Gas exchange efficiency slopes to assess exercise tolerance in chronic obstructive pulmonary disease

Abstract

Background

In patients with chronic obstructive pulmonary disease (COPD), the clinical use of the minute ventilation-carbon dioxide production (_E-_CO2) slope has been reported as a measure of exercise efficiency, but the oxygen uptake efficiency slope (OUES), i.e., the slope of oxygen uptake (_O2) versus the logarithmically transformed _E, has rarely been reported.

Methods

We hypothesized that the _E-_CO2 slope is more useful than OUES in clinical use for the pathophysiological evaluation of COPD. Then, we investigated the cardiopulmonary exercise testing parameters affecting each of these slopes in 122 patients with all Global Initiative for Chronic Obstructive Lung Disease (GOLD) COPD grades selected from our database.

Results

Compared with the GOLD I-II group (n = 51), peak _O2 (p < 0.0001), OUES (p = 0.0161), _E at peak exercise (p < 0.0001), and percutaneous oxygen saturation (SpO₂) at peak exercise (p = 0.0004) were significantly lower in the GOLD III-IV group (n = 71). The GOLD III-IV group was divided into two groups by the exertional decrease in SpO₂ from rest to peak exercise: 3% or less (the non-desaturation group: n = 23), or greater than 3% (the desaturation group: n = 48). OUES correlated only weakly with peak _O2, _E at peak exercise, and the difference between inspired and expired mean O₂ concentrations (ΔF_O2) at peak exercise, i.e., an indicator of oxygen consumption ability throughout the body, in the GOLD III-IV group with exertional hypoxemia. In contrast, the _E-_CO2 slope was significantly correlated with ΔF_O2 at peak exercise, regardless of the COPD grade and exertional desaturation. Across all COPD stages, there was no correlation between the _E-_CO2 slope and _E at peak exercise, and stepwise analysis identified peak _O2 (p = 0.0345) and ΔF_O2 (p < 0.0001) as variables with a greater effect on the _E-_CO2 slope.

Conclusions

The OUES may be less useful in advanced COPD with exertional hypoxemia. The _E-_CO2 slope, which is independent of _E, focuses on oxygen consumption ability and exercise tolerance in COPD, regardless of the exertional hypoxemia level and COPD grade. Therefore, the _E-_CO2 slope might be useful in establishing or evaluating tailor-made therapies for individual patient’s pathologies in COPD as an indicator focusing on oxygen consumption ability.

Supplementary Information

The online version contains supplementary material available at10.1186/s12890-024-03312-2. 

Background

With the aging of the population, there is an urgent need to improve and maintain exercise tolerance irrespective of disease severity in chronic obstructive pulmonary disease (COPD) patients [1–3]. In cardiopulmonary exercise testing (CPET) using expiratory gas analysis, all parameters are calculated from directly measured ventilation flow and carbon dioxide (CO₂) and oxygen (O₂) concentrations [4]. As shown in the online supplementary Methods, oxygen uptake (_O2), is calculated using the following equation:

where F_IO2 and F_EO2 are inspired and expired mean O₂ concentrations, respectively, _E is minute ventilation, and _CO2 is carbon dioxide output. Based on the above equation and defining ΔF_O2 = F_IO2 − F_EO2, _O2 is considered dependent on _E and ΔF_O2. ΔF_O2 is considered to reflect the ability of oxygen consumption processed by the entire body, including the muscles [4, 5], as shown in online supplementary Figure S1. Given that the primary indicator of exercise tolerance in CPET is peak _O2[4, 6], information on the individual contributions of _E or ΔF_O2 to exercise tolerance, might be important when formulating treatment plans by considering which of the two factors has greater potential for improving exercise intolerance. However, this assumption has not been adequately evaluated in COPD.

We previously reported that ΔF_O2, which is directly measured in incremental exercise at peak exercise, has a strong negative correlation with ventilatory efficiency, indicated (1) by the _E-_CO2 slope, which indicates the degree of CO₂ clearance during exercise [7, 8], and (2) by the _E/_CO2 nadir, which is defined as the lowest value of the ratio between _E and _CO2 during exercise, but is not well correlated with _E/_{CO2 intercept}, that is, the Y intercept from the _E versus _CO2 regression line [9]. Furthermore, the only indicators that were independent of _E at peak exercise were ΔF_O2 and the _E-_CO2 slope [9]. Therefore, of the three ventilatory efficiency indicators, this study focused on the _E-_CO2 slope. We also previously reported that, after pulmonary rehabilitation in patients with advanced COPD, the increased ΔF_O2 at peak exercise in incremental exercise was correlated with the improvement in not only ventilatory efficiency, but also peak _O2. However, the increase in _E does not affect the improvement in the _E-_CO2 slope [10]. Furthermore, no significant correlation was confirmed between the change in _E and that in ΔF_O2 at peak exercise following pulmonary rehabilitation. Interestingly, the calculated ratio of oxygen uptake to ventilation (i.e., _O2/_E), which is almost equivalent, but not equal, to actually measured ΔF_O2, is known as Oxygen Uptake Efficiency, and its highest plateau (OUEP) during incremental exercise has been reported as a useful factor for assessing the severity of dysfunction and as a prognostic factor for mortality and morbidity in patients with chronic heart failure, though it has not been evaluated in COPD [11]. More in depth knowledge of specialized indicators that reflect the ability of oxygen consumption to increase ΔF_O2 during exercise separately from ventilatory ability to increase _E would be useful for establishing therapeutic strategies and assessing exercise therapy targeting peripheral muscles not only in COPD patients, but in all patients for whom exercise therapy is indicated.

