Oxygen Uptake and Lactate Kinetics in Patients with Chronic Obstructive Pulmonary Disease during Heavy Intensity Exercise: Role of Pedaling Cadence

Abstract

Oxygen uptake slow component (${\dot{V}}_{{\mathrm{O}}_2}$_sc) is associated with lactate accumulation, likely a contribution of poorly oxidative muscle fibers. We aimed to test the hypothesis that higher muscle tension during slow pedaling rates would yield more prominent ${\dot{V}}_{{\mathrm{O}}_2}$_sc in healthy subjects, but not in COPD patients. Eight severe COPD patients and 8 age-matched healthy individuals performed 4 rest-heavy exercise transitions at 40 and 80 RPM. Work rates at the two cadences were balanced. Venous blood was sampled for measurement of lactate concentration at rest and every 2 minutes until the end of exercise. ${\dot{V}}_{{\mathrm{O}}_2}$ kinetics were analyzed utilizing nonlinear regression. ${\dot{V}}_{{\mathrm{O}}_2}$ phase II amplitudes at the two cadences were similar in both groups. In healthy individuals, ${\dot{V}}_{{\mathrm{O}}_2}$_sc was steeper at 40 than 80 RPM (46.6 ± 12.0 vs. 29.5 ± 11.7 mL/min², p = 0.002) but not in COPD patients (16.2 ± 14.7 vs. 13.3 ± 7.6 mL/min²). End-exercise lactate concentration did not differ between cadences in either group. In healthy individuals, greater slow-cadence ${\dot{V}}_{{\mathrm{O}}_2}$_sc seems likely related to oxidative muscle fiber recruitment at higher muscular tension. COPD patients, known to have fast-twitch fiber predominance, might be unable to recruit oxidative fibers at high muscle tension, blunting ${\dot{V}}_{{\mathrm{O}}_2}$_sc response.

Keywords:

• ${\dot{V}}_{{\mathrm{O}}_2}$ slow component
• oxygen uptake
• heavy intensity exercise
• lactate
• pedaling rate
• chronic obstructive pulmonary disease
• cycle ergometer
• gas exchange

Introduction

The kinetics of oxygen uptake (${\dot{V}}_{{\mathrm{O}}_2}$) at the onset of constant work-rate (CWR) exercise has been characterized as being composed of three phases. Phase I is considered to be the result of the immediate cardiovascular response after which ${\dot{V}}_{{\mathrm{O}}_2}$ increases monoexponentially toward a steady state (variably termed the primary, fundamental, or phase II kinetics). If the CWR is in the heavy intensity domain, that is, above the lactate threshold (LT), an additional delayed “slow component” (${\dot{V}}_{{\mathrm{O}}_2}$sc) develops (1, 2). In healthy individuals, the ${\dot{V}}_{{\mathrm{O}}_2}$sc has been characterized by an increase in ${\dot{V}}_{{\mathrm{O}}_2}$ between the 3rd and 6th minute of exercise (${\dot{V}}_{{\mathrm{O}}_2}$_(6-3)) (3, 4). The ${\dot{V}}_{{\mathrm{O}}_2}$sc initially becomes evident after about 2–3 minutes of heavy-intensity exercise, is predominantly due to a decrease in work efficiency within the exercising muscles (5, 6), and is associated with progressive motor unit recruitment (7) and muscle fatigue (8). The involvement of substrate level phosphorylation serves as a source of increased serum lactate concentration during heavy-intensity exercise (3, 4, 6). Several studies have shown that there is a close correlation between the magnitude of the ${\dot{V}}_{{\mathrm{O}}_2}$sc and the change in blood lactate concentration during supra-LT exercise (3, 9–12).

Current consensus suggests a mechanistic link of ${\dot{V}}_{{\mathrm{O}}_2}$sc to the recruitment of Type II muscle fiber contraction that are apparently less efficient energetically than Type I fiber contraction, thus contributing to a higher ratio of blood [lactate] to [pyruvate] (1, 6, 13). The ${\dot{V}}_{{\mathrm{O}}_2}$sc associated with muscle fatigue and increasing inefficiency in the very heavy intensity domain is logically associated with developing fatigue, exercise intolerance, and task failure. Barstow et al. showed that the magnitude of the ${\dot{V}}_{{\mathrm{O}}_2}$sc during heavy intensity CWR exercise was negatively correlated with the proportion of Type I fibers over a range of pedaling rates (14). Endo et al. (15), involving functional magnetic resonance imaging during heavy- and very heavy-intensity exercise in healthy individuals, demonstrated that increasing ${\dot{V}}_{{\mathrm{O}}_2}$sc magnitude was associated with a progressive heterogeneous muscle recruitment. There was a positive correlation between the magnitude of ${\dot{V}}_{{\mathrm{O}}_2}$sc and the activation of the muscles in the whole thigh, but the nature of the fibers recruited during ${\dot{V}}_{{\mathrm{O}}_2}$sc (i.e., whether type I or type II) has yet to be determined. Available evidence suggests that, for the same external power output, Type II muscle fibers may be recruited to a greater extent at low pedaling frequency (e.g., 40 RPM) (16, 17). Since work rate is the product of torque applied to the crank of the cycle and the number of revolutions per minute, it follows that higher muscle tension is required at low-pedaling frequency for a given work rate. Thus, an additional increase of ${\dot{V}}_{{\mathrm{O}}_2}$ at low pedaling frequency could be caused by recruitment of additional muscle fibers during slow movement, presumably related to higher muscle tension (16, 18). However, in COPD patients, muscle wasting affects primarily Type I fibers and there is a relative increase in number of Type II muscle fibers (19–23). We hypothesized that, if the ${\dot{V}}_{{\mathrm{O}}_2}$sc response is associated with additional fiber recruitment, then the slope of the ${\dot{V}}_{{\mathrm{O}}_2}$sc and the rate of lactate accumulation in severe COPD patients might be different from healthy individuals during heavy intensity CWR exercise at slow pedaling rates as compared to higher pedaling rates. If the ${\dot{V}}_{{\mathrm{O}}_2}$sc is associated with additional fiber recruitment, then it must involve both Type I and Type II recruitment and an increased lactate accumulation at slow pedaling rates as compared to higher pedaling rates. The aim of this study was, therefore, to examine the effect of pedaling cadence on the ${\dot{V}}_{{\mathrm{O}}_2}$sc and lactate accumulation when exercising at the same effective work rate in severe COPD compared with healthy individuals. Further, we hypothesized that the slope of the ${\dot{V}}_{{\mathrm{O}}_2}$sc in severe COPD patients might be lower than that of healthy individuals during heavy constant work-rate exercise due to the lower proportion the available Type I fibers in the COPD patients.

