Pulmonary Capillary Blood Volume Response to Exercise Is Diminished in T Mild Chronic Obstructive Pulmonary Disease Vincent Tedjasaputraa,B, Sean Van Diepenc, Devin B

Respiratory Medicine 145 (2018) 57–65

Contents lists available at ScienceDirect

Respiratory Medicine

journal homepage: www.elsevier.com/locate/rmed

Pulmonary capillary blood volume response to exercise is diminished in T mild chronic obstructive pulmonary disease Vincent Tedjasaputraa,b, Sean van Diepenc, Devin B. Phillipsa,b, Eric Y.L. Wonga, Mohit Bhutania, ∗ Wade W. Michaelchuka,b, Tracey L. Bryana, Michael K. Sticklanda,d, a Division of Pulmonary Medicine, Faculty of Medicine and Dentistry, University of Alberta, Canada b Faculty of Kinesiology, Sport, and Recreation, University of Alberta, Canada c Department of Critical Care and Division of Cardiology, Faculty of Medicine and Dentistry, University of Alberta, Canada d G.F. MacDonald Centre for Lung Health, Covenant Health, Edmonton, Alberta, Canada

ARTICLE INFO ABSTRACT

Keywords: Background: Previous work suggests that mild chronic obstructive pulmonary disease (COPD) patients have Pulmonary capillary blood volume greater lung dysfunction than previously appreciated from spirometry alone. There is evidence of pulmonary Diffusing capacity microvascular dysfunction in mild COPD, which may reduce diffusing capacity (DLCO) and increase ventilatory Mild COPD inefficiency during exercise. The purpose of this study was to determine if DLCO, pulmonary capillaryblood Exercise volume (Vc), and membrane diffusing capacity (Dm) are diminished during exercise in mild COPD, and whether Dyspnea this is related to ventilatory inefficiency and dyspnea.

Methods: Seventeen mild COPD patients (FEV1/FVC: 64 ± 4%, FEV1 = 94 ± 11%pred) and 17 age- and sex- matched controls were recruited. Ten moderate COPD patients were also tested for comparison

(FEV1 = 66 ± 7%pred). DLCO, Vc, and Dm were determined using the multiple-fraction of inspired oxygen (FIO2) DLCO method at baseline and during steady-state cycle exercise at 40W, 50%, and 80% of VO2peak. Using expired gas data, ventilatory inefficiency was assessed by VE/VCO 2. Results: Compared to controls, mild COPD had lower DLCO at baseline and during exercise secondary to diminished Vc (P < 0.05). No difference in Dm was observed between controls and mild COPD at rest orduring exercise. Patients with high (i.e. ≥34) had lower Vc and greater dyspnea ratings compared to control at 40W. Moderate COPD patients were unable to increase Vc with increasing exercise intensity, suggesting further pulmonary vascular impairment with increased obstruction severity. Conclusion: Despite relatively minor airflow obstruction, mild COPD patients exhibit a diminished DLCOand capillary blood volume response to exercise, which appears to contribute to ventilatory inefficiency and greater dyspnea.

1. Introduction inefficiency (VE/VCO 2), increased dyspnea, and reduced VO2peak compared to control [3,5]. Importantly, increased ventilatory inefficiency Chronic obstructive pulmonary disease (COPD) is characterized by (i.e. a nadir during exercise ≥34) is predictive of mortality in progressive partially reversible airway obstruction and exertional dys- COPD [6]. Some have suggested the increased during exercise pnea [1], which impairs a patient's ability to complete day-to-day tasks in mild COPD is due to greater deadspace [5], as a result of pulmonary and reduces quality of life [2]. Evolving work in mild COPD (GOLD vascular dysfunction [7–9]. Pulmonary vascular dysfunction reduces stage 1, forced expiratory volume in 1s (FEV1)/forced vital capacity pulmonary microvascular perfusion [8], decreases surface area for gas (FVC) < 70%, FEV1 ≥ 80% predicted) have demonstrated greater exchange, increases deadspace ventilation, and thus requires greater functional impairment than previously appreciated [3,4], suggesting minute ventilation to maintain alveolar ventilation (i.e. greater that the physiological impact of the disease is worse than indicated by during exercise). spirometry alone. Despite their relatively minor airway impairment, During exercise, diffusing capacity must increase in order to aug- during exercise, mild COPD patients demonstrate increased ventilatory ment gas exchange [10]. The main components of diffusing capacity are the membrane diffusing capacity (Dm), which reflects the surface area

∗ Corresponding author. 3-135 Clinical Sciences Building, Edmonton, Alberta, T6G 2J3, Canada. E-mail address: [email protected] (M.K. Stickland). https://doi.org/10.1016/j.rmed.2018.10.015 Received 6 August 2018; Received in revised form 11 October 2018; Accepted 18 October 2018 Available online 19 October 2018 0954-6111/ © 2018 Elsevier Ltd. All rights reserved. V. Tedjasaputra et al. Respiratory Medicine 145 (2018) 57–65 available for gas exchange and alveolar membrane thickness, and pul- 2.5. Cardiopulmonary exercise test monary capillary blood volume (Vc) [11]. During exercise, there is typically recruitment and distention of pulmonary capillaries, which act Subjects then performed an incremental cycle test (Ergoselect II to increase Vc and Dm, as well as blunt the rise in pulmonary artery 1200 Ergoline, Blitz, Germany) to determine [19]. Inspiratory pressure with exercise [12]. Capillary recruitment would reduce capacity (IC) measurements were performed at rest and every 2 min deadspace ventilation, and therefore reduce VE/VCO 2 during exercise. during the exercise test [20]. Participants rated the intensity of their Compared to age-matched controls, resting diffusing capacity (DLCO) is “breathing discomfort” and “leg discomfort” at rest and every 2 min reduced by 20% in mild COPD [3], and lower resting DLCO is corre- during exercise using the modified 10-point Borg Scale [21]. The nadir lated with lower VO2peak and a greater exercise nadir in mild was determined as the lowest 30-s value achieved during the COPD [13]. However, it is unclear how DLCO, and its components Vc incremental exercise test. and Dm, respond to exercise in mild COPD. Accordingly, the purpose of this study was to examine the DLCO, Vc, and Dm responses to exercise 2.6. Exercise diffusing capacity in mild COPD. We hypothesized that mild COPD would have a blunted DLCO and Vc response to exercise, suggesting pulmonary vascular DLCO was determined using the single-breath breath-hold tech- dysfunction which would contribute to deadspace ventilation (i.e. nique [18] at baseline and during exercise. The relative exercise in- ) and dyspnea. With an increase in the severity of COPD, there tensities of 50% and 80% of were determined by linear re- may be further changes to the pulmonary vascular response to exercise. gression of work rate and VO2 obtained during the incremental exercise Therefore, 10 moderate COPD patients were also studied to describe the test on the preliminary day, similar to our previous work [15]. Prior to continuum of responses to increasing obstruction severity. data collection, subjects were coached to avoid Müller or Valsalva

maneuvers during the breath hold maneuver. To ensure alveolar PO2 2. Material and methods was stable, each subject was given five breaths of gas at the respective