On the other hand, the Oxygen Uptake Efficiency Slope (OUES) in the regression equation _O2 = OUES × log_10E + intercept is considered an objective estimate of cardiopulmonary functional reserve even at submaximal exercise. OUES has been proposed as a surrogate for maximal oxygen uptake (_O2 max.) based on the concept of OUES [12], because OUES has the ability to interrogate responses at submaximal exercise, which can be beneficial in patients whose peak _O2 is low (i.e., due to premature exercise termination because of symptoms or failure of _O2 to increase due to oxygen extraction difficulties). Although the physiological meaning of OUES has been highlighted in healthy subjects [13], children [12], and adult patients with cardiac disease [14], it has not been evaluated in COPD. Traditionally, OUES and OUEP have been used for outcome prediction in cardiovascular patients in whom respiratory limitation is not an issue. Given the advantage of the logarithm step, OUES emphasizes the physical condition in _E in the early exercise phase. Therefore it might be a different exercise indicator from the later exercise phase evaluation of COPD such as the _E-_CO2 slope [7, 8] and peak _O2, because the severity of ventilatory disorders due to airflow limitation is greater at the end of exercise. In other words, the cardiopulmonary functional reserve suggested by OUES in advanced COPD with severe airflow limitation might be an overestimation that cannot actually be reached.

We hypothesized that, in COPD patients with exertional dyspnea, the _E-_CO2 slope is more useful than OUES in understanding their exercise pathophysiology and formulating a specific treatment plan for each patient based on that pathophysiology. In this study, we evaluated patients with all Global Initiative for Chronic Obstructive Lung Disease (GOLD) COPD grades, to examine the clinical implications of the _E-_CO2 slope and OUES, and to determine which of the pathophysiological indicators, especially ventilatory variables and ΔF_O2 during exercise are affected by them.

Methods

Subjects

This was a retrospective study using existing clinical records. Of a cumulative total of 1212 patients who required evaluation of their exertional dyspnea during routine medical care and underwent CPET to understand their exercise pathophysiology or before treatment such as pulmonary rehabilitation at the NHO Osaka Toneyama Medical Centre from January 2015 to August 2022, 122 patients diagnosed with COPD ranging from mild to very severe according to the GOLD criteria, were selected and their data were retrospectively analyzed. Patients were excluded if: (1) they were diagnosed with bronchial asthma, active infection, or severe heart disease; (2) had a history of lung resection; (3) there were changes in their drug regimen within 4 weeks before CPET; or (4) they were unstable due to other reasons. This study was performed in accordance with the Declaration of Helsinki and was approved by the institutional review board of the NHO Osaka Toneyama Medical Center (approval number: TNH-A-2022018). Each patient provided, written, informed consent before CPET evaluation during routine medical care, with the understanding that CPET was performed because it was considered necessary in daily clinical practice, and that the data obtained may be used in clinical research or other settings.

Pulmonary function tests (PFTs)

Post-bronchodilator spirometry was performed (CHESTAC 8800; CHEST M.I. Inc., Tokyo, Japan) according to the recommendations of the American Thoracic Society [15]. PFTs were performed within 2 weeks before or after CPET.

Cardiopulmonary exercise testing (CPET)

A CPET system (Aero monitor AE310S; Minato Medical Science Co., Ltd, Osaka, Japan) was used for symptom-limited incremental exercise tests performed in 1-minute or 2-minute stages with a 10-watt step protocol using an electrically braked cycle ergometer (CV-1000SS; Lode, Groningen, The Netherlands) [10, 16]. Calibrations for ventilatory measurement and gas measurements (CO₂ and O₂) were performed before each CPET. Pre-exercise resting measurements were obtained during the steady-state period after at least 3 min of rest after preparation for CPET. Ventilatory variables were measured on a breath-by-breath basis using a face mask, and they are presented as 30-s averages at rest, at 2-min intervals during exercise, and at the end of exercise. Dyspnea levels were assessed at rest, during the last 15 s of each exercise stage, and at the end of exercise using a 10-point modified Borg category-ratio scale. ΔF_O2 was calculated from the difference between F_IO2 and F_EO2 measured during exercise. As measurements, the maximum value of ΔF_O2 during exercise (ΔF_O2 max) and ΔF_O2 at peak exercise were examined. In addition, from January 2015 to August 2022, CPET was performed using the same expiratory gas analyzer based on the above protocol without any changes. The _E-_CO2 slope and _E/_CO2 nadir were calculated and defined as previously described [9]. Based on the assumption that log-transforming _E will make the relationship between _E and _O2 more linear than using _E, as shown in Fig. 1 and online supplemental Figure S2, OUES was calculated as the slope of the regression line relating _O2 (mL ⋅ min^− 1) to the common logarithm of _E (L ⋅ min^− 1) by the equation: _O2 (mL ⋅ min^− 1) = slope × log _{10 E} (L ⋅ min^− 1) + intercept, where the constant ‘slope’ is OUES, using breath-by-breath gas-exchange measurements [12]. Basically, to avoid possible irregular breathing patterns, data from the 3-min resting period, from the first minute of exercise, and from a plateau in _O2 were excluded from the calculation of OUES. The anaerobic threshold (AT) was determined using the V-slope method [4]. Predicted maximal voluntary ventilation (MVV) was calculated as FEV₁ × 35 [4]. Predicted maximum heart rate was calculated as 220–age in years [4].

