23 Metabolic exercise testing in chronic heart failure

Chronic heart failure (HF) is characterized by exercise intolerance usually due to breathlessness or fatigue in the presence of cardiac dysfunction. Hence when assessing a patient with such symptoms, in addition to appropriate cardiac imaging, some form of standardized exercise testing is important to measure objectively the degree and nature of the symptoms and confirm their aetiology. In addition, exercise capacity is a powerful predictor of mortality and is used as a marker of the need for cardiac transplantation. Although useful data can be gained from a standard treadmill-based exercise test,¹ or a corridor walk test, additional and independent information is available when the test is performed while measuring metabolic gas exchange.²

In patients with chronic HF, an exercise test with or without metabolic gas analysis gives important information about ischaemia, inducible arrhythmias, and prognosis.^3,–6 Metabolic gas analysis during exercise is accurate and reproducible⁷ and also a robust predictor of outcomes.^8,9

During an incremental exercise test, patients exercise to exhaustion while wearing a tight-fitting mask or mouthpiece, and expired air is sampled to measure metabolic gas exchange. This can either be done by sampling from a large bag at intervals (Douglas bag method) or more commonly on a breath-by-breath basis. Oxygen uptake, minute ventilation (as a product of tidal volume and frequency of ventilation), and CO₂ production can be measured (Fig. 23.1). Exercise duration, heart rate, blood pressure changes, and peak heart rate are often quoted, but the variables most commonly used to describe the exercise response are peak O₂ consumption (pVo₂), anaerobic threshold (AT), and the derived variables of the relationship of ventilation to CO₂ production (Ve/Vco₂ slope) and the ratio of CO₂ output to O₂ uptake (Vco₂/Vo₂; respiratory exchange ratio or RER).

Skeletal muscle cellular activity requires energy, which is stored in skeletal muscle myocytes in the form of creatine phosphate and glycogen. Creatine phosphate is rapidly accessible, but stores are sufficient for only a few seconds of work. The currency of energy transfer is in the breaking and reformation of the terminal phosphate bond of ATP. The energy released when this bond is broken is used for a cycle of linking and releasing of the two elements of the contractile structure, actin and myosin. The linking and subsequent release leads to conformational changes in the cell. If neighbouring cells perform this activity in a controlled and coordinated manner, contraction of the muscle can take place. Other cellular activities such as biosynthesis and active transport are also supported by the energy released from the hydrolysis of ATP.

The production of units of ATP, which must continuously be regenerated to allow cellular work to continue, depends largely upon the oxidation in the mitochondria of carbohydrates, fatty acids, and, in conditions of starvation, protein. Carbohydrate sources (such as glycogen) are converted initially to pyruvate, and then enter the tricarboxylic acid cycle (TCA or Krebs cycle), as acetyl-CoA, eventually forming CO₂, high-energy electrons, and hydrogen ions. On the other hand, fatty acids undergo β-oxidation but then also enter the TCA cycle as acetyl-CoA units. The release of the protons and the entry of the electrons into the electron transport chain is dependent on the consumption of O₂ and leads to the production of water and the regeneration of ATP (Fig. 23.2). An important part of the process is the reduction of nicotinamide adenine dinucleotide (NAD) to NADH. NADH is the route through which electrons enter the electron transport chain, in the process being oxidized again to NAD.

In conditions where the supply of O₂ is not sufficient to keep up with demand, the impaired flow of electrons into the electron transfer chain would eventually inhibit energy production in the cell. In the absence of O₂, pyruvate cannot enter the TCA cycle, so, in order to allow the reoxidization of NADH to NAD, pyruvate is instead converted to lactate. This is termed anaerobic metabolism and leads to the production of fewer ATP molecules for each molecule of substrate than aerobic metabolism. When O₂ becomes available again, the lactate can be converted back to pyruvate and pass into the TCA cycle. Cellular metabolism can therefore continue in circumstances of relative O₂ deficiency, but is less efficient.

In most circumstances, both aerobic and anaerobic metabolism occur simultaneously in all cells, but in skeletal muscle cells, particularly at high workloads, anaerobic metabolism provides a greater proportion of the energy than at rest.

The ratio of O₂ consumption to CO₂ production in the tissues is a function of the substrate used. When using glucose, each molecule of O₂ used leads to the formation of one molecule of CO₂:

C₆H₁₂O₆ + 6O₂ → 6H₂O + 6CO₂

This leads to a respiratory quotient (RQ: Qco₂/Qo₂) of 1. Lipids are much more reduced than carbohydrate, and so when fatty acids are oxidized, the CO₂ production is lower than the rate of O₂ consumption with a correspondingly lower RQ of around 0.7:

C₁₆H₃₂O₂ + 23O₂ → 16H₂O + 16CO₂ and 16/23 = 0.7where C₁₆H₃₂O₂ is palmitic acid, a commonly used fatty acid.

In humans at rest, there is a preponderance of fatty acid metabolism so the CO₂ production is lower than O₂ consumption and the RQ is around 0.7. The higher the proportion of carbohydrate used, the greater the required ventilatory response to eliminate the CO₂.

With metabolic gas exchange measurements, ‘whole body’ gas exchange is measured from the difference between inspired and expired O₂ and CO₂ at the mouth and is not a direct measure of cellular metabolism. As with RQ, O₂ consumption and CO₂ production measured at the mouth can be expressed as a ratio—the RER.

