A Simplified Approach to Select Exercise Endurance Intensity for Interventional Studies in COPD

ABSTRACT

Time to exercise limitation (Tlim) in response to constant work rate (CWR) is sensitive to interventions in chronic obstructive pulmonary disease (COPD). This is particularly true when the pre-intervention test lasts between 3 and 8 min (Tlim_3′–8′). There is, however, no simple method to select a work rate which is consistently associated with Tlim_3′–8′ across the spectrum of COPD severity.

We assessed 59 GOLD stages II–IV patients who initially cycled to Tlim at 75% peak. In case of short (<3 min, low-endurance) or long (>8 min, high-endurance) tests, patients exercised after 60 min at 50% or 90%, respectively (CWR_{50%⇐75%⇒90%}).

Critical mechanical constraints and limiting dyspnea at 75% were reached within the desired timeframe in 27 “mid-endurance” patients (46%). Increasing work rate intensity to 90% hastened the mechanical-ventilatory responses leading to Tlim_3′–8′ in 23/26 (88%) “high-endurance” patients; conversely, decreasing exercise intensity to 50% slowed those responses leading to Tlim_3′–8′ in 5/6 (83%) “high-endurance” patients. Repeating the tests at higher (60%) or lower (80%) intensities fail to consistently produce Tlim_3′–8′ in “low-” and “high-endurance”, respectively (p > 0.05). Compared to a fixed work rate at 75%, CWR_{50%⇐75%⇒90%} significantly decreased Tlim's coefficient of variation; consequently, the required N to detect 100 s or 33% improvement in Tlim decreased from 82 to 26 and 41 to 14, respectively.

This simplified approach to individualized work rate adjustment (CWR_{50%⇐75%⇒90%}) might allow greater sensitivity in evaluating interventional efficacy in improving respiratory mechanics and exercise tolerance while simultaneously reducing sample size requirements in patients with COPD.

KEYWORDS:

• exercise
• COPD
• dyspnoea
• lung function
• endurance

Introduction

Exercise intolerance has been associated with chronic disability, poor health-related quality of life and mortality in COPD (1,2). In fact, strategies aimed at improving patients' ability to sustain a given physical task for longer (i.e., endurance) has been consistently related to positive clinical outcomes (3). As outlined in a recent European Respiratory Society taskforce (4), a high-intensity constant work rate (CWR) test to time to limitation (Tlim) (5) has evolved as a particularly useful approach to unravel the beneficial effects of interventions on exercise endurance in this patient population.

The above-mentioned taskforce (4), however, recognizes a key caveat regarding to CWR testing in intervention trials: how to select a work rate which is consistently sustained for a “reasonable” duration by most COPD patients? (3,6,7) Underlying this question is the notion that the test should last sufficiently long to be limited by abnormal mechanical-ventilatory responses and their sensory consequences (i.e., dyspnea) (8). Tests which might end up being too short (e.g., <3 min) might be precociously limited by leg pain and/or peripheral muscle fatigue. Conversely, the test should not be sustained for prolonged periods of time to avoid exercise termination for reasons other than critical mechanical constraints and limiting dyspnea, e.g., dry mouth, seat discomfort, boredom (9). Moreover, large inter-subject variation in baseline Tlim not only increases the sample required for intervention studies but also complicates the interpretation of efficacy (5,7). Based on these considerations, the European Respiratory Society taskforce recommended a pre-intervention Tlim between 3 and 8 min (Tlim_3′–8′) (4).

Unfortunately, there is no practical method to select a work rate which is consistently associated with Tlim_3′–8′ in patients with COPD. This important caveat largely stems from the fact that tolerance to a given work rate decreases hyperbolically above individual's highest sustainable work rate, i.e., critical power (7,9–12). Thus, Tlim is expected to vary greatly among subjects depending on where the selected work rate lies in the individual's power-duration relationship (5,6,7). For instance, if the selected work rate is substantially above the critical power, Tlim might be shorter than 3 min. Conversely, a sub-critical power test can be sustained for prolonged periods of time, e.g., longer than 8 min. Unfortunately, obtaining critical power in each patient is not clinically feasible as it demands several tests on different days (13). Thus, establishing a practical strategy to obtain Tlim_3′–8′ in a pre-intervention CWR test remains an important unmet clinical need in COPD.