Material and methods

Study participants

We studied 8 clinically stable patients with severe COPD who were no more than mildly hypoxemic (i.e., oxygen saturation measured by pulse oximetry [SpO₂] at rest >92%, and SpO₂ during exercise >89% when respiring room air); they had no significant co-morbidity. Eight age-matched healthy individuals were also recruited. Included subjects had no significant co-morbidity. Specifically, subjects with significant cardiovascular diseases such as ischemia, congestive heart failure, or arrhythmias were excluded. The institutional review board at the Los Angeles Biomedical Research Institute at Harbor-UCLA Medical Center approved the protocol and all participants gave signed written informed consent.

Pulmonary function testing

Before exercise testing, study participants underwent pulmonary function testing for spirometry (Vmax 229, version 5-2A; SensorMedics; Yorba Linda, CA). Albuterol (200 μg) was administered to the COPD patients from a metered-dose inhaler 20 minutes before testing. Normal standards established by Hankinson et al. (24) were used to generate percent predicted values. Maximal voluntary ventilation (MVV) was calculated as forced expiratory volume in one second (FEV₁) times 40.

Cycle ergometer calibration

The cycle ergometer (ERGOLINE-800, Sensormedics, Yorba Linda, CA) was physically calibrated (VACUMED, Ventura, CA) and found to be accurate within between 10 and 200 W; the difference between the actual and set work rate averaged 0.55 ± 0.71 W. Based on the hyperbolic function of the ergometer, the calculated torque at 40 RPM was 1.5 times that at 60 rpm and at 80 RPM torque was 0.75 times that at 60 rpm.

Incremental exercise tests

A symptom-limited incremental exercise test was conducted on an electromagnetically braked cycle ergometer (Ergoline-800; SensorMedics) with a ramp slope of 5 or 10 W/min for the COPD patients and a ramp slope of 15 or 20 W/min for the control participants. During the incremental test, pedaling rate was kept constant at 60 RPM. Peak ${\dot{V}}_{{\mathrm{O}}_2}$ was determined as the average ${\dot{V}}_{{\mathrm{O}}_2}$ over the last 30 seconds of exercise. Lactate threshold (LT) was estimated from gas exchange measurements by the modified V-slope method (25). Seat height was recorded for the first test and reproduced for all subsequent tests. Subjects were instructed to avoid consumption of alcohol and caffeine 4 hours before each test and to avoid strenuous exercise 24 hours prior to testing.

Square-wave transitions

Each subject performed four square-wave transitions from rest to a power output calculated to require 30%Δ for each of the pedaling frequencies. The 30%Δ was calculated using the following formula: 30%Δ =${\dot{V}}_{{\mathrm{O}}_2}$_LT + 0.3 × (${\dot{V}}_{{\mathrm{O}}_2}$_peak − ${\dot{V}}_{{\mathrm{O}}_2}$_LT).

Since the ${\dot{V}}_{{\mathrm{O}}_2}$ is appreciably different between 40 and 80 RPM at the same external work rate, on the day of incremental test, each subject completed a 6 minute CWR exercise (20 W) at each pedaling rate in order to quantify the metabolic difference that is accounted for at the two pedaling rates. Assuming a 10 mL ${\dot{V}}_{{\mathrm{O}}_2}$/W increment, half of the calculated work rate difference was then added to the target work rate to be performed at 40 RPM and subtracted from the target work rate to be performed at 80 RPM. In this way, we targeted similar metabolic rates at the two pedaling cadences. The adjusted final work rates for each pedaling rate and group are presented in Table 2.

On each testing day, participants performed the transitions to each of the two cadences, separated by at least 2 hours, in a randomized order. Each square-wave transition consisted of 3 minutes of rest, an abrupt transition to 8 minutes of exercise at power output resulting in 30%Δ, and 10 minutes recovery at rest. In order to eliminate the effect of initial inertia of the ergometer, before the beginning of exercise the flywheel was manually accelerated above the speed corresponding to the desired cadence, and the exercise transition started the moment the participant started moving the legs. Fifteen minutes prior each exercise, albuterol (200 µg) was given to the COPD patients to achieve a stable state of bronchodilation.

