- Email: [email protected]
Local dynamic stability during long-fatiguing walks in people with multiple sclerosis
Gait & Posture 76 (2020) 122–127
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Full length article
Local dynamic stability during long-fatiguing walks in people with multiple sclerosis
T
I. Arpana,b,*, P.C. Finoc, B.W. Flingd, F. Horaka a
Department of Neurology, Oregon Health & Science University, Portland, OR, United States Advanced Imaging Center, Oregon Health & Science University, Portland, OR, United States c Department of Health, Kinesiology, & Recreation, University of Utah, Salt Lake City, UT, United States d Department of Health and Exercise Science, Colorado State University, Fort Collins, CO, United States b
A R T I C LE I N FO
A B S T R A C T
Keywords: Multiple sclerosis Six-minute walk test Motor fatigue Local dynamic stability
Background: Altered balance/stability during walking is common in people with multiple sclerosis (PwMS). While dynamic gait stability has been related to falling and localized muscle fatigue, it has rarely been studied in MS. Specifically, the effects of walking-related fatigue on dynamic stability are unclear in PwMS. Research questions: 1) Are temporal changes in dynamic stability during long-walks different among PwMS and healthy controls (HC)? 2) Is there a relationship between stability and walking performance changes in PwMS? Methods: Twenty-five PwMS and ten HC participated in the six-minute walk test (6MWT) wearing six-wireless inertial sensors. Local dynamic stability (LDS) during gait was quantified by maximum-finite-time Lyapunov exponents (λS), where larger λS indicates less stable dynamics. Linear mixed models were fit to compare changes in LDS and walking performance over time among two groups. Additionally, the percent changes in λS and distance from minute 1 to 6 were recorded as Dynamic Stability Index (DSI6-1) and Distance-Walked Index (DWI6-1) respectively. Finally, Pearson correlation compared the association between DSI6-1 and DWI6-1. Results: A significant group*time interaction was found for LDS. PwMS did not have different LDS than HC until minute-4 of walking, and differences persisted at minute-6. Further, PwMS walked significantly shorter distances and demonstrated a greater decline in walking performance (DWI6-1) during the 6MWT. Finally, DSI6-1 and DWI6-1 were significantly correlated in PwMS. Significance The dynamic stability differences among PwMS and HC were only apparent after 3-minutes of walking and ∼60% of PwMS became less stable over time, supporting the use of long walks in MS to capture stability changes during the motor task performance. A significant relationship between the decline in stability and poor walking performance over time during the 6MWT suggested a possible role of walking-related fatigue in the worsening of balance during long walks in PwMS.
1. Introduction Walking is critical to the quality of life and independence, and reduced mobility in people with neurological disorders is an independent risk factor for morbidity and mortality [1,2]. In multiple sclerosis (MS), walking is impacted from the beginning of the disease, even in the absence of clinical signs of pyramidal dysfunction [3]. In people with MS (PwMS), walking patterns change due to declines in sensorimotor, visual and cognitive processes associated with reduced muscle activation, weakness, spasticity, impaired balance, and poor coordination [4]. Fall rates in MS (∼50 %) [5] are even higher than the often researched older adult population (30 %) [6]. Therefore, comprehensive investigations of different aspects of walking impairments are critical in PwMS to help them walk comfortably and safely.
⁎
Local dynamic stability (LDS) is an approach to evaluate the ability of the locomotor system to accommodate infinitesimally small perturbations that occur naturally during walking [7]. Specifically, LDS reflects the rate at which variability is controlled by the locomotor system. Poor LDS is an early predictor of fall risk in the elderly population [8]. Since maintaining stability is a complex task requiring the integration of sensorimotor resources [9,10], the prevalence of sensory disturbances may pose a serious threat of falling while walking in PwMS. Previous studies addressing the sensorimotor control of stability in MS primarily focused on balance assessments during rather static conditions (postural sway) [11,12], which may not represent dynamic daily-routines. Furthermore, demyelination in MS disrupts the transport of signals along the nerve pathways [13], and these neurodegenerative changes may slow down an individual’s response to perturbations
Corresponding author at: Balance Disorders Laboratory, Oregon Health & Science University, Portland, OR, United States. E-mail address: [email protected] (I. Arpan).
https://doi.org/10.1016/j.gaitpost.2019.10.032 Received 13 February 2019; Received in revised form 18 September 2019; Accepted 23 October 2019 0966-6362/ © 2019 Elsevier B.V. All rights reserved.