In the present^{Figure 1}study, we prospectively tested the yield of a simplified approach in selecting a constant work rate which is consistently associated with Tlim_3′–8′ in COPD patients. Based on our previous experience with this testing modality (e.g., 14–16), we hypothesized that 75% peak (WR_75%) would lead to Tlim_3′–8′ in a sizeable fraction of patients. In those with longer (>8 min) or shorter (<3 min) Tlim, we anticipated that a test performed after 1 hour (13) at WR_90% or WR_50% would lead to Tlim in the desired range of 3 to 8 minutes. We reasoned that showing the value of such pragmatic approach would contribute to a more widespread use of exercise testing in the evaluation of pharmacological and non-pharmacological efficacy in COPD.

Methods

Subjects

Clinically-stable GOLD II-IV^{Table 1}COPD patients (1) were recruited from outpatient respiratory clinics at Hospital Sao Paulo (Sao Paulo, Brazil), Hospital de Clinicas de Porto Alegre (Porto Alegre, Brazil), and Kingston General Hospital (Kingston, Canada). Patients were required to present with postbronchodilator forced expiratory volume in 1 s (FEV₁)/forced vital capacity (FVC) and FEV₁ below the lower limit of normal (1). They were also required to have had no exacerbations or changes in respiratory medications over the last 2 months before study entry. Patients were excluded in the presence of any cardiovascular, neuromuscular or orthopedic co-morbidity known to impair exercise capacity and/or ability to pedal, need of long-term oxygen therapy or any contraindication to clinical exercise testing. This study received ethical approval from the Research Ethics Board of Hospital Sao Paulo, Brazil (#1397213), Kingston General Hospital, Canada (#DMED-872-05) and^{Figure 1}Hospital de Clinicas, Brazil (#17-0158). All subjects gave written informed consent.

Study protocol

After an initial visit^{Figure 2, Figure 3}in which detailed medical assessment, full pulmonary function tests and a ramp-incremental cycle exercise test to tolerance were performed, patients undertook a constant test at 75% of the pre-established peak work rate peak on a different visit (CWR_75%). Patients were not told that we targeted Tlim_3′–8′. This initial intensity was selected based on our own (14–16)), and others (17–20), experience with the test and the European Respiratory Society recommendations (4). Thus, we a priori anticipated that 75% peak would be associated with Tlim_3′–8′ in a sizeable fraction of “mid-endurance” patients. In case of “low endurance” at CWR_75% (i.e., Tlim < 3 min), the test was repeated at 50% peak (CWR_50%) after 60 min; conversely, “high-endurance” at CWR_75% (Tlim > 8 min) prompted another test at 90% peak (CWR_90%) (Figure 1). “High-endurance” patients were randomly told to stop cycling at the 8^th min or allowed to proceed up to Tlim; thus, we were able to properly analyze Tlim distribution (before work rate adjustment) in a sizeable fraction of patients. In order to assure that CWR_50% and CWR_90% did not substantially under- or overestimate endurance capacity in “low-”- and “high-endurance” patients, these subjects were invited to an additional visit when they exercised at slightly higher or lower intensities (CWR_60% and CWR_80%, respectively). Between-day repeatability and responsiveness of the work rate associated with Tlim were assessed, on different visits, after placebo or short-acting bronchodilators in a subset of patients (Figure 1).

Figure 1. Study protocol. CPET: cardiopulmonary exercise test; CWR: constant work rate; Tlim: time to exercise intolerance; BD: bronchodilator.

Procedures

Pulmonary function testing included routine spirometry, body plethysmography and transfer factor for carbon monoxide (TL_CO) using automated testing equipment (Platinum Elite Series™ Body Plethysmograph Systems, MGC Diagnostics Corporation, USA). Maximum voluntary ventilation (MVV) was calculated as FEV₁ times 40. The same reference values were used in all laboratories (21–23)

