2 Heart failure syndromes

Heart failure (HF) is a protean condition, presenting acutely to hospital in most cases, but presenting with a more insidious course to primary care physicians in many cases.¹ Patients may only be diagnosed as having a primary cardiac problem after being seen by respiratory physicians or even, on occasion, after gastrointestinal workup for hepatomegaly (with or without jaundice) or weight loss. Nevertheless, there are common presenting clinical syndromes in patients with HF (Box 2.1) which should prompt different initial treatment strategies.

As a pragmatic definition, acute HF is HF necessitating emergency admission to hospital. Attempts have been made to classify acute HF into different types,² but the classification schemes often read as arbitrary, and resemble the Borges classification system for animals.³ For the majority of patients, the problem of acute HF is that of ‘fluid in the wrong place’; if that fluid is in the lungs, the patient has pulmonary oedema, but if predominantly in the tissues, the patient may present with anasarca (Greek ανα-, throughout; σαρχ, σαρκ-, flesh). Of course, patients will lie somewhere along a spectrum. Most patients will have some degree of pulmonary congestion even if the dominant problem is one of fluid retention; conversely, many patients with frank pulmonary oedema will have some evidence of ankle oedema.

A large number of patients presenting with acute HF will have a background history of antecedent stable chronic HF. For patients presenting with pulmonary oedema, there will often be an obvious precipitant of the immediate crisis, and the trigger should be sought and treated (Box 2.2). Failure of compliance is the commonest identified trigger in several studies, with perhaps half of all admissions being potentially preventable if compliance had been better.^4,5

Other common triggers include further ischaemic events in patients with ischaemic heart disease underlying their HF, and arrhythmia. Particularly in older patients, intercurrent illness, and especially chest infection, is a common precipitant (Fig. 2.1).⁶ The immediate precipitant does affect prognosis. Where uncontrolled hypertension is the culprit, the prognosis is good: however, patients admitted because of pneumonia, worsening renal function, or ischaemia have a worse prognosis.⁶

The fact that poor compliance is such a common trigger in all populations studied emphasizes the importance of patient education and follow-up to try to prevent recurrences. Although it is difficult to provide proof, increased numbers of admissions are certainly associated with a worse long-term prognosis.⁷

If the left ventricle fails acutely, cardiac output is maintained by the Frank–Starling mechanism: an increase in the left ventricular end-diastolic pressure, representing the preload of the left ventricle, leads to an increase in stroke work. However, the increase in pressure inevitably causes an increase in pulmonary venous, and then capillary, pressure.

The balance of forces keeping fluid within blood vessels is largely a balance between the hydrostatic pressure tending to force fluid out and the colloid osmotic pressure tending to keep fluid in. If the left ventricle fails, the rise in pulmonary capillary pressure required to maintain left ventricular output will exceed the combined resistance of the colloid osmotic pressure and the alveolar basement membrane and the capacity of the pulmonary lymphatics to drain tissue fluid.⁸ At this point, fluid will start to accumulate in the pulmonary interstitium, then the alveoli, and ultimately the airways (see Fig. 2.2).

As fluid accumulates, so the lungs become stiffer and the work of breathing increases; bronchospasm (so-called ‘cardiac asthma’) can be a prominent feature; and at the same time, gas exchange is hampered by fluid filling the alveoli. The sympathetic response worsens the situation by causing tachycardia and peripheral vasoconstriction, thereby increasing the afterload against which the failing left ventricle is trying to eject blood.

Pulmonary oedema is an acute medical emergency and an exceptionally alarming experience for the patient. The typical clinical picture is well known. Symptoms tend to be of very abrupt onset, typically appearing over the course of less than an hour, but may be preceded by a day or so of worsening breathlessness and nocturnal dyspnoea. The patient rapidly becomes extremely breathless and distressed; speaking more than a few words at a time becomes impossible, and the need to breathe becomes overwhelming. The fearful sensation of impending death, angor animi, is very common. The patient needs to sit upright, often forwards, and might die if forced to lie flat. As the alveoli fill with fluid, the patient will cough, often violently, and will expectorate quantities of pink-tinged frothy fluid.

There is invariably a huge sympathetic nervous system response: the periphery becomes shut down due to vasoconstriction with associated pallor and coldness of the skin. Profuse sweating is commonly seen.

