Paced-Walk and Step Tests to Assess Exertional Dyspnea in COPD

Abstract

This study reports on the development and test-retest reproducibility of a modified shuttle walking and step testing protocols to assess exertional dyspnea in patients with COPD. Patients with COPD randomly performed four externally paced 3-min bouts of shuttle walking at 1.5, 2.5, 4.0 and 6.0 km·h^{− 1} or stepping at constant rate of 18, 22, 26 and 32 steps·min^{− 1}. Walking and stepping procedures were repeated once, on a separate occasion. Ventilation, heart rate, gas exchange parameters and dyspnea Borg scores were obtained at the end of each exercise bout. Most patients completed walking or stepping at the slow and middle speed cadences but only 33% completed walking at 6.0 km·h^{− 1} and 40% completed stepping at 32 steps·min^{− 1}. Walking and stepping at progressively faster pace caused a progressive increase in ventilation and dyspnea. Test-retest Pearson correlation coefficients for ventilation, heart rate, gas exchange parameters and dyspnea scores over the four exercise bouts, all exceeded 0.80. Intra-class correlation coefficients were at least as strong as Pearson coefficients. Bland & Altman representation showed that for repeated Borg scores, 92 and 96% of points lied within 2 SD of the mean difference between test-retest values for walking and stepping. The majority of moderate and severe COPD patients completed the 3-min shuttle waking testing at 1.5 to 4.0 km·h^{− 1} and the 3-min of step testing at rates between 18 and 26 steps·min^{− 1}. Both field tests were highly reproducible in patients with COPD.

Keywords:

• Chronic obstructive pulmonary disease
• reproducibility
• agreement
• step test
• field test

INTRODUCTION

The diagnosis of chronic obstructive pulmonary disease (COPD) is based on spirometric measurement of the 1-second forced expiratory volume (FEV₁). Although necessary for diagnostic purposes and useful to follow the evolution of the disease, FEV₁ correlates poorly with symptom intensity, exercise capacity and health-related quality of life ([1]). Moreover, physicians rarely rely only on FEV₁ thresholds to make therapeutic decisions. Rather, the general recommendation is that treatment effectiveness should be based on the assessment of patient-perceived outcomes such as dyspnea, exercise capacity and perceived health ([2]).

Timed-walking tests are commonly carried out in patients with chronic diseases to provide a general assessment of exercise tolerance and physical capabilities on the basis of the distance walked in a pre-determined time frame, usually between 6 and 12 minutes ([3]). However, the lack of standardization of walking speed during these tests may compromise the evaluation of treatment effect upon repeated tests. This issue was illustrated in a study showing that, in COPD patients, the 6-minute walking test was less responsive to bronchodilation than an endurance shuttle walk, during which the walking speed was externally imposed ([4]). Moreover, the fact that the primary outcome of these tests is the “distance covered” rather than dyspnea reduces their ability to assess the impact of therapeutic interventions on exertional dyspnea especially given the lack of standardization in walking speed.

Stair climbing may also contribute to dyspnea during activities of daily living in patients with COPD. Step testing, an exercise modality consistent with this activity of daily living, could be useful to assess exertional dyspnea in patients with COPD since it is known to markedly enhance metabolic and ventilatory requirements due to the involvement of large muscle groups. To date, step testing has been mainly used for the prediction of maximal work capacity from the stepping heart rate ([5]) in the general healthy adult and pediatric populations ([6],7,8). The test has also been adapted for use in senior populations ([9]). The use of stepping in clinical practice has been limited to the detection of heart disease ([10], [11]) or exercise induced asthma ([12], [13]) and to the assessment of exertional oxygen desaturation in patients with various pulmonary diseases ([14],15,16). To date however, this testing methodology has not been applied for the assessment of exertional dyspnea.

The first objective of this study was therefore to develop walking and step tests that could effectively assess the degree of exertional dyspnea in COPD patients with a broad range of disease severity COPD and eventually monitor the efficacy of therapeutic interventions.

However, in order to confirm the meaningfulness of these tests in the clinical setting, good measurement properties are also needed in terms of validity, reproducibility and reliability. The second aim of this article was therefore to demonstrate the reproducibility of two newly designed externally paced field tests for the assessment of exertional dyspnea in patients with moderately severe COPD. We report on two parameters of reproducibility: reliability, the degree to which subjects can be distinguished from each other, despite measurement error ([17]) and agreement, the absolute measurement error ([18],19).

