“Real-world prognostic performance of different severe and advanced heart failure definitions: Data from the Swedish Heart Failure Registry”

Background A valid definition of severe heart failure (HF) is essential for earlier identification, timely referral for advanced therapies, and to optimize clinical trial design. Objectives To assess the prevalence, prognostic performance, and 1-year outcomes associated with different definitions of severe and advanced HF (AdvHF) in patients with heart failure with reduced ejection fraction (HFrEF). Methods We included 15,153 patients with EF '40%, HF duration '6 months, and no prior left ventricular assist device or heart transplant, from the Swedish Heart Failure Registry. Several definitions of severe/AdvHF were evaluated: a simplified Heart Failure Association (HFA) definition (NYHA III–IV, EF '30%, ≥2 HF hospitalizations in 12 months), the GALACTIC-Severe-definition (NYHA III–IV, EF '30%, ≥1 hospitalization in 6 months), and variations adding/removing criteria such as NT-proBNP levels and diuretic dose to the HFA-AdvHF definition. Predictive performance and 1-year outcomes were analyzed. Results The HFA-definition identified 6.4% of patients, with a 1-year cardiovascular death/HF hospitalization risk of 70% and all-cause mortality of 45% (AUC = 0.72). The GALACTIC-Severe definition identified 12.6% of patients with slightly lower but still substantial event rates (59% and 38%, respectively; AUC = 0.73). Definitions incorporating NT-proBNP ≥2000 pg/ml had the highest prognostic accuracy (AUC = 0.75). Conclusions The HFA-AdvHF definition selected a smaller, high-risk group, while the GALACTIC-Severe criteria identified a broader population, with a lower, but still high risk. NT-proBNP, HF hospitalization history, and diuretic dose might represent valuable enrichment tools for future trials.