Exercise tests were performed on electronically braked cycle ergometers with standard metabolic and ventilatory data being measured by the Vmax™ 29 Encore CPET System in all centers (SensorMedics Corporation, Yorba Linda, CA, USA). Physiological data were averaged over 20-second intervals. Oxyhemoglobin saturation was monitored by pulse oximetry (SpO₂, %). Subjects rated symptoms of dyspnea and leg effort on a modified 10-point Borg scale every minute. Operating lung volumes were derived from inspiratory capacity (IC, L) measurements (14). The time course of continuous physiological responses was estimated by half-time (t^1/2, s); for discrete variables, we obtained the time elapsed to individual patients to reach benchmarks which has been associated with low endurance (70% VT/IC and dyspnea score “5”) (24). For the subset of subjects who repeated the work rate associated with Tlim_3′–8′ after inhaled placebo or bronchodilator, patients were required to withhold their currently COPD medication according to standard recommendations. In these visits, patients cycled to Tlim 20 min after inhaling 5 mL of nebulized saline or 5 mL albuterol 2.5 mg plus 0.5 mg ipratropium bromide.

Statistical analysis

Variables were summarized according to variable distribution (IBM™ SPSS™ Statistics version 20). One-way ANOVA or Student's t-test compared normally distributed variables: Kruskal-Wallis or Mann-Whitney tests compared nonparametric variables between groups. Within group comparisons at specific timepoints (Tlim and t½) were made using paired-samples T-test or Wilcoxon signed-rank test when appropriate. Two-way ANOVA with repeated measures contrasted physiological and perceptual responses across groups as a function of time. Limits of agreement of physiological and perceptual responses at Tlim pre- and postplacebo were analyzed in Bland-Altman plots. A p < 0.05 level of significance was used for all analyses.

Results

Patient characteristics

Eighty-two patients were screened for study participation. Eighteen patients were excluded due to at least one of the following: recent COPD exacerbation (N = 5), postbronchodilator FEV₁ above the upper limit of normal (N = 6), inability to cycle (N = 5), severe oxyhemoglobin desaturation on exertion (SpO₂ < 75%) and/or cardiovascular limitation on incremental CPET (N = 2). Five patients refused participation. The remaining 59 patients (35 males) presented, on average, with moderate to severe airflow limitation, significant gas trapping and moderate decrement in transfer factor. Incremental CPET revealed moderate impairment in exercise tolerance which was associated with ventilatory limitation and mild exertional hypoxemia (Table 1).

Table 1. Resting and exercise characteristics of the whole sample and subgroups based on tolerance to the initial endurance test (CWR_75%).

Variable	All (N = 59)	Low endurance (N = 6)	Mid-endurance (N = 27)	High endurance (N = 26)
Demographic/anthropometric
Age (yrs)	62 ± 1	64 ± 2	63 ± 2	60 ± 2
Gender (M:F)	35: 24	1: 5	13: 14^†	21: 5^*
BMI (kg/m^2)	27.4 ± 0.8	23.6 ± 1.7	26.1 ± 1.0	29.7 ± 1.3^*
Pack-years	47 ± 3	52 ± 14	50 ± 4	43 ± 4^*
Lung function
FEV1 (L)	1.3 ± 0.1	0.8 ± 0.1^‡	1.3 ± 0.1^†	1.6 ± 0.1^*
FEV1 (% pred)	49 ± 2	38 ± 3	49 ± 2	52 ± 3
GOLD spirometric stage
II	34 (57.6%)	1 (16.7%)	16 (59.3%)	17 (65.4%)
III – IV	25 (42.4%)	5 (83.4%)	11 (40.7%)	9 (34.6%)^*
FVC (L)	3.0 ± 0.1	2.1 ± 0.2	2.9 ± 0.1	3.4 ± 0.2^*
FVC (% pred)	81 ± 2	75 ± 5	82 ± 3	81 ± 3
FEV1/FVC	0.46 ± 0.02	0.37 ± 0.04	0.46 ± 0.02	0.47 ± 0.02
IC (% pred)	80 ± 2	74 ± 6	77 ± 3	85 ± 3
RV (% pred)	172 ± 7	187 ± 18	175 ± 12	165 ± 8
TLC (% pred)	109 ± 2	104 ± 7	107 ± 3	106 ± 3
RV/TLC	0.52 ± 0.13	0.60 ± 0.02	0.53 ± 0.19	0.50 ± 0.19
TLCO (% pred)	67 ± 4	45 ± 7	63 ± 4	77 ± 6^*
Peak exercise
Work rate (W)	80 ± 4	43 ± 9	71 ± 5^†	99 ± 6^*
Work rate (% pred)	57 ± 3	55 ± 11	56 ± 3	58 ± 2
O2 (ml/min/kg)	16.8 ± 0.8	12.3 ± 1.8	16.3 ± 0.8	18.5 ± 1.3^*
O2 (% pred)	73 ± 3	65 ± 10	75 ± 4	72 ± 4
E /MVV	0.84 ± 0.03	0.80 ± 0.08	0.85 ± 0.03	0.84 ± 0.04
SpO2 (%)	93.1 ± 0.7	91.2 ± 2.1	92.9 ± 1.1	93.2 ± 0.8
Borg dyspnoea	6.0 ± 0.3	4 ± 1	6.2 ± 0.5	6.3 ± 0.4
Borg leg discomfort	6.5 ± 0.3	4 ± 1^‡	6.5 ± 0.5	7.0 ± 0.4^*

