Abstract
Impaired aerobic function is a potential mechanism of exercise intolerance in patients with combined cardiorespiratory disease. We investigated the pathophysiological and sensory consequences of a low change in oxygen uptake (ΔV′O2)/change in work rate (ΔWR) relationship during incremental exercise in patients with coexisting chronic obstructive pulmonary disease (COPD) and systolic heart failure (HF).
After clinical stabilisation, 51 COPD–HF patients performed an incremental cardiopulmonary exercise test to symptom limitation. Cardiac output was non-invasively measured (impedance cardiography) in a subset of patients (n=18).
27 patients presented with ΔV′O2/ΔWR below the lower limit of normal. Despite similar forced expiratory volume in 1 s and ejection fraction, the low ΔV′O2/ΔWR group showed higher end-diastolic volume, lower inspiratory capacity and lower transfer factor compared to their counterparts (p<0.05). Peak WR and peak V′O2 were ∼15% and ∼30% lower, respectively, in the former group: those findings were associated with greater symptom burden in daily life and at a given exercise intensity (leg discomfort and dyspnoea). The low ΔV′O2/ΔWR group presented with other evidences of impaired aerobic function (sluggish V′O2 kinetics, earlier anaerobic threshold) and cardiocirculatory performance (lower oxygen pulse, lower stroke volume and cardiac output) (p<0.05). Despite similar exertional hypoxaemia, they showed worse ventilatory inefficiency and higher operating lung volumes, which led to greater mechanical inspiratory constraints (p<0.05).
Impaired aerobic function due to negative cardiopulmonary–muscular interactions is an important determinant of exercise intolerance in patients with COPD–HF. Treatment strategies to improve oxygen delivery to and/or utilisation by the peripheral muscles might prove particularly beneficial to these patients.
Abstract
Impaired aerobic function due to negative cardiopulmonary–muscular interactions is an important determinant of exercise intolerance in patients with comorbid COPD and heart failure http://ow.ly/gveW30nxsUG
Introduction
Heart failure with reduced left ventricular ejection fraction (HF) is a common and disabling comorbidity of chronic obstructive pulmonary disease (COPD) [1, 2]. There is growing recognition that, despite similar respiratory and cardiac impairment at rest, patients with coexisting COPD–HF have poorer exercise tolerance than their counterparts with COPD or HF [3–5]. Advancing the knowledge on the mechanisms leading to patient's disability is crucial to lessen symptom burden and improve their ability to cope with the demands of daily life [6].
In this context, previous work from our group showed that impairments in leg muscle oxygenation and blood flow on exertion were closely related to a low cardiac output and a heightened sense of leg discomfort in a subset of patients with COPD–HF [3]. In addition, we found that an increased ventilatory response to exercise in COPD–HF (i.e. ventilatory inefficiency [5, 7] and exertional oscillatory ventilation [8]) was instrumental to increase the operating lung volumes leading to earlier mechanical-inspiratory constraints and worse dyspnoea at a given work rate (WR) [5, 8]. It is noteworthy that “breathing in excess” at higher lung volumes has important negative haemodynamic consequences (reviewed in [6] and [9]), particularly in patients with already-compromised cardiac function [10]. Impaired oxygen (O2) delivery to and/or utilisation by the working skeletal muscles (i.e. aerobic dysfunction) [11, 12] constitutes the corollary of those deleterious interactions [13]. It is therefore conceivable that COPD–HF patients with worse haemodynamics at rest who ventilate excessively at higher operating lung volumes on exertion [14] are particularly prone to aerobic dysfunction and to report higher leg discomfort and dyspnoea than patients with preserved aerobic function.
In the present study, we contrasted the sensory and cardiopulmonary responses to exertion in COPD–HF patients presenting or not with impaired aerobic function as indicated by a low versus preserved change in oxygen uptake (ΔV′O2)/ΔWR [15, 16] relationship during incremental exercise. We hypothesised that, compared to those with preserved aerobic function, patients showing a low ΔV′O2/ΔWR ratio would present with worse cardiac function at rest, greater exertional ventilation, higher operating lung volumes and, consequently, a higher burden of symptoms on exertion and on daily life. Confirmation of these hypotheses would shed new light on the key traits of COPD and HF which need to be addressed therapeutically to lessen the growing disability associated with this devastating coexistence [6].