METHODS

Patient population

Forty-three stable patients with COPD (36M/7F) participated in the study. Subjects were recruited from the Montreal Chest Institute and the Institut Universitaire de Cardiologie et de Pneumologie de Québec. Patients were older than 50 years, had a smoking history of ≥ 10 pack-years, an FEV₁ ≤ 80% of predicted, a FEV₁/FVC ratio ≤ 70%, and had no change in treatment regimen over the preceding 4 weeks or during the course of the study. Patients were excluded if they had a history of asthma, if they used oxygen therapy, had a SpO₂ < 80% upon exertion, musculoskeletal problems or any other contraindication to exercise. The study protocol was approved by our respective institutional research ethics boards and written informed consent was given by all patients.

Experimental protocol

Each patient visited the laboratory on 5 separate occasions over a 2–3-week period. On their first visit, patients completed pulmonary function testing according to the ATS/ERS standards ([20],21,22) (Medisoft body box 5500®) and received medical clearance to participate in the study. Patients were then asked to perform the Incremental Shuttle Walk Test ([23]) to assess peak exercise capacity. Following a 30-minute rest period, patients were familiarized with the shuttle walking and stepping procedures at paces of 1.5 km·h^{− 1} and 18 steps·min^{− 1}, respectively. During Visit 2, the patient performed either the Shuttle Walk test or Step Test, in a randomized order. The same exercise test that was performed at Visit 2 was repeated at Visit 3, within 2 to 4 days of Visit 2. The same procedure was repeated at Visits 4 and 5 with the other exercise testing modality. On the second stepping test visit, patients were also asked to complete a 3-minute bout of self-paced stair climbing, one hour after having completed the step test.

Shuttle walk test

The Shuttle Walk Test protocol consisted of 4 bouts of 3-minute each at constant walking speeds of 1.5, 2.5, 4.0 and 6.0 km·h^{− 1}, respectively. The test was performed on a 10 m shuttle walk course in the hospital corridor with cones placed at 0.5 and 9.5 m to allow for a 0.5 m turning radius at either end. The patient started walking upon the start of the audio track and to arrive at the cone by the next auditory signal. Patients were instructed to maintain the imposed cadence for the entire 3-minute exercise bout or until they became symptom limited and felt unable to maintain the cadence, or were asked to stop if they fell behind the imposed pace.

Step test

The Step Test protocol was designed to cover a range of energy requirements equivalent to V˙O₂ between 15 and 25 mlO₂·kg^{− 1}·min^{− 1} considering the estimated oxygen cost of stepping ([24]) and the peak V˙O₂ typical of patients with mild to moderately severe COPD. It consisted of four bouts of 3-minute each, at constant stepping rates of 18, 22, 26 and 32 steps·min^{− 1}, respectively. Patients were instructed to start stepping on hearing the audio instructions for “step-up” indicating to place both feet up onto the first stair of the step, with the rate being targeted to the movement of each foot and “step-down” indicating to step back down to the floor, one foot after the other.

A 10-minute rest period was given between each walking or stepping bout to allow for breathing and heart rate to return to baseline. The tests were terminated once patients were unable to complete a given bout or after having completed all four bouts. Prior to, and at the end of each exercise bout, patients were asked to rate their level of breathlessness on the 10-point Borg dyspnea scale ([25]). At baseline and throughout all exercise tests, patients were equipped with a compact portable telemetric system (Jaeger Oxycon Mobile®) for monitoring of ECG and ventilatory and gas exchange parameters as well as a finger oximeter for measurement of transcutaneous oxygen saturation (SpO₂).

Three-minute self-paced stair climbing

After 1 hour of quiet sitting following the second step testing protocol, patients were asked to ascend and descend a standard flight of stairs of known height and step number for 3 minutes while wearing the portable metabolic measurement system. The total number of steps ascended and descended, the frequency and total length of pauses were recorded. Heart rate, ventilatory and gas exchange parameters were continuously monitored throughout the testing period.

