13 The pathophysiology of heart failure

The classical definition of heart failure (HF) is fundamentally a pathophysiological one. It is the ‘inability of the heart to provide sufficient oxygen to the metabolizing tissues despite and adequate filling pressure’.¹ Initially, the abnormalities found in the HF syndrome were described in terms of their haemodynamic effects. However, as the relationship between the pathophysiology of HF and its therapy has emerged over the last 20 years, it is now clear that the pathophysiology of HF is highly complex. It also involves neurohormonal and inflammatory adaptations which initially help the situation but chronically contribute to progression of the HF syndrome and adversely affect the structure and function of the heart itself. This chapter reviews the pathophysiology of HF by outlining what is known about its key players: haemodynamic abnormalities, ventricular remodelling, neurohormonal activation, and inflammatory responses.

Although any cardiac pathology can ultimately lead to HF, most is known about the pathophysiology of HF due to myocardial failure leading to left ventricular systolic dysfunction (LVSD), which concerns most of this chapter. Much work is needed to elucidate further the pathophysiology of HF when it occurs in the presence of normal systolic function.

In response to a reduction in myocardial contractility and/or in the presence of an excessive haemodynamic load, the heart employs a number of adaptive mechanisms to maintain cardiac output. The most important of these is the Frank–Starling mechanism whereby an increase in preload helps augment cardiac output (Fig. 13.1).² Secondly, myocardial hypertrophy begins to provide greater contractility.³ Activation of neurohormonal systems, in particular the renin–angiotensin–aldosterone system (RAAS) and sympathetic nervous system, also stimulate contractility and increase preload by their effects on volume homeostasis and support of arterial pressure and perfusion.⁴ These mechanisms initially improve cardiac performance: however, chronically, they become maladaptive.¹

Remodelling is the term most commonly used to describe the changes in size, shape, and function of the left ventricle that occur as a result of the initial cardiac pathology and its subsequent progression with the activation of the neurohormonal systems described in more detail below. Pathological changes occurring at cellular, organ, and systemic levels drive the process.⁵

The initial insult and its duration determine the broad category of remodelling. There are two distinct types. The first is concentric remodelling, when there is a generalized increase in LV wall thickness and mass. Ventricular dilatation does not occur initially but does subsequently with time. The second type is often referred to as eccentric remodelling, a hallmark of which is dilation of the ventricle, decreased systolic function, and consequent mitral, tricuspid, and aortic valve regurgitation. Eccentric remodelling is classically seen following myocardial infarction, in states of volume overload, valve regurgitation and dilated cardiomyopathies.^5,6

In both kinds of remodelling, an important cellular change is myocyte hypertrophy, which can be triggered by increased load, neurohormonal driven signalling pathways, inflammation, and oxidative stress. Ultimately, these processes lead to well-described post-translational modifications which result in a myocyte phenotype similar to that seen during foetal development, consequent upon activation of the ‘foetal gene programme’. The changes include the generation of new sarcomeres, an increase in the size of myocytes, and a change in their substrate preference from free fatty acids to glucose.⁷ Initially whether these processes occur concentrically or eccentrically, they minimize ventricular wall stress. Over time, however, the changes lead to progressive contractile dysfunction and chamber dilation with a subsequent change in the shape of the left ventricle from elliptical to spherical.⁸

Other events at cellular level are also occurring during ventricular remodelling. There is on going cell death by both necrosis and apoptosis (programmed cell death).⁹ Apoptosis can be triggered by many of the same stimuli that lead to myocardial hypertrophy, i.e. load, neurohormonal driven signalling pathways, inflammation, and oxidative stress (see Chapter 12).¹⁰ In addition, there are changes in the interstitial matrix with an increase in fibrosis and collagen turnover.¹¹ There is an increase in activity of matrix metalloproteinases (MMPs) and a decrease in their endogenous inhibitors, tissue metalloproteinases (TIMPS).¹² The end result is increasing ventricular dilatation (Table 13.1).

