Statistics from Altmetric.com
In developed countries, at least 38–54% of patients with heart failure show preserved left ventricular (LV) ejection fraction.1–3 The prevalence of heart failure with preserved ejection fraction (HFPEF) is steadily increasing and its prognosis is poor.1–3 LV diastolic dysfunction, either alone or in combination with other factors (figure 1), is the major underlying mechanism of HFPEF.3–5 In the general population, the presence of even mild clinical diastolic dysfunction has been associated with pronounced increases in all cause mortality.6 Hence, the diagnosis of clinical (prognostic) diastolic dysfunction leading to HFPEF is of critical importance.
The European Society of Cardiology guidelines based definition of HFPEF requires a presence of signs and symptoms of heart failure, preserved LV ejection fraction of a non-dilated left ventricle, and evidence of diastolic dysfunction (impaired LV relaxation or increased LV diastolic stiffness) with elevated LV filling pressures.3 LV diastolic dysfunction and filling pressures can be assessed non-invasively using Doppler echocardiography.7 This article will focus on the clinical use of Doppler echocardiography in the diagnosis of HFPEF.
Diagnosis of HFPEF is easy
In patients presenting with acute decompensated heart failure, the differential diagnosis of dyspnoea does not usually pose any significant problems. These patients have clear signs and symptoms of congestion, and elevated LV filling pressures and B-type natriuretic peptides (BNP, N terminal pro-BNP (NT-proBNP)). Echocardiography usually shows a pseudonormal or restrictive transmitral flow pattern and, in cases with diastolic heart failure, a reduced peak early diastolic velocity of the mitral annulus longitudinal motion (e′).
Discrimination between a normal and a pseudonormal transmitral flow pattern may sometimes represent a dilemma if the assessment of echocardiography findings is done in isolation without being directly involved with the patient's clinical management. Patients with the pseudonormal transmitral flow usually show clinical signs of heart failure. However, this is not the focus of this article.
The other ‘easy’ diagnostic group includes patients with inherited or acquired restrictive cardiomyopathy, and, less often, cases of hypertrophic cardiomyopathy. These patients have notably increased LV stiffness (a left and upward shift of the end-diastolic pressure–volume relationship) with a restrictive LV filling pattern (E/A ratio >2 along with a deceleration time <160 ms) and a low e′. Such patients, however, represent only 3% of subjects with HFPEF.
Diagnosis of HFPEF is difficult
The differential diagnosis of HFPEF is challenging in stable elderly patients with equivocal symptoms. In contrast to younger, more active, individuals, symptomatology in the elderly is atypical or limited, partly also due to impaired mobility. These patients may report chronic moderate dyspnoea (New York Heart Association (NYHA) functional class II–III), fatigue or occasional palpitation, or they may simply report that they do not feel normal or that something is wrong, but without a clear articulation of symptoms. These complaints are associated with a variety of diseases but also with heart failure. In fact, the majority of patients with HFPEF are elderly individuals (mean age 75 years) clearly fulfilling the above mentioned characteristics.3 ,6 ,8 Since they do not show signs of congestion the European Society of Cardiology definition of HFPEF cannot be used. Furthermore, these patients show subtle abnormalities in multiple indices of systolic and diastolic function, but without a single leading pathology.9 LV stiffness is only mildly to moderately increased. The resting LV filling pressures and NT-proBNP are normal or slightly elevated. Echocardiography shows an impaired relaxation transmitral flow pattern which is a non-specific finding in the elderly (figure 2). Finally, the high prevalence of non-cardiac comorbidities in this age group makes the differential diagnosis challenging (figure 1).1 ,2 ,10 In clinical practice, many causes of moderate dyspnoea, without the presence of a significant pathology on routine examinations, remain unexplained or are attributed simply to deconditioning or being mildly overweight. Other patients, without a defined diagnosis, are administered thiazide-type diuretics. The complexity of this problem is illustrated in the following cases.
