Combining Dynamic Hyperinflation with Dead Space Volume during Maximal Exercise in Patients with Chronic Obstructive Pulmonary Disease.

Physiological dead space volume (VD) and dynamic hyperinflation (DH) are two different types of abnormal pulmonary physiology. Although they both involve lung volume, their combination has never been advocated, and thus their effect and implication are unclear. This study aimed (1) to combine VD and DH, and (2) investigate their relationship and clinical significance during exercise, as well as (3) identify a noninvasive variable to represent the VD fraction of tidal volume (VD/VT). Forty-six male subjects with chronic obstructive pulmonary disease (COPD) and 34 healthy male subjects matched for age and height were enrolled. Demographic data, lung function, and maximal exercise were investigated. End-expiratory lung volume (EELV) was measured for the control group and estimated for the study group using the formulae reported in our previous study. The VD/VT ratio was measured for the study group, and reference values of VD/VT were used for the control group. In the COPD group, the DHpeak/total lung capacity (TLC, DHpeak%) was 7% and the EELVpeak% was 70%. After adding the VDpeak% (8%), the VDDHpeak% was 15% and the VDEELVpeak% was 78%. Both were higher than those of the healthy controls. In the COPD group, the VDDHpeak% and VDEELVpeak% were more correlated with dyspnea score and exercise capacity than that of the DHpeak% and EELV%, and had a similar strength of correlation with minute ventilation. The VTpeak/TLC (VTpeak%), an inverse marker of DH, was inversely correlated with VD/VT (R2 ≈ 0.50). Therefore, we recommend that VD should be added to DH and EELV, as they are physiologically meaningful and VTpeak% represents not only DH but also dead space ventilation. To obtain VD, the VD/VT must be measured. Because obtaining VD/VT requires invasive arterial blood gases, further studies on noninvasive predicting VD/VT is warranted.


Introduction
In the alveolar dead space (V D ) of the three component (Riley) model [1], if alveolar V D exists, residual volume is expected to increase, potentially causing air trapping and hyperinflation of the lung. However, the physiological V D refers to ventilation not involved in gas exchange and involved in unperfused or underperfused alveoli [2] and includes anatomical and alveolar V D s [1]. Acute dynamic hyperinflation (DH) refers to a temporary increase in operating lung volume above the resting value, i.e., end-expiratory lung volume at peak exercise (EELV peak ) [3][4][5][6] minus resting EELV (EELV rest ) [7]. Because the definitions of alveolar V D and DH are different, physiological V D would not cause DH, and thus their relationship is unclear.
The physiological V D /tidal volume ratio (V D /V T ) can be calculated using the Bohr-Enghoff equation [2]. Therefore, V D can be considered to be a part of V T , and anatomical V D can be assumed to occur at the beginning of V T . Accordingly, as EELV is immediately followed by tidal breathing, beginning V D not included in EELV should be added.
In patients with chronic obstructive pulmonary disease (COPD), the V D /V T is often highly increased at rest and usually mildly decreased during exercise as compared with normal subjects. This phenomenon has been hypothesized to be due to a small increase in V D and a small expansion in V T , as V T is constrained by DH. V T "floats" above DH and is concomitantly limited by the ceiling of total lung capacity (TLC) and causes reductions in inspiratory reserve volume and O'Donnell threshold [8]. This is in contrast to healthy subjects, in whom a small change in V D and a large increase in V T are usually noted.
Although the definition and mechanism of V D and DH are quite different, both are volumes; DH, i.e., EELV peak minus EELV rest has been reported to be correlated with the V D /V T ratio [3,9,10] (see the Appendix A Table A1), and EELV peak has been shown to be inversely related to V Tpeak /TLC (V Tpeak %) [11]. Hence, the aims of this study were as follows: (1) to combine V D with DH; (2) to investigate the relationship between DH and V D ; (3) to investigate the relationship between V D DH and dyspnea, exercise capacity, and ventilation capability; and (4) to investigate the relationship between V D /V Tpeak and V Tpeak % during maximal exercise in order to find a surrogate for V D /V Tpeak , which is an invasive variable. This study could help clinicians better understand the relative positions of EELV, DH, V D , and V T in TLC, and show that V D and DH together are unfavorable lung volumes during exercise [9,10]. Using the easily calculated V Tpeak % during exercise, testing could possibly reflect the invasively measured V D /V Tpeak , and thus clinicians could use the V Tpeak % as an indicator of DH and also V D /V Tpeak . To the best of our knowledge, this is the first study to integrate the concept of dead space ventilation and DH during exercise.

