The lungs lie in series with the heart, share a common surface area, and are exposed to similar intrathoracic pressure changes. Hemodynamically, the lungs accept nearly all of the cardiac output and have afferent pathways that may modulate heart rate. Conversely, there are cardiac and vascular receptors that modulate ventilation and common receptors to both systems that are influenced by the activation of the rennin-angiotensin-adrenergic systems. Thus, it is apparent that as the disease process of chronic heart failure (HF) progressěs, the pulmonary system is appreciably altered. As a result, there are marked adaptations in the way HF patients breathe during exercise. The underlying premise of this review is that the pulmonary system begins to play a major role in the exercise intolerance associated with HF. To this end, we will briefly summarize the major changes in the pulmonary system in patients with stable, moderate to severe (New York Heart Association [NYHA] classes II and III) systolic HF and review the respiratory system adaptations or maladaptations to exercise. Less appreciated causes for the loss of lung function (e.g., due to genetic variation or alterations in cardiac size) and abnormalities in response to exercise (e.g., atypical regulation of end-expiratory lung volume [EELV] and exercise-induced pulmonary edema) will be discussed.
HF-INDUCED CHANGES IN THE PULMONARY SYSTEM AT REST
Resting Pulmonary Function
Figure 1 summarizes the major alterations in pulmonary function of patients with HF. The pathogenesis of HF is typically initiated with the onset of reduced cardiac output and a rise in left atrial pressure or pulmonary wedge pressure that leads to a subsequent rise in pulmonary venous pressure and pulmonary arterial pressure. Although this scenario causes the initial rise in pulmonary vascular pressures, there is some evidence for a reactive component from the pulmonary blood vessels leading to variable increases in pulmonary vascular resistance, which is not always linked to the severity of systolic dysfunction. A number of mediators related to the activation of the renin-angiotensin-adrenergic systems (e.g., angiotensin II, endothelin 1) in the setting of genetic variation or periods of mild sustained or intermittent hypoxia (e.g., ventilation-perfusion inhomogeneities or central sleep apnea, respectively) likely contribute to the variability in the degree of secondary pulmonary hypertension. The chronically elevated pulmonary venous pressures and mediators related to the activation of the renin-angiotensin-adrenergic systems appear to alter the alveolar-capillary membrane and cause remodeling of the vessel walls of the pulmonary circulation. These changes coupled with gradual increases in heart size, subclinical interstitial pulmonary edema, and respiratory muscle weakness lead to reduced lung volumes, a decrease in lung compliance, and the restrictive changes commonly observed in this population. Figure 2 shows the mean maximal flow-volume envelope of chronic HF patients (n = 11, age = 62 yrs, left ventricular ejection fraction = 24%, NYHA classes II and III, minimal smoking history) relative to age- and sex-matched control participants (5). Importantly, the capacity for increasing flow and volume is reduced in HF, although traditionally, a ventilatory mechanical limitation to breathing during exercise has not been considered an important limiting factor in these patients (see exercise lung mechanics). A less appreciated finding is the role that cardiac volume within a closed thoracic cavity may have on the more compliant lungs. Figure 3 shows the relationship of heart size (relative to total thoracic volume) to the observed changes in total lung capacity, in a cohort of 44 HF patients (NYHA classes I-IV) (11). In this study, cardiac volume contributed significantly to the reduction in lung volume. Heart size is clearly linked to the severity of disease, and thus, it is likely that as cardiac volume increases, there are also secondary changes in pulmonary venous pressure and congestion that further contribute to this relationship.
There is a debate over the degree of obstructive changes observed in HF. Clearly, obstructive changes are observed with acute HF; however, these changes are not as apparent in stable patients where smoking history, previous coronary bypass surgery, and obesity are carefully controlled. As HF progresses and true lung congestion occurs (causing bronchial congestion as well as the potential for interstitial edema), there appears to be some compression of the airways. The response to methacholine (used to test for airway reactivity) may also be enhanced in HF that in turn contributes to the reduced expiratory flows. Another potential cause of obstructive changes may be related to desensitization of receptors in the airways responsible for maintaining bronchial tone, such as the β2-adrenergic receptors, due to chronically elevated catecholamine levels.