A 72-year-old woman was referred for chronic effort dyspnoea (NYHA II–III). Her history was remarkable for mild hypertension with normal blood pressure while on medication. Physical examination revealed bilateral perimalleolar oedema and normal jugular filling without hepatojugular reflux. Blood analysis showed mild anaemia (haemoglobin (Hb) 10.2 g/dl) and normal NT-proBNP (88 pg/ml). She was in sinus rhythm without reversible myocardial ischaemia during MIBI scintigraphy. Echocardiography showed borderline wall thickness with mildly increased LV mass (101 g/m2), normal LV ejection fraction of a non-dilated left ventricle, mildly enlarged left atrium (volume index 29 ml/m2), and borderline systolic pulmonary artery pressure (sPAP) (35 mm Hg). An impaired relaxation transmitral flow pattern was noted (figure 2). The question became whether the dyspnoea was of cardiac origin or, to put it differently, whether the patient had HFPEF.
In the differential diagnosis of HFPEF, the following systematic approach should be applied (figure 3). The first step is to assess LV diastolic function. Impaired LV relaxation is a marker of HFPEF regardless of the underlying mechanism. If LV relaxation is normal, HFPEF is unlikely. In the population of ambulatory patients, who are older than 65 years and have normal LV ejection fraction, approximately 50% has normal age appropriate diastolic function while the majority of the remaining patients show a low e′ suggesting reduced LV relaxation.6 ,11 If LV relaxation is reduced, the second step is to establish whether the degree of impairment has any clinical relevance for a patient. In other words, to evaluate whether the diastolic dysfunction leads to increased LV filling pressures during the individual's daily routines and thus to HFPEF. This may be very challenging because the low e′ may be associated either with a preclinical (asymptomatic, latent) diastolic dysfunction or with HFPEF (clinical diastolic dysfunction). The preclinical diastolic dysfunction is characterised by reduced LV relaxation at rest but preserved diastolic reserve to maintain normal LV filling pressures during the individual's daily activities. Preclinical diastolic dysfunction causes no symptoms or congestion, and it portends a good prognosis. The majority of these patients will never develop HFPEF.12 In contrast, the clinical diastolic dysfunction is associated with reduced diastolic reserve and it leads to increased LV filling pressures and thus to HFPEF with a poor prognosis.
LV relaxation can be assessed using the average e′ from the septal and lateral corner of the mitral annulus.7 The recording of e′ is performed in the apical four chamber view using a pulsed wave tissue Doppler. The e′ cut-off values for different age groups were described in a recent set of recommendations.7 This particular patient had, with regard to her age, preserved myocardial relaxation (average e′ ≥9 cm/s) and the left atrium was only mildly dilated. Combination of the preserved e′ with either a normal or mildly dilated left atrium (left atrium volume index <34 ml/m2) defines normal LV diastolic function.7 Hence, her symptoms were not likely to be of cardiac origin (figure 4). Some less frequent causes of HFPEF should still be considered —for example, constrictive pericarditis, chronotropic incompetence, dynamic mitral regurgitation (MR), severe dyssynchrony, and dyspnoea as an angina pectoris equivalent.
A 67-year-old man with mild dyspnoea underwent echocardiography which showed an impaired relaxation transmitral flow pattern and a low e′ suggesting impaired LV relaxation (figure 5). It is well known that increasing age is associated with decreasing myocardial relaxation. Does this case represent clinical diastolic dysfunction —that is, a disease with a poor prognosis?6 Or is this a case of a preclinical benign diastolic dysfunction? The distinction between HFPEF and the isolated echocardiography abnormality is critical.7 ,13
The clinical diastolic dysfunction is associated with permanent or intermittent increases in LV filling pressures leading to left atrial (LA) dilatation, atrial fibrillation, and elevation of sPAP. The ratio of transmitral flow E wave to e′ (E/e′) can be used to assess LV filling pressures. The E wave has been shown to have a positive correlation with preload and a negative correlation with LV relaxation. In contrast, e′ shows a stronger relationship with LV relaxation.14 ,15 Thus, in patients with impaired myocardial relaxation and increased LV filling pressures (HFPEF, clinical diastolic dysfunction) the E wave increases and the e′ decreases. This leads to an elevation of the E/e′ ratio. Several studies have demonstrated a significant correlation between E/e′ and pulmonary capillary wedge pressure or LV end-diastolic pressure.14 ,16 In contrast, the accuracy of the E/e′ ratio to predict elevated LV filling pressures has been recently questioned by other authors who showed only a weak relationship between the e′ and LV