Study Design
In this observational cross-sectional study, we analyzed lung function data and cardiopulmonary exercise with inspiratory capacity maneuver data from subjects with COPD and healthy controls at the Chung Shan Medical university hospital. The relationships between V Tpeak % and V D /V T were investigated in the subjects with COPD. V D , V T , and EELV as % of TLC were illustrated using percentages. Signed informed consent was obtained from each participant. The local Institutional Review Board of the institution (CS16174) approved this study, which was conducted in compliance with the Declaration of Helsinki.

Subjects
Subjects aged ≥40 years without any chronic diseases including uncontrolled diabetes mellitus, uncontrolled hypertension, anemia (hemoglobin <13 g/dL), and no acute illnesses in the recent period of 1 month were enrolled. Anthropometric measurements, leisure/sports activities, and cigarette smoking were recorded. Subjects with a body mass index ≤18 kg/m 2 or ≥32 kg/m 2 or with laboratory findings of cardiovascular, hematological, metabolic, or neuromuscular diseases were excluded. All of the participants performed lung function and cardiopulmonary exercise tests (CPET). Subjects who did not have sufficient motivation to perform CPET were also excluded.

Study Group
Male adult subjects who underwent spirometry, plethysmography, and diffusing capacity were enrolled if their forced expired volume in one second (FEV 1 )/forced expired capacity (FVC) was <0.7 [12]. The diagnosis of COPD was made according to the global initiative for chronic obstructive lung disease (GOLD) criteria [12]. As few female subjects met the criteria of COPD, they were not included in this study.

Control Group
A group of healthy subjects was recruited among the hospital staff and from the local community through personal contacts. Healthy male subjects reported no chronic diseases.

Functional Daily Activity
The oxygen cost diagram (OCD) was used to evaluate the participants' functional activity. The participants were asked to indicate a point on an OCD, a 100 mm long vertical line with everyday activities listed alongside the line, above which breathlessness limited them [13]. The distance from zero was measured and scored.

Pulmonary Function Testing
Cigarette smoking, drinking coffee, tea, or alcohol, and taking medications were not permitted 24 h before any test. Bronchodilators were not administered within 3 h for short-acting beta agonists and 12 h for long-acting beta agonists before the tests [14,15]. FEV 1 , TLC, residual volume (RV), and diffusing capacity for carbon monoxide (D L CO) were measured using spirometry, body plethysmography, and the single-breath technique, respectively, in accordance with the currently recommended standards [16][17][18]. All of the spirometry data were obtained before and after inhaling a standard dose of fenoterol HCl. Post-dose measurements were performed 15 min after inhalation. Static lung volume data and D L CO data were obtained before inhaling fenoterol. Simple volume calibration was conducted and accuracy checks for body plethysmograph mouth flow and pressure and box pressure were performed as reported previously [14,15].

Cardiopulmonary Exercise Testing (CPET)
Each subject completed an incremental exercise test using a cycle ergometer to the limit of the symptom. Work rate was selected at a rate of 5-20 W/min based on a derived protocol formula according to the oxygen-cost diagram scores [19]. Oxygen uptake (VO 2 ) (mL/min), CO 2 output (VCO 2 ) (mL/min), and minute ventilation (V E ) were continuously measured. V O2peak was symptom-limited peak V O2 , because V O2max , which was the plateau of V O2 , was likely not attained in the participants with COPD. The ratio of compartment of TLC and TLC was remarked as the % of TLC such as EELV%, DH%, V D %, and V T %. A dyspnea score was obtained using the Borg scale by asking the patients about their dyspnea levels while they were performing the ramp-pattern exercise at the end of each minute and at peak exercise.

Dynamic Inspiratory Capacity (IC) Measurement
The techniques used for performing and accepting IC measurements of our previous study [11] were modified from a previous report [7]. Dynamic IC was measured at the end of a steady-state resting baseline, near the middle of loaded exercise (supposed to be near anaerobic threshold, AT), and near peak exercise. Dynamic IC near AT was measured approximately 5-6 min after the start of loaded exercise. EELV was calculated as TLC minus dynamic IC [5,6,20,21]. DH referred to end-expiratory lung volume at AT or peak exercise (EELV AT or peak ) minus resting EELV (EELV rest ). In this study, EELV was estimated for subjects with COPD using the formulae from the data of our previous report [11]. EELV rest % = 0.7235 − 1.0053 × V Trest %; EELV AT % = 0.9877 − 2.0132 × V T AT %; EELV peak % = 0.9491 − 1.35178 × V Tpeak %; O'Donnell threshold (OT) = TLC -EELV − V Tpeak (see O'Donnell threshold in Reference [22]).