Gait & Posture 76 (2020) 122–127
I. Arpan, et al.
speed, aiming to cover as much distance as possible,” [17] wearing sixwireless inertial sensors (Opal, v1, APDM Inc., Portland Oregon). One sensor was positioned on the low back, two on the feet, one on the sternum, and two on the wrists. The sensors transmitted raw data at 128 Hz wirelessly to the laptop data collection using the MobilityLab. The primary outcome measure of the study, dynamic stability, was assessed using LDS, which reflects the rate at which an individual controls infinitesimally small internal or external perturbations. Raw 3D accelerometer and gyroscope data were extracted from the sternum, lumbar spine, and feet inertial sensors. For each walk test, data were parsed into non-overlapping minute intervals. Each minute of gait was then analyzed using the same procedure. First, bouts of straight gait were obtained by removing turns using an angular-velocity based threshold [22] and removing one stride immediately preceding and following each turn. Second, heel contact, toe-off, and mid-swing events were detected based on the frequency content of the foot acceleration and angular velocity signals after implementing a continuous wavelet transformation, as described by Cain et al. [23]. Each straight walking bout was then divided into non-overlapping segments of 7 consecutive, straight walking strides, with each bout time-normalized to 7 × 130 points to maintain equal data-length across segments. If a walking bout did not include at least 7 straight strides, it was excluded from the remainder of the analysis. LDS was then estimated for each time-normalized walking segment of 7 strides following Kantz’s algorithm [24] and previous reports for estimating LDS over short bouts of gait [25,26]. Briefly, a 9D state space was constructed from the threedimensional trunk accelerations and their twice time-delayed copies using a fixed time delay of 25 % of the average stride time. The average distance to the two nearest neighbors of the trajectory was tracked for one stride for each point. Mean log divergence curves were created by mapping the average distance across all points as a function of the percentage of a normalized stride. For each walking segment of 7 strides, the short-term maximum finite-time Lyapunov exponent, λS, was estimated from the slope of the mean log divergence curves from 0 to 0.5 strides, where larger λS values indicate less stability. To obtain the LDS estimate for each minute of gait, the median of λS for each minute was retained for statistical analysis. In addition, participants were videotaped during walking, and distances walked during each minute and total distance were recorded by two independent reviewers. The decline in walking performance was documented by the distance walked index (DWI), which was calculated using the following equation [18]:
during walking. A very recent study found that LDS is associated with prospective falls in PwMS [14]. Since LDS provides assessments of stability and fall risk that are not available through standard measures such as gait speed, understanding LDS during walking in PwMS may provide unique insights into the development of effective gait training and rehabilitation interventions for fall prevention. Fatigue is another critical factor in MS that can directly impact the rate of sensorimotor control of movement [15]. Approximately 70–90% of PwMS experience difficulties in initiating and/or sustaining physical activities (motor fatigue) during daily life [16]. Recently the six-minute walk test (6MWT), a reliable measure of motor function in MS, has been used to demonstrate motor fatigue [17,18]. PwMS not only have shorter 6MWT distances than healthy controls (HC) but also more than onethird exhibit a decline in walking distance during the 6MWT [18]. It has been suggested that diffuse demyelination and axonal loss in PwMS results in a less central motor drive for sustained muscle contractions and coordination, leading to walking-related motor fatigue [18]. Importantly, motor fatigue can worsen walking stability (LDS), as shown in healthy individuals[19], increasing their risk of falls. However, despite the well-documented stability impairments and high prevalence of motor fatigue in MS, none of the previous studies have looked into the effects of walking-related fatigue on dynamic stability in PwMS. Identification of the stability changes in patient’s physical movement like walking as a function of fatigue, might result in early prediction of falls during activities of daily living, tailored rehabilitation programs, and better management for PwMS. The primary purpose of this study was to characterize LDS in HC and PwMS during the long walking task. Specifically, we investigated minute by minute changes in LDS during the 6MWT. We hypothesized that there will be a significant difference in the temporal changes in stability during the 6MWT between PwMS and HC. Further, we investigated the association between stability changes and change in walking performance during the 6MWT as a secondary analysis. We hypothesized that there would be a significant relationship between the decline in walking distance and the decline in LDS over the duration of the 6MWT in PwMS. 2. Methods 2.1. Study design This cross-sectional research protocol complied with the principles of the Declaration of Helsinki and was reviewed and approved by the institutional review board of the Oregon Health & Science University (OHSU). All participants provided informed consent prior to participation.
DWI 6-1: {(Distance covered in 6th min- Distance covered in 1st min)/ (Distance covered in 1st min)} * 100
2.2. Participants
3. Statistical analysis
PwMS and HC were recruited on a convenience basis from the OHSU MS Clinic and the local community. Inclusion criteria were ages 18–65 years, an absence of any orthopedic or neurologic problems other than MS, and the ability to walk for 6-minutes without an assistive device. Exclusion criteria were MS exacerbation or the use of corticosteroids within 30 days of screening. Participants were instructed not to take caffeine in the morning of the testing and all testings were done between 10 am-noon. Additionally, PwMS were told not to take fatigue-related medication for 24 h before testing. Demographic and clinical characteristics, including age, gender, body mass index, disease duration, MS phenotype, self-rated Expanded Disability Status Scale (EDSS) [20] and Multiple Sclerosis Impact Scale (MSIS-29) [21] were collected.
Independent t-tests were used to assess group differences in the demographic features between HC and PwMS. To assess whether LDS and walking performance changed differently over time in PwMS compared to HC, linear mixed models were fit. Fixed effects of group, time, and the group*time interaction were included in the models, and random intercepts and slopes over time were included to account for within-subject correlations. The HC was the reference group in all models and time was modeled as continuous. The group*time interaction was the primary effect of interest. To explore whether the group*time interaction remained after adjusting for other covariates, age, MFIS total score, and gait speed were added as covariates, and retained if significant at a p < 0.05 level. In addition, the percent change in LDS (DSI6-1) from first to last minute was calculated. This calculation matched the formula used to compute DWI6-1, which has been traditionally used to assess motor fatigue in PwMS [18]. Pearson correlation coefficients compared whether the percent change from first to last minute in LDS (DSI6-1) was