Values are presented as mean ± SE or n (%).

^* p < 0.05 low vs. high endurance.

^† p < 0.05 mid vs. high endurance.

^‡ p < 0.05 low vs. mid endurance.

Definitions and abbreviations: BMI: Body mass index; FEV1: forced expiratory volume in 1 second; GOLD: Global Initiative for Chronic Obstructive Lung Disease; FVC: forced vital capacity; RV: residual volume; TLC: total lung capacity; T_LCO: transfer factor for carbon monoxide; ̇VO₂: total oxygen consumption; ̇V_E: ventilation; MVV: maximum voluntary ventilation; SpO₂: peripheral oxygen saturation.

Tolerance to CWR_75%

Twenty-seven (46%) “mid-endurance” patients exercised within the desirable time range in response to CWR_75% (Tlim_3′–8′). The remaining 32 patients either exercised for less than 3 min (N = 6 (10%); “low-endurance”) or at least 8 min (N = 26 (44%); “high-endurance”). As expected, “low-endurance” subjects presented with worse resting lung function and peak exercise capacity than their counterparts; conversely, “high-endurance” subjects had the best functional performance (Table 1).

Effect of changes in work rate in “low-” and “high-endurance” groups

Imposing lower (CWR_50%) and higher (CWR_90%) work rates led to Tlim_3′–8′ in 5/6 (83%) and 23/26 (88%) of “low-” and “high-endurance” patients, respectively. Median [interquartile range] decrease and increase in work rate were −10 [−10 to −7.5] W and 14 [12 to 17) W, respectively. As a consequence of these adjustments, Tlim increased by 120 [96] s or 107 [85]% and decreased by −260 [261] s or −50 [20]%, respectively (Figure 2A). Overall, Tlim_3′–8′ was eventually reached in 55/59 patients (93.2%) with the CWR_{50%⇐75%⇒90%} approach. Tlim's coefficient of variation decreased from 73.3% (mean (standard deviation) = 476 ± 349 seconds, coefficient of skewness = 1.402; p < 0.001) in response to CWR_75% to 29.6% (306 ± 90 seconds, coefficient of skewness = 0.479; p > 0.05) after work rate adjustment in the “low-” and “high-endurance” groups, respectively (Figure 2B). In the subset of patients who accepted repeating the test at slightly higher (CWR_60% instead of CWR_50%, N = 6) or lower (CWR_80% instead of CWR_90%, N = 17) work rate, only 1 “low-endurance” patient and no “high-endurance” patient stopped exercising within the desired timeframe (Tlim_3′–8′).

Figure 2. Individual time to limitation (Tlim) in response to constant work rate testing at 75% peak (CWR_75%) and after intensity adjustment (CWR_{50%⇐75%⇒90%}) in patients whom CWR_75% did not provide Tlim between 3 and 8 min (Tlim_3′–8′) (N = 32, panel A). Tlim distribution in response to CWR_75% and CWR_{50%⇐75%⇒90%} is shown in panel B. Note that a fraction of “high-endurance” patients were told to stop exercising at the 8th minute during CWR_75% (N = 8): this explains higher Tlim frequency near this time point. Tlim: time to intolerance; Tlim_3′–8′: time to intolerance within 3 to 8 minutes; CWR: constant work rate.