FIO2 for each DLCO test gas [15,16]. Diffusing capacity was adjusted for 2.1. Ethical approval hemoglobin concentration [22] (HemoCue 201+, HemoCue AB, An- gelholm, Sweden) as well as CO [23], which was estimated in real-time This study was approved by the Human Research Ethics Board of the using non-invasive CO-oximetry (Radical 7, Masimo, Irvine, CA, USA), University of Alberta (protocol #54658). All subjects gave written, in- for each exercise workload. formed consent to participate in the study. To assess Vc and Dm, multiple-FIO2 DLCO breath holds were performed with three different FIO2 values (0.21, 0.40, 0.60) during steady 2.2. Subjects state exercise at 40 W, 50% and 80% of similar to our previous work [15,16]. Steady state was determined by exercising for a minimum

Seventeen patients with mild COPD (post-bronchodilator FEV1/ of 3 min at the previously determined intensity, with a change in heart FVC < lower limit of normal (LLN) and FEV1 ≥ 80% predicted) and 17 rate ≤3 bpm in 1 min. Vc and Dm were calculated from the equation 1= 1 + 1 age-, sex-, and height-matched control subjects, as well as 10 moderate [11] DLCO Dm CO × Vc , and theta was calculated using the equa- 1 COPD patients (FEV1 50–80% predicted) were recruited for this study. = ×POA 2 + tion CO , with the values α = 0.0058 and β = 0.73, Four mild and six moderate COPD patients were current smokers, and based on moderate red cell permeability [15,16,24]. For each workload, were instructed to withhold from smoking 24 h prior to testing day. the relationship between 1/DLCO and 1/theta for the three FIO2 values The sample size was derived from a similar study that demonstrated were plotted, and a regression equation was then calculated, determining a 12% difference in pulmonary capillary blood volume between oldand the values of 1/Vc as the slope and 1/Dm as the y-intercept of the re- young male subjects during exercise [14], and a priori power calcula- sulting linear equation. The minimum acceptable r2 was 0.95, and DLCO tion to determine that a sample size of n = 15 was appropriate to detect maneuvers were repeated when r2 fell below this value, or if quality a between-group difference (α = 0.05, β = 0.8). assurance standards as discussed below were not met.

2.3. Study overview 2.7. Quality assurance of breath hold maneuvers

In total, subjects completed three visits to the laboratory. In visit Recognizing the challenge of a 6-s breath hold during exercise in one, subjects completed full pulmonary function testing followed by an COPD patients, we spent considerable time and effort in training sub- incremental cardiopulmonary exercise test to determine on a jects to the proper breath holding technique and timing to adhere to our cycle ergometer. At least 48 h later, subjects returned to the laboratory quality standards. Mouth pressure measured at the mass flow sensor for the exercise DLCO sessions. Subjects performed multiple-FIO2 DLCO was monitored throughout the breath hold to be within ± 20 mmHg to maneuvers [11] at rest and during steady state exercise [15–17]. Par- ensure no Valsalva or Müller maneuvers and an open glottis. Acceptable ticipants returned at least one week later for the echocardiography breath hold time during trials was 6.0 ± 0.3 s. Exhaled methane was session to estimate cardiac output and pulmonary artery systolic pres- monitored to ensure a horizontal tracing throughout the exhalation, sure at rest and during exercise. which suggests adequate mixing of test gas in resident alveolar air. The

breath hold was repeated if measured VA was not within 5% of the 2.4. Pulmonary function testing previous trial, or if any of the quality standards described above were not met. Participants completed a full pulmonary function test, including pre- and post-bronchodilator spirometry (400 μg Salbutamol), de- 2.8. Echocardiography termination of lung volumes by plethysmography, and resting DLCO (V62J Body Plethysmograph, SensorMedics, Yorba Linda, CA) in ac- At least one week after the exercise DLCO, participants returned for cordance with clinical practice guidelines [18]. an exercise echocardiogram. All echocardograms were performed by

58 V. Tedjasaputra et al. Respiratory Medicine 145 (2018) 57–65 one experienced sonographer (Vivid Q, GE Healthcare, Fairfield, CT, Table 1 USA) and were later interpreted by a cardiologist blind to experimental Descriptive characteristics, pulmonary function, and resting lung volumes. conditions (ECHOPac, GE Healthcare, Fairfield, CT, USA). Cardiac Control Mild COPD P values output (Q̇) was determined using the cross-sectional area and pulse wave velocity-time integral of the left ventricular outflow tract multi- Subjects (male/female) 7/10 7/10 plied by heart rate. Pulmonary artery systolic pressure (PASP) was es- Age 63 ± 13 65 ± 13 0.770 Height, m 1.68 ± 0.09 1.66 ± 0.08 0.524 timated using peak tricuspid regurgitant jet velocity using the equation Weight, kg 76.7 ± 12.2 71.2 ± 13.8 0.226 2 PASP = 4(VTR) + RAP, where VTR is the peak tricuspid regurgitant jet BMI 27.0 ± 2.9 25.6 ± 3.6 0.217 velocity in meters per second, and RAP is right atrial pressure in mmHg Smoking history, pack-years 3 ± 4 36 ± 19 < 0.001* . −1 [25]. RAP was determined after assessment of inferior vena cava (IVC) VO2peak Abs,L min 2.33 ± 0.79 1.73 ± 0.56 0.017* diameter, measured proximal to the entrance of hepatic veins. The , % pred 124 ± 35 95 ± 19 0.005* . −1. −1 value of RAP was 5 mmHg if the diameter of the IVC was ≤2.1 cm and VO2peak Rel, mL kg min 30.1 ± 8.8 24.3 ± 6.2 0.034* collapses > 50% with a sniff test; if IVC diameter was > 2.1 cmand , % pred 119 ± 36 97 ± 25 0.038* collapsed < 50% with a sniff test, RAP value was 10 mmHg[25]. SPO2 rest, % 98.1 ± 1.1 96.4 ± 2.7 0.026* SPO2 peak, % 96.0 ± 3.2 95.5 ± 2.81 0.653 V /VCO 28.4 ± 3.2 34.3 ± 5.6 0.001* 2.9. Statistical analysis E 2 nadir FEV1,L 2.97 ± 0.85 2.37 ± 0.47 0.017*

