Abstract
The purpose of this study was to determine if a dissociation existed between respiratory drive, as estimated by diaphragmatic electromyography (EMGdi), and its pressure-generating capacity during exercise in mild chronic obstructive pulmonary disease (COPD) and whether this, if present, had negative sensory consequences.
Subjects meeting spirometric criteria for mild COPD (n=16) and age and sex-matched controls (n=16) underwent detailed pulmonary function testing and a symptom limited cycle test while detailed ventilatory, sensory and respiratory mechanical responses were measured.
Compared with controls, subjects with mild COPD had greater ventilatory requirements throughout submaximal exercise. At the highest equivalent work rate of 60 W, they had a significantly higher: total work of breathing (32±17 versus 16±7 J·min−1; p<0.01); EMGdi (37.3±17.3 versus 17.9±11.7% of maximum; p<0.001); and EMGdi to transdiaphragmatic pressure ratio (0.87±0.38 versus 0.52±0.27; p<0.01). Dyspnoea–ventilation slopes were significantly higher in mild COPD than controls (0.17±0.12 versus 0.10±0.05; p<0.05). However, absolute dyspnoea ratings reached significant levels only at high levels of ventilation.
Increased respiratory effort and work of breathing, and a wider dissociation between diaphragmatic activation and pressure-generating capacity were found at standardised work rates in subjects with mild COPD compared with controls. Despite these mechanical and neuromuscular abnormalities, significant dyspnoea was only experienced at higher work rates.
Abstract
Mild COPD patients experience respiratory mechanical abnormalities during exercise despite relatively preserved FEV1 http://ow.ly/yHukJ
Introduction
Despite having relatively preserved spirometry, patients with milder chronic obstructive pulmonary disease (COPD) show increased mortality [1], increased hospitalisations, and decreased health-related quality of life [2–5] compared with nonsmoking controls. Many of these patients also experience exertional dyspnoea [6], which may partially explain their reduced physical activity levels [7]. Despite these findings, this subpopulation has not been extensively studied and no evidence-based recommendations currently exist for their management beyond smoking cessation. Therefore, the optimal management of patients with mild COPD awaits a better understanding of their underlying pathophysiology.
The clinical presentation of patients with mild airway obstruction is highly variable and depends on the integration of several factors such as the nature and extent of their physiological impairment, the attendant compensatory respiratory system adjustments, and complex behavioural adaptations (e.g. activity avoidance). The heterogeneous physiological abnormalities of patients with mild airway obstruction include: increased peripheral airway resistance, early airway closure, maldistribution of ventilation, impaired pulmonary gas exchange and gas trapping [8–12]. Recent studies have shown that these physiological abnormalities are amplified during exercise [6, 13, 14]. Thus, the combination of ventilatory inefficiency and abnormal ventilatory mechanics can force the respiratory system to prematurely reach its physiological limits at lower exercise intensities than in healthy controls [13].
The current study extends previous work [15–17] to determine if a dissociation exists between diaphragmatic activation and pressure generation or between respiratory muscle effort and the mechanical output of the respiratory system during exercise in patients with mild COPD and whether such factors have negative sensory consequences. Therefore, the current study tested the hypotheses that: 1) dissociation between respiratory neural drive and the muscular/mechanical response of the respiratory system would be present in mild COPD at lower work rates than in healthy controls; and 2) these compensatory adjustments to support the increased ventilatory demands of exercise would be associated with an earlier onset of intolerable dyspnoea.
Methods
Subjects
The study included 16 subjects with Global Initiative for Chronic Obstructive Lung Disease grade 1 COPD (forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC) <0.70 and FEV1 ≥80% predicted) and 16 healthy age- and sex-matched controls. Subjects were recruited through advertisements and through the respiratory clinic at Kingston General Hospital (Kingston, ON, Canada). Individuals with cardiovascular, metabolic, neuromuscular or any other disease that could contribute to dyspnoea or interfere with exercise testing were excluded.