Considering the marked influence of pre-intervention variability on the number of observations required to uncover postintervention improvement (25), we calculated the minimal required sample size (N) to power a hypothetical study according to Tlim distribution after CWR_75% and CWR_{50%⇐75%⇒90%}. Decreasing Tlim's coefficient of variation with CWR_{50%⇐75%⇒90%} diminished the required N from 82 to 26 (to detect 100 seconds improvement) (26) and 41 to 14 (to detect 33% improvement) (26) at 5% significance level and 80% statistical power (27).

Physiological and sensory mechanisms of Tlim variability

As shown in Figure 3 and Table 2, decreasing exercise intensity from CWR_75% to CWR_50% led to lower metabolic, mechanical-ventilatory and perceptual adjustments to exercise in the “low-endurance” group. Conversely, those responses were accelerated in response to CWR_90% compared to CWR_75% in the “high-endurance” group (p < 0.05). Thus, similar end-exercise responses were obtained at longer and shorter timeframes, respectively (p < 0.05). Of note, work rate adjustments approximate the rate of change of physiological and sensory responses across the groups (Figure 4 and Table 2).

Table 2. Physiological and sensory responses to constant work rate (CWR) test at different percentage of peak WR in COPD patients separated by tolerance to the initial endurance test (CWR_75%).

	Low endurance			High-endurance
			Mid-endurance
	CWR75%	CWR50%	CWR75%	CWR75%	CWR90%
At Tlim
Time (s)	125 ± 13	225 ± 17	307 ± 17	778 ± 80	325 ± 21
̇VO2 (mL/min)	620 ± 76 ^†	586 ± 64	1100 ± 68	1458 ± 89	1536 ± 68 ^*
̇VO2(% peak)	101 ± 5	97 ± 5	97 ± 2	100 ± 2	106 ± 3 ^*
HR (% pred)	69 ± 4	65 ± 4	78 ± 3	78 ± 2	78 ± 2
̇VE/MVV	0.87 ± 0.08	0.87 ± 0.08	0.82 ± 0.03	0.83 ± 0.04	0.88 ± 0.04 ^*
VT/IC (%)	76 ± 7	70 ± 6	74 ± 2	75 ± 3	77 ± 2
IRV (L)	0.30 ± 0.06	0.42 ± 0.10	0.47 ± 0.04	0.58 ± 0.07	0.52 ± 0.06
EILV/TLC	0.92 ± 0.05	0.93 ± 0.07	0.92 ± 0.07	0.92 ± 0.06	0.92 ± 0.05
Borg dyspnea	5.8 ± 1	6.3 ± 0.5	6.4 ± 0.4	6.9 ± 0.5	6.9 ± 0.5
Time course
̇VO2 t^1/2 (s)	28 ± 4 ^†	38 ± 5	49 ± 3 ^‡	86 ± 6	43 ± 5^*
̇VE t^1/2 (s)	34 ± 5 ^†	46 ± 7	58 ± 9 ^‡	93 ± 7	51 ± 7^*
Time at VT/IC 70%	58 ± 6 ^†	188 ± 12^*	173 ± 44	201 ± 10	121 ± 9^*
Time at dyspnea “5”	118 ± 10 ^†	237 ± 12^*	243 ± 16 ^‡	431 ± 21	251 ± 14^*

Values are presented as mean ± SE.

^* p < 0.05 for intragroup comparisons.

^† p < 0.05 low vs. mid and high endurance at CWR_75%.

^‡ p < 0.05 mid vs. high endurance at CWR_75%. ̇VO₂: oxygen consumption; HR: heart rate; ̇V_E: ventilation; MVV: maximum voluntary ventilation; V_T: tidal volume; IC: inspiratory capacity; IRV: inspiratory reserve volume.

Figure 3. Metabolic, cardiovascular, mechanical-ventilatory and sensory responses in patients whom initial CWR_75% was shorter than 3 min (“low-endurance”) or longer than 8 min (“high-endurance”) and the effect of repeating the test at lower (CWR_50%) or higher (CWR_90%) intensities, respectively. ̇VO₂: oxygen consumption; HR: heart rate; IRV: inspiratory reserve volume; V_T: tidal volume; IC: inspiratory capacity; ̇V_E: ventilation; MVV: maximum voluntary ventilation. ^*p < 0.05 for CWR_90% vs. CWR_75% at isotime. ^ǂp < 0.05 for CWR_50% vs. CWR_75% at isotime.

