The walk ratio: Investigation of invariance across walking conditions and gender in community-dwelling older people

Accepted Manuscript Title: The Walk Ratio: Investigation of invariance across walking conditions and gender in community-dwelling older people Authors: B˚ard Bogen, Rolf Moe-Nilssen, Anette Hylen Ranhoff, Mona Kristin Aaslund PII: DOI: Reference:

S0966-6362(18)30101-2 https://doi.org/10.1016/j.gaitpost.2018.02.019 GAIPOS 5971

To appear in:

Gait & Posture

Received date: Revised date: Accepted date:

29-6-2017 17-2-2018 19-2-2018

Please cite this article as: Bogen B˚ard, Moe-Nilssen Rolf, Ranhoff Anette Hylen, Aaslund Mona Kristin.The Walk Ratio: Investigation of invariance across walking conditions and gender in community-dwelling older people.Gait and Posture https://doi.org/10.1016/j.gaitpost.2018.02.019 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

The Walk Ratio: Investigation of invariance across walking conditions and gender in community-dwelling older people.

SC R

IP T

Authors: 1. Bård Bogen (MSc, PhD-student) Department of Global Public Health and Primary Care, University of Bergen Kalfarveien 31 5018 Bergen Norway e-mail: [email protected]

M

A

N

U

2. Rolf Moe-Nilssen (PhD, Professor) Department of Global Public Health and Primary Care, University of Bergen Kalfarveien 31 5018 Bergen Norway e-mail: [email protected]

PT

ED

3. Anette Hylen Ranhoff (Dr. med, Professor) Department of Clinical Science, University of Bergen Postbox 7804, 5020 Bergen Norway e-mail: [email protected]

A

CC E

4. Mona Kristin Aaslund (PhD, Research fellow) Department of Global Public Health and Primary Care, University of Bergen Kalfarveien 31 5018 Bergen Norway e-mail: [email protected] Research highlights   

The Walk ratio is invariant across different speeds and on an uneven surface. The Walk ratio increases when counting backwards. The women in our study increased speed by cadence and not step length.

Abstract Background The step length-cadence ratio, also called the walk ratio (WR; cm/steps/min) is a measure of cautious gait, poor balance control or impaired gait, but has not been investigated for both genders in a general population of older adults across different speeds and conditions.

Method

IP T

The participants were community-dwelling volunteers between 70-81 years. They walked 6.5 meters under four different conditions: At preferred speed, fast speed, during a dual task condition and on an uneven surface. Step length (cm) and cadence (steps/minute) was captured using a body-worn sensor. Both cadence and step lengths were adjusted for body height.

Results

N

U

SC R

70 older adults participated (mean age 75.5 (SD 3.4), 60 percent women). The WR was 0.60 cm/steps/min (SD 0.07) during preferred speed walking, 0.58 cm/steps/min (SD 0.07) during fast walking, 0.68 cm/steps/min (SD 0.18) during dual task-walking and 0.59 cm/steps/min (0.07) during uneven surface-walking. In planned pairwise comparisons, the WR during dual task was significantly different from preferred speed walking (mean difference -0.087 cm/steps/min, 95% CI -0.140, 0.033), from fast speed walking (mean difference -0.098 cm/steps/min, 95% CI -0.154, -0.041) and uneven surface walking (mean difference 0.092 cm/steps/min, 95% CI 0.040, 0.145). There were no gender differences except during the fast walking condition, where women had a significantly lower WR than the men (0.56 cm/steps/min vs 0.61 cm/steps/min, p=0.002).

A

Discussion

ED

M

We found that the WR is invariant during different speeds, and during an uneven surface condition, but is affected during a dual task-condition, when attention must be divided between a cognitive and a motor task.

Background:

PT

Keywords: Walk ratio, older, automaticity, gender

A

CC E

Gait speed is widely used as a measure of health and function in older people (1), but is also unspecific and reveals little about gait quality. Mobility limitations are known to increase with age (2), and detection of early signs of gait pathology would be helpful for providing appropriate health care (3). Easily administered measurements of gait quality in older adults are therefore needed. Gait speed is determined by step length and step frequency (cadence), and changes in gait speed come from varying those gait features. Interestingly, the relationship between step length and cadence (cm/steps/min), termed the walk ratio (WR), appears stable for all but very low speeds (4, 5). This suggests that the WR is an invariant feature of gait, allowing for an optimal balance between energy expenditure and forward propulsion, for stability, and for minimal attentional demand during walking (4). Normal WRs lie in a range around 0.65 cm/steps/min, and WRs that are lower suggest a strategy for maintaining speed that typically involves increasing cadence while lowering step length. This may be indicative of cautious gait, poor balance control or impaired central control of gait (6, 7). Higher WRs than normal suggest increased step length or lower cadence, or both. In previous studies on clinical populations, the WR has been found to discriminate between persons with and without multiple sclerosis (MS), between different severities of MS, and between persons