2.3. Study outcomes Participants were instructed to complete the 6MWT “at their fastest 123
Gait & Posture 76 (2020) 122–127
I. Arpan, et al.
Table 1 Demographic characteristics of the study population. Data is displayed as mean ± SE except for self-rated EDSS*, which is presented as median (range). Healthy controls
People with MS
10 47.6 ± 3.3 yrs 6F/4M 173.5 ± 4.6 83.7 ± 7.9 13.9 ± 4.2 5.7 ± 2.0 7.1 ± 2.0 1.1 ± 0.4
25 51.1 ± 1.6 yrs 20 F/5M 171.9 ± 2.1 85.5 ± 4.0 39.7 ± 3.9 18.2 ± 1.7 18.4 ± 2.1 3.1 ± 0.5
Relapsing Remitting Primary Progressive Secondary Progressive
NA NA NA NA NA
21 2 2 14.4 ± 1.4 yrs 3.5 (2.5-5.0)
Physical impact score Psychological impact score
NA NA 585 ± 27 m
27.7 ± 3.8 30.2 ± 3.4 452 ± 17 m
Count Age Gender (F/M Height (cm) Weight (kg) Total MFIS score Physical fatigue Cognitive fatigue Psychological fatigue
p value
0.3 0.7 0.8 < 0.01 < 0.01 < 0.01 < 0.05
Disease type
Disease duration Self-rated EDSS* MSIS-29 scores
6MWT distance
< 0.01
associated with changes in walking performance (DWI6-1). All statistical analysis was performed in MATLAB (r2017b, Mathworks, Inc., Natick, MA, USA) using a 0.05 significance level. 4. Results 4.1. Demographics Twenty-five PwMS and ten HC were included in the study. There was no difference in the age, height, and weight of PwMS and HC (Table 1). Disease severity in the MS group was characterized by two different self-reported questionnaires: self-rated EDSS and MSIS-29. A moderately significant association (r = 0.51, p < 0.01) was found for the disability scores obtained from the EDSS and the physical impact sub-scale of MSIS-29. 4.2. Dynamic stability There was no main effect of group, but a significant group*time interaction was found for LDS (Table 2). The group*time interaction remained significant (p = 0.011) after adjusting for covariates of age, gait speed, and total MFIS score. Covariates of age (p = 0.447), gait speed (p = 0.608), and total MFIS score (p = 0.299) did not have a significant effect on LDS. Fully adjusted model results are presented in the supplemental material. Post-hoc analysis revealed that LDS did not differ between PwMS and HC at the beginning of the 6MWT, but healthy controls became more stable (decrease in λS) over time, relative
Fig. 1. Median Local dynamic instability estimate, λS, for each participant at each minute block of time. Thick lines indicate group means. Thin lines depict individual subject data. Higher values of λS indicate less stability.
to PwMS (Fig. 1). Descriptively, all (but one) HC demonstrated an increase in the LDS (decline in λS) from first to last minute (DSI6-1: −4 to −20 %, without outlier). In contrast, changes in LDS from minute 1 to minute 6 were more heterogeneous in the MS group. About 8 PwMS exhibited an improvement in LDS (DSI6-1: −1 to −21 %), 3 with no change, 5 PwMS showed < 10 % decrease in LDS (DSI6-1: 4–5%), and 9 showed > 10 % decline in LDS (DSI6-1: 7–43%).
Table 2 Summary results of unadjusted linear mixed effects models for local dynamic stability (LDS) and distance walked. Bolded p values indicate significance at a 0.05 level. Beta Estimate Local Dynamic Stability Intercept Time Group Group*Time Distance Walked Intercept Time Group Group*Time
SE
p value
Lower 95% CI
Upper 95% CI
0.443 −0.008 −0.017 0.012
0.022 0.004 0.026 0.005
< 0.001 0.035 0.514 0.009
0.400 −0.016 −0.068 0.003
0.486 −0.001 0.034 0.021
94.7 0.4 −17.6 −0.8
4.4 0.3 5.2 0.3
< 0.001 0.197 < 0.001 0.017
86.0 −0.2 −27.9 −1.5
103.3 0.9 −7.4 −0.1
4.3. Walking distance Due to technical issues during the video recording of the 6MWT in 3 PwMS, walking distance for each minute could only be obtained in 22 PwMS. Both group effect and group*time interaction were significant, indicating that there were significant differences in walking distance between two groups and over time (Table 2). The group (p = 0.043) and group*time interaction (p = 0.016) remained significant after adjusting for covariates of age, gait speed, and total MFIS score. Covariates of age (p = 0.786) and total MFIS score (p = 0.439) did not have a significant effect on walking distance, but gait speed had a significant, positive effect on walking distance (p < 0.001). Fully adjusted model 124
Gait & Posture 76 (2020) 122–127
I. Arpan, et al.
Fig. 2. A. Walking distance covered by people with MS (PwMS) and Healthy Control (HC) groups during each minute of walking during the 6MWT. Thick lines indicate unadjusted group means; thin lines indicate individual subject data. B. Comparison of change in walking distance from minute 1 to minute 6 (DWI6-1) of the 6MWT among people with MS (PwMS) and Healthy Controls (HC). Greater the decline in walking distance from minute 1 to minute 6, the more likely the participants were subjected to walking-related motor fatigue. Data is displayed as mean ± SE for each group.