Study design
This study received ethical approval from the Queen’s University and Affiliated Hospitals Research Ethics Board (DMED-1285-10). Subjects participated in two testing visits. On visit one, a detailed medical history was completed, followed by chronic activity-related dyspnoea questionnaires, pre- and post-bronchodilator (400 μg salbutamol) pulmonary function testing and a symptom-limited incremental cycle test for familiarisation purposes. On visit two, subjects performed pulmonary function testing followed by the same exercise protocol but with additional measurements of diaphragmatic electromyography (EMGdi) and respiratory pressures. COPD subjects refrained from using their short- and long-acting bronchodilators for 6 and 24 h prior to testing, respectively, and avoided caffeine, heavy meals and alcohol for at least 4 h prior to testing.
Pulmonary function
Spirometry, body plethysmography, single-breath diffusing capacity of the lung for carbon monoxide (DLCO), single-breath nitrogen test, maximal mouth pressures and impulse oscillometry were performed using an automated testing system (Vs62j Body Plethysmograph, Vmax229d Encore and Masterscreen IOS; SensorMedics, Yorba Linda, CA) according to established guidelines and expressed as % pred of normal values.
Cardiopulmonary exercise testing
Exercise was performed on an electronically-braked cycle ergometer (Ergometrics 800S; SensorMedics) using a cardiopulmonary testing system (Vmax229d; SensorMedics). Testing involved a rest period and a 1 min warm-up of unloaded pedalling followed by stepwise increases of 20 W per 2 min until symptom limitation. Standard cardiopulmonary measurements were collected on a breath-by-breath basis. Blood pressure, heart rate and oxyhaemoglobin saturation were monitored using sphygmomanometry, electrocardiography, and pulse oximetry, respectively. Inspiratory capacity (IC) measurements were performed at rest, every second minute during exercise and at the end of exercise [18]. Subjects rated the intensity of their “breathing discomfort” and “leg discomfort” at rest, every second minute during exercise and at the end of exercise using the modified 10-point Borg scale [19]. Subjects were asked to report their primary reason for stopping exercise and to select qualitative phrases that described how their breathing felt at peak exercise [20]. An inflection in tidal volume (VT) relative to minute ventilation (V′E) was determined for each subject during exercise [21]. The highest equivalent work rate (HEWR) was defined as the highest work rate achieved by all subjects.
Diaphragmatic EMG and respiratory pressures
A combined electrode-balloon catheter was passed through the nose and positioned as previously described [22]. Raw EMGdi signals were sampled at 2000 Hz (PowerLab, model ML880; ADInstruments, Castle Hill, Australia), band-pass filtered between 20–1000 Hz (Bio-amplifier model RA-8; Guangzhou Yinghui Medical Equipment Co. Ltd, Guangzhou, China) and converted into a root mean square. The mean root mean square for each inspiration was determined by manually selecting the EMGdi signal that fell between cardiac QRS complexes. The electrode pair with the largest EMGdi amplitude was used for the analysis. Maximal EMGdi (EMGdi,max) was determined as the highest EMGdi amplitude from any IC manoeuvre [22, 23]. The ratio between EMGdi/EMGdi,max and VT/(predicted vital capacity (VCpred)) was used as an index of neuromechanical dissociation of the respiratory system.
Oesophageal and gastric balloon catheters were connected to calibrated differential pressure transducers (model DP15-34; Validyne Engineering, Northridge, CA, USA). Transdiaphragmatic pressure (Pdi) was calculated as the difference between Pga and Pes. Maximal Pdi (Pdi,max) and Pes (Pes,max) were determined during maximal IC and sniff manoeuvres, respectively. The work of breathing was determined as the area within an averaged tidal Pes–volume loop and was further subdivided into the inspiratory resistive and inspiratory elastic work of breathing [24]. The breath selection occurred over the same 30 s as all other cardiorespiratory and EMG data (i.e. the first 30 s of each second minute of exercise). An index of ventilatory muscle recruitment was calculated as the difference between end-inspiratory and end-expiratory Pga divided by the difference between end-inspiratory and end-expiratory Pes (ΔPga/ΔPes) [25].