Fig. 1.

Calculation of OUES. a) The relationship between oxygen uptake (_O2: mL ⋅ min⁻¹) and the common logarithm of minute ventilation (_E: L ⋅ min⁻¹) (b) the relationship between _O2 (mL ⋅ min⁻¹) and _E (L ⋅ min⁻¹). _O2 and _E obtained from breath-by-breath gas-exchange measurements in one COPD patient with GOLD I (75 years old, 64.2 kg) were used. OUES was calculated as the slope of the regression line relating _O2 (mL ⋅ min^− 1) to the common logarithm of _E (L ⋅ min^− 1) by the equation: _O2 = slope × log _10E + intercept. where the constant ‘slope’ is the OUES. Basically, to avoid possible irregular breathing patterns, data from the 3-min resting period, from the first minute of exercise, and from a plateau in _O2 were excluded from the calculation of the OUES. Green cross: exercise starting point; blue crosses: period during which OUES was calculated

Statistical analysis

Variables are presented as means (standard deviation) or medians (interquartile range) depending on the normality of the distribution (Shapiro–Wilk test) unless otherwise stated. Depending on the normality of the distribution, for continuous variables, (1) an unpaired t-test or the Wilcoxon rank-sum test was used for comparison between patients with GOLD grades I-II and III-IV, and (2) univariate analysis using Pearson’s correlation or Spearman’s rank correlation coefficient was used to evaluate correlations between clinical variables. R squared was used to confirm how close the data were to the fitted regression line. The chi-squared test was used for categorical variables. Bidirectional stepwise variable selection was performed to determine variables more correlated with the _E-_CO2 slope of the exertional variables related to oxygen uptake including COPD grade, which was divided into GOLD grades I-II and III-IV, and the desaturation level at peak exercise, with variance inflation factors (VIFs) less than 3. In the stepwise regression models, the p values were set at 0.20 for variables to both enter and stay. A p value of less than 0.05 was considered significant. Statistical analyses were performed using JMP software, version 11 (SAS Institute Inc., Cary, NC, USA).

Results

The study sample consisted of 122 patients distributed across all GOLD stages (Table 1). There were no significant differences in age, sex, smoking history, and the choice of inhalation therapy between the GOLD I-II and III-IV groups, although BMI (p = 0.0322) was lower, and especially, the frequency of triple inhalation therapy (p = 0.0008) was higher in the GOLD III- IV group than in the GOLD I-II group (Table 1).

Table 1.

Baseline characteristics of COPD patients classified according to Global Initiative for Chronic Obstructive Lung Disease (GOLD) criteria

	All COPD patients (n = 122)	GOLD spirometric severity
		I + II (n = 51)	III + IV (n = 71)	p-value
Age, y	74 (6) [75 (70; 78)]	75 (5)	73 (6) [74 (70; 78)]	0.2253
Sex, male/female (n)	112/10	47/4	65/6	0.9040
BMI, kg ⋅ m− 2	21.5 (3.3)	22.2 (3.1)	21.0 (3.4)	0.0322
Pulmonary function test
FEV1, L	1.27 (0.58) [1.09 (0.86; 1.58)]	1.81 (0.52) [1.71 (1.49; 2.14)]	0.88 (0.19)	< 0.0001
%FEV1, % predicted	50.8 (22.4) [45.6 (35.2; 63.7)]	72.0 (18.0) [67.9 (57.6; 82.6)]	35.6 (8.4)	< 0.0001
FEV1/FVC, %	42.2 (12.2) [40.6 (31.9;49.3)]	51.8 (10.0) [50.0 (43.8; 60.3)]	35.3 (8.2) [32.9 (29.9; 40.0)]	< 0.0001
VC, L	3.05 (0.77)	3.50 (0.68)	2.72 (0.65)	< 0.0001
%VC, %	99.0 (23.1)	113.8 (122.1)	88.5 (20.6)	< 0.0001
IC, L	1.96 (0.52)	2.27 (0.48)	1.73 (0.43)	< 0.0001
Medications (n)
LAMA	102	37	65	0.0052
LABA	92	31	61	0.0015
ICS	64	21	43	0.0344
Triple inhalation therapy	45	10	35	0.0008

Data are presented as means (standard deviation) or [medians (interquartile range: 25th percentile to 75th percentile)], unless otherwise stated. BMI: body mass index; FEV₁: forced expiratory volume in one second; FVC: forced vital capacity; GOLD: Global Initiative for Chronic Obstructive Lung Disease; IC: inspiratory capacity; ICS: inhaled corticosteroid; LABA: long-acting β₂-agonist; LAMA: long-acting muscarinic antagonist; VC: vital capacity. Medications are presented separately. Pulmonary function tests were performed after bronchodilator administration

