Comparison of laboratory- and field-based exercise tests for COPD: a systematic review

Abstract

Exercise tests are often used to evaluate the functional status of patients with COPD. However, to the best of our knowledge, a comprehensive systematic comparison of these tests has not been performed. We systematically reviewed studies reporting the repeatability and/or reproducibility of these tests, and studies comparing their sensitivity to therapeutic intervention. A systematic review identified primary manuscripts in English reporting relevant data on the following exercise tests: 6-minute walk test (6MWT) and 12-minute walk test, incremental and endurance shuttle walk tests (ISWT and ESWT, respectively), incremental and endurance cycle ergometer tests, and incremental and endurance treadmill tests. We identified 71 relevant studies. Good repeatability (for the 6MWT and ESWT) and reproducibility (for the 6MWT, 12-minute walk test, ISWT, ESWT, and incremental cycle ergometer test) were reported by most studies assessing these tests, providing patients were familiarized with them beforehand. The 6MWT, ISWT, and particularly the ESWT were reported to be sensitive to therapeutic intervention. Protocol variations (eg, track layout or supplemental oxygen use) affected performance significantly in several studies. This review shows that while the validity of several tests has been established, for others further study is required. Future work will assess the link between these tests, physiological mechanisms, and patient-reported measures.

Keywords:

• 6MWT
• 12MWT
• COPD
• walk test
• repeatability
• reproducibility
• shuttle walk test
• cycle ergometer test

Introduction

COPD is a leading cause of death worldwide, and the prevalence of the disease is projected to increase as the population ages and as exposure to risk factors, such as smoking, continues.^1–3 COPD is characterized by symptoms of breathlessness and reduced exercise capacity.^4,5 Decrements in exercise capacity can result in reduced ability to perform activities of daily living, and the resultant inactivity and sedentary lifestyle can further exacerbate exercise impairment (the COPD “vicious circle”).⁶

In clinical practice, spirometry is recommended by the Global Initiative for Chronic Obstructive Lung Disease (GOLD) for the diagnosis of COPD.⁵ However, the results of spirometry alone poorly predict disability and quality of life in patients with COPD⁷ and correlate only weakly with dyspnea, exercise capacity, and health status.^8–10 Recent guidelines on the diagnosis and treatment of COPD indicate that assessment of disease severity is improved by using additional functional criteria such as exercise capacity.^4,5,11 Quantification of the degree of functional impairment is therefore important for the assessment of response to treatment and as an outcome for clinical trials.

There are a number of laboratory- and field-based tests currently used for the assessment of exercise capacity, including the 6- and 12-minute walk tests (6MWT and 12MWT, respectively), the incremental and endurance shuttle walk tests (ISWT and ESWT, respectively), the incremental and endurance cycle ergometer tests (ICET and ECET, respectively), and the incremental and endurance treadmill tests (ITT and ETT, respectively). However, there is no consensus about which test is the most appropriate for use in patients with COPD. These tests have different primary outcomes (eg, endurance time, distance, oxygen consumption) that may reflect different physiological parameters. It is therefore difficult to compare results across studies, limiting interpretation of the published literature in this field. Furthermore, the relative merits of different tests have not been established.

The systematic review presented here therefore evaluated evidence of the “repeatability” (defined as consistency of results when multiple tests are conducted on the same day) and the “reproducibility” (consistency of results when tests are conducted on different days) of the tests. The review also assessed the relative properties of the eight commonly used exercise tests and their sensitivity to therapeutic intervention (such as rehabilitative, pharmacological, or surgical procedures). In addition, the effect of protocol variations within each test was assessed across studies. When possible, results were placed in the context of available minimal clinically important difference (MCID) values, which have thus far been ascertained for the 6MWT,¹² ISWT,¹³ ESWT,¹⁴ and ICET.¹² Investigation of these factors will be useful in guiding test selection in clinical practice and for outcome measures in clinical trials. As these tests are often also used as interventions, evaluation of exercise testing modalities in patients with COPD will also inform the clinical development of optimal exercise rehabilitation strategies.

Methods

Search strategy

Literature searches were conducted using Ovid^® (Ovid Technologies Inc., New York, NY, USA), incorporating Ovid Medline^® (US National Library of Medicine, Bethesda, MD, USA), for the period from 1948 to January 22, 2013, Ovid Embase™ (Elsevier Inc., Philadelphia, PA, USA) for 1974 to January 22, 2013, and The Cochrane Library (John Wiley and Sons Ltd, Hoboken, NJ, USA) for 1962 to January 22, 2013 (see Tables S1–S3). Search strings were constructed to identify studies reporting primary data on the outcomes of the following exercise tests in patients with COPD: the 6MWT, 12MWT, ISWT, ESWT, ICET, ECET, ITT, and ETT. The full search strings are presented in the “Supplementary materials” section.

Study selection

Study selection followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines for performing a systematic literature review.¹⁵ Review articles and studies not published in English were excluded using search-engine filters. Studies confounded by comorbidities (such as cancers, diabetes, and non-COPD respiratory-tract diseases) were excluded on review of title/abstract. The remaining studies were screened based on titles and abstracts, and full articles were reviewed when their relevance was unclear from the abstract. Screening was performed by a single author (GM) and records were initially reviewed by title/abstract; a full paper review was subsequently undertaken for publications that could not be excluded by title/abstract. Included records were verified by a second author (IF). A 30% random sample of excluded records was also reviewed by the second author (IF). Disagreements were settled by consultation with the remaining authors.

When reviewing abstracts or full papers, records were excluded if they were reviews, were not in the English language, studied patients with confounding comorbidities (eg, cancers or diabetes), did not use an exercise test as an outcome measure, or examined an intervention other than our interventions of interest (pulmonary rehabilitation, bronchodilator therapy, and lung-volume reduction surgery). Specific inclusion criteria included any definition of COPD (including emphysema- and bronchitis-specific studies); interventions were included only in our comparison of sensitivity and limited to pulmonary rehabilitation, bronchodilation, and lung-volume reduction surgery. Included test outcomes are outlined in the “Data abstraction” section. Following screening, studies were subsequently included for assessment if they reported data on:

• repeatability (studies reporting data from two or more performances of the same test[s] on the same day under the same conditions)

• reproducibility (studies reporting data from two or more performances of the same test[s] on different days under the same conditions)

• comparisons of sensitivity (studies reporting responses of two or more tests to the following therapeutic interventions: pulmonary rehabilitation, bronchodilator therapy, or lung-volume reduction surgery)

• protocol variations (studies reporting two or more performances of a test when protocol parameters have been modified).

Data abstraction

Data were primarily abstracted by a single author (GM) and reviewed by all co-authors. A randomly generated selection of 30% of all articles was reviewed by a second author (IF) for quality-control purposes.

The following outcomes of exercise tests were recorded: distance or stages achieved for the 6MWT, 12MWT, and ISWT; duration of exercise for the ESWT, ECET, and ETT; and the highest recorded volume of oxygen consumption (peak VO₂) and/or maximum workload (W_max) for the ICET and ITT. Articles merited inclusion in this review if they reported: outcomes of the specified tests when performed repeatedly under the same conditions, either on the same day (repeatability) or on different days (reproducibility); changes in response before and after therapeutic intervention (comparison of sensitivity); or effects of within-test variations in protocol (protocol variation).

Studies comparing the sensitivity of tests were also assessed for expression by the authors of preference for any specific test. When distances were reported in feet, values were converted to meters using standard conversion criteria stated by the International Bureau of Weights and Measures (0.3048 meters per foot). Within each publication, tests for which results are available are referred to as “test 1”, “test 2”, etc; occasions on which a test has been described by the authors, but results are not reported (such as for practice tests), are referred to as “familiarizations”.