Materials and methods
Subjects
We included incremental cardiopulmonary exercise testing data (CPET) from all patients (n=51) who were prospectively enrolled in studies addressing the pathophysiology of coexisting COPD–HF from March 2015 to December 2018. The specific outcome of the present report (aerobic dysfunction during ramp-incremental CPET) has never been explored in our previous investigations which involved a fraction of these patients (i.e. those assessed up to September 2017); thus, there is no overlap between the current report and the data previously shown in a subset of these patients [3–5]. All patients had an established clinical and functional diagnosis of COPD (post-bronchodilator forced expiratory volume in 1 s/forced vital capacity ratio below the lower limit of normal and Global Initiative for Chronic Obstructive Lung Disease spirometric stages 2–3) [17] and documented heart failure with reduced left ventricular ejection fraction (≤40% at the time of diagnosis). They underwent a variable period of clinical stabilisation (ranging from 2 to 8 months) in which their treatment was carefully optimised by cardiologists and respirologists working in academic centres in Brazil and Canada. The original prospective studies which provided the data from the current reported had received ethical approval from the research ethics boards of the Federal University of Sao Paulo Hospital (Sao Paulo, Brazil; #1151/2015) and Queen's University Affiliated Teaching Hospitals (Kingston, ON, Canada; DMED-1588-13).
Procedures
Functional capacity and dyspnoea in daily life were assessed by the New York Heart Association (NYHA) classification and the modified Medical Research Council (mMRC) scale, respectively. After transthoracic echocardiography and pulmonary function tests (spirometry, static lung volumes, transfer factor, arterial blood gases) (1085 ELITE D; Medical Graphics, St Paul, MN, USA in Brazil and Vmax229d; SensorMedics, Yorba Linda, CA, USA in Canada), ramp-incremental CPET (SensorMedics Vmax229d system in both laboratories) was conducted on a different day. The rate of WR increase (Ergoline 800s; SensorMedics) was individually selected (5–15 W·min−1) based on patient's reported level of disability. Key measurements included standard breath-by-breath cardiorespiratory and breathing pattern parameters [18], dynamic operating lung volumes calculated from inspiratory capacity manoeuvres [14], assessment of dyspnoea and leg discomfort intensity (10-point category-ratio Borg scale) and, on a subset of patients, cardiac output by signal morphology impedance cardiography [19]. As detailed in the supplementary material, parameters of aerobic function (V′O2 kinetics delay at early exercise [15], ΔV′O2/ΔWR [15] and peak V′O2 (figure 1) and V′O2 at the estimated lactate threshold [20]) and ventilatory efficiency (Δ minute ventilation (V′E)–Δ carbon dioxide output (V′CO2) slope and intercept by linear regression and V′E/V′CO2 nadir) [21] were obtained following standard recommendations [22]. A low ΔV′O2/ΔWR was defined according to the sex-specific lower limits of normal [23].
Oxygen uptake (V′O2) profile (10-s mean) as a function of work rate (WR) in two representative patients with coexistent chronic obstructive pulmonary disease–heart failure: patients A and B, presenting with preserved and impaired aerobic function, respectively. Note that as a result of a slower initial “lag phase” (122 s versus 66 s) and a shallow ΔV′O2/ΔWR relationship, peak V′O2 is appreciably more impaired than peak WR in patient B than in patient A.
Statistical analysis
The statistical software package used was IBM SPSS Statistics version 24 (IBM, Armonk, NY, USA). Unpaired t-test (or Mann–Whitney test when appropriate) were used to compare between-subject differences. The Chi-squared test was used to compare frequencies. Association between selected continuous variables was investigated by Pearson's product-moment correlation test. Two-way ANOVA with repeated measures were used to compare symptoms intensity and cardiorespiratory, metabolic, gas exchange and operating lung volumes at rest and iso-WR. A p<0.05 level of significance was used for all analyses.