V D /V T Calculation
Brachial artery blood samples were drawn via an arterial catheter connected to a pressure transducer within the last 15 s of each minute after the start of exercise to the peak of exercise [23].
At rest, near the anaerobic threshold, and at the peak of exercise, the physiological V D /V T was calculated using a standard formula as follows [24]: V D /V T = (P a CO 2 − PĒCO 2 )/P a CO 2 − V Dm /(V T − V Dm ), where PĒCO 2 = VCO 2 /V E × (P B − 47 mmHg) and PB is barometric pressure measured daily and V Dm is breathing valve dead space. Hemoglobin and biochemistry data were provided. In normal subjects, mean values of V D /V T are 0.30 ± 0.08 at rest, 0.20 ± 0.07 at AT, and 0.19 ± 0.07 at peak [2].

Statistical Analysis
Data were summarized as mean ± standard deviation. The sample size was estimated to be at least 17 for each group when the population mean difference in V D /V T was 0.1 with a standard deviation for the normal and COPD groups of 0.1 and with a significance level of 0.05 and a power of 0.8. The unpaired t-test was used to compare the means between two groups. The paired t-test was used to compare two related means between two different time points with Bonferroni correction. Pearson's correlation coefficients were further used when appropriate for quantifying the pairwise relationships among the interested variables. All statistical analyses were performed using SAS statistical software 9.4 (SAS Institute Inc., Cary, NC, USA). Statistical significance was set at p < 0.05 and p < 0.017 for Bonferroni correction.

Results
A total of 81 male subjects were enrolled, including 46 subjects (mean age 65.2 ± 5.8 years) with COPD after excluding one subject due to poor motivation, and 34 healthy subjects matched for age and height (mean age 62.2 ± 9.2 years) ( Table 1 and Figure 1). Most of the COPD subjects had GOLD stages II and III with hyperinflation and air trapping, normocapnia, and borderline hypoxemia at rest and could perform daily brisk walking on the level. Compared to the healthy controls during exercise, most of the COPD subjects had mildly impaired exercise capacity due to ventilatory limitation with poor lung expansion, significant oxyhemoglobin desaturation, and exercise hyperventilation ( Table 2).     3.1. The % of TLC: EELV%, DH%, V D %, V T %, V D DH%, V D EELV%, and V T EELV% (or End-Inspiratory Lung Volume, EILV) In the COPD group, EELV rest % was 63% ± 2% and EELV peak was 70% ± 7% as compared with 48% ± 13% and 46% ± 13% in the healthy group ( Figure 2, group comparisons, both p < 0.0001). Hence, DH peak % was 7% ± 7% as compared with 1% ± 10% in the healthy group (p = 0.03). In the COPD group, V Drest % was 5% ± 1% and V Dpeak % was 8% ± 2% as compared with 4% ± 2% and 6% ± 1% in the healthy group (Figure 2, group comparisons: p < 0.01 and p < 0.0001). In the COPD group, DH peak % was similar to V Dpeak % at peak exercise (7% ± 7% vs. 8% ± 2%, p = 0.61).

Relationships among the Compartments of TLC 195
VDpeak% was moderately positively correlated with VTpeak% (  Open circles, end-expiratory lung volume (EELV); solid circles, dead space volume (V D ) plus EELV; down triangles, tidal volume (V T ) plus EELV (i.e., end-inspiratory lung volume, EILV); vertical bars, standard error of estimate; OT, O'Donnell threshold; DH, dynamic hyperinflation indicating EELV at AT or peak exercise minus EELV at rest; dashed line, EELV at rest. Comparisons of each compartment between COPD patients and normal controls at rest, AT and peak exercise, respectively, all p < 0.0001 except V T EELV at rest, p < 0.01 and V T EELV at peak exercise, p < 0.001. In COPD patients, comparisons of each compartments of TLC between two time points, all p < 0.0001 except EELV at AT versus EELV at peak exercise, p < 0.001 and V D EELV at AT versus V D EELV at peak exercise, p = 0.046, which was insignificant.