Peak exercise variables

Incremental exercise parameters at peak exercise in COPD patients in the GOLD I-II and III-IV groups are shown in Table 2. Compared with the GOLD I-II group, peak_O2 (p < 0.0001), OUES (p = 0.0161), O₂ pulse (p < 0.0001), percutaneous oxygen saturation (SpO₂) (p = 0.0004) at peak exercise, and _E (p < 0.0001) at peak exercise were significantly lower in the GOLD III-IV group with reduced breathing reserve, as expected by 1-_E/maximal voluntary ventilation (MVV). Reduced _E in the GOLD III-IV group was caused by expiratory airflow limitation, indicated by lower values of tidal volume (V_T) (p < 0.0001), respiratory frequency (f_R) (p = 0.0079), and inspiratory duty cycle (T_I/Ttot, where Ti is the inspiratory time of a single breath in seconds and Ttot is the total time of a single respiratory cycle in seconds) (p < 0.0001). There were no significant differences in ΔF_O2 max during exercise, ΔF_O2 at peak exercise, and the _E-_CO2 slope between the GOLD I-II and III-IV groups. However, the dependence of ΔF_O2 at peak exercise on peak _O2, expressed as the square of the correlation coefficient, was highest in the GOLD IV group (GOLD I (n = 18): r² = 0.21; GOLD II (n = 33): r² = 0.29; GOLD III (n = 51): r² = 0.12; GOLD IV (n = 20): r² = 0.68). In contrast, OUES (p = 0.0161) was lower in the GOLD III-IV group than in the GOLD I-II group.

Table 2.

Incremental exercise variables in COPD patients classified according to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) criteria

	GOLD spirometric severity
	I + II (n = 51)	III + IV (n = 71)	p-value
At peak exercise
Dyspnea, Borg scale	5.6 (1.8) [5.0 (4.0; 7.0)]	6.0 (1.9) [6.0 (5.0; 8.0)]	0.2126
O2, mL⋅min−1 ⋅ kg−1	14.9 (3.8) [15.2 (11.7; 17.3)]	10.6 (2.3)	< 0.0001
R	1.10 (0.10)	1.01 (0.08) [1.01 (0.96; 1.05)]	< 0.0001
E, L ⋅ min−1	43.7 (12.6)	28.0 (6.6)	< 0.0001
VT, mL	1397 (320)	987 (239)	< 0.0001
fR, breaths ⋅ min− 1	31.7 (6.1)	28.9 (5.3)	0.0079
TI/Ttot	0.41 (0.05) [0.41 (0.38; 0.43)]	0.34 (0.05)	< 0.0001
E/O2	49.7 (10.1)	48.0 (7.4) [47.0 (44.0; 51.2)]	0.2892
E/CO2	45.1 (8.7)	47.8 (8.4) [46.8 (42.7; 51.0)]	0.1623
1−E/MVV, %	29.2 (15.4)	8.4 (19.0) [11.9 (2.9; 19.4)]	< 0.0001
HR, beats ⋅ min− 1	124 (17.6)	112 (16.8)	0.0005
HR/predicted maximum HR, %	80.8 (12.4)	80.0 (10.7)	0.7007
SpO2, %	93 (4) [95 (91; 97)]	91 (5) [92 (87; 94)]	0.0004
O2 pulse, mL ⋅ beats− 1	7.3 (1.9)	5.4 (1.6) [5.1 (4.2; 6.4)]	< 0.0001
OUES	1319 (420)	1135 (440) [1115 (880; 1328)]	0.0161
ΔFO2, %	2.58 (0.50) [2.52 (2.28; 2.74)]	2.59 (0.38)	0.5014
During exercise
ΔFO2 max., %	2.88 (0.57) [2.78 (2.47; 3.11)]	2.72 (0.41)	0.2062
E-CO2 slope	39.1 (8.9)	40.3 (10.9) [39.2 (34.2; 43.2)]	0.8073
E/CO2 nadir	43.1 (7.9)	46.8 (7.6) [46.0 (41.0; 50.0)]	0.0294
Number of patients reaching AT, n (%)	49 (96)	51 (72)	0.0006
O2 at AT, mL ⋅ min−1	601 (178) [575 (490; 670)]	481(104)	0.0001
Number of patients reaching RCP, n (%)	19 (37)	7 (10)	0.0003

Data are presented as means (standard deviation) or [medians (interquartile range: 25th percentile to 75th percentile)], unless otherwise stated. AT: anaerobic threshold; ΔF_O2: difference between inspired mean oxygen concentration and expired mean oxygen concentration; ΔFO₂ max.: the highest value during exercise; f_R: breathing frequency; HR: heart rate; O₂ pulse: _O2/HR; predicted maximum HR: 220−age (y); OUES: oxygen uptake efficiency slope (see the Methods for details); R: gas exchange ratio; RCP: respiratory compensation point; T_I/Ttot: inspiratory duty cycle; _CO2: carbon dioxide output; _E: minute ventilation; and _E/_CO2-slope: the slope was determined by linear regression analysis of _E to _CO2 obtained during exercise (see the Methods for details); _E/_CO2-nadir: the lowest value during exercise (see the Methods for details); _O2: oxygen uptake; V_T: tidal volume. Estimated maximal voluntary ventilation (MVV) (L·min^− 1) was equal to forced expiratory volume in one second (FEV₁)×35

Correlations of OUES with exercise variables

In all GOLD grades, and in the GOLD I-II and GOLD III-IV groups, OUES was not strongly correlated with peak _O2, _E, ventilatory efficiency, ΔF_O2 at peak exercise, and resting pulmonary function; the square of the correlation coefficient was not greater than 0.5 in all groups. (Table 3; Fig. 2a-d). The GOLD III-IV group was divided into two groups depending on the exertional decrease from rest to peak exercise in SpO₂ into the non-desaturation group (3% or less, n = 23) and the desaturation group (greater than 3%, n = 48), as shown in Table 4. The correlation coefficients of OUES with exertional variables, such as peak _O2, _E, oxygen pulse, and ΔF_O2 at peak exercise, were higher in the non-desaturation group than in the desaturation group (Fig. 3a-d). In the desaturation group of the GOLD III-IV group, _E at peak exercise (p = 0.0124) was significantly lower, and O₂-pulse at peak exercise (p = 0.0818) tended to be lower than in the non-desaturation group, but peak _O2 (p = 0.4505) and ΔF_O2 at peak exercise (p = 0.2839) were not (Table 4; Fig. 4).