During exercise there is a several-fold increase in O₂ consumption by skeletal muscle cells. The physiological adaptations to exercise include increased cardiac output from a resting level of 3.5 L/min to typically 20 L/min, with local arterial vasodilation to increase O₂ delivery to exercising muscles. There is a rightward shift of the oxyhaemoglobin dissociation curve encouraging unloading of O₂ in areas of acidosis, and there is increased ventilation up to 100 L/min.

The body’s upper limit of O₂ utilization is determined by the maximal cardiac output,¹⁰ arterial O₂ content, the fractional distribution of cardiac output to the exercising muscle,¹¹ and the ability of the muscle to extract O₂.¹² Ventilation is a limiting factor only when the ventilatory capacity is insufficient to eliminate the CO₂ produced by aerobic metabolism and the bicarbonate buffering of lactic acid,¹³ or at the rarely achieved maximum voluntary ventilation (MVV).

The measurement of O₂ uptake at the mouth is traditionally recorded as O₂ consumption (Vo₂). Strictly, O₂ is consumed at the cellular level, and only a small percentage of resting O₂ uptake is consumed by skeletal muscles. During exercise, most of the increase in O₂ uptake at the mouth reflects increased O₂ consumption in the skeletal muscles. Hence O₂ uptake at the mouth and O₂ consumption are assumed to be equivalent and the terms are used interchangeably. The value is presented either as an absolute (mL/min) or referenced to body weight (mL/kg per min).

Physiologists refer to the concept of maximal O₂ consumption (‘Vo_{2 max}’) as the O₂ consumption plateau reached where an increase in imposed workload no longer elicits an increase in Vo₂. However, on most occasions, even normal individuals cannot tolerate the discomfort long enough to achieve a plateau in Vo₂, and HF patients are almost never able to exercise to such plateau. A flattening of the Vo₂–work rate relationship is therefore not seen. The term ‘peak Vo₂’ should be used when a plateau is not reached, and is used as an index of peak exercise capacity.

Metabolic activity produces CO₂ and water as waste products. The amount of CO₂ produced for a given energy release, is determined by the substrate and how it is metabolized. As with O₂ consumption, the CO₂ production occurs in the metabolizing tissues but is measured at the mouth.

Ventilation (in L/min) is a product of frequency (f  ) and tidal volume (V_T) at the mouth. The relationship between O₂ consumption and workload during submaximal exercise is linear.¹⁴ However, the ventilatory response is not related linearly to O₂ consumption (Vo₂). Instead, there is a close relationship between the production of CO₂ (Vco₂) and minute ventilation (Ve) (Fig. 23.3).^15,–17 This relationship, termed the Ve/Vco₂ slope, becomes steeper above the AT (see below).¹⁶

The relation between ventilation and CO₂ production is given by

 Ve/Vco₂Vco₂ = 863/Pao₂ ×x (1 – Vd/Vt)(Equation 23.1)

where 863 is a constant to standardize volume measurements, Paco₂ is the arterial tension of CO₂ and Vd/Vt is dead space as a fraction of tidal volume. A consequence is that, in the short term at least, Vco₂ is determined by ventilation: if an individual hyperventilates, Vco₂ increases as the CO₂ passing through the lungs is blown off.¹⁸

Equation 23.1 includes a calculation for dead space—lung tissue that is ventilated but not perfused, which includes bronchi, trachea, and underperfused alveoli. As a result of perfusion changes and the increased frequency of ventilation,¹⁹ dead space ventilation as a fraction of tidal volume is greater in patients with chronic HF than in control subjects. This might contribute to exercise limitation and symptoms in patients with chronic HF; hence Vd/Vt is frequently presented within a cardiopulmonary exercise report.

Several variables are commonly derived from the basic measurements. It must be borne in mind that derived variables especially suffer from a multiplication of errors when calculated from poorly performed tests.

The ventilatory equivalents for O₂ (Veqo₂ or Ve/Vo₂) and CO₂ (Veqco₂ or Ve/Vco₂) give an impression of the instantaneous ventilation required at a particular time point for the metabolic gas in question and are usually plotted against work rate or time in a progressive test. Both decline slightly early during progressive exercise until the AT, at which point both increase. Both ventilatory equivalents for O₂ and CO₂ are higher in patients with chronic HF than in controls throughout exercise.

A problem with incremental testing in a population of individuals unused to maximal exercise tests is determining whether a maximum has been reached, or whether exercise is ‘submaximal’. Where a genuine plateau in O₂ consumption is reached, then a maximum test can be inferred confidently. For most patients with HF, the RER (Vco₂/Vo₂) is usually used.