In addition to the restrictive and mild obstructive lung changes that occur in HF, there is a fall in the lung diffusing capacity for carbon monoxide (DLCO, an index of the functional surface area of the lungs for gas exchange) due primarily to a fall in alveolar-capillary conductance, but which may also be due to a decline in pulmonary capillary blood volume secondary to a fall in cardiac output (pulmonary blood flow) or altered perfusion in well-ventilated regions of the lungs, despite the rise in pulmonary wedge pressure. Ventilation and perfusion inhomogeneities as well as anemia (common in HF) may also contribute to a reduction in the resting DLCO, although DLCO is typically corrected for hemoglobin. Aggressive management of HF with angiotensin-converting enzyme (ACE) inhibitors or aldosterone inhibitors has been shown to improve DLCO. However, the improvements in DLCO after heart transplant are variable, consistent with chronic alterations at the alveolar-capillary membrane that may be difficult to reverse. The DLCO is important in HF because it correlates with ventilatory efficiency (increased ventilation for a given metabolic demand, VE/VCO2) as well as exercise capacity (peak V˙O2), both demonstrating significant prognostic value (4).
Pulmonary Function and Genetic Susceptibility
Recent studies have suggested that deterioration in resting pulmonary function with HF may also be influenced by genetic variation (1,14). The gene encoding ACE exhibits an insertion/deletion polymorphism resulting in 3 genotypes (DD, ID, and II), which affects serum and tissue ACE activity as well as other vasoactive substances (e.g., bradykinin). Patients with the DD genotype appear to have the highest serum ACE activity and plasma angiotensin II levels with significantly reduced forced vital capacity, forced expiratory volume in 1 s, and DLCO relative to the II genotype, despite chronic ACE suppression (1). Similarly, a polymorphism of the β2-adrenergic receptor (amino acid 16, where an individual can be homozygous for arginine, glycine, or heterozygous) may influence agonist-mediated desensitization of the receptor or receptor density in airways or the pulmonary vasculature (14). HF patients homozygous for arginine at this position appear to have greater deterioration in pulmonary function relative to HF patients homozygous for glycine at this amino acid. Larger studies will be necessary to confirm these findings and to determine if there are combinations of genes that may increase susceptibility to pulmonary function changes in the HF population.
Resting Gas Exchange
Arterial oxygen saturation (SaO2) in resting, clinically stable HF patients is typically normal or may be at the low end of the normal range and may deteriorate further in sicker patients, patients with coexisting lung disease, or those with acute HF exacerbations. Interestingly, the partial pressure of oxygen in the arterial blood (PaO2) may be somewhat reduced, despite relatively normal SaO2 values. This is due to the shape of the O2 dissociation curve and most likely a small leftward shift due to respiratory alkalosis, although there is evidence to suggest that 2,3-diphosphoglycerate may increase in HF as a result of venous hypoxia that would tend to shift this curve back to the right.
A hallmark of HF is the reduced partial pressure of arterial carbon dioxide (PaCO2) because of chronic hyperventilation (15). There are a number of afferent neural pathways that may alter ventilatory drive contributing to the chronic hyperventilation. These afferent pathways include vascular receptors located in the right and left atrium, right ventricle, and pulmonary arteries; mechanoreceptors located in the respiratory muscles of the chest wall; pulmonary receptors including visceral receptors in the bronchi; receptors near the alveoli and pulmonary capillaries (e.g., C fibers); and juxtapulmonary receptors (Jreceptors), which respond to mild congestion, heightened activity from skeletal muscle (e.g., chemoreceptors or metaboreceptors), and input from both central and peripheral chemoreceptors due to increased catecholamines or other augmented mediators in HF. In addition, mild carotid body hypoxia (due to hypoperfusion, slowed circulation times, or the mildly reduced PaO2 values) may contribute to the augmented ventilatory drive.
Other contributors to altered gas exchange at rest include mild ventilation-perfusion inhomogeneities in the lungs that may influence dead space ventilation and the alveolar to arterial oxygen difference. Dead space ventilation can also be influenced by a more rapid-shallow breathing pattern, so that the dead space to tidal volume ratio increases (15). The combination of an increased dead space ventilation and an increased drive to breathe leads to a poorer ventilatory efficiency, defined by the minute ventilation relative to the carbon dioxide produced (VE/VCO2) or O2 consumed (VE/V˙O2). The abnormalities in gas exchange may worsen in the supine position as venous return increases and central blood volume rises, resulting in an increase in left atrial pressure and increased congestion in the pulmonary and bronchial circulations.
