Diastolic Heart Failure Mechanisms and Assessment Revisited

Abstract

The syndrome of heart failure (HF) with preserved ejection fraction (HFpEF) makes up about half of the HF population. The HF mechanisms in these patients are varied and not fully understood. In addition, the term “diastolic HF” was switched to HFpEF because of difficulties in measuring the left ventricular (LV) diastolic performance. In the late stages, HFpEF carries a prognosis that is as bad as or worse than that of HFrEF. Hence, it is important to recognize LV diastolic impairment at an earlier stage so that the causal mechanisms, if any, can be treated to retard its progression. Despite the availability of numerous disease-modifying agents for HFrEF, there are hardly any available treatments for HFpEF. With our aging population, there will be an epidemic of HFpEF and hence this entity needs attention and respect. In this paper, we review the fundamental mechanisms of HFpEF, the physiology of LV filling and how LV diastolic function can be comprehensively measured. We also speculate how this may help with the early recognition of diastolic HF and its treatment.

1. Introduction

The fact that heart failure (HF) can occur despite a normal left ventricular (LV) ejection fraction (EF) was first described in the early 1980s [1,2]. It was observed in these patients, based on radionuclide ventriculography showing LV time–volume curves, that early LV filling was impaired, with a greater contribution from atrial contraction, and the term “diastolic heart failure” was coined (Figure 1). Subsequently, this was changed to HF with preserved EF (HFpEF) because of the complexities of the measurement of LV diastolic function [3]. Diastolic HF was initially thought to be associated with a better prognosis compared to systolic HF based on Framingham data [4], but it was subsequently shown to be as sinister as or worse than systolic HF, especially in hospitalized class IV patients with HF (Figure 2) [5]. We would like to summarize the growth in the understanding of LV diastolic function (

Table 1. History of concept developments in terms of diastolic HF.

100 AC	Galen proposed that the heart filled by dilation of the right ventricle
1628	Harvey proposed that most ventricular filling occurred during atrial systole
1877	Francois-Franck proposed that most LV filling occurred in early diastole
1906	Henderson described three phases of diastole
1921	Wiggers and Katz reported that the atrial contribution to LV filling varied
1927	Meek proposed that the “contraction remainder” extended into diastole
1949	Wiggers coined the term “inherent elasticity” to describe the diastolic properties of the myocardium
1970s	Studies were performed on LV relaxation and passive LV properties, and these were characterized
1970s	Extensive studies of hypertrophic cardiomyopathy describing severely abnormal diastole and potential benefits of verapamil
1980s	Diastolic heart failure was described as an entity
1992	Monograph on “Diastolic heart failure” published by Shah and Pai
2000s	The term “Heart failure with preserved ejection fraction (HFpEF)” was coined because of uncertainties about evaluating LV diastolic function

).

However, it has to be recognized that clinical HF is a late stage of the disease process. It is important to recognize early abnormalities in LV diastolic function. Diastole is more vulnerable than systole to a variety of disease processes and insults. Hence, early recognition of LV diastolic dysfunction may provide earlier insights into myocardial processes that may eventually produce HF. This opens up opportunities to treat the disease in the early stages and potentially prevent progression to clinical HF and reduce the HF burden in the community. Some of the examples of this include hypertensive heart disease and infiltrative disorders such as cardiac sarcoidosis and amyloidosis. Many other pathophysiological bases of diastolic HF are poorly understood.

In this review, we will cover the spectrum of LV diastolic dysfunction (Figure 3), physiology of diastole, factors affecting LV diastolic function and measurement of LV diastolic function.

2. Physiology of Diastole

Diastole starts with aortic valve closure (A2 component of S2) and extends to the LV isovolumic contraction phase. It is made of four distinct phases (Figure 4).

• LV isovolumic relaxation from aortic valve closure to mitral valve opening
• Early phase of rapid passive filling
• Diastasis
• Atrial systole

There are multiple processes that control or affect different phases of diastole, and they are described below (2).

2.1. LV Relaxation Process

The early part of diastole is determined by the process of active LV relaxation, which is an energy-dependent process and is the reversal of the contractile process. It starts in late systole and extends into the early LV filling phase. In patients with hypertrophic cardiomyopathy, where relaxation is very abnormal, it can extend further—occasionally resulting in a fall in the LV diastolic pressure into mid-diastole despite LV filling [6]. It depends not only on the active relaxation of individual muscle fibers but also on the spatial and temporal synchrony of relaxation [7,8]. Asynchrony would impair the rate of LV pressure decay. Factors affecting the LV relaxation process are listed in

Table 2. Factors affecting LV relaxation.