When glucose is the metabolic substrate, the RER is 1.0 (1 mole of CO₂ is produced for each mole of O₂ consumed). The RER for lipid metabolism is around 0.7, as lipid is more highly reduced than glucose. At rest in patients and normal individuals alike, the ratio is around 0.7, representing the balance between fatty acid and glucose metabolism. However, as exercise progresses, ATP is generated increasingly from anaerobic metabolism. The shift away from aerobic metabolism leads to an increase in lactate production which is buffered in the blood by bicarbonate (HCO₃^–) ions. Carbonic acid dissociates to water and CO₂, which is then blown off at the mouth. As a consequence, CO₂ output increases relative to O₂ consumed, and the RER gradually rises above 1.0. A test is usually taken to be maximal when the RER exceeds 1.1 (although fit individuals may attain an RER of 1.4) and in practice in HF populations, an RER of 1.0 or more is often accepted. Higher RER levels improve the prognostic information of the data collected,²⁰ whereas those with low RER levels are much less informative.²¹

At low workload, aerobic metabolism is able to support energy production completely, and the need for anaerobic metabolism is low. During incremental exercise, there comes a point where the circulation can no longer deliver sufficient O₂ to the exercising muscle, and anaerobic metabolism supplies an increasing proportion of the ATP. At this point, lactate is generated, which is buffered by bicarbonate in the blood. Bicarbonate in turn is converted to water and CO₂, which is added to the metabolic CO₂ production and blown off at the lungs in order to maintain plasma pH at physiological levels. The result is a nonlinear increase in Vco₂ relative to Vo₂, detectable from the metabolic gas exchange data.²² The point at which this begins is termed the AT and can be identified by plotting Vo₂ against Vco₂, known as the V-slope method (Fig. 23.4). AT is recorded as the O₂ consumption at this point.

The identification of a discrete AT point is arbitrary as the increase in CO₂ production is gradual,²³ but it can nevertheless be reasonably reliably estimated by expired air analysis.²⁴ There is a high correlation between the AT and pVo₂,²⁵ making the anaerobic threshold a potentially useful submaximal measure of exercise capacity that is independent of patient motivation.²⁶

There are other methods of identifying the AT; a second method depends upon the changing ventilatory response to CO₂ during exercise.²⁷ The relationship of Ve to Vco₂ (ventilatory equivalent for CO₂ or Veqco₂) can also be plotted against time, and the nadir gives another estimation of the AT (Fig. 23.5).

The relationship of ventilation to CO₂ is linear throughout early exercise, but after the AT, an increase in ventilation occurs out of proportion to the production of CO₂. This was initially thought to be a consequence of the acidosis due to lactic acid,²⁸ but plasma pH and CO₂ remain stable,²⁹ and the reason for the change in the ratio remains elusive. Nevertheless, some laboratories use the lowest point in the ratio before the rise as the AT. This method is highly reproducible. The ‘crossing’ method, using the point at which the ratio of CO₂ production and O₂ consumption passes 1.0, allows an estimation of the AT in more patients than the other methods but gives a higher value.³⁰

During exercise, both CO₂ output and ventilation increase steadily. The relationship between the two is linear,³¹ but in patients with HF, the slope of the relationship is increased (the slope is steeper), so that for a given CO₂ production, there is more ventilation (Fig. 23.4).³² The Ve/Vco₂ slope is directly related to both mortality and morbidity.³³ Peak Vo₂ and the Ve/Vco₂ slope are inversely related to each other,^33,34 so that the more reduced the exercise capacity, the greater the ventilatory response to exercise (Fig. 23.6).

Additional variables can contribute further information. Cyclic fluctuations in ventilation, known as early oscillatory breathing (EOV), are exacerbated in patients with chronic HF both at rest and during exercise.^35,36 There are at least two definitions of cyclical breathing, but patients with EOV by either definition have a worse prognosis.³⁷

Although the relation between ventilation and O₂ uptake is not linear, the O₂ uptake efficiency slope (OUES) is derived by plotting Vo₂ as a function of log₁₀VE, which is an approximately linear relation.³⁸ The steeper the slope, the more O₂ is taken up for a given unit ventilation (Fig. 23.7). One advantage of the OUES is that it can be measured from submaximal data and does not depend upon reaching peak exertion. The OUES is predictive of prognosis even from submaximal tests.³⁹

Combining peak variables occasionally offers greater prognostic power. For example, peak cardiac power output, the product of cardiac output and mean arterial blood pressure at peak exercise, relates to exercise capacity and outcome.⁴⁰ However, not all combinations of peak variables are of greater predictive value than their constituent measurements.⁴¹

The cause of exercise intolerance and the abnormal physiology during exercise in chronic HF (such as the increased Ve/Vco₂ slope), and the relative contributions of central haemodynamics, peripheral vasculature, and skeletal muscle adaptations, are incompletely understood. Many variables, such as ventilation, cardiac power output, metabolic gas exchange kinetics,⁴² heart rate and, of course, time, increase during an exercise test as a function of exercise load (and exercise duration in an incremental test). Thus, the exercise measures and derived variables will all correlate highly with each other, and all will correlate to some degree with prognosis.⁴³ Which is the most ‘powerful’ predictor will vary from dataset to dataset by random chance. There is a temptation to plot one variable as a function of another, and assume that the mathematically dependent variable plotted on the y-axis is somehow determined by the variable on the x-axis: this is certainly not the case. Correlation does not imply cause and effect. An absurd example is the observation that hair length (if measured accurately) will also increase as a function of duration of an exercise test, and would correlate closely with peak Vo₂: however, increase in hair length is clearly not the determinant of peak Vo₂. A more controversial example is the influence of chronotropic incompetence and whether it is a limitation of heart rate increase that determines exercise capacity. Chronotropic incompetence is more frequent and more severe in patients taking β-blockers than in those not taking them, yet they do not induce a reduction in peak O₂ consumption.^44,45 It cannot be assumed that any variable derived from an exercise test determines exercise capacity. It makes as much sense to state that exercise capacity, for example, determines peak cardiac power output or peak heart rate.