EXERCISE RESPONSES IN HF
Figure 4 provides an overview of the changes in the pulmonary system associated with HF and the resultant impact on lung mechanics, gas exchange, and ventilatory control at rest and during exercise. Whole-body exercise requires a host of adaptive mechanisms that challenge the HF patient. Most important is the need to increase cardiac output to meet the metabolic demands of the locomotor muscles, at the same time maintaining alveolar ventilation to preserve blood gas homeostasis with minimal cost to the respiratory muscles. This becomes more challenging as ventilatory demands increase and disease severity progresses, leading to further alterations in lung mechanics, gas exchange, and ventilatory control mechanisms and ultimately to competition for blood flow between the respiratory and locomotor muscles.
During exercise, patients with HF breathe more rapidly with a reduced tidal volume that may accentuate as intensity increases or with disease severity (5,15). This tachypneic pattern during exercise has been related to lung stiffness and is achieved with an increase in duty cycle because of a greater than normal decrease in expiratory time. Healthy adults increase both tidal volume and frequency early in exercise with primarily an increase in tidal volume up to 50% to 60% of the vital capacity as exercise continues. Conversely, patients with HF have a smaller vital capacity at rest and use a smaller percentage of this capacity during exercise, often plateauing at a tidal volume closer to 40% of the vital capacity (5).
Figure 5 shows the flow-volume response to progressive exercise in a group of classes II and III HF patients (right) along with age- and sex-matched untrained healthy adults (left) (5). The tidal breaths are plotted within the maximal flow-volume loop obtained before exercise and immediately after exercise. The healthy adult responds to increasing ventilatory demands by encroaching on both the inspiratory and expiratory reserve volumes. This keeps the tidal breath along the linear portion of the pressure-volume relationship of the lungs and chest wall and minimizes the increase in the work and cost of breathing. The fall in EELV is also important because it improves the length-tension relationship of the inspiratory muscles so that for a given neural input to these muscles, there is greater force generation. The expiratory airflows during exercise in the healthy participants generally fall within the maximal airflows defined by the maximal flow-volume envelope, even during heavy exercise. In HF, however, even at rest, EELV tends to be near residual volume. With increasing ventilatory demands, the tidal volume is increased by encroaching primarily on the inspiratory reserve volume. Figure 6 shows a more exaggerated example of the flow-volume and pressure-volume response to exercise in an NYHA class III HF patient. Despite significant room to increase the tidal breath and ventilation by encroaching further on the inspiratory reserve volume, this patient continues to breathe at extremely low lung volumes, so that the majority of the expiratory flow produced during tidal breathing meets or exceeds the maximal available expiratory flows. This patient also produces expiratory pleural pressures that exceed the maximal effective pressure (pressure that produces the highest flow, Pmaxe), resulting in a significant amount of wasted expiratory effort. The most likely reason for an altered breathing strategy such as this in HF patients (outside of obesity, which is associated with a similar breathing strategy) may relate to the stiff lungs and avoidance of an (inspiratory) elastic load that would significantly increase the work and cost of breathing. Interestingly, healthy adults (Fig. 5) also demonstrated a large exercise-induced bronchodilation that was not observed in HF. Although speculative, the lack of a bronchodilation in the HF group may be because of desensitized receptors (e.g., β2-adrenergic receptors), from high catecholamine levels or small increases in bronchial blood volume due to progressive increases in pulmonary venous pressure with exercise.
Although HF patients appear to avoid a high elastic load to breathing by reducing the EELV and breathing rapid and shallow, there is still evidence that the work and cost of breathing are increased. Using a classic measure of diaphragmatic work, the tension time index, HF patients were shown to produce markedly higher pressures in relationship to the pressure generating capacity of the diaphragm, and this pressure approached a threshold that increases the likelihood for fatigue (7). The capacity for respiratory muscle force generation is dependent on muscle length (lung volume) and velocity of muscle shortening (inspiratory flow) and thus falls during the increased ventilatory demands of exercise. There is also evidence of respiratory muscle weakness in HF as well as exercise-induced deoxygenation of the accessory respiratory muscles with light activity. However, there is also histochemical and biochemical evidence that the diaphragm adapts to the higher work demands, and it does not appear that there is significant diaphragmatic fatigue with exercise. This would suggest that diaphragmatic fatigue is an unlikely contributor to exercise intolerance in the HF population.