Results
Clinical and resting characteristics
As expected from the increased prevalence of coronary artery disease in COPD [6], most patients were middle-aged or elderly males with HF secondary to ischaemic heart disease (43 (84.3%) out of 51). Most frequent comorbidities included non-insulin dependent diabetes mellitus and chronic kidney failure. Patients were under currently recommended therapy for COPD [17] and HF [24] (table 1). From a functional perspective, patients typically presented with moderate to severe airflow limitation, gas trapping and a low transfer factor (table 2).
General characteristics of the whole sample of patients with coexistent chronic obstructive pulmonary disease (COPD)–heart failure (HF) and patients separated by the presence or not of aerobic dysfunction based on a low or preserved change in oxygen uptake (ΔV′O2)/change in work rate (ΔWR), respectively
Resting functional characteristics of chronic obstructive pulmonary disease (COPD)–heart failure (HF) patients separated by the presence or not of aerobic dysfunction based on a low or preserved change in oxygen uptake (ΔV′O2)/change in work rate (ΔWR), respectively
A low ΔV′O2/ΔWR relationship was found in 27 (52.9%) out of 51 patients (see figure 1 for representative subjects and figure 2a for mean data). No significant differences were found for most clinical and functional variables when the groups with preserved versus low ΔV′O2/ΔWR were contrasted (p>0.05). However, the latter group reported worse functional capacity and dyspnoea in daily life (table 1); moreover, they had lower inspiratory capacity and transfer factor (table 2). Among the echocardiographic variables, only a higher left ventricular end-diastolic diameter separated the low ΔV′O2/ΔWR group from their counterparts (table 2) (p<0.05).
a, b and c) Metabolic; d and e) cardiovascular; and f) sensory responses to incremental exercise in patients presenting or not with impaired aerobic function (change in oxygen uptake (ΔV′O2)/change in work rate (ΔWR) below lower limit of normal). V′CO2: carbon dioxide output; RER: respiratory exchange ratio, HR: heart rate. *: p<0.05 (between-group differences at a given work rate.
Metabolic and cardiovascular responses to exertion
Peak exercise capacity, either expressed as WR or V′O2, was significantly reduced in the low ΔV′O2/ΔWR group (p<0.001). However, owing to definition criterion, whereas the former was on average ∼15% (∼10 W) lower in this group, the latter was ∼30% (∼0.35 L·min−1) inferior compared to the preserved ΔV′O2/ΔWR group (figure 2a, table 3). A plateau on the V′O2 response was found in six patients, all in the low ΔV′O2/ΔWR group. Thus, whereas no patient with preserved ΔV′O2/ΔWR had a severely reduced peak V′O2 (<50% pred), this finding was observed in 12 (44.4%) out of 27 patients in the low ΔV′O2/ΔWR group. Conversely, 12 (46.1%) out of 26 and three (11.1%) out of 27 patients with preserved or low ΔV′O2/ΔWR had peak >70% pred, respectively (p<0.05).
Physiological and sensory responses to incremental cardiopulmonary exercise testing in chronic obstructive pulmonary disease (COPD)–heart failure (HF) patients separated by the presence or not of aerobic dysfunction based on a low or preserved change in oxygen uptake (ΔV′O2)/change in work rate (ΔWR), respectively
The low ΔV′O2/ΔWR group presented other evidence of impaired aerobic function, i.e. slower V′O2 kinetics and, in those with an identifiable lactate threshold (n=27), a lower V′O2 lactate threshold (table 3); moreover, the respiratory exchange ratio was higher at a given WR (figure 2c) due to a lower V′O2 but similar V′CO2 (figure 2b) (p<0.05). Of note, lower oxygen pulse in this group (figure 2e) was associated with higher submaximal heart rate (figure 2d); moreover, stroke volume and cardiac output were reduced in this group compared to their counterparts with preserved ΔV′O2/ΔWR (figure 3). Those metabolic and cardiocirculatory abnormalities were associated with higher submaximal ratings of leg discomfort as exercise progressed (figure 2f) (p<0.05).
