Today, I review, link to, and excerpt from How to diagnose heart failure with preserved ejection fraction: the HFA-PEFF diagnostic algorithm: a consensus recommendation from the Heart Failure Association (HFA) of the European Society of Cardiology (ESC) [PubMed Abstract] [Full-Text HTML] [Full-Text PDF]. Eur Heart J. 2019 Oct 21;40(40):3297-3317. doi: 10.1093/eurheartj/ehz641.
There are 255 similar articles in PubMed.
The above article has been cited by 451 articles in PubMed.
All that follows is from the above resource.
Abstract
Making a firm diagnosis of chronic heart failure with preserved ejection fraction (HFpEF) remains a challenge. We recommend a new stepwise diagnostic process, the ‘HFA-PEFF diagnostic algorithm’. Step 1 (P=Pre-test assessment) is typically performed in the ambulatory setting and includes assessment for HF symptoms and signs, typical clinical demographics (obesity, hypertension, diabetes mellitus, elderly, atrial fibrillation), and diagnostic laboratory tests, electrocardiogram, and echocardiography. In the absence of overt non-cardiac causes of breathlessness, HFpEF can be suspected if there is a normal left ventricular ejection fraction, no significant heart valve disease or cardiac ischaemia, and at least one typical risk factor. Elevated natriuretic peptides support, but normal levels do not exclude a diagnosis of HFpEF. The second step (E: Echocardiography and Natriuretic Peptide Score) requires comprehensive echocardiography and is typically performed by a cardiologist. Measures include mitral annular early diastolic velocity (e’), left ventricular (LV) filling pressure estimated using E/e’, left atrial volume index, LV mass index, LV relative wall thickness, tricuspid regurgitation velocity, LV global longitudinal systolic strain, and serum natriuretic peptide levels. Major (2 points) and Minor (1 point) criteria were defined from these measures. A score ≥5 points implies definite HFpEF; ≤1 point makes HFpEF unlikely. An intermediate score (2-4 points) implies diagnostic uncertainty, in which case Step 3 (F1: Functional testing) is recommended with echocardiographic or invasive haemodynamic exercise stress tests. Step 4 (F2: Final aetiology) is recommended to establish a possible specific cause of HFpEF or alternative explanations. Further research is needed for a better classification of HFpEF.
Keywords: HFpEF; Heart failure; biomarkers; diagnosis; echocardiography; exercise echocardiography; natriuretic peptides.
Published on behalf of the European Society of Cardiology. All rights reserved. © The Author(s) 2019. For permissions, please email: journals.permissions@oup.com.
Introduction
In the general population aged ≥60 years, 4.9% were identified to have heart failure with preserved ejection fraction (HFpEF),1 implying several millions of affected individuals in Europe. This number is expected to increase further as people live longer and obesity and diabetes become more common.1–3 Heart failure with preserved ejection fraction already accounts for more than half of all heart failure (HF) hospital admissions.1 Providing effective management is a major unmet clinical need that will depend on a clear diagnosis.
Understanding of the pathophysiology of HFpEF has advanced,9–13 diagnostic options have evolved,14–17 and this novel information needs to be integrated into a new comprehensive diagnostic algorithm for suspected HFpEF.
A writing committee initiated by the HFA of the ESC has therefore produced an updated consensus recommendation—the HFA–PEFF diagnostic algorithm (Figure 1). Its key elements are (i) the concept that identification of HFpEF involves all levels of care, including general practitioners, internists, general cardiologists, HF specialists, and invasive cardiologists; (ii) a stepwise diagnostic approach from initial clinical assessment to more specialized tests will therefore be useful; (iii) the diagnosis is not always straightforward, so the integration of distinct parameters from complementary diagnostic domains into a new diagnostic score is recommended; (iv) for the subset of patients with an inconclusive score, definitive diagnosis (or exclusion) will require invasive haemodynamics and/or non-invasive or invasive exercise stress tests; and (v) underlying pathophysiological alterations (such as chronotropic incompetence, reduced LV compliance) and specific aetiologies (such as amyloidosis18) have to be considered. A precise diagnosis is increasingly important since new targeted therapies are becoming available for defined subsets of HFpEF patients.
Why new diagnostic recommendations for heart failure with preserved ejection fraction?
The key criteria in the previous HFA recommendations were: (i) symptoms and/or signs of HF, (ii) normal or only mildly abnormal LV systolic function, and (iii) LV diastolic dysfunction.4 Diagnostic parameters were invasive measurements, echocardiographic indices of LV diastolic function and filling pressures, LV hypertrophy (LVH), left atrial (LA) enlargement, serum natriuretic peptides (NP), and atrial fibrillation (AF).4 Over time, both advantages and disadvantages of this approach have been reported.
