Heart failure with preserved ejection fraction: the impact of stricter definitions

Aims To determine the impact of internationally recognized criteria for the diagnosis of heart failure with preserved ejection fraction (HF-PEF) on the feasibility of recruitment into a HF-PEF trial. Methods and results We reviewed the medical records of 5883 consecutive patients admitted with heart failure across eight hospitals in the UK. From those who had an EF >= 50% (filter 1), patients who had an alternative cardiac cause of symptoms or an over-riding co-morbidity were excluded (filter 2). The remaining patients had to satisfy the following selection criteria (filter 3): the presence of a non-dilated left ventricle; structural remodelling (LV hypertrophy or dilated left atrium); and evidence of impaired diastolic function or raised BNP. Overall, 20.5% (n=1203) of patients admitted with heart failure had an EF >= 50%. There was significant heterogeneity between the hospitals as this proportion varied from 42.4% in the teaching hospitals to 13.8% in the district general hospitals (P < 0.0001). Applying filter 2, the cohort size was reduced from 1203 to 122 patients. Of the latter 122, only 45 people had evidence of structural cardiac remodelling and evidence of either raised cardiac filling pressures on echocardiography or raised BNP consistent with HF-PEF. Conclusion Applying our filters, only 0.8% of all patients admitted with heart failure would be eligible for a HF-PEF trial. These data have important implications for the feasibility of clinical trials for HF-PEF and for the generalizability of the results of such studies.