Abstract
Background Gait characteristics are important risk factors for falls, hospitalisations and mortality in older adults, but the impact of COPD on gait performance remains unclear. We aimed to identify differences in gait characteristics between adults with COPD and healthy age-matched controls during 1) laboratory tests that included complex movements and obstacles, 2) simulated daily-life activities (supervised) and 3) free-living daily-life activities (unsupervised).
Methods This case–control study used a multi-sensor wearable system (INDIP) to obtain seven gait characteristics for each walking bout performed by adults with mild-to-severe COPD (n=17; forced expiratory volume in 1 s 57±19% predicted) and controls (n=20) during laboratory tests, and during simulated and free-living daily-life activities. Gait characteristics were compared between adults with COPD and healthy controls for all walking bouts combined, and for shorter (≤30 s) and longer (>30 s) walking bouts separately.
Results Slower walking speed (−11 cm·s−1, 95% CI: −20 to −3) and lower cadence (−6.6 steps·min−1, 95% CI: −12.3 to −0.9) were recorded in adults with COPD compared to healthy controls during longer (>30 s) free-living walking bouts, but not during shorter (≤30 s) walking bouts in either laboratory or free-living settings. Double support duration and gait variability measures were generally comparable between the two groups.
Conclusion Gait impairment of adults with mild-to-severe COPD mainly manifests during relatively long walking bouts (>30 s) in free-living conditions. Future research should determine the underlying mechanism(s) of this impairment to facilitate the development of interventions that can improve free-living gait performance in adults with COPD.
Tweetable abstract
Gait impairment of adults with COPD is only observed during relatively long walking bouts (>30 s) in free-living conditions but not during shorter (≤30 s) walking bouts in either laboratory or free-living settings https://bit.ly/44vBpYp
Introduction
COPD is characterised by persistent respiratory symptoms [1], airflow limitation [1] and extrapulmonary disease manifestations such as fatigue [2], weight loss [3] and comorbidities [4]. Adults with COPD are also prone to falls [5], which can lead to long-lasting pain, functional impairment, disability and death [6]. Even just fear of falling can lead to physical activity avoidance [7], which may in turn increase risk for exacerbations, hospital admissions and mortality in adults with COPD [8–10]. Since altered gait is one of the main risk factors for falls [11, 12], characterising gait is critical for preventing falls and their devastating consequences.
Walking speed is arguably the most intensively studied and validated gait characteristic [13]. It is an important predictor of multiple falls [11], incident disability and mortality in older adults [14]. Walking speed depends on two other gait characteristics: stride length (one stride is two consecutive steps) and cadence (the number of steps taken per minute). Another gait characteristic, double support duration, represents the duration of simultaneously having both feet in contact with the ground while walking and is generally increased in older adults [15]. Decreased stride length and cadence, and increased double support duration are all associated with an increased fall risk in older adults [11, 12]. Finally, gait variability measures (e.g. coefficients of variation of stride duration, stride length or double support duration) represent movement consistency and stability (i.e. gait not leading to falls, in spite of perturbations) [16, 17]. It has been argued that increased gait variability might even better predict increased fall risk in older adults than previously discussed mean values [11, 18].
Despite the importance of the seven abovementioned gait characteristics as risk factors for falls, hospitalisations and mortality in older adults, relatively few studies have examined gait characteristics in COPD. Furthermore, the available literature shows inconsistencies on the impact of COPD on gait. Some studies identified a reduced walking speed, cadence, stride length and a higher double support duration for adults with COPD compared to healthy controls [19–21], while other studies did not [22, 23]. Based on a 2021 scoping review, we hypothesise that adults with COPD walk with a slower walking speed, lower cadence, shorter stride length, a longer double support duration and similar gait variability compared to healthy controls [13]. By analogy with other chronic conditions such as Parkinson's disease [24] and multiple sclerosis [25], we hypothesise that the degree of gait impairment in COPD is different during longer (>30 s) compared to shorter (≤30 s) walking bouts, as symptoms that could affect gait, such as leg fatigue or dyspnoea [26], are more likely to be present during longer walking bouts. Another limitation of previous research is that differences in gait characteristics between adults with COPD and controls have almost exclusively been examined in a laboratory setting [27], without considering any of the complex movements (e.g. starting to walk from a seated position) or obstacles (e.g. a step) that are inherent to free-living walking outside of the laboratory. Consequently, available information on gait in COPD represents gait capacity, rather than free-living gait performance. We hypothesise that gait impairment in COPD is more pronounced during more complex activities and that therefore the impact of COPD on gait can mainly be observed in free-living conditions.