FEV1, % pred 110 ± 14 94 ± 11 0.001*

Pulmonary function variables presented as percent predicted are FEV1, z-score 0.39 ± 1 −0.71 ± 0.62 0.001* based on normative data [26]. Spirometric data are also presented as z- FVC, L 3.99 ± 1.10 3.68 ± 0.74 0.335 scores using the Global Lungs Initiative 2012 equations [27]. For all FVC, % pred 106 ± 16 106 ± 9 0.901 FVC, z-score 0.81 ± 1.25 0.41 ± 0.63 0.257 inferential analyses, the probability of a type I error was set at 0.05. FEV1/FVC, % 75 ± 5 64 ± 4 < 0.001*

Values are expressed as means ± standard error of the mean unless FEV1/FVC, % pred 101 ± 7 83 ± 6 < 0.001* otherwise indicated. Unpaired Student's T-test were used to compare FEV1/FVC, z-score −0.50 ± 0.66 −1.77 ± 0.61 < 0.001* subject characteristics (age, height, weight, etc.) between groups. Sta- TLC, L 6.12 ± 1.37 5.72 ± 1.40 0.410 tistical analysis was performed using two-way repeated measures TLC, % pred 103 ± 16 100 ± 13 0.484 FRC, L 3.17 ± 0.67 3.42 ± 0.71 0.303 ANOVA (SigmaPlot, v.13, Systat Software, San Jose, CA, USA) to FRC, % pred 103 ± 16 117 ± 16 0.010* evaluate the effect of mild COPD (mild COPD vs. control) on thedif- RV, L 1.95 ± 0.56 2.07 ± 0.84 0.635 fusing capacity response (dependent variables: DLCO, Dm, Vc) to ex- RV, % pred 91 ± 26 100 ± 37 0.447 DLCO, mL.min−1.mmHg−1 22.9 ± 4.0 18.5 ± 6.1 0.018* ercise (4 levels: baseline, 40W, 50%, and 80% of VO2peak). Where a DLCO (% pred) 94 ± 10 77 ± 17 0.001* main effect of disease state was found, a Holm-Sidak post-hoc testwas performed to locate differences between groups and exercise intensities. Values expressed as Means ± SD. VO2peak, peak oxygen consumption; SPO2,

The moderate COPD response was compared to control at 3 levels: oxygen pulse saturation; VE/VCO 2 nadir, ventilatory equivalent for CO2 at baseline, 40W and 80% of using two-way repeated measures lowest point during graded exercise test; FEV1, forced expired volume in 1 s; ANOVA. The 50% of exercise workload was removed from the FVC, forced vital capacity; TLC, total lung capacity; FRC, functional residual moderate COPD analysis because the workload was nearly identical to capacity; RV, residual volume assessed by plethysmography; DLCO, diffusing that of 40W in these patients. capacity at rest; Z-score calculated using Global Lung Initiative 2012 equations To further examine the COPD response, the mild COPD group was [27]. *Significantly different from control. subdivided by the nadir VE/VCO 2 observed during the incremental exercise test, defined as < 34 vs. ≥ 34 [6], as well as whether dynamic hyperinflation developed with exercise, as defined byade- mild COPD at 50% (P = 0.006) and 80% of compared to control crease in their IC from baseline to peak exercise [28,29]. One-way (P = 0.016). PASP and Q were not significantly different between ANOVA was used to determine differences between these subgroups at groups at rest or during exercise, although there were fewer compar- an identical workload of 40W, at which the metabolic demand would isons of PASP due to a lack of subjects with sufficient quality of tri- be similar. An additional comparison was made at the relative workload cuspid regurgitant jet Doppler envelope during exercise (control of 50% of as it was the closest data point to the observed nadir n = 11, mild COPD n = 9). Dyspnea was not different between groups (60 ± 10% of ). at rest. During exercise, dyspnea was greater in the mild COPD group as compared to controls at 40 W (P = 0.002, Table 3). 3. Results 3.2. Exercise diffusing capacity Descriptive characteristics of the study sample are provided in Table 1. As expected, mild COPD patients had lower FEV (P = 0.017) 1 At baseline, DLCO was lower in the mild COPD group compared to and greater functional residual capacity % compared to control pred control (P = 0.049). With incremental exercise, both mild COPD and (P = 0.010), but there was no significant difference in total lung ca- controls increased DLCO with increasing oxygen consumption pacity (P = 0.410, Table 1). (P < 0.001). However, DLCO was higher in control subjects at all exercise intensities (Fig. 1A). 3.1. Cardiopulmonary response to exercise in mild COPD

Cardiopulmonary responses to steady state exercise at 40W, 50%, 3.3. Pulmonary capillary blood volume and 80% of are given in Table 2. Absolute VO2 was not different between mild COPD and controls at rest (P = 0.645) or at 40W At baseline, Vc was lower in the mild COPD group compared to (P = 0.294), but was lower in the mild COPD group at 50% (P = 0.001) control (P = 0.033, Fig. 1B). With incremental exercise, both mild and 80% of (P < 0.001, Table 2). Similarly, HR was lower in COPD and control groups increased Vc with increasing oxygen

59 V. Tedjasaputra et al. Respiratory Medicine 145 (2018) 57–65

Table 2 Cardiopulmonary response to exercise in mild COPD.

Baseline 40W 50% 80%

PO, Control 0 40 ± 0 75 ± 40 127 ± 36 W COPD 0 40 ± 0 46 ± 20 83 ± 32

P NA NA < 0.001* < 0.001*

VO2, Control 0.34 ± 0.09 0.94 ± 0.21 1.27 ± 0.36 1.92 ± 0.56 L.min−1 COPD 0.29 ± 0.06 0.83 ± 0.20 0.91 ± 0.30 1.36 ± 0.43

P 0.645 0.294 0.001* < 0.001*

VCO2, Control 0.31 ± 0.07 0.82 ± 0.20 1.16 ± 0.30 1.98 ± 0.53 L.min−1 COPD 0.25 ± 0.05 0.70 ± 0.20 0.76 ± 0.28 1.33 ± 0.40

P 0.573 0.236 < 0.001* < 0.001*

PETO2, Control 96.4 ± 17.9 91.4 ± 15.4 91.4 ± 15.5 99.3 ± 4.0 mmHg COPD 101.8 ± 4.0 97.0 ± 3.5 96.7 ± 3.9 98.6 ± 5.3