Cut-offs for key non-invasive parameters are often based on limited data, and may fall in a non-diagnostic intermediate range. The non-invasive diagnosis or exclusion of HFpEF will not depend on a single parameter above or below a certain cut-off, but on a combination of parameters derived from clinical, laboratory, and imaging tests that together will give a probability for the diagnosis. A recent example of such an approach was a composite HFpEF diagnostic score, derived retrospectively from clinical characteristics (age >60 years, obesity, atrial fibrillation, treatment with ≥2 antihypertensive drugs) and echocardiographic measurements [E/e′ >9, pulmonary artery systolic pressure (PASP) >35 mmHg].5
Echocardiographic criteria for diagnosing heart failure with preserved ejection fraction
Left ventricular ejection fraction (LVEF) estimates global function but does not indicate LV volume or stroke volume. Despite a preserved LVEF, patients with HFpEF have impaired LV long-axis systolic function, which can be measured using mitral annular systolic excursion or systolic velocities or LV global longitudinal strain (GLS).19 As well as global diastolic dysfunction, they have long-axis diastolic dysfunction which can be measured from the velocity of long-axis lengthening of the LV in early diastole (from mitral annular velocity, e′). These were not considered in the previous HFA recommendations.4
A mean E/e′ index ≥15 at rest has good diagnostic value for identifying a high mean pulmonary capillary wedge pressure (mPCWP), supporting the likelihood of HFpEF,20 , 21 but an E/e′ ratio within the intermediate range (9–14) is less sensitive.22 The E/e′ ratio has limitations that are relevant in routine clinical practice23–29 and its use as a single diagnostic index above all other non-invasive measures of filling pressures (such as retrograde pulmonary venous flow) cannot be recommended. In consequence, HFpEF cannot be diagnosed from a single echocardiographic measure, and inclusion of recently validated functional and structural parameters into a diagnostic score may better define this heterogeneous disorder.
Usefulness of natriuretic peptides
In general, NP levels are higher in patients presenting with acute shortness of breath for cardiac reason or in acute HF, than in patients who have chronic HF.30 , 31 Of note, our recommendations target stable symptomatic HFpEF, and natriuretic peptide levels can be normal in these patients even with invasively confirmed HFpEF. In consequence, normal NP levels do not exclude HFpEF, especially in the presence of obesity.32 , 33 Interpretation depends also on whether the patient is in sinus rhythm (SR) or has AF, which itself is associated with increased NP levels even in the absence of HF.34 , 35
Besides obesity, sex, age, and renal function affect NP levels,36 , 37 but using stratified cut-points only marginally improves diagnostic accuracy (net reclassification index 3%),38 at the expense of less everyday utility. The variability of repeated measurements in individual patients is up to 100%, so a rise or fall of ≤100% may not necessarily indicate recovery or progression of disease.39 , 40
Diagnostic algorithms for heart failure with preserved ejection fraction
The concept of a diagnostic algorithm that incorporates imaging and biomarkers (NPs) was recommended by the HFA in 2007,4 and adapted by others.41 It allowed parallel diagnostic pathways starting from haemodynamic measurements, echocardiography, or NPs,4 that could yield different results for the same patients. In addition, the proportion of non-classifiable patients was substantial. Thus, our revised algorithm (see below) proposes a novel stepwise diagnostic approach that has only one entry point, and all patients will be classifiable.
Defining aetiology and pathophysiology
Heart failure with preserved ejection fraction typically evolves from a combination of risk factors and comorbidities, including advanced age, female sex, obesity, systemic arterial hypertension, diabetes mellitus, renal dysfunction, anaemia, iron deficiency, sleep disorders, and chronic obstructive pulmonary disease.1,2,11,42–44 Heart failure with preserved ejection fraction ‘masqueraders’ such as heart valve disease, arrhythmias, and pericardial constriction need to be excluded.
Similarly, a patient with a normal LVEF and HF-like symptoms caused by significant coronary artery disease (CAD) is also not considered to have HFpEF.Similar to current practice for heart failure with reduced ejection fraction (HFrEF), we recommend applying the descriptive term HFpEF for both the classical form with typical risk factors and comorbidities, and for rarer cases with a specific aetiology, provided that the key diagnostic criteria are met. Specific aetiologies that may be treatable include inherited or acquired infiltrative, restrictive, inflammatory, or genetic cardiomyopathies45–48 (Table 2).