The measurement of peak O₂ consumption (pVo₂) requires a form of exercise equipment, usually either a programmable treadmill or a stationary electromagnetically braked cycle, and a metabolic gas analyser. A 12-lead ECG monitor is used for safety and to determine heart rate. Ideally, the laboratory should be air-conditioned to achieve a stable temperature and humidity. Repeated exercise tests should take place at about the same time in the day, and not within 3 h of a main meal. Caffeine should be avoided for 6 h before a test. Subjects should be advised to wear comfortable shoes, and women should wear a loose shirt or vest under which the ECG electrodes can be applied to the skin.

All metabolic gas analysis systems must measure ventilation and analyse the concentrations of O₂ and CO₂ in expired air. In a typical system, subjects wear a mask or breathe through a mouthpiece (with a nose-clip) attached to a flow meter which allows the measurement of tidal volume (Vt) and frequency of ventilation (f  ). These are used to derive total minute ventilation (Ve). Connected to the flow meter is a sample tube which continuously aspirates a small volume of the expired and inspired air to measure the concentrations of O₂ and CO₂. By knowing Ve and the concentrations of the metabolic gases, volumes of CO₂ produced and O₂ consumed can be derived.

Mass spectrometry is the gold standard for gas analysis. It provides rapid and reliable online breath-by-breath assessment of O₂ and CO₂. Sampled gases are subjected to an electron beam and thereby converted to charged ions. These are then accelerated in an electric field and their direction altered by a magnetic field. Detectors produce an output according to the numbers of ions striking them. However, the equipment is expensive, and requires regular maintenance. Alternatives to mass spectrometry include paramagnetic analysers or gas chromatography linked to thermal conductivity detectors for O₂ analysis and infrared absorption detectors for CO₂.

The most common and maintenance-free devices for measuring flow are low-resistance turbines which are less affected by ambient temperature and humidity than alternative devices, but resistance and inertia of the vane can reduce sensitivity. Alternatives include pneumotachometers which measure a pressure drop across two capillary tubes, and anemometers which calculate air flow based upon the temperature change in a thin electrified wire stretched across the tube.

Although initial work was performed using intermittent bag collection of expired air, most equipment now measures ventilation and gas concentrations on a breath-by-breath basis. This requires a temporal adjustment of the gas analysis, which takes longer than the flow assessment, to maintain accuracy.⁴⁶

Small errors in sampling are magnified by the calculations performed during and after the test. The equipment should therefore be calibrated at least daily and preferably before each test. Mass spectrometry devices commonly use indicator dilution to give the flow signal, hence only one device needs to be calibrated to get all the signals of interest, whereas with the typical metabolic cart, there are three different devices—flow, O₂ and CO₂ meters. Exercise tests should be carried out and interpreted by experienced personnel, with appropriate support from equipment manufacturers. Equipment should be serviced regularly.

The procedure must be explained in full to each subject. When a maximal test is being undertaken, it should be made clear before the test starts that the aim is to assess peak exercise capacity and that the subject will be encouraged to exercise to their limit. During the test, they should be asked to score their symptoms of breathlessness or fatigue according to a recognized rating of perceived exertion scale (such as the Borg score, Table 23.1),⁴⁷ and they should be discouraged from talking during the test. Simple signals should be agreed upon, such as raising of the left hand to stop the test. The subject should receive standardized encouragement at the midpoint of each stage, such as ‘You’re doing well’, ‘Keep it up’.

Subjects respond differently to the mask or the mouthpiece. The mouthpiece commonly causes distress and hypersalivation, but is associated with fewer gas leaks. In contrast, masks are more comfortable but have a higher incidence of leaks and a larger dead space volume. Both options should be available.

Throughout the test, a 12-lead ECG should be monitored, and a hard copy is usually printed with each exercise stage or at appropriate intervals. Blood pressure and heart rate should also be recorded at these time points. Some devices allow intermittent determination of cardiac output by inert gas dilution. It is important to explain and demonstrate the procedure before the test.

Basic assessment of pulmonary status can be performed routinely on any subject undergoing an exercise test. Using the capnograph, and software within the metabolic cart, a maximal flow volume loop can recorded and forced expiratory volume in one second (FEV₁), forced vital capacity (FVC), and inspiratory volumes measured. The spirometry test should be judged to be satisfactory by the joining of the inspiratory and expiratory loops. The FEV₁ can be used to calculate MVV (= FEV₁ × 35).⁴⁸

Before exercise begins, 3–5 min of resting data should be collected. This allows the subject to become familiar with the mouthpiece or mask, and the investigators to identify leaks and analyser problems. Baseline O₂ consumption should be between 3 and 5 mL/kg per min and the respiratory exchange ratio below 0.9. Baseline heart rate and blood pressure should be measured at the end of this phase.