Therefore, the objectives of this study were to identify differences in gait characteristics between adults with COPD and healthy controls during 1) laboratory tests that include complex movements (e.g. timed up and go) or obstacles (e.g. walking over a step), 2) simulated daily-life activities (supervised) and 3) free-living daily-life activities (unsupervised). The generated insights will help to develop and evaluate early intervention and prevention strategies for improving free-living gait performance of adults with COPD, and consequently reduce their fall risk and fall-related complications.
Methods
Study design and participants
This multicentre, observational case–control study used data from the technical validation study of the IMI2-JU-funded Mobilise-D project (https://www.mobilise-d.eu/), which aimed to validate a new digital method for remote monitoring of mobility in different cohorts, including adults with COPD and healthy older adults [28, 29]. The study was approved by the London-Bloomsbury Research Ethics committee (19/LO/1507), medical faculty of Kiel University (D540/19), medical faculty of the University of Tübingen (647/2019BO2) and Helsinki Committee of the Tel Aviv Sourasky Medical Center (0551-19TLV). All participants provided written informed consent.
Clinically stable adults with COPD (post-bronchodilator forced expiratory volume in 1 s (FEV1) to forced vital capacity (FVC) ratio <0.70; smoking history equivalent to at least 10 pack-years; 4 weeks without antibiotics and/or oral corticosteroids to treat a moderate or severe exacerbation) were recruited from The Newcastle upon Tyne Hospitals NHS Foundation Trust (UK) and Sheffield Teaching Hospitals NHS Foundation Trust (UK). Candidates were excluded if they had undergone major lung surgery, had a lung tumour or a respiratory disease other than COPD, or had orthopaedic, neurological or other complaints that significantly impaired normal biomechanical movement patterns, as judged by the investigator. Healthy controls (>65 years) were recruited from The Newcastle upon Tyne Hospitals NHS Foundation Trust (UK), the University of Kiel (Germany), Robert Bosch Foundation for Medical Research (Germany) and the Tel Aviv Sourasky Medical Center (Israel). Data were collected between July 2020 and July 2021.
Participants were excluded if they were unable to walk 4 m independently with or without walking aid, had shoe size <36 European Union (3 UK), Montreal Cognitive Assessment score ≤15 or an occurrence of any of the following within 3 months prior to inclusion: myocardial infarction, hospitalisation for unstable angina, stroke, coronary artery bypass graft, percutaneous coronary intervention or implantation of a cardiac resynchronisation therapy device.
Procedures
Gait characteristics were obtained during laboratory tests, simulated daily-life activities and free-living daily-life activities using the INertial module with DIstance Sensors and Pressure insoles (INDIP) system, a validated multi-sensor system including pressure insoles, distance sensors at the ankles, and magneto-inertial sensors on the shoes and lower back [28, 30–33] (supplementary figure E1).
Laboratory tests included five structured mobility tasks with increasing complexity, performed at the participants' preferred walking speed (figure 1) [28]. When needed, the use of arm rests or handrails was permitted during the tests. Simulated daily-life activities were based on a series of supervised daily-life tasks performed in the laboratory while sitting and moving around a room, such as picking up objects and setting the table [28] (figure 2).
Visualisation of laboratory tests. Reproduced and modified from [29] with permission.
Visualisation of simulated daily-life activities. Reproduced and modified from [29] with permission.