Central haemodynamic responses to incremental exercise in the subgroup of chronic obstructive pulmonary disease (COPD)–heart failure (HF) patients with cardiac impedance measurements. a) Stroke volume; b) heart rate; c) cardiac output. Patients are separated by the presence or not of impaired aerobic function (change in oxygen uptake (ΔV′O2)/change in work rate (ΔWR) below the lower limit of normal; n=11 and n=10, respectively). *: p<0.05.
Pulmonary gas exchange and ventilatory responses to exertion
The presence and severity of exertional hypoxaemia did not differ between the groups (table 3). The low ΔV′O2/ΔWR group showed consistently higher ventilation at a given WR and metabolic demand (V′CO2) (figure 4a and b). The higher ventilatory response was associated with a lower tidal volume (figure 4c) and a progressively lower inspiratory capacity; in fact, only the low ΔV′O2/ΔWR group showed a significant decrease in inspiratory capacity from rest to exercise termination (table 3 and figure 4d) (p<0.05). Thus, they presented with higher operating lung volumes throughout exercise, leading to earlier attainment of critical inspiratory constraints (figure 4e). Those ventilatory and mechanical abnormalities were associated with higher submaximal ratings of dyspnoea as exercise progressed (figure 4f) (p<0.05).
a and b) Ventilatory; c) breathing pattern; d and e) operating lung volumes; and f) sensory responses to incremental exercise in patients presenting or not with impaired aerobic function (change in oxygen uptake (ΔV′O2)/change in work rate (ΔWR) below lower limit of normal). V′E: minute ventilation; V′CO2: carbon dioxide output; VT: tidal volume; IC: inspiratory capacity; EILV: end-inspiratory lung volume; EELV: end-expiratory lung volume; TLC: total lung capacity. *: p<0.05 (between-group differences at a given work rate). #: dynamic lung volumes commonly associated with critical inspiratory constraints [14].
Discussion
The main original findings of the present study involving patients with coexisting COPD–HF indicate that impaired aerobic function during incremental CPET, as primarily indicated by a low ΔV′O2/ΔWR ratio and confirmed by a cluster of other parameters, was associated with 1) lower limits for tidal volume expansion (lower inspiratory capacity), impaired gas exchange efficiency (lower transfer factor) and a higher left ventricular end-diastolic diameter at rest; 2) excessive exertional ventilation relative to metabolic demand; 3) higher operating lung volumes leading to earlier attainment of critical inspiratory constraints; and 4) a severely impaired peak aerobic capacity. Of note, those patients provided higher ratings of leg discomfort and breathlessness at a given WR and reported lower functional capacity (NYHA class) and a greater burden of chronic dyspnoea (mMRC scale). Collectively, our results indicate that a low ΔV′O2/ΔWR ratio on CPET signals deleterious cardiopulmonary–peripheral muscular interactions which are germane to patients' functional impairment in daily life.
The ramp-incremental protocol for CPET has the clinical advantage of providing estimates of the key parameters of aerobic function (V′O2 kinetics delay, ΔV′O2/ΔWR ratio, estimated V′O2 lactate threshold and peak V′O2) [15] in a short time frame. This is particularly desirable for the assessment of disabled patients with cardiopulmonary diseases to whom longer or repeated constant WR tests are not feasible options. Using this testing format, we were able to identify, for the first time, a subgroup of COPD–HF patients in whom those parameters uniformly indicated poor aerobic function. Thus, after a sluggish start, V′O2 increased less than expected for the change in power output (i.e. shallower ΔV′O2/ΔWR), leading to a more marked impairment in peak V′O2 than peak WR (table 3). Moreover, there was an earlier shift to a predominantly anaerobic metabolism in these patients compared to their counterparts. Those abnormalities were probably instrumental to explain why those patients provided higher ratings of leg discomfort at a given WR (figure 2f).