Competition Between Competing Muscle Beds
The elevation in respiratory muscle work would increase competition for blood flow between the respiratory and working locomotor muscles. This competition is enhanced in HF with the limited blood flow reserve, despite evidence that HF patients may more aggressively vasoconstrict nonactive beds. In an animal model of HF, the diaphragm has been shown to preferentially recruit blood flow from working locomotor muscles, whereas in healthy humans, an inspiratory load will cause a reflex reduction in leg blood flow (8). O'Donnell et al. (9) compared exercise time and symptoms during exercise in HF patients who performed cycle ergometry under three conditions: continuous positive airway pressure (CPAP), proportional assist ventilation (PAV), and with no form of ventilatory support (control). They found that inspiratory unloading (CPAP or PAV) increased exercise time and reduced leg fatigue, suggesting that unloading the respiratory muscles may result in enhanced leg blood flow, which in turn reduces symptoms of leg fatigue and allows for increased exercise performance (Fig. 7). Other studies have used low-density gas to reduce the work and cost of breathing in HF patients and also have found improved exercise endurance (6). The latter finding is interesting given that the major source of ventilatory work is elastic rather than flow-resistive work. Although not previously tested, it may be that the low-density gases reduce expiratory flow limitation and subsequently the expiratory pressure work along with a modest reduction in flow-resistive work.
Gas Exchange and Ventilatory Control
Although DLCO is usually only mild to moderately reduced in HF, some patients with long-standing disease have markedly reduced values, to the point where it is remarkable that these patients do not demonstrate more significant arterial oxygen desaturation during exercise. The majority of studies, however, examining blood gases in clinically stable HF patients have not observed exercise-induced reductions in SaO2. This is likely because DLCO is linked to disease severity and sicker patients have poorer cardiac output (4). Thus, despite the limited surface area for gas exchange, the low cardiac output in the setting of prolonged circulation times assures adequate time for end capillary oxygen diffusion and general maintenance of SaO2.
The most prominent gas exchange abnormality observed in HF influencing exercise ventilation appears to be the high dead space ventilation (15). Some studies have observed dead space to tidal volume ratio values that are almost twice their normal healthy counterpart. Such a high dead space to tidal volume ratio in the HF population can result in up to half of every breath devoted to dead space ventilation and thus having no role in gas exchange. This elevated dead space to tidal volume ratio combined with a chronic mild hyperventilation results in a high VE/VCO2 ratio, which may be exaggerated during exercise (Fig. 8). The VE/VCO2 slope is often calculated by plotting VE versus VCO2 over the course of light to moderate exercise (close to the ventilatory or anaerobic threshold). Although some studies discriminate between the two measures, the VE/VCO2 ratio is similar to the more labor-intensive measure of slope near the ventilatory or "anaerobic threshold" or the nadir of VE/VCO2 in an incremental exercise test. A high VE/VCO2 slope has been shown to be as prognostic or more so than the commonly reported peak oxygen consumption, although it also correlates with this measure in HF (12).
EXERCISE AND HF: UNRESOLVED ISSUES IN EXERCISE LIMITATION AND CARDIORESPIRATORY INTERACTIONS
Does Interstitial Edema Develop During Exercise in HF Patients?
It remains controversial if a subclinical interstitial edema exists in relatively stable NYHA classes II and III patients at rest or during exercise. Given the high pulmonary wedge pressure associated with HF and the common occurrence of pulmonary edema with acute decompensation, it seems likely that the lungs may be in a chronic state of altered lung fluid balance. However, Agostoni et al. (2) assessed DLCO and its components (alveolar-capillary conductance and pulmonary capillary blood volume) before and after ultrafiltration, to remove fluid, and did not observe significant changes in these measures. These authors suggested that in typically stable HF patients at rest, the lungs are dry and that the reduced alveolar-capillary conductance must be related to chronic changes at the alveolar-capillary membrane. A number of studies have examined the changes in pulmonary wedge pressure with exercise and have not found a significant relationship with exercise capacity or the symptoms of dyspnea. However, the development of interstitial edema is dependent not only on factors that contribute to fluid accumulation (i.e., pulmonary venous pressure), but also on factors influencing fluid removal. Fluid is typically removed through lymphatic drainage. In addition, there is likely a small flux of fluid even in the normal lung across the alveolus that is removed through epithelial sodium channels via stimulation of the β2-adrenergic receptors. It remains unclear how these pathways of fluid removal may be altered in the HF population and if this makes this population more or less susceptible to the development of interstitial pulmonary edema, particularly during exercise.