Free-living daily-life activities consisted of 2.5 h of unsupervised activities in a habitual environment (home/work/community) chosen by the participant. The duration of the measurements was chosen as a trade-off between experimental, clinical and technical requirements [28]. The free-living activities were unstructured, but participants were encouraged to complete several specific tasks, such as rise from a chair and walk to another room, walk up and down a set of stairs, and walk outdoors. Hence, these activities should not be interpreted as entirely regular daily-life routines.
Variables
Participant characteristics were collected for all participants including age, sex, height, body mass index, falls in the previous year, walking aid usage, living arrangement, education and pain while walking. Additionally, the COPD group completed the COPD assessment test and performed a 6-min walk test according to American Thoracic Society (ATS) guidelines. Pulmonary function (FEV1 and FVC) for participants with COPD was obtained from medical records completed within 6 months prior to participant inclusion.
For the laboratory tests, simulated daily-life activities and free-living daily-life activities, we collected: the number of walking bouts, walking bout duration (s), walking bout length (cm) and the number of turns per walking bout. A walking bout was defined as a walking sequence containing at least two consecutive strides of both feet. The start and end of a walking bout were determined by a resting period or any non-walking activity for at least 3 s [30, 34].
For every walking bout, seven gait characteristics were calculated using the INDIP system: 1) walking speed (cm·s−1), 2) cadence (steps·min−1; number of steps taken per minute), 3) stride length (cm; length of two consecutive steps), 4) double support duration (s; duration of simultaneously having both feet in contact with the ground), 5) stride duration (duration of two consecutive steps) variability (%; coefficient of variation of stride duration), 6) stride length variability (%; coefficient of variation of stride length) and 7) double support duration variability (%; coefficient of variation of double support duration). Of note, the acceleration and deceleration phase at the start and end of each laboratory test was included for the calculation of the gait characteristics (see the definition of a walking bout above), as this more closely reflects free-living conditions [34].
Statistical analyses
Statistical power to identify differences in gait characteristics between adults with COPD and healthy controls was estimated as 86%, based on walking speed data from previous studies (standard deviation of 20 cm·s−1 and expected differences of 20 cm·s−1 [19, 21, 22, 26, 35–37]), α=0.05, and our sample size of 17 adults with COPD and 20 healthy controls, using the GRANMO power calculator (https://www.imim.es/ofertadeserveis/software-public/granmo/).
Participant and gait characteristics are presented as mean±SD for normally distributed continuous variables and as n (%) for categorical variables. Normality was tested for all variables using histograms. Participant characteristics, the number of walking bouts per participant and the total time spent walking during daily-life activities were compared between adults with COPD and healthy controls using unpaired t-tests, Wilcoxon rank-sum or Fisher's exact tests, as appropriate.
Gait characteristics and walking bout duration were compared between adults with COPD and healthy controls using linear regression models adjusted for age, height and walking bout length. For the five laboratory tests (all including one walking bout by design), regression analyses were performed with one observation per participant. For the simulated and free-living daily-life activities (consisting of multiple walking bouts per participant), mixed effect regression models with a random intercept for individuals were used. To test whether the impact of COPD on gait performance was affected by walking bout duration, the previous models were additionally stratified by walking bout duration, using a cut-off based on scientific literature (i.e. shorter (≤30 s) and longer (>30 s) walking bouts) [24, 25]. Two sensitivity analyses were performed: 1) the mixed effect regression models for simulated and free-living daily-life activities were additionally adjusted for the number of turns per walking bout; and 2) models were stratified using additional cut-offs at 20 and 40 s.
Data preparation and statistical analyses were performed in JupyterLab using the Python 3.5 and the R 4.1.2 programming languages.
Results
Participant characteristics
17 adults with mild-to-severe COPD and 20 healthy age-matched controls were included in the study. There were no statistically significant differences between the groups regarding sex distribution, age, height, body mass index, falls in the previous year, walking aid usage, living arrangement, education and pain while walking (table 1). Three adults with COPD and seven healthy controls had an insufficient number of strides during the straight walking test to identify a walking bout, and one healthy control did not perform the simulated daily-life activities.
