Heart failure with mid-range or mildly reduced ejection fraction

Abstract

Left ventricular ejection fraction (EF) remains the major parameter for diagnosis, phenotyping, prognosis and treatment decisions in heart failure. The 2016 ESC heart failure guidelines introduced a third EF category for an EF of 40-49%, defined as heart failure with mid-range EF (HFmrEF). This category has been largely unexplored compared with heart failure with reduced EF (HFrEF; defined as EF <40% in this Review) and heart failure with preserved EF (HFpEF; defined as EF ≥50%). The prevalence of HFmrEF within the overall population of patients with HF is 10-25%. HFmrEF seems to be an intermediate clinical entity between HFrEF and HFpEF in some respects, but more similar to HFrEF in others, in particular with regard to the high prevalence of ischaemic heart disease in these patients. HFmrEF is milder than HFrEF, and the risk of cardiovascular events is lower in patients with HFmrEF or HFpEF than in those with HFrEF. By contrast, the risk of non-cardiovascular adverse events is similar or greater in patients with HFmrEF or HFpEF than in those with HFrEF. Evidence from post hoc and subgroup analyses of randomized clinical trials and a trial of an SGLT1-SGLT2 inhibitor suggests that drugs that are effective in patients with HFrEF might also be effective in patients with HFmrEF. Although the EF is a continuous measure with considerable variability, in this comprehensive Review we suggest that HFmrEF is a useful categorization of patients with HF and shares the most important clinical features with HFrEF, which supports the renaming of HFmrEF to HF with mildly reduced EF.

Fig. 1. Phenotype, risk of cause-specific outcomes and effect of therapies in HFrEF, HFmrEF and HFpEF.

Phenotype, risk of cause-specific outcomes and demonstrated or potential effect of treatments across the left ventricular ejection fraction categories of heart failure. Heart failure with mildly reduced ejection fraction (HFmrEF) shares features with both heart failure with reduced ejection fraction (HFrEF), such as a higher prevalence of ischaemic heart disease and less frequent renal impairment, and heart failure with preserved ejection fraction (HFpEF), such as hypertension, milder heart failure symptoms and lower levels of natriuretic peptides. HFmrEF is intermediate between the two categories for age and prevalence of atrial fibrillation. Cardiovascular mortality is lower in patients with HFmrEF and those with HFpEF than in patients with HFrEF. Non-cardiovascular mortality is lower in patients with HFmrEF and those with HFrEF than in patients with HFpEF. Post hoc and subgroup analyses of trials of heart failure suggest a potential benefit of therapy with a mineralocorticoid-receptor antagonist (MRA), angiotensin-receptor blocker–neprilysin inhibitor (ARNI) or sodium–glucose cotransporter 2 inhibitor (SGLT2i) in patients with HFmrEF. ↑ and ↓ denote higher or more common and lower or less common, respectively, than in an age-matched control population, with the exception of age, in which ↑ denotes higher than average among adults; + denotes strength of benefit; ± denotes insufficient evidence. CRT, cardiac resynchronization therapy; ICD, implantable cardioverter–defibrillator; RAS, renin–angiotensin system.

Fig. 2. HFmrEF clinical characteristics and similarities to HFrEF and HFpEF in major HF registries.

Summary of data from major registries of heart failure (HF) across the ejection fraction (EF) spectrum^,,,,–,. Data from the HF with mildly reduced EF (HFmrEF) category is shown with a blue background when resembling data from the HF with reduced EF (HFrEF) category, in red when resembling data from the HF with preserved EF (HFpEF) category and blue/red when intermediate. Values are mean (median for N-terminal pro-B-type natriuretic peptide (NT-proBNP)) and percentages. AF, atrial fibrillation; CCS, chronic coronary syndrome; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; SBP, systolic blood pressure. ^aPercentage of patients in each cohort or trial who were in the respective EF category. ^bSerum creatinine level (mg/dl); eGFR not provided. ^cIschaemic heart disease as HF aetiology. ^dBNP provided instead of NT-proBNP. ^eTwo HFrEF categories (EF <25% and 25–40%), HFmrEF defined as EF 40–55% and HFpEF defined as EF ≥55%. ^fHistory of myocardial infarction.

Fig. 3. HFmrEF clinical characteristics and similarities to HFrEF and HFpEF in major RCTs of HF.

Summary of data from randomized clinical trials (RCTs) of heart failure (HF) across the ejection fraction (EF) spectrum^,,,,. Data from the HF with mildly reduced EF (HFmrEF) category is shown in blue when resembling the HF with reduced EF (HFrEF) category, in red when resembling the HF with preserved EF (HFpEF) category and blue/red when intermediate. Values are mean, median or percentage, as reported in the original trial publications. AF, atrial fibrillation; CCS, chronic coronary syndrome; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; NT-proBNP, N-terminal pro-B-type natriuretic peptide; SBP, systolic blood pressure. ^aPercentage of patients in each cohort or trial who were in the respective EF category. ^bPatients in sinus rhythm. ^cFour HFrEF categories (EF <20%, 20–25%, 26–34% and 35–39%). ^dHistory of myocardial infarction. ^eSerum creatinine level (mg/dl); eGFR not provided. ^fEF range ≤42.5% for HFrEF, >42.5–52.5% for HFmrEF and >52.5% for HFpEF (chosen to avoid bias towards EF values ending in 0 or 5); three HFrEF categories (EF ≤22.5%, 22.5–32.5% and >32.5–42.5%) and two HFpEF categories (EF >52.5–62.5% and >62.5%). ^gThree HFpEF categories (EF 50.00–54.99%, 55.00–59.99% and ≥60.00%).

Fig. 4. Outcomes according to EF in patients with HF in major registries and RCTs.