relaxation (τ).17–,19 Kasner et al20 have performed the largest invasive PV (pressure–volume) loop controlled study in humans with 43 HFPEF patients and 12 matched controls. In this study, the lateral e′-based indices (e′ <0.08 m/s, E/e′ ≥8, e′/a′ <1) showed high sensitivity and specificity to identify subjects with prolonged Tau, increased LV end-diastolic pressure, and LV stiffness. Assessment of the E/e′ is easy and highly feasible in most patients. Therefore, the joint guidelines of the European Association of Echocardiography and the American Society of Echocardiography recommend determining the E/e′ ratio as part of the assessment of LV filling pressures.7 The e′ and the E/e′ ratio should not be used in the presence of an extensive mitral annulus calcification, mitral stenosis, moderate to severe MR, surgical mitral rings and prosthetic valves. In patients with left bundle branch block, the septal e′ should be avoided. In patients with regional wall motion abnormality, the e′ from the adjacent mitral annulus corner should not be taken. Given these limitations, it needs to be emphasised that the e′ should be interpreted in the clinical context and in relation with other Doppler echocardiography indices. The detailed description of all the exceptions can be found in the guidelines.7
High E/e′ ratio ≥13 (using average e′ from the septal and lateral mitral annulus) suggests increased LV filling pressures (LV end-diastolic pressure >16 mm Hg, mean capillary wedge pressure >12 mm Hg), while low E/e′ ≤8 usually indicates normal LV filling pressures. In patients with atrial fibrillation, the different cut off value (septal E/e′ ≥11) has been proposed to confirm increased LV filling pressures.7 The majority of patients with HFPEF have an E/e′ in a non-diagnostic range (between 8–13).16 This suggests that the E/e′ has reduced sensitivity, at rest, to confirm HFPEF. In such patients with a non-diagnostic E/e′ ratio, a multi-parametric strategy to assess LV filling pressure needs to be applied (figure 3). Prolonged isovolumic relaxation time (>100 ms) or very low E/A (<0.5) together with a long deceleration time (>240 ms) usually suggest normal LV filling pressures, except in cases of hypertrophic cardiomyopathy. In contrast, short isovolumic relaxation time (<60 ms) or high E/A (>2) along with a short deceleration time (<160 ms) and a low e′ are associated with increased LV filling pressures. LA volume reflects the cumulative burden and chronicity of elevated LV filling pressures. A biplane LA volume index ≥34 ml/m2 is a strong predictor of an adverse outcome; hence, the use of this cut off has been recommended to distinguish between preclinical (isolated echocardiography abnormality) and prognostic (HFPEF) diastolic dysfunction.7 ,21 In a study by Lam, the presence of pulmonary hypertension was shown to be a highly accurate indicator for distinguishing between HFPEF and controls.22 Elevated LA pressure is usually associated with some degree of pulmonary hypertension. Not surprisingly, elevated LA pressure is the most common cause of pulmonary hypertension in adults. Assessment of NT-proBNP can also be useful, although most often in cases involving acute decompensation or in HFPEF cases with increased LV stiffness.3
In typical clinical settings, many cases of ambulatory dyspnoea are challenging because of borderline and controversial examination results despite employing the multi-parametric approach recommended in the current guidelines.7 The resting E/e′ ratio and NT-proBNP are often in a non-diagnostic range. LA dilatation and pulmonary hypertension can be caused by conditions other than diastolic dysfunction (eg, arrhythmias, valvular diseases, lung disease, etc). Moreover, a significant proportion of patients with HFPEF have normal LV filling pressures at rest, with an increase only occurring during physical activity.23 ,24 Therefore, diastolic stress echocardiography should be attempted whenever possible.23 ,24 The patients with HFPEF have reduced diastolic reserve characterised by a failure to increase the rate of myocardial relaxation during exercise. Delayed myocardial relaxation cannot assure low pressure LV filling during increasing heart rates when diastolic filling time becomes shorter. This is illustrated in figures 6 and 7, which show typical findings during diastolic stress echocardiography in patients with normal and reduced diastolic reserves, respectively. It is noteworthy that these two patients could not be correctly classified as having preclinical diastolic dysfunction or HFPEF using the recommended multi-parametric approach. In subjects with a normal relaxation reserve, exercise leads to an increase in both the transmitral flow E wave and the e′, and thus, the E/e′ ratio remains low. In contrast, in patients with an impaired LV relaxation reserve, a significant increase in the E wave is accompanied by only minimal augmentation of the e′, therefore the E/e′ ratio increases into the diagnostic range (table 1).