ED

M

A

N

U

SC R

IP T

with MS with and without a history of falls (8, 9). In a recent randomized controlled trial, patients with hip fracture receiving in-hospital comprehensive geriatric assessments had higher WRs than control subjects 12 months after surgery (10). Low WRs have also been found to be associated with falls (11, 12), and with mild to moderate dementia (13). In a study of a healthy young, middle-aged and old volunteers, the WR was found to be unaffected by age and almost invariant of speed; however, walking under conditions with added demands for attention or stability was not examined. In addition, no gender specific analyses were performed in this study (7). In another study including physically active older women recruited from recreational sports groups, the WR decreased during performance of some dual tasks (walking while carrying a tray, and walking while carrying a tray and performing a Stroop task), but not while performing the Stroop task alone, suggesting that attentional load perturbs the automaticity of gait. In their study, WRs appeared to vary between young-old and old-old at slow and preferred speeds (6). In these two studies WRs were not adjusted for height (6) or only step length was adjusted (7). As many gait characteristics are dependent on stature, failure to do this may give misleading results, particularly for between-subject comparisons. Also, the gender differences were not examined in the first study (7), and the second study included only women (6). Sekiya and co-authors found no gender differences in young adults (4), while Ko and co-authors described how older women walked with higher cadences and shorter stride lengths than older men (14). In light of this, exploration of eventual gender differences in WRs in older adults seem warranted. Thus, there is a need to confirm the invariance of WRs for both older men and women, adjusting for stature. The aim of this study was therefore to examine the invariance of WRs adjusted for stature during gait at preferred and fast speeds, and under different environmental and task constraints in communitydwelling older adults. We hypothesized that WRs adjusted for stature would differ between older men and women, would remain stable at preferred and fast speeds, but vary under conditions that challenged postural stability or that included secondary cognitive demands.

Method Design

CC E

Participants

PT

This study has a cross-sectional repeated measures design where subjects were tested on one occasion under differing task and environmental constraints.

A

400 names and addresses of women and men between 70-81 years old from the municipality of Bergen were retrieved from the electoral roll. They were then invited by mail and telephone to attend testing at a movement lab at the University of Bergen. To be included, participants had to be able to walk 10 meters without walking aids, live in their own homes and be able to give informed, written consent.

Procedures Walking tests were performed on a 6.5 meter walkway with photoelectric cells at each end. In addition, participants walked 2 meters before and after the photoelectric cells, to avoid inclusion of acceleration and deceleration phases in the recording. Recording started and stopped automatically at the passing of the photoelectric cells. The participants wore a triaxial Xsens accelerometer (MTx, Xsens Technologies B.V., Enschede) attached to the lower back with an elastic belt. The acceleration signal was transmitted wirelessly to a laptop via Bluetooth, and processed by in-house software.

IP T

Cadence was calculated as 60*frequency (Hz)/samples per step. Similarly, step length (cm) was calculated as walking speed (cm/s)*samples per step/frequency (Hz). A detailed description of the procedure has been published previously (15). Following laws of pendulum mechanics, where step length is proportional to body height, while cadence is proportional to the inverse of the square root of body height, adjustments for height were made. The body-height adjusted WR was calculated using the following formulas by Sekiya (4): Body height-adjusted step length=(step length/body height) x (average body height). Body height-adjusted cadence=cadence x (body height/average body height)1/2. Body height-adjusted walk ratio= body height-adjusted step length/body height-adjusted cadence.

N

U

SC R

The participants were allowed one practice walk at preferred speed for familiarization. Then, while recording, they walked back and forth once during the following four conditions: i) at preferred speed (“Walk as you normally would”) ii) at fast speed “(Walk as fast as you can without running or losing balance”) iii) during dual task while counting backwards from 50, with intervals of three (no instruction for prioritizing tasks) iv) on an uneven surface made by a rubber mat with unevenly spaced convex circular bulges, covered by another mat (the participants were made aware of the bulges).

Statistical analysis

ED

M

A

Data was exported to Microsoft Excel 2013 by the in-house software, and analyses were performed in IBM SPSS Statistics 2015 software. WRs, cadences, step lengths, and walking speeds for all conditions are presented as means and standard deviations, for both men and women. To investigate within-subject WR invariance across conditions, we used repeated measures analysis of variance (ANOVA) with gender as a between-subject factor, and planned within-subject comparisons of WRs in accordance with the main hypotheses, and with Bonferroni-correction for multiple comparisons.