Factors impairing LV relaxation:

Aging

LV hypertrophy

Myocardial ischemia

LV dyssynchrony; e.g., RV pacing, LV hypertrophy, LBBB

Myocardial disarray as in hypertrophic cardiomyopathy

Increased LV afterload causing increased myocyte calcium load

Factors that improve LV relaxation:

Higher sympathetic tone and catecholamines

Calcium channel blockers

Specific LV lusiotropic agents such as levosimendan and calcium channel blockers in hypertrophic cardiomyopathy

.

2.2. Early Diastolic Suction Due to LV Recoil

This sort of suction force is especially prominent in mitral stenosis, severe volume loss and hypercontractile ventricles such as hypertrophic cardiomyopathy. The suction force would facilitate early diastolic filling. This is one of the reasons why, in hyperdynamic LVs, even with LV hypertrophy, the early passive filling could be normal despite markedly impaired LV relaxation—a phenomenon frequently seen in hypertrophic cardiomyopathy. When the LV volume is very small because of excessive shortening of the muscle fibers, their springing back into the resting state generates recoil. The recoil of the LV twist may contribute to this phenomenon as well.

2.3. Left Ventricular Compliance

Compliance is the ratio of change in volume over pressure or the amount of increase in volume per unit increase in distending pressure in the passive stage. The reciprocal of this or the amount of increase in pressure per unit change in volume is stiffness. Compliance comes into play in the latter half of LV diastole [9,10]. Its measurement will be described later. Compliance can decrease with age, LV hypertrophy, fibrosis, infiltrative disorders, etc. (

Table 3. Processes associated with increased LV diastolic stiffness.

Increasing age

LV hypertrophy

Interstitial fibrosis

Myocardial scarring

Infiltrative disorders such as cardiac amyloid, sarcoid and other storage disorders

) [11].

2.4. Myocardial Stiffness

This is measured as the modulus of myocardial stiffness through the stress–strain relationship, as described later, and measures the passive elastic properties of muscle fibers corrected for geometry. This is a direct measure of the passive properties and is increased in muscle fiber hypertrophy, fibrosis and infiltration.

2.5. Erectile Effects of Coronary Filling

Because of the squeezing of intramyocardial coronaries during systole, epicardial coronaries store blood and energy during this phase, which is dissipated in diastole. This is thought to aid in early diastolic LV relaxation and filling.

2.6. Factors Affecting Pericardium or Intrapericardial Pressure

The normal intrapericardial pressure is the same as the atmospheric pressure. Hence, the transmural LV diastolic pressure, which is the true LV-filling pressure, is the same as the cavitary LV diastolic pressure. For example, if the LV end-diastolic pressure (LVEDP) is 15 mmHg, the transmural LV diastolic pressure is 15 mmHg. It is not so if the intrapericardial pressure is higher. For example, if the intrapericardial pressure is 15 mmHg, as may happen in cardiac tamponade, to produce the same transmural LV-filling pressure, the LVEDP has to be 30 mmHg to achieve equivalent LV filling. The intrapericardial pressure is elevated not only in cardiac tamponade but also in pericardial constriction, situations in which pericardial constraint is invoked and situations where the mediastinal pressure is elevated. The pericardial constraint is invoked when there is an acute increase in the volume of any of the cardiac chambers such that the unyielding pericardium would cause an increase in the intrapericardial pressure. Examples of this include acute valve leaks because of infection, trauma or infarction causing acute regurgitation of mitral/aortic/tricuspid valves, right ventricular infarction or acute massive pulmonary embolism where the right ventricle dilates because of dysfunction or acute afterload mismatch. This may also occur in acute pericarditis because of edema and cellular infiltrates in the pericardium making it stiffer. The conditions that increase the mediastinal pressure, which is transmitted to the pericardium, include large pleural effusions, massive tumors, tension pneumothorax and use of positive end-expiratory pressure (PEEP) ventilation. When PEEP is used, or if auto PEEP occurs because of expiratory obstruction, about 50% of the raised intrapleural pressure is transmitted to the pericardial space. Hence, it is important to remember to factor PEEP into the interpretation of hemodynamics obtained invasively in the intensive care unit. These conditions may mimic LV diastolic failure and must be kept in mind as differential diagnoses as its mimickers (

Table 4. Diastolic heart failure mimickers.