Participant characteristics
Gait characteristics during laboratory tests
Walking speed ranged between 89 and 106 cm·s−1 for adults with COPD and between 91 and 106 cm·s−1 for healthy controls during the different tests (table 2). No statistically significant differences in gait characteristics during the laboratory tests were found between adults with COPD and healthy controls, except for less stride duration variability (−1.1%, 95% CI: −2.1 to −0.1) for adults with COPD during the surface test (table 2; figures 3 and 4). All walking bouts during the laboratory tests were shorter walking bouts (i.e. ≤30 s), except for the surface test of one participant with COPD (walking bout of 34 s). Therefore, no stratification was conducted.
Gait characteristics during laboratory tests
Differences in walking speed, cadence, stride length and double support duration between adults with COPD and healthy controls during laboratory tests, expressed as β coefficient (95% CI) of a linear regression model adjusted for age, height and walking bout length. The healthy controls were used as the reference group.
Differences in gait variability measures between adults with COPD and healthy controls during laboratory tests, expressed as β coefficient (95% CI) of a linear regression model adjusted for age, height and walking bout length. The healthy controls were used as the reference group.
Gait characteristics during simulated daily-life activities
Participants with COPD and healthy controls performed 88 and 96 walking bouts with an average walking speed of 59±21 cm·s−1 and 60±24 cm·s−1, respectively, during the simulated daily-life activities (table 3). Adults with COPD had less stride length variability (−5.8%, 95% CI: −11.2 to −0.4; figures 5 and 6). All walking bouts were shorter (i.e. ≤30 s) and therefore no stratification was conducted. Sensitivity analyses in which the models were additionally adjusted for the number of turns provided similar results (supplementary figures E2 and E3).
Gait characteristics (on the walking bout level) during the simulated and free-living daily-life activities (analyses for all free-living bouts combined, and for shorter (≤30 s) and longer (>30 s) walking bouts separately)
Differences in walking speed, cadence, stride length and double support duration between adults with COPD and healthy controls during simulated and free-living daily-life activities, expressed as β coefficient (95% CI) of a linear mixed effects model with random intercepts for participants. The healthy controls were used as the reference group. Models were adjusted for age, height and walking bout length.
Differences in gait variability measures between adults with COPD and healthy controls during simulated and free-living daily-life activities, expressed as β coefficient (95% CI) of a linear mixed effects model with random intercepts for participants. The healthy controls were used as the reference group. Models were adjusted for age, height and walking bout length.
Gait characteristics during free-living daily-life activities
Participants with COPD and healthy controls performed 1035 (90% shorter and 10% longer) and 1330 (88% shorter and 12% longer) walking bouts, respectively, during the 2.5 h of unsupervised, free-living activities (table 3). All participants performed both shorter and longer walking bouts. Healthy controls spent more time walking during the 2.5 h of free-living activities than adults with COPD, originating from more time spent walking during, and longer durations of, longer walking bouts (table 3; supplementary figure E4).
Walking speed during shorter bouts was 57±20 cm·s−1 for patients with COPD and 54±21 cm·s−1 for healthy controls, increasing to 71±23 cm·s−1 for COPD and 85±29 cm·s−1 for controls during longer bouts (table 3). No statistically significant differences in gait characteristics were observed between adults with COPD and healthy controls when all free-living walking bouts were considered together, or during shorter free-living walking bouts (figures 5 and 6). During longer free-living walking bouts, adults with COPD had a slower walking speed (−11 cm·s−1, 95% CI: −20 to −3) and lower cadence (−6.6 steps·min−1, 95% CI: −12.3 to −0.9) compared to healthy controls, and a trend towards a shorter stride length (p=0.08) (figure 5). Sensitivity analyses 1) additionally adjusting the models for the number of turns (supplementary figures E2 and E3) and 2) using cut-offs at 20 s or 40 s for separating shorter and longer walking bouts provided similar results (supplementary figures E5 and E6).