They should always be considered once a diagnosis of HFpEF has been made (Table 2, Supplementary material online, S2–S4). It has been suggested that patients with HFrEF share a common mechanism that responds to common treatment (inhibition of the renin-angiotensin system)3 but there are other treatments for subsets of patients with HFrEF that
are specific (such as treating ischaemia when there is hibernating myocardium, using targeted antiviral therapy or immune modulation in inflammatory HFrEF, and corticosteroid therapy in sarcoidosis-related HFrEF); in that respect, our proposed use of the generic term HFpEF is similar and should include specific myocardial aetiologies.Basic mechanisms affecting the myocardium in HFpEF include
myocyte hypertrophy, systolic and diastolic dysfunction, energetic abnormalities, interstitial fibrosis, inflammation, increased oxidative stress, endothelial dysfunction, and impaired density and autoregulation of the microcirculation.9,10,12,45–48,154,155 Cardiovascular pathophysiological processes include increased systemic vascular resistance, increased conduit arterial stiffness, abnormal ventriculararterial coupling, reduced LV long-axis systolic function, slowed early diastolic relaxation, reduced LV compliance with increased enddiastolic stiffness, reduced LA reservoir and contractile function, impaired right ventricular (RV) function, and chronotropic incompetence.52,156–164 Patients often have reduced reserve of stroke volume, heart rate, and cardiac output (CO), and the increase in CO relative to oxygen consumption is blunted.165 Heart failure with preserved ejection fraction patients typically have high LV filling pressures, whether at rest and/or on exercise, and they may develop fluid retention and an expanded plasma volume.28,159,164,166,167 All these mechanisms might be targets for treatment.In a meta-analysis, exercise capacity in HFpEF was related to chronotropic incompetence, high mPCWP, blunted augmentation of arteriovenous oxygen-content difference (implying inadequate perfusion of exercising skeletal muscles), reduced stroke volume reserve, and pulmonary hypertension.168 Changes in pulmonary artery pressure (PAP) on exercise are determined by the interplay between
CO, PA compliance, pulmonary vascular resistance, and mPCWP. The increase in PAP is flow-dependent so it is best reported in relation to the increase in CO; the upper limit of normal is þ3 mmHg/L/min.169 There are haemodynamic differences between patients with pre- and post-capillary pulmonary hypertension.164We recommend that the pathophysiological phenotype(s) prevailing in an individual HFpEF patient are determined, as that may allow the selection of specific therapies (see diagnostic Step 4 below).
The new Heart Failure Association diagnostic recommendations
The flowchart (Figure 2) provides an overview of the new diagnostic algorithm.
Step 1(P): Pre-test assessment
Step 1(P) should be performed in any patient who presents with symptoms and/or signs compatible with a diagnosis of HF. It requires a detailed clinical and demographic history; an electrocardiogram (ECG); blood tests; standard echocardiography to exclude other causes such as HFrEF or heart valve disease; and investigations for ischaemia, arrhythmias, anaemia, or pulmonary disease (Figure 2). NP levels can be obtained if the assay is available; elevated levels suggest heart disease but normal levels do not exclude HFpEF. Step 1(P) mirrors the 2016 ESC HF guidelines concerning initial HF diagnostic workup.3
Symptoms and signs
Breathlessness on exertion (New York Heart Association Class II or III) is highly sensitive for a diagnosis of HF but only moderately specific (about 50%) for a cardiac cause.170 Orthopnoea is quite specific but relatively insensitive. Patients with HFpEF often report reduced exercise capacity and fatigue, out of proportion to cardiac abnormalities at rest. In elderly, overweight and deconditioned persons, poor exercise capacity, dyspnoea on exertion, and peripheral oedema may also have a non-cardiac origin.