Results

Overview of identified studies

The search methodology used to identify relevant articles is summarized in Figure 1. Of 1,781 unique articles screened, 71 were ultimately deemed eligible for inclusion in this review.

Figure 1 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) diagram detailing the identification and inclusion process of the articles. Some studies are included in more than one analysis category; consequently, the aggregate number of studies in the repeatability, reproducibility, comparative, and protocol variation groups adds up to more than 71.

Studies assessing the repeatability and reproducibility of tests

Clinical practice is influenced by factors such as the repeatability and reproducibility of exercise tests in patients with COPD. These factors have been extensively assessed for the 6MWT and 12MWT; data are more limited for the ISWT, ESWT, ITT, and ETT (23 references for the 6/12MWT; 12 for the IWST, ESWT, ITT, and ETT; seven for the ICET and ECET; this made 37 references in total owing to overlap of these categories). Table 1 summarizes the results of studies assessing the repeatability and reproducibility of the 6MWT and 12MWT; Table 2 focuses on the ISWT, ESWT, ETT, and ITT; and Table 3 presents data on the ICET and ECET.

Table 1 Repeatability and reproducibility of the 6- and 12-minute walk tests

Study	Patients, n (males, n)	Disease severity (mean ± SD FEV1 [mean ± SD% predicted] or COPD grading, unless otherwise stated)	Comparison	Additional information
6-minute walk test
Repeatabilitya
Andersson et al^Citation16	47 (16)	1.20±0.49 (46.0±17.0)	First versus second test	Significant increase in distance from first to second test (Δ22 m [5.3%]) (ICC =0.94)
Bansal et al^Citation17	27 (15)	0.8±0.2 (38.1±14.3)	First versus second test	Significant increase in distance from first to second test, for linear (Δ14 m [3.5%]) and circuit (Δ12 m [3.0%]) track layouts
Eiser et al^Citation18	57 (30)	NR (35.0±12.0)	First versus second versus third test	Significant increase in distance from first to second test (Δ7 m [1.6%]); NSD from second to third test
Jenkins and Cecins^Citation19	245 (162)	1.06±0.5 (41.0±18.0)	First versus second test	Significant increase in distance from first to second test (Δ37 m [9.5%])
Stevens et al^Citation20	21 (9)	1.07±0.53 (NR)	First versus second versus third test	Hallway: significant increase in distance from first to second test (approximate Δ32.9 m [9.0%]); NSD in distance from second to third test Treadmill: significant increase in distance from first to second test (approximate Δ16.2 m [9.9%]); NSD in distance from second to third test
Vagaggini et al^Citation21	18 (15)	NR (48.0±14.0)	Familiarization then first versus second test	NSD in distance from first to second test after familiarization (Δ15 m [3.4%])
Reproducibilityb
Bansal et al^Citation17	27 (15)	0.8±0.2 (38.1±14.3)	First versus second test	NSD in distance from first to second day for straight (Δ12 m [2.9%]) and circular (25 m [6.0%]) track layouts
Behnke et al^Citation26	TG: 66 (51)	TG: 1.30±0.49 (41.9±13.9)	Familiarization then first versus mean of second and third tests	TG: significant increase in distance from first to mean of second and third test (Δ209.5 m [72.9%])
	CG: 22 (20)	CG: 1.45±0.51 (46.9±15.1)		CG: significant increase in distance from first to mean of second and third test (Δ57.5 m [18.9%])
Brooks et al^Citation27	18 (10)	1.0±0.3 (42.0±8.0)	First versus second test	NSD in distance from first to second test (NR)
Chatterjee et al^Citation28	88 (37)	Median ± IQR FEV1, L (mean ± SD%) 1.21±0.55 (52.0±19.4)	First versus second versus third test	NSD in distance from first to second test (Δ32 m [9.8%]); NSD in distance from second to third test (Δ4 m [1.1%])
Eiser et al^Citation18	57 (30)	NR (35.0±12.0)	First versus second versus third test	Significant decrease in mean distance from first to second day (Δ–9 m [–2.1%]); significant increase in mean distance from second to third day (Δ8 m [1.9%])
Kozu et al^Citation29	45 (38)	1.10±0.50 (45.0±12.0)	First versus second test	Significant increase in distance from first to second day (Δ13 m [4.3%])
Poulain et al^Citation30	10 (NR)	Moderate COPD	Familiarization then first versus second test	NSD in distance from first to second day after familiarization (Δ4.7 m [0.94%])
Rejeski et al^Citation31	209 (117)	1.57; 0.58 (57.1±17.0)	First versus second test	Significant increase in distance from first to second day (Δ22 m [4.6%]); significant correlation between tests (r=0.91)
Roomi et al^Citation32	15 (6)	NR (49.0±5.0)	First versus second test	NSD in distance from first to second day (Δ−1 m [0.0%])
Sciurba et al^Citation33	470 (287)	0.75±0.24 (26.3±7.4)	First versus second test	Significant increase in distance from first to second day (Δ20 m; baseline NR) (ICC =0.88)
Spencer et al^Citation34	44 (22)	NR (56.0±19.0)	First versus second test	Significant increase in distance from first to second day (Δ27 m [5.9%])
Troosters et al^Citation35	20 (NR)	1.36±0.46 (45.0±14.0)	First versus second test	NSD in distance from first to second day (Δ15 m [2.6%])
12-minute walk test
Repeatabilitya
O’Reilly et al^Citation22	10 (10)	Chronic bronchitis or radiological emphysema	Familiarization then first versus second test	NSD in distance from first to second test after familiarization (mean variation 3.1%; distance NR)
Reproducibilityb
Arnardóttir et al^Citation36	EIH group: 19 (8)	EIH group: 0.9±0.1 (34.6±2.4)	First versus second versus third test	EIH group: NSD in distance from first to second to third day
	Non-EIH group: 38 (19)	Non-EIH group: 1.0±0.1 (38.9±1.8)		Non-EIH group: significant increase in distance from first to second day (Δ92 m [12%]; baseline NR); significant increase in distance from second to third day (Δ28 m [4%]; baseline NR)
Beaumont et al^Citation37	12 (10)	1.03±0.27 (NR)	First versus second versus third test	Significant increase in distance from first to second day (Δ46 m [6.5%]); NSD in distance from second to third day (Δ37 m [4.9%])
Berger and Smith^Citation38	10 (10)	Moderate/severe COPD	Familiarization then first versus second test	NSD in distance from first to second test after familiarization (Δ13 m [2.2% using values from Beaumont et al^Citation37])
McGavin et al^Citation39	35 (35)	Chronic bronchitis	First versus second versus third test	Significant increase in distance from first to second day (Δ64 m [7.2%]); NSD in distance from second to third day (Δ41 m [4.3%])
Mungall and Hainsworth^Citation40	13 (13)	1.5±0.4 (NR); chronic bronchitis or radiological emphysema	First versus second versus third test (with three further tests NR in detail)	Significant increase in distance in third test compared with first and second; NSD distance on three subsequent tests (distances NR)
O’Reilly et al^Citation22	10 (10)	Chronic bronchitis or radiological emphysema	Familiarization then first versus second test	NSD in distance from first to second day after familiarization (Δ38 m [4.8%])
Swinburn et al^Citation41	17 (6)	0.77±0.30 (NR)	First versus second versus third versus fourth test	Progressive, significant increases in distance from first to fourth day (16%; baseline NR)

Notes:

a Repeatability = similarity of test results when performed on the same day;

b reproducibility = similarity of test results when performed on different days.