Relationships among the Compartments of TLC
V Dpeak % was moderately positively correlated with V Tpeak % ( Table 3, r = 0.66, p <0.0001) and moderately negatively correlated with the other compartments at peak exercise (r = −0.47 to −0.68, p <0.01 to <0.0001). Table 3. Relationships among the compartments of total lung capacity (TLC) and correlations of seven components of total lung capacity (TLC) with oxygen uptake (VO 2 ), minute ventilation (V E ), and dyspnea at peak exercise in 46 patients with COPD.

Relationships between the % of TLC and Oxygen Uptake, Minute Ventilation, and Dyspnea
In the % of TLC, V D EELV peak % and V D DH peak % showed the best correlations with ∆Borg/∆VCO 2 and, and a similar strength of correlation with V Epeak ( Table 3). The higher the V D DH peak % and V D EELV peak %, the higher the dyspnea score and the lower the VO 2peak % and V Epeak .

V Tpeak % versus V D /V Tpeak
In the COPD group, V Trest % was 9% ± 2% and V Tpeak % was 18% ± 5% as compared with 13% ± 7% and 32% ± 54% in the healthy group (Figure 2, group comparisons p < 0.01 and p < 0.0001). In the COPD group, there was a negatively significant relationship between V T % and V D /V T at rest, anaerobic threshold, and peak exercise, and this was stronger as the exercise intensity increased (see the Appendix A Table A2, r = −0.34 to −0.64, p = 0.02 to p < 0.0001). When pooling the data of these two variables at the three time points together, the relationship was much closer (r = −0.72, p < 0.0001).

Discussion
There are four main findings in this study. First, V D and DH (V D DH) and V D and EELV (V D EELV) could be combined. Secondly, we found that in the patients with COPD, V D and DH were similar in size, and that V D EELV rest accounted for 68% of the TLC and V D EELV peak accounted for up to 78%. Third, compared to DH peak % and EELV peak %, V D DH peak % and V D EELV peak % were more closely related to dyspnea and exercise capacity and had a similar power in relation to ventilation capability. Lastly, V Tpeak %, a recently reported marker of DH peak [11], was moderately negatively correlated with V D /V Tpeak . To the best of our knowledge, these findings have not previously been published.

The % of TLC
The importance of EELV peak % has been reported when the EELV peak is ≥75% of TLC, a threshold value which can maximize the sensitivity and specificity of detecting ≤5.5 mL/heartbeat change in oxygen pulse (∆O 2P ) and ≤10,000 oxygen uptake efficiency slope (OUES) during exercise [25], where ∆O 2P and OUES are markers of cardiovascular function. In addition to EELV peak % >75% [25], the reciprocal IC peak /TLC <25% [26] has also been associated with lower O 2P and exercise capacity in patients with severe COPD. IC peak /TLC <23% has also been associated with lower O 2P and exercise capacity in patients with severe COPD [27]. Although OUES was not measured in this study, our previous study reported that IC peak /TLC was significantly correlated with O 2P and ∆O 2P (r = 0.35-0.36, both p < 0.05) [28]. These results support an interaction between hyperinflation and decreased cardiac function that can contribute to exercise limitation in these patients. A greater amount of trapped gas in the lung increases the intrinsic positive end-expiratory pressure, and this compresses the heart and impedes venous return causing further heart impairment [25,26]. It has recently been reported that this compression can occur even at rest [29].
DH has been shown to increase with exercise in patients with COPD [3][4][5][6]9,10,[20][21][22], and thus EELV caused failure of V T to expand, as in the healthy subjects in this study (0.6 ± 0.31 L versus 1.12 ± 0.57 L, p < 0.0001). A high level of V D EELV "buoyed" the expandable basic lung volume above its position, meaning that V T had limited room to expand downwards so that it could not help but invade upwards to the OT or near its limit (Figure 2). In COPD, decreased OT [3,22] and increased DH have been reported to be possible causes of exercise limitation [30], although some studies have questioned whether DH occurs in all COPD patients [31][32][33]. These previous studies have measured DH peak but not included V Dpeak . In this study, V D DH peak % and V D EELV peak % were slightly better than DH peak % and EELV peak % with regards to the correlation with ∆Borg/∆VO 2 and VO 2peak % and had a similar power with regards to the correlation with V Epeak (Table 3). Therefore, it could be reasonable to combine V Dpeak with DH peak and to combine V Dpeak with EELV peak . In this study, V D EELV peak %, an unfavorable lung volume, was elevated to as high as 78% ± 6% of TLC.
In the patients with COPD in this study, although V Dpeak % was small as compared with EELV peak % but similar to DH peak % in size, V D DH peak % accounted for 15% of TLC. The majority of the increase in physiological V D must have come from alveolar V D , as the increase in anatomical V D was estimated to be only 12 mL and 20 mL in the COPD and control groups, respectively, based on the estimation that anatomical V D would increase 20 mL per liter increase in EELV [1]. Hence, the remaining increase in physiological V D must have come from alveolar V D , which is strongly influenced by lung pathology but less influenced by other factors such as age, sex, body size (1 mL of physiological dead space per pound of weight reported by Radford), posture, low cardiac output, pulmonary emboli, and posture [1].
V D % and EELV% were moderately negatively correlated (Table 3). This is because V D % and V T % were moderately positively correlated and V T % and EELV% were highly negatively correlated (r = −0.83, p < 0.0001) [11]. V D % was positively correlated with V T % because V D is calculated by V D /V T multiplied by V T . Hence, the larger the V T , the larger the V D , and the smaller the EELV. It is clear that V D is different from EELV and DH in the direction of correlation, that these volumes can be combined, and that the combinations are more related to exercise capacity and exertional dyspnea sensation, although V D is small. Interestingly, V D % alone was poorly related to exercise tolerance and dyspnea. However, the relationships between DH% and EELV% versus exercise tolerance and dyspnea were slightly improved after adding V D % (Table 3).