Measurement of pulmonary gas exchange

Pulmonary gas exchange (${\dot{V}}_{{\mathrm{O}}_2}$ and ${\dot{V}}_{{\text{CO}}_2}$) and minute ventilation (${\dot{V}}_{\mathrm{E}}$) were measured breath-by-breath (Vmax Spectra, V 12-2, SensorMedics, Yorba Linda, CA). The mass flow sensor and the gas analyzers were calibrated immediately before each exercise test. In order to assess reproducibility, a metabolic simulator (Metabolic Calibration Kit No.17051 Vacumed, Ventura, CA) was used before each test to simulate predefined levels of ${\dot{V}}_{\mathrm{E}}$,${\dot{V}}_{{\mathrm{O}}_2}$, and${\dot{V}}_{{\text{CO}}_2}$(26) at four respiratory rates (20, 30, 45, and 55 breaths per minute). The coefficient of variations for${\dot{V}}_{{\mathrm{O}}_2}$,${\dot{V}}_{{\text{CO}}_2}$, and ${\dot{V}}_{\mathrm{E}}$ from the metabolic simulator were 2.4, 2.0, and 5.2% in the tests corresponding to the tests at 40 RPM (62 tests), and 2.9, 2.3, and 5.3% in the tests corresponding to the tests at 80 RPM (61 tests). In addition, we used a biologic control participant for system validation in a two-level subthreshold CWR (20 W and 70 W at 60 RPM for 6 minutes each) on a weekly basis to assess the coefficient of variation of ${\dot{V}}_{{\mathrm{O}}_2}$ (27). The mean (±SD) of ${\dot{V}}_{{\mathrm{O}}_2}$for CWR at 20 W was 0.705 (±0.054) L/min and at 70 W was 1.272 (±0.054) L/min with the coefficient of variations being 7.6% and 4.2%, respectively (calculated from 8 tests). This coefficient of variation was similar to what has been found previously in this laboratory (27).

Heart rate (HR) and oxygen saturation (SpO₂) were recorded with continuous ECG (Cardiosoft; SensorMedics) and pulse oximetry (N-200 pulse oximeter; Nellcor; Hayward, CA). On one of the test days, before the CWR exercise was performed, a 21–gauge catheter was placed in a dorsal hand vein. An infrared bulb was used to warm the hand, thereby arterializing the venous samples. This procedure has been found to yield pCO₂, pH, and lactate levels that approximate arterial values (28). Samples were taken within 60 seconds before the onset, and at 2, 4, 6, and 8 minutes after the onset of exercise. Blood samples were taken in heparinized syringes and were placed in iced water immediately. Each sample was analyzed within 30 minutes after blood sampling for pH and standard bicarbonate (model 1640; Instrumentation Laboratories; Lexington, MA). A portion of each blood sample was centrifuged to obtain plasma for lactate measurement using ion chromatography (Metrohm-Peak, Inc., Houston, TX).

Lactate measurement

Whole blood samples of 1.5 mL each were centrifuged immediately after the test and the plasma was frozen at −80 °C and stored. Lactate concentration was measured using 0.2 mL of plasma diluted in absolute alcohol by a factor of 10 in order to precipitate plasma proteins. The supernatant was evaporated in nitrogen stream and was reconstituted in 2 mL of deionized water. Samples were then filtered using 25 µm syringe filter (Fisherbrand). Ion chromatography (Metrohm 761 ion chromatograph with anion suppression, Metrosep A, Supp 5 250 mm column with a 0.02 mL volume loop; eluent: flow =7 mL/min, 3.2 mM Na₂CO₃, 1.0 mM NaHCO₃) was used. One milliliter of the prepared solution was loaded onto a 0.02 mL loop and run in the anion suppression column for 30 minutes. The difference between end exercise lactate and the resting lactate concentration for each pedaling rate was expressed as a delta value (ΔLAC) in mmol/liter (mM).

Data analysis

Breath-by-breath data for each step-transition was linearly interpolated at 1 second intervals, time aligned to the onset of exercise and bin-averaged for 5 second intervals to provide on- and off-transitions for each exercise test. The four repetitions were then ensemble averaged to improve the signal to noise ratio of the underlying physiological response. The time course of ${\dot{V}}_{{\mathrm{O}}_2}$ was described in terms of exponential function that was fit to the data using nonlinear regression technique by the Marquardt-Levenberg algorithm (SigmaPlot 10.0 Systat Software, Inc., San Jose, CA).

Time course of the gas exchange response

The mathematical model for the ${\dot{V}}_{{\mathrm{O}}_2}$ response was based on a two exponential function with 1 time delay and fitted to the averaged response for each subject starting at the visually selected beginning of phase II response (29) as follows: (1) $$f\left(t\right)=y\left(b\right)+A_1\times \left[1-{\mathrm{e}}^{(-t/\tau 1)}\right]+A_2\times \left[1-{\mathrm{e}}^{(-\left(t-\text{TD}2\right)/\tau 2)}\right]$$(1) where A₁×[1 − e^(−t/τ1)] denotes phase II (i.e., primary component) and A₂×[1 − e^{(−(t-TD2)/τ2)}] denotes the slow component phase, y(b) is the ${\dot{V}}_{{\mathrm{O}}_2}$ value at the end of phase I; A₁ and A₂ are the asymptotic values for the exponential terms; τ₁ and τ₂ are the time constants, and TD₂ is the time delay identified at the start of the ${\dot{V}}_{{\mathrm{O}}_2}$sc. In this model, phase II continues beyond the start of the ${\dot{V}}_{{\mathrm{O}}_2}$sc; therefore, at the end-exercise (t_lim), the total amplitude was set equal to baseline + A₁' + A₂'. Further, phase II ${\dot{V}}_{{\mathrm{O}}_2}$ amplitude was calculated at the start of phase III (at TD₂) as A₁″ in the form: (2) $$A_1″ = A_1\times \left(1-{\mathrm{e}}^{(-\text{TD}/\tau 1)}\right)$$(2)

These components of the nonlinear function are depicted in a control participant at 40 RPM (Figure 1a) and 80 RPM (Figure 1b).

Figure 1. Schematic showing the two-exponential model used to describe responses of ${\dot{V}}_{{\mathrm{O}}_2}$ during exercise in a healthy subject (a) and in a COPD patient (b) at 40 RPM (a), 80 RPM (b). Parameters correspond to those in Eq. 1. BL, baseline; EE${\dot{V}}_{{\mathrm{O}}_2}$, ${\dot{V}}_{{\mathrm{O}}_2}$at end exercise; A₁′: amplitude of primary component of ${\dot{V}}_{{\mathrm{O}}_2}$ at end of exercise; A₁′: amplitude of primary component of ${\dot{V}}_{{\mathrm{O}}_2}$ at time TD₂ (time delay); A₂': amplitude of the slow component ${\dot{V}}_{{\mathrm{O}}_2}$.