Treatment of data

Outcome measures taken at the end of each bout of shuttle walking or stepping were: exertional dyspnea Borg score, heart rate (HR), oxygen consumption (V˙O₂), ventilation (V˙_E), breathing frequency (F_b), tidal volume (V_T. Baseline values were taken from the last 30-second of a continuous 3-minute quiet sitting rest prior to beginning the test. Average values were calculated from the last 30-second of each 3-minute bout recording. For the self-paced stair climbing test, individual recordings were examined; patients with a continuous pause exceeding 45 seconds over the entire 3-minute period or 30 second over the last minute of the test were excluded from the analysis.

Statistical analysis

Data are shown as means ± standard deviation. Mean comparisons for the effect of walking or stepping rates on dyspnea or cardiorespiratory variables were achieved using a 1 × 5 (baseline and walking or stepping rates) ANOVA for repeated measures. Repeated measures ANOVA were used to compare cardiorespiratory parameters during stair climbing and paced stepping. Post-hoc analyses were carried out using Turkey HSD.

The overall test-retest reproducibility of each field test was assessed by calculating a Pearson correlation coefficient between trials 1 and 2 for each dependent outcome measure. Test-retest reliability was evaluated using Intraclass Correlation Coefficient (ICC) as it reflects both systematic and random differences in test measures ([26]). An ICC, if higher than 0.75, was judged to be excellent ([27]). To assess the homogeneity and the consistency of each separate bout of shuttle walking or stepping, test-retest Intraclass Correlation Coefficients (ICC) were calculated. In addition, in order to test for the homogeneity of the relationship given the potential discriminative effects of the various movement cadences, specific test-retest correlation coefficients were calculated for each cadence and compared to the overall correlation using Fisher-Z-transformations and Chi square analyses. All analyses were carried out using Statistica 6.0 (StatSoft®). Statistical significance was set at p < 0.05.

Finally, the limits of agreement between-trials on dyspnea scores measurement were established using a Bland and Altman representation for both walking and stepping. The proportion of scores ± 2 standard deviations of the mean difference between test-retest values was taken as a parameter of agreement.

RESULTS

Description of patient population

Patients' characteristics and pulmonary function data are presented in Table 1. Results show on average, moderate to severe airflow obstruction, hyperinflation and impaired diffusion capacity. The majority of patients showed moderate to severe disease as attested by their classification into GOLD stages II and III (37 of 43). Six of the 43 patients were classified as GOLD stage IV.

Table 1 Patients' characteristics and pulmonary function.

	Total (N = 43)
Age (years)	65 ± 7
Gender (M/F)	36/7
BMI (kg/m^2)	27 ± 6
FEV1 (L)	1.4 ± 0.5
FEV1 (% predicted)	49 ± 16
FEV1/FVC (%)	42 ± 11
TLC (L)	7.5 ± 1.4
TLC (% predicted)	124 ± 18
FRC (L)	5.2 ± 1.2
FRC (% predicted)	158 ± 32
IC (L)	2.5 ± 0.7
DLCO (% predicted)	59 ± 16
GOLD II n (%)	22 (51)
GOLD III n (%)	15 (35)
GOLD IV n (%)	6 (14)

Abbreviations: BMI: Body Mass Index; FEV₁: Forced expiratory volume in one second; FVC: Forced Vital Capacity; TLC: Total Lung Capacity; FRC: Functional Residual Capacity; IC: Inspiratory Capacity; DLCO: Diffusion capacity of carbon dioxide; GOLD: Global Initiative for Chronic Obstructive Lung Disease classification of COPD severity.

Values are mean ± standard deviation.

Peak exercise capacity

Peak exercise variables during the incremental shuttle walk test are shown in Table 2. As expected for patients with COPD, peak HR,V˙O₂, and ventilatory reserve were lower than the age predicted normal values. Peak SpO₂ showed exercise induced hypoxemia while most patients experienced severe dyspnea at end-exercise.

Table 2 Peak data from the incremental shuttle walk test.

	Patients (N = 43)
HR (beats·min^− 1)	120 ± 16
HR (% pred. max)	78 ± 11
V˙O2 (ml·kg^− 1· min^− 1)	16.9 ± 4.8
V˙E(L·min^− 1)	41 ± 13
V˙E/ MVV (%)	93 ± 26
Fb (breaths·min^− 1)	32 ± 6
VT (L)	1.3 ± 0.3
SpO2 (%)	89 ± 6
BORG dyspnea score	5.5 ± 2.7

Values are mean ± SD. HR: heart rate; V˙O₂ : oxygen consumption; V˙_E: ventilation; MVV: maximal voluntary ventilation; Fb: breathing frequency; Vt: tidal volume; SpO₂: oxygen pulse saturation.