V T % versus V D /V T
V D /V T has been reported to be the most consistent gas exchange abnormality in smokers with only mild abnormalities in spirometry [3]. However, invasive methods to obtain arterial blood gases are needed to measure V D /V T . In this study, V T %, an inverse marker of DH [11], was inversely correlated with V D /V T (R 2 ≈ 0.50) (see the Appendix A Table A2). However, Mahut et al. reported that V D /V Tpeak was only mildly correlated to DH (r = −0.45, p = 0.004) [10], where DH was represented by IC peak % predicted [10]. This difference in correlation between DH and V D /V T in these two studies could be due to the different criteria used for DH, i.e., IC peak % predicted versus V T %. Predicted IC data were obtained from the general population, whereas V T % was directly measured in the participants. In addition, Mahut et al. reported that the alveolar volume (V A )/TLC ratio was significantly correlated with V D /V Trest but much less significantly correlated with V D /V Tpeak (see the Appendix A Table A1) [10]. V A is usually measured using the single breath helium dilution method at rest and is equal to TLC − V D [34]. Therefore, V A would underestimate TLC in subjects with poorly communicating airways or disequilibrium of ventilation. V A /TLC measured at rest cannot reflect DH peak , so that it was poorly correlated with V D /V Tpeak . Moreover, in this study, the relationship between V T % and V D /V T was strongest when data at rest, anaerobic threshold, and peak exercise were pooled (see the Appendix A Table A2, r = −0.72, p < 0.0001). The mechanism underpinning the stronger relationship between V Tpeak % and V D /V Tpeak with increasing exercise intensity could be due to the common factor V Tpeak being highly constrained at peak exercise. The stronger relationship between V T % and V D /V T after pooling different stages of exercise is comparable to a previous study in which V E /VCO 2 was used instead of V T % in healthy subjects and patients with COPD [3].
Nevertheless, Paoletti et al. reported that V Tpeak /FEV 1 > 1 (or V Tpeak /IC = 0.96 ± 0.05), emphysema, the slope of V E /VCO 2 , and P ET CO 2peak values were colinear [35] (Figure 3). In their study, the patients with COPD had high RV% predicted and high emphysema score measured with high resolution computed tomography (HRCT). They hypothesized that V Tpeak /FEV 1 > 1 or elevated V Tpeak /IC was due to DH occurring at peak exercise in patients with severe emphysema, which is comparable with our study and another study using V Tpeak /SVC to assess the severity of emphysema evaluated with HRCT [36] (Figure 3). However, it has been reported that the change in V D /V T from rest to peak exercise was not related to the severity of emphysema [35]. In the current study, V Tpeak /FEV 1 > 1 and V Tpeak /SVC were correlated with V Tpeak %, respectively (Figure 3, r = −0.36 and 0.66, p = 0.001, p < 0.0001), however neither were correlated with V D /V Tpeak . Nevertheless, V Tpeak % was correlated with V D /V Tpeak (r = −0.64, p < 0.0001), suggesting that V Tpeak % could be more powerful than V Tpeak /FEV 1 and V Tpeak /SVC (Figure 3).  (Figure 3). In their study, the patients with 290 COPD had high RV% predicted and high emphysema score measured with high resolution computed 291 tomography (HRCT). They hypothesized that VTpeak/FEV1 > 1 or elevated VTpeak/IC was due to DH 292 occurring at peak exercise in patients with severe emphysema, which is comparable with our study and 293 another study using VTpeak/SVC to assess the severity of emphysema evaluated with HRCT [36] (Figure  294 3). However, it has been reported that the change in VD/VT from rest to peak exercise was not related to 295 the severity of emphysema [35]. In the current study, VTpeak/FEV1 > 1 and VTpeak/SVC were correlated with 296 VTpeak%, respectively ( bolded boxes, from this study; blue boxes, from references [35] and [36]. Solid lines, significantly correlated; 303 dashed lines, not significantly correlated. Black lines, from this study; blue lines, from reference [35]; green 304 line, from reference [36]. VT%, tidal volume and total lung capacity (TLC) ratio; EELV%, end-expiratory lung 305 volume and TLC ratio; VT/SVC, VT and slow vital capacity ratio; VT/FEV1, VT and forced expired volume in 306 one second ratio; HRCT, high resolution computed tomography; RV%, residual volume predicted %; Δ VE/Δ 307 VCO2, slope of minute ventilation and CO2 output; PETCO2, end-tidal CO2 pressure. 308