Results

A

CC E

PT

The sample consisted of 70 men and women (60 percent women, average age 75.5 years, SD 3.4). None of the participants used walking aids habitually, and all were community ambulators. Mean WR was between 0.58 - 0.60 cm/steps/min for preferred speed, fast speed and uneven surface, while mean WR for dual task was 0.68 cm/steps/min (Table 1). Mean walking speed differed from 87 - 147 cm/s between conditions, cadence from 89.2 - 122.6 steps/min, and step length from 58 - 71 cm (Table 1). Since the assumption of sphericity was violated (Mauchly’s test p<0.001), multivariate tests were used to investigate within-subjects effects in the repeated measures ANOVA model, which demonstrated significant main effect of walking condition (p<0.001) on WR, and also a significant condition*gender interaction (p = 0.045) To investigate specific hypotheses, planned pairwise comparisons were performed and showed that the working hypothesis of invariance of WR between preferred and fast walking speed was confirmed. Likewise the working hypothesis of a change in WR under a condition that included a secondary cognitive task was confirmed. However, the hypothesis of a change in WR under a condition that challenged postural stability was not confirmed. See Table 2. To further explore the condition*gender interaction, we did a post hoc pairwise comparison of gender differences between for the four conditions using Bonferroni correction. Only the fast walking

speed condition returned a significant gender difference in WR (p=0.02) with mean WR (SD) of 0.61 cm/steps/min (0.08) for men and 0.56 cm/steps/min (0.05) for women. Gender-specific WR-data for the participants have been published previously(16). Table 1. Approximately here Table 2. Approximately here

Discussion

A

CC E

PT

ED

M

A

N

U

SC R

IP T

In this study, we hypothesized that the WR would be invariant at preferred and fast gait speeds, and that dual task and uneven surface walking would affect WR by being more challenging tasks. We found, in line with previous studies, that the WR did not change between the preferred and fast walking speed conditions. This suggests that the WR is an invariant feature of gait, to optimize energy expenditure and stability (8). Also, the WR offers the opportunity of comparing individuals who walk at different preferred speeds. Contrary to our hypothesis, the WR during walking on an uneven surface was similar to walking at preferred and fast speeds, even though walking speed was lower. Menz (17) found that WRs increased when young subjects walked across uneven surfaces compared to over even surfaces; however, in that study they walked at the same speed at both the even and uneven surfaces. Our participants slowed down, from 117 cm/s on the even surface to 102 cm/s on the uneven surface. It is possible that the young subjects could have been intent on keeping the same speed, and choose a strategy of longer but fewer steps, whereas our participants choose a strategy that involved caution, with lower speed but similar step length-cadence relationship. It should be noted that not many of our participants found the uneven surface difficult. Many expressed familiarity with the condition, likening it to walking in a forest or similar. WRs during the dual task condition were significantly different from during the other conditions. The dual task paradigm is based on the assumption that postural control during walking and other motor behaviors is not automatic but reliant on attention, and that dividing attention will add a cost to the motor task (18). In our study, this cost was expressed as increased WR (Table 1). Zijlstra (6) and co-authors found that WR at preferred speed was unchanged during a Stroop task, and decreased during a task of carrying a tray, and carrying a tray and performing a Stroop task simultaneously. This may reflect that different added tasks may impact the motor task differently. Similarly, in a study by Nordin et al (19) it was found that increased step width during a tray carrying task was an indication of reduced fall risk, whereas increased step width during a counting task was indicative of increased fall risk. In a counting task such as in our study, temporal factors come into play and the normal rhythm of steps slow down for the participant to have time to calculate the next number. A task of carrying a tray however, requires diligence and alertness, and would be better solved by shortening steps to allow for shorter swing times and relatively shorter periods of single support, thereby increasing stability. It is a common observation that women and men walk differently, which may be explained both culturally and anatomically. However, there is relatively little empirical evidence that substantiate these differences. We have found no other studies of older adults where WRs were analyzed according to gender. Ko and co-authors found that during preferred speed walking, women walked with higher cadences and lower step lengths than men (14). Chui and Lusardi compared different age groups with regards to gender differences in step length and cadence, at both preferred and fast speeds. While differences at different speeds were not compared statistically, it seemed that men increased both step length and cadence more than women when fast walking (20). In principle, in

