Abstract Background/Introduction Heart failure with preserved ejection fraction (HFpEF) refers to a heterogeneous group of patients with unexplored subtypes in terms of a guide to treatment or prognosis. Left ventricular (LV) remodeling is very important as not only an index of structural and functional changes, but also a prognosticator in HF patients. Aims To investigate whether patients with HFpEF can be classified into subgroups by degree of LV remodeling and whether it is related to survival. Methods Patients with HFpEF in the China Heart Failure Center Registry (an ongoing large, nationwide, multicenter registry enrolling the admitted HF patients from secondary and above hospitals since 2017) who had data on the in-hospital LV end-diastolic dimension (LVEDD), ejection fraction (LVEF) by echocardiography, and the 1-year survival were recruited. Patients with major structural heart diseases were excluded. The primary endpoint was all-cause mortality during 1 year after discharge. Kaplan-meier survival curve and Cox regression analysis were adopted to compare among the patients stratified by the LVEDD and LVEF. Results A total of 29727 patients with HFpEF (mean age 73±12 years, 51.3% male) were included, of whom 69.8% was in hypertension, 30.4% in diabetes, 37.3% in atrial fibrillation or flutter, and 14.1% in chronic kidney disease. The mean LVEF of the study population was 60±6%. When the patients were divided into 5 groups as every 5% increase in the LVEF, the risk of all-causemortality was increased in the groups of 50-54% [23.3% patients, age- and gender-adjusted hazard ratio (HR)=1.414, p<0.001] and55-59% (28.8%,1.281, p=0.001), while not different in the groups of 60-64% (26.1%,1.056, p=0.493) and 65-69% (14.3%,0.969,p=0.720) when compared with the group of ≥70% (7.6%) as the reference (Figure 1a). The mean LVEDD was 48±7mm. When the patients were divided into 6 groups as every 5mm increase in the LVEDD, the groups of <40mm (7.7% patients, age-and gender-adjusted HR=1.548, p<0.001) and ≥60mm (5.3%, 1.509, p<0.001) had the highest risk ofmortality, followed by the group of 40-44mm (22.2%,1.214, p=0.001), while the groups of 45-49mm (33.0%,0.991, p=0.863) and 55-59mm (10.0%,1.100, p=0.200) had similar risk when compared with the group of 50-54mm (21.7%) as the reference (Figure 1b). Furthermore, the patients were divided into 8 phenotypes of LV remodeling by using the LVEDD (<40mm, 40-44mm, 45-59mm, ≥60mm) and the LVEF divisions (<60%, ≥60%). When the phenotype of LVEDD of 45-59mm and LVEF of ≥60% taken as the reference, all other phenotypes had an increased risk of death by age and sex adjusted Cox regression analysis (Table1). Conclusion(s) Both the LVEF and LVEDD can provide risk stratification for patients with HFpEF. Those patients with a LVEDD of mid-ranged 45-59mm and a high LVEF ≥60% had the lowest risk of mortality. This concept of phenotyping HFpEF by LV remodeling could be considered in future clinical practice and research.KM survival stratified by EF and EDDRisk of death stratified by EF&EDD