Clinical Implications of VDDHpeak% and VDEELVpeak%, and VTpeak% 309
Since DH may not occur in all COPD patients [31][32][33], as VDDHpeak% and VDEELVpeak% are substantially 310 larger and slightly more related to dyspnea [31] and exercise capacity than DH% and EELV%, and as VTpeak% 311 can be obtained easily and noninvasively, these three markers could potentially be used to evaluate the 312 effect of bronchodilator or lung volume reduction surgery on dyspnea and exercise tolerance. 313

Study Limitations 314
Airflow obstruction should be defined as a FEV1/VC ratio below the fifth percentile (z-score −1.645) 315 of the distribution of a reference population [17] according to the 2019 ATS-ERS technical statement 316 [16]. In the present study, the use of GOLD criteria to define COPD could have introduced age, sex, 317 and height selection bias. However, the severity of most of the subjects with COPD in this study 318 had GOLD stages II-IV (93.5%), and thus the likelihood of underdiagnosing COPD was small. 319 Although OCD is not a commonly used tool to evaluate physical activity for patients with COPD, 320 previous studies have suggested that the OCD and the COPD assessment test should be used 321 simultaneously when undertaking clinical evaluations of patients with COPD, and that the OCD in 322 ramp-slope selection should be used for dyspneic patients when undertaking CPET [13,19]. 323 Black bolded boxes, from this study; blue boxes, from References [35,36]. Solid lines, significantly correlated; dashed lines, not significantly correlated. Black lines, from this study; blue lines, from reference [35]; green line, from reference [36]. V T %, tidal volume and total lung capacity (TLC) ratio; EELV%, end-expiratory lung volume and TLC ratio; V T /SVC, V T and slow vital capacity ratio; V T /FEV 1 , V T and forced expired volume in one second ratio; HRCT, high resolution computed tomography; RV%, residual volume predicted %; ∆ V E /∆ VCO 2 , slope of minute ventilation and CO 2 output; P ET CO 2 , end-tidal CO 2 pressure.

Clinical Implications of V D DH peak % and V D EELV peak %, and V Tpeak %
Since DH may not occur in all COPD patients [31][32][33], as V D DH peak % and V D EELV peak % are substantially larger and slightly more related to dyspnea [31] and exercise capacity than DH% and EELV%, and as V Tpeak % can be obtained easily and noninvasively, these three markers could potentially be used to evaluate the effect of bronchodilator or lung volume reduction surgery on dyspnea and exercise tolerance.