Abbreviations: CG, control group; CI, confidence interval; EIH, exercise-induced hypoxia; FEV₁, forced expiratory volume (L) in 1 second; ICC, intra-class correlation coefficient; IQR, interquartile range; NR, not reported; NSD, no significant difference; SD, standard deviation; TG, training group; m, meters; min, minutes.

Table 2 Repeatability and reproducibility of incremental shuttle walk test, endurance shuttle walk test, incremental treadmill test, and endurance treadmill test

Study	Patients, n (males, n)	Disease severity (mean ± SD FEV1 [mean ± SD% predicted] or COPD grading, unless otherwise stated)	Comparison	Additional information
Incremental shuttle walk test
Repeatabilitya
Eiser et al^Citation18	57 (30)	NR (35.00±12.00)	First versus second versus third test	Significant increase in distance from first to second test (Δ13 m; 4.7%]); NSD in distance from second to third test (Δ6 m [2.1%])
McKeough et al^Citation23	53 (34) 31 (NR)	NR (55.00±19.00) NR (51.00±16.00)	First versus second test Third versus fourth test	Significant increase in distance from first to second test (Δ20 m [6.3%])c Significant increase in distance from third to fourth test (Δ18 m [4.8%])d
Vagaggini et al^Citation21	18 (15)	NR (48.00±14.00)	Familiarization then first versus second test	Significant increase in distance from first to second test, after familiarization (Δ40 m [14.6%])e
Reproducibilityb
Arnardóttir et al^Citation44	93 (26)	Moderate/severe COPD	First versus second test	NSD in distance from first to second day (Δ9 m [2.9%])
Campo et al^Citation42	30 (18)	Mild/moderate/severe/very severe COPD	Familiarization then first versus second test	NSD in distance from first to second day (ICC =0.88) (distances NR)
Eiser et al^Citation18	57 (30)	NR (35.00±12.00)	First versus second versus third test	Significant increase in distance from first to second day (Δ13 m; 4.6%]); NSD in distance from second to third day (Δ−1 m [0.3%])
Perrault et al^Citation43	43 (36)	1.40±0.50 (49.00±16.00)	Familiarization then first versus second test	NSD in VO2, HR, VE, VT from first to second test at each of the four cadences (ICC >0.93); distance NR
Singh et al^Citation45	35 (25)	Group A: 0.50 (0.36–1.45) Group B: 1.10 (0.60–2.10)	First versus second versus third test	Downgraded protocol: NSD in distance on first, second, and third day Modified protocol: significant increase in distance from first to second day (Δ31 m [9.0%]); NSD in distance from second to third day
Endurance shuttle walk test
Repeatabilitya
Revill et al^Citation24	44 (33)	0.94±0.40 (37.00±13.00)	First versus second test	NSD in duration from first to second test (Δ12 s [6.2%])
McKeough et al^Citation23	53 (34) 31 (NR)	NR (55.00±19.00) NR (51.00±16.00)	First versus second test Third versus fourth test	NSD in duration from first to second test (Δ−2 s [5.8%]) NSD in duration from third to fourth test (Δ44 s [8.7%])
Reproducibilityb
Revill et al^Citation47	44 (22)	Group A: 1.01±0.36 (35.00±4.00) Group B: 0.79±0.21 (34.00±4.00) Group C: 0.80±0.18 (35.00±8.00)	First versus second versus third test	Significant increase in duration from first to second day (Δ59 s; 23.5%]); NSD in duration from second to third day (Δ15 s [4.8%])
Revill et al^Citation46	23 (13)	0.81±0.27 (33.00±12.00)	Familiarization then first versus second test	NSD in duration from first to second test after familiarization (tests performed with supplemental oxygen)
Endurance treadmill test
Reproducibilityb
Cooper et al^Citation52	470 (NR)	Moderate/severe/very severe COPD	First versus second test	NSD in duration from first to second test (Δ24 s [7.6%]) (ICC =0.85)
Incremental treadmill test
Reproducibility
Mathur et al^Citation50	8 (6)	0.69±0.16 (NR)	First versus second test	Numerical increase in peak VO2 from first to second test (Δ1.0 mL/min/kg [8.3%]; statistical test NR)

Notes:

a Repeatability = similarity of test results when performed on the same day

b reproducibility = similarity of test results when performed on different days

c P<0.001

d P≤0.05

e P≤0.01

Abbreviations: FEV₁, forced expiratory volume (L) in 1 second; HR, heart rate; ICC, intra-class correlation coefficient; NR, not reported; NSD, no significant difference; SD, standard deviation; V_E, expired volume; VO₂, oxygen consumption; V_T, tidal volume; m, meters; min, minutes; s, time in seconds.

Table 3 Repeatability and reproducibility of cycling tests

Study	Patients, n (males, n)	Disease severity (mean ± SD FEV1 [mean ± SD% predicted] or COPD grading, unless otherwise stated)	Comparison	Additional information
Incremental cycle ergometer test
Repeatabilitya
Brown et al^Citation25	11 (NR)	1.51±0.59 (NR)	First versus second test	NSD in peak VO2 from first to second test (Δ53 mL/min [4.0%]); NSD in Wmax from first to second test (Δ0 W [0%])
Reproducibilityb
Brown et al^Citation25	11 (NR)	1.51±0.59 (NR)	First versus third test	NSD in peak VO2 from first to third test (Δ93 mL/min [7.1%]); NSD in Wmax from first to third test (Δ0 W [0%])
Covey et al^Citation48	56 (40)	NR (49.00±16.00)	First versus second test	NSD in peak VO2 from first to second day (Δ0.011 L/min [0.9%]); NSD in Wmax from first to second day (Δ1 W [1.5%])
Cox et al^Citation49	11 (8)	Individually listed in paper	First versus second test	NSD in peak VO2 from first to second day (Δ0.04 L/min [2.0%]); NSD in Wmax from first to second day (Δ1 W [0.6%])
Mathur et al^Citation50	8 (6)	0.69±0.16 (NR)	First versus second test	Numerical increase in peak VO2 from first to second test (Δ0.8 mL/min/kg [6.8%]; statistical test NR)
Poulain et al^Citation30	10 (NR)	Moderate COPD	Familiarization then first versus second test	NSD in %-predicted peak VO2 from first to second day (Δ2.5%-predicted VO2 [3.9%]); NSD in Wmax from first to second day (Δ1 W [1.3%])
Swinburn et al^Citation41	17 (6)	0.77±0.30 (NR)	First versus second versus third versus fourth test	Progressive and significant increases in duration from first to fourth day (29%; baseline values NR) (VO2/Wmax NR)
Endurance cycle ergometer test
Reproducibilityb
van’t Hul et al^Citation51	60 (46)	Moderate/severe COPD	First versus second test	NSD in duration from first to second test (Δ−12 s [approximately −4.4%]) (ICC =0.85)

Notes:

a Repeatability = similarity of test results when performed on the same day;

b reproducibility = similarity of test results when performed on different days.

Abbreviations: FEV₁, forced expiratory volume (L) in 1 second; ICC, intra-class correlation coefficient; NR, not reported; NSD, no significant difference; SD, standard deviation; VO₂, oxygen consumption; W_max, maximum workload; W, watts; min, minutes; s, time in seconds.