Common physical findings include sinus tachycardia, or arrhythmia, commonly atrial fibrillation or ventricular tachycardia. Hypertension is common, either as a precipitant or as a consequence of the sympathetic activity. The jugular venous pressure may be raised, but there are often no signs of peripheral oedema as the syndrome develops abruptly: there has been no time for the patient to become fluid overloaded. The problem is not one of excess fluid; rather, fluid in the wrong body compartment.

The cardiac findings depend upon the previous history, and may include a displaced and dyskinetic apex beat. A gallop rhythm is very common, with third, fourth, and summation sounds difficult to distinguish given the tachycardia. The chest may be silent in extremis, but is usually filled with a variety of fine and coarse crackles, and wheezes. In cases presenting early or with mild pulmonary oedema, the classical finding of fine late inspiratory crackles at the bases may be heard (see Fig. 2.3).

Modern treatment of acute pulmonary oedema has changed the natural history of pulmonary oedema, and the outlook depends upon the severity of the syndrome as well as the underlying causes. Grading systems for recording severity are available, with the Killip class⁹ and an assessment based on the combination of perfusion and congestion¹⁰ commonly used (Table 2.1). The gradings are primarily designed for use in people with HF following acute myocardial infarction, but are helpful in assessing prognosis whatever the underlying cause of the pulmonary oedema.

Patients with acute pulmonary oedema typically present outside office hours, and it is striking that they either improve rapidly or die, so that within a few hours the immediate clinical outcome is obvious.

At the other end of the spectrum of acute HF are patients presenting with fluid retention. This is a far more gradual process that that underlying acute pulmonary oedema. By the time patients present, they may have accumulated over 20 Lof excess fluid (and it requires approximately 5 L excess before ankle oedema appears).

The underlying pathophysiology is the neurohormonal response to poor renal perfusion and fall in arterial blood pressure. The kidneys ‘try’ to maintain normal perfusion by the release of renin, ultimately leading to aldosterone release and salt and water retention by the kidneys. In addition, antidiuretic hormone (ADH; arginine vasopressin) is released from the anterior pituitary gland.

ADH is high relative to serum sodium, and causes water retention and the production of hypertonic urine, coupled with thirst, which results in increased fluid intake.¹¹

The excess fluid increases the venous hydrostatic pressure which results in the Starling forces in the capillaries favouring fluid loss from the vessels and accumulation in the tissues.

Where the excess fluid accumulates is a function of gravity. The ankles are usually first affected, commonly with swelling that increases during the day and may have gone by the next morning as a consequence of several hours’ leg elevation. The oedema progressively rises up the legs, and then affects the abdominal wall. Pleural effusions and ascites are common at this stage, and pericardial effusions may become large.

The prominent physical finding is, of course, peripheral oedema, which is pitting. Sinus tachycardia or atrial fibrillation are usual findings together with low systemic blood pressure. The jugular venous pressure is invariably raised, and there may be evidence of tricuspid regurgitation in the jugular venous waveform. There is often a dilated heart with prominent third heart sound. The lung fields may be clear, or there may be some evidence of pulmonary oedema.

There is some evidence that strict bed rest might result in a reduction in oedema,¹² but without therapy anasarca or ‘dropsy’ becomes a chronic state. Surprisingly large volumes of excess fluid are sometimes tolerated for many months before a patient finally presents (Fig. 2.4). Modern diuretic therapy means that the majority of patients progress at this stage to having chronic HF.

The vast majority of patients with HF receive active treatment so that following a presentation with an acute episode of HF, congestion is removed. The chronic HF syndrome is what affects patients with heart failure once they are taking appropriate combination therapy with diuretics (as needed), angiotensin converting enzyme (ACE) inhibitors, β-blockers, and sometimes aldosterone antagonists. For these patients, the term ‘congestive’ HF is inappropriate—they should not be congested at all with suitable use of diuretics.