Pericardial constriction

Acute pericarditis

Acute aortic regurgitation

Acute mitral regurgitation

Acute tricuspid regurgitation

Right ventricular infarction

Acute massive or submassive pulmonary embolism

Large pleural effusions

Tension pneumothorax

Large mediastinal tumors

Use of positive end-expiratory pressure (PEEP) and auto PEEP

).

2.7. Atrial Systolic Function

Late diastolic LV filling is caused by an atrial booster pump. It accomplishes what early diastole may not do in terms of the optimal ventricular filling and preload. If early diastolic filling is impaired because of impaired LV relaxation, atrial systole compensates for that with enhanced booster function (e.g., older individuals) and vice versa. If early filling is efficient because of supernormal LV relaxation, resulting in exaggerated early LV filling (resulting in physiological S3) as occurs in young children, the atrial contribution will be minimal and such patients can tolerate very high heart rates without hemodynamic consequences as LV filling occurs rapidly and adequately in the first 100 ms of diastole. The relationship between LV early filling (mitral E wave) and the atrial contribution can be compared to that between a medical intern and a resident! If the intern is efficient, the resident has to do very little. If the intern is bad, the resident (atrial systole) has to work hard to compensate. If both are bad, there is trouble and the attending (elevated left atrial pressure) has to come into play!!! The atrial transport function and volume depend upon three things according to the Sarnoff model: atrial preload or state of the left atrial filling, atrial afterload or the pre-A left ventricular diastolic pressure and atrial mechanical function [12]. For example, if the LV diastolic pressure is high, the left atrial afterload increases, resulting in reduced left atrial ejection. Examples of reduced left atrial contractility include atrial myopathy and immediate post-cardioversion state.

2.8. Heart Rate, Diastolic Filling Rate and Diastolic Stress

An increase in the heart rate stresses systole and diastole differentially, placing an excessive burden on diastole, as shown in Figure 5. This is because of the disproportionate shortening of diastole compared to systole, as explained in the figure legend. This is one of the reasons why diastole is more sensitive than systole to a variety of myocardial insults and why diastolic dysfunction is an early manifestation of disease. This underscores the importance not only of recognizing HFpEF but also of being able to pick up on the early signs of deranged LV diastolic mechanisms. Diastolic stress may be able to uncover covert LV diastolic dysfunction.

2.9. Functional and Structural Basis of Diastolic Impairment

As we alluded to before, it is also important to identify the underlying mechanisms of LV diastolic dysfunction as it is a syndrome consisting of clinical HF with preserved EF. Many of the underlying mechanisms may be reversible. Examples include treating blood pressure to reduce the LV afterload to promote relaxation and reduce myocardial energy expenditure, relieving obstruction in hypertrophic obstructive cardiomyopathy for similar reasons, treatment of ischemia, addressing LV dyssynchrony if possible, looking for myocardial infiltrative disorders, especially cardiac amyloid, with Tc-PYP scans, cardiac MRI and serological testing (as amyloid is treatable), looking for cardiac sarcoid when there are other features, etc.

3. Assessment of Diastole

3.1. Invasive Measure

There are three fundamental invasive measures of LV diastolic function. These are (1) the measures of LV early diastolic relaxation measured as Tau, (2) the measure of LV late diastolic stiffness measured as the modulus of chamber stiffness (k), and (3) the modulus of myocardial stiffness. The first two affect LV cavitary behaviors and the last one is the passive property of the muscle fiber derived from the LV stress–strain relationship.

3.1.1. Measurement of LV Relaxation

This is fundamentally described by Tau, which mathematically describes the rate of LV pressure decay after aortic valve closure or following peak negative dP/dt. There are multiple ways researchers have fitted equations to this pressure–time curve, but the fit has to be perfect, with a correlation coefficient of >0.99. Figure 6 summarizes some of these methods. The peak negative dP/dt is another measure that is used as a measure of LV relaxation, but it is a one-point measure rather than covering the entire duration of pressure decay and is less desirable. Typically, the normal peak negative dP/dt is in the range of −1500 mmHg/s and Tau, depending on the method used, is 30–60 ms. The longer the Tau, the worse the LV early diastolic relaxation.

