Clinical characteristics and outcomes of heart failure with reduced versus preserved ejection fraction and relationship with body mass index in a nationwide multi-ethnic Asian cohort

Background

Heart failure with reduced (HFrEF) and preserved ejection fractions (HFpEF) have heterogeneous clinical presentations. Obesity is a recognised risk factor for heart failure (HF). Despite increasing prevalence of HF in Asia, contemporary population-level data assessing HF phenotypes and relationship with body mass index (BMI) remain limited. We aim to describe clinical profiles and outcomes of patients with HFrEF and HFpEF, further categorised into BMI categories, in a nationwide multi-ethnic Asian cohort.

Methods

We analysed data from a national HF registry, including HFpEF and HFrEF patients admitted to eight tertiary hospitals from 2016-2024. HFpEF was defined as left ventricular ejection fraction (LVEF)≥50% and HFrEF as LVEF≤40%. Baseline demographics, characteristics and outcomes were compared across HF phenotypes and Asian BMI cut-offs: underweight (<18.5kg/m²), normal (18.5-22.9kg/m²), overweight (23-27.4kg/m²) and obese (≥27.5kg/m²). Descriptive statistics summarised frequency distributions, mean values and differences between groups.

Results

A total of 7963 patients were included (571 HFpEF and 7,392 HFrEF); mean age 70±23 years and 69% male. HFpEF patients were older (77±12 years) and more often female (58%), while HFrEF patients younger (70±23 years) and predominantly male (71%). Obese patients were younger (64±14.1 years) than underweight patients (77±11 years). Chinese ethnicity was most common in both HFpEF and HFrEF (70% vs. 64%). Comorbidities were similarly prevalent in HFpEF and HFrEF, comprising hypertension (82% vs. 81%, respectively), dyslipidaemia (80% vs. 75%), ischaemic heart disease (68% vs. 72%) and diabetes (49% vs. 56%). Overweight and obese patients had highest burden of metabolic comorbidities including diabetes and dyslipidaemia (Figure 1A).

Obesity was associated with lower NT-proBNP levels, reflecting obesity-related suppression NT-proBNP (Figure 1B). NT-proBNP was lower in HFpEF than HFrEF (median 5,300 vs. 12,000pg/mL). All-cause mortality progressively increased with decreasing BMI categories (obese 34%, overweight 42%, normal 50% and underweight 56%), and was highest in underweight and lowest in obese patients across HF phenotypes (underweight vs. obese: HFpEF 33% vs. 7%; HFrEF 57% vs. 37%), consistent with the "obesity paradox."

Conclusion

HFpEF and HFrEF demonstrated distinct demographic, metabolic and biomarker profiles. Consistent with prior studies, underweight patients had higher all-cause mortality and emerged as a high-risk phenotype across HF phenotypes, despite lower prevalence of traditional cardiovascular risk factors. This possibly reflects associated frailty, sarcopenia or cachexia. In contrast, overweight-obese patients had better survival despite higher comorbidities burden. These findings highlight the importance of personalised, phenotype-driven approach in HF, incorporating nutritional and frailty assessments to guide treatment strategies and optimise weight management.For image description, please refer to the figure legend and surrounding text.