Figure 4. Metabolic, cardiovascular, mechanical-ventilatory, and sensory responses patients whom initial CWR_75% was within the desired time range (Tlim_3′–8′) and those who reached Tlim_3′–8′ after repeating the test at CWR_50% (“low-endurance”) or CWR_90%(“high-endurance”).̇VO₂: oxygen consumption; HR: heart rate; IRV: inspiratory reserve volume; V_T: tidal volume; IC: inspiratory capacity; ̇V_E: ventilation; MVV: maximum voluntary ventilation. ^*p < 0.05 for low-endurance at CWR_50% vs. mid-endurance at CWR_75% and high-endurance at CWR_90%.

Between-day repeatability and responsiveness

Tlim increased by ∼11% after placebo (N = 12); however, postplacebo Tlim was invariably within 3 and 8 minutes. As shown in Figure 5 and e-Figure 1, key mechanical-ventilatory variables and dyspnea at Tlim showed acceptable repeatability (related to mean values). Tlim increased by 72 [11–107]% or 190 [30–365] seconds after broncodilators compared to baseline (e-Figure 2) (coefficient of skewness = 0.873 and 0.360, respectively; p > 0.05). Of note, improvement in Tlim after bronchodilators was associated with slower kinetics of ventilation and mechanical constraints, leading to a sluggish rate of dyspnea increase over time (Figure 5).

Figure 5. Mechanical-ventilatory and sensory responses in the subgroup of patients (N = 12) who exercised at the work rate associated with Tlim_3′–8′ after placebo and bronchodilator. IRV: inspiratory reserve volume; ̇V_E: ventilation; MVV: maximum voluntary ventilation; V_T: tidal volume; IC: inspiratory capacity. ^*p < 0.05 for postbronchodilator vs. initial/postplacebo tests at isotime.

Discussion

This study tested the success of a simple individualized approach in selecting a work rate leading to critical mechanical constraints and intolerable dyspnea within 3 minutes and to 8 minutes (Tlim_3′–8′) in COPD. We found that selecting 75% peak as the initial work rate (CWR_75%) (14–20) with subsequent upward (CWR_90%) or downward (CWR_50%) adjustments in those with long (>8 minutes) or short (<3 minutes) tests, respectively, led to Tlim_3′–8′ in more than 90% of patients. This was largely a result of the intensity-dependency of the mechanical-ventilatory (and resulting dyspnea) kinetics in COPD. Thus, at the work rate frequently selected in previous clinical trials CWR_75% successfully lead to Tlim_3′–8′ in “mid-endurance” patients (∼45%). Conversely, it required CWR_90% or CWR_50% to hasten or delay the responses to an extent that similar physiological and sensory limits were reached within the desired time range in “high-” and “low-endurance” patients, respectively (Figure 3). This simplified approach (CWR_{50%⇐75%⇒90%}) constitutes a practical alternative to elicit critical mechanical constraints and limiting dyspnea within the timeframe recommended by the European Respiratory Society for intervention studies in COPD (4).

The challenges in selecting a high-intensity constant work rate cycle test which allows sufficient time to adequately stress the respiratory system and to determine the impact of a therapeutic intervention have been extensively discussed elsewhere (e.g., 3,6,7). For instance, critical power as a percentage of peak work rate has a wide range in patients with moderate to severe COPD (33–90%) (7,9–12). In this context, we are aware of the concerns expressed by Whipp and Ward (5) and, more recently, by van der Vaart et al. (7) about the risks of setting endurance testing as a fixed percentage of peak work rate. Recognizing these complexities, we followed their recommendation to perform an initial endurance test based on a targeted fraction of the peak work rate and then, if the test duration falls outside a desired limit, the measured Tlim is used to select an adjusted work rate to repeat the test (5, 7).