3.1.2. Measurement of Passive Elastic Properties of LV

As the LV fills passively, the pressure rises in a curvilinear fashion, with the rate of the pressure rise being higher at a higher level of filling. This behavior is best described by fitting an exponential equation, as shown on Figure 7. At higher rates of filling, the LV viscous properties may come into play as well—i.e., the pressure rise may be higher at the instant of rapid filling because of the slower relaxation of the muscle fibers, which is attributable to the viscous properties of the LV. This may come into play during atrial systole or whenever the LV fills very rapidly. In other words, the dynamic compliance may be lower than the static compliance.

3.1.3. Measurement of Myocardial Stiffness

The modulus of myocardial stiffness is obtained by using the stress–strain relationship. Stress is the deforming force per unit area and can be computed given the LV diastolic pressures, wall thickness, length and using geometric assumptions about LV cavity. Strain is the deformation produced by stress and is a passive property. With the advent of speckle-tracking echocardiography, it has become easier to compute both the regional and global LV strain, both in the long axis and in other directions. Because of the limited application and the complexity of its derivation, we will not discuss this any further.

Table 5. Commonly used terminology addressing LV diastolic function with the relevant definitions.

Force:	An agent that causes a change in momentum
Stress:	Force per unit cross-sectional area that causes deformation
Strain:	Fractional change from unstressed dimension
Creep:	Time-dependent elongation of a material held at constant stress
Stress relaxation:	Diminution of stress when held at constant strain
Elasticity:	Property of recovery of a material from a stressed state when the stress is removed
Viscosity:	Property of the material that retards deformation in response to stress
Viscoelasticity:	Property of the material when stress depends upon strain, e.g., myocardium
Hookean material:	Material that follows Hooke’s law, which states that the stress–strain relationship is linear
Non-Hookean material:	Material that does not follow Hooke’s law, e.g., myocardium
Modulus of stiffness:	Slope of the stress–strain relationship
Compliance:	Change in volume per unit change in pressure for a chamber
Stiffness:	Reciprocal of compliance

lists some commonly used terminology in this area with the meaning to help readers understand the literature better.

3.2. Echo-Doppler Techniques

In practice, echocardiography is the cornerstone of assessing LV diastolic function and filling pressures, and to some extent, gaining insights into the myocardial structure [13,14,15]. The tools available include the transmitral flow, pulmonary vein flow, Doppler tissue velocities of the mitral annulus, mitral E and A wave propagation inside the LV, velocity profiles or mitral and aortic regurgitation signals, left atrial size and dynamics, ultrasound elastography and myocardial backscatter analysis, etc.

3.2.1. Transmitral Flow

The transmitral flow is recorded at the tip of the mitral leaflets using a small sample size (2–3 mm) of pulsed wave Doppler to obtain a pure signal at that point, optimizing the filter and gain to see the profile clearly down to the baseline without noise. This will help obtain an accurate E wave deceleration time and A wave duration. This should not be obtained either distally or proximally, as the profile varies and the profile at the annulus is only obtained for volumetric purposes. A distal sample will provide a different profile, which is slightly delayed without augmentation from flow-dependent mitral leaflet opening. Figure 8A,B show an illustration of the genesis of mitral flow and various measures that can be obtained.

Table 6. Factors affecting LV IVRT, E wave amplitude and A wave amplitude.

Factors affecting LV IVRT:

Rate of LV relaxation increases with impaired relaxation

LBBB increases IVRT

RV pacing increases IVRT

High LA pressure reduces IVRT

Large left atrial V wave reduces IVRT

Factors affecting E wave amplitude:

Impaired LV relaxation reduces E wave amplitude

High LA pressure increases its amplitude

Exaggerated LV recoil increases its amplitude

Factors affecting mitral A wave amplitude:

Left atrial contractile function: reduced in atrial cardiomyopathy, post cardioversion

Atrial preload: larger LA volume at the time of systole boosts the atrial transport function

Atrial afterload: Higher LVEDP or pre-A pressure will reduce atrial contribution

lists various factors that can affect the LV isovolumic relaxation time, E wave amplitude and A wave amplitude. It is important to consider these factors and integrate them with other measures of diastolic function.