P 0.101 0.091 0.114 0.844

PETCO2, Control 32.4 ± 3.6 35.2 ± 3.2 37.4 ± 3.7 36.5 ± 4.0 mmHg COPD 30.0 ± 1.7 33.4 ± 2.6 34.2 ± 3.2 35.3 ± 3.9

P 0.056 0.161 0.013* 0.344

VE, Control 13.3 ± 2.7 26.4 ± 6.7 35.3 ± 9.4 60.0 ± 18.5 L.min−1 COPD 12.1 ± 2.1 25.9 ± 6.4 27.8 ± 8.5 45.0 ± 11.9

P 0.723 0.876 0.040* < 0.001*

VE/VCO 2 Control 44 ± 5 35 ± 3 31 ± 3 30 ± 3 COPD 51 ± 9 38 ± 5 37 ± 5 35 ± 7

P < 0.001* 0.005* 0.001* 0.009*

SPO2, Control 98.1 ± 1.1 97.2 ± 1.7 97.0 ± 1.7 96.1 ± 2.3 % COPD 96.4 ± 2.7 96.1 ± 2.6 96.0 ± 2.9 94.9 ± 3.4

P 0.075 0.244 0.271 0.220

Q̇, Control 3.47 ± 0.73 5.50 ± 1.24 7.23 ± 1.88 8.99 ± 2.27 L.min−1 COPD 3.69 ± 1.23 5.69 ± 1.33 5.96 ± 0.94 8.25 ± 2.31

P 0.651 0.800 0.053 0.440

HR, Control 77 ± 10 99 ± 14 112 ± 15 137 ± 18 bpm COPD 76 ± 0 96 ± 11 100 ± 10 119 ± 13

P 0.684 0.558 0.006* 0.001*

SV, Control 48 ± 9 58 ± 13 66 ± 16 68 ± 18 mL COPD 48 ± 15 62 ± 14 63 ± 13 73 ± 23

P 0.949 0.492 0.543 0.577

PASP, Control 22.8 ± 10.4 34.4 ± 15.7 42.4 ± 18.5 50.0 ± 19.1 mmHg COPD 23.8 ± 7.5 32.4 ± 12.3 34.1 ± 13.0 35.14 ± 14.5

P 0.807 0.770 0.362 0.124

Values means ± SD. PO, power output; VO2, oxygen consumption; VCO2, carbon dioxide production; PETO2, partial pressure of end tidal O2;PETCO2, partial pressure of end tidal CO2; VE, minute ventilation; VE/VCO 2, ventilatory equivalent for carbon dioxide; SPO2, oxygen pulse saturation; Q̇, cardiac output; HR, heart rate; SV, stroke volume; PASP, pulmonary artery systolic pressure. *Significantly different than control. consumption (P < 0.001). During exercise, the mild COPD group had increasing VO2 (P < 0.001), but Dm was not significantly different significantly lower Vc at 40W (P = 0.033), 50% (P = 0.029), and 80% between groups during exercise at any intensity (P=0.269, Fig. 1C). of VO2peak (P = 0.010) compared to control. 3.5. Grouping the mild COPD response by ventilatory inefficiency

3.4. Membrane diffusing capacity VE/VCO 2 was greater in mild COPD at rest and during all exercise intensities (Table 2). Furthermore, the mean nadir was sig- At baseline, Dm was not different between mild COPD and controls nificantly higher in mild COPD compared to control (Table 1). When (P = 0.643). With incremental exercise, both groups increased Dm with the mild COPD patients were split based on their nadir , nine

60 V. Tedjasaputra et al. Respiratory Medicine 145 (2018) 57–65

Table 3 Lung volumes during exercise in mild COPD.

Baseline 40W 50% 80%

VA, Control 5.49 ± 1.37 5.56 ± 1.41 5.52 ± 1.38 5.56 ± 1.47 L Mild COPD 4.76 ± 1.04 4.90 ± 1.03 4.78 ± 1.05 4.95 ± 0.96

P 0.090 0.134 0.090 0.173

VT, Control 0.79 ± 0.18 1.21 ± 0.26 1.50 ± 0.33 2.13 ± 0.52 ∗ L Mild COPD 0.68 ± 0.16 1.21 ± 0.37 1.36 ± 0.54 1.76 ± 0.40

P 0.352 0.992 0.234 0.003*

IC, Control 2.65 ± 0.67 2.65 ± 0.73 2.86 ± 0.73 2.80 ± 0.63 ∗ L Mild COPD 2.48 ± 0.75 2.58 ± 0.79 2.55 ± 0.72 2.38 ± 0.74

P 0.502 0.349 0.118 0.050*

IC/TLC, Control 43 ± 7 46 ± 7 48 ± 7 47 ± 5 % Mild COPD 44 ± 13 46 ± 13 45 ± 11 42 ± 9

P 0.737 0.994 0.397 0.104

IRV, Control 1.86 ± 0.68 1.60 ± 0.77 1.45 ± 0.59 0.78 ± 0.37 L Mild COPD 1.80 ± 0.67 1.36 ± 0.76 1.20 ± 0.51 0.69 ± 0.45

P 0.797 0.267 0.232 0.654

Dyspnea Control 0.0 ± 0.0 1.6 ± 0.6 2.7 ± 0.8 4.4 ± 0.7 ∗ Mild COPD 0.2 ± 0.6 2.5 ± 0.9 2.9 ± 0.9 4.4 ± 1.2