Protocols for both stationary cycle and treadmill exercise are available and the choice of equipment is often dictated by the experience of the personnel and the space available in the laboratory. Treadmill-based exercise is more natural and often leads to a higher peak O₂ consumption since subjects also carry their own weight. In contrast, cycle exercise allows haemodynamic monitoring and more gradual increments in workload. Whatever form of exercise is used, whether it consists of steady and progressive increases in work, or more stepwise increases, the rate of increase of work must be slow enough to allow an adaptation of the slow metabolic gas kinetics seen in chronic HF. The protocol should aim for an exercise time of between 6 and 12 min, as this maximizes peak Vo₂.⁴⁹

The Bruce protocol⁵⁰ modified by the addition of a ‘stage 0’ at onset consisting of 3 min of exercise at 1.61 km/h (1 mile/h) with a 5% gradient is often used (Table 23.2). However, the steps between each stage are large, such that many subjects stop exercise immediately after a stage begins. Alternatives are the Balke protocol⁵¹ which has a constant speed and gradual increase in incline (1% per minute), and the Naughton⁵² protocol which has shorter stages with smaller increments in both speed and incline. Each of these has been criticized for excessive duration of exercise. More rapidly incremental protocols are more discriminatory between patients, but do not allow steady state metabolic gas exchange to be reached. During the test, the subject should be encouraged not to use the handrails except as a guide and balance. Dependence on handrails alters the work being performed and the reliability of the O₂ uptake data.⁵³

Preparation of the subject for exercise on the stationary cycle proceeds as for treadmill exercise. The resistance (watts) against which the subject pedals is increased gradually until exhaustion. The ramp protocol is the most commonly employed and consists of continuous increments of workload aiming for 10 min of exercise.

Exercise tests should be symptom-limited maximal tests. At peak exercise, heart rate and blood pressure should be measured and an

ECG printed. The rating of perceived exertion and the reason for stopping should be noted. Once the subject has signalled exhaustion, cycle resistance should be reduced to 0 or the treadmill should be slowed to minimum with no incline, and cycling or walking should continue for a further minute or so. Monitoring should continue until heart rate and gas exchange have returned to resting values.

The point at which an increased load does not lead to an increase in O₂ consumption, is defined as ‘Vo_{2 max}’. In most untrained subjects (and particularly HF patients), maximal O₂ consumption is rarely reached before exercise is discontinued due to fatigue or dyspnoea.⁵⁴ The O₂ consumption at termination of exercise is peak Vo₂ (pVo₂).The point from which pVo₂ is calculated is important. An RER of less than 1 is regarded as a submaximal test and the data must be interpreted in light of this.⁵⁴

An RER consistently above 1.0 at the end of exercise suggests that close to peak exercise has been performed.⁵⁵ The prognostic value of peak O₂ consumption depends largely on the RER at peak.^21,56 The AT can be used to extrapolate the maximal O₂ consumption, but although useful, extrapolated Vo_{2 max} is less reliable than peak O₂ consumption.⁵⁷

The data from the last 10 s of exercise are averaged to give peak ventilation (Ve), O₂ consumption (Vo₂), and CO₂ production (Vco₂). The AT is also calculated (Box 23.1).

Peak symptom-limited exercise testing using either a treadmill or a bicycle in a controlled setting is safe. In the largest published series of exercise testing in over 6000 men assessed for possible ischaemic heart disease, there were no deaths and the incidence of ventricular tachycardia, defined as three or more consecutive beats, was 1.1%.⁵⁸ In another series of 289 patients with severe left ventricular dysfunction (〈35%), only one resuscitated cardiac arrest with ventricular fibrillation occurred, and nonsustained ventricular tachycardia occurred in 20%, with hypotension in 5%.⁵⁹ In a larger population of 607 patients enrolled in the Veterans Administration Cooperative Study-Vasodilator Heart Failure Trial (VHeFT), there was an incidence of arrhythmias necessitating termination of the test in only 1.6% of patients and no cardiac arrest or death.⁶⁰

Patients with chronic HF typically have a reduced exercise time, lower than predicted peak O₂ consumption, and an elevated slope of the relationship between ventilation and CO₂ output (Fig. 23.3). The Ve/Vco₂ slope and pVo₂ are inversely related to each other and independently related to prognosis (Figs 23.6 and 23.8).

The aetiology of the impaired exercise capacity in patients with chronic HF remains incompletely understood. The haemodynamic hypothesis suggests that in particular individuals there is a single dominant pathology, either increased pulmonary fluid leading to breathlessness or poor skeletal muscle perfusion causing fatigue. However, regardless of the limiting symptom experienced during an exercise test, objective measures of exercise tolerance are similar in patient cohorts whichever symptom is dominant,⁶¹ and the prognosis is related to the peak O₂ consumption achieved and not the nature of the symptoms experienced.⁶² Furthermore, the type of exercise performed seems to influence whether individuals suffer breathlessness or fatigue. Slowly incrementing tests and cycle exercise are more likely to lead to fatigue;⁶³ in contrast, rapidly incremental exercise tests more frequently lead to breathlessness, even if the workload is standardized. Cycle exercise is more often stopped by fatigue than breathlessness compared with treadmill exercise, even when the same level of exercise is performed.^64,65 These findings suggest that there is a common underlying pathology resulting in symptoms; and that the symptoms are variably reported by patients depending upon context.