Data analysis
Between-group comparisons of subject characteristics, and physiological and sensory variables at each time point during exercise were performed using unpaired t-tests. Qualitative descriptors of dyspnoea at peak exercise and the reasons for stopping exercise were analysed using Fisher’s exact test. To evaluate the relationship between dyspnoea intensity (dependent variable) and relevant independent variables during exercise, the following were included in a multivariable linear regression model: the independent variable of interest, group as a categorical effect, an interaction term to determine whether the relationship being tested was similar across groups (independent variable×group), and subjects were treated as random effects to account for serial measurements (subject nested within group). Results are reported as mean±sd, unless otherwise specified. A p-value <0.05 was used for statistical significance.
Results
Subjects
Descriptive characteristics of the mild COPD subjects (n=16) and controls (n=16) are summarised in table 1. Subjects with mild COPD had a significant smoking history (range 15–111 pack-years) including 12 ex-smokers and four current smokers. Only two control subjects had a smoking history of 2.5 and 2 pack-years each, while all other controls were never-smokers. Nine COPD subjects used respiratory medications on a regular basis: short-acting β2-agonist as needed (n=7), short-acting anticholinergic (n=2), long-acting anticholinergic (n=5), inhaled corticosteroid (n=1), and a long-acting β2-agonist combined with an inhaled corticosteroid (n=6). Comorbidities in the COPD group included the following: hypertension (n=10), diabetes mellitus type 2 (n=4), hypercholesterolaemia (n=3), osteoarthritis (n=3), anxiety (n=2), gastro-oesophageal reflux (n=1), and sleep apnoea (n=1). Comorbidities in the healthy control group included the following: hypertension (n=3), osteoarthritis (n=3), hypercholesterolaemia (n=1), sleep apnoea (n=1), and anaemia (n=1). None of the COPD subjects had participated in pulmonary rehabilitation.
All but one COPD subject was below the lower limit of normal for FEV1/FVC [26]. Compared with controls, subjects with mild COPD had significantly greater activity-related dyspnoea, airway resistance, N2 slope, resonant frequency, functional residual capacity, residual volume, and a greater difference between total lung capacity and alveolar volume. They also had a significantly reduced IC, DLCO and static respiratory muscle strength.
Physiological responses to exercise
Peak work rate and oxygen uptake (V′O2) were significantly lower in subjects with mild COPD compared with controls. Selected sensory and physiological parameters at rest, the HEWR, and peak exercise are shown in table 2. figure 1 shows the metabolic and ventilatory responses to exercise. Compared with controls, subjects with mild COPD had greater ventilatory inefficiency (fig. 1b), higher ventilatory requirements (fig. 1c), and higher operating lung volumes during exercise (fig. 1f). A discernible VT/V′E inflection occurred in all subjects except for one subject with mild COPD. The VT/V′E inflection (denoted as triangles in the figure) occurred at a significantly lower V′O2, V′E and work rate in the subjects with mild COPD compared with controls.
Metabolic and ventilatory responses to exercise. a) Oxygen uptake (V′O2), b) minute ventilation (V′E)/carbon dioxide production (V′CO2), c) V′E, d) respiratory frequency (fR), e) tidal volume (VT) and f) operating lung volumes at different work rates. COPD: chronic obstructive pulmonary disease; TLC: total lung capacity; EILV: end-inspiratory lung volume; EELV: end-expiratory lung volume. Data are presented as mean±sem. Triangles represent the VT/V′E inflection. *: p<0.05.
Respiratory mechanics and diaphragmatic function
The total work of breathing and the inspiratory elastic and resistive work of breathing were significantly elevated in mild COPD throughout submaximal exercise compared with controls (fig. 2). At the HEWR, the total work of breathing was two-fold higher in mild COPD versus controls. Peak exercise work of breathing was higher in the controls reflecting their higher peak V′E. EMGdi and mean Pdi were higher in mild COPD at rest and throughout submaximal exercise (fig. 2d and 2e, respectively). The ratio between EMGdi/EMGdi,max and Pdi/Pdi,max was also significantly higher during exercise in subjects with mild COPD (fig. 2f). Subjects with mild COPD had lower VT and V′E, and a modestly higher Pdi for any given EMGdi (fig. 3a–c). Total respiratory effort (Pes/Pes,max) was similar between groups for any given EMGdi (fig. 3d). Neuromechanical dissociation of the respiratory system, whether expressed as (EMGdi/EMGdi,max)/(VT/VCpred) or (Pes/Pes,max)/(VT/VCpred), was significantly higher in mild COPD for any given work rate (fig. 4) or V′E compared with the controls.