Table 3.

Square of the correlation coefficient of parameters related to OUES in COPD patients (n = 122)

	All COPD patients	GOLD I-II (n = 51)	GOLD III-IV (n = 71)
Peak incremental exercise parameters
Dyspnoea, Borg scale	0.02	0.00	0.03
O2, mL ⋅ min−1 ⋅ kg	0.35	0.60	0.19
E, L ⋅ min−1	0.27	0.25	0.26
VT, mL	0.27	0.30	0.19
fR, breaths ⋅ min− 1	0.01	0.01	0.00
TI/Ttot	0.03	0.01	0.00
E/CO2	0.26	0.31	0.14
HR, beats ⋅ min− 1	0.05	0.09	0.01
O2 pulse, mL ⋅ beats− 1	0.47	0.66	0.35
ΔFO2, %	0.21	0.33	0.17
During exercise
ΔFO2 max., %	0.18	0.34	0.07
E−CO2 slope	0.25	0.29	0.20
E/CO2 nadir	0.22	0.28	0.08
AT (O2), mL ⋅ min−1	0.39	0.46	0.33
Pulmonary function and others
FEV1, L	0.16	0.19	0.12
VC, L	0.19	0.08	0.07
FVC, L	0.18	0.07	0.14
IC, L	0.17	0.16	0.10
Age, y	0.00	0.00	0.02
BMI, kg ⋅ m− 2	0.14	0.13	0.15

AT: anaerobic threshold; BMI: body mass index; FEV₁: forced expiratory volume in one second; FVC: forced vital capacity; ΔF_O2: difference between inspired mean oxygen concentration and expired mean oxygen concentration; ΔFO₂ max.: the highest ΔF_O2 value during exercise; f_R: breathing frequency; HR: heart rate; IC: inspiratory capacity; O₂ pulse:  _O2/HR; OUES: oxygen uptake efficiency slope (see the Methods for details); T_I/Ttot: inspiratory duty cycle; VC: vital capacity; _CO2: carbon dioxide output; _E: minute ventilation; _E/_CO2-nadir: the lowest value during exercise (see the Methods for details); and _E−_CO2-slope: the slope was determined by linear regression analysis of _E to _CO2 obtained during exercise (see the Methods for details); _O2: oxygen uptake; V_T: tidal volume

Fig. 2.

In all COPD stages, correlations of the oxygen uptake efficiency slope (OUES) (upper panel) and the _E/_CO2 slope (bottom panel) with exertional variables during cardiopulmonary exercise testing. Correlations of OUES with peak _O2 (a), _E at peak exercise (b), O₂ pulse at peak exercise (c), and ΔF_O2 at peak exercise (d) are shown. Correlations of the _E/_CO2 slope with peak _O2 (e), ΔF_O2 at peak exercise (f), and ΔF_O2 max. during exercise (g) are shown. ΔF_O2: difference between inspired and expired mean oxygen concentrations; ΔF_O2 max.: highest ΔF_O2 value during exercise; O₂ pulse: _O2/heart rate; _E: minute ventilation; _O2: oxygen uptake. Closed circle: Global Initiative for Chronic Obstructive Lung Disease (GOLD) I-II group (n = 51). Open circle: GOLD III-IV group (n = 71). The shaded area indicates the CI. The correlation coefficient (r) was obtained in the analysis of all COPD stages

Table 4.

Incremental exercise variables in GOLD III-IV COPD patients divided into two groups by the desaturation level at peak exercise

	Non-desaturation group: 3% or less (n = 23)	Desaturation group: greater than 3% (n = 48)	p-value
At peak exercise
Dyspnea, Borg scale	5.7 (1.7)	6.2 (1.9) [6.5 (5.0; 8.0)]	0.2520
O2, mL⋅min−1 ⋅ kg−1	11.0 (3.0)	10.5 (1.9)	0.4505
R	1.02 (0.10) [1.02 (0.95; 1.07)]	1.00 (0.07)	0.7119
E, L ⋅ min−1	30.7 (7.2)	26.6 (5.9)	0.0124
VT, mL	1029 (240)	967 (238) [917 (794; 1082)]	0.2796
fR, breaths ⋅ min− 1	30.3 (5.2)	28.2 (5.2)	0.1058
TI/Ttot	0.35 (0.06)	0.34 (0.04)	0.2806
E/O2	49.5 (8.1)	47.3 (7.0)	0.2310
E/CO2	48.9 (9.8) [48.0 (43.0; 52.0)]	47.3 (7.6) [46.0 (41.4; 51.0)]	0.6055
1−E/MVV, %	2.2 (15.9) [10.2 (−8.1; 17.8)]	11.3 (13.9) [12.7 (3.3; 20.3)]	0.2169
HR, beats ⋅ min− 1	111 (17)	113 (17)	0.5799
HR/predicted maximum HR, %	78.8 (10.4)	80.6 (11.0)	0.5280
SpO2, %	95 (2)	88 (5)	ND
O2 pulse, mL ⋅ beats− 1	6.0 (2.1)	5.1 (1.3)	0.0818
OUES	1086 (437)	1158 (444) [1129 (880; 1345)]	0.4317
ΔFO2, %	2.52 (0.39)	2.63 (0.37)	0.2839
During exercise
ΔFO2 max., %	2.68 (0.43)	2.73 (0.41)	0.6083
E-CO2 slope	41.7 (11.4) [39.1 (34.6; 44.8)]	39.6 (10.7) [39.3 (34.1; 42.8)]	0.5972
E/CO2 nadir	47.4 (8.1)	46.5 (7.3) [45.5 (41.0; 49.0)]	0.6387
Number of patients reaching AT, n (%)	16 (70)	35 (73)	0.7689
O2 at AT, mL ⋅ min−1	515 (144)	464(78)	0.1109