All-cause mortality (panels a,d), cardiovascular mortality (panels b,e) and hospitalization for heart failure (HF) (panels c,f) according to left ventricular ejection fraction (EF) in patients with HF in major registries and randomized clinical trials (RCTs). Data from refs^,,,,,,,,,. ESC-HF-LT included prevalent and incident HF; 1-year rates are shown for all outcomes. SwedeHF included prevalent and incident HF; 1-year rates of all-cause death are shown; no data on cardiovascular mortality or hospitalization for HF were available. GWTG-HF included in-hospital new-onset and worsening HF; 1-year all-cause mortality and 1-year hospitalization for HF estimated from the 1-year time point on the x-axis and percentage on the y-axis of Kaplan–Meier curves in ref.; data on 5-year all-cause mortality and 5-year rate of hospitalization for HF shown in table 3 of the original article; no data on cardiovascular mortality were available. TIME-CHF included prevalent HF; rate of 1-year hospitalization for HF or death is shown; rates estimated from the 1-year time point on the x-axis and percentage on the y-axis of Kaplan–Meier curves in ref.; data on all-cause mortality and rate of hospitalization for HF during the overall follow-up reported in the main text of the original article; no data on cardiovascular mortality were available. CHART-2 (ref.) included prevalent and incident HF; 1-year rates for all outcomes estimated from the 1-year time point on the x-axis and percentage on the y-axis of Kaplan–Meier curves in ref.. TOPCAT and TOPCAT Americas trials: 1-year rates are shown for all outcomes, patients with EF ≥50% were divided into subcategories (50.00–54.99%, 55.00–59.99% and ≥60.00%); we reported the ranges of event rates if different across subcategories (asterisks); event rates for patients with EF <50% are provided in table 3 of ref.. BB-meta-HF: rates at a median follow-up of 1.3 years are shown for all outcomes in patients in sinus rhythm; patients with EF <40% were divided into subcategories (<20%, 20–25%, 26–34% and 35–39%); we reported the ranges of event rates if different across subcategories (asterisks); event rates in the 40–49% and ≥50% categories were directly provided in the trial publication, and we converted them into an estimated event rate per 100 patient-years by dividing by 1.3; data on the rate of first hospitalization for HF were not provided. DIG: 1-year rates shown for all outcomes; we reported the ranges of event rates if different between the treatment and placebo groups (asterisks). CHARM: 1-year rates provided for all outcomes. PARADIGM-HF–PARAGON-HF: no global follow-up duration was provided in the trial publication; therefore, we estimated rates per 100 patient-years by dividing by 2.25 (the median follow-up duration of PARADIGM-HF) for the EF <40% category and by dividing by 2.9 (the median follow-up duration of PARAGON-HF) for the EF 40–49% and EF ≥50% categories; data on all-cause mortality were not provided. HFmrEF, heart failure with mildly reduced ejection fraction; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; NA, not available.

Fig. 5. Trajectories and changes in EF and outcomes over time in patients with HF.

Data from a retrospective, nationwide, registry study of patients with heart failure (HF). a | Each bar segment shows the proportion of patients with changes in left ventricular ejection fraction (EF) at follow-up and the associated hazard ratio (HR) and 95% confidence interval (CI) of all-cause death or hospitalization for HF by EF category at baseline. In each category, the group of patients with a stable EF over time was used as reference. b | The risk of all-cause death or hospitalization for HF relative to EF changes. adj, adjusted; HFmrEF, heart failure with mildly reduced ejection fraction; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction. Parts c and d adapted with permission from ref., Elsevier.

Fig. 6. Outcomes in patients with HF according to EF.

Forest plot depicting hazard ratio (HR) and 95% confidence interval (CI) for intervention versus control according to left ventricular ejection fraction (EF) in patients with heart failure (HF) with reduced EF (HFrEF), HF with mildly reduced EF (HFmrEF) and HF with preserved EF (HFpEF) for the specified trials and outcomes. In the CHARM programme^,, therapy with candesartan reduced the risk of cardiovascular (CV) death and hospitalization for HF (HHF) in patients with HFrEF or HFmrEF but not in those with HFpEF. In PEP-CHF, perindopril therapy was not effective in reducing the risk of all-cause death or HHF in elderly patients with mildly impaired systolic dysfunction. A retrospective analysis of the DIG trial did not detect a benefit of digoxin therapy when the EF was >40%. The PEACE trial^, did not show a reduction in the risk of CV death and HHF with angiotensin-converting enzyme inhibitors in patients with ischaemic HF with mildly reduced EF. In the TOPCAT Americas substudy, the risk of the primary composite end point of CV death, HHF or aborted cardiac arrest (CA) was reduced with spironolactone therapy in patients with EF at the lower end of the included EF spectrum (EF ≥45%). The BB-meta-HF study, a large meta-analysis of trials on β-blockers, demonstrated a lower all-cause and CV mortality in patients in sinus rhythm with HFrEF or HFmrEF but not in those with HFpEF. In a prespecified subgroup analysis of the PARAGON-HF trial, therapy with sacubitril–valsartan was effective in patients with EF lower than or the same as the median (EF 57%). For the HFrEF (EF <40%) category of the BB-meta-HF study, we report the ranges of HR and 95% CI across the subgroups of HFrEF considered in the study (EF <20%, 20–25%, 26–34% and 35–39%). For the HFrEF and HFpEF categories of the PARADIGM-HF–PARAGON-HF study, we report the ranges of HR and 95% CI across the subgroups of HFrEF (EF ≤22.5%, 22.5–32.5% and >32.5–42.5%) and HFpEF (EF >52.5–62.5% and >62.5%) considered in the study.