Six studies presented repeatability data for the 6MWT (Table 1).^16–21 Of these, five reported a significant increase in 6MWT distance from the first to the second test;^16–20 the remaining study found no differences between results, though patients had been previously familiarized with the tests.²¹ The three studies clearly reporting the results of three 6MWTs performed on the same day found that there was no significant difference between the second and third tests.^18,20,21 The only study that presented intra-class correlation coefficients (ICCs) for repeated 6MWTs on the same day reported excellent repeatability (ICC =0.94), but also observed that the second test was significantly higher.¹⁶ For the 12MWT, the distance achieved was reported to be repeatable in the only study in which patients were retested on the same day.²² Equivocal results were reported for the repeatability of the ISWT by three studies;^18,21,23 one found that the distance was repeatable after familiarization,¹⁸ but the other two reported poor repeatability even after familiarization (Table 2).^21,23 For the ESWT, exercise duration was reported to be repeatable in the two studies in which patients were retested on the same day.^23,24 One study reported that peak VO₂ and W_max were repeatable for the ICET (Table 3).²⁵ No repeatability data were found for the ECET, ITT, or ETT.

Reproducibility of the 6MWT was assessed in 12 studies,^{17,18,26–35} six of which reported that distances achieved in the 6MWT demonstrated good reproducibility between the first and second tests (Table 1).^{17,27,28,32,35} Of two studies presenting reproducibility results from three 6MWTs,^18,28 only one reported reproducibility between the second and third tests.²⁸ Two further studies reported results of tests after familiarization;^26,30 only one found 6MWT results to be reproducible.³⁰ The only study presenting ICC data between the first and second 6MWT performance showed high reproducibility (ICC =0.88), but also that there was a significant increase in distance in the second 6MWT.³³ Reproducibility of the 12MWT was assessed in seven studies.^22,36–41 Five of these presented the results of three or more tests, and reported that the 12MWT distance increased significantly from the first to the second test;^{36,37,39–41} one of these studies reported that in a subset of patients who readily experienced exercise-induced hypoxia, the 12MWT distance did not significantly change from the first to the second to the third test.³⁶ Two additional studies, in which there had been prior familiarization with the test, reported no significant change in 12MWT distance between subsequent first and second tests.^22,38 For the ISWT, five studies presented data assessing reproducibility: from the first to the second test with prior familiarization,^42,43 from the first to the second test without prior familiarization,^44,45 and from the second to the third test without prior familiarization (Table 2).¹⁸ Two further studies evaluating the ESWT reported reproducibility (either from the first to the second test with prior familiarization,⁴⁶ or from the second to the third test without familiarization).⁴⁷ Good reproducibility of the ICET from the first to the second test was reported by four studies both with³⁰ and without prior familiarization (Table 3).^25,48,49 Two further studies reported an increase in ICET duration from the first to the second test with no familiarization,^41,50 with one of these studies reporting progressive increases in ICET duration from the first through to the fourth test.⁴¹ In the only study reporting data for the ECET, duration was found to have excellent reproducibility (ICC =0.85).⁵¹ Reproducibility was also found to be excellent in the only study reporting such data for the ETT (no significant increases from the first to the second test, ICC =0.85),⁵² but less so for the ITT (increased peak VO₂ from the first to the second test, statistical test not reported).⁵⁰

Several studies were identified that compared the repeatability and/or reproducibility of two or more exercise tests. One study observed that repeatability for the 6MWT and the ISWT was comparable, but that the ISWT was more reproducible.¹⁸ However, another study showed that the ISWT was more repeatable than the 6MWT.²¹ One study reported that both the 6MWT and the ICET were reproducible.³⁰ The ESWT was reported to be more repeatable than the ISWT in one study, when measured in two sessions before and two sessions after pulmonary rehabilitation.²³ In another study, both 12MWT distance and ICET performance were found to increase significantly and progressively over four tests; the ICET was found to have no obvious advantages over the 12MWT when assessing exercise performance.⁴¹ The final study to report reproducibility of more than one test reported that peak VO₂ increased from test 1 to test 2 in both the ICET and the ITT; however, the authors did not report the statistical tests used.⁵⁰ Three studies^18,28,30 were found that compared the reproducibility of two or more exercise tests. Of these, two reported that the 6MWT was found to have similar reproducibility to the ISWT¹⁸ and the ICET.³⁰

Studies comparing responses to interventions among exercise tests

In total, 23 studies were identified that compared responses of two or more exercise tests after one of the following interventions: pulmonary rehabilitation (16 studies),^{23,47,53–66} administration of bronchodilators (six studies),^{14,18,67–70} and lung-volume reduction surgery (one study)⁷¹ (Table 4). Of the 16 studies that assessed pulmonary rehabilitation in patients with COPD, the most commonly assessed test was the 6MWT, which was reported by eleven studies.^{53–55,57–59,61,63–66}