If the response was fundamentally monoexponential, the second exponential term in Eq. 1(1) $$f\left(t\right)=y\left(b\right)+A_1\times \left[1-{\mathrm{e}}^{(-t/\tau 1)}\right]+A_2\times \left[1-{\mathrm{e}}^{(-\left(t-\text{TD}2\right)/\tau 2)}\right]$$(1) was dropped out when F-tests confirmed that the root mean square (RMS) was not significantly reduced using the more complex model (30). The amplitude of the ${\dot{V}}_{{\mathrm{O}}_2}$sc relative to the end-exercise ${\dot{V}}_{{\mathrm{O}}_2}$ was calculated as A₂′/EE${\dot{V}}_{{\mathrm{O}}_2}$ (where EE${\dot{V}}_{{\mathrm{O}}_2}$ is the ${\dot{V}}_{{\mathrm{O}}_2}$ at end exercise). The ${\dot{V}}_{{\mathrm{O}}_2}$sc slope was calculated A₂'/EE${\dot{V}}_{{\mathrm{O}}_2}$divided by 8 minutes minus TD₂.

Statistical analyses

Results are presented as mean ± SD unless stated otherwise. Comparisons between responses at different pedaling cadences for both controls and COPD patients were made using a two-way repeated measures ANOVA with post hoc Holm-Sidak paired t-tests. The difference in ${\dot{V}}_{{\mathrm{O}}_2}$sc slope between 40 and 80 RPM (ΔSlope_40–80) and the difference of lactate concentration (end-exercise − resting) between 40 and 80 RPM (ΔLAC_40–80) in the two subject groups were compared using unpaired t-test or Mann-Whitney U test where appropriate. Statistical significance was accepted at p < 0.05. The coefficient of variation, expressed as a percentage, was used as a measure of variation in repeated measurements of ${\dot{V}}_{{\mathrm{O}}_2}$, ${\dot{V}}_{{\text{CO}}_2}$, and ${\dot{V}}_{\mathrm{E}}$ in physiologic and metabolic pump calibrations.

Results

Table 1 shows participant characteristics and pulmonary function tests at rest and during exercise in both COPD patients and healthy individuals. The anthropometric values of the age-matched healthy group were not different from COPD patients (Table 1).

Table 1. Participant characteristics, lung function and responses to incremental exercise.

	Healthy (N = 8)	COPD (N = 8)
Age (year)	57 ± 6	63 ± 9
Gender M:F	4:4	6:2
Height (cm)	173 ± 7	170 ± 13
Weight (kg)	76.5 ± 10.3	75.0 ± 10.2
Body mass index (kg/m^2)	25.4 ± 2.6	25.8 ± 2.0
FEV1 (L/min)	3.13 ± 0.87	1.07 ± 0.45*
FEV1 (% predicted)	95 ± 14	35 ± 10*
FVC (L/min)	4.17 ± 1.01	2.62 ± 0.88*
FVC (% predicted)	99 ± 13	64 ± 11*
FEV1/FVC (%)	75 ± 3.5	40 ± 8*
${\dot{V}}_{{\mathrm{O}}_2}$peak (mL/min)	1,745 ± 426	1,154 ± 225*
Peak work rate (W)	144 ± 27	73 ± 19*
Peak heart rate (beat/min)	151 ± 18	136 ± 14
LT ${\dot{V}}_{{\mathrm{O}}_2}$ (mL/min)	1,163 ± 269	781 ± 249*
LT ${\dot{V}}_{{\mathrm{O}}_2}$/${\dot{V}}_{{\mathrm{O}}_2}$peak (%)	67 ± 6	67 ± 15
${\dot{V}}_{\mathrm{E}}$ peak (L/min)	63.9 ± 17.8	39.3 ± 10.4*
${\dot{V}}_{\mathrm{E}}$ peak/MVV (%)	52 ± 13	98 ± 25*

Values are mean ± SD; ${\dot{V}}_{{\mathrm{O}}_2}$ peak, ${\dot{V}}_{{\mathrm{O}}_2}$ at end of incremental exercise; LT${\dot{V}}_{{\mathrm{O}}_2}$, ${\dot{V}}_{{\mathrm{O}}_2}$ at lactate threshold; ${\dot{V}}_{\mathrm{E}}$ peak, ${\dot{V}}_{\mathrm{E}}$ at end of incremental exercise; MVV, maximal voluntary ventilation.

* p < 0.05 healthy subjects versus COPD.

Table 2. ${\dot{V}}_{{\mathrm{O}}_2}$ kinetic parameters in COPD and healthy subjects during heavy-intensity cycle ergometry at 40 and 80 RPM.

	Healthy (N = 8)		COPD (N = 8)
	40 rpm	80 rpm	40 rpm	80 rpm
Work rate (W)	109 ± 24.1	94 ± 18.5	56 ± 17.4	42 ± 12.5
BL ${\dot{V}}_{{\mathrm{O}}_2}$ (L/min)	0.61 ± 0.06	0.61 ± 0.14	0.52 ± 0.12	0.54 ± 0.14
τ1 (seconds)	42 ± 12	41 ± 11	49 ± 17	53 ± 9
TD2 (seconds)	156 ± 44	152 ± 38	172 ± 46	204 ± 53^‡
A1′ (L/min)	1.46 ± 0.29^†	1.43 ± 0.24^†	1.04 ± 0.2	1.07 ± 0.23
A1′′ (L/min)	1.43 ± 0.3^†	1.41 ± 0.24^†	1.02 ± 0.21	1.06 ± 0.23
A2′ (L/min)	0.22 ± 0.07*^†	0.14 ± 0.07^†	0.08 ± 0.08	0.05 ± 0.03
EE${\dot{V}}_{{\mathrm{O}}_2}$ (L/min)	1.68 ± 0.29^†	1.58 ± 0.24^†	1.12 ± 0.26	1.12 ± 0.23
A1′ (%EE${\dot{V}}_{{\mathrm{O}}_2}$)	78.9 ± 7.6	82.1 ± 8.7	87.4 ± 8.9	88.1 ± 7.5
A1″ (%EE${\dot{V}}_{{\mathrm{O}}_2}$)	62.3 ± 5.7	64.3 ± 5.5	62.3 ± 3.8	63.6 ± 4.2
A2′ (%EE${\dot{V}}_{{\mathrm{O}}_2}$)	13.2 ± 4.4^¶	9.2 ± 4.5**	6.8 ± 5.4	4.7 ± 3.0
${\dot{V}}_{{\mathrm{O}}_2}$sc slope (mL/min^2)	46.6 ± 12*^†	29.5 ± 11.7^†	16.2 ± 14.7	13.3 ± 7.6