3.2.2. Pulmonary Vein Flow

On transthoracic echo, this is typically obtained from the apical view and from the right upper pulmonary vein. A 2–3 mm pulsed wave Doppler sample should be placed inside the vein, guided by color, to obtain a pure profile without noise from the tissues. During transesophageal echocardiography, it can be obtained from all four veins, and the profiles would vary and would depend on the body position as well. Figure 9 shows such a profile and describes its genesis.

Figure 10 illustrates how the mitral and pulmonary vein flows can be used to characterize the LV diastolic function and filling pressures. This, combined with the mitral annular velocities, provides powerful insights into the LV diastolic performance.

3.2.3. Doppler Tissue Imaging (DTI)

Traditionally, these velocity profiles are obtained from the medial and lateral mitral annulus using pulsed wave Doppler technology, setting the bandpass filter to capture only low-velocity, high-amplitude signals because the annular velocities are in about the 5–20 cm/s range and the signals are very intense, unlike the blood pool signals produced by moving red blood cells [16]. Figure 11 describes these signals and their genesis. In a normal heart, there is striking synchrony of these signals obtained from different quadrants of the mitral annulus and these are closely linked to both LV filling and ejection functions, as shown in Figure 12 and Figure 13. In myocardial disease, there may be dyssynchrony or incoordination among different myocardial walls, leading to the impairment of both systolic power generation and relaxation because of the incoordination in time and space. The annular e’ velocity is a good marker of LV relaxation and is reduced in impaired relaxation. The mitral E/annular e’ velocity ratio is a good reflection of the mean left atrial pressure, where a ratio ≥15 reflects elevated LA pressure.

Figure 14, Figure 15 and Figure 16 show some examples of how to combine these measurements and also cover some nuances concerning how to interpret these signals.

3.2.4. Left Atrial Volume and Dynamics

The left atrium can dilate because of atrial fibrillation, atrial myopathy or chronically elevated left atrial pressure [17,18]. In other words, an increased LA volume as a marker of chronic LA pressure elevation is comparable to “HbA1C” as a marker of hyperglycemia over a period of time. If the operative left atrial pressure is high, it operates at the steeper portion of the left atrial pressure–volume curve and fills only a little during systole when the mitral valve is closed and the LA receives blood from the pulmonary veins. The higher the LA pressure, the less the systolic filling of LA and the chamber enlarges minimally. Consequently, the passive LA strain will be reduced and the LA “ejection fraction” will be lower.

3.2.5. Mitral E and A Wave Propagation in LV and Their Changing Profiles as Measures of LV Relaxation and Stiffness

Slower mitral E wave propagation into the LV correlates with reduced LV relaxation and faster mitral A wave propagation into LV outflow tract accompanied by its reduced duration correlates with late diastolic LV stiffness [19], but these are somewhat cumbersome to use. These have been supplanted by mitral annular velocities, which provide valuable information about the LV diastolic function a lot more easily.

3.3. Assessment of Myocardial Structure

The ultimate goal of all these exercises is to find out what is happening in the myocardium in an effort the retard or reverse the process that has been causing LV diastolic function. However, this area is still quite primitive. Cardiac MRI is perhaps the most valuable tool for characterizing the myocardium as it can provide insights into the wall thickness, fibrosis, patten of fibrosis, edema, lipid or iron contents and extracellular volume [20,21,22,23]. Cardiac CT HU units can offer some insights as a normal myocardium is about 40–60 HU, fatty metaplasia will be much lower in the negative range, and fibrosis in 60–100 HU units [24,25]. Intramyocardial calcification can be 100–400 HU. Acoustic elastography has been used to measure the stiffness of the liver, but it has not worked well for the heart [26,27]. Integrated backscatter analysis has the potential to differentiate normal and viable muscle from fibrous tissue and other infiltrates [28], but these ultrasound applications are far from routine in clinical practice. The application of AI may perhaps help advance this field.

4. Summary and Conclusions

We have reviewed the clinical relevance and importance of LV diastolic function, the principles behind its assessment and the tools available to assess these in practical terms. Further work in this area and in the field of myocardial tissue characterization will be very important to understand, diagnose, prevent and treat the syndrome of “diastolic HF” or HFpEF. Our goal was to simplify and convey these concepts effectively rather than to perform an in-depth analysis of all the existing literature. We have not delved into the evolving thoughts on treating “diastolic HF”. We would encourage readers to view these publications to gain insights into these [29,30,31,32].