Feasibility of 3-min shuttle walk and step tests

The totality of patients completed shuttle walking at 1.5 km·h^{− 1}, 40 (93%) at 2.5 km·h^{− 1}, 36 (84%) at 4.0 km·h^{− 1} and only 14 (33%) patients were able to fully complete shuttle walking at 6.0 km·h^{− 1} at either one of the trials. Thirty-five of the 43 patients (81%) completed stepping at 18 steps·min^{− 1}, 34 (79%) at 22 steps·min^{− 1}, 30 (70%) at 26 steps·min^{− 1} and only 17 (40%) patients were able to fully complete stepping at 32 steps·min^{− 1} at either one of the trials. Patients with severe disease found walking or stepping challenging with only 3 of 6 GOLD stage IV patients completing walking at 2.5 km·h^{− 1} and 1 of 6 GOLD stage IV patients completing the full 3 minute period at the slowest rate of stepping (18 steps·min^{− 1}). No incident were reported during exercise testing.

Cardiorespiratory responses

Mean Borg dyspnea scores at baseline and at the end of each period of walking and stepping are shown in Figure 1. The lowest walking speed and stepping rate induced a significant increase in the intensity of dyspnea from baseline with further significant increases in dyspnea intensity being observed for each subsequent increase in walking speed or stepping rate. Dyspnea intensity during stepping at 32 steps·min^{− 1} was similar to dyspnea intensity at the end of the incremental shuttle walking test.

Figure 1 Mean dyspnea score at the end of each walking speed (panel A) and stepping rate (panel B). The dotted line represents the mean of the peak incremental shuttle walk test for the entire group. *p < 0.05 from baseline; †p < 0.05 from previous walking speed or stepping rate. Values are mean ± standard deviation.

Table 3 shows V˙O₂, HR, V˙_E and breathing patterns for each walking and stepping rate. There were significant increases (p < 0.05) at each successive walking and stepping rate for all parameters except V_T during stepping. Thus, the stepping -induced increase in V˙_E was explained by a significant increase in Fb. HR, V˙O₂ and V˙_E measured during stepping at rates of 26 and 32 steps·min^{− 1} and during walking at 6 km·h^{− 1} were equal to or exceeded values recorded during the peak incremental shuttle walk test.

Table 3 Mean V˙O₂, HR, V˙_E, VT and Borg dyspnea scores from trials 1 and 2 of the 3-min Shuttle Walk and the Step Test.

Shuttle Walk	Trial	1.5 km·h−1	2.5 km·h−1	4.0 km·h−1	6.0 km·h−1
V˙O2 (ml·kg·min^− 1)	1	8.7 ± 1.2	10.4 ± 1.5	13.7 ± 2.1	20.8 ± 3.0
	2	8.6 ± 1.4	10.3 ± 1.5	13.7 ± 2.1	20.8 ± 3.1
HR (beats·min^− 1)	1	90 ± 12	93 ± 13	100 ± 12	122 ± 13
	2	91 ± 14	93 ± 14	100 ± 12	122 ± 13
V˙E (L·min^− 1)	1	20.8 ± 4.1	24.7 ± 5.1	30.5 ± 6.9	49.3 ± 12.3
	2	20.6 ± 4.6	24.6 ± 5.5	30.5 ± 6.8	49.7 ± 12.6
VT (L)	1	1.0 ± 0.2	1.0 ± 0.2	1.2 ± 0.3	1.5 ± 0.4
	2	1.0 ± 0.2	1.1 ± 0.2	1.2 ± 0.3	1.5 ± 0.4
Dyspnea Borg score	1	1.4 ± 1.6	1.5 ± 1.6	2.0 ± 1.5	4.6 ± 1.9
	2	1.4 ± 1.4	1.5 ± 1.9	2.0 ± 1.5	4.1 ± 1.4
Step Test	Trial	18 steps·min−1	22 steps·min−1	26 steps·min−1	32 steps·min−1
V˙O2 (ml·kg·min^− 1)	1	15.8 ± 1.6	17.9 ± 1.9	20.1 ± 2.0	23.1 ± 2.6
	2	15.4 ± 1.9	17.2 ± 2.1	19.8 ± 2.4	22.9 ± 3.3
HR (beats·min^− 1)	1	108 ± 12	116 ± 14	125 ± 15	131 ± 14
	2	107 ± 12	115 ± 12	122 ± 14	129 ± 14
V˙E (L·min^− 1)	1	35.5 ± 7.4	40.5 ± 7.8	47.1 ± 8.3	58.1 ± 11.7
	2	34.5 ± 7.0	39.3 ± 7.2	44.7 ± 6.9	54.9 ± 11.0
VT (L)	1	1.3 ± 0.3	1.4 ± 0.3	1.4 ± 0.3	1.4 ± 0.4
	2	1.3 ± 0.3	1.4 ± 0.3	1.4 ± 0.3	1.4 ± 0.3
Dyspnea Borg score	1	2.9 ± 1.3	3.7 ± 1.4	5.0 ± 2.0	5.4 ± 2.5
	2	2.4 ± 1.3	3.2 ± 1.5	4.5 ± 2.0	5.6 ± 2.5