Electrocardiographic abnormalities
Patients may have electrocardiographic features of LVH (such as a Sokolov-Lyon Index ≥3.5 mV; abnormal repolarisation) and/or LA enlargement, but there are no pathognomonic signs and the diagnostic value of an ECG to identify HFpEF is poor.5 The most important indication is to detect atrial fibrillation (AF), which is highly predictive of underlying HFpEF.5 , 148
Laboratory tests
Several tests are recommended, including: sodium, potassium, urea, and creatinine (with an estimated glomerular filtration rate); liver function tests; HbA1c (metabolic syndrome and type 2 diabetes are common comorbidities); thyroid stimulating hormone; and full blood count, ferritin, transferrin saturation, and for anaemia. Anaemia associated with HFpEF aggravates symptoms and exercise intolerance.171 , 172
Natriuretic peptides
Multiple studies in primary care have shown that serum levels <125 pg/mL (or ng/L) for N-terminal pro-brain natriuretic peptide (NT-proBNP) or <35 pg/mL for BNP, have high negative predictive values (NPV; 95–99%) for excluding any heart failure.39 , 40 , 121 , 173–177 The main trigger for release of NPs is high LV end-diastolic wall stress, which is inversely proportional to wall thickness. It is therefore understandable that the excellent NPV of NPs is true particularly for HFrEF with a dilated LV, but not necessarily for HFpEF where LVH tends to normalize wall stress. In consequence, it has become clear that up to 20% of patients with invasively proven HFpEF have NPs below these diagnostic thresholds,28 , 178–180 which represents a limitation to the use of NPs. Therefore, it is important to understand that with our SCORE approach HFpEF can still be diagnosed even, if NP cut-offs (stratified by SR vs. AF) are below the given thresholds’.28 , 178–180
Echocardiography
Standard echocardiography should be performed in every breathless patient in whom there is clinical suspicion of HF, unless all the factors listed in Table 1 are absent or negative. Echocardiography may exclude alternative causes of dyspnoea such as HFrEF, valve disease, primary pulmonary hypertension, or pericardial effusion.181 , 182
Left ventricular ejection fraction should be measured, not estimated, ideally from biplane or three-dimensional images. Only small variations in normal ranges for EF by age, gender, and ethnic group have been reported, so it is recommended that a single cut-point of ≥50% is applied to define a ‘preserved’ EF. Left ventricular diameters and volumes should also be recorded. A diagnosis of HFpEF is suggested if there is a non-dilated LV with a normal EF, concentric remodelling or LVH, and left atrial enlargement. Echocardiographic findings at rest compatible with this HFpEF phenotype are often found in asymptomatic patients, who are at risk of progressing to overt HFpEF.183 , 184 Of note, the presence of structural alterations on echocardiography supports, but its absence does not exclude HFpEF. A more detailed or advanced echocardiographic study (see Step 2(E); Supplementary material online, S1) is not necessary at this step, but if it can be performed then only one examination will be needed.
Exercise tests
Coexisting epicardial stenotic coronary artery disease in patients with HFpEF impacts on mortality and should be detected and treated.133 Coronary microvascular dysfunction is part of the HFpEF pathophysiology134 so non-invasive stress testing can give false-positive results.134 , 186 Nonetheless, a bicycle or treadmill exercise test, or tests with higher sensitivity to detect ischaemia such as dobutamine stress echocardiography, cardiac magnetic resonance (CMR) imaging, or myocardial scintigraphy, or an anatomical approach using coronary computed tomography (CT) angiography or invasive angiography, should be considered if CAD is suspected.187
A stress test provides information about exercise capacity, the blood pressure response to exercise (which may be hypertensive), and the heart rate response. Chronotropic incompetence is present in 33–77% of HFpEF patients,188 , 189 and defined as the failure to reach 70–80%188–190 of the predicted maximal heart rate. Reduced heart rate recovery after exercise has prognostic value.191–193 Reduced exercise capacity can be defined as a peak workload ≤75% of the value predicted for age. In elderly patients with suspected HFpEF a 6-minute walk test (6MWT) distance ≤300 m can be considered abnormal193 but 6MWT performance is affected by non-cardiac as well as cardiopulmonary conditions.193 , 194
In selected cases, advanced cardiopulmonary exercise testing (CPET) with spiro-ergometry may be performed. Reduced exercise capacity is defined as a peak oxygen consumption (VO2 max) ≤20 mL/kg/min, and ventilatory inefficiency as a VE/VCO2 slope ≥30.166 , 195 Cardiopulmonary exercise testing provides objective evidence of exercise capacity and may differentiate between cardiac and non-cardiac causes (pulmonary, peripheral) for dyspnoea,157 , 166 , 196–200 but its value to distinguish between HFpEF and non-cardiac causes may be limited.166 Cardiopulmonary exercise testing is not a typical element in the initial HFpEF workup (see below).
If HFpEF is suspected after Step 1(P), a more specific assessment may confirm or exclude the diagnosis (Step 2(E)).
Step 2(E): Echocardiographic and natriuretic peptide heart failure with preserved ejection fraction diagnostic score
There is no single non-invasive diagnostic criterion for HFpEF so we recommend a combination of echocardiographic measurements of cardiac structure and function, and NP levels. Some may already be available from Step 1(P).
Many of these measurements are continuously distributed within a population, from normal to possibly abnormal and to overtly abnormal values. Diagnostic cut-points may vary according to age, gender, body weight, renal function, and the presence of atrial fibrillation. To take account of these factors, we recommend the use of major and minor diagnostic criteria according to the severity of an abnormality and the presence of modifiers. Major criteria (and cut-points) have been selected for their high specificity, while minor criteria should be more sensitive. Cut-points were derived particularly from studies that compared echocardiographic parameters against invasive haemodynamic data.5 , 28 , 166
In one cohort with 64% prevalence of HFpEF determined by invasive measurements, the univariable sensitivity of septal e′ velocity <7 cm/s to diagnose HFpEF, without adjusting for age or other variables, was 46%, while its specificity was 76%.5 The sensitivity and specificity of an E/e′ ratio >9 were 78% and 59%, compared with 46% and 86% for E/e′ >13. The sensitivity and specificity of LA volume index >30 mL/m2 were about 70%. Measurements of LV mass had low sensitivity (26%) for HFpEF but high specificity (86%) if LVH was present. PAP >35 mmHg [derived from tricuspid regurgitation (TR) velocity] was 46% sensitive and 86% specific for HFpEF,5 which makes it an important diagnostic criterion. The utility of GLS* <16% was moderate (sensitivity 62% and specificity 56%5).201
*GLS (Global Longitudinal Strain): Beginner’s Guide to Strain:
What should be in your lab in 2018 [PDF].