The alteration in the geometry of the left ventricle ultimately contributes to increased wall stress (by the law of Laplace) and leads to dilation of the mitral valve annulus and stretching and remodelling of the papillary muscles causing shortening of the posterior mitral valve leaflet thereby causing functional or ischaemic mitral regurgitation (Box 13.1, Table 13.2). The mitral regurgitation further exacerbates ventricular dilation by introducing an element of volume overload into the equation. The presence of mitral regurgitation in HF is, not surprisingly, associated with an adverse prognosis.^13,14

Adverse left ventricular remodelling is associated with increased mortality rates irrespective of the underlying cardiac pathology and is a target for many of the therapeutic advances directed at HF due to LVSD. The remodelling that occurs following myocardial infarction, where there is initial infarct expansion at the border zone between infarcted and normal myocardium, is also a therapeutic target: perfusion strategies limit infarct size and the afterload reducing effects of angiotensin converting enzyme (ACE) inhibitors reduce wall stress which can limit the remodelling process (Fig. 13.2).^15,–17

When coronary artery disease is the cause of the HF, further adverse remodelling can be exacerbated by intercurrent ischaemic events. Left ventricular dysfunction may have a component of hibernating myocardium where myocytes in poorly perfused areas shut down their metabolic activities and cease to contract. Hibernation is potentially reversible by restoring perfusion. This is possibly, in part, one of the mechanisms underlying the beneficial effects of β-blockers.¹⁸

The adverse remodelling process can lead to both electrical and mechanical dyssynchrony which then leads to a vicious cycle of a further reduction in cardiac output, augmented neurohormonal activation, reduction in left ventricular function, more severe mitral incompetence, and further ventricular remodelling. Dyssynchrony manifests as severe chronic heart failure (CHF) and has a poorer outlook and is now the target of cardiac resynchronization therapy.^19,–21

It seems logical to discuss HF with normal ejection fraction (HeFNEF) under the heading of remodelling as the main differences between predominantly systolic dysfunction and HeFNEF seem to be related to the type of remodelling which the heart undergoes: in HeFNEF, the left ventricular hypertrophy phenotype predominates.

Up to 50% of patients presenting with HF have preserved (normal) left ventricle systolic ejection fraction. The syndrome has several other names: diastolic HF, and HF with preserved systolic function (see Chapter 26). The main abnormality occurs in diastole and is due to impaired relaxation or stiffness.

Less is known about the pathophysiology of this form of HF. A lot of work has focused on describing diastolic function in terms of haemodynamics by measuring left ventricular pressure–volume loops and trying to quantify relaxation, filling abnormalities, and chamber stiffness.²²

Abnormalities of diastolic function happen in most patients with LVSD, but in patients with isolated diastolic HF, the only abnormality in the left ventricular pressure loops occurs in diastole, where the curve is shifted upwards and to the left (Fig. 13.3A), so that there is increased diastolic pressure with normal diastolic volumes. In patients with systolic HF (Fig. 13.3B), there are changes in the pressure volume loop that include a reduction in both LVEF and stroke volume. The diastolic portion is not normal either in that there is increased diastolic pressure. In mixed systolic and diastolic dysfunction (Fig. 13.3C), there is usually a modest decrease in LVEF, and an increase in end-diastolic volume and pressure, reflecting decreased chamber compliance.²²

Noninvasive ways of assessing diastolic dysfunction are covered in Chapter 19.

In diastolic HF, complex abnormalities occur at the level of the myofilaments, myocytes, matrix, and the heart.²³ In the myofilaments, abnormal stiffness and relaxation can be due to changes in the proteins within the contractile thick and thin filaments, myosin-binding protein C (MyBPC), and the linkage

protein titin. Titin is a key player and changes in the PEVK region of its isoforms can alter stiffness and elastic recoil in the myocardium. At the myocyte level, calcium signalling and interaction with myofilaments play an important role. Expression and post-translational modifications of the sarcoplasmic reticulum calcium release channel (RyR), Ca²⁺ uptake proteins (PLB, SERCA), sarcolemma exchanger (NCX), and ion pumps can all occur. They are summarized in Fig. 13.4.²⁴

Importantly, the neurohormonal changes seen in systolic HF (discussed in the next section. ‘Neurohormonal activation’) also occur in HeFNEF, although to a lesser extent (Fig. 13.5).²⁵

As compared with patients with predominantly systolic dysfunction, patients with HeFNEF tend to be older, are more likely to be female, and often have hypertension as the main aetiology, although coronary artery disease (CAD) is a common finding as well.²⁶ It is, of course, the main type of HF found in hypertrophic cardiomyopathy and infiltrative cardiac diseases.