The symptoms of chronic HF are most commonly breathlessness and fatigue on exertion, leading to exercise limitation and consequent decline in quality of life. The severity of symptoms is most commonly measured using the New York Heart Association (NYHA) scale (Table 2.2).¹³ Unfortunately, the NYHA system is only weakly related to measures of exercise capacity, and bears no relation to left ventricular function at rest. It is often not clear from clinical studies whether the patients themselves are recording the score (which should surely be the case, as it is a subjective scoring system) or the physicians caring for the patients. When physicians score the patients, the NYHA system becomes a composite score of overall severity of HF rather than being a pure symptom score.¹⁴

Another limitation is that patients are forced into one of four categories, and in practice, most patients recruited to clinical trials are in either class II or class III (those in class I have no symptoms and might thus be thought not to have HF; those in class IV are bed-bound). Further, there is a temptation to describe populations of patients by their ‘average’ NYHA class. This is inappropriate—no individual patient can have anything other than an integer score, and the scale is nonlinear.

Other scoring systems are better matched to the complexity of symptom assessment, and are better able to define subtle differences both between patients and in response to therapy. They are more cumbersome to administer in practice than the NYHA score. The Minnesota Living with Heart Failure self-assessment questionnaire is the most widely used, and is a series of 21 questions, each scored from zero to 5.^15,16 The Kansas City questionnaire¹⁷ has the advantage of asking patients about how symptoms have changed and gives a better idea of the trajectory of an individual’s clinical course.

A functional assessment is very helpful in trying to get an objective measure of a patient’s symptoms. Incremental exercise tests with metabolic gas exchange measurements are often thought to be the best single assessment, but the equipment required is not universally available. Many patients are unable to manage an incremental exercise test. The six-minute walk test^18,19 is easy to administer, can be attempted by the great majority of patients, and is reproducible.

Why chronic HF causes shortness of breath and fatigue has traditionally been attributed to abnormal central haemodynamics. It might be supposed that ‘forward’ failure leads to inability adequately to perfuse exercising skeletal muscle, thereby resulting in fatigue; and ‘backward’ failure leads to a rise in pulmonary venous pressure, stiff (or even oedematous) lungs, thereby resulting in breathlessness. However, against this hypothesis is the fact that there is no relation between exercise capacity and central haemodynamics (at least at rest); some patients with very severe left ventricular dysfunction have near normal exercise capacity;²⁰ acute correction of central haemodynamics (e.g. with positive inotropic drug therapy or even heart transplantation²¹) does not result in acute correction of exercise limitation. During early stages of exercise, the cardiac output responses are often normal in HF.

Some light is thrown on the issue by the observation that different kinds of exercise can lead to different symptoms in the same individual: rapidly incremental tests are more likely to cause limiting breathlessness,²² whereas slower tests, although eliciting the same exercise performance, are more likely to cause fatigue. Cycle exercise is more often stopped by fatigue than breathlessness than is treadmill exercise, even when the same level of exercise is performed.²³, ²⁴

Some work has suggested that right ventricular function and pulmonary haemodynamics might be key determinants of exercise capacity, but some patients with the Fontan circulation (who thus have no right ventricle in the circulation) have near normal exercise capacity.²⁵

The lungs are abnormal in many patients with chronic HF, in terms both of spirometric variables and of diffusion capacity.²⁶ In some studies, exercise capacity correlates closely with some spirometric variables.²⁷ However, about one-third of patients being assessed for transplantation will have normal spirometry and diffusion.²⁸

One possibility is that pulmonary dead space might be increased. Dead space is that component of air in the respiratory tract not available for gas exchange. Anatomical dead space is the fixed dead space formed from the airways. It could plausibly be increased by an altered ventilatory pattern: the same minute ventilation achieved with double the respiratory rate and half the tidal volume will double anatomical dead space. Physiological dead space, on the other hand, is made up of alveoli that are ventilated but not perfused—‘wasted’ or inefficient ventilation.

However, there is no dead space receptor that might sense the increase and drive an excessive ventilatory response. In contrast to what might be expected, patients with chronic HF have better than normal arterial blood gases during exercise,²⁹ suggesting that the primary abnormality driving an excessive ventilatory response to exercise must lie elsewhere.

Abnormalities of skeletal muscle in chronic HF range from ultrastructural,³⁰ through histological³¹ and metabolic,³² to changes in gross function (weakness and early fatigue³³) and reduction in bulk.³⁴ The key feature distinguishing patients with abnormal exercise capacity is differences in skeletal muscle function: those with normal exercise capacity have normal (or near normal) skeletal muscle.²⁰ That these changes might cause the sensation of fatigue is easy to picture.