Work rate refers to the adjusted balanced external work rate applied to the cycle ergometer. Values are mean ± SD; BL${\dot{V}}_{{\mathrm{O}}_2}$ baseline ${\dot{V}}_{{\mathrm{O}}_2}$ at the beginning of phase II response; EE${\dot{V}}_{{\mathrm{O}}_2}$ end-exercise ${\dot{V}}_{{\mathrm{O}}_2}$, A₂' amplitude of the slow component of ${\dot{V}}_{{\mathrm{O}}_2}$ calculated using two-exponential model fit, ${\dot{V}}_{{\mathrm{O}}_2}$sc: ${\dot{V}}_{{\mathrm{O}}_2}$ slow component, ^*p < 0.05 40 rpm versus 80 RPM in healthy subjects, ^†p < 0.05 healthy subjects versus COPD at 40 or 80 RPM, ^‡p < 0.05 COPD versus control at 80 RPM, ^**p < 0.05 healthy subjects versus COPD (regardless of RPM), ^¶p < 0.05 40 versus 80 RPM (regardless of subject group).

Incremental exercise test

The COPD patients had severe obstruction as measured by low FEV₁ (35 ± 10% predicted) and FEV₁/FVC ratio (40 ± 8%) (Table 1). Accordingly, their severely decreased exercise capacity was due to ventilatory limitation as ${\dot{V}}_{\mathrm{E}}$ approached maximum voluntary ventilation (MVV) (98 ± 25%), while healthy subjects had appreciable ventilatory reserve at peak exercise (${\dot{V}}_{\mathrm{E}}$/MVV =52 ± 13%). The low LT in COPD occurred at the same percentage of ${\dot{V}}_{{\mathrm{O}}_2}$_peak as healthy individuals (67 ± 15% and 67 ± 6%, respectively) (Table 1).

${\dot{V}}_{{\mathrm{O}}_2}$ response kinetics to square-wave work rate transitions

Figure 2 shows the group average ${\dot{V}}_{{\mathrm{O}}_2}$ kinetic response to heavy intensity exercise at 40 and 80 RPM cadence in healthy subjects (Figure 2a) and COPD patients (Figure 2b). Table 2 shows the ${\dot{V}}_{{\mathrm{O}}_2}$ kinetic parameters for the square-wave transitions to heavy intensity exercise at each pedaling cadence. The phase II response amplitude of ${\dot{V}}_{{\mathrm{O}}_2}$ at 40 and 80 RPM did not differ in either group (1.46 ± 0.29 vs. 1.43 ± 0.24 L/min in healthy subjects and 1.04 ± 0.21 vs. 1.07 ± 0.23 L/min in COPD patients). Thus, we were successful in accounting for the difference in the oxygen cost of internal work due under the different cadence conditions. The ratio of A₁'/EE${\dot{V}}_{{\mathrm{O}}_2}$ was 78.9 ± 7.6% at 40 RPM and 82.1 ± 8.7% at 80 RPM in healthy subjects, and 87.4 ± 8.9% at 40 RPM and 88.1 ± 7.5% at 80 RPM in COPD patients. COPD patients were ventilatory limited in both cadences as end exercise ${\dot{V}}_{\mathrm{E}}$ approached their MVV at the end of heavy intensity CWR exercise at both pedaling rates (EE${\dot{V}}_{\mathrm{E}}$/MVV 96 ± 20% and 92 ± 21% at 40 RPM and 80 RPM, respectively). The EE${\dot{V}}_{{\mathrm{O}}_2}$ in the healthy group was 97 ± 6.8% of ${\dot{V}}_{{\mathrm{O}}_2}$_peak at 40 RPM and 92 ± 7.6% of the ${\dot{V}}_{{\mathrm{O}}_2}$_peak at 80 RPM. The EE${\dot{V}}_{{\mathrm{O}}_2}$in COPD group was 98 ± 11% of ${\dot{V}}_{{\mathrm{O}}_2}$_peak at 40 RPM and 98 ± 12% of ${\dot{V}}_{{\mathrm{O}}_2}$_peak at 80 RPM. Furthermore, the average slope of ${\dot{V}}_{{\mathrm{O}}_2}$ of the slow component was steeper at 40 RPM than at 80 RPM in the healthy group (46.6 ± 12 vs. 29.5 ± 11.7 mL/min², p = 0.002) but that of COPD patients was not different (16.2 ± 14.7 vs. 13.3 ± 7.6 mL/min², p = 0.53). Importantly, the ${\dot{V}}_{{\mathrm{O}}_2}$sc slope was significantly steeper in healthy group than in COPD patients at both pedaling frequencies (p = 0.0001 and p = 0.01 at 40 RPM and 80 RPM, respectively).

Figure 2. Group averaged ${\dot{V}}_{{\mathrm{O}}_2}$ response kinetics to heavy intensity exercise in healthy subjects (a) at 40 RPM (solid triangle) and 80 RPM (open triangle) and in COPD patients (b) at 40 RPM (solid circle) and 80 RPM (open circle). Point interval =10 seconds. Dotted lines (- - - - -) show the average peak ${\dot{V}}_{{\mathrm{O}}_2}$ from the incremental test in each group. In the ensemble averages, time zero represents the start of Phase II and the vertical lines represent start of phase II and end of exercise.