A variety of neurohormonal systems are activated in response to the alteration in cardiac function. These are well-developed evolutionary responses designed to protect against exsanguination. They therefore increase blood pressure and critical organ perfusion in the short term but ultimately also become maladaptive when chronically stimulated.

Several mechanisms are involved in sympathetic nervous system activation, which occurs very early in HF, when the cardiac output drops. The fall in cardiac output (and blood pressure) is sensed by mechanoreceptors in the aortic arch, carotid sinus, left ventricular, and renal afferents leading to increased central sympathetic flow and raised circulating concentrations of noradrenaline. There is both increased neuronal release of noradrenaline and decreased reuptake. Initially the sympathetic activation supports the failing circulation by encouraging myocyte hypertrophy, an increased heart rate, vasoconstriction, and lusitropy.^27,28 However, the effects are ultimately deleterious. Both adrenaline and noradrenaline are directly toxic to the myocardium, promoting apoptosis and calcium overload.^29,30 Myocardial oxygen consumption is increased and further adverse remodelling takes place. Down-regulation of β₁-adrenoreceptors occurs, which further inhibits the ability of the heart to respond to a catecholamine surge.^31,32

The parasympathetic nervous system is also disturbed in HF. There is a reduction in vagal tone. The primary abnormality is though to be a reduction in baroreceptor sensitivity.³³

This link between sympathetic activation and HF was described by Cohn’s group, who noted that raised circulating concentrations of noradrenaline were associated with an adverse prognosis in HF.²⁸ The final endorsement of the deleterious effects of the SNS in HF came about with the β-blocker treatment trials that demonstrated improved outcomes for patients (Fig. 13.6).^34,35

The drop in cardiac output and blood pressure also stimulates the RAAS. This does not have the immediate effects of SNS activation. Indeed, there is little RAAS activation in patients with asymptomatic left ventricular dysfunction and more profound activation comes as the syndrome progresses (Fig. 13.7). The activation is also markedly increased by diuretic use.³⁶

Release of renin from the juxtaglomerular apparatus in the kidney occurs through two main mechanisms: first, adrenergic stimulation of β-receptors and secondly, the decreased renal blood flow sensed by mechanoreceptors.³⁷ Renin circulates and converts angiotensinogen already in the circulation (synthesized by the liver) into angiotensin I. Angiotensin I is converted to angiotensin II principally through the action of ACE in the pulmonary circulation and other vascular beds (Fig. 13.8).³⁸ However, angiotensin II is also produced in numerous tissues (including myocytes) by non-ACE-dependent pathways via proteases such as chymase, kallikrein, and cathepsin. The alternative pathways are thought to be the routes by which ‘ACE and aldosterone escape’ occur when patients are treated with ACE inhibitors.³⁹

Angiotensin II is the principal effector hormone of the RAAS. It causes vasoconstriction, myocyte and vascular hypertrophy, and aldosterone release. In addition, it is responsible for the release of other neurohormones such as vasopressin, endothelin and cathecholamines. It also causes thirst by a direct cerebral effect (Fig. 13.9).

The effects are mediated by activation of two receptors. The angiotensin II type 1 receptor (AT₁) is thought to be the one through which most of the deleterious effects occur. However, some adverse effects, such as apoptosis, can occur via stimulation of the type 2 (AT₂) receptor.