A unifying picture to explain the origin of symptoms comes from the ergoreflex (see Fig. 1.3). The ergoreflex is neurally mediated and arises from exercising muscle in proportion to work done. The strength of the signal is also proportional to the amount

of muscle doing the work—the stimulus is greater when arm muscle is used to perform a given workload compared with leg muscle.³⁵

Stimulation of the ergoreflex both increases ventilation and causes sympathetic nervous system activation. In patients with chronic HF, the ergoreflex is enhanced in proportion to the degree of exercise limitation and ventilatory abnormality.³⁶

The ergoreflex model explains the two common chronic HF symptoms, and also helps explains other features of the syndrome. The origin of the sympathetic activation is not immediately obvious as the baroreflexes are down-regulated in HF.³⁷ Ergoreflex stimulation causes sympathetic activation. In addition, the chemoreflexes are enhanced in chronic HF, and they, too, are associated with baroreflex down-regulation.³⁸ Indeed, the peripheral model gives a new understanding of autonomic nervous system changes in chronic HF;³⁹ in the normal state, the main inputs into autonomic nervous system control from the cardiovascular system are the baroreceptors and cardiopulmonary receptors, with parasympathetic modulation being the major output; in HF, chemoreceptors and ergoreceptors are the most important inputs, and sympathetic activation results.

There is, of course, nothing ‘natural’ about the outcome of patients with chronic HF. The marked improvements in prognosis that have come with modern medical therapy⁴⁰ (Fig. 2.5) are shown by the falling event rates among patients in the placebo groups of clinical trials. For very many patients, chronic HF can be stable for many years, but for some, it can be a progressive illness resulting in early death or transplantation.

That chronic heart disease can result in cachexia has been known for many hundreds of years. Quite how it comes about remains unknown, and anecdotally its frequency seems to be falling, perhaps as a consequence of widespread use of β-blockers. Part of the difficulty in discussing the syndrome is the lack of a universally recognized definition of cachexia. Clinicians know it when they see it, but defining it is a different matter. It is best thought of as a process of active weight loss rather than referring to a patient who is simply thin; but how much weight loss, and loss from which body compartment (fat, muscle, or bone) is not satisfactorily determined.

Partly as a result of the lack of an agreed definition, the epidemiology of cardiac cachexia is unclear. Data from clinical trial databases suggests that weight loss is common,⁴¹ with over 40% of patients losing 5% or more of their body weight during 3 years of follow-up in the SOLVD trial (Fig. 2.6).

The weight loss in cachexia is from all body compartments, not simply lean muscle. Muscle loss is common from early in the course of chronic HF,⁴² but loss of nonlean tissue is also seen⁴³ and patients are more prone to osteoporosis than normal individuals.⁴⁴ Patients with cachexia tend to have more advanced HF. The loss of bulk contributes to the general sense of fatigue and the activation of the ergoreflexes outlined above.

Chronic HF seems to be an inherently catabolic state.^45,46 This is seen even at the level of increased hepatic fibrinogen synthesis.⁴⁷ Part of the explanation may be the continuous neurohormonal activation of HF. Sympathetic activation causes an increase in basal metabolic rate,^48,49 glycogenolysis, and lipolysis.⁵⁰ In animal models, high levels of angiotensin II are also associated with profound weight loss.^51,52 In normal individuals, infusions of catabolic hormones (hydrocortisone, glucagon, and adrenaline) induce hyperglycaemia, hyperinsulinaemia, insulin resistance, and negative nitrogen balance—precisely the changes seen in the cachexia syndrome.^53,54

Other neurohormonal changes are commonly seen in chronic HF which are much more prevalent in patients with cachexia. In general, there seems to be a shift in the normal balance between catabolic and anabolic hormonal factors, so that patients develop resistance to the effects of both insulin⁵⁵ and growth hormone⁵⁶ and a decrease in the ratio of anabolic to catabolic steroid.⁵⁷ Additional procatabolic changes include the production of tumour necrosis factor (TNFα),⁵⁸ which is itself related to the changes in neurohormones.⁵⁹ These changes are strongly related to the alterations in body compartments,⁶⁰ suggesting that there is indeed an aetiological link between neurohormonal activation and weight loss.