The time constant of the phase II ${\dot{V}}_{{\mathrm{O}}_2}$ responses was not different between 40 RPM and 80 RPM (42 ± 12 seconds vs. 41 ± 11 seconds, respectively, p = 0.95) in healthy subjects. In COPD, the ${\dot{V}}_{{\mathrm{O}}_2}$ time constant was appreciably, but not significantly, longer than in healthy subjects (49 ± 17 seconds for 40 RPM vs. 53 ± 9 seconds for 80 RPM, p = 0.51, see Table 2).

Lactate and pH responses to square wave work rate transitions

During the 8 minutes of constant work rate exercise the lactate concentration increased in a linear fashion such that at 40 RPM the LAC increase was somewhat steeper in the healthy group than at 80 RPM, this difference, however did not reach statistical significance. In the COPD group, the lactate concentrations were virtually identical at each time point on both pedaling frequencies (Figure 3). This resulted in no significant difference in ΔLAC between pedaling rates in either group (5.5 ± 2.1 mM vs. 4.3 ± 2.2 mM at 40 and 80 RPM in healthy group (p = 0.35), and 3.3 ± 1.9 mM vs. 3.0 ± 1.3 mM at 40 and 80 RPM in COPD patients (p = 0.69).

Figure 3. Group averaged blood lactate concentration over time during heavy-intensity exercise in healthy subjects and COPD patients during cycling at 40 RPM and 80 RPM. Mean and SE are shown, as well as mean regression between 2 and 8 minutes of exercise to illustrate the difference in trends.

The average decline in end-exercise blood pH from rest at 40 RPM did not differ significantly from that of 80 RPM in either group (−0.07 ± 0.03 vs. −0.04 ± 0.03 in healthy individuals and −0.04 ± 0.03 vs. −0.03 ± 0.03 in COPD patients). Moreover, there was no difference in change of standard bicarbonate (SBC) from baseline between 40 and 80 RPM in either group (−2.2 ± 0.85 vs. −2.2 ± 3.35 mEq/L in healthy subjects and −2.3 ± 1.8 vs. −2.1 ± 4.5 mEq/L in COPD patients).

Relationships between the ${\dot{V}}_{{\mathrm{O}}_2}$sc slope and ΔLAC

The slope differences between 40 and 80 RPM are shown on Figure 4. It is apparent, that the ΔSlope_40–80 was greater in healthy subjects than in COPD patients (17.1 ± 9.1 mL/min², 2.9 ± 15.9 mL/min², respectively, p = 0.046). The end exercise ΔLAC_40–80 tended to be higher in the healthy group, but there was essentially no difference in the COPD group. (Figure 4).

Figure 4. The difference in Phase III ${\dot{V}}_{{\mathrm{O}}_2}$ slope between 40 and 80 RPM (ΔSlope _{40–80 RPM} phase III ${\dot{V}}_{{\mathrm{O}}_2}$) (circle) and the difference of lactate concentration (end exercise – resting) between 40 and 80 RPM (ΔLAC_{40–80 RPM}) (triangle) for each healthy subject (H) and each COPD patient (C). *p < 0.05 ΔSlope _{40–80 RPM} phase III ${\dot{V}}_{{\mathrm{O}}_2}$in healthy subjects versus ΔSlope_{40–80 RPM} phase III ${\dot{V}}_{{\mathrm{O}}_2}$ in COPD patients. Individual data points, means, and SE are shown.

Importantly, however, there was a strong positive correlation between ${\dot{V}}_{{\mathrm{O}}_2}$sc with ΔLAC in both healthy (r = 0.637, p = 0.047) and COPD patients (r = 0.709, p = 0.002) with near-identical slopes (4.31 mL O₂/min²/mM for the pooled data), whereas the intercept was significantly higher in the healthy versus the COPD group (19.6 ± 9.0 vs. −0.32 ± 4.5 mL/min², respectively; mean ± SEM) (Figure 5).

Figure 5. Correlation between the slopes of ${\dot{V}}_{{\mathrm{O}}_2}$_sc and the change in lactate concentration between end exercise lactate and the resting (ΔLAC) at 40 and 80 RPM in the healthy group (triangles) and in the COPD group (circles). The slopes are not different (p = 0.6678), but the intercepts are significantly different (19.5 ± 9.02 vs. −0.32 ± 4.51, p = 0.0008).

Discussion

The main finding of this study is that in the group of patients with COPD the slopes of the ${\dot{V}}_{{\mathrm{O}}_2}$sc were virtually identical at low (40 RPM) and high (80 RPM) pedaling rates, while in the healthy group the slope was significantly steeper at low pedaling rate than that of at high pedaling rate. Thus, there is a distinct difference in the response pattern of ${\dot{V}}_{{\mathrm{O}}_2}$ between healthy subjects and COPD patients in the heavy exercise domain depending on the cycling cadence. These findings are consistent with our primary hypothesis.

Method comparison

In this study, we successfully compensated for the internal work rate associated with the pedaling rate by adjusting the work rates assuming a 10 mL/W increase in ${\dot{V}}_{{\mathrm{O}}_2}$. Thus, the amplitude of the primary Phase II component of ${\dot{V}}_{{\mathrm{O}}_2}$ was not influenced by pedaling rate either in the healthy group or in COPD patients. However, these amplitudes were larger in healthy subjects than in COPD patients because the work rate equivalent to 30% delta was higher in healthy group. We believe that the precisely matched metabolic rate for the two different pedaling rates in each of the two groups serves a reasonable platform for comparing the slow component kinetic responses in the heavy intensity domain.