Table 4 Between-test comparisons of sensitivity to interventions

Study	N	Result	Additional information
Pulmonary rehabilitation
6-minute walk test
Borghi-Silva et al^Citation54	8 (Car) 8 (Pla)	Significant increase in distance: Δ87 m (19.8%) (>MCID) Significant increase in distance: Δ34 m (7.3%) (>MCID)	Training with concomitant carnitine versus placebo; ITT conducted but peak VO2 and Wmax NR
Borghi-Silva et al^Citation53	20	Significant increase in distance: Δ105 m (27.8%) (>MCID)	Training significantly improved peak VO2 compared with control group
Carrieri-Kohlman et al^Citation57	51	Significant increase in distance: Δ52.1 m (11.7%) and Δ50.3 m (12.7%) (>MCID)	Improvements were similar in both tests whether patients were given a monitored or coached pulmonary rehabilitation program
Cooper^Citation55	7 (GPT) 9 (IMT)	Significant increase in distance: Δ32 m (6.9%) (>MCID) IMT, Δ23 m (4.8%)	Patients in IMT group showed significant improvement in pulmonary function test; increase in GPT group was not significant
Cortopassi et al^Citation58	71	NSD in distance	ITT and ETT both responded significantly to PR, but 6MWT did not
Eaton et al^Citation59	20	Significant increase in distance: Δ47 m (17%) (>MCID)	Standardized mean differences after PR: 6MWT, 0.32 versus ESWT, 0.54 The authors note: “The ESWT has potential advantages in that it may be more responsive than the 6MWT” Values taken from text (do not match Eaton et al)^Citation59
Ngaage et al^Citation61	14	Significant increase in distance: Δ68.3 m (36.3%) (>MCID)	ISWT values taken from Ngage et al;^Citation61 changes in both 6MWT and ISWT after PR were clinically significant
Ong et al^Citation63	37	Significant increase in distance: Δ36 m (9.0%) (>MCID)	6MWT did not correlate with ICET or ECET responses to PR; moderate correlation between ICET Wmax and ECET duration (r=0.406; P=0.013)
Ries et al^Citation64	1,218	Significant increase in distance: Δ23.0 m (6.6%) (non-MCID)	Patients who had never previously received PR responded more favorably than those who had (Δ6MW distance: 31.0 versus 18.5 m; ICET Wmax: 4.3 versus 2.4 W)
Van Helvoort et al^Citation65	18	Significant increase in distance: Δ63 m (14.8%) (>MCID)	6MWT, ICET, and ECET had greater validity in assessing responses to PR than cardiorespiratory measures
Van Ranst et al^Citation66	389	Significant increase in distance: Δ52 m (14.2%) (>MCID)	Clinically relevant response after PR for both tests; greater sensitivity in the ECET than in the 6MWT (not assessed statistically)
12-minute walk test
Arnardóttir et al^Citation56	28 (IT) 32 (CT)	Significant increase in distance: Δ75 m (9.0%) Significant increase in distance: Δ94 m (10.8%)	Similar improvements in 12MWT distance and Wmax whether training was IT or CT; peak VO2 was higher in CT, but VO2 values at isotime were lower in IT
Incremental shuttle walk test
Greening et al^Citation60	601	Significant increase in distance: Δ61 m (28.9%) (>MCID)	Clinically relevant response after PR for both tests; greater sensitivity in the ESWT than in the ISWT (not assessed statistically)
McKeough et al^Citation23	31	Significant increase in distance: Δ46 m (15%) (non-MCID)	Results are calculated from the better of two pre-PR test results versus the better of two post-PR test results Values taken from Table 2
Ngaage et al^Citation61	14	Significant increase in distance: Δ57.1 m (52.9%) (>MCID)	ISWT values taken from Ngage et al;^Citation61 changes in both 6MWT and ISWT after PR were clinically significant
O’Farrell et al^Citation62	85	Significant increase in distance: Δ52.9 m (33.3%) (>MCID)	ESWT more responsive than ISWT to determine improvements in exercise capacity following PR
Revill et al^Citation47	44	Significant increase in distance: Δ36 m (20.2%) (non-MCID)	Clinically relevant response after PR in ESWT but not ISWT; greater sensitivity in ESWT than in ISWT (effect size 2.90 versus 0.41)
Endurance shuttle walk test
Eaton et al^Citation59	20	Significant increase in distance: Δ302 m (92%) (>MCID); significant increase in duration: Δ270 s (88%) (>MCID)	Standardized mean differences after PR: 6MWT, 0.32 versus ESWT, 0.54 The authors note: “The ESWT has potential advantages in that it may be more responsive than the 6MWT” Values taken from text (do not match Eaton et al)^Citation59
Greening et al^Citation60	601	Significant increase in duration: Δ408 s (205.4%) (>MCID)	Clinically relevant response after PR for both tests; greater sensitivity in the ESWT than in the ISWT (not assessed statistically)
McKeough et al^Citation23	31	Significant increase in duration: Δ182 s (58%)	Results are calculated from the better of two pre-PR test results versus the better of two post-PR test results Values taken from McKeough et al^Citation23
O’Farrell et al^Citation62	85	Significant increase in distance: Δ271.0 m (106.7%) (>MCID); significant increase in duration: Δ293.1 s (no percentage change or baseline value provided) (>MCID)	ESWT more responsive than ISWT to determine improvements in exercise capacity following PR
Revill et al^Citation47	44	Significant increase in distance: Δ334 m (140%) (>MCID); significant increase in duration: Δ404 s (150%) (>MCID)	Clinically relevant response after PR in ESWT but not ISWT; greater sensitivity in ESWT than in ISWT (effect size 2.90 versus 0.41)
Incremental cycle ergometry test
Arnardóttir et al^Citation56	28 (IT) 32 (CT)	Significant increase in peak VO2: Δ53 mL/min (5.4%); significant increase in Wmax : Δ11 W (18.0%) (>MCID) Significant increase in peak VO2 : Δ146 mL/min (15.0%); significant increase in Wmax: Δ11 W (17.2%) (>MCID)	Similar improvements in 12MWT distance and Wmax whether training was IT or CT; peak VO2 was higher in CT, but VO2 values at isotime were lower in IT
Cooper^Citation55	7 (GPT) 9 (IMT)	NSD in peak VO2 or Wmax NSD in peak VO2 or Wmax	Patients in IMT group showed significant improvement in pulmonary function test, increase in GPT group was not significant
Ong et al^Citation63	37	Significant increase in peak VO2: Δ172 mL/min (20.0%); significant increase in Wmax: 6 W (12.0%) (>MCID)	6MWT did not correlate with ICET or ECET responses to PR; moderate correlation between ICET Wmax and ECET duration (r=0.406; P=0.013)
Ries et al^Citation64	1,218	Significant increase in Wmax : Δ3.1 W (8.6%) (non-MCID)	Patients who had never previously received PR responded more favorably than those who had (Δ6MWT distance: 31.0 versus 18.5 m; ICET Wmax: 4.3 versus 2.4 W)
Van Helvoort et al^Citation65	18	NSD in Wmax and peak VO2	6MWT, ICET, and ECET had greater validity in assessing responses to PR than cardiorespiratory measures
Endurance cycle ergometry test
Ong et al^Citation63	37	Significant increase in duration: Δ322 s (73.5%)	6MWT did not correlate with ICET or ECET responses to PR; moderate correlation between ICET Wmax and ECET duration (r=0.406; P=0.013)
Van Helvoort et al^Citation65	18	Significant increase in duration: Δ6.5 min (166%)	6MWT, ICET, and ECET had greater validity in assessing responses to PR than cardiorespiratory measures
Van Ranst et al^Citation66	389	Significant increase in duration: Δ241 s (84.6%)	Clinically relevant response after PR for both tests; greater sensitivity in the ECET than in the 6MWT (not assessed statistically)
Incremental treadmill test
Borghi-Silva et al^Citation54	8 (Car) 8 (Pla)	Peak VO2 and Wmax NR Peak VO2 and Wmax NR	Training with concomitant carnitine versus placebo; ITT conducted but peak VO2 and Wmax NR
Borghi-Silva et al^Citation53	20	Significant increase in peak VO2 : Δ2 mL/min/kg (14.3%)	Training significantly improved peak VO2 compared with control group
Carrieri-Kohlman et al^Citation57	51	NSD in peak VO2 in either group; significant increase in maximum work (exercise stage): Δ1.6 levels (34.0%) and Δ1.5 levels (34.1%)	Improvements were similar in both tests whether patients were given a monitored or coached pulmonary rehabilitation program
Cortopassi et al^Citation58	71	Significant increase in duration: Δ184 s (27.4%); significant increase in distance: Δ190 m (22.1%); VO2 and Wmax NR	ITT and ETT both responded significantly to PR, but 6MWT did not
Endurance treadmill test
Cortopassi et al^Citation58	71	Significant increase in duration: Δ534 s (106.0%); VO2 and Wmax NR	ITT and ETT both responded significantly to PR, but 6MWT did not
Bronchodilator therapy
6-minute walk test
Cazzola et al^Citation67	22	Significant increase in distance: Δ53.6 m (baseline values NR) (>MCID)	“6-MWT seems to be a more appropriate instrument than 12-MWT for assessing the exercise response to a bronchodilator”
Eiser et al^Citation18	54 (6MWT)	Significant increase in distance: Δ37 m (8.7%) (>MCID)	Sensitivity index: 6MWT, 0.84; ISWT, 0.76; distance calculated by number of shuttles performed (1 shuttle =10 m) “There is no important difference in either the reproducibility or sensitivity of self- paced or externally paced walking tests”
Oga et al^Citation68	38	Significant increase in distance: Δ6 m (1.2%) (non-MCID)	Percentage improvement: 6MWT, 1%; ICET, 3%; ECET, 19% “Among the frequently used post-PR exercise tests, the most responsive index, as measured by the percentage change from baseline, is the endurance time. The correlation between the post-PR changes in these exercise indices is poor”
Pepin et al^Citation14	14	NSD in distance	“The endurance shuttle walk is more responsive to the acute effects of bronchodilation than the 6MWT”
12-minute walk test
Cazzola et al^Citation67	22	Significant increase in distance: Δ59.9 m (baseline values NR) (>MCID)	“6-MWT seems to be a more appropriate instrument than 12-MWT for assessing the exercise response to a bronchodilator”
Incremental shuttle walk test
Eiser et al^Citation18	50 (ISWT)	Significant increase in distance: Δ30 m (11.1%) (non-MCID)	Sensitivity index: 6MWT, 0.84; ISWT, 0.76; distance calculated by number of shuttles performed (1 shuttle =10 m) “There is no important difference in either the reproducibility or sensitivity of self- paced or externally paced walking tests”
Endurance shuttle walk test
Pepin et al^Citation14	14	Significant increase in distance: Δ144 m (40.0%) (>MCID); duration NR	“The endurance shuttle walk is more responsive to the acute effects of bronchodilation than the 6MWT”
Pepin et al^Citation69	17	Significant increase in duration: 164 s (>MCID)	“Standardized response mean larger for walking than cycling (0.93 and 0.20, respectively)”; % change and baseline NR
Incremental cycle ergometry test
Oga et al^Citation68	38	NSD in peak VO2 ; significant increase in Wmax : 3 W (3.4%)	Percentage improvement: 6MWT, 1%; ICET, 3%; ECET, 19% “Among the frequently used post-PR exercise tests, the most responsive index, as measured by the percentage change from baseline, is the endurance time. The correlation between the post-PR changes in these exercise indices is poor”
Endurance cycle ergometry test
Oga et al^Citation68	38	Significant increase in duration: 34 s (18.0%)	Percentage improvement: 6MWT, 1%; ICET, 3%; ECET, 19% “Among the frequently used post-PR exercise tests, the most responsive index, as measured by the percentage change from baseline, is the endurance time. The correlation between the post-PR changes in these exercise indices is poor”
Pepin et al^Citation69	17	NSD in duration	“Standardized response mean larger for walking than cycling (0.93 and 0.20, respectively)”; % change and baseline NR
Zhang et al^Citation70	20	Significant increase in duration: Δ157 s (47.1%)	Values taken from Zhang et al^Citation70 (percentage estimates in text are different); greater percentage change in ECET than in TT
Endurance treadmill test
Zhang et al^Citation70	20	Significant increase in duration: Δ110 s (30.4%); Wmax NR	Values taken from Table 4 (percentage estimates in text are different); greater percentage change in ECET than in TT
Lung-volume reduction surgery
6-minute walk test
Lederer et al^Citation71	23	Significant increase in distance: Δ38 m (9.1%) (>MCID)	Measurements taken 1 year after surgery; clinically relevant change in 6MWT; greater percentage change in ICET than in 6MWT
Incremental cycle ergometry test
Lederer et al^Citation71	23	NSD in peak VO2; significant increase in Wmax: 6 W (15.4%)	Measurements taken 1 year after surgery; clinically relevant change in 6MWT; greater percentage change in ICET than in 6MWT