Work of breathing and diaphragmatic function in response to exercise. a) Total work of breathing (Wb), b) inspiratory elastic Wb, c) inspiratory resistive Wb, d) ratio of diaphragmatic electromyography (EMGdi) to maximal EMGdi (EMGdi,max), e) ratio of transdiaphragmatic pressure (Pdi) to maximal Pdi (Pdi,max) and f) ratio of EMGdi/EMGdi,max to Pdi/Pdi,max at different work rates. Data are presented as mean±sem. Triangles represent the tidal volume/minute ventilation inflection. COPD: chronic obstructive pulmonary disease. *: p<0.05.
Ventilatory and respiratory pressure responses for a given level of diaphragmatic activation. a) Ratio of tidal volume (VT) to predicted vital capacity (VCpred), b) minute ventilation (V′E), c) ratio of transdiaphragmatic pressure (Pdi) to maximal Pdi (Pdi,max) and d) ratio of tidal oesophageal pressure (Pes) to maximal Pes (Pes,max) at different diaphragmatic electromyography (EMGdi) to maximal EMGdi (EMGdi,max) ratios. Data are presented as mean±sem. Triangles represent the VT/V′E inflection. Statistical comparisons were performed at the VT/V′E inflection and at peak exercise. COPD: chronic obstructive pulmonary disease. *: p<0.05.
Neuromechanical dissociation of the respiratory system. As exercise progresses there is an increasing disparity between a) the ratio of diaphragmatic electromyography (EMGdi) to maximal EMGdi (EMGdi,max) or b) the ratio of tidal oesophageal pressure (Pes) to maximal Pes (Pes,max) and the tidal volume (VT) response in mild chronic obstructive pulmonary disease (COPD) compared with healthy controls. VCpred: predicted vital capacity. Data are presented as mean±sem. Triangles represent the VT/V′E inflection. *: p<0.05.
Sensory responses to exercise
Breathing discomfort, alone or in combination with leg discomfort was selected as the primary reason for stopping exercise in 69% of subjects with mild COPD versus 25% of healthy controls (p=0.07). Leg discomfort alone was the primary reason for stopping exercise in 19% of subjects with COPD versus 69% of healthy controls (p<0.05). Other reasons for stopping exercise in mild COPD included “general fatigue” (n=2). One healthy control subject stopped due to a “dry mouth and throat”. Phrases alluding to “inspiratory difficulty” were selected more frequently in subjects with mild COPD versus controls at peak exercise (69% versus 25% of subjects, p<0.05). Dyspnoea–V′E slopes were significantly higher in subjects with mild COPD versus controls (0.17±0.12 versus 0.10±0.05; p<0.05). Dyspnoea intensity ratings were also significantly higher in mild COPD subjects at the HEWR (fig. 5a). Despite this, absolute dyspnoea ratings in the mild COPD group remained relatively low up to the VT/V′E inflection, corresponding to a V′E of 40.8±13.2 L·min−1. Dyspnoea increased similarly in both groups as a function of increasing EMGdi throughout exercise (fig. 5b). Dyspnoea intensity during exercise correlated best with V′E (% of maximum voluntary ventilation), Pes/Pes,max and EMGdi/EMGdi,max (all p<0.0005); these relationships were similar across groups. Indices of neuromechanical dissociation (i.e. Pes/VT and EMG/VT ratios) were more weakly associated with dyspnoea, but these relationships showed a significant (p<0.005) group interaction.
Dyspnoea intensity in response to exercise. Dyspnoea measured using the modified 10-point Borg scale in response to exercise measured as a) work rate and b) the ratio of diaphragmatic electromyography (EMGdi) to maximal EMGdi (EMGdi,max). Data are presented as mean±sem. Triangles represent the tidal volume/minute ventilation inflection. COPD: chronic obstructive pulmonary disease. *: p<0.05.