Data are presented as means (standard deviation) or [medians (interquartile range: 25th percentile to 75th percentile)], unless otherwise stated. AT: anaerobic threshold; ΔF_O2: difference between inspired mean oxygen concentration and expired mean oxygen concentration; ΔF_O2 max.: the highest value during exercise; f_R: breathing frequency; HR: heart rate; O₂ pulse: _O2/HR; predicted maximum HR: 220−age (y); ND: not done; OUES: oxygen uptake efficiency slope (see the Methods for details); R: gas exchange ratio; RCP: respiratory compensation point; T_I/Ttot: inspiratory duty cycle; _CO2: carbon dioxide output; _E: minute ventilation; and _E/_CO2-slope: the slope was determined by linear regression analysis of _E to _O2 obtained during exercise (see the Methods for details); _E/_CO2-nadir: the lowest value during exercise (see the Methods for details); _O2: oxygen uptake; V_T: tidal volume. Estimated maximal voluntary ventilation (MVV) (L·min^− 1) was equal to forced expiratory volume in one second (FEV₁)×35

Fig. 3.

In COPD Global Initiative for Chronic Obstructive Lung Disease (GOLD) III and IV stages, the correlations of oxygen uptake efficiency slope (OUES) (upper panel) and the _E/_CO2 slope (bottom panel) with exertional variables during cardiopulmonary exercise testing classified by oxygen (O₂) desaturation level. Correlations of OUES with peak _O2 (a), _E at peak exercise (b), O₂ pulse at peak exercise (c), and ΔF_O2 at peak exercise (d) are shown. Correlations of the _E/_CO2 slope with peak _O2 (e), ΔF_O2 at peak exercise (f), and ΔF_O2 max. during exercise (g) are shown. ΔF_O2: difference between inspired and expired mean oxygen concentrations; ΔF_O2 max.: highest ΔF_O2 value during exercise; O₂ pulse: _O2/heart rate; _E: minute ventilation; _O2: oxygen uptake. Closed triangle and solid line: group with O₂ desaturation of 3% or less during exercise (n = 23). Open triangle and dotted line: group with O₂ desaturation greater than 3% during exercise (n = 48). The shaded area indicates the CI. The correlation coefficient (r) was obtained in the analysis for each group

Fig. 4.

Relationship between oxygen uptake (_O2: mL ⋅ min⁻¹) and minute ventilation (_E: L ⋅ min⁻¹) in (a) GOLD III-IV without exertional desaturation, 3% or less decrease (n = 23) and (b) GOLD III-IV with exertional desaturation, greater than 3% decrease (n = 48). Plots from left to right at rest, at 2 min, and at peak exercise. To see whether it is inappropriate to represent the relationship of _O2 to _E as a line, the dotted line is a straight line connecting the resting and 2-minute plots. GOLD: Global Initiative for Chronic Obstructive Lung Disease. Mean ± standard error. *: p < 0.05 between the non-desaturation and desaturation groups in _E (x-component)

Correlation of the _E-_CO2 slope with exercise variables

In all GOLD grades, and in the GOLD I-II and GOLD III-IV groups, there was no correlation between _E-_CO2 slope and _E at peak exercise ( Table 5). There were, however, correlations between the _E-_CO2 slope and oxygen uptake-related variables (peak _O2, _O2 at the AT, OUES, ΔF_O2 at peak exercise, and maximum ΔF_O2 during exercise) (Table 3 5; Fig. 2e-g). In the GOLD III-IV group, the correlation coefficients of the _E-_CO2 slope with exertional variables, such as peak _O2, ΔF_O2 at peak exercise, and maximal ΔF_O2 during exercise were almost the same between the groups with and without exertional desaturation (Fig. 3e-g). Of the four exercise variables related to oxygen uptake (peak _O2, _O2 at the AT, OUES, ΔF_O2 at peak exercise), and including COPD grade and SpO₂ at peak exercise, investigations of more influential variables that correlated with the _E-_CO2 slope across all GOLD grades were performed using stepwise variable selection. The analysis identified peak _O2 (F value = 4.6, p=0.0345) and ΔF_O2 at peak exercise (F value = 148.4, p < 0.0001) as more influential variables correlated with the _E-_CO2 slope.

Table 5.