Abbreviations: 6MWT, 6-minute walk test; 12MWT, 12-minute walk test; Car, patients receiving carnitine supplements; CT, continuous training; ECET, endurance cycle ergometer test; ESWT, endurance shuttle walk test; ETT, endurance treadmill test; GPT, general physical training; ICET, incremental cycle ergometer test; IMT, inspiratory muscle training; ISWT, incremental shuttle walk test; IT, interval training; ITT, incremental treadmill test; MCID, minimal clinically important difference; NR, not reported; NSD, no significant difference; pla, patients receiving placebo; VO₂, oxygen consumption; PR, pulmonary rehabilitation; TT, treadmill test; VAS, visual analog scale; V_E, ventilation; W_max, maximum workload; W, watt.

Two studies compared the response to the 6MWT and ITT after pulmonary rehabilitation; both reported significant increases in 6MWT distance and ITT performance (peak VO₂⁵³ and work-level completed).⁵⁷ However, the latter study did not observe a significant response in peak VO₂ during the ITT after pulmonary rehabilitation.⁵⁷ One further study assessed the 6MWT and ITT during nutritional supplementation and placebo, and reported that the 6MWT distance was sensitive to pulmonary rehabilitation (>MCID); but these authors did not present peak VO₂ or W_max data for the ITT.⁵⁴ Another study assessed the 6MWT, ETT, and ITT, and found that both the ETT and the ITT were sensitive to pulmonary rehabilitation, whereas the 6MWT was not (again, the authors did not report peak VO₂ or W_max data for the ITT).⁵⁸ Several further studies reported equivocal findings when comparing the 6MWT with the ICET^55,63–65 after pulmonary rehabilitation. All three studies comparing the ECET with the 6MWT found the ECET to be more responsive to pulmonary rehabilitation.^63,65,66 One study assessing responses to pulmonary rehabilitation reported similar sensitivities for the 6MWT and the ISWT, with both giving responses that exceeded the MCID.⁶¹ All four studies assessing the sensitivity of the ISWT and the ESWT to pulmonary rehabilitation reported a significant improvement in performance for both tests;^23,47,60,62 however, in all four studies the response of the ESWT was greater and in two the ISWT response did not reach the MCID.^23,47 An additional study suggested that although both tests showed a significant response that was above the MCID, the ESWT was more responsive to pulmonary rehabilitation than the 6MWT.⁵⁹ Equivocal sensitivity was observed in response to pulmonary rehabilitation when using the 12MWT and the ICET.⁵⁶

Of the six studies comparing the responses of two or more exercise tests to bronchodilator therapy,^{14,18,67–70} one reported the 6MWT to be more responsive to pharmacological intervention than the 12MWT⁶⁷ and one reported the 6MWT to be more responsive to pharmacological intervention than the ISWT.¹⁸ The standardized (percentage) increase in response to the ECET was higher than that of either the 6MWT or the ICET in one study,⁶⁸ and higher than that of the ETT in another.⁷⁰ One study reported the response of the ESWT to bronchodilation to be greater than that of the 6MWT¹⁴ and one study reported the response of the ESWT to bronchodilation to be greater than that of the ECET.⁶⁹

Finally, one study assessed exercise test performance 1 year after lung-volume reduction surgery, and reported a 9.1% increase in 6MWT distance and a 15.4% increase in ICET W_max, but noted that ICET peak VO₂ did not increase significantly.⁷¹

Minimal clinically important differences in responses to interventions

MCIDs have been thus far ascertained for the 6MWT (26 meters),¹² the ISWT (48 meters),¹³ the ESWT (45 seconds–85 seconds or 60 meters–115 meters [MCID calculated after bronchodilatory intervention]),⁷² and the ICET (4 watts).¹² Of the eleven studies assessing the 6MWT in response to pulmonary rehabilitation, nine reported an increase in excess of the recognized MCID,^{53–55,57,59,61,63,65,66} with another reporting a significant increase in distance of less than the MCID⁶⁴ (the remaining study reported no significant change in 6MWT distance after pulmonary rehabilitation⁵⁸); these increases ranged in magnitude from 4.8% to 36.3%. Five studies reported a significant response of the ISWT to pulmonary rehabilitation ranging from 15.0% to 52.9%,^{23,47,60–62} with three finding that the distance observed reached the MCID.^60–62 All five studies assessing the sensitivity of the ESWT to pulmonary rehabilitation reported that distance and/or duration increased in excess of the MCID (increases ranged from 88.0% to 205.4% when expressed as time [seconds];^{23,47,59,60,62} and from 92.0% to 140.0% when expressed as distance [meters]).^47,59,62 Of five studies assessing the ICET before and after pulmonary rehabilitation,^{55,56,63–65} two reported W_max responses in excess of the MCID,^56,63 and another reported significant changes in W_max that did not reach the MCID⁶⁴ (the increased ICET performance observed across these studies ranged from 5.4% to 20.0% for peak VO₂ and 8.6% to 18.0% for W_max). The two remaining studies did not observe a significant change in ICET peak VO₂ or W_max after pulmonary rehabilitation.^55,65 All three studies reporting the response of the ECET to pulmonary rehabilitation observed significant increases in duration (ranging from 73.5% to 166.0%).^63,65,66 Four studies assessed the ITT before and after pulmonary rehabilitation; two did not present data for either peak VO₂ or W_max.^54,58 Of the remaining two studies, one reported a significant increase in peak VO₂ of 14.3%,⁵³ while the other found no significant increase in peak VO₂ after pulmonary rehabilitation.⁵⁸ The only study assessing the ETT before and after pulmonary rehabilitation reported significant increases in duration (27.4%) and distance (22.1%).⁵⁸