The type and level of exercise used in stress echocardiography should be adjusted according to the patient's daily routine and comorbidities. In younger, more active individuals, the preferred option is a tilting bicycle allowing continuous echocardiography monitoring. Indices of transmitral flow, e′ and the tricuspid regurgitation jet are recorded at rest and during submaximal exercise before the waves fuse (typically at heart rates between 100–120 beats/min). In older sedentary patients with limiting comorbidities or orthopaedic problems, the performance of 10 sit-ups may represent a similar level of physical exercise as cycling for younger patients. Sit-ups are performed on the echocardiography table, so the recording of the transmitral flow and e′ can be done immediately after exercise. Some patients are unable to perform any sort of submaximal exercise. In these patients, a multi-parametric strategy should be employed to demonstrate structural or haemodynamic changes associated with periodic elevation of LV filling pressures (see below).
Table 2 shows the variables associated with normal and abnormal diastolic function. Patients with unequivocally normal diastolic function are characterised by preserved e′, a normal sized left atrium, and normal haemodynamics. In contrast, patients with clear clinical diastolic dysfunction have a reduced e′, a dilated left atrium and increased LV filling pressures. These represent patients with a poor prognosis. In clinical practice, the separation of patients with a normal versus an abnormal diastolic function (case 1) does not usually pose a problem. In contrast, in ambulatory patients presenting with dyspnoea and reduced e′, it is challenging to distinguish those with HFPEF from those with the isolated echocardiography abnormality. These patients often fall into the intermediate group (table 2) with reduced resting myocardial relaxation (e′) but without LV hypertrophy, LA dilatation or haemodynamic impairment (case 2). The majority of such patients will have normal diastolic reserve and no HFPEF; however, in a minority of them, the isolated diastolic dysfunction will deteriorate to HFPEF within 10 years.12 It is worth noting that in the patient in case 2, there was a significant increase in e′ during submaximal bicycle exercise, suggesting preserved diastolic reserve (figure 6). So it seems that, in this case, the reduced myocardial relaxation at rest has no structural or haemodynamic impact on the patient. This is a case of the preclinical or isolated echocardiography abnormality.
Figure 8 shows two elderly patients with chronic dyspnoea. It is evident that both patients have reduced myocardial relaxation but only one of them suffers from HFPEF. Of interest, despite a notably different prognosis, both patients are classified as stage I (mild, impaired relaxation pattern) diastolic dysfunction according to the current guidelines.7 The abnormality leading to HFPEF with increased mortality cannot be graded as mild. This example points out limitations of the currently used indices to assess LV diastolic function. This example also clearly demonstrates that e′ should not be used as the sole parameter to draw conclusions about LV diastolic function. In contrast, LV relaxation (e′) should be interpreted in the context of LV and LA morphology, function, haemodynamics, and response to exercise.
Practical approach to diagnosing HFPEF
Figure 3 shows an integrated approach to confirm the diagnosis of HFPEF in stable elderly patients with preserved LV ejection fraction and chronic moderate dyspnoea. In the minority of these individuals, the resting E/e′ ratio will be elevated (≥13) and the resting e′ will be reduced (≤9 cm/s) to the diagnostic range, which suggests an increase in LV filling pressures in the setting of diastolic LV dysfunction, and thus clear HFPEF. Furthermore, these patients will likely show some degree of congestion during the physical examination accompanied by a pseudo-normal transmitral flow pattern. However, the vast majority of stable outpatients with dyspnoea will present with an impaired relaxation transmitral flow pattern, reduced e′ (<9 cm/s), but a low E/e′ ratio (<13), and thus they fall into the border zone using the recommended multi-parametric strategy. These individuals could still have HFPEF that presents with either normal (E/e′≤8) or borderline (E/e′ 9–12) LV filling pressures at rest. In such patients exercise echocardiography should be attempted whenever possible. The type and level of exercise should be adjusted according to the patient's daily routine and comorbidities. One of the advantages of using exercise is that the patient serves as his or her own control. This may reduce the confounding effects of various factors on the relationship between the e′ related indices and LV relaxation. In patients who are unable to exercise, a multi-parametric strategy should be employed (figure 3, table 2). Diagnostic accuracy improves with increasing numbers of positive markers.