The utility of NP levels varies according to several factors including cardiac rhythm. For NT-proBNP >275 pg/mL, a sensitivity of 59% and a specificity of 77% were reported (accuracy 68%).5 Sensitivity decreased to 46% while specificity increased to 85% if the cut-off was increased to >450 pg/mL (accuracy 66%). At our lowest recommended cut-off of 125 pg/mL (minor criterion, if the patient is in sinus rhythm), the sensitivity reported in that study was 77% and the specificity 53% (accuracy 65%). Of note, 39% of patients in that study were in AF or had a history of paroxysmal AF.5 Combining the results of E/e′ and NT-proBNP can increase their predictive value, notably their sensitivity to diagnose HFpEF.202
Echocardiographic measurements of function and morphology
In Step 1(P) we recommend standard echocardiography, at least to assess LVEF and LV diameter. In Step 2(E) we recommend more detailed echocardiographic measurements (Supplementary material online, S1). These could all be obtained during a single study. The echocardiographic criteria in the HFA–PEFF score, listed below, mirror consensus recommendations for the diagnosis of LV diastolic function.41
Septal and lateral mitral annular peak early diastolic velocity (e′)
Major criterion: septal e’ <7 cm=s; or lateral e’ <10 cm/s subjects aged <75 years
Major criterion: septal e’ <5 cm=s; or lateral e’ <7 cm/s subjects aged >_75 years
The main determinant of e’ , the early diastolic velocity of mitral annular motion, is LV relaxation. It reflects LV lengthening and is influenced by preload.203,204 Left ventricular longitudinal e0 velocity declines with age;205 normative ranges reported from elderly participants were found to be lower than those given in the 2007 HFA consensus.206 We include age-specific e0 criteria in the HFA–PEFF score, measured as recommended.41
Average septal-lateral E/e’ ratio
Major criterion: average septal–lateral E=e’ ratio >_15
Minor criterion: average septal–lateral E=e’ ratio 9 -14
The ratio of the peak velocity of mitral inflow during early diastole (E), recorded by pulsed Doppler between the tips of the mitral leaflets, over the average of septal and lateral mitral annular early diastolic peak velocities (e’) recorded by pulsed tissue Doppler, reflects the mPCWP.41 The mitral E/e’ index correlates with LV stiffness and fibrosis 20,21 and is less age-dependent than e’.206 It also has diagnostic value during exercise.28,158 The E/e’ index is little influenced by
changes in volume but it is influenced by the severity of LVH.23,24.Tricuspid regurgitation peak velocity or
pulmonary arterial systolic pressureMajor criterion: TR peak velocity >2:8 m=s
Major criterion: Pulmonary artery systolic pressure >35 mmHg
Pulmonary arterial systolic pressure is calculated from the modified Bernoulli equation as 4 x peak TR velocity plus estimated right atrial pressure. Elevated PASP and reduced RV function are important predictors of mortality in HFpEF.207–211 Even a moderate increase in PASP can lead to increased ventricular interaction since a leftward shift of the ventricular septum impedes LV filling.212 A PASP >35 mmHg discriminates HFpEF from hypertensives and controls.207 A TR peak velocity >2.8 m/s indicates increased PASP41,213 and is an indirect marker of LV diastolic dysfunction.41
Left ventricular global longitudinal
systolic strainMinor criterion: GLS < 16%
Left ventricular peak systolic GLS is not angle-dependent, unlike
myocardial velocities recorded by tissue Doppler.186 It is measured using speckle-tracking echocardiography as the average of systolic strain obtained from all LV segments in the apical 4-chamber, apical 2-chamber, and apical long-axis views.214Reduced LV longitudinal systolic strain and LV early diastolic strain rate have both been identified in HFpEF.19,215,216 Impaired GLS predicts HF hospitalization, cardiovascular death, or cardiac arrest.216,217 It correlates with invasive measurements of LV stiffness and with NP levels.19,204,218 All strain values are dimensionless and are expressed as percentages. For ease of use in these recommendations, we suggest a cut-point of 16% in absolute values;219–222 and a value below 16% (e.g. 14%) is recommended as a minor criterion.