One problem with determining if interstitial pulmonary edema occurs during exercise is the difficulty in measuring changes in lung water. Theoretically, DLCO should be sensitive to a widening of the alveolar-capillary membrane. Alveolar-capillary conductance should rise in concert with pulmonary capillary blood volume as cardiac output rises during exercise. However, the ratio of alveolar-capillary conductance to pulmonary capillary blood volume has been shown to fall early after exercise in patients with HF. Similarly, we have observed a blunted rise in DLCO with exercise and a subsequent plateau or slight fall in DLCO as exercise progresses (10). Figure 9 shows DLCO relative to cardiac output during exercise in HF patients and healthy age- and sex-matched participants. This ratio initially falls early in exercise, consistent with a greater increase in cardiac output as opposed to an increase in lung surface area; however, as cardiac output continues to rise, there is a point at which lung surface area begins to increase proportionately, suggesting the recruitment or distension of capillaries. Interestingly, with continued exercise in HF patients, the ratio of DLCO relative to cardiac output falls, suggesting that lung surface area for diffusion has reached a plateau or fallen, whereas it continues to be maintained in the healthy population. The small difference in DLCO/cardiac output slope, coupled with a drop observed in the HF population, may be evidence for mild interstitial pulmonary edema developing during exercise that resolves relatively quickly during the cool-down recovery phase. Enhanced lymphatic flow due to increased intrathoracic pressure swings, tidal volume, and breathing frequency combined with exercise-related catecholamine stimulation of epithelial sodium channels through the β2-adrenergic receptors may participate in rapid fluid removal. Whether the development of interstitial edema contributes to the exaggerated symptoms of dyspnea commonly associated with exercise in the HF population is not clear; however, it would most likely result in J-receptor stimulation, stiffer lungs, a more tachypneic pattern of breathing, and a higher VE/VCO2 ratio, all factors that would contribute to an increased work and cost of breathing as well as heightened symptoms during exercise.
Pulmonary and Cardiovascular System Interactions
With breathing, particularly during exercise, there are large dynamic changes in intrathoracic pressure, lung volume, and respiratory rate and changes in lung mechanics that have been shown to influence cardiovascular function. The fact that the lungs and heart share a common surface area within a fixed thoracic cavity suggests that lung inflation may influence cardiac filling. The degree of lung inflation is dependent not only on changes in tidal volume, but also on the regulation of EELV. Although it is unlikely that lung inflation significantly influences cardiac filling in healthy individuals, where the lungs are much more compliant than the heart, in HF (i.e., increased lung stiffness and cardiac hypertrophy), this potential exists. In addition, driving each change in tidal volume is the swing in pleural pressure. A more negative swing in pressure would normally drop intrathoracic pressure and increase the gradient for venous return while at the same time increasing cardiac afterload. Conversely, a positive swing in pleural pressure (e.g., Fig. 6, left) would increase intrathoracic pressure and reduce the gradient for venous return and in turn reduce cardiac afterload. Classic studies have suggested that in healthy individuals, the heart is more sensitive to factors that alter preload and less sensitive to factors that alter afterload. Importantly, the converse is true in HF patients. Thus, avoiding large negative swings in intrathoracic pressure (such as during unloading with CPAP or PAP) may improve cardiac output in HF. Although speculative, positive swings in pleural pressure may actually further reduce afterload. These effects are complicated by the timing of the cardiac cycle (systole vs diastole) within the ventilatory cycle (inspiration vs expiration). There also appears to be ventricular interdependence (amount of blood flow into one ventricle is dependent on the amount of blood flow into the other ventricle) that may be accentuated with large swings in lung inflation and intrathoracic pressure during exercise and as the heart enlarges in HF (13). Finally, rapid shallow breathing may accentuate sympathetic nerve activity, causing a release of catecholamines, whereas slow deeper breathing would reduce this activity. Thus, to some extent, the augmented ventilatory drive and tachypneic pattern of breathing in HF may contribute to the pathophysiology of the disease.
Figure 10 shows an example (preliminary data from our laboratory) of the ventilatory response to exercise in a CHF patient who exhibits phasic oscillations in ventilation. This appears typically because of oscillations in tidal volume in most patients; however, some will demonstrate oscillations in respiratory rate as well. The amplitude and cycle length in the ventilatory oscillations may vary across patients with an average cycle length at rest of approximately 60 s. Although the example in Figure 10 shows persistent oscillations in ventilation with exercise, often the amplitude will diminish or the phasic pattern will disappear as exercise intensity increases. Whether the oscillations represent pure alterations in ventilatory control or are linked to cardiorespiratory interactions is not entirely clear, because it is difficult to tease out the influence of one system on the other. This likely represents ventilatory instability caused by a combination of altered ventilatory drive in the setting of slowed circulation times, although there may be a component related to a primary central oscillation similar to Mayer waves that also are known to alter sympathetic outflow in an oscillatory manner. Others have suggested that there are primary alterations in cardiac output that contribute to variation in the ventilatory pattern. The latter suggestion is based on similar oscillations in V˙O2 and the phasic differences between V˙O2 oscillations with ventilatory oscillations. In addition, it has been suggested that cardiac output may vary beat by beat; however, this could simply be changing based on the respiratory alterations in cardiac output due to changes in tidal volume and pleural pressure. The presence of oscillations in breathing in HF patients during wakefulness or during exercise has been associated with poor prognosis and may be more prognostic than the periodic breathing observed during sleep (central sleep apnea) (3).