Discussion
Our study was the first to examine differences in free-living gait performance between adults with mild-to-severe COPD and age-matched healthy controls. Walking speed and cadence were significantly reduced in the COPD group during longer (>30 s) free-living walking bouts, and there was a non-significant trend towards a shorter stride length for adults with COPD during these bouts. Opposed to our hypothesis, no differences were observed in walking speed, cadence or stride length during shorter walking bouts in either laboratory or free-living settings. Double support duration and gait variability measures were generally comparable between adults with COPD and healthy controls.
The most important finding of the current study is that mild-to-severe COPD was associated with walking at a slower speed and lower cadence during longer free-living walking bouts, while no differences were observed during shorter bouts. Specifically, walking speed changed from 54±21 cm·s−1 in shorter bouts to 85±29 cm·s−1 in longer bouts among healthy participants, while in COPD it only increased from 57±20 cm·s−1 to 71±23 cm·s−1. The underlying mechanisms of this attenuated increase in walking speed and cadence during longer walking bouts of adults with COPD could be pathological (e.g. due to poor exercise capacity [19, 38]) and/or behavioural (e.g. a coping mechanism to maintain a tolerable dyspnoea sensation [26] or a compensatory strategy to maintain balance and reduce fall risk [39]).
Previous research has provided contradictory results on the presence/absence of differences in gait characteristics between adults with COPD and healthy controls during laboratory tests [19–23, 26, 35, 36]. Careful screening of scientific literature did not identify consistent differences in sample size, disease severity or walking bout durations between these studies that could explain the observed inconsistencies. These inconsistencies might still originate from other methodological or clinical differences between studies that were not systematically reported, such as differences in the starting procedure (static versus dynamic start) [40] or dyspnoea levels of participants [41]. Hence, comparisons of current results with previous literature should only be done cautiously. Nevertheless, the observed slower walking speed and lower cadence during free-living conditions in our study corresponds with the results of the largest laboratory-based study to date on gait in COPD (n=196 adults with COPD walking on an electronic walkway), suggesting that the reduced walking speed in COPD is related to a decrease in cadence [36]. The walking speed during the straight walking test of the healthy controls in the current study was slower than for healthy older adults in previous reports [42]. This can most likely be attributed to the inclusion of the acceleration and deceleration phase at the start and end of each laboratory test in the current study (i.e. a static start), while other laboratory-based studies often excluded this acceleration/deceleration phase (i.e. a dynamic start).
We generally did not observe differences in double support duration or gait variability measures between adults with COPD and healthy controls. In combination with the absence of an association between double support duration/gait variability measures and fall history in COPD [36], this could suggest that stability during walking is not markedly affected in COPD and that these gait characteristics might be of secondary importance in this population. Of note, we observed less stride duration variability in adults with COPD during the surface test, and less stride length variability during the simulated daily-life activities, which could be the result of an inability to adjust gait when faced with a change in walking surface or obstacle [17], or a chance finding.
A clear implication of the present study is the need to consider walking bout duration and include a sufficient number of longer bouts in future studies on gait in COPD. Furthermore, examining the distribution of bout duration over multiple days might reveal differences in walking behaviour between adults with COPD and healthy controls, as already suggested by the 2.5 h of free-living measurements in the current study (supplementary figure E4). From a clinical perspective, the observed gait impairment suggests that free-living gait performance of adults with COPD can be improved, potentially by improving exercise capacity (e.g. through pulmonary rehabilitation [43]) and/or reducing exertional dyspnoea (e.g. through bronchodilator therapy [44]). Ultimately, this could reduce fall risk and increase quality of life and survival of adults with COPD.
The main strength of this study was the collection of gait characteristics in free-living conditions in adults with COPD. By using a multi-sensor wearable system and combining laboratory tests, simulated daily-life activities and free-living daily-life activities, we were able to show that adults with mild-to-severe COPD experience gait impairment during free-living activities that may not be observable during short laboratory tests or simulated daily-life activities. Other strengths of the study include the use of a control group and the adjustment for both participant characteristics (age and height) and walking bout length in the regression analyses.