[27,28]. The unique aspect of this study was the quantification of changes in LDS at one-minute intervals during the 6MWT task in PwMS and HC. Though group* time results showed that, on a group level, PwMS became less stable over time compared to HC, individual participants’ trajectories over time were heterogeneous. Unlike HC, PwMS did not adapt their walking pattern to conserve LDS during the 6MWT, and about 60 % of PwMS became less stable over time. More importantly, 9 PwMS demonstrated > 10 % change in LDS, which has been estimated as a clinically significant difference for LDS between fall-prone and healthy elderly [29], vestibular sensory disruption [25], and concussion [26]. These findings suggest that PwMS may be at a higher risk of falls during long, sustained activities due to a decline in stability over time. Consistent with this hypothesis, at least three different studies have identified LDS as a pertinent marker for retrospective and future falls in MS population [14,30,31]. In contrast to our finding of no differences in stability among PwMS and HC for the first few minutes of the 6MWT (about 250 m), a previous study by Huisinga et al. reported worse average trunk stability in PwMS during a short walk of 30 s at a self-selected pace [32]. These differences can most likely be attributed to walking speed and analysis methods. In our study, participants were instructed to walk at a faster than normal speed and to maintain that speed over the whole sixminutes [17], while subjects in the previous study walked at their selfselected pace. Secondly, differences in sampling frequency and methods to calculate LDS likely also contributed to the discrepancy across studies. Here, we used methods that have been consistently used to estimate short bouts of gait in clinical tests [8,25,26]. Due to these methodological differences, our results demonstrating no immediate differences between PwMS and HC when walking fast cannot easily be compared to results from Huisinga et al. [32]. Furthermore, a very recent study by Tajali and colleagues found that LDS was able to predict future falls in a cohort of PwMS using two minutes of normal pace treadmill walking [14]. However, the subjects were instructed to walk for five minutes prior to the testing procedure to familiarize themselves with the treadmill and to select the appropriate belt speed. Our results complement this procedural decision and encourage future studies investigating the optimal duration of walking assessments in PwMS. Another key insight from the study was the investigation of stability changes over the 6MWT in PwMS and its relationship to changes in the walking performance. The modified 6MWT has been previously used in multiple MS studies to characterize the occurrence of motor fatigue [17,18]. The decline in performance during the 6MWT in PwMS can be attributed to both peripheral and central mechanisms [33,34]. Specifically, the inflammatory processes in MS result in demyelination or partial impairment of neural pathways; while neuronal functions may be partly preserved, axonal demyelination can slow down the neural conduction resulting in fatigue especially under high demand or highly repetitive requirements. In addition to affecting the rate of performance of a motor task, fatigue is also known to diminish the precision of motor control [19]. Since LDS quantifies the rate of divergence of movement trajectories, we speculate that the decline in LDS (i.e., faster divergence
results are presented in the supplemental material. PwMS walked significantly shorter distances compared to HC at each minute of the 6MWT (Fig. 2A). Further, compared to HC, PwMS demonstrated a greater decline in walking distance from minute 1 to minute 6 of the 6MWT (DWI6-1; Fig. 2B), potentially indicating a higher incidence of walking-related fatigue in the MS group. 4.4. Relationship between changes in dynamic stability and walking performance from minute 1 to minute 6 A significant relationship (r= -0.46, p = 0.03; Fig. 3) was observed between the change in dynamic stability (DSI6-1) and the change in distance (DWI6-1) from minute 1 to minute 6 of the 6MWT in PwMS, suggesting that a decline in dynamic stability was associated with a decline in walking performance due to motor fatigue in MS. 5. Discussion The primary aim of this study was to characterize minute-by-minute changes in dynamic stability during the 6MWT in PwMS and HC. The key finding was that PwMS did not have different stability than HC until minute 4 of the 6MWT and differences in LDS persisted at minute 6. This finding supports the use of long walks in MS as shorter walking tests may not capture changes in dynamic stability over time. In addition, we found a significant relationship between the decline in stability and decline in walking performance during the 6MWT in the MS group, suggesting that walking-related fatigue may be one of the contributing factors to worsening of balance control. Walking impairments are well documented in PwMS. PwMS are known to walk slower, with fewer, shorter steps and spend a greater percentage of a gait cycle on double support time compared to HC
Fig. 3. Correlation between the percent change in local dynamic stability (DSI6and the percent change in walking distance (DWI6-1) from minute 1 to minute 6 of the 6MWT in people with MS (PwMS). 1)
125
Gait & Posture 76 (2020) 122–127
I. Arpan, et al.
inappropriately influenced this work. The research in this proposal employs the technology of APDM, a company in which author, FH, has financial interest. This potential conflict has been reviewed and managed by OHSU.
and slower control) may be an indication of the slowing of motor unit firing patterns in fatigued PwMS. The fact that decline in walking performance in this study was significantly related to the decline in LDS supports the hypothesis that there is a diminished central reserve for sustained muscle contractions and coordination during the long walking tasks in PwMS. However, it is unlikely that a single mechanism is responsible for the trends that we observed over time. Even within our cohort of MS participants, we saw between-subject variation in the trends of LDS over time – it is possible there were mechanistic subgroups within our MS cohort. Future work should further explore the mechanistic causes of the decline in motor task performance and its potential relationships with LDS in a larger sample of PwMS. Our findings have important implications for rehabilitation in today’s health care environment. Falls can result in severe injuries resulting in loss of independence, institutionalization and even death. As most falls occur during walking, assessing fall-risk based on gait parameters may be a promising approach. In addition, traditional fatigue and balance interventions programs should include gait stability retraining for fall prevention in patients that demonstrate decline in stability during the motor task performance, so that PwMS do not have to prioritize fatigue management over LDS or vice versa while long continuous activities. And finally, this study supports the use of long walks by clinicians and researchers to study stability changes during walking tasks in PwMS. Though 6-minute bout of continuous walking is uncommon, it is equally important to test and train PwMS to confidently cover large distances during activities such as recreation, shopping, etc. without experiencing a decline in physical resources.