PT

ED

M

A

N

U

SC R

IP T

both these studies, the WR would remain stable. The female participants in our study walked with significantly lower WR during the fast speed condition. As WRs were adjusted for height in our study, anthropometric factors such as leg length are unlikely to explain this difference. In previous studies, healthy adults who choose a strategy of increasing stride length to maximize speed had higher lower extremity joint moments at fast speeds, than participants who choose a strategy of increasing cadence (21). Also, maximizing speed by increasing stride lengths involve a more out-of-phase pelvicthorax counter rotation strategy (22). Whether these mechanisms had any influence on the strategies chosen by the women in our study is conjecture, but the lower WRs during fast walking suggest that women choose a slightly more cautious approach to fast walking. The participants in our study had a preferred speed WR of 0.60 cm/steps/min, which is higher than in MS-patients (8) and comparable to young adults (4) and physically active older women (6). This could suggest that our participants overall were well-functioning and in reasonably good health. This is mirrored in an average preferred walking speed of 1.17 m/s, which is above the cut-off for indicating good health and function (23), although it should be mentioned that the procedure in our study involved a dynamic start, which tends to give higher walking speeds than static starts (24). Also, participants were living in their own homes, and had sufficient mental and physical capacity to volunteer and show up for the testing. Therefore, generalizability may be limited, and further studies investigating the WR in frailer old adults is warranted. It is noteworthy that accumulating evidence suggest that there is a “normal” range of WRs, which is helpful in identifying gait pathology. This study is also limited by our procedures: Walking speed and WRs were measured using photoelectric cells and accelerometers. Part of what makes the WR appealing is that it can be measured using only a measuring tape and a stop watch, but estimating the WR like this requires the tester to be able to operate the stop watch while counting steps at the same time. While this is quite feasible, it was not the procedure that was used in our study. While we have confidence in the precision of our procedure, our results may not be directly comparable to WRs measured in a nonlaboratory setting. In summary, the WR was invariant across preferred and fast speeds, and also during an uneven surface walking condition. However, the WR increased significantly during a dual task-condition, requiring the participants to divide their attention. The WR is a simple test of gait quality and may offer a valuable addition to other clinical gait tests, such as gait speed.

Conflict of interest

CC E

The authors (B Bogen, R Moe-Nilssen, AH Ranhoff, MK Aaslund) declare that they have no conflict of interest.

Acknowledgement

A

This study was funded by the Norwegian Fund for Postgraduate Training in Physiotherapists, and the Kavli Center for Research on Aging and Dementia (Haraldsplass Diaconal Hospital). REFERENCES: 1. 2. 3.

Middleton A, Fritz SL, Lusardi M. Walking speed: the functional vital sign. J Aging Phys Act. 2015;23(2):314-22. Rantakokko M, Manty M, Rantanen T. Mobility decline in old age. Exerc Sport Sci Rev. 2013;41(1):19-25. Bridenbaugh SA, Kressig RW. Laboratory review: the role of gait analysis in seniors' mobility and fall prevention. Gerontology. 2011;57(3):256-64.

11. 12. 13.

14.

15. 16.

17. 18.

IP T

CC E

19.

SC R

10.

U

9.

N

8.

A

7.

M

6.

ED

5.