A key interpretative issue relates to the potential mechanism(s) underlying the defining feature of the group with impaired aerobic function: a low ΔV′O2/ΔWR ratio. Whereas in healthy subjects V′O2 parallels the increase in WR, thereby allowing work efficiency to be estimated [15], a pathological decrease in ΔV′O2/ΔWR implies a lower V′O2 cost to perform a given WR and/or progressively slower V′O2 kinetics as WR increases, i.e. a gradually longer time for muscle V′O2 to be represented “at the mouth” [12, 16]. The first hypothesis is consistent with the notion of impaired muscle oxygen utilisation secondary to profound abnormalities in muscle oxidative metabolism [25]. Of note, those abnormalities have been described in patients with HF [26] or COPD [27–29]; thus, it is conceivable that severe peripheral muscular derangements impairing oxygen extraction may have contributed to decrease ΔV′O2/ΔWR ratio in selected patients [30].
A large body of evidence has also been accumulated in favour of a role of impaired muscle oxygen delivery (i.e. lower cardiac output under preserved arterial oxygen saturation) to decrease the ΔV′O2/ΔWR relationship in patients with HF [31–33]. In fact, we found that the group with impaired aerobic function presented with similar oxygen saturation by pulse oximetry, but lower oxygen pulse (figure 2e) and stroke volume (figure 3a) than their counterparts with preserved aerobic function. The higher resting left ventricular end-diastolic diameter in the former group suggests more advanced HF [34], which may have contributed to the lower ΔV′O2/ΔWR. It is also conceivable that ΔV′O2/ΔWR has been negatively impacted by the deleterious central haemodynamic consequences of increased operating lung volumes (figure 4e) [6]. The most noticeable findings in previous investigations carried out in hyperinflated COPD patients (without HF) point to an increased right ventricular afterload and a relative underfilling of the left ventricle [35, 36]. It is noteworthy that patients with more impaired pulmonary microvascular blood flow in the study by Hueper et al. [37] presented with a low transfer factor, one of the few resting functional findings which differed the patients with low versus preserved ΔV′O2/ΔWR (table 2), The greater inspiratory constraints found in the former group (figure 4e) signal severe neuromechanical dissociation and higher swings in intrathoracic pressure; the latter being an important mechanism of increasing left ventricular transmural pressure [13, 38]. Indeed, we found a temporal association between oxygen pulse (figure 2e) and stroke volume (figure 3a) plateauing as the inspiratory reserve volume became critically low (after ∼30 W) (figure 4e). Those abnormalities were conceivably more relevant to those in need of a higher left ventricular filling volume, i.e. patients with aerobic dysfunction who presented with higher left ventricular end-diastolic diameter at rest (table 1).
Another noticeable finding that characterised the low ΔV′O2/ΔWR group was an increased ventilatory response to a given V′CO2 compared to the preserved ΔV′O2/ΔWR group [5]. Of note, patients' higher V′E was increased well before any indirect evidence of the lactate threshold (in fact, even at rest) (figure 4a). In other words, increased V′E in the low ΔV′O2/ΔWR group was not only a response to a greater lactic acidotic drive (table 3), but was also influenced by higher dead space/tidal volume ratio and/or a lower arterial carbon dioxide tension (PaCO2) set point [11]. A higher dead space/tidal volume could be partially explained by a lower tidal volume (figure 4c) due to greater inspiratory constraints (figure 4e) [39]. Other sources of afferent stimuli apart from those involved in the regulation of the physiological dead space may have also been involved, including heightened ergoreceptor [40, 41] and sympathetic activation [42] in the setting of poorer muscle oxygenation in patients with aerobic impairment.
The patients in the low ΔV′O2/ΔWR group presented with higher operating lung volumes throughout exercise (figure 4f). In addition to a trend towards greater “static” hyperinflation, it is noteworthy that they had worse dynamic hyperinflation than their counterparts (figure 4d). The higher submaximal ventilation (figure 4a) may have accelerated the rate of dynamic hyperinflation in these patients, leading to earlier attainment of critical inspiratory constraints and a downward shift in tidal volume (figure 4c) [39]. Thus, patients in the low ΔV′O2/ΔWR group developed an unfortunate combination of abnormalities which are central to the genesis of dyspnoea in COPD [43]: a high respiratory neural drive (∼higher V′E) (figure 4a), which is only partially rewarded by tidal expansion (greater inspiratory constraints) (figure 4c and e) [44].