Declaration of Competing Interest The authors (IA, PCF, & BWF) declare that there are no financial or personal relation with people or organizations that have inappropriately influenced this work. The research in this proposal employs the technology of APDM, a company in which author, FH, has financial interest. This potential conflict has been reviewed and managed by OHSU. Acknowledgements The authors would like to thank research assistants, Spencer Smith, BS and Maddy Dunn, BS, for their support in data collection. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.gaitpost.2019.10.032. References [1] J.A. Knight, Physical inactivity: associated diseases and disorders, Ann. Clin. Lab. Sci. 42 (2012) 320–337. [2] R.W. Motl, P.L. Sandroff, BM, The importance of physical fitness in multiple sclerosis, J. Nov. Physiother. 3 (2013) 7, https://doi.org/10.4172/2165-7025. 1000141. [3] C.L. Martin, et al., Gait and balance impairment in early multiple sclerosis in the absence of clinical disability, Mult. Sclerosis (Houndmills, Basingstoke, England) 12 (2006) 620–628, https://doi.org/10.1177/1352458506070658. [4] M.H. Cameron, J.M. Wagner, Gait abnormalities in multiple sclerosis: pathogenesis, evaluation, and advances in treatment, Curr. Neurol. Neurosci. Rep. 11 (2011) 507–515, https://doi.org/10.1007/s11910-011-0214-y. [5] S. Coote, J.J. Sosnoff, H. Gunn, Fall incidence as the primary outcome in multiple sclerosis falls-prevention trials: recommendation from the international MS falls prevention research network, Int. J. MS Care 16 (2014) 178–184, https://doi.org/ 10.7224/1537-2073.2014-059. [6] M. McCarthy, Falls are leading cause of injury deaths among older people, US study finds, BMJ 354 (2016) i5190, https://doi.org/10.1136/bmj.i5190. [7] J.B. Dingwell, J.P. Cusumano, P.R. Cavanagh, D. Sternad, Local dynamic stability versus kinematic variability of continuous overground and treadmill walking, J. Biomech. Eng. 123 (2001) 27–32. [8] K.S. van Schooten, et al., Ambulatory fall-risk assessment: amount and quality of daily-life gait predict falls in older adults, J. Gerontol. Ser. A 70 (2015) 608–615, https://doi.org/10.1093/gerona/glu225. [9] J.G. Nutt, F.B. Horak, B.R. Bloem, Milestones in gait, balance, and falling, Mov. Disord. 26 (2011) 1166–1174, https://doi.org/10.1002/mds.23588. [10] K. Takakusaki, Functional neuroanatomy for posture and gait control, J. Mov. Disord. 10 (2017) 1–17, https://doi.org/10.14802/jmd.16062. [11] J. McLoughlin, C. Barr, M. Crotty, S.R. Lord, D.L. Sturnieks, Association of postural sway with disability status and cerebellar dysfunction in people with multiple sclerosis: a preliminary study, Int. J. MS Care 17 (2015) 146–151, https://doi.org/ 10.7224/1537-2073.2014-003. [12] A.J. Solomon, J.V. Jacobs, K.V. Lomond, S.M. Henry, Detection of postural sway abnormalities by wireless inertial sensors in minimally disabled patients with multiple sclerosis: a case-control study, J. Neuroeng. Rehabil. 12 (2015) 74, https:// doi.org/10.1186/s12984-015-0066-9. [13] A. Alizadeh, S.M. Dyck, S. Karimi-Abdolrezaee, Myelin damage and repair in pathologic CNS: challenges and prospects, Front. Mol. Neurosci. 8 (2015), https://doi. org/10.3389/fnmol.2015.00035 35-35. [14] S. Tajali, et al., Impaired local dynamic stability during treadmill walking predicts future falls in patients with multiple sclerosis: a prospective cohort study, Clin. Biomech. (2019), https://doi.org/10.1016/j.clinbiomech.2019.05.013. [15] I. Cogliati Dezza, et al., Functional and structural balances of homologous sensorimotor regions in multiple sclerosis fatigue, J. Neurol. 262 (2015) 614–622, https:// doi.org/10.1007/s00415-014-7590-6. [16] S.R. Schwid, et al., Quantitative assessment of motor fatigue and strength in MS, Neurology 53 (1999), https://doi.org/10.1212/wnl.53.4.743 743-743. [17] M.D. Goldman, R.A. Marrie, J.A. Cohen, Evaluation of the six-minute walk in multiple sclerosis subjects and healthy controls, Mult. Scler. 14 (2008) 383–390, https://doi.org/10.1177/1352458507082607. [18] C. Leone, et al., Prevalence of walking-related motor fatigue in persons with multiple sclerosis: decline in walking distance induced by the 6-Minute walk test, Neurorehabil. Neural Repair 30 (2016) 373–383, https://doi.org/10.1177/
6. Limitations One of the major limitations of the study was the small sample size of both HC and PwMS. Therefore, findings from this exploratory analysis should be rigorously tested and replicated in future studies. In addition to the issues surrounding the lower sample size, another limitation was the lack of falls data. Since not everybody in the MS group demonstrated the decline in dynamic stability during walking, it would have been interesting to see if PwMS, who maintained LDS during the duration of the 6MWT, were at a lower risk of falling. Further, by separating the 6MWT into one-minute bouts, we retained only an average of 5 bouts of straight gait in each minute. This small number of gait bouts within each minute likely increased the minute-to-minute variability in our LDS estimates [35]. However, we repeated our analysis using two-minute long segments and obtained similar results – between-group differences were not present initially, but emerged after at least two minutes of walking. These results using two-minute long segments are presented in the supplemental material. While our findings were robust, future studies using different segmentations of the 6MWT should confirm our results. 7. Conclusions PwMS decreased stability over time relative to controls during longwalking tasks. Further, the decline in stability and motor fatigue were related to each other, suggesting a possible role of fatigue in worsening of walking balance in PwMS. Still, the underlying mechanisms are not completely resolved and further research incorporating measurements of cortical gait control and its relationship to worsening of the motor task performance might be promising. Funding This research was supported by the Collins Medical Trust (Portland, OR), the Medical Research Foundation (Portland, OR), and the National MS Society’s Mentored Postdoctoral Fellowship Award. The authors (IA, PCF, & BWF) declare that there are no financial or personal relation with people or organizations that have 126