The importance of RAAS activation in the progression of HF has been greatly highlighted and further understood by the landmark treatment trials with ACE inhibitors, which were the first drugs to alter the mortality of HF.^40,–42 More recent trials with angiotensin receptor blockers have emphasized the importance of RAAS activation.⁴³

Angiotensin II causes release of the mineralocorticoid hormone aldosterone. Under normal conditions, its production is regulated via adrenocorticotropic hormone (ACTH) and serum potassium concentrations. Aldosterone encourages sodium retention and potassium excretion. The potassium loss contributes to the arrhythmia burden in HF. Aldosterone release is also stimulated by endothelin, cathecholamines, and vasopressin. Plasma concentrations rise in proportion to the severity of the disease.⁴⁴ More recently, its role in causing fibrosis in the myocardium by promoting collagen turnover has been highlighted by the beneficial effects of treatment with aldosterone antagonists both on survival in patients with HF and on reducing fibrosis in both human and animal models.^45,–47

In a similar fashion to angiotensin, endothelin is produced as a pre-pro-peptide that is cleaved by a furin protease into Big Endothelin. Further cleavage by endothelin converting enzyme (a neutral endopeptidase) occurs to produce a family of three endothelins. Endothelin 1 is the predominant isoform expressed in humans. It is mainly produced in endothelial cells. It acts via two receptors, A and B. The receptors are G-protein coupled and activate phospholipase C, eventually leading to calcium release from the sarcoplasmic reticulum. Endothelin’s principal effects are potent vasoconstriction, growth promotion, inotropy, and aldosterone and angiotensin II release.^48,–51

Circulating concentrations of endothelin are elevated two- to threefold in HF patients in proportion to haemodynamic and functional disease severity (Fig. 13.10).^51,52 It is thought to be a key player in the remodelling process. High plasma concentrations are independently associated with an adverse outcome in HF. They also track closely with the degree of pulmonary hypertension.^53,54

Despite the early promise of endothelin A receptor blockade, which improved myocyte hypertrophy, and mixed A and B receptor blockade with bosentan, which favourably altered haemodynamics, larger clinical trials with endothelin antagonists have been disappointing^55,–57.

Vasopressin, also known as arginine vasopressin (AVP—in contrast to lysine vasopressin, found in pigs), or antidiuretic hormone (ADH), is a peptide released in response to a decrease in cardiac output, atrial stretch, and a drop in plasma osmolality. Angiotensin II and noradrenaline also augment its release in HF. Vasopressin’s main site of action is the renal distal collecting tubule where it acts via the V₂ receptor to mobilize aquaporin-2 channels to the cell surface. Aquaporin-2 allows water reabsorption from the urine in the collecting duct down the osmotic concentration gradient in the renal medulla. Vasopressin also acts via V₁ and V₂ receptors to cause vasoconstriction in the pulmonary and peripheral circulation.

Although AVP is normally released in response to an increase in plasma osmolality, it is paradoxically raised in patients in proportion to the severity of HF³⁶ despite such patients often having hyponatraemia (which might be expected to reduce AVP). Nonosmotic stimuli to AVP release, such as angiotensin II and prostaglandins, become dominant in these circumstances.

Short treatments with the AVP antagonists conivapatan and tolvaptan improve haemodynamics, fluid removal, and electrolyte balance, but a recent long-term trial has not shown any conclusive effects on HF morbidity and mortality.^58,59 It may be that the AVP antagonists are of particular benefit in patients with hyponatraemia, but such a proposal has not been specifically tested.

The natriuretic peptides (NP), hormones produced by the heart, are also activated in HF. The most important ones in HF are atrial natriuretic peptide (ANP), mainly produced in the atria, and B-type natriuretic peptide (BNP), which is predominantly found in the ventricular myocardium.⁶⁰ Their release is stimulated by increased wall stress and stretch as well as by the circulating neurohormones, endothelin, angiotensin II, aldosterone, vasopressin, and noradrenaline. They act as counter-regulatory hormones and have a major role in controlling cardiovascular homeostasis via their key effector functions: natriuresis, diuresis, and vasodilation. Plasma concentrations of the NPs are raised in proportion to the severity of HF and are powerful predictors of a poor prognosis.^61,–64 Although the counter-regulatory system is ultimately too weak to prevent progression of HF, the NPs have become very useful biomarkers to aid in diagnosis and assignment of prognosis, and as potential tools for monitoring therapy in HF. Manipulation of the system either by inhibiting of NP breakdown with neutral endopeptidase inhibitors or by augmenting their effects by use of intravenous exogenous recombinant BNP has

not proved very successful to date, although further trials of these strategies are under way.^65,66

