Differences in clinical characteristics and outcomes in real-world HFpEF patients according to clinical definitions

Abstract Background Heart failure with preserved ejection fraction (HFpEF) is characterised by its heterogeneity, which extends to its definition. How HFpEF is defined has important consequences for recruitment into clinical trials and clinical outcomes, and several definitions have been used across clinical guidelines and trials. Purpose The aim of this study was to characterise a cohort of newly-diagnosed community HFpEF patients and their outcomes according to HFpEF definitions used in recent guidelines and outcome trials. Methods We conducted a single-centre study of patients who underwent echocardiography for suspected symptomatic HF and elevated NT-proBNP (>125 pg/ml). Patients were classified as to whether they met the HFpEF diagnostic criteria using the ESC 2016 HF Guidelines,1 the H2FPEF criteria (score ≥6) 2 and definitions used in CHARM-Preserved,3 I-Preserve,4 TOPCAT,5 PARAGON-HF6, HFA PEFF7 and EMPEROR-Preserved.8 The primary outcome was time to mortality or cardiovascular hospitalisation. Results In total, there were 282 patients evaluated (mean age 78±9 years; 63.5% female; median NT-proBNP 1199 pg/ml). As expected there was a high prevalence of comorbidities (68% hypertension, 49% obesity, 36% atrial fibrillation, 21% ischaemic heart disease and 17.4% diabetes). All patients met the CHARM-Preserved criteria, while the H2FPEF criteria were the most restrictive, with only 69 patients (24.5%) were defined as HFpEF (Table). HFA-PEFF criteria identified 86 patients (30.5%) as HFpEF. One hundred and eighty three patients (65%) met the ESC 2016 criteria. Recent clinical trials' definitions included a wide range of patients (PARAGON-HF 58%, EMPEROR 62%, TOPCAT 72% and I-Preserved 76%). Among eight definitions and criteria for HFpEF, more AF (95.7%), obesity (73.9%) and diabetes (26.1%) were identified in H2FPEF group compare to the others.(Picture1) Median follow-up was 18±9 months. Over the follow-up period, in the whole cohort 46 patients suffered the primary outcome (16.3%), including 11 deaths and 36 cardiac related hospitalisations. The incidence of the primary outcome was highest in patients meeting the H2FPEF and EMPEROR definition (24%). The I-PRESERVE and TOPCAT criteria were most discriminatory. Clinical trial definitions gave similar event rates to the H2FPEF criteria but included substantially more patients (Picture 2). Conclusions There is significant variation in the clinical characteristics and outcomes of HFpEF patients depending on