Many studies have failed to show a significant link between exercise performance and left ventricular performance (Fig. 23.9).^66,–72 Furthermore, acute increases in resting cardiac contractility through inotropes^73,74 or cardiac transplantation,^75,76 increased cardiac output following mitral valvuloplasty,⁷⁷ and conversely, reductions in contractility and heart rate rise with β-blockers,⁴⁴ have no immediate impact on exercise capacity. In contrast, although exercise training can lead to impressive increases in exercise capacity which appears after a prolonged period of training, it has no effect on cardiac function.^78,79

Patients with chronic HF have a lower heart rate rise during exercise. This has been suggested as an aetiological factor in exercise intolerance in chronic HF. However, heart rate limitation seems not to reduce peak O₂ consumption in chronic HF patients,^44,45 and increasing heart rate during exercise has no beneficial effect.⁸⁰ It is therefore unlikely that chronotropic incompetence is a mediator, but rather a marker of impaired exercise capacity.^81,82 As such however, it is of course related to outcome in the same manner as other peak variables.

Many investigators have described abnormalities of pulmonary function in patients with chronic HF. Spirometric variables are variously reported to be abnormal in heart failure,⁸³ although large numbers of patients have normal spirometry.^84,85 Nevertheless, some spirometric indices do correlate with exercise capacity,⁸⁶ particularly in patients with more modest symptoms.⁸⁷

Large airways function is abnormal in some patients with chronic HF,⁸⁸ but although there is some reversibility of airways obstruction with nebulized β-agonists, this has no effect on exercise capacity.⁸⁹

Bronchoconstriction is a major component of acute pulmonary oedema.^90,91 Cabanes et al.⁹² suggested that there was an increase in bronchial reactivity in HF patients. This effect could be partially reversed with albuterol, a β-stimulant, and, paradoxically, with inhaled methoxamine, a bronchoconstrictor.⁹³ The hypothesis here was that if left atrial pressure increased as a result of exercise, the veins draining the distal bronchi into the left atrium might become congested, and methoxamine would improve exercise capacity by constricting these vessels. However, others have found no increase in bronchial hyperesponsiveness to stimulation with methacholine or sodium metabisulphite, no change in cough responsiveness to capsaicin, and no evidence for exercise-induced bronchospasm.⁹⁴

Patients with chronic HF also have impaired transalveolar diffusion as measured by transcapillary carbon monoxide diffusion (DLco).⁹⁵  DLco is directly related to exercise capacity,⁹⁶ and the increased ventilatory response during exercise,⁹⁷ and hence is a prognostic marker.⁹⁸ It can be improved by ACE inhibitors (an effect countered by aspirin),⁹⁹, exercise training,¹⁰⁰ and sildenafil,¹⁰¹ but not by ultrafiltration,¹⁰² suggesting that the cause of the impaired diffusion is vascular, possibly endothelial, rather than being related to lung blood volume or haemodynamic status.

Diaphragmatic weakness has been implicated in the symptomatology of chronic HF. The histological abnormalities seen in the skeletal muscle of HF patients (see ‘Skeletal muscle’, below) are also seen in the diaphragm.¹⁰³ The ability to generate negative intrathoracic pressure is slightly reduced in heart failure patients when compared with controls.¹⁰⁴ This weakness, combined with reduced lung compliance leading to increased diaphragmatic work particularly in the supine position,¹⁰⁵ suggests that respiratory muscle could contribute to exercise limitation. However, overall diaphragmatic strength at rest appears to be well preserved,^106,107 and contractile diaphragmatic fatigue is uncommon in chronic HF patients during an exercise test.¹⁰⁸

Finally, patients ventilate at the same proportion of their maximum ventilatory volume at peak exercise as control subjects, further refuting the suggestion that ventilatory capacity limits exercise in chronic HF.¹⁰⁹ However, as will be discussed in relation to skeletal muscle generally below, diaphragmatic fatigue causes sympathetic activation and a reduction in leg blood flow.^110,111 These changes could potentially contribute to the cycle of impaired muscle perfusion, worsening of the skeletal myopathy, and the abnormal ventilatory response to exercise.

Ventilation–perfusion matching in the lung is commonly abnormal in patients with chronic HF. Ventilation, V, and perfusion, Q, are ideally equal, leading to a V/Q ratio of 1.0 in an idealized lung unit. Where perfusion is greater than ventilation (V/Q 〈 1.0), there is effective shunting of venous blood; where ventilation is greater than perfusion (V/Q 〉 1.0), there is ‘wasted ventilation’, or dead space. At rest in upright normal individuals, for example, there is a gradient of V/Q ratio from apex to base of the lung leading to dead space in the apices.

Increased dead space may be a cause of the elevated Ve/Vco₂ slope. Equation 23.1 suggests that when Ve/Vco₂ is increased, if Paco₂ is constant, then Vd/Vt must be increased. However, while this may be true, it does not imply that increased dead space ventilation is the cause of the increased Ve/Vco₂ slope. Several lines of evidence suggest that it may not be. Firstly, the exercise hyperventilation seen during exercise in chronic HF patients is present from the outset of exercise,^112,113 whereas any abnormalities in blood chemistry and increases in dead space ventilation would worsen as exercise progresses. Secondly, although during submaximal exercise Vd/Vt is higher in patients than controls, at peak exercise controls have a higher absolute dead space ventilation than patients.⁸⁵ Furthermore, some data suggest that apical lung perfusion is increased in heart failure patients at rest compared with controls¹¹⁴ (which leads to better V/Q matching) and it does not deteriorate during exercise.¹¹⁴ Thirdly, β-blockade with carvedilol reduces hyperventilation without changing Vd/Vt.¹⁸ Finally, correcting for the increased dead space ventilation due to breathing frequency in chronic HF patients does not weaken the correlation between the increased ventilatory response to exercise and peak O₂ consumption,³² and increasing anatomical dead space in normal individuals does not cause an increase in the Ve/Vco₂ slope.¹¹⁵