V˙O₂ :Oxygen Consumption; HR: Heart Rate; V˙_E: Ventilation; VT: tidal volume. Values are means ± standard deviation.

There were significant increases (P < 0.05) at each successive walking and stepping rate for all parameters except V_T during stepping (the statistical symbols are not shown for the purpose of clarity).

Self-paced 3-minute stair-climbing test

On average, patients maintained a stepping rate of 35 ± 9 steps·min^{− 1}, with a range of 15 to 50 steps·min^{− 1}. Eight of the 43 patients paused for more than 30 seconds during the 3rd minute resulting in their data being excluded from the analysis. Of the remaining 35 patients, 15 (43%) stepped continuously over the 3-minute period without taking any pause. Of the 20 patients who paused, 50% paused more that once with an average pause time of 12 ± 10 seconds. Dyspnea intensity at the end of the self-paced test (5.7 ± 2.3) was not significantly different that at the end of the step tests at imposed paces of 26 and 32 steps·min^{− 1} (Table 3).

Similarly, there were no significant differences in HR (122 ± 14 beats·min^{− 1}), V˙O₂ (18.5 ± 3.3 ml·kg^{− 1}· min^{− 1}) and V˙_E (44 ± 11 L·min^{− 1}) at the end of the self-paced stair climbing test compared with stepping at 26 steps·min^{− 1}; values for self-paced stair climbing being higher than those at 22 steps·min^{− 1} but lower than those at 32 steps·min^{− 1} (p < 0.05). Tidal volume and breathing frequency were not significantly different between stair climbing and stepping at any stepping rate.

Test-retest of shuttle walk and step tests

As can be seen in Table 3, mean differences between trials 1 and 2 for V˙O₂, HR, V˙_E, VT and Borg dyspnea at each of the four cadences were less than 5% for all dependent variables and were not statistically significant.

Reliability of step test and shuttle walk test

Figure 2 shows the individual data for trials 1 and 2 obtained for for V˙O₂, V˙_E, V_T, F_b for both tests over the four cadences. Overall, the test-retest correlation coefficients for all variables were equally strong and statistically significant, with values never falling below 0.93 in all cases. Test-retest correlation coefficients for dyspnea scores over all cadences are shown in Figure 3 with the size of the symbol adjusted for the number of overlapping individual data points. The test-retest correlation coefficients obtained between Trials 1 and 2 for shuttle walking and stepping, were statistically significant and equally strong for both modalities.

Figure 2 Test-retest for (A) V˙ O₂, (B) V˙_E, (C) tidal volume (V_T) and (D) breathing frequency (F_b) (breaths.min⁻¹) for the Shuttle Walk (left) and the Step Test (right).

Figure 3 Borg dyspnea scores obtained at trials 1 and 2 for the 3-min Shuttle Walk (A) and Step Test (B) and Bland & Altman representation of difference in dyspnea Borg scores from trial 1 to trial 2 versus mean individual dyspnea Borg scores for Shuttle Walk (C) and Step Test (D). The size of each point is indicative of the number of data points it represents. Thus the larger the point, the more data points it represents. The solid horizontal lines within the Bland & Altman graphs represent the bias whereas the broken lines represent the upper and lower limits of agreement (±2SD of the mean value between trial 1 and 2).