P 0.374 0.002* 0.655 1.000

VA, alveolar volume; VT, tidal volume; IC, inspiratory capacity; IC/TLC, inspiratory capacity as a percent of total lung capacity; IRV, inspiratory reserve volume; Dyspnea, Borg Scale of Breathlessness (1-10). *Significantly different than control, P < 0.05. mild COPD patients had a nadir VE/VCO 2 ≥ 34, and eight mild COPD 4. Discussion patients had a nadir < 34. Of note, there were no control subjects who had a nadir ≥ 34. During exercise at 40W, the The present study examined the diffusing capacity, pulmonary ca- high group had lower Vc (P = 0.030) and greater dyspnea pillary blood volume, and membrane diffusing capacity responses to (P = 0.012), but Dm was not different between subgroups (P = 0.350). exercise in mild COPD patients and age-, sex-, and height-matched VO At 50% of 2peak, pulmonary capillary blood volume was lower in the controls. DLCO was lower in mild COPD at rest and during exercise high group compared to control (P = 0.004), but there was no compared to controls as a result of lower Vc. A reduced Vc, and thus difference in Dm (P = 0.359) or dyspnea (P = 0.540) between groups lower surface area for gas exchange would increase deadspace venti- (see Fig. 2). lation and lead to increased ventilatory demand (i.e. greater ) to maintain adequate alveolar ventilation. In keeping with this idea, the current study found that in addition to the reduced exercise Vc, the 3.6. Grouping the mild COPD response by hyperinflation mild COPD group had both greater and dyspnea during exercise at 40W compared to controls. As an additional comparison, Dynamic hyperinflation with exercise was observed in 11 mild moderate COPD patients were also studied, and these patients failed to COPD patients (ΔIC from rest = −0.35 ± 0.25 L), while six did not increase Vc with exercise, suggesting further pulmonary vascular im- hyperinflate with exercise (ΔIC = 0.35 ± 0.18 L). When mildCOPD pairment with greater COPD severity. These data suggest that despite patients were grouped by level of dynamic hyperinflation, there were the relatively mild perturbation in airflow obstruction (i.e. FEV1), mild no between-group differences in Vc (P = 0.101) or Dm (P = 0.898) COPD patients exhibit early pulmonary vascular impairment that con- during exercise at 40W, and no differences in Vc (P = 0.596) or Dm tributes to an exaggerated ventilatory and dyspnea response to exercise. (P = 0.998) during exercise at 50% of 4.1. The pulmonary vasculature in mild COPD

3.7. Moderate COPD In both controls and mild COPD patients, DLCO rises during exercise to meet the increased metabolic demand in part through an increase in Additional information for the moderate COPD group can be found Vc, which reflects recruitment and distension of pulmonary capillaries. in the online supplement. Mean FEV1 in the moderate COPD group was The slope of Vc response from rest to 40W appears similar between mild 69 ± 8%%pred, and moderate patients had greater percent predicted COPD and controls, suggesting that COPD patients started at a lower functional residual capacity %pred (P = 0.006) and residual volume Vc. The comparison during exercise at 40 W allows us to evaluate the (P = 0.003) compared to control. At baseline, the moderate COPD Vc and Dm response between groups at the same metabolic demand, group had lower DLCO (Supplement Fig. 1A, P = 0.023) compared to where the O2 diffusion requirement would be the same in both mild controls. Compared to rest, exercise intensity increased DLCO COPD and controls groups. Given that cardiac output and VO2 were not (P < 0.001), secondary to increased Dm (Supplement Fig. 1C, different between groups, it can be concluded that the lower Vcob- P = 0.003). Moderate COPD patients failed to demonstrate an increase served in mild COPD at 40W is independent of global pulmonary blood in Vc with exercise (Supplement Fig. 1B, P = 0.222). flow or2 O requirement. The lower Vc in mild COPD could be a result of

61 V. Tedjasaputra et al. Respiratory Medicine 145 (2018) 57–65

inefficiency [7]. Taken together, despite the increase in Vc from rest to exercise, these data suggest that mild COPD patients likely have pulmonary vascular dysfunction or destruction which prevents the appropriate DLCO response to exercise. The reduction in Vc could also be a consequence of emphysema, which would diminish gas exchange surface area and increase deadspace ventilation [32,33]. Our finding of reduced Vc with a high VE/VCO 2 is consistent with previous research showing that COPD patients with greater emphysema severity have reduced resting diffusing capacity and greater ventilatory inefficiency during exercise [32]. In- terestingly, pulmonary capillary blood flow has also found to be reduced in non-emphysematous areas of the lung, suggesting microvascular impairment within an intact vascular network in mild COPD [8]. However, the current study was not able to distinguish vascular impairment from vascular destruction (i.e. emphysema). Taken together, these data suggest that mild COPD patients likely have impairment or destruction of their pulmonary vasculature which attenu- ates the increase in diffusing capacity needed for exercise. Future research should assess pulmonary capillary blood volume in targeted cohorts of COPD with known emphysema to further identify if pulmonary microvascular blood flow is reduced independent of, or con- current to emphysema. It is unlikely that the low capillary blood volume in COPD is an adaptive response to poor regional ventilation, given that matching is typically worse in COPD compared to healthy individuals. In the current study, was greater in the mild COPD group as compared to control at rest and at every exercise intensity (Table 2), which parallels previous work in mild COPD [5,32,34]. When the mild COPD group was subdivided based on a nadir above or below 34, the group with a nadir ≥ 34 had significantly lower exercise Vc. This is consistent with the concept that high areas and deadspace (and thus ventilation inefficiency), are secondary to reduced capillary blood volume in COPD [7].

4.2. Implications for exertional dyspnea

Previous research in mild COPD has demonstrated that exertional dyspnea is the result of an exaggerated ventilatory response to exercise (i.e. increased minute ventilation relative to carbon dioxide production, ) and airflow limitation (i.e. expiratory flow limitation and resulting dynamic hyperinflation) [5,6,28,35,36]. In addition to the lower Vc and increased , mild COPD patients in the current study also reported greater dyspnea at 40W. Further, when mild COPD patients were subdivided by ventilatory inefficiency, the high patients had worse dyspnea (Fig. 2C) compared to controls, while dyspnea in the low COPD group was not significantly different from controls. These findings suggest that impairments in the pulmonary vasculature may be linked to dyspnea in mild COPD. Specifi- cally, the diminished exercise Vc in mild COPD reduces surface area for Fig. 1. A. Diffusion capacity response to exercise. B. Pulmonary Vc response to gas exchange, which increases deadspace ventilation. This increased exercise. C. Membrane diffusing capacity response to exercise. *Mild COPD had deadspace necessitates greater minute ventilation (i.e. greater significantly lower DLCO and Vc compared to control P < 0.05. ) in order to maintain alveolar ventilation, which translates to greater perceived dyspnea in these patients. The lower Vc in mild COPD may result in greater pulmonary vascular resistance, which would thus require greater pulmonary artery diminished blood flow proximal to the capillaries, secondary to vas- pressure for a given cardiac output. However, in the current study, cular dysfunction. Previous research has demonstrated thickening of pulmonary artery systolic pressure was not found to be significantly smooth muscles in the pulmonary arteries and arterioles of mild COPD elevated in mild COPD compared to control. It is likely that the parallel patients [30], which would increase vascular resistance and thus reduce nature of the pulmonary microvasculature provides redundancy such perfusion of the pulmonary capillaries. More recent work using con- that a significant reduction in capillary perfusion does not result ina trast-enhanced magnetic resonance imaging has shown a 30% reduc- larger increase in pulmonary vascular resistance. Unfortunately, high tion in resting blood flow to the pulmonary microvasculature in mild quality Doppler signals of the tricuspid regurgitant jet could only be COPD [8]. Pulmonary vascular dysfunction in COPD patients would obtained in 9 mild COPD patients and 11 controls during exercise, and reduce capillary perfusion and result in more zone 1 conditions [31]. thus the current study may have been underpowered to detect a sig- This would disrupt VA/Q (ventilation-perfusion) matching, thereby in- nificant between-group difference in pulmonary artery systolic pres- creasing deadspace ventilation and thus augmenting ventilatory sure.