The time constant of Phase II ${\dot{V}}_{{\mathrm{O}}_2}$ response was the same with different pedaling rates, but in the COPD group, the ${\dot{V}}_{{\mathrm{O}}_2}$ time constant tended to be longer than in healthy group (though this difference did not achieve statistical significance [Table 2]). It has been shown that the time constant of the main component is independent of the actual work rate domain (31). However, the amplitudes of the response pattern (Phase II and slow component) are highly dependent on the metabolic requirements of the given work rate. This was the reason why we attempted to adjust the work rate such that we achieve similar metabolic rates in our study in order to have a solid basis of kinetic analysis in which the different primary amplitudes at different cadences were not factors to consider.

In this study, the subjects performed four individual transitions with the expectation that, with averaging, the breath-by-breath noise would be roughly halved (32). Using nonlinear regression analysis, we achieved a high confidence in model parameter values partially because we started the exponential function at the start of Phase II, thus eliminating the need to identify additional parameters related to Phase I, and partially because of the good signal-to-noise ratio achieved by averaging several transitions (32).

Previous studies of oxygen uptake kinetics (14, 33) employed a contiguous model involving 3 exponential components corresponding to all three phases. In our studies, exercise started with pre-spun flywheel and the work rate was applied as soon as the subject started to pedal. This helped us to isolate Phase I and start the kinetic analysis from the beginning of Phase II. This practice helped effectively eliminating the need to identify two additional parameters in the model (i.e., the time constant for Phase I and the primary time delay [TD1]). Furthermore, characterizing slow component with a third time constant in the model inherently assumes that all units contributing to the slow component kinetics are ‘turned on’ at the same time (i.e., at time of TD2). However, the same response profile can potentially be described without presuming that the slow component is a distinct component caused by well-characterized functional and structural contributors. Theoretically, it is possible to construct a model that would consider newly, but asynchronously recruited units characteristically with the same time constant as Phase II (34). We acknowledge that our modeling strategy did not attempt to model units asynchronously joining to perform the external work.

Role of pedaling cadence on oxygen uptake kinetics during heavy-intensity exercise

Other investigators have studied the influence of pedaling rate on oxygen uptake kinetics. Barstow et al. reported that changing the pedaling rate between 45 and 90 RPM had no effect on the amplitude of the ${\dot{V}}_{{\mathrm{O}}_2}$sc (14). Pringle et al. demonstrated that the amplitude of the ${\dot{V}}_{{\mathrm{O}}_2}$sc was greater at higher pedal rate (115 RPM) but there was no difference in amplitude of ${\dot{V}}_{{\mathrm{O}}_2}$sc with pedaling rates between 75 and 35 RPM (35). In contrast, in the present study, we have shown opposite results in that we found a steeper slope of the ${\dot{V}}_{{\mathrm{O}}_2}$sc in healthy individuals at slow pedaling rate and no difference in COPD patients between low and high pedaling rates. We see two potential reasons to explain this difference: first, we successfully adjusted for the internal work difference imposed by the two pedaling rates. In the study of Pringle et al. (35), however, the Phase II ${\dot{V}}_{{\mathrm{O}}_2}$ component averaged 1.47 ± 0.13 L/min at 35 RPM and 1.54 ± 0.15 L/min at 75 RPM which, though not a significant difference, suggests that the actual work rate was lower at the lower pedaling rate, perhaps affecting their results. Second, recruiting additional oxidative fibers can cause an additional rise of ${\dot{V}}_{{\mathrm{O}}_2}$ at low pedaling rates that is presumably related to higher muscle tension (2, 16, 18, 36). Each recruited fiber initially needs energy to overcome internal friction and to stretch before delivering external work (18). Sargeant et al. has demonstrated in healthy individuals that Type I fibers have lower power-generating capability compared to Type II fibers. However, Type II fibers are fatigue sensitive fibers and are recruited at a higher fraction of maximum power after type I fibers have already been fully recruited (17). Indeed, it has been suggested that the more fatigue sensitive Type II fibers are recruited earlier than would be expected (7, 37–39). Using functional magnetic resonance, Endo et al. (15) recently demonstrated that increasing ${\dot{V}}_{{\mathrm{O}}_2}$sc magnitude was associated with a progressive heterogeneous recruitment of muscle fibers. It is further suggested that progressive recruitment of muscle fibers yields a poor ratio of force production to ${\dot{V}}_{{\mathrm{O}}_2}$ and leads to an elevated O₂ cost at high external power output resulting in appearance of ${\dot{V}}_{{\mathrm{O}}_2}$sc. The influence of changes in pedaling rate on the ${\dot{V}}_{{\mathrm{O}}_2}$ response to heavy intensity exercise in subjects with differences in muscle fiber type distribution, however, is not known. The reported effect of pedaling rate on muscle activation is conflicting in the available literature. In healthy individuals, Barstow et al. (14) did not find any interactions between muscle fiber type and changes in pedaling rate.

One might argue that the different absolute work rate between the two subject groups might be responsible for the differences in slow component amplitude. The lower the absolute work rate, the less is the slow component amplitude, so is the lactate concentration. However, we found that the ${\dot{V}}_{{\mathrm{O}}_2}$ at end exercise normalized to the individual peak ${\dot{V}}_{{\mathrm{O}}_2}$ was higher in the COPD group (143 ± 17% at 40 RPM and 145 ± 19% at 80 RPM in COPD vs. 133 ± 13% at 40 RPM and 128 ± 14% at 80 RPM in healthy subjects). Hence the relative work rate was—if anything—higher in the COPD group than in the healthy subjects, hence encroaching more the heavy-very heavy work rate domain, which in turn should have contributed to a more pronounced slow component in the COPD group. In fact, referring to the significant difference in the intercepts of ${\dot{V}}_{{\mathrm{O}}_2}$sc versus ΔLAC on Figure 5, one might conclude that it is the absolute and not the relative work rate that is associated with lactate production, although the question why similar lactate production results in much different ${\dot{V}}_{{\mathrm{O}}_2}$sc still remains unanswered.