Left atrial volume index
Major criterion: >34 mL/m² in sinus rhythm
Major criterion: >40 mL/m² in atrial fibrillation
Minor criterion: 29-34 mL/m² in sinus rhythm
Minor criterion: 34–40 mL/m² in atrial fibrillation
The maximal volume of the LA, measured at end-systole from bi- . plane or three-dimensional images and indexed to body surface area [left atrial volume index (LAVI)] is an indirect correlate of LV filling pressures.41 It is more accurate as a marker of chronic LA remodelling than either LA area or diameter223–225 and it correlates with other echocardiographic indices of LV diastolic function.226 A LAVI
of 29–34 mL/m2 is considered as a minor criterion since it represents the upper limit in healthy subjects.227,228In patients without AF or heart valve disease, LAVI >34 mL/m2 independently predicts death, heart failure, AF, and ischaemic
stroke.229–231 In patients with HFpEF and permanent AF, LAVI was 35% more enlarged than it was in HFpEF patients in SR.34 Patients with permanent AF may have a large LAVI even if they have no LV diastolic dysfunction.34,41 We therefore recommend separate cutoffs for LAVI in SR vs. AF.Left ventricular mass index and relative
wall thicknessMajor criterion:
LVMI ≥149 g/m² in men or ≥122 g/m²
in women and RWT >0:42Minor criterion:
LVMI ≥115 g/m² in men or ≥95 g/m² in women
or RWT >0:42 or LV end-diastolic wall thickness ≥12 mmIncreased LV diastolic wall thickness in a non-dilated heart implies that the patient has LVH. It develops first in the basal segments of the ventricular septum,232 and a wall thickness ≥12 mm at that site is common in elderly people. Localized septal hypertrophy may be a consequence of abnormal ventricular–arterial coupling but it is not sufficient to indicate that there is significant global LV remodelling or hypertrophy.
Left ventricular geometry is often classified using relative wall
thickness (RWT), calculated as twice the LV posterior wall thickness divided by the LV internal diameter at end-diastole (LVPW 2/LVIDD), and using left ventricular mass index (LVMI) normalized to body surface area or height.Four patterns are described:
normal (normal LVMI, RWT ≥0.42),
concentric remodelling (normal LVMI, RWT >0.42),
concentric hypertrophy (increased LVMI, RWT >0.42),
and eccentric hypertrophy (increased LVMI, RWT ≤0.42).41,233,234 In patients with HFpEF, both concentric LVH and concentric remodelling can be observed.235
The absence of LVH on echocardiography does not exclude
HFpEF.5 We therefore recommend the finding of concentric hypertrophy (increased LVMI and increased RWT) as a major criterion, or any one of a lesser degree of LVH, RWT, and LV end-diastolic wall thickness as a minor criterion.227,234,236Natriuretic peptides
Major criterion:
NT-proBNP >220 pg=mL;
or BNP >80 pg/mL in sinus rhythm
Major criterion:
NT-proBNP >660 pg/mL or
BNP >240 pg/mL in atrial fibrillationMinor criterion:
NT-proBNP 125–220 pg/mL; or
BNP 35–80 pg/mL in sinus rhythmMinor criterion:
NT-proBNP 375–660 pg/mL; or
BNP 105–240 pg/mL in atrial fibrillationIn Step 1(P), a single low cut-point was recommended in order to have a sensitive marker for cardiac abnormalities. In this step, in order to increase specificity, a higher cut-off value is recommended as a major criterion, in agreement with ESC guidelines.3 Cut-offs are also stratified for the presence of SR or AF.
Natriuretic peptide levels should always be interpreted in context.180 Definitive cut-offs to diagnose HFpEF in patients with SR or in AF are not well established, and trials have used different values.237 , 238 In the setting of screening, average NPs have been reported to be 3–3.5 fold higher in patients with AF than in patients in SR.239 Average NPs were found to be threefold higher in patients with AF than in patients in SR.34 , 35 , 240 In prevalent symptomatic HFpEF with AF, levels tend to be even higher.241 For diagnosing HFpEF, we hence recommend values in patients with AF that are three times higher than used for patients in SR.
Calculating and interpreting the HFA–PEFF score
The score has functional, morphological, and biomarker domains. Within each domain, a major criterion scores 2 points or a minor criterion 1 point (Figure 3; Supplementary material online, Table S1). Each domain can contribute maximally 2 points, if any major criterion from this domain is positive, or 1 point if no major but any minor criterion is positive. If several major criteria within a single domain are positive, this domain still contributes 2 points; and if no major but several minor criteria are positive the contribution still is 1 point. Major and minor criteria are not additive in a single domain. Points are added only when they come from different domains.