62 V. Tedjasaputra et al. Respiratory Medicine 145 (2018) 57–65

Fig. 2. Control and mild COPD grouped by nadir VE/VCO 2. Mild COPD patients were subdivided based on their nadir < 34 and ≥34 during incremental exercise. A. Pulmonary Capillary Blood Volume (Vc) response at 40W, B. Membrane diffusing capacity (Dm) response at 40W, C. Dyspnea response at 40W, D. Vc response at50%of VO2peak, E. Dm response at 50% of , F. Dyspnea response at 50% of . *Significantly different than control, P <0.05.

4.3. Hyperinflation in mild COPD the pulmonary vasculature with advancing COPD (Supplement Figure 1). There is previous evidence that pulmonary microvascular perfusion The increased FRC observed in mild COPD in the current study is reduced in moderate COPD at rest compared to mild COPD [8], and provides some evidence of static lung hyperinflation (Table 1). During previous work using the multiple inert gas elimination technique has exercise, dynamic hyperinflation could mechanically collapse or reduce found increased ventilation-perfusion mismatch at rest in moderate the transverse surface of pulmonary capillaries, thereby decreasing COPD, secondary to areas of high VA/Q ratio [7], consistent with in- capillary blood volume [30,37]. In the current study, exercise Vc was creased deadspace and ventilatory inefficiency. Taken together, these not different between hyperinflators compared to non-hyperinflators, data appear to support the concept that pulmonary vascular dysfunc- which suggests that the lower Vc in mild COPD is unlikely to result from tion worsens with COPD disease severity, consequently reducing the differences in operating lung volume. This finding is consistent with diffusing capacity during exercise. However, it is unclear if the lackof previous work demonstrating no change in VE/VCO 2 when dynamic increase in Vc during exercise in moderate COPD is predominantly due hyperfinflation is reduced with either heliox and/or short-acting β- to emphysema or vascular impairment. agonist administration during exercise in COPD [38–40]. Taken together, the current study supports the view that dynamic hyperinflation 4.5. Considerations does not directly affect Vc nor [40] response during exercise in mild COPD. The partial pressure of end-tidal CO2 (PETCO2) was substituted for partial pressure of arterial CO2 (PaCO2) to calculate theta CO [15,16], 4.4. Moderate COPD which introduces inherent uncertainty of the PAO2 calculation. The arterial-end tidal CO2 difference a-ET(P CO2) is greater in mild COPD The moderate COPD group demonstrated no increase in Vc with compared to healthy controls [5,41], and thus adjusting for the po- increasing exercise intensity, which may indicate further impairment of tentially increased Pa-ETCO2 in COPD would have resulted in a greater