Potentially, inflammation may play an important role in the pathophysiology of HF (see Chapter 8). As in other phases of cardiac pathology, there is a nonspecific indication of underlying generalized inflammation given by a raised CRP concentration.⁶⁷

The key inflammatory players investigated to date are the inflammatory mediators, cytokines. Initial work in patients with advanced HF and cardiac cachexia showed that plasma concentrations of tumour necrosis factor α (TNFα) were elevated.^68,69 Since then, many studies have shown increases in other cytokines such as interleukin-1β (IL-1β), IL-2, IL-6, Fas ligand, monocyte chemoattractant protein-1 (MCP-1) and macrophage inflammatory protein α (MIP-1α).⁷⁰ In a similar manner to neurohormonal activation, prolonged inflammatory activation with an imbalance between pro- and anti-inflammatory cytokines is ultimately deleterious.

TNFα is, to date, the most studied cytokine in HF. It is predominantly released from macrophages, although neutrophils, lymphocytes, platelets, mast cells, and endothelial cells can all produce it. In addition, it can be released by myocytes subject to stretch.

It is increased, as are its soluble receptors, in proportion to the severity of HF (Fig. 13.11).⁷¹ Both TNFα and IL-6 have been shown to be independent predictors of prognosis in HF.⁷² TNFα has a negatively inotropic action as well as toxic effects on the myocardium that lead to apoptosis, myocyte hypertrophy, and matrix remodelling, contributing to progressive adverse ventricular remodelling. Similarly, both IL-1 and IL-6 are negatively inotropic in vitro and their circulating concentrations are increased in proportion to the clinical and haemodynamic severity in HF patients.^73,74

Less is known about the role of other chemokines in HF. The production of MCP-1 is stimulated by IL-1, IL-6, TNFα, and angiotensin II. Levels of MCP-1 are increased in animal models of HF.⁷⁵ More recently, MCP-1, MIP-1α, and RANTES (regulated upon activation, normal T cell expressed and secreted) have been shown to be elevated in human HF patients and correlated with traditional markers of disease severity such as NYHA class and ejection fraction.⁷⁶ Furthermore, chemokines and their receptors are found in biopsies taken from failing myocardium.⁷⁷ They may well play an important role in the pathogenesis of the condition.

To date, however, modulation of inflammatory mediators in HF by statins, TNFα receptor blockers or monoclonal antibodies directed to TNFα has failed to demonstrate beneficial effects in large studies in HF patients.^78,79

Many other biomarkers are elevated in patients with HF (Table 13.3). Whether these are mere epiphenomena or whether they are involved in the pathophysiology of the syndrome is as yet unclear. Some seem to be obvious candidates for a major role: for example, ST2, the soluble receptor for IL-33, seems to be linked to early fibrosis in the heart,^80,81 and apelin, the endogenous ligand of the APJ receptor, is a potent inodilator produced in the heart but down regulated in HF and up-regulated in models of reverse remodelling.^82,–84 Manipulation of these systems will unravel their roles.

The HF syndrome does not just affect the heart. As a result of the extensive compensatory activity, structural changes take place in vascular arterioles with increasing stiffness of vessels and endothelial dysfunction. Morphological and functional changes occur in skeletal muscle and respiratory function is affected with an increase in physiological dead space and airways obstruction.^85,–87

The pathophysiological processes at play in the HF syndrome combine to produce a progressive state of reduced left ventricular function, left ventricular adverse remodelling, and dilatation which leads to death from progressive pump failure or ventricular arrhythmia. Except in cases where the myocardial function returns to normal or near normal, neurohormonal blockade leads to some reverse remodelling, associated with an improved outlook for patients. However, the natural history for many patients is that treatment merely delays deterioration in left ventricular function which is then associated with many comorbidities (particularly renal dysfunction) which ultimately lead to progressive multiorgan dysfunction and death.