Study Limitations
Airflow obstruction should be defined as a FEV 1 /VC ratio below the fifth percentile (z-score −1.645) of the distribution of a reference population [17] according to the 2019 ATS-ERS technical statement [16].
In the present study, the use of GOLD criteria to define COPD could have introduced age, sex, and height selection bias. However, the severity of most of the subjects with COPD in this study had GOLD stages II-IV (93.5%), and thus the likelihood of underdiagnosing COPD was small. Although OCD is not a commonly used tool to evaluate physical activity for patients with COPD, previous studies have suggested that the OCD and the COPD assessment test should be used simultaneously when undertaking clinical evaluations of patients with COPD, and that the OCD in ramp-slope selection should be used for dyspneic patients when undertaking CPET [13,19]. However, the International Physical Activity Questionnaire and accelerometry could also be helpful in this case [37,38]. A novel analytical method reported calculating shunt V D by subtracting respiratory V D (i.e., anatomical V D and alveolar V D ) from physiological V D [39]. We did not calculate shunt V D , as this method is sophisticated and the shunt V D level was expected to be small. Tidal flow limitation measured with negative expiratory pressure has been shown to play a role in reducing the IC at rest, during which tidal flow limitation constrains V T expansion during exercise thereby causing an elevation in V D /V T at peak exercise [40]. Although tidal flow limitation was not measured in this study, it can be anticipated to occur in the subjects with more severe airflow obstruction and higher air trapping with a lower IC [41]. In the COPD group in this study, EELV was estimated using the formulae reported in our previous study [11], and thus the estimated DH% and EELV% values may not be exactly the same as the measured data. In the healthy controls, data on V D /V T at rest, AT, and peak exercise were retrieved from reference subjects, as it was difficult to obtain permission from our Institutional Review Boards to perform arterial catheterization for exercise testing. The emphysematous phenotype could be related to V D DH. However, as there were relatively few subjects and emphysema was not evaluated using HRCT in this study, further studies are warranted to address these issues. Lastly, V D cannot be obtained without using invasive method in patients with COPD, and thus its clinical implication could be limited. Studies to investigate the development of a novel noninvasive method to obtain V D or V D /V T are warranted. Finally, using Jones' and Bohr's equations to estimate V D /V T in subjects with COPD is not suitable, as P ET CO 2 used in the equations cannot be used as a surrogate for P a CO 2 or alveolar PCO 2 [42,43].

Conclusions
Although the definitions of V D and DH are quite different, this study shows the utility of their combination, and that it could play a role in physiology with regards to the evaluation of exertional dyspnea and exercise capacity in subjects with COPD. In addition, V T % was significantly correlated with V D /V T , suggesting that V T % is not only a convenient marker for DH as reported previously, but also a potential noninvasive marker for V D /V T .
Author Contributions: M.-L.C. initiated and designed the study, analyzed and interpreted the data, wrote the manuscript. All authors have read and agreed to the published version of the manuscript Funding: The study was supported in part by the Minister of Science and Technology, Taiwan (MOST 106-2314-B-040-025). The funding body had no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.

Conflicts of Interest:
The author declares no competing financial interests.  Table A1. Summary of the correlation coefficient (r) between the dead space fraction (V D /V T ) and some physiological variables reported by Mahut et al. [10] and Elbehairy et al. [3].

Rest Peak
V A /TLC [10] −0.6 −0.2 V E peak /MVC% [10] NA 0.32 IC peak % predicted [10] NA −0.45 V E /VCO 2 [3] 0.78 ** NA KCO [10] −0.52 −0.43 D L CO% predicted [10] NA * NA* PaO 2 peak [10] NA −0.66 Borg peak /%VO 2peak [10] NA 0.33 V A , alveolar volume measured during diffusing capacity for carbon monoxide (DLCO) measurement; TLC, total lung capacity; IC, inspiratory capacity; V E , minute ventilation; CO 2 , CO 2 output; KCO, the diffusing constant of Krogh, i.e., D L CO/V A without considering barometric pressure, where V A is alveolar volume in BTPS equal to TLC measured by single breath helium dilution method after subtracting anatomic dead space [34]; Borg, Borg score. * p < 0.05 reported in reference [10], but r values are not reported, ** data involving rest and submaximal exercise in healthy subjects and mild COPD subjects. NA: not available. Table A2. Pearson correlations (r) pairwise deletion between dead space and tidal volume ratio (V D /V T ) and tidal volume and total lung capacity ratio (V T %) at different phases of exercise test in participants with chronic obstructive pulmonary disease. AT: anaerobic threshold, * p < 0.05, ** p < 0.01, † p < 0.0001, All: V T % at rest, AT, and peak and V D /V T at rest, AT, and peak were pooled together.