Gait & Posture 76 (2020) 122–127
I. Arpan, et al.
1016/j.gaitpost.2012.10.012. [28] J.J. Sosnoff, B.M. Sandroff, R.W. Motl, Quantifying gait abnormalities in persons with multiple sclerosis with minimal disability, Gait Posture 36 (2012) 154–156, https://doi.org/10.1016/j.gaitpost.2011.11.027. [29] T.E. Lockhart, J. Liu, Differentiating fall-prone and healthy adults using local dynamic stability, Ergonomics 51 (2008) 1860–1872, https://doi.org/10.1080/ 00140130802567079. [30] R. Hilfiker, et al., Local dynamic stability as a responsive index for the evaluation of rehabilitation effect on fall risk in patients with multiple sclerosis: a longitudinal study, BMC Res. Notes 6 (2013), https://doi.org/10.1186/1756-0500-6-260 260260. [31] A.T. Peebles, A.P. Bruetsch, S.G. Lynch, J.M. Huisinga, Dynamic balance is related to physiological impairments in persons with multiple sclerosis, Arch. Phys. Med. Rehabil. 99 (2018) 2030–2037, https://doi.org/10.1016/j.apmr.2017.11.010. [32] J.M. Huisinga, M. Mancini, R.J. St George, F.B. Horak, Accelerometry reveals differences in gait variability between patients with multiple sclerosis and healthy controls, Ann. Biomed. Eng. 41 (2013) 1670–1679, https://doi.org/10.1007/ s10439-012-0697-y. [33] K.R. Sharma, J. Kent-Braun, M.A. Mynhier, M.W. Weiner, R.G. Miller, Evidence of an abnormal intramuscular component of fatigue in multiple sclerosis, Muscle Nerve 18 (1995) 1403–1411, https://doi.org/10.1002/mus.880181210. [34] A.T. White, J.N. Lee, A.R. Light, K.C. Light, Brain activation in multiple sclerosis: a BOLD fMRI study of the effects of fatiguing hand exercise, Mult. Scler. 15 (2009) 580–586, https://doi.org/10.1177/1352458508100034. [35] K.S. van Schooten, S.M. Rispens, P.J. Elders, J.H. van Dieen, M. Pijnappels, Toward ambulatory balance assessment: estimating variability and stability from short bouts of gait, Gait Posture 39 (2014) 695–699.
1545968315597070. [19] P.-C. Kao, M.A. Pierro, K. Booras, Effects of motor fatigue on walking stability and variability during concurrent cognitive challenges, PLoS One 13 (2018), https:// doi.org/10.1371/journal.pone.0201433 e0201433-e0201433. [20] J. Bowen, L. Gibbons, A. Gianas, G.H. Kraft, Self-administered Expanded Disability Status Scale with functional system scores correlates well with a physician-administered test, Mult. Sclerosis (Houndmills, Basingstoke, England) 7 (2001) 201–206, https://doi.org/10.1177/135245850100700311. [21] J. Hobart, D. Lamping, R. Fitzpatrick, A. Riazi, A. Thompson, The Multiple Sclerosis Impact Scale (MSIS-29): a new patient-based outcome measure, Brain 124 (2001) 962–973, https://doi.org/10.1093/brain/124.5.962. [22] M. El-Gohary, et al., Continuous monitoring of turning in patients with movement disability, Sensors (Basel, Switzerland) 14 (2013) 356–369, https://doi.org/10. 3390/s140100356. [23] S.M. Cain, M.V. Porter, L. Ojeda, N.C. Perkins, Accurate and robust gait event detection using foot-mounted inertial measurement units, Proceedings of the American Society of Biomechanics, (2019). [24] H. Kantz, A robust method to estimate the maximal Lyapunov exponent of a time series, Phys. Lett. A 185 (1994) 77–87. [25] L.H. Sloot, et al., Sensitivity of local dynamic stability of over-ground walking to balance impairment due to galvanic vestibular stimulation, Ann. Biomed. Eng. 39 (2011) 1563–1569, https://doi.org/10.1007/s10439-010-0240-y. [26] P.C. Fino, A preliminary study of longitudinal differences in local dynamic stability between recently concussed and healthy athletes during single and dual-task gait, J. Biomech. 49 (2016) 1983–1988, https://doi.org/10.1016/j.jbiomech.2016.05.004. [27] M.J. Socie, R.W. Motl, J.H. Pula, B.M. Sandroff, J.J. Sosnoff, Gait variability and disability in multiple sclerosis, Gait Posture 38 (2013) 51–55, https://doi.org/10.
127
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Full length article
Local dynamic stability during long-fatiguing walks in people with multiple sclerosis
T
I. Arpana,b,*, P.C. Finoc, B.W. Flingd, F. Horaka a
Department of Neurology, Oregon Health & Science University, Portland, OR, United States Advanced Imaging Center, Oregon Health & Science University, Portland, OR, United States c Department of Health, Kinesiology, & Recreation, University of Utah, Salt Lake City, UT, United States d Department of Health and Exercise Science, Colorado State University, Fort Collins, CO, United States b
A R T I C LE I N FO
A B S T R A C T
Keywords: Multiple sclerosis Six-minute walk test Motor fatigue Local dynamic stability
Background: Altered balance/stability during walking is common in people with multiple sclerosis (PwMS). While dynamic gait stability has been related to falling and localized muscle fatigue, it has rarely been studied in MS. Specifically, the effects of walking-related fatigue on dynamic stability are unclear in PwMS. Research questions: 1) Are temporal changes in dynamic stability during long-walks different among PwMS and healthy controls (HC)? 2) Is there a relationship between stability and walking performance changes in PwMS? Methods: Twenty-five PwMS and ten HC participated in the six-minute walk test (6MWT) wearing six-wireless inertial sensors. Local dynamic stability (LDS) during gait was quantified by maximum-finite-time Lyapunov exponents (λS), where larger λS indicates less stable dynamics. Linear mixed models were fit to compare changes in LDS and walking performance over time among two groups. Additionally, the percent changes in λS and distance from minute 1 to 6 were recorded as Dynamic Stability Index (DSI6-1) and Distance-Walked Index (DWI6-1) respectively. Finally, Pearson correlation compared the association between DSI6-1 and DWI6-1. Results: A significant group*time interaction was found for LDS. PwMS did not have different LDS than HC until minute-4 of walking, and differences persisted at minute-6. Further, PwMS walked significantly shorter distances and demonstrated a greater decline in walking performance (DWI6-1) during the 6MWT. Finally, DSI6-1 and DWI6-1 were significantly correlated in PwMS. Significance The dynamic stability differences among PwMS and HC were only apparent after 3-minutes of walking and ∼60% of PwMS became less stable over time, supporting the use of long walks in MS to capture stability changes during the motor task performance. A significant relationship between the decline in stability and poor walking performance over time during the 6MWT suggested a possible role of walking-related fatigue in the worsening of balance during long walks in PwMS.