The Test-Retest Pearson and Intraclass Correlation Coefficients (ICC) calculated for all dependent variables over the four cadences are provided in Table 4. Results show Pearson coefficients approximating 0.90 and higher for all physiological variables as well as the BORG score and ICC to be similar to the Pearson coefficients for both modalities. The test-re-test ICC calculated for the shuttle walking and stepping as a function of movement cadence are shown in Table 5. Results indicate excellent reliability across all movement cadences with no difference between cadences.

Table 4 Overall Pearson and Intraclass Correlation Coefficients for the Test-Retest 3-min Shuttle Walk and the Step Test.

	Shuttle Walk		Step Test
	Pearson r	Intraclass r	Pearson r	Intraclass r
V˙O2 (ml·kg·min^− 1)	0.97	0.98	0.93	0.93
HR (beats·min^− 1)	0.94	0.83	0.83	0.93
V˙E (L·min^− 1)	0.98	0.99	0.95	0.96
VT (L)	0.97	0.97	0.95	0.95
Fb (breaths·min^− 1)	0.94	0.92	0.93	0.92
Borg score	0.89	0.91	0.90	0.91

V˙O₂: Oxygen Consumption; HR : Heart Rate; V˙_E: Ventilation; V_T: tidal volume; Fb: breathing frequency.

Table 5 Intraclass Correlation Coefficient for the Test-retest on the 3-min Shuttle Walk and the Step Test as a function of movement cadence.

	Movement cadence
Shuttle Walk	1.5 km·h−1	2.5 km·h−1	4.0 km·h−1	6.0 km·h−1
V˙O2 (ml·kg·min^− 1)	.84	.86	.92	.82
HR (beats·min^− 1)	.86	.95	.93	.96
V˙E (L·min^− 1)	.94	.95	.96	.95
VT (L·breaths)	.94	.98	.96	.98
Fb (breaths·min^− 1)	.88	.92	.95	.91
Borg score	.88	.85	.78	.80
Step Test	18 steps·min−1	22 steps·min−1	26 steps·min−1	32 steps·min−1
V˙O2 (ml·kg·min^− 1)	.78	.83	.88	.87
HR (beats·min^− 1)	.87	.92	.93	.95
V˙E (L·min^− 1)	.92	.91	.92	.94
VT (L·breaths)	.94	.96	.95	.97
Fb (breaths·min^− 1)	.85	.89	.85	.95
Borg score	.79	.84	.92	.94

V˙O₂ : Oxygen Consumption; HR : Heart Rate; V˙_E: Ventilation; V_T: tidal volume; Fb: breathing frequency.

Agreement of step test and shuttle walk test

Figure 3 shows a Bland & Altman representation of the dyspnea scores obtained on trials 1 and 2 for both field tests. In both cases, more than 90% of data points lie within 2SD of the average values of the two trials, indicating good agreement between the two trials. The inter-trial agreement in dyspnea Borg score was relatively narrow with a mean difference (±2 SD) of 0.05 ± 1.76 for the Shuttle Walk and 0.42 ± 1.71 for the Step Test. There was no correlation between the mean Borg score of the two trials and the differences in Borg scores between trial 1 and 2.

DISCUSSION

This study establishes the feasibility of 3-min walking and step testing protocol to assess exertional dyspnea in patients with moderate to severe COPD. Results indicate good reproducibility, for both the walking and the stepping protocols with respects to dyspnea Borg scores. The two exercise tests were highly reproducible showing very good ICC-determined reliability. We also confirm their discriminative property as well as high agreement suggesting that these field tests are adequate evaluative tools. Results also showed good reproducibility for test-retests measurements of V˙O₂, V˙_E, VT and heart rate suggesting that the tests are adequate to capture the standard physiological responses to the exercise stimulus. Both externally paced walking and stepping induced levels of dyspnea that could eventually be amenable to pharmacological or rehabilitation interventions.