63 V. Tedjasaputra et al. Respiratory Medicine 145 (2018) 57–65 difference in calculated Vc between mild COPD and controls. Appendix A. Supplementary data The multiple-FIO2 DLCO method assumes that alveolar PO2 does not affect Dm or Vc in a significant way[11,24,42]. Despite the brief ex- Supplementary data to this article can be found online at https:// posure to 40% and 60% O2, there is potential concern that hyperoxia doi.org/10.1016/j.rmed.2018.10.015. might disengage hypoxic pulmonary vasoconstriction in COPD. If this were the case in the current study, an increase of blood flowing to the References capillaries would result in an overestimation of Vc on room air, and thus the difference in Vc between healthy controls and COPD during exercise [1] M.B. Parshall, R.M. Schwartzstein, L. Adams, R.B. Banzett, H.L. Manning, would be greater than we observed. Therefore, it is extremely unlikely J. Bourbeau, et al., An official American Thoracic Society statement: update onthe mechanisms, assessment, and management of dyspnea, Am. J. Respir. Crit. Care that the finding of blunted pulmonary capillary blood volume inCOPD Med. (2012) 435–452, https://doi.org/10.1164/rccm.201111-2042ST. is the result of the transient exposure to hyperoxia. [2] M.R. Maleki-Yazdi, C.K. Lewczuk, J.M. Haddon, N. Choudry, N. Ryan, Early de- Another potential methodological limitation is the challenge of the tection and impaired quality of life in COPD GOLD stage 0: a pilot study, COPD 4 (2007) 313–320, https://doi.org/10.1080/15412550701595740. breath-hold maneuver. Exhalation to residual volume could be impeded [3] D. Ofir, P. Laveneziana, K.A. Webb, Y.-M. Lam, D.E. O'Donnell, Mechanisms of with airway obstruction, and thus gas trapping may decrease the dyspnea during cycle exercise in symptomatic patients with GOLD stage I chronic available surface area for diffusion. However, alveolar volume did not obstructive pulmonary disease, Am. J. Respir. Crit. Care Med. 177 (2008) 622–629, change with increasing exercise intensity in either group, suggesting https://doi.org/10.1164/rccm.200707-1064OC. [4] S.J.T. Zanoria, R. ZuWallack, Directly measured physical activity as a predictor of that all subjects were able to reach residual volume prior to each breath hospitalizations in patients with chronic obstructive pulmonary disease, Chron. hold maneuver. Respir. Dis. 10 (2013) 207–213, https://doi.org/10.1177/1479972313505880. This study was designed to compare Vc and Dm during exercise both [5] A.F. Elbehairy, C.E. Ciavaglia, K.A. Webb, J.A. Guenette, D. Jensen, S.M. Mourad, et al., Pulmonary gas exchange abnormalities in mild chronic obstructive pul- at an absolute workload (40 W) and at exercise workloads relative to monary disease. Implications for dyspnea and exercise intolerance, Am. J. Respir. each individual's VO2peak. The 40 W comparison allows us to compare Crit. Care Med. 191 (2015) 1384–1394, https://doi.org/10.1164/rccm.201501- the physiological responses between groups at an identical metabolic 0157OC. [6] J.A. Neder, A. Alharbi, D.C. Berton, F.F. Arbex, D.M. Hirai, K.A. Webb, et al., demand, whereas the relative intensity comparison at 50% and 80% of Exercise ventilatory inefficiency adds to lung function in predicting mortality in allows for a better understanding of the diffusing capacity re- COPD, COPD 13 (2016) 416–424, https://doi.org/10.3109/15412555.2016. sponse to increased oxygen requirement, at similar percentages of 1158801. [7] R. Rodriguez-Roisin, M. Drakulovic, D.A. Rodríguez, J. Roca, J.A. Barberà, . We would suggest that the inclusion of both absolute and re- P.D. Wagner, Ventilation-perfusion imbalance and chronic obstructive pulmonary lative intensity comparisons is a strength of the study. disease staging severity, J. Appl. Physiol. 106 (2009) 1902–1908, https://doi.org/ Lastly, the control group had significantly higher compared 10.1152/japplphysiol.00085.2009. [8] K. Hueper, J. Vogel-Claussen, M.A. Parikh, J.H.M. Austin, D.A. Bluemke, J. Carr, to COPD. Our previous investigation showed that Vc was not different et al., Pulmonary microvascular blood flow in mild chronic obstructive pulmonary between aerobic trained and untrained individuals for a given work rate disease and emphysema. The MESA COPD study, Am. J. Respir. Crit. Care Med. 192 [15]. However, the current study observed lower Vc in COPD compared (2015) 570–580, https://doi.org/10.1164/rccm.201411-2120OC. to controls at the same absolute work rate of 40W, suggesting that the [9] J.A. Barberà, A. Riverola, J. Roca, J. Ramirez, P.D. Wagner, D. Ros, et al., Pulmonary vascular abnormalities and ventilation-perfusion relationships in mild difference in between groups is unlikely to explain the differ- chronic obstructive pulmonary disease, Am. J. Respir. Crit. Care Med. 149 (1994) ences in DLCO and Vc during submaximal exercise. 423–429, https://doi.org/10.1164/ajrccm.149.2.8306040. [10] M.K. Stickland, M.I. Lindinger, I.M. Olfert, G.J.F. Heigenhauser, S.R. Hopkins, Pulmonary gas exchange and acid-base balance during exercise, Comp. Physiol. 3 (2013) 693–739, https://doi.org/10.1002/cphy.c110048. 5. Conclusion [11] F.J. Roughton, R.E. Forster, Relative importance of diffusion and chemical reaction rates in determining rate of exchange of gases in the human lung, with special Patients with mild COPD demonstrated a lower diffusing capacity reference to true diffusing capacity of pulmonary membrane and volume of bloodin the lung capillaries, J. Appl. Physiol. 11 (1957) 290–302. response to exercise compared to control subjects due to reduced pul- [12] J.T. Reeves, J.H. Linehan, K.R. Stenmark, Distensibility of the normal human lung monary capillary blood volume. The lower Vc during exercise appeared circulation during exercise, Am. J. Physiol. Lung Cell Mol. Physiol. 288 (2005) to be associated with increased VE/VCO 2, suggesting pulmonary vas- L419–L425, https://doi.org/10.1152/ajplung.00162.2004. cular impairment leading to ventilatory inefficiency and dyspnea in [13] A.F. Elbehairy, A. Faisal, J.A. Guenette, D. Jensen, K.A. Webb, R. Ahmed, et al., Resting Physiological Correlates of Reduced Exercise Capacity in Smokers with Mild mild COPD. Moderate COPD patients showed further impairment with Airway Obstruction, (2017), pp. 1–10, https://doi.org/10.1080/15412555.2017. an inability to increase Vc during exercise. Despite their relatively 1281901. minor airway obstruction, mild COPD patients appear to have adverse [14] B.J. Taylor, A.R. Carlson, A.D. Miller, B.D. Johnson, Exercise-induced interstitial pulmonary edema at sea-level in young and old healthy humans, Respir. Physiol. changes to their pulmonary vascular response that affects their ability Neurobiol. 191 (2014) 17–25, https://doi.org/10.1016/j.resp.2013.10.012. to perfuse pulmonary capillaries and increase diffusion appropriately [15] V. Tedjasaputra, M.M. Bouwsema, M.K. Stickland, Effect of Aerobic Fitness on during exercise. Capillary Blood Volume and Diffusing Membrane Capacity Responses to Exercise vol 594, (2016), pp. 4359–4370, https://doi.org/10.1113/JP272037. [16] M.M. Bouwsema, V. Tedjasaputra, M.K. Stickland, Are there sex differences in the capillary blood volume and diffusing capacity response to exercise? J. Appl. Financial support Physiol. 122 (2017) 460–469, https://doi.org/10.1152/japplphysiol.00389.2016. [17] V. Tedjasaputra, S. van Diepen, S.É. Collins, W.M. Michaelchuk, M.K. Stickland, Funding for this study was provided by the University of Alberta Assessment of pulmonary capillary blood volume, membrane diffusing capacity, and intrapulmonary arteriovenous anastomoses during exercise, J Vis Exp (2017), Hospital Research Foundation, and the Canadian Respiratory Research https://doi.org/10.3791/54949 e54949–e54949. Network. Vincent Tedjasaputra was supported by a studentship from [18] B.L. Graham, V. Brusasco, F. Burgos, B.G. Cooper, R. Jensen, A. Kendrick, et al., The Lung Association of Alberta and Northwest Territories. 2017 ERS/ATS standards for single-breath carbon monoxide uptake in the lung, Eur. Respir. J. 49 (2017) 1600016, https://doi.org/10.1183/13993003.00016- 2016. [19] S.M. Holm, W. Rodgers, R.G. Haennel, G.F. MacDonald, T.L. Bryan, M. Bhutani, Acknowledgements et al., Effect of modality on cardiopulmonary exercise testing in male andfemale COPD patients, Respir. Physiol. Neurobiol. 192 (2014) 30–38, https://doi.org/10. 1016/j.resp.2013.11.009. The authors gratefully acknowledge the contributions of Desi Fuhr [20] J.A. Guenette, R.C. Chin, J.M. Cory, K.A. Webb, D.E. O'Donnell, Inspiratory capacity (Clinical Physiology Laboratory, University of Alberta), ultra- during exercise: measurement, analysis, and interpretation, Pulm. Med. 2013 sonographers Qiong Yin and Bernadette Bellosillo, and the medical (2013) 956081, https://doi.org/10.1155/2013/956081. support of Drs. Marc Bibeau, Ben Chiam, Peter Wei, Dominic Carney [21] G.A. Borg, Psychophysical bases of perceived exertion, Med. Sci. Sports Exerc. 14 (1982) 377–381. (The Lung Health Clinic, Edmonton, AB), and Dr. Ron Damant (Division [22] R.M. Marrades, O. Diaz, J. Roca, J.M. Campistol, J.V. Torregrosa, J.A. Barberà, of Pulmonary Medicine, University of Alberta). et al., Adjustment of DLCO for hemoglobin concentration, Am. J. Respir. Crit. Care