Square of the correlation coefficient of parameters related to _E-co₂ slope in COPD patients (n = 122)

	All COPD patients	GOLD I-II (n = 51)	GOLD III-IV (n = 71)
Peak incremental exercise parameters
Dyspnoea, Borg scale	0.00	0.03	0.00
O2, mL ⋅ min−1 ⋅ kg	0.15	0.34	0.14
E, L ⋅ min−1	0.00	0.00	0.00
VT, mL	0.04	0.07	0.06
fR, breaths ⋅ min− 1	0.05	0.04	0.06
TI/Ttot	0.00	0.02	0.00
E/O2	0.57	0.55	0.60
HR, beats ⋅ min− 1	0.13	0.23	0.11
O2 pulse, mL ⋅ beats− 1	0.08	0.13	0.06
ΔFO2, %	0.61	0.57	0.63
During exercise
ΔFO2 max., %	0.55	0.69	0.45
E/CO2 nadir	0.61	0.81	0.54
AT (O2), mL ⋅ min−1	0.07	0.14	0.09
Pulmonary function and others
FEV1, L	0.00	0.04	0.00
IC, L	0.00	0.01	0.00
Age, y	0.00	0.00	0.01
BMI, kg ⋅ m− 2	0.02	0.00	0.04

AT: anaerobic threshold; BMI: body mass index; FEV₁: forced expiratory volume in one second; ΔFO₂: difference between inspired mean oxygen concentration and expired mean oxygen concentration; ΔF_O2 max.: the highest ΔF_O2 value during exercise; f_R: breathing frequency; HR: heart rate; IC: inspiratory capacity; O₂ pulse: _O2/HR; T_I/Ttot: inspiratory duty cycle; _CO2: carbon dioxide output; _E: minute ventilation; _E/_CO2-nadir: the lowest value during exercise (see the Methods for details); _E-_CO2-slope: the slope was determined by linear regression analysis of _E to _CO2 obtained during exercise (see the Methods for details); _O2: oxygen uptake; V_T: tidal volume

Discussion

We investigated the clinical implications of OUES and the _E-_CO2 slope as indicators of pathophysiological conditions during incremental exercise in patients with all COPD grades. The main findings of this study were as follows. First, OUES was not strongly correlated across all GOLD grades with exercise tolerance, ventilatory variables, ventilatory efficiency, and oxygen consumption ability. In particular, these correlations were weaker in the GOLD III-IV group with exertional oxygen desaturation. In contrast, the _E-_CO2 slope was specifically correlated with oxygen consumption ability and exercise tolerance regardless of COPD grade and exertional oxygen desaturation, having only a minimal correlation with _E.

Admittedly, though OUES and the _E-_CO2 slope are commonly used indicators of gas exchange efficiency evaluated by CPET, there have been no reports of their clinical use in COPD patients based on the differences between the two indicators. In particular, little is known about what OUES really means. Based on the formula, _O2 = OUES × log_10E + intercept, given that the value of log_10E is smaller than that of the actually measured value of _E, and that OUES represents the absolute rate of increase in _O2 per log_10E, even if minor ventilatory abnormalities occur up to submaximal exercise before reaching peak exercise, OUES allows detection of the resultant ventilatory variations more sensitively than does the _E-_CO2 slope, which is evaluated using the actual measured value of _E. Indeed, Barron et al. [17] reported that OUES, rather than the _E-_CO2 slope, was useful for differentiating between heart failure requiring excessive ventilatory compensation and COPD resulting in reduced ventilatory ability with the same exercise tolerance, which might indicate that OUES is more sensitive to the exertional ventilatory response during exercise. In the present study, regardless of COPD severity, compared with the _E-_CO2 slope, OUES correlated with _E at peak exercise, although the correlation was not strong (Tables 3 and 5). Certainly, in the population investigated in the present study, OUES may have been extensively, but never strongly, correlated with not only _E and ΔF_O, but also with peak _O2, and O₂-pulse as an indicator of cardiac function. However, if COPD progressed to GOLD III and IV with exertional hypoxia, OUES correlated only weakly with the above exercise variables (Fig. 3). In the GOLD III-IV group with exertional hypoxia, _E at peak exercise was lower, and O₂-pulse tended to be lower, but there were no significant differences in HR at peak exercise and HR at peak exercise/predicted maximal HR between the groups with and without exertional hypoxia (Table 4). These findings may not suggest that cardiac function is decreasing in the group with exertional hypoxia, but they may suggest that exertional hypoxia is involved as one of the reasons for the reduced O₂-pulse during exercise [4]. Furthermore, though Fig. 1 and online supplement Figure S2 demonstrate that the relationship between _O2 and _E during exercise is such that using the log transformation of _E was valid in the present study, a lack of increase in _E during exercise in advanced COPD might decrease the utility of the logarithmic conversion of _E, because the relationship between _E and _O2 was more linear in the desaturation group of the GOLD III-IV group (Fig. 4). Given that most advanced COPD is associated with reduced _E and exertional hypoxia, the clinical utility of OUES, especially in advanced COPD may be less.