What are the practical implications of our results? Firstly, we provided novel evidence that measurements of a submaximal, effort-independent parameter of aerobic function (ΔV′O2/ΔWR) during incremental CPET provide clinically relevant information pertaining to exertional symptoms and daily functioning in patients with COPD–HF. Secondly, the negative haemodynamic consequences of higher operating lung volumes in patients with poorer cardiac function calls for the importance of maximising the deflating effects of bronchodilator therapy in this subgroup of patients with COPD [45]. Thirdly, strategies geared towards an improvement in oxygen delivery to and utilisation by the skeletal muscles (e.g. sildenafil [46], nitrate supplementation [47] and aerobic and muscle training [48]) are probably relevant to lessen patients' breathlessness (by lowering the ventilation stimuli) and leg muscle discomfort. Finally, the combination of a higher left ventricular end-diastolic diameter in association with lower inspiratory capacity and transfer factor (in addition to a lower PaCO2, as per our previous findings) [5] should be valued to identify the most disabled patients with COPD–HF.
We recognise some limitations of our study. Due to its non-invasive nature, our study is not particularly informative regarding to the specific effects of lung mechanical abnormalities on central haemodynamics. Despite the inherent limitation of impedance cardiography [19], it is noteworthy that stroke volume and cardiac output trajectories and values were commensurate to those expected in patients with HF (figure 3). To our knowledge, no study to date has prospectively exposed COPD–HF patients to invasive haemodynamics on exertion. In any circumstance, a study with those features would be a formidable challenge in such an unstable population. Our patients underwent a prolonged period of clinical stabilisation before study entry; moreover, most presented with repeated COPD and/or HF exacerbations, thereby postponing the CPET. It follows that our sample size, though small for epidemiological standards, reflects the real-life challenges in assessing on exertion an extremely frail population.
In conclusion, impaired aerobic function, as primarily indicated by a low ΔV′O2/ΔWR during incremental CPET, discriminates a subgroup of patients with COPD–HF who are particularly symptomatic and disabled. Negative cardiopulmonary interactions conspire against a normal oxygen delivery to and utilisation by the skeletal muscles in these patients. The latter abnormalities increase the ventilatory response to exercise and, indirectly, the operating lung volumes; thus, they ultimately potentiate the central haemodynamic abnormalities. Effective strategies to fight the devastating consequences of COPD–HF should address the continuum of cardiocirculatory, pulmonary and muscular abnormalities that culminates in poor exercise tolerance and quality of life in this growing population.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary material ERJ-02386-2018.Supplement
Footnotes
This article has supplementary material available from erj.ersjournals.com
Conflict of interest: A. Rocha has nothing to disclose.
Conflict of interest: F.F. Arbex has nothing to disclose.
Conflict of interest: P.A. Sperandio has nothing to disclose.
Conflict of interest: F. Mancuso has nothing to disclose.
Conflict of interest: M. Marillier has nothing to disclose.
Conflict of interest: A-C. Bernard has nothing to disclose.
Conflict of interest: M.C.N. Alencar has nothing to disclose.
Conflict of interest: D.E. O'Donnell has nothing to disclose.
Conflict of interest: J.A. Neder has nothing to disclose.
Support statement: William Spear Start Endowment Fund, Queen's University; financial support to A. Rocha was provided by Capes, Brazil; financial support to M. Marillier was provided by a long-term research fellowship from the European Respiratory Society; J.A. Neder was funded by a New Clinician Scientist Program from the Southeastern Ontario Academic Medical Association (SEAMO), Canada. The funders had no role in the study design, data collection and analysis or preparation of the manuscript.
- Received December 16, 2018.
- Accepted January 22, 2019.
- Copyright ©ERS 2019
References