Individuals with severe COPD who have muscle wasting have extremes of muscle fiber type distribution (19, 22) and might have homogeneous pattern of distribution; they therefore might respond differently to changes in pedaling rate. This might also mean occurrence of additional fiber recruitment at higher muscle tension that might manifest at slow pedaling rates. This might also explain why the slow component oxygen uptake response is steeper in control participants because Type II fiber predominance has been seen even in COPD patients without clear signs of weight loss or muscle wasting. This is supported by the comparable ΔLAC levels in the two groups. Several other studies have demonstrated significant correlation between % Type II muscle fibers and the relative amplitude of the ${\dot{V}}_{{\mathrm{O}}_2}$sc (14, 33). However, no direct evidence exists that the mechanism responsible for the ${\dot{V}}_{{\mathrm{O}}_2}$sc is the progressive recruitment of Type II fibers.

An alternative hypothesis to explain the ${\dot{V}}_{{\mathrm{O}}_2}$sc is that fatigue in the Type II fiber pool as exercise proceeds necessitates an increased activity of Type I fibers (35, 40), which have been reported to be relatively less efficient at higher force requirements (41, 42); therefore, a higher oxygen requirement would be expected for the same rate of tension generation. This was supported by the study of Cannon et al. (8) that showed that muscle fatigue measured by a reduction in force output occurs in healthy subjects performing heavy- and very heavy-intensity exercise and initiates the rise in the O₂ cost of exercise where the ${\dot{V}}_{{\mathrm{O}}_2}$sc is present, the amplitude of which was independent of pedaling rates (60, 90, and 120 RPM). It is worth noting, however that in this experiment the fatigue-related reduction of velocity-specific peak power was greatest at 120 RPM, three times as much as at 60 RPM. The reduction of peak power would have been less at different cadence, had they applied the work rate adjustment technique to balance the effect of cadence-dependent pedal force. Nevertheless, these results cannot readily be extrapolated to the lower cadence rate of 40 RPM. We are not aware of any data to support this notion in COPD.

Our results were consistent with other previous studies that have demonstrated a correlation between the slope of ${\dot{V}}_{{\mathrm{O}}_2}$sc and lactate concentration increase during heavy intensity exercise (3, 9–11). However, some studies suggested that pathway of hepatic glycogen resynthesis could only contribute minimally to the additional ${\dot{V}}_{{\mathrm{O}}_2}$ (2, 43). The intercept of the relationship between ${\dot{V}}_{{\mathrm{O}}_2}$sc slope and lactate increase was higher in control participants than COPD patients. COPD patients, because of their known fast-twitch fiber predominance, might be unable to recruit oxidative fibers at high muscle tension, blunting ${\dot{V}}_{{\mathrm{O}}_2}$_sc response.

Clinical implications

Muscle dysfunction has been found to be prevalent in the COPD population. This contributes to exercise intolerance, which is often the chief complaint of these patients. To some extent, this reflects deconditioning related to a sedentary lifestyle. However, as shown in the differences in response to alteration of pedaling rate between COPD and age-matched healthy subjects study, fundamental differences in physiological responses to exercise exist that seem unlikely to be solely by deconditioning.

Limitations

We acknowledge a series of limitations in that we did not employ fiber typing; there was no fatigue assessment or EMG analysis. Furthermore, there was no assessment of blood flow through the working muscles, which has been shown to affect muscle oxygen uptake and the heterogeneity of the blood flow could contribute to the dynamics of oxygen uptake (for detailed review of this topic the reader is referred to an in-depth review of the heterogeneity of intramuscular blood flow by Koga et al. (44)). Optimally, all of the above factors could be considered if one intends to attempt addressing the long-debated question of the mechanism of ${\dot{V}}_{{\mathrm{O}}_2}$sc. Future studies might incorporate muscle biopsy so that differences in muscle characteristics between COPD and healthy subjects could be more definitively assessed. Finally, our results are specific to the case in which alterations in pedaling rate were made under iso-metabolic conditions (sum of external and internal work rate was constant). Studies in which changes in pedaling rate are made with external work rate was held constant might yield different results.

Conclusion

All of these should help to gain deeper insight of the mechanisms behind the major findings, that (1) ${\dot{V}}_{{\mathrm{O}}_2}$sc is larger at slow cadence in healthy subjects, (2) ${\dot{V}}_{{\mathrm{O}}_2}$sc is shallower in COPD, and (3) there is no difference in ${\dot{V}}_{{\mathrm{O}}_2}$sc at 40 RPM and 80 RPM in patients with COPD.

Declaration of Interest

All authors declare no conflicts of interest.

Abbreviations
ΔLAC	=	Difference between end exercise lactate and resting lactate concentration
ΔLAC40-80	=	Difference of lactate concentration (end-exercise − resting) between 40 and 80 RPM
ΔSlope40-80	=	Difference in ${\dot{V}}_{{\mathrm{O}}_2}$ slope between 40 and 80 RPM
ANOVA	=	Analysis pf variance
COPD	=	Chronic obstructive pulmonary disease
CWR	=	Constant work-rate
EE${\dot{V}}_{\mathrm{E}}$	=	Minute ventilation at end exercise
EE${\dot{V}}_{{\mathrm{O}}_2}$	=	Oxygen uptake at end exercise
FEV1	=	Forced expiratory volume in 1 second
FVC	=	Forced vital capacity
LT	=	Lactate threshold
MVV	=	Maximal voluntary ventilation; ${\dot{V}}_{{\text{CO}}_2}$
	=	Carbon dioxide output; ${\dot{V}}_{\mathrm{E}}$: Minute ventilation; ${\dot{V}}_{{\mathrm{O}}_2}$: Oxygen uptake; ${\dot{V}}_{{\mathrm{O}}_2}$
	=	Oxygen uptake slow component

Additional information

Funding

R. Casaburi occupies the Grancell/Burns Chair in the Rehabilitative Sciences.