Sekiya N. NH, Ito H., Furuna T. The invariant relationship between step length and step rate during free walking. Journal of Human Movement Studies. 1996;30:241-57. Murakami R, Otaka Y. Estimated lower speed boundary at which the walk ratio constancy is broken in healthy adults. J Phys Ther Sci. 2017;29(4):722-5. Zijlstra A, de Bruin ED, Bruins N, Zijlstra W. The step length-frequency relationship in physically active community-dwelling older women. Eur J Appl Physiol. 2008;104(3):427-34. Egerton T, Danoudis M, Huxham F, Iansek R. Central gait control mechanisms and the stride length - cadence relationship. Gait Posture. 2011;34(2):178-82. Rota V, Perucca L, Simone A, Tesio L. Walk ratio (step length/cadence) as a summary index of neuromotor control of gait: application to multiple sclerosis. Int J Rehabil Res. 2011;34(3):265-9. Kalron A. Construct validity of the walk ratio as a measure of gait control in people with multiple sclerosis without mobility aids. Gait Posture. 2016;47:103-7. Thingstad P, Taraldsen K, Saltvedt I, Sletvold O, Vereijken B, Lamb SE, et al. The long-term effect of comprehensive geriatric care on gait after hip fracture: the Trondheim Hip Fracture Trial--a randomised controlled trial. Osteoporos Int. 2016;27(3):933-42. Callisaya ML, Blizzard L, McGinley JL, Srikanth VK. Risk of falls in older people during fastwalking--the TASCOG study. Gait Posture. 2012;36(3):510-5. Barak Y, Wagenaar RC, Holt KG. Gait characteristics of elderly people with a history of falls: a dynamic approach. Phys Ther. 2006;86(11):1501-10. Schwenk M, Zieschang T, Englert S, Grewal G, Najafi B, Hauer K. Improvements in gait characteristics after intensive resistance and functional training in people with dementia: a randomised controlled trial. BMC Geriatr. 2014;14:73. Ko SU, Tolea MI, Hausdorff JM, Ferrucci L. Sex-specific differences in gait patterns of healthy older adults: results from the Baltimore Longitudinal Study of Aging. J Biomech. 2011;44(10):1974-9. Moe-Nilssen R, Helbostad JL. Estimation of gait cycle characteristics by trunk accelerometry. J Biomech. 2004;37(1):121-6. Bogen B, Moe-Nilssen R, Aaslund MK. The Walk Ratio: Gender differences or height differences? Gait & Posture. 2017;57(Supplement 1, proceedings from the annual meeting of the European Society for Movement Analysis in Adults and Children (ESMAC) ):49. Menz HB, Lord SR, Fitzpatrick RC. Acceleration patterns of the head and pelvis when walking on level and irregular surfaces. Gait Posture. 2003;18(1):35-46. Woollacott M, Shumway-Cook A. Attention and the control of posture and gait: a review of an emerging area of research. Gait Posture. 2002;16(1):1-14. Nordin E, Moe-Nilssen R, Ramnemark A, Lundin-Olsson L. Changes in step-width during dualtask walking predicts falls. Gait Posture. 2010;32(1):92-7. Chui KK, Lusardi MM. Spatial and temporal parameters of self-selected and fast walking speeds in healthy community-living adults aged 72-98 years. J Geriatr Phys Ther. 2010;33(4):173-83. Ardestani MM, Ferrigno C, Moazen M, Wimmer MA. From normal to fast walking: Impact of cadence and stride length on lower extremity joint moments. Gait Posture. 2016;46:118-25. Huang Y, Meijer OG, Lin J, Bruijn SM, Wu W, Lin X, et al. The effects of stride length and stride frequency on trunk coordination in human walking. Gait Posture. 2010;31(4):444-9. Abellan van Kan G, Rolland Y, Andrieu S, Bauer J, Beauchet O, Bonnefoy M, et al. Gait speed at usual pace as a predictor of adverse outcomes in community-dwelling older people an International Academy on Nutrition and Aging (IANA) Task Force. J Nutr Health Aging. 2009;13(10):881-9.

PT

4.

20.

A

21. 22. 23.

Graham JE, Ostir GV, Kuo YF, Fisher SR, Ottenbacher KJ. Relationship between test methodology and mean velocity in timed walk tests: a review. Arch Phys Med Rehabil. 2008;89(5):865-72.

A

CC E

PT

ED

M

A

N

U

SC R

IP T

24.

Tables: Table 1: Walking speed, height-adjusted cadence, step length and WR for all walking conditions. Walking speed (cm/s)

Cadence (steps/min)

Step length (cm)

WR (cm/steps/min)

SD

Mean

SD

Mean

SD

Mean

SD

Pref. speed

117

0.20

108.22

9.21

64

0.08

0.60

0.07

Fast speed

147

0.24

122.63

10.23

71

0.08

0.58

0.07

Dual task

87

0.27

89.19

20.03

58

0.09

0.68

0.18

Uneven surface

102

0.25

101.20

11.80

60

0.09

0.59

0.07

WR = Height adjusted walk ratio

Table 2: Planned pairwise comparisons of WR during different conditions. Std. error

0.011

0.005

-0.087*

0.020

0.006

0.004

Dual task

-0.098*

Uneven surface

-0.005

Fast walking speed Dual task

A

CC E

PT

ED

M

Based on estimated marginal means. * The mean difference is significant at the 0.05 level. a. Bonferroni adjustment for multiple comparisons. WR = Height adjusted walk ratio.

A

Fast walking speed vs.

Lower bound Upper bound

0.13

-0.002

<0.001

-0.140

1.00

-0.006

0.018

0.021

<0.001

-0.154

-0.041

0.006

1.00

-0.020

0.010

N

Uneven surface

p-valuea

U

Preferred walking speed vs.

95% CI difference

SC R

Mean difference

Conditions

IP T

Mean

0.024 -0,033