A conceptual problem with all hypotheses for the origin of the increased Ve/Vco₂ in HF that rely on pulmonary pathology is the absence of a physiological signal. Arterial blood gas tensions are normal or supranormal during exercise in patients with HF with a higher Pao₂ and lower Paco₂ than in normal subjects.^116,117 Indeed, where blood gas tensions are abnormal, there is usually another explanation for the symptoms of breathlessness.¹¹⁷ There is no dead space ventilation detector, and how increased dead space ventilation would be communicated to the respiratory centres in order to stimulate ‘compensatory’ hyperventilation is unclear. On the other hand, if there were a stimulus to excess ventilation arising outside the lungs, there would be a tendency to an excessive fall in arterial CO₂. Increasing dead space ventilation might thus be a physiological response to prevent even greater reductions in Paco₂.

In chronic HF, despite impaired ventricular function, the blood pressure response to exercise is not usually abnormal (particularly at submaximal workloads),¹¹⁸ as a consequence of increased peripheral resistance.^119,120 The increased peripheral vascular resistance is a result of chronically increased sympathetic tone and renin–angiotensin system activation. There is arterial smooth muscle hypertrophy and activation of fibroblasts with hyalinosis of the vascular wall.¹²¹ These changes lead to reduced arterial compliance¹²² and hence to poor vasodilatation in skeletal muscle arterioles during exercise.¹²³

In addition, there is a reduced response to endogenous vasodilatory stimuli,^124,125 to infused hyperosmolar solutions,¹²⁶ and to pharmacological agents.¹²⁷ There are also increased in levels of endothelin, a powerful vasoconstrictor.¹²⁸

The skeletal muscles are abnormal in patients with chronic HF. There is general reduction in muscle bulk early in the course of heart failure,¹³⁸ and there is a shift from type I (slow twitch) to type IIb (fast twitch) fibres within skeletal muscles.¹²⁹ Type IIb fibres are more easily fatiguable and have less aerobic capacity than type I fibres. Capillary density is also reduced.¹³⁰ Hence there is an earlier need for anaerobic metabolism during exercise than normal. It is also becoming clear that perfusion within muscles and fibre recruitment may be impaired in chronic HF,¹³¹ and muscle strength is also reduced,¹³² as is endurance.¹³³ Exercise capacity in heart failure patients is related to both muscle strength and bulk,^134,–136, and the reduction in endurance correlates with exercise performance.¹³⁷

From a patient’s point of view, the ability to perform repeated submaximal exercise (endurance) may be more important than peak force generation (strength), and early quadriceps fatiguability has been reported.^133,138 Fatigue is independent of acute changes in blood flow^163,139 and of central factors.¹⁷⁰ Fatiguability has been shown in a very small muscle group¹⁶⁷ in which blood flow is unlikely to be limited by cardiac reserve, suggesting that intrinsic muscle factors mediate fatiguability of muscle. It is not difficult to imagine how such abnormal muscles might lead to fatigue, but it is less clear how they might lead to breathlessness.

Ergoreceptors are muscle receptors sensitive to work. Stimulation of the receptors during exercise leads to ventilation and sympathetic activation—the ergoreflex. The degree of activation is in part related to the work performed per unit of muscle mass and to the metabolic state of the muscle. In normal subjects, for example, the ventilatory response to a given workload is much greater if the work is performed by arms rather than legs.¹⁴⁰ Furthermore, when normal subjects exercise with cuffs inflated to suprasystolic pressure, thus preventing the wash-out of the metabolic products of exercise, the ventilatory response to exercise is greater than in the control situation.¹⁴⁰

The ergoreflex can be quantified by experiments involving cuff inflation around an exercising limb. During exercise in a normal subject, ventilation progressively increases with an increase in vascular resistance in the nonexercising limbs. There is a swift return to normal at the end of exercise. If a cuff is inflated at peak exercise proximal to the exercising muscle, the muscle is ‘frozen’ in its exercising state.¹⁴¹ The consequent persisting increase in ventilation compared with the control response is the ergoreflex response.¹⁴² In HF patients, presumably as a result of skeletal muscle abnormalities, the ergoreflex response generated by postexercise regional circulatory occlusion is much greater, both in terms of haemodynamic responses and in terms of the ventilatory response.^143,144 The reflex is sensitive to metabolic stimulation rather than movement.¹⁴⁵ The increased ventilatory response to exercise is proportional to the increased ergoreflex activity,¹⁴³ and a training programme can reduce the contribution of the ergoreflex to ventilation,¹⁴³ presumably by increasing the quality of the skeletal muscle.