Of the four studies assessing the 6MWT before and after bronchodilator therapy, two reported improvements in distance that exceeded the MCID (one found an increase in 6MWT distance of 8.7%;¹⁸ the other reported a higher absolute distance increase [53.6 meters], but did not present baseline values; therefore the percentage increase cannot be calculated⁶⁷). The third study reported a small (1.2%) but significant increase in 6MWT distance,⁶⁸ while the remaining study found no significant difference in 6MWT distance.¹⁴ One study found a significant increase in 12MWT distance after bronchodilation in excess of the MCID (59.9 meters), but did not provide baseline values.⁶⁷ One study reported a significant increase in ISWT performance after bronchodilators (30 meters) that did not reach the MCID.¹⁸ Two studies reported significant increases in ESWT performance above the MCID, of 144 meters¹⁴ and 164 seconds,⁶⁹ respectively. One study reported a small but significant improvement in ICET W_max of 3.4% after bronchodilator therapy, but noted that peak VO₂ did not increase significantly.⁶⁸ Three studies assessed ECET performance after bronchodilators; two found significant increases in duration of 18.0%⁶⁸ and 47.1%,⁷⁰ with the remaining study reporting no significant improvement in ECET duration.⁶⁹

In the only study assessing lung-volume reduction surgery, a 6MWT improvement in excess of the MCID was seen; however, ICET improvements (whether peak VO₂ or W_max) did not reach MCID.⁷¹

Studies assessing within-test variations in protocol

Eighteen studies were identified that assessed minor variations in protocol within a specific exercise test (Table 5).^{17,20,33,46,73–86} Variations (such as track environment or layout and the type of encouragement provided by the investigators to the patient) affected test outcomes (and consequently, their repeatability and reproducibility).

Table 5 Protocol variations reported to affect performance of exercise capacity test

Test	Variation	Study
6MWT	Track layout	Sciurba et al^Citation33 Casas et al^Citation76 Bansal et al^Citation17
	Supplemental oxygen provided	Revill et al^Citation46 Borghi-Silva et al^Citation75 Davidson et al^Citation77 Ozalevli et al^Citation83
	Corridor versus treadmill	Stevens et al^Citation20
	Indoor versus outdoor environment	de Almeida et al^Citation78
	Wheeled walker support provided	Honeyman et al^Citation80
	Verbal encouragement provided	Crisafulli et al^Citation73
12MWT	Corridor versus treadmill	Swerts et al^Citation86
ISWT	Verbal encouragement	Rosa et al^Citation84
	Supplemental oxygen provided	Sandland et al^Citation85
ESWT	Supplemental oxygen provided	Sandland et al^Citation85
ICET	Incremental workload gradient used	Benzo et al^Citation74 Miyahara et al^Citation82
ECET	Supplemental oxygen provided	Dean et al^Citation79
ITT	Incremental workload gradient used	Hsia et al^Citation81

Abbreviations: 6MWT, 6-minute walk test; 12MWT, 12-minute walk test; ECET, endurance cycle ergometer test; ESWT, endurance shuttle walk test; ICET, incremental cycle ergometer test; ISWT, incremental shuttle walk test, ITT, incremental treadmill test.

Discussion

A number of laboratory- and field-based exercise tests are used to assess the degree of functional impairment in patients with COPD. However, the choice of which test to use in clinical trials historically seems to have been made on a practical basis (tests such as the 6MWT and 12MWT require little time, organization, or equipment), or without necessarily taking into account how representative the exercise modality used is to activities of daily living for these patients (eg, cycling tests) or the likely impact of the intervention on the outcome of the test. Equally there are few data describing the relative merits of these tests employed simultaneously to evaluate interventions such as rehabilitation and bronchodilator therapy. Results from this systematic review indicate that there is an extensive body of published literature regarding the performance of the eight exercise tests that are widely used.

Repeatability data were found for only the 6MWT, 12MWT, ISWT, ESWT, and ICET. As could be anticipated, we did not identify studies assessing the repeatability of the ECET, ITT, or ETT; these studies test to exhaustion and would be impractical for patients with COPD to perform repeatedly on the same day. Of those reported, the 6MWT was by far the most thoroughly assessed. This may reflect its simplicity and relevance to daily life.⁸⁷ However, a substantial proportion of the reported data does not explicitly support 6MWT repeatability. Some studies suggested that the ISWT was more repeatable²¹ and reproducible¹⁸ than the 6MWT, while another found that the ESWT was, in turn, more repeatable than the ISWT.²³ The comparative results of exercise tests are inconsistent. Furthermore, there are only very limited data to support the repeatability and/or reproducibility of all cycle tests and treadmill tests. Repeatability and reproducibility were generally improved with familiarization in all types of test. Although this review assesses the influence of protocol variations, it is possible that some studies in which variations were not the primary focus may not have been identified. However, it is clear that even in ostensibly identical tests (eg, two studies reporting the 6MWT), responses can be significantly affected by subtle variations in track layout or environment, or by encouragement from the researchers conducting the test.

When studies reporting the sensitivity of two or more exercise tests to therapeutic intervention were reviewed, there was no consistent evidence supporting the use of one test over any other. Nine of the eleven studies assessing the 6MWT after pulmonary rehabilitation,^{53–55,57,59,61,63,65,66} two of the four studies assessing 6MWT after bronchodilators,^18,67 and the only study assessing the 6MWT after lung-volume reduction surgery⁷¹ all reported distance improvements greater than the MCID. Additionally, three of the five studies assessing ISWT^60–62 and all five of the studies assessing ESWT after pulmonary rehabilitation^{23,47,59,60,62} reported distance improvements greater than the MCID. Performance improvements were also observed to be in excess of the MCID in the only two studies assessing the ESWT following bronchodilator therapy;^14,69 this was therefore the only exercise test reported by multiple papers that consistently responded to therapeutic interventions to a clinically relevant degree. It must be considered that these exercise tests have differing physiological demands, and it may be that the benefits of each intervention are measured differently by each test. However, limited data for bronchodilator therapy and lung-volume reduction surgery make it difficult to identify any obvious differences in the responses of tests to these interventions.

The review has several limitations that must be acknowledged. As well as identifying whether or not test responses have exceeded the MCID, we have also reported percentage changes in exercise test performance whenever possible to enable a very crude comparison of test outcomes recorded in different units. However, we are aware that the validity of this comparison relies on a direct relationship between these test outcomes, which is unlikely to be true: a large percentage change in one test result may not be equivalent to a large percentage change in another.

In an attempt to assess the validity of exercise tests in patients with COPD as comprehensively as possible, we have collated data from studies that have used various definitions of COPD, which include distinct subcategories such as emphysema and chronic bronchitis, and that have employed diverse methods of assessing diagnosis and severity. Moreover, the study designs included are very diverse. These issues make meaningful meta-analysis difficult. We have, however, tried to provide the ranges of responses (both absolute and percentage changes) when possible, to provide an indication of the magnitude of exercise test responses. A further consequence of the comprehensive nature of this review is that the sample sizes of the identified studies vary widely. For this reason, we have included study sample sizes in our tables.

Decisions regarding which test to use are also influenced by the practicalities of routine clinical practice. It is reasonable to assume that walking is more representative of daily life than cycling for patients with COPD. Furthermore, given the equivocal evidence for the use of the ESWT over the ISWT, clinicians may wish to consider that the ESWT requires a prior “workload setting” ISWT to be performed by the patient, requiring additional time and resource considerations. Any test of exercise capacity should be highly repeatable and reproducible and also should be able to detect changes in performance after interventions aiming to improve exercise capacity.

