously, higher EAT attenuation was independently associated with increased risk of the primary outcome (HR 1.35 per 5-HU increase; 95% CI 1.14-1.60; p < 0.001), as was greater EAT volume (HR 1.06 per 10-mL increase; 95% CI 1.01-1.11; p = 0.019). Addition of EAT volume to the clinical model did not improve model fit (Δχ2 = 0.87; p = 0.352), whereas subsequent addition of EAT attenuation significantly improved model fit (Δχ2 = 11.98; p < 0.001). Kaplan-Meier analysis demonstrated lower event-free survival among patients with higher EAT attenuation (log-rank p = 0.03). CONCLUSIONS: EAT attenuation and volume were independently associated with adverse cardiovascular outcomes in HF when considered simultaneously. EAT attenuation provided incremental prognostic information beyond established clinical risk markers and EAT volume, supporting the potential relevance of CT-derived adipose tissue composition for cardiovascular risk assessment. LAY SUMMARY: Fat surrounding the heart undergoes important biological changes that may influence cardiovascular risk. In this study of patients with heart failure, both the amount and CT-derived composition of epicardial adipose tissue (EAT) were independently associated with adverse cardiovascular outcomes. Higher EAT attenuation, reflecting altered adipose tissue composition, provided additional prognostic information beyond clinical risk factors and EAT volume. These findings suggest that assessment of EAT composition may complement conventional measures of adipose tissue burden for identifying patients at higher cardiovascular risk.