The precise exercise product responsible for triggering the ergoreflex is not clear. Possibilities include potassium, particular as arterial potassium rises on exercise,^146,147 mirroring closely the increase in ventilation.¹⁴⁸ However, potassium rises during exercise do not demonstrate an inflection similar to that seen in the rise in ventilation during heavy exercise.¹⁴⁹ Furthermore, β-blockers lead to a greater rise in arterial potassium for a given workload, but do not increase minute ventilation.¹⁵⁰ Another possible contributor is prostaglandin production, as local prostaglandin levels in exercising muscle correlate with the ventilatory response.¹⁵¹ However, patients taking regular aspirin, a powerful inhibitor of inflammatory prostaglandin production, do not demonstrate a clinically significant difference in ventilatory response to exercise compared with those not taking aspirin.¹⁵²

In addition to a metabolic response to exercise, there are suggestions that nerve fibres sensitive to stretch might be involved in the response to ventilation. Muscle sympathetic nerve activity (MSNA) is related to the Ve/Vco₂ slope in chronic HF patients,¹⁵³ and passive limb movement leads to an increase in MSNA in chronic HF patients but not controls,¹⁵⁴ proportional to heart failure severity.¹⁵⁵ This suggests the existence of a mechanoreflex in addition to the metaboreflex.

The concept that peripheral muscle receptors sensitive to work, whether they be metabo- or mechanoreceptors, contribute to ventilation, and that this reflex is abnormal in patients with chronic HF, provides an elegant and unifying hypothesis explaining both breathlessness and fatigue in these patients.¹⁵⁶ On exercise, patients with heart failure have weak, structurally abnormal skeletal muscles, which are performing more work per unit muscle volume than normal muscle. In turn, this results in greater production of the metabolic products of exercise in the context of impaired perfusion. This leads to an increased stimulation of ergoreceptors sensitive to metabolic products, which stimulate increased ventilation relative to CO₂ production thus increasing the Ve/Vco₂ slope and reducing Paco₂.

The increased sympathetic activation of chronic HF may also contribute to the ventilatory response to exercise. Acute and chronic β-blockade, although having little effect on maximal¹⁵⁷ or submaximal exercise capacity,^158,159 reduces submaximal and maximal ventilation during exercise, and also reduces the Ve/Vco₂ slope.^44,45 On the other hand, increasing presynaptic catecholamine levels leads to an increase in ventilation in normal subjects, perhaps by sensitizing the ergoreceptors.¹⁶⁰

In subjects referred for investigation of symptoms of breathlessness and fatigue, a carefully performed exercise test provides an objective measure of exercise capacity. Cardiopulmonary exercise testing gives an overall assessment of the severity of the pathophysiology of chronic HF, integrating skeletal muscle, lung, endothelial, and cardiac function.

Perhaps more than other single assessment, peak Vo₂ is as a sensitive predictor of morbidity and mortality.^82,161 A peak Vo₂ of less than 14 mL/kg per min has been used as a functional criterion for selecting patients for cardiac transplantation.¹⁶² However, since the incorporation of β-blockers into routine management, patients have a better outlook for a given exercise capacity, and a pVo₂ of around 10–12 mL/kg per min is increasingly used as the cut-off identifying patients at especially high risk who might therefore benefit from transplantation.¹⁶³ Nevertheless, the value of pVo₂ as a risk assessment tool remains in the presence of optimal medical therapy,^164,–166 and referral for transplant assessment can safely be deferred in patients with a pVo₂ greater than 18mL/kg per min.

More recently, it has become clear that an elevated Ve/Vco₂ slope provides additional and independent prognostic information.¹⁶⁷ Combining peak O₂ consumption and Ve/Vco₂ slope improves risk assessment in some¹⁶⁸ but not all studies.⁴¹

Cardiopulmonary exercise testing can not only aid with prognostic assessment but can also assist in establishing the cause of symptoms of exercise intolerance in an individual. The most common differential diagnoses are chronic obstructive pulmonary disease (COPD), obesity, psychogenic hyperventilation, and poor effort. Data collected during a routine cardiopulmonary exercise test can add to clinical information and data from noninvasive imaging. Table 23.3 shows the cardiopulmonary exercise test variables seen with important differential diagnoses.

With aggressive neurohormonal blockade, appropriate device therapy, and careful hospital and community-based management programmes, mortality rates of less than 10% and hospitalization rates of less than 20% per year can be achieved for many patients with chronic HF. Contemporary randomized, placebo-controlled studies exploring the effects of new therapies on mortality and morbidity therefore require ever-increasing numbers of subjects. As a noninvasive, reproducible, and repeatable surrogate for outcome and an objective measure of heart failure severity, exercise testing with or without metabolic gas exchange is increasingly used as an endpoint in studies in chronic HF. Improvements in exercise capacity in smaller studies can also help plan larger mortality studies. So, for example, early studies of cardiac resynchronization therapy showed improvements on in exercise capacity,^169,–171 allowing better planning of the CARE-HF study using the earlier studies as a guide.¹⁷²

In addition to helping with diagnosis in patients presenting with symptoms of exercise intolerance, cardiopulmonary metabolic exercise testing (CPET) provides reliable and objective information on prognosis in patients with chronic HF and can be used to stratify patients requiring transplant assessment. CPET is noninvasive and low risk and can be easily repeated in order to monitor responses to therapy, in both daily clinical practice and research settings. However, the test must be carried out and interpreted in a controlled manner by experienced staff.