Conclusion

This review of the published literature has found good evidence to support the repeatability and reproducibility of all tests, particularly the 6MWT, as long as a prior familiarization is conducted. There is consistent evidence to suggest that the ESWT is highly sensitive to therapeutic intervention. Sensitivity data that are available for other tests are largely inconsistent, and the 6MWT and ISWT appear to be less sensitive to intervention than the ESWT and ICET. These factors, allied to practical aspects, must be considered when planning interventional trials.

Author contributions

All authors contributed to the conception and design of the study, the analysis and interpretation of data, and revision of the manuscript. All authors approved the final version of the manuscript for publication.

Acknowledgments

The study was funded by GlaxoSmithKline UK. Writing support was provided by Martin Bell of Oxford Pharma-Genesis™ Ltd, funded by GlaxoSmithKline UK. Georgina Meakin and Iain Fotheringham are employees of Oxford PharmaGenesis Ltd, which has received funding from GlaxoSmithKline UK. Yogesh Suresh Punekar, John Riley, and Sarah Cockle are current employees of GlaxoSmithKline, Uxbridge, UK. Sally J Singh was involved with the development of the incremental shuttle walk test, and has served on advisory boards for GlaxoSmithKline. Sally J Singh was part funded by the National Institute for Health Research (NIHR) Collaboration for Leadership in Applied Health Research and Care East Midlands. Support was also provided by the NIHR Leicester Respiratory Biomedical Research Unit. The views expressed are those of the authors and not necessarily those of the National Health Service, the NIHR, or the Department of Health.

Supplementary materials

Complete Ovid search strings

Table S1 Embase™ search strings, search conducted January 22, 2013

Search	Search strings	Results
1	chronic obstructive pulmonary disease.mp. or exp chronic obstructive lung disease/	68,878
2	chronic obstructive lung disease.mp.	65,527
3	chronic obstructive respiratory disease.mp.	119
4	copd.mp.	33,438
5	chronic obstructive airway disease.mp.	378
6	chronic bronchitis.mp. or exp chronic bronchitis/	16,146
7	lung emphysema.mp. or exp lung emphysema/	19,606
8	chronic obstructive airway disease.mp.	378
9	chronic airflow obstruction.mp.	627
10	1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9	103,379
11	exp cardiopulmonary exercise test/or exp exercise tolerance/or exp aerobic exercise/or exp exercise/or exp exercise test/	228,181
12	(exercise or exercise capacity or exercise tolerance).mp.	302,038
13	(clinically important difference or mcid or clinically meaningful).mp.	6,903
14	11 or 12 or 13	311,378
15	(‘6 minute walk’ or ‘6m walk’ or ‘6 m walk’ or ‘six minute walk’ or ‘6MWD’).mp.	4,644
16	(‘12 minute walk’ or ‘12m walk’ or ‘12 m walk’ or ‘twelve minute walk’ or ‘12MWD’).mp.	90
17	(‘shuttle walk’ or ‘shuttle walking’ or ‘iswt’ or ‘10 metre walking’ or ‘10 meter walking’ or ‘ESWT’).mp.	1,010
18	cycle ergometer.mp. or exp bicycle ergometer/	5,540
19	bicycle ergometry.mp. or exp bicycle ergometry/	6,840
20	(cycle ergometry or cycle ergometric or bicycle ergometric).mp.	1,721
21	treadmill.mp. or exp treadmill/	29,379
22	(short physical performance battery or sppb).mp.	367
23	15 or 16 or 17 or 18 or 19 or 20 or 21 or 22	45,996
24	10 and 14 and 23	1,626
25	(test or assessment or capacity).mp.	3,225,394
24	10 and 14 and 23	1,408
25	(test or assessment or capacity).mp.	1,099

Table S2 Medline^® search strings, search conducted January 22, 2013

Search	Search string	Results
1	chronic obstructive pulmonary disease.mp. or exp pulmonary disease, chronic obstructive/	31,239
2	chronic obstructive lung disease.mp.	2,686
3	exp lung diseases, obstructive/or chronic obstructive respiratory disease.mp.	150,257
4	copd.mp.	21,567
5	chronic obstructive airway disease.mp.	239
6	chronic bronchitis.mp. or exp bronchitis, chronic/	8,867
7	lung emphysema.mp.	298
8	chronic obstructive airway disease.mp.	239
9	chronic airflow obstruction.mp.	502
10	1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9	161,227
11	exp exercise tolerance/or exp exercise test/or exercise.mp. or exp exercise/	239,508
12	(exercise capacity or exercise tolerance).mp.	16,495
13	(clinically important difference or mcid or clinically meaningful).mp.	4,675
14	11 or 12 or 13	243,911
15	(‘6 minute walk’ or ‘6m walk’ or ‘6 m walk’ or ‘six minute walk’ or ‘6mwd’).mp.	2,563
16	(‘12 minute walk’ or ‘12m walk’ or ‘12 m walk’ or ‘twelve minute walk’ or ‘12mwd’).mp.	62
17	(‘shuttle walk’ or ‘shuttle walking’ or ‘iswt’ or ‘10 metre walking’ or ‘10 meter walking’ or ‘eswt’).mp.	644
18	$cycle ergomet*.mp.	4,838
19	treadmill.mp.	20,912
20	(short physical performance battery or sppb).mp.	247
21	15 or 16 or 17 or 18 or 19 or 20	28,642
22	10 and 14 and 21	1,278
23	(test or assessment or capacity).mp.	1,939,771
24	10 and 14 and 21 and 23	1,065
25	limit 24 to (English language and humans)	928

Table S3 The Cochrane Library search strings, search conducted January 22, 2013

Search	Search string	Results
1	MeSH descriptor: [Pulmonary Disease, Chronic Obstructive] explode all trees	1,845
2	chronic obstructive pulmonary disease	4,929
3	MeSH descriptor: [Lung Diseases, Obstructive] explode all trees	12,759
4	chronic obstructive respiratory disease	3,340
5	copd	6,244
6	chronic obstructive airway disease	4,348
7	MeSH descriptor: [Bronchitis, Chronic] explode all trees	77
8	chronic bronchitis	1,969
9	lung emphysema	517
10	chronic obstructive airway disease	4,348
11	chronic airflow obstruction	416
12	#1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9 or #10 or #11	18,973
13	MeSH descriptor: [Exercise Tolerance] explode all trees	1,431
14	MeSH descriptor: [Exercise Test] explode all trees	5,816
15	MeSH descriptor: [Exercise] explode all trees	11,528
16	exercise or exercise capacity or exercise tolerance	38,220
17	clinically important difference or mcid or clinically meaningful	7,450
18	#13 or #14 or #15 or #16 or #17	45,480
19	‘6 minute walk’ or ‘6m walk’ or ‘6 m walk’ or ‘six minute walk’ or ‘6mwd’	3,377
20	‘12 minute walk’ or ‘12m walk’ or ‘12 m walk’ or ‘twelve minute walk’ or ‘12mwd’	2,233
21	‘shuttle walk’ or ‘shuttle walking’ or ‘iswt’ or ‘10 metre walking’ or ‘10 meter walking’ or ‘eswt’	716
22	*cycle ergomet*	2,392
23	treadmill	3,759
24	short physical performance batter or sppb	41
25	#19 or #20 or #21 or #22 or #23 or #24	9,217
26	test or assessment or capacity	211,756
27	#12 and #18 and #25 and #26	779

Disclosure

Other than the funding outlined in the “Acknowledgments” section, the authors declare no conflicts of interest in this work.