Discussion
The novel findings of this study are as follows. 1) The resistive and elastic loads on the respiratory muscles were significantly increased resulting in increased diaphragmatic activation, respiratory effort and work of breathing in mild COPD relative to controls at standardised work rates. 2) In COPD, there was a widening dissociation between indices of respiratory neural drive and the muscular and mechanical response of the respiratory system at lower work rates compared with controls. 3) Despite the mechanical and neuromuscular adaptations needed to support increased ventilation, subjects with mild COPD did not experience significant dyspnoea during moderate submaximal work rates.
Although the mild COPD group had only minor spirometric abnormalities and minimal lung hyperinflation at rest, they had evidence of significant peripheral airway dysfunction as indicated by a reduction in mid-volume maximal expiratory flows, maldistribution of ventilation, pulmonary gas trapping and negative frequency dependence of resistance. These subjects also had mild-to-moderate activity-related dyspnoea and moderately impaired health status. Absolute peak work rate and V′O2 were significantly lower in the COPD group relative to controls and predicted values. Collectively, these results demonstrate that subjects with mild airway obstruction can have important physiological and clinical abnormalities despite having relatively preserved spirometry.
Our results confirm previous reports of significant ventilatory abnormalities in subjects with mild COPD during exercise (fig. 1) [6, 14, 27]. For example, greater ventilatory inefficiency (high V′E/carbon dioxide production (V′CO2)) was present at rest and during exercise in mild COPD, likely reflecting an increased physiological dead space due to regional increases in ventilation/perfusion ratios. Our mild COPD group also had evidence of dynamic hyperinflation at peak exercise (∼0.3 L), along with greater tachypnoea and an earlier VT inflection, indicating that the end-inspiratory lung volume encroached on total lung capacity (TLC) (88% of TLC) at a lower work rate (102 versus 178 W) and V′E (55.7 versus 92.7 L·min−1) than in the controls.
Compared with controls, subjects with mild COPD had greater resistive and elastic loading of the inspiratory muscles resulting in an increased total work of breathing for any given submaximal work rate. At the HEWR, the total work of breathing was two-fold greater in mild COPD relative to controls and this was predominantly driven by a higher inspiratory resistive work of breathing. Respiratory muscle effort was also markedly higher in mild COPD subjects across all submaximal work rates. Thus, to perform the same standardised physical task, subjects with mild COPD consistently had greater erosion of their respiratory reserve than healthy individuals.
Diaphragmatic function and dynamic respiratory mechanics
To our knowledge, this is the first study to examine detailed measures of diaphragmatic function in individuals with mild airway obstruction. EMGdi was markedly higher in this group at rest and throughout all submaximal work rates. The ratio of EMGdi to Pdi was also significantly higher relative to controls, indicating some degree of neuromuscular dissociation of the diaphragm. This higher ratio was driven, in part, by the more rapid increase in EMGdi since the rate of increase in Pdi was similar in both groups and increased linearly with increasing exercise intensity (fig. 2e). This is in contrast with the work by Sinderby et al. [17] in moderate-to-severe COPD, where a plateau in Pdi occurred relatively early during exercise despite a continued increase in EMGdi. The authors proposed that this was the result of reduced pressure generating capacity of the diaphragm due to lung hyperinflation. Patients with moderate-to-severe COPD have been shown to have an EMGdi of 70% of maximum at peak incremental exercise corresponding to a V′E of 32.3±9.6 L·min−1 using the same catheter as used in the present study [28]. At a similar absolute V′E (33.9±8.1 L·min−1), we found that patients with mild COPD have an EMGdi of 37% of maximum, suggesting that respiratory neural drive increases dramatically with advancing COPD severity. Accordingly, the relatively preserved pressure generating capacity of the diaphragm in our mild COPD subjects probably reflects the presence of less severe mechanical constraints. In keeping with this, the pattern of ventilatory muscle recruitment during exercise was not different in mild COPD and controls. Nevertheless, the higher diaphragmatic activation and EMGdi to Pdi ratio in mild COPD indicates the presence of important compensatory physiological adaptations. Specifically, the combination of increased mechanical loading and pulmonary gas exchange abnormalities led to an increase in neural activation of the diaphragm and total respiratory muscle contractile effort, which permitted subjects to achieve the alveolar ventilation required to meet the metabolic demands of moderate physical work.