1. Introduction Walking is critical to the quality of life and independence, and reduced mobility in people with neurological disorders is an independent risk factor for morbidity and mortality [1,2]. In multiple sclerosis (MS), walking is impacted from the beginning of the disease, even in the absence of clinical signs of pyramidal dysfunction [3]. In people with MS (PwMS), walking patterns change due to declines in sensorimotor, visual and cognitive processes associated with reduced muscle activation, weakness, spasticity, impaired balance, and poor coordination [4]. Fall rates in MS (∼50 %) [5] are even higher than the often researched older adult population (30 %) [6]. Therefore, comprehensive investigations of different aspects of walking impairments are critical in PwMS to help them walk comfortably and safely.
⁎
Local dynamic stability (LDS) is an approach to evaluate the ability of the locomotor system to accommodate infinitesimally small perturbations that occur naturally during walking [7]. Specifically, LDS reflects the rate at which variability is controlled by the locomotor system. Poor LDS is an early predictor of fall risk in the elderly population [8]. Since maintaining stability is a complex task requiring the integration of sensorimotor resources [9,10], the prevalence of sensory disturbances may pose a serious threat of falling while walking in PwMS. Previous studies addressing the sensorimotor control of stability in MS primarily focused on balance assessments during rather static conditions (postural sway) [11,12], which may not represent dynamic daily-routines. Furthermore, demyelination in MS disrupts the transport of signals along the nerve pathways [13], and these neurodegenerative changes may slow down an individual’s response to perturbations
Corresponding author at: Balance Disorders Laboratory, Oregon Health & Science University, Portland, OR, United States. E-mail address: [email protected] (I. Arpan).
https://doi.org/10.1016/j.gaitpost.2019.10.032 Received 13 February 2019; Received in revised form 18 September 2019; Accepted 23 October 2019 0966-6362/ © 2019 Elsevier B.V. All rights reserved.
Gait & Posture 76 (2020) 122–127
I. Arpan, et al.
speed, aiming to cover as much distance as possible,” [17] wearing sixwireless inertial sensors (Opal, v1, APDM Inc., Portland Oregon). One sensor was positioned on the low back, two on the feet, one on the sternum, and two on the wrists. The sensors transmitted raw data at 128 Hz wirelessly to the laptop data collection using the MobilityLab. The primary outcome measure of the study, dynamic stability, was assessed using LDS, which reflects the rate at which an individual controls infinitesimally small internal or external perturbations. Raw 3D accelerometer and gyroscope data were extracted from the sternum, lumbar spine, and feet inertial sensors. For each walk test, data were parsed into non-overlapping minute intervals. Each minute of gait was then analyzed using the same procedure. First, bouts of straight gait were obtained by removing turns using an angular-velocity based threshold [22] and removing one stride immediately preceding and following each turn. Second, heel contact, toe-off, and mid-swing events were detected based on the frequency content of the foot acceleration and angular velocity signals after implementing a continuous wavelet transformation, as described by Cain et al. [23]. Each straight walking bout was then divided into non-overlapping segments of 7 consecutive, straight walking strides, with each bout time-normalized to 7 × 130 points to maintain equal data-length across segments. If a walking bout did not include at least 7 straight strides, it was excluded from the remainder of the analysis. LDS was then estimated for each time-normalized walking segment of 7 strides following Kantz’s algorithm [24] and previous reports for estimating LDS over short bouts of gait [25,26]. Briefly, a 9D state space was constructed from the threedimensional trunk accelerations and their twice time-delayed copies using a fixed time delay of 25 % of the average stride time. The average distance to the two nearest neighbors of the trajectory was tracked for one stride for each point. Mean log divergence curves were created by mapping the average distance across all points as a function of the percentage of a normalized stride. For each walking segment of 7 strides, the short-term maximum finite-time Lyapunov exponent, λS, was estimated from the slope of the mean log divergence curves from 0 to 0.5 strides, where larger λS values indicate less stability. To obtain the LDS estimate for each minute of gait, the median of λS for each minute was retained for statistical analysis. In addition, participants were videotaped during walking, and distances walked during each minute and total distance were recorded by two independent reviewers. The decline in walking performance was documented by the distance walked index (DWI), which was calculated using the following equation [18]:
during walking. A very recent study found that LDS is associated with prospective falls in PwMS [14]. Since LDS provides assessments of stability and fall risk that are not available through standard measures such as gait speed, understanding LDS during walking in PwMS may provide unique insights into the development of effective gait training and rehabilitation interventions for fall prevention. Fatigue is another critical factor in MS that can directly impact the rate of sensorimotor control of movement [15]. Approximately 70–90% of PwMS experience difficulties in initiating and/or sustaining physical activities (motor fatigue) during daily life [16]. Recently the six-minute walk test (6MWT), a reliable measure of motor function in MS, has been used to demonstrate motor fatigue [17,18]. PwMS not only have shorter 6MWT distances than healthy controls (HC) but also more than onethird exhibit a decline in walking distance during the 6MWT [18]. It has been suggested that diffuse demyelination and axonal loss in PwMS results in a less central motor drive for sustained muscle contractions and coordination, leading to walking-related motor fatigue [18]. Importantly, motor fatigue can worsen walking stability (LDS), as shown in healthy individuals[19], increasing their risk of falls. However, despite the well-documented stability impairments and high prevalence of motor fatigue in MS, none of the previous studies have looked into the effects of walking-related fatigue on dynamic stability in PwMS. Identification of the stability changes in patient’s physical movement like walking as a function of fatigue, might result in early prediction of falls during activities of daily living, tailored rehabilitation programs, and better management for PwMS. The primary purpose of this study was to characterize LDS in HC and PwMS during the long walking task. Specifically, we investigated minute by minute changes in LDS during the 6MWT. We hypothesized that there will be a significant difference in the temporal changes in stability during the 6MWT between PwMS and HC. Further, we investigated the association between stability changes and change in walking performance during the 6MWT as a secondary analysis. We hypothesized that there would be a significant relationship between the decline in walking distance and the decline in LDS over the duration of the 6MWT in PwMS. 2. Methods 2.1. Study design This cross-sectional research protocol complied with the principles of the Declaration of Helsinki and was reviewed and approved by the institutional review board of the Oregon Health & Science University (OHSU). All participants provided informed consent prior to participation.