Feasibility of the shuttle walk and the step test

Self-paced timed walking tests are commonly used to obtain a one-time assessment of functional capacity in patients with chronic disorders because of their simplicity of use. The 6-minute walking test has been, and continues to be extensively used in COPD as it shows a good prognostic value for disease staging ([28], [29]). Similarly, step testing is one of the oldest field testing modalities first introduced to characterize the level of physical fitness based on the heart rate response to stepping ([7]). Because it requires little space and simple equipment to monitor, interest has been growing to use symptom-limited stair climbing in the clinical setting. It has been used successfully for pre-operative screening ([30], [31]), or to diagnose the presence of ischemic heart disease through the assessment electrocardiogram ([7]).

Our findings clearly establish the feasibility of both the walking and the stepping tests in patients with GOLD stage II to IV COPD as most patients completed at least 3 of the 4 incremental movement cadences. It may be suggested however that the cadences may need to be reduced for COPD patients with GOLD stage IV since only a few successfully completed more than the first lower movement cadences. The proposed tests appeared to be safe since no incident were reported with the current study population. However, experience gained in a larger population will be necessary before making a final statement on this issue.

A key assumption for an adequate assessment of dyspnea is that the test triggers a ventilatory response sufficient to reproduce the extent of symptoms reported by patients. Using the current 3-minute constant rate walking and stepping protocols, a significant increase in ventilation from baseline was seen. Increasing the walking speed by 1.0–2.0 km·h^{− 1} or the stepping rates by 4 steps·min^{− 1} resulted in significant increases in ventilation which also translated in significant increases in dyspnea intensity. The administration of the test could thus offer flexibility in selecting the appropriate level of ventilatory stimulus by selecting a lower or higher imposed walking or stepping rate.

Values of V˙O₂ and V˙_E measured in this study are in the range of those generally measured in COPD patients of similar disease severity using cycle ergometry or treadmill testing ([32], [33]) indicating that the walking and stepping protocols used induce a level of ventilatory stimulation similar to other tests used to examine the responses of therapeutic interventions on exertional dyspnea ([34]).

In the present study multiple bouts of stepping or waking at different paces were used as a preliminary approach to validate the ability of patients to sustain the selected stepping rates. It is conceivable however that these tests could be administered in such a way as to involve a single or at most two stepping or walking rates to ensure that the triggered breathing stimulus is adequate to assess exertional dyspnea. An algorithm could indeed be developed from pulmonary function data or MRC dyspnea score that would identify the walking or stepping rates for which a significant level of exertional dyspnea is likely to occur.

Similarity with self-paced stair climbing

The extent to which the selected stepping rates reflect stair climbing was assessed by comparing cardiorespiratory requirements measured during stepping and self-paced stair climbing. Measurements of V˙O₂, heart rate and V˙_E at the end of the standard stair climbing test were similar to those found for stepping at 26 steps·min^{− 1}, while dyspnea intensity was closer to that recorded for paced stepping at 32 steps·min^{− 1}. Presumably patients respected the instructions provided and selected a pace similar to that routinely used in their daily activities. Consequently, this would indicate that the step test at rates between 26 and 32 steps·min^{− 1} adequately captures the level of breathing constraints faced by patients with moderate to severe COPD when going up and down stairs as a component of their daily activities.

Reproducibility of the shuttle walk and the step test

Learning effect

Our results indicate no significant mean differences in either dyspnea or ventilatory parameters between trials 1 and 2 suggesting that no significant learning effect has occurred once patients have been familiarized with the procedure. This observation contrasts with results from studies using self paced walk tests showing that these may entail a significant learning effect leading to 7 to 9% improvements in distance walked under test-retest conditions ([28]). In the present study, the field tests combined both the externally-imposed pacing and a set time duration thus standardizing both total distance covered and cadence, and as a result, exercise intensity. Thus, standardizing movement pace would appear to reduce the potential for a significant learning effect to be observed, enhancing the observed repeatability.

Reliability

The ICC is the most suitable reliability parameter for continuous variable. Although the Pearson correlation coefficient is often used, this measure is limited by the fact that systemic differences are not taken into account. In our study, the two parameters have been used to allow comparison with other studies.