It has long been understood that there is a widening disparity between the drive to breathe and the corresponding mechanical output of the impaired respiratory system during exercise in advanced COPD [29]. Since respiratory neural drive cannot be measured directly, we rely on indirect estimates such as measures of effort (i.e. Pes/Pes,max). Unfortunately, this approach can lead to significant underestimation in the setting of severe respiratory mechanical impairment [17, 30, 31]. The additional use of EMGdi as a surrogate for respiratory neural drive in mild COPD [32] is unique to this study. Increases in EMGdi/EMGdi,max and Pes/Pes,max were closely correlated (fig. 3d) throughout exercise in health and mild COPD, thus supporting the contention that these measures are reasonable surrogates of respiratory neural drive, at least in this population. Remarkably, even in the setting of mild spirometric abnormality, VT expansion and, thus, V′E was significantly reduced for a given EMGdi (and respiratory muscle effort) compared with controls (fig. 3a and 3b). This inability to appropriately expand VT is probably explained by the prevailing mechanical ventilatory constraints. This latter contention is bolstered by the previous finding that bronchodilators, by decreasing lung hyperinflation, allow greater VT expansion during exercise in mild COPD [33, 34].
Dyspnoea
Despite the compensatory increases in diaphragmatic activation and respiratory effort needed to maintain V′E in pace with increasing metabolic demands, minimal dyspnoea was experienced up to 80 W or a V′E of ∼40 L·min−1. The V′O2 at this work rate in mild COPD corresponds to approximately five metabolic equivalents. Thus, by extrapolation, such patients are likely to be able to undertake many daily activities requiring moderate exertion without experiencing significant respiratory discomfort [35]. However, it is clear that more sustained or strenuous tasks would evoke severe dyspnoea. In both groups, dyspnoea increased as a function of increasing EMGdi, Pes and V′E, all relative to maximum. Thus, the higher dyspnoea intensity at equivalent work rates in mild COPD compared with controls, generally reflects their higher drive to breathe and respiratory muscle effort requirements. We speculate that this increasing dissociation between neural drive and the muscular/mechanical response of the respiratory system might explain the more frequent selection of the qualitative descriptor inspiratory difficulty at a much lower peak work rate in mild COPD compared with controls [16].
Limitations
This study had a small sample size due to the complexity and invasiveness of the EMGdi and respiratory pressure measurements. Despite this, our sample size was sufficient to detect significant differences in dyspnoea intensity and a wide range of physiological and respiratory mechanical outcome variables of interest. The mild COPD subjects who participated in this study had, on average, mild-to-moderate activity-related dyspnoea. Thus, our results may not be generalisable to asymptomatic subjects with mild COPD [36].
Conclusions and implications
In patients with a relatively preserved FEV1, extensive mechanical and respiratory neuromuscular adaptations were already in place, which allowed them to adequately meet the metabolic demands of moderate physical activity. However, maximal exercise capacity was diminished in these patients, and our study is the first to show that this was associated with the inability of the respiratory system to appropriately increase ventilation in the face of increasing diaphragmatic activation and contractile respiratory muscle effort. These compensatory adjustments in mild COPD appeared to have little negative influence on respiratory sensation until higher levels of ventilation were required to support increased exercise intensities. These results highlight that exclusive reliance on spirometry, to assess severity of airway obstruction in patients with mild COPD, may result in underestimation of clinically important physiological impairment.
Footnotes
Support statement: This study received funding from a William Spear/Richard Start Endowment Fund, Queen’s University, and the Ontario Thoracic Society. J.A. Guenette was supported by postdoctoral fellowships from the Natural Sciences and Engineering Research Council of Canada (NSERC), the Canadian Thoracic Society and the Canadian Lung Association, as well as a New Investigator Award from the Providence Health Care Research Institute and St. Paul’s Hospital Foundation. R.C. Chin was supported by the Queen’s Graduate Award and the Queen Elizabeth II Graduate Scholarship in Science and Technology. P.B. Dominelli was supported by an NSERC Canada Graduate Scholarship.
Conflict of interest: Disclosures can be found alongside the online version of this article at erj.ersjournals.com
- Received February 20, 2014.
- Accepted June 14, 2014.
- ©ERS 2014
References