64 V. Tedjasaputra et al. Respiratory Medicine 145 (2018) 57–65

Med. 155 (2011) 236–241, https://doi.org/10.1164/ajrccm.155.1.9001318. Emphysema on thoracic CT and exercise ventilatory inefficiency in mild-to-mod- [23] B.L. Graham, J.T. Mink, D.J. Cotton, Effects of increasing carboxyhemoglobin on erate COPD, COPD 14 (2017) 210–218, https://doi.org/10.1080/15412555.2016. the single breath carbon monoxide diffusing capacity, Am. J. Respir. Crit. Care Med. 1253670. 165 (2002) 1504–1510, https://doi.org/10.1164/rccm.2108071. [33] M. Kirby, D. Pike, D.D. Sin, H.O. Coxson, D.G. McCormack, G. Parraga, COPD: do [24] M.L. Ceridon, K.C. Beck, T.P. Olson, J.A. Bilezikian, B.D. Johnson, Calculating al- imaging measurements of emphysema and airway disease explain symptoms and veolar capillary conductance and pulmonary capillary blood volume: comparing the exercise capacity? Radiology 277 (2015) 872–880, https://doi.org/10.1148/radiol. multiple- and single-inspired oxygen tension methods, J. Appl. Physiol. 109 (2010) 2015150037. 643–653, https://doi.org/10.1152/japplphysiol.01411.2009. [34] J.A. Neder, F.F. Arbex, M.C.N. Alencar, C.D.J. O'Donnell, J. Cory, K.A. Webb, et al., [25] L.G. Rudski, W.W. Lai, J. Afilalo, L. Hua, M.D. Handschumacher, Exercise ventilatory inefficiency in mild to end-stage COPD, Eur. Respir. J.45 K. Chandrasekaran, et al., Guidelines for the echocardiographic assessment of the (2014) 377–387, https://doi.org/10.1183/09031936.00135514. right heart in adults: a report from the american society of echocardiography en- [35] D.E. O'Donnell, K.A. Webb, The major limitation to exercise performance in COPD is dorsed by the european association of echocardiography, a registered branch of the dynamic hyperinflation, J. Appl. Physiol. 105 (2008), https://doi.org/10.1152/ european society of cardiology, and the canadian society of echocardiography, J. japplphysiol.90336.2008b 753–5– discussion 755–7. Am. Soc. Echocardiogr. 23 (2010), https://doi.org/10.1016/j.echo.2010.05.010 [36] P. Laveneziana, K.A. Webb, K. Wadell, J.A. Neder, D.E. O'Donnell, Does expiratory 685–713– quiz 786–8. muscle activity influence dynamic hyperinflation and exertional dyspnea inCOPD? [26] C. Gutierrez, R.H. Ghezzo, R.T. Abboud, M.G. Cosio, J.R. Dill, R.R. Martin, et al., Respir. Physiol. Neurobiol. 199 (2014) 24–33, https://doi.org/10.1016/j.resp. Reference values of pulmonary function tests for Canadian Caucasians, Cancer Res. 2014.04.005. J. 11 (2004) 414–424. [37] P. Harris, N. Segel, I. Green, E. Housley, The influence of the airways resistance and [27] P.H. Quanjer, S. Stanojevic, T.J. Cole, X. Baur, G.L. Hall, B.H. Culver, et al., Multi- alveolar pressure on the pulmonary vascular resistance in chronic bronhcitis, ethnic reference values for spirometry for the 3-95-yr age range: the global lung Cardiovasc. Res. 2 (1968) 84–92. function 2012 equations, Eur. Respir. J. 40 (2012) 1324–1343, https://doi.org/10. [38] P. Palange, G. Valli, P. Onorati, R. Antonucci, P. Paoletti, A. Rosato, et al., Effect of 1183/09031936.00080312. heliox on lung dynamic hyperinflation, dyspnea, and exercise endurance capacity in [28] J.A. Guenette, K.A. Webb, D.E. O'Donnell, Does dynamic hyperinflation contribute COPD patients, J. Appl. Physiol. 97 (2004) 1637–1642, https://doi.org/10.1152/ to dyspnoea during exercise in patients with COPD? Eur. Respir. J. 40 (2012) japplphysiol.01207.2003. 322–329, https://doi.org/10.1183/09031936.00157711. [39] M.M. Peters, K.A. Webb, D.E. O'Donnell, Combined physiological effects of [29] D.E. O'Donnell, S.M. Revill, K.A. Webb, Dynamic hyperinflation and exercise in- bronchodilators and hyperoxia on exertional dyspnoea in normoxic COPD, Thorax tolerance in chronic obstructive pulmonary disease, Am. J. Respir. Crit. Care Med. 61 (2006) 559–567, https://doi.org/10.1136/thx.2005.053470. 164 (2001) 770–777, https://doi.org/10.1164/ajrccm.164.5.2012122. [40] A.F. Elbehairy, K.A. Webb, P. Laveneziana, N.J. Domnik, J.A. Neder, [30] J.L. Wright, L. Lawson, P.D. Paré, R.O. Hooper, D.I. Peretz, J.M. Nelems, et al., The D.E. O'Donnell, et al., Acute bronchodilator therapy does not reduce wasted ven- structure and function of the pulmonary vasculature in mild chronic obstructive tilation during exercise in COPD, Respir. Physiol. Neurobiol. (2018) 1–29, https:// pulmonary disease. The effect of oxygen and exercise, Am. Rev. Respir. Dis.128 doi.org/10.1016/j.resp.2018.03.012. (1983) 702–707, https://doi.org/10.1164/arrd.1983.128.4.702. [41] J.S. Williams, T.G. Babb, Differences between estimates and measured PaCO2 [31] J.B. West, C.T. Dollery, Distribution of blood flow and ventilation-perfusion ratio in during rest and exercise in older subjects, J. Appl. Physiol. 83 (1997) 312–316, the lung, measured with radioactive carbon dioxide, J. Appl. Physiol. 15 (1960) https://doi.org/10.1164/ajrccm/136.5.1285. 405–410. [42] C.C.W. Hsia, Recruitment of lung diffusing capacity: update of concept and appli- [32] J.H. Jones, J.T. Zelt, D.M. Hirai, C.V. Diniz, A. Zaza, D.E. O'Donnell, et al., cation, Chest 122 (2002) 1774–1783.