Results from 1 to 2 week test-re-test reproducibility of the 6-Minute Walk Test in older adults, patients suffering from heart failure, COPD, peripheral artery disease, idiopathic interstitial, pneumonia, and children with cystic fibrosis generally show Pearson correlation coefficient on distance walked (ranging from 0.90 to 0.98 overall) and the intra-class coefficients (ranging from 0.82 to 0.99) ([35],36,37). Previous data on the Incremental Shuttle Walk field test ([32]) show Pearson correlation coefficient of 0.98 on two trials one week apart in patients with chronic heart failure ([38]) and intraclass coefficients of 0.88 and 0.87 in clinically stable patients with COPD ([39]) and intermittent claudication ([40]). Thus the present coefficients approximating 0.90 compares well with those of the 6-Minute Walk Test and incremental Shuttle Walk field tests.

Validation of step testing against directly measured peak oxygen uptake has been reported with Pearson correlation coefficients exceeding 0.88 in healthy adult populations ([6], [7]) and well as in healthy elderly adults (≥65 years of age) using a self-determined stepping pace ([9]). There are however only few reports on test-retest reliability. In healthy older adults an intra-class correlation coefficient of 0.92 was computed from stepping heart rate measured on two tests performed 1–2 weeks apart ([41]) while a good inter-trial reliability was found for the prediction of peak V˙O₂, heart rate and rate of perceived exertion in firefighters ([42]). Thus, the coefficients approximating 0.90 computed in the present COPD population compare well to previous findings in healthy population.

In the present study, the main outcome was the exercise-related symptom of dyspnea. There is little data if at all, regarding the test-retest reproducibility of field tests on physiological outcomes such as ventilation or gas exchange parameters. Based on the measured indices of reliability, our results show that the field tests are good discriminator tools. No significant differences outcome measures have been observed from one week to the next for any of the imposed movement cadences and strong Pearson and intra-class correlation coefficients were found for all physiological responses including the BORG scores. These observations thus suggest that whatever the cadence selected for use, reproducible physiological responses and resulting exertional dyspnea will be recorded.

Agreement

Agreement concerns the absolute measurement error ([18]). For evaluative purposes in which one wants to distinguish clinically important changes from measurement error a small measurement error is required. Our results show a mean difference in Borg ratings of less than 0.5 units from trial 1 to trial 2 for both the Shuttle Walk and the Step Test with the majority of scores remaining within a 1-point unit of the Borg scale. This difference is small as it below the minimum clinically significant difference for this variable ([43]), Thus, it appears that the two measurement tools may be ascribed good evaluative properties with an appropriate discriminative property.

Walking vs Stepping tests: is one better than the other?

The need to consider the appropriateness of one exercise modality over another when attempting to detect changes in exertional dyspnea has recently been demonstrated. In a study comparing cycling to externally paced walking it was shown that while cycle endurance testing may be appropriate to detect changes following the administration of a bronchodilator ([44],45,46), its responsiveness to detect changes in dyspnea may be less than that of a walking-based exercise test ([47]).

In the present study, equally strong reproducibility was found for both externally paced walking and stepping. Both tests bear an advantage over cycle endurance testing in that they more closely match the nature of activities of daily living inducing exertional dyspnea in these patients. Thus, while the walking protocol may be better suited than cycling our results suggest that stepping may be equally appropriate to detect changes in exertional dyspnea in patients with COPD. On the basis of limits of agreements, a somewhat smaller difference was found for the step-test suggesting tighter control on exertional dyspnea. Similarly, from a practical perspective, the argument could be made that given the less space requirements of step testing; this may be better suited for use in the primary care setting.

CONCLUSIONS

Our study showed that both field tests using three-minute bouts of externally paced walking or stair stepping were highly reproducibility to monitor exercise-induced physiological responses and dyspnea in patients with stable moderate to severe COPD. These tests using short bouts of constant load exercise may be used to provide physicians with an accurate, reproducible, and well-tolerated evaluation of the level of exertional dyspnea encountered by their patients in the course of their daily activities. Thus within a period of 5 or 10 minutes, the clinical management team could eventually monitor the evolution of disease or assess the outcome of therapeutic interventions. Follow up studies are however required to determine responsiveness of these tests to therapeutic interventions.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

ACKNOWLEDGMENTS

The authors would like to thank Éric Nadreau, Marthe Bélanger, Marie-Josée Breton, Josée Picard, Brigitte Jean and Carmen Darauay.

Disclosure of funding: This work has been supported by an unrestricted grant from Boehringer Ingelheim and Pfizer.

François Maltais is a research scholar of the Fonds de la Recherche en Santé du Québec.