The study also had some potential limitations. Firstly, the sample size was relatively small. Nevertheless, the performed power calculation indicated that our sample size had high power (i.e. 86%) to respond to the study objectives. Secondly, data were collected during the COVID-19 pandemic, which might have prevented adults with very severe COPD from participating, and which might have affected the activities performed during the free-living measurements. The COVID-19 pandemic also disrupted the recruitment strategy (which intended to distribute the recruitment of participants over five different centres to ensure generalisability [28]), and participants were unevenly recruited by the different centres. Thirdly, gait characteristic values obtained during the straight walking test cannot directly be compared with values obtained in other laboratory-based studies, as most studies exclude the acceleration/deceleration phase at the start/end of a straight walking test. Fourthly, the relatively short duration of free-living measurements (i.e. 2.5 h) might not be representative of all daily-life activities and routines of the participants. Finally, free-living gait characteristics could have been affected by environmental factors (e.g. slope, house/apartment size, meteorological factors), which were not assessed or controlled for in the current study.
In conclusion, gait impairment in adults with mild-to-severe COPD occurred during longer walking bouts (>30 s) with an attenuated increase in walking speed and cadence, compared to healthy controls. These differences were not observed during shorter walking bouts (≤30 s). Double support duration and gait variability measures were generally comparable between the two groups. Future research should replicate our analyses in larger studies and determine the underlying mechanism(s) of this gait impairment during longer bouts to facilitate the development of interventions that can improve free-living gait performance in adults with COPD.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary material 00159-2023.SUPPLEMENT
Acknowledgements
The authors acknowledge all the members of the Mobilise-D work package 2 for continuous discussion and critical input. They are particularly grateful to the participants in the study for their time and enthusiastic contribution, especially during the pandemic. Full membership of the Mobilise-D consortium is available on the website https://mobilise-d.eu/about-us/.
Footnotes
Provenance: Submitted article, peer reviewed.
Support statement: This work was supported by the Mobilise-D project that has received funding from the Innovative Medicines Initiative 2 Joint Undertaking (JU) under grant agreement number 820820. This JU receives support from the European Union's Horizon 2020 research and innovation programme, and the European Federation of Pharmaceutical Industries and Associations (EFPIA). S. Del Din and L. Rochester were also supported by the Innovative Medicines Initiative 2 (IMI2) JU project IDEA-FAST – grant agreement 853981. S. Del Din, L. Alcock and L. Rochester were also supported by the National Institute for Health Research (NIHR) Newcastle Biomedical Research Centre (BRC) based at The Newcastle upon Tyne Hospital NHS Foundation Trust, Newcastle University and the Cumbria, Northumberland and Tyne and Wear NHS Foundation Trust, and by the NIHR/Welcome Trust Clinical Research Facility infrastructure at Newcastle upon Tyne Hospitals NHS Foundation Trust. C. Mazzà and E. Buckley were also supported by the Sheffield BRC based at the Sheffield Teaching Hospital and the University of Sheffield. H. Demeyer is a post-doctoral research fellow of the Flemish Research Foundation (FWO Flanders). The Barcelona Institute for Global Health (ISGlobal) receives support from the Spanish Ministry of Science, Innovation and Universities through the “Centro de Excelencia Severo Ochoa 2019–2023” Programme (CEX 2018-000806-S), and support from the Generalitat de Catalunya through the CERCA Programme. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR, the Department of Health, the IMI, the European Union, the EFPIA or any Associated Partners. Funding information for this article has been deposited with the Crossref Funder Registry.
Conflict of interest: A. Mueller is an employee of Novartis and holds shares in this company. L. Palmerini is co-founder of and owns shares in mHealth Technologies (https://mhealthtechnologies.it/). S. Del Din reports consultancy activity with Hoffmann-La Roche Ltd. outside the scope of this study. I. Vogiatzis is an associate editor of this journal. The other authors have nothing to disclose.
- Received March 14, 2023.
- Accepted June 29, 2023.
- Copyright ©The authors 2023
This version is distributed under the terms of the Creative Commons Attribution Non-Commercial Licence 4.0. For commercial reproduction rights and permissions contact permissions{at}ersnet.org
References