For example, 2 major (E/e′ >15, and TR >2.8 m/s) and 1 minor (GLS <16) criteria, all in the functional domain, will lead to a total score from that domain of 2 points. The total score would be 5, if at least one minor criterion (LAVI <34 mL/m2; LV wall thickness >12 mm) and one major criterion (BNP in SR >80 pg/mL) would be present coming from the morphological and biomarker domains, respectively. It is important to understand that not all parameters from each domain need to be recordable (which is typically the case). The HFA-PEFF score can be calculated even if not all parameters are obtained, which adds to the practical utility of the score.
A total score ≥5 points is considered to be diagnostic of HFpEF, while a score of ≤1 point is considered to make a diagnosis of HFpEF very unlikely and to mandate investigations for alternative causes. Patients with an intermediate score (2–4 points, Figures 2 and 3) need further evaluation (Step 3(F1); Figures 4 A,B).
fig7 when server ready [fig 4 A,B]
If LAVI, LVMI, or wall thickness cannot be assessed by echocardiography, we recommend using measurements obtained from CMR imaging instead. Of note, there are some systematic differences in measurements of LV volumes and LVEF between imaging modalities.242 In one comparative study, LV volumes were larger and LVEF was lower but not statistically different with CMR compared with other imaging modalities.243Step 3 (F1): Functional testing
Symptoms compatible with HF can be confirmed to originate from the heart if haemodynamic abnormalities such as reduced stroke volume, reduced CO, and elevated LV filling pressures are detected either at rest or during exercise. In a typical elderly patient with multiple comorbidities, the presence or absence of isolated cardiac structural and/or functional abnormalities at rest does not always establish or exclude the diagnosis of HFpEF. If invasive testing demonstrates a high LV filling pressure [left ventricular end-diastolic pressure (LVEDP) ≥16 mmHg, PCWP ≥15 mmHg] at rest, then the diagnosis may be confirmed; otherwise, assessment during exercise is recommended, either by non-invasive exercise stress echocardiography or by invasive haemodynamics (Figures 2 and 4 A,B).
Exercise stress echocardiography: the diastolic stress test
During exercise in healthy people, enhanced LV untwisting and early diastolic suction maintain or increase stroke volume despite shortening of the filling time and without increasing LV filling pressures. In patients with HFpEF, impaired early diastolic relaxation, reduced increments in suction, and poor LV compliance lead to inadequate increases in stroke volume and CO on exercise, increased LV filling pressures, and increased PASP.28 , 41 , 244–248 High LV filling pressures and inadequate CO responses during exercise can also impair RV reserve.52.
Many patients with HFpEF have symptoms mainly on exertion that are usually attributed to the increase in LV filling pressures which is needed to maintain adequate filling and stroke volume.159 , 249 Acquiring echocardiographic data during exercise can unmask LV diastolic and systolic dysfunction. The parameters that have been studied most often, during or immediately after exercise, are the mitral E/e′ ratio and the TR peak velocity, which indicate increases in mPCWP and PASP, respectively.28 , 41 , 244–248 , 250
Ideally a semi-supine bicycle test with imaging during exercise, or else a treadmill or upright bicycle exercise protocol with imaging at or immediately after peak stress, is recommended41 , 244 but there are no universally adopted protocols. The European Association of Cardiovascular Imaging and the American Society of Echocardiography recommend a stepped protocol, starting at 25 W at 60 r.p.m. with the load increasing by 25 W every 3 min until the patient has reached his maximal predicted workload and/or maximal predicted heart rate (220—age in years) and/or developed limiting symptoms.244 Some patients cannot perform that protocol, and a ramped exercise test on a semi-supine bicycle at 60 r.p.m. starting at 15 W and with increments of 5 W every minute has also been proposed, to a submaximal target heart rate of 100–110/min or until the patient develops limiting symptoms.245 None of these protocols have been shown to be superior to others.
The mitral E/e′ ratio and peak TR velocity should be acquired at baseline, during each stage including peak exercise, and during a submaximal stage before fusion of the mitral E and A velocities213 or during the first 2 min of the recovery phase when mitral E and A velocities are no longer fused and LV filling pressures remain elevated.41 , 244 Changes in CO can be assessed by measuring the velocity integral of flow in the LV outflow tract, multiplied by the HR.
Exercise echocardiography should be considered abnormal if average E/e′ ratio at peak stress increases to ≥15, with or without a peak TR velocity >3.4 m/s.28 , 41 , 244 An increase only in TR velocity should not be used to diagnose HFpEF because it might be caused simply by a normal hyperdynamic response to exercise (with increased pulmonary blood flow) in the absence of LV diastolic dysfunction.251
An average E/e′ ratio during exercise ≥15 adds 2 points to the HFA–PEFF score. An average E/e′ ratio ≥15 with a peak TR velocity >3.4 m/s adds 3 points to the previous score from Step 2(E). If the combined score from Step 2(E) and Step 3(F1) is ≥5 points, then the diagnosis of HFpEF can be confirmed.