The proposed approach (CWR_{50%⇐75%⇒90%}) was based on previous studies which found that CWR_75% provides Tlim_3′–8′ in a sizeable fraction of patients (14–20). This initial exercise intensity was also suggested by the European Respiratory Society taskforce (4). We opted to set the subsequent work rate at 90% peak for “high-endurance” patients as previous investigations showed that Tlim distribution at CWR_75% is skewed to high values (as confirmed in the present study; Figure 2B, upper panel) (14). Moreover, a near-maximal work rate might be required to elicit Tlim_3′–8′ in some ventilatory-limited patients (10,28). In fact, CWR_90% led to Tlim_3′–8′ in ∼90% of “high-endurance” patients; conversely, Tlim remained too long (i.e., >8 min) when exercise intensity was decreased to 80% in a subgroup of these patients. Our experience with more severe, “low-endurance” patients at CWR_75% (Table 1) indicated that Tlim_3′–8′ is more frequently obtained after substantial decrease in exercise intensity. Indeed, CWR_50%, but not CWR_60%, led to Tlim_3′–8′ in most patients.

Our results expose the key mechanisms explaining the large effect of work rate adjustments on Tlim in COPD. Thus, by definition, mechanical-inspiratory (e.g., a critical IRV) (24) and ventilatory (MVV) (29) ceilings are closer to resting values in patients than normal subjects. Consequently, modulation of the responses' time course in “low” and “high-endurance” patients had a profound effect on the time elapsed before attainment of relatively-fixed constraints at Tlim (Figure 3) (30). Of note, decreasing or increasing the target work rate slowed or accelerated the rate of increase in ventilation: as elegantly showed by Puente-Maestu and co-workers, this has a marked influence in the time course of operating lung volumes.(12). It is also noteworthy that the high between-day repeatability of key “quantitative” (̇VE/MVV) and “qualitative” (VT/IC, EILV/TLC) indexes of exercise ventilation at Tlim (Figure S1). These data provide the physiological underpinnings of the remarkable similarity of end-exercise dyspnea scores (Figure 3F).

Our results have some important practical implications for efficacy assessment in interventional studies in COPD. Provided that the previous incremental test is adequately performed (29), the approach herein described (CWR_{50%⇐75%⇒90%}) can be easily applied in multicenter studies. In the minority of patients (<10% in this study) in whom Tlim_3′–8′is not reached after two same-day tests, another work rate might be tried on a different visit: Figure 2 suggests that Tlim_3′–8′ would require a maximal or even supra-maximal work rate for most of these patients. Importantly, narrowing Tlim distribution led to a large reduction in the required sample size to power a potential study looking at the effects of interventions on exercise tolerance.(6) This has been cited as an important caveat of endurance testing (6,7): the CWR_{50%⇐75%⇒90%} approach, therefore, seems particularly valuable for large randomized controlled trials involving patients with COPD.

The present study has some limitations to be highlighted. Although the sample size compares well with studies which used Tlim as an outcome in COPD (also summarized in (4) and (16)) it remains unclear whether our yield in eliciting Tlim_3′–8′ (>90%) would be reproduced in large randomized trials. However, we tried to minimize this caveat by including patients with a large range of resting functional abnormalities (Table 1). It should be specifically noted that we do not claim that the selected work rates (CWR_{50%⇐75%⇒90%}) are invariably the “best” choices to test exercise endurance across the whole spectrum of COPD severity. At least in the present multi-center study involving patients with moderate to very severe COPD, we observed that this simplified protocol led to Tlim_3′–8′ in the vast majority of patients. Whether alternative, or complementary, approaches might also prove clinically useful remains to be prospectively tested.

In conclusion, a feasible method to select the work rate for endurance testing in a single postincremental exercise visit (CWR_{50%⇐75%⇒90%}) elicited critical mechanical constraints and limiting dyspnea within the timeframe recommended by the European Respiratory Society for patients with COPD (Tlim_3′–8′) (4). This simplified protocol might allow greater sensitivity in evaluating efficacy of bronchodilator in improving respiratory mechanics and exercise tolerance while simultaneously reducing sample size requirements in this patient population.

Acknowledgments

The authors thank Boehringer Ingelheim, Canada (in particular Dr. Alan Hamilton) for supporting part of this study. They are also indebted to Dr. Darcy Marciniuk, University of Saskatchewan who contributed to patients' recruitment.

Declaration of interest

The authors report no conflicts of interest.

Additional information

Funding

Boehringer Ingelheim, Canada [#381-11].