The differential diagnosis of dyspnoea in elderly, stable outpatients with normal LV ejection fraction is difficult. The majority of these individuals have no signs of congestion during the physical examination, while Doppler echocardiography shows an impaired relaxation transmitral flow pattern, reduced e′ and a low E/e′ ratio in a non-diagnostic range. To distinguish between the isolated echocardiography abnormality and the clinical diastolic dysfunction, LV relaxation (e′) should be interpreted in the context of LV and LA morphology, and haemodynamics. In the majority of patients, a multi-parametric strategy including diastolic stress echocardiography needs to be used to establish a diagnosis of HFPEF. However, there is a clear clinical need for novel non-invasive approaches.
Diagnosis of heart failure with preserved ejection fraction using Doppler echocardiography: key points
The majority of stable patients with heart failure with preserved ejection fraction (HFPEF) have no signs of congestion during the physical examination, while Doppler echocardiography shows an impaired relaxation transmitral flow pattern, a reduced e′ and a low E/e′ ratio in a non-diagnostic range.
The presence of both impaired left ventricular (LV) relaxation and increased LV filling pressures is required to establish a diagnosis of HFPEF.
To distinguish between preclinical diastolic dysfunction—that is, a benign isolated echocardiography abnormality—and clinical diastolic dysfunction—that is, HFPEF with a poor prognosis—the e′ derived indices should be interpreted in the context of left ventricular–atrial morphology and function, haemodynamics, and response to exercise.
In the majority of patients, a multi-parametric strategy including diastolic stress echocardiography needs to be applied to provide evidence of increased LV filling pressures and thus to establish a diagnosis of HFPEF.
You can get CPD/CME credits for Education in Heart
Education in Heart articles are accredited by both the UK Royal College of Physicians (London) and the European Board for Accreditation in Cardiology—you need to answer the accompanying multiple choice questions (MCQs). To access the questions, click on BMJ Learning: Take this module on BMJ Learning from the content box at the top right and bottom left of the online article. For more information please go to: http://heart.bmj.com/misc/education.dtl
▸ RCP credits: Log your activity in your CPD diary online (http://www.rcplondon.ac.uk/members/CPDdiary/index.asp)—pass mark is 80%.
▸ EBAC credits: Print out and retain the BMJ Learning certificate once you have completed the MCQs—pass mark is 60%. EBAC/ EACCME Credits can now be converted to AMA PRA Category 1 CME Credits and are recognised by all National Accreditation Authorities in Europe (http://www.ebac-cme.org/newsite/?hit=men02).
Please note: The MCQs are hosted on BMJ Learning—the best available learning website for medical professionals from the BMJ Group. If prompted, subscribers must sign into Heart with their journal's username and password. All users must also complete a one-time registration on BMJ Learning and subsequently log in (with a BMJ Learning username and password) on every visit.
- ↵These recommendations summarise the criteria for the diagnosis of HFPEF.
- This was the largest invasive study proving the concept that increased LV stiffness contributes to increased LV end-diastolic pressures during handgrip exercise. LV diastolic function was assessed using invasive pressure–volume loop analysis at rest, during handgrip exercise, and during atrial pacing.
- ↵This landmark paper provided evidence that patients with heart failure and a normal ejection fraction have abnormalities in the diastolic properties of the left ventricle that are sufficient to explain the patients’ haemodynamic abnormalities and occurrence of heart failure.
- ↵This large survey provided the evidence that even mild clinical diastolic dysfunction was associated with pronounced increases in all cause mortality.
- ↵Joint guidelines of the European Association of Echocardiography and the American Society of Echocardiography for the evaluation of LV diastolic function.
- ↵This was the landmark paper introducing the E/e′ ratio as an index allowing estimation of LV filling pressures.
- ↵Review on LA volume as a predictor of cardiovascular outcome.
- ↵This large study involving 244 HFPEF patients and 719 controls demonstrated high prevalence of significant pulmonary hypertension in HFPEF. The presence of pulmonary hypertension discriminated with high accuracy between HFPEF and controls.
- ↵Validation of exercise E/e′ against invasive measurements.
- ↵This paper was the first to introduce the concept of diastolic stress echocardiography.
Contributors All authors contributed significantly to the submitted paper.
Competing interests In compliance with EBAC/EACCME guidelines, all authors participating in Education in Heart have disclosed potential conflicts of interest that might cause a bias in the article. The authors have no competing interests.
Provenance and peer review Commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.