In advanced COPD, due to the existing ventilatory disorders, even with maximal ventilatory effort only reduced _E can be achieved, which is a factor that limits exercise performance and reduces exercise tolerance [18, 19]. In such cases, to enable higher _O2, the subjects often rely on the compensatory increase in ΔF_O2 due to the reduction in F_EO2, that is, higher oxygen consumption ability throughout the body [9, 10]. In the present study, the dependence of ΔF_O2 on peak _O2, expressed as the square of the correlation coefficient, was higher in the GOLD IV group than in the other stages. The following two conflicting findings in our previous studies might be very useful to suggest that specific CPET parameters related to exertional oxygen consumption ability would enable planning of treatment strategies and their evaluation. First, increasing ΔF_O2 at peak exercise in pulmonary rehabilitation with exercise therapy in severe and very severe COPD patients led to improvement in both ventilatory efficiency and peak _O2 [10]. Second, in contrast, a treatment specifically aimed at improving ventilation using expiratory pressure load training, resulted in improved ventilation without improving ΔF_O2 at peak exercise in patients with COPD, improving peak _O2, but with no significant change in ventilatory efficiency [20]. When assessing exertional variables using CPET in many diseases, many reports have shown that the _E-_CO2 slope is related to exercise intolerance [8, 21]. So far, the _E-_CO2 slope has been considered to mainly reflect lung efficiency as a physiological mechanism. However, the _E-_CO2 slope was also found to be affected by the oxygen consumption ability resulting from circulatory disturbance, in addition to excessive ventilation in the lung [22], and Nayor et al. [23] reported that _E/_CO2 nadir is related to the arteriovenous O₂ difference at peak exercise, suggesting that ventilatory efficiency correlates with the oxygen extraction ability of peripheral muscles. In the present study, ΔF_O2 at peak exercise correlated with ventilatory efficiency not only across all COPD grades (Table 5), but also in advanced COPD, regardless of exertional hypoxemia (Fig. 3). In addition, stepwise variable analysis, including the COPD grade and the SpO₂ level at peak exercise, identified ΔF_O2 at peak exercise rather than peak _O2 as a more influential variable correlated with the _E-_CO2 slope across all GOLD grades. Although this needs to be studied in other diseases that result in excess ventilation during exercise, the above findings suggest that, at least in ventilation-limited diseases such as COPD, the _E-_CO2 slope is a specific indicator of oxygen consumption ability that is independent of ventilatory ability to increase _E. Given that (1) for advanced COPD with exertional hypoxia, assessment of exercise variables by OUES appears to have limitations, as shown in the present study, and that (2) in advanced COPD, pulmonary rehabilitation is often required to focus on improving oxygen consumption ability rather than ventilation due to the presence of a severe ventilatory disorder, the _E-_CO2 slope might be useful for planning and assessing personalized pulmonary rehabilitation.

This study has some limitations. First, since the present study involved COPD patients at a single center, a multicenter evaluation including not only COPD patients, but also those with many other diseases, is needed to confirm the physiological mechanisms reflected by the _E-_CO2 slope and OUES. Second, assessment of blood circulation, such as by echocardiography or right heart catheterization, would have been useful [24, 25]. Given that there was no strong relationship between the arterial desaturation level and ΔF_O2 at peak exercise in the non-desaturation group and the desaturation group in the GOLD III-IV group in the present study, and that among the variables related to ΔF_O2, F_IO2 is almost constant at 21%, and, therefore, ΔF_O2 depends primarily on F_EO2, it was assumed that F_EO2 is probably closely related to mixed venous oxygen pressure, which is obtained by right heart catheterization. This assumption has not been fully investigated and requires further study. Third, treatment strategies targeted at improving oxygen consumption ability are underdeveloped [3, 26, 27]. If such treatments become available, the importance of the _E-_CO2 slope in relation to oxygen consumption ability might become more apparent. Increasing muscle mass while improving oxygen extraction ability, rather than simply increasing muscle mass as the endpoint, is necessary to improve exercise intolerance, which, as of now, still remains an unresolved issue.

Conclusions

We demonstrated the characteristics of OUES and the _E-_CO2 slope as indicators of pathophysiological exertional conditions in patients with all COPD grades. OUES is a comprehensive exercise index extensively involving pulmonary, cardiovascular, and muscle crosstalk in the body in COPD, but its association with them was never strong, and it may not be useful in cases of advanced COPD with exertional hypoxemia. In contrast, the _E-_CO2 slope is an index that is independent of ventilatory ability, and it assesses oxygen consumption ability and exercise tolerance in COPD cases, regardless of the exertional hypoxemia level and COPD grade. Therefore, the _E-_CO2 slope might be useful in establishing tailor-made pulmonary rehabilitation strategies for individual pathologies in COPD, serving as an indicator that focuses more on oxygen consumption ability independent of ventilatory ability.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

CPET

Cardiopulmonary exercise testing

ΔF_O2

Difference between inspired and expired mean oxygen concentrations

GOLD

Global Initiative for Chronic Obstructive Lung Disease

ICS

Inhaled corticosteroid

LABA

Long-acting β2-agonist

LAMA

Long-acting muscarinic antagonist

OUES

Oxygen-uptake efficiency slope

_E-_CO2 slope

Minute ventilation-carbon dioxide production slope

Author contributions

All authors contributed substantially to this article. M.K. was responsible for the study conception and design. Y. H., M.K., K.K., M.S., M.Y., N.Y., H.K., H.H., F.M., M.T., Y.R., S.S., N.T., M.T., T.K., and K.H. were responsible for data acquisition, analysis and interpretation. Y.H. was responsible for drafting the article. Y. H., M.K., K.K., M.S., M.Y., N.Y., H.K., H.H., F.M., M.T., Y. R., S.S., N.T., M.T., T.K., and K.H. were responsible for revising the article. Each author approved the submission of this manuscript for publication.

Funding

This work was supported by a Grant-in-Aid for Clinical Research from the National Hospital Organization (number, not applicable). The funder had no role in the study design, data collection and analysis, or preparation of the manuscript.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The institutional review board of the National Hospital Organization Osaka Toneyama Medical Center approved the study protocol (approval number: TNH-R-2022018) and the protocol was in accordance with the Declaration of Helsinki for experiments involving human subjects. The patients/participants provided their written informed consent for study participation before cardiopulmonary exercise testing.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.