However, echocardiographic stress tests also have limitations. It was reported that E/e′ was not measurable in about 10% of subjects during submaximal exercise (20 W) and in about 20% of HFpEF patients during peak exercise, and that TR velocity was measurable in only 50%; about 20% of controls were considered to have false-positive tests.28 Data from stress echocardiography are not sufficient to substitute for invasive haemodynamic data under all circumstances. If the score remains <5 points or if exercise echocardiography cannot be performed, we recommend an invasive haemodynamic stress test in any case of doubt, especially if a therapeutic decision depends on the results.
Invasive haemodynamic tests at rest and with exercise
Left ventricular end-diastolic pressure LVEDP in the resting supine position is typically obtained in the context of left heart catheterization and bears important diagnostic information in the workup of unexplained dyspnoea. In selected patients, LV compliance and stiffness can be determined directly by using a multiple-loop conductance catheter to record the end-diastolic pressure–volume relationship (EDPVR) during preload reduction, giving a volume-independent parameter for LV stiffness (constant of chamber stiffness, b, normal <0.27252).21 , 159 , 253–255 Invasive demonstration of impaired LV relaxation at rest, measured by high-fidelity pressure catheters as the time constant of LV relaxation (tau, τ > 48 ms4) or of elevated LV filling pressures at rest (LVEDP ≥16 mmHg) confirms definite evidence of HFpEF.
Right heart catheterization should be considered for the structured workup of suspected HFpEF, especially when left heart pressures are not available. When resting mPCWP, measured using a Swan-Ganz catheter, is elevated in the presence of a normal LV end-diastolic volume index, then usually LV end-diastolic distensibility is reduced. A resting mPCWP ≥15 mmHg3 confirms definite evidence of HFpEF.
However, normal LVEDP or mPCWP levels at rest do not exclude HFpEF. In compensated HFpEF, haemodynamic alterations may be detected only during exercise or when the patient deteriorates.28 , 179 , 230 , 256 , 257 Also, volume depletion or intensified diuretic treatment may shift the diastolic pressure–volume relationship to the left, without changing LV compliance (dV/dP; the inverse of LV stiffness252) If resting filling pressures are normal, exercise right heart catheterization is recommended for the definite workup of unexplained exertional dyspnoea,179 especially if the patient has an intermediate Score in Step 2(E) or if exercise echocardiography is inconclusive or not feasible (Figures 2 and 4 B). Specialized centres may perform exercise right heart catheterization upfront in the absence of exercise echocardiography, depending on the individual experience of the site.
During supine exercise in healthy control subjects, cut-offs for peak PCWP and LVEDP are <20–23 mmHg258 , 259 and <25 mmHg,260 , 261 respectively. Patients with values <25 mmHg during peak exercise are classified as having non-cardiac dyspnoea. A steep increase in PCWP during exercise is a typical haemodynamic response in HFpEF,256 , 262 indicating that the dyspnoea on exertion is mainly of cardiac origin. Patients with peak exercise PCWP ≥25 mmHg are classified as having HFpEF (Supplementary material online, S2). An increase in LV filling pressure during exercise that is not accompanied by increases in end-diastolic volume, indicates limitation to LV filling or the development of pericardial constraint.256
A high resting mPCWP and a pathological increase in mPCWP during exercise predict poor outcomes from HFpEF.168 , 249 , 263 Patients with a normal mPCWP at rest (<12 mmHg) but a steep increase during exercise (to ≥25 mmHg) have a two-fold increase in mortality.263 Ten-year mortality was 6.6% if resting mPCWP was ≤12 mmHg and peak exercise mPCWP was <25 mmHg; 28.2% in patients with low mPCWP at rest and high exercise mPCWP; and 35.2% in those with high resting mPCWP and high peak exercise mPCWP (≥25 mmHg).263
Exercise mPCWP reclassifies patients with a normal resting mPCWP and stratifies risk. If other investigations have been inconclusive, invasive measurement of mPCWP or LVEDP is considered as the clinical reference investigation for diagnosing HFpEF28 (see Supplementary material online, S7 about how to perform an invasive stress test). Other causes such as significant CAD, mitral stenosis, or pericardial constriction must be excluded.
Step 4(F2): Final aetiology
Most cases of HFpEF are related to common risk factors and comorbidities, but the possibility of a specific underlying aetiology should always be considered (Table 2, Supplementary material online, Tables S2–S4; Figure 5 A,B). We postulate that identification of specific HFpEF aetiologies will advance the field of targeted therapies.
Start here.