The intimate relationship between the heart and lungs is enhanced in patients with HF primarily due to alterations in cardiac size, increased pulmonary venous pressure, and neurohumoral activation. Thus, baseline lung function and the ventilatory responses to exercise are significantly altered. These changes in the pulmonary system appear to parallel disease severity and thus provide insight on disease status and prognosis. Moreover, we submit that a major cause of exercise intolerance in patients with HF is related to both chronic and acute alterations related to the pulmonary system.
1. Abraham, M.R., L.J. Olson, M.J. Joyner, S.T. Turner, K.C. Beck, andB.D. Johnson. Angiotensin-converting enzyme genotype modulates pulmonary function and exercise capacity in treated patients with congestive stable heart failure. Circulation
2. Agostoni, P.G., M. Guazzi, M. Bussotti, M. Grazi, P. Palermo, and G.Marenzi. Lack of improvement of lung diffusing capacity following fluid withdrawal by ultrafiltration in chronic heart failure. J. Am. Coll. Cardiol.
3. Corra, U., A. Giordano, E. Bosimini, A. Mezzani, M. Piepoli, A.J. Coats, and P. Giannuzzi. Oscillatory ventilation
during exercise in patients with chronic heart failure: clinical correlates and prognostic implications. Chest
4. Guazzi, M., G. Reina, G. Tumminello and M.D. Guazzi. Alveolar-capillary membrane conductance is the best pulmonary function correlate of exercise ventilation
efficiency in heart failure patients. Eur. J. Heart Fail.
5. Johnson, B.D., K.C. Beck, L.J. Olson, K.A. O'Malley, T.G. Allison, R.W. Squires, and G.T. Gau. Ventilatory constraints during exercise in patients with chronic heart failure. Chest
6. Mancini, D., L. Donchez, and S. Levine. Acute unloading of the work of breathing extends exercise duration in patients with heart failure. J. Am. Coll. Cardiol.
7. Mancini, D.M. Pulmonary factors limiting exercise capacity in patients with heart failure. Prog. Cardiovasc. Dis.
8. Musch, T.I. Elevated diaphragmatic blood flow during submaximal exercise in rats with chronic heart failure. Am. J. Physiol. Heart Circ. Physiol.
9. O'Donnell, D.E., C. D'Arsigny, S. Raj, H. Abdollah, and K.A. Webb. Ventilatory assistance improves exercise endurance in stable congestive heart failure. Am. J. Respir. Crit. Care Med.
10. Olson, L.J., Snyder, E.M., K.C. Beck, and B.D. Johnson. Reduced rate of alveolar-capillary recruitment and fall of pulmonary diffusing capacity during exercise in patients with heart failure. J. Card. Fail.
11. Olson, T.P., K.C. Beck, J.B. Johnson, and B.D. Johnson. Competition for intrathoracic space reduces lung capacity in patients with chronic heart failure: a radiographic study. Chest
12. Ponikowski, P., D.P. Francis, M.F. Piepoli, L.C. Davies, T.P. Chua, C.H. Davos, V. Florea, W. Banasiak, P.A. Poole-Wilson, A.J. Coats, and S.D. Anker. Enhanced ventilatory response to exercise in patients with chronic heart failure and preserved exercise tolerance: marker of abnormal cardiorespiratory reflex control and predictor of poor prognosis. Circulation
13. Santamore, W.P., and L.J. Dell'Italia. Ventricular interdependence: significant left ventricular contributions to right ventricular systolic function. Prog. Cardiovasc. Dis.
14. Snyder, E.M., S.T. Turner, and B.D. Johnson. βeta-2 adrenergic receptor genotype and pulmonary function in patients with heart failure. Chest
(in press, 2006).
15. Wasserman, K., Y.Y. Zhang, A. Gitt, R. Belardinelli, A. Koike, L.Lubarsky, and P. Agostoni. Lung function and exercise gas exchange in chronic heart failure. Circulation