One fascinating potential explanation which explains the otherwise slightly mysterious rise in TNFα is the possibility that bowel wall oedema, possibly caused by recurrent episodes of decompensation, allows the translocation of bacterial endotoxin across the bowel wall.⁶¹ Bacterial endotoxin is the most potent natural stimulus for TNFα production. In support of this notion, circulating endotoxin is high during episodes of decompensation, and declines with treatment.⁶² A further observation is that the endotoxin hypothesis may explain the apparent protective effects of cholesterol in patients with chronic HF:⁶³ endogenous lipoproteins act as a sump for endotoxin.⁶⁴

Other potential contributors to cachexia are poor dietary intake, although there is only small-scale evidence for such a phenomenon.^65,66 There is some evidence that malabsorption (possibly as a consequence of gut oedema) may be a cause of impaired nutrition and fat malabsorption in particular.⁶⁷ Malaise, lethargy, nausea, lack of motivation, and poor mobility may contribute, particularly in elderly people.⁶⁸

Hyperalimentation does not seem to offer any substantial benefit to stable patients with chronic HF,⁶⁹ but there is no large-scale intervention trial looking at its possible benefits in a cachectic population. There is some evidence that micronutrient supplementation may be helpful.⁷⁰ Other dietary approaches are being considered (reviewed⁷¹ with many suggestions focusing on possible anti-inflammatory strategies), but none has so far proved to be effective.

Conventional HF therapy does affect cachexia. ACE inhibitors, or at least enalapril, reduce the risk of weight loss (Fig. 2.7).^72,73 Similar effects have been reported with the angiotensin receptor antagonist candesartan.⁷⁴ β-Adrenoceptor antagonism causes a fall in basal metabolic rate,⁷⁵ an effect that may underlie the increase in weight seen in some patients on long-term β-blocker therapy.^76,77 There is some evidence that β-blockers may reduce the risk of cardiac cachexia developing,⁷⁸ and may reverse it once it has occurred.⁷⁹

The major clinical impact of cachexia is on outcome. Defined as unintentional weight loss of at least 7.5%, cachexia was associated with a mortality of 50% at 18 months.⁸⁰ In fact, decreasing body mass and not just an active process of cachexia is inversely related to survival.^81,82 Increasing body mass is strongly associated with survival following left ventricular assist device implantation.⁸³ The situation following cardiac transplantation is more complex. Weight at transplantation does not have a major impact on outcome,⁸⁴ but because thinner patients have a worse prognosis, there is more to be gained from transplantation for underweight patients.

It may seem odd to consider ‘sudden death’ to be a clinical syndrome, but one of the peculiarities of chronic HF is that patients are at risk of dying suddenly at any point in their clinical course. Approximately half the patients dying from HF die from progressive disease, but the others die suddenly. The mode of death in HF depends upon the severity of the HF syndrome (see Fig. 2.8). With worsening NYHA class of symptoms, so the likelihood of a death being sudden declines, with sudden death predominating as the mode of death in patients with milder symptoms. Note, however, that the likelihood of dying is much lower in patients with mild symptoms, so the absolute number of sudden deaths increases with worsening symptoms.

Patients with chronic HF are prone to tachyarrhythmias, both atrial and ventricular. The cause of sudden death has traditionally been considered to be a ventricular arrhythmia—either ventricular tachycardia or fibrillation (Fig. 2.9). However, it is important to remember that conduction system disease is very common in HF, and so patients are at risk of bradycardia as well.

A difficulty in understanding the pathophysiology of sudden death is the lack of an agreed definition of sudden death.⁸⁵ A further consideration is that most patients with chronic HF have underlying coronary heart disease, and so are potentially at risk from further ischaemic events. Although sudden deaths are commonly presumed to be due to arrhythmia, post-mortem studies of patients with HF dying suddenly show that very many are secondary to ischaemic events.^86,87 These ischaemic events are mostly not detected in life, leading to a false impression of how common arrhythmic death is. The importance lies in appreciating that therapies targeted specifically at sudden death (e.g. with implantable cardioverter-defibrillators) are not able to eradicate sudden death, which will still continue to happen.