DWI 6-1: {(Distance covered in 6th min- Distance covered in 1st min)/ (Distance covered in 1st min)} * 100
2.2. Participants
3. Statistical analysis
PwMS and HC were recruited on a convenience basis from the OHSU MS Clinic and the local community. Inclusion criteria were ages 18–65 years, an absence of any orthopedic or neurologic problems other than MS, and the ability to walk for 6-minutes without an assistive device. Exclusion criteria were MS exacerbation or the use of corticosteroids within 30 days of screening. Participants were instructed not to take caffeine in the morning of the testing and all testings were done between 10 am-noon. Additionally, PwMS were told not to take fatigue-related medication for 24 h before testing. Demographic and clinical characteristics, including age, gender, body mass index, disease duration, MS phenotype, self-rated Expanded Disability Status Scale (EDSS) [20] and Multiple Sclerosis Impact Scale (MSIS-29) [21] were collected.
Independent t-tests were used to assess group differences in the demographic features between HC and PwMS. To assess whether LDS and walking performance changed differently over time in PwMS compared to HC, linear mixed models were fit. Fixed effects of group, time, and the group*time interaction were included in the models, and random intercepts and slopes over time were included to account for within-subject correlations. The HC was the reference group in all models and time was modeled as continuous. The group*time interaction was the primary effect of interest. To explore whether the group*time interaction remained after adjusting for other covariates, age, MFIS total score, and gait speed were added as covariates, and retained if significant at a p < 0.05 level. In addition, the percent change in LDS (DSI6-1) from first to last minute was calculated. This calculation matched the formula used to compute DWI6-1, which has been traditionally used to assess motor fatigue in PwMS [18]. Pearson correlation coefficients compared whether the percent change from first to last minute in LDS (DSI6-1) was
2.3. Study outcomes Participants were instructed to complete the 6MWT “at their fastest 123
Gait & Posture 76 (2020) 122–127
I. Arpan, et al.
Table 1 Demographic characteristics of the study population. Data is displayed as mean ± SE except for self-rated EDSS*, which is presented as median (range). Healthy controls
People with MS
10 47.6 ± 3.3 yrs 6F/4M 173.5 ± 4.6 83.7 ± 7.9 13.9 ± 4.2 5.7 ± 2.0 7.1 ± 2.0 1.1 ± 0.4
25 51.1 ± 1.6 yrs 20 F/5M 171.9 ± 2.1 85.5 ± 4.0 39.7 ± 3.9 18.2 ± 1.7 18.4 ± 2.1 3.1 ± 0.5
Relapsing Remitting Primary Progressive Secondary Progressive
NA NA NA NA NA
21 2 2 14.4 ± 1.4 yrs 3.5 (2.5-5.0)
Physical impact score Psychological impact score
NA NA 585 ± 27 m
27.7 ± 3.8 30.2 ± 3.4 452 ± 17 m
Count Age Gender (F/M Height (cm) Weight (kg) Total MFIS score Physical fatigue Cognitive fatigue Psychological fatigue
p value
0.3 0.7 0.8 < 0.01 < 0.01 < 0.01 < 0.05
Disease type
Disease duration Self-rated EDSS* MSIS-29 scores
6MWT distance
< 0.01
associated with changes in walking performance (DWI6-1). All statistical analysis was performed in MATLAB (r2017b, Mathworks, Inc., Natick, MA, USA) using a 0.05 significance level. 4. Results 4.1. Demographics Twenty-five PwMS and ten HC were included in the study. There was no difference in the age, height, and weight of PwMS and HC (Table 1). Disease severity in the MS group was characterized by two different self-reported questionnaires: self-rated EDSS and MSIS-29. A moderately significant association (r = 0.51, p < 0.01) was found for the disability scores obtained from the EDSS and the physical impact sub-scale of MSIS-29. 4.2. Dynamic stability There was no main effect of group, but a significant group*time interaction was found for LDS (Table 2). The group*time interaction remained significant (p = 0.011) after adjusting for covariates of age, gait speed, and total MFIS score. Covariates of age (p = 0.447), gait speed (p = 0.608), and total MFIS score (p = 0.299) did not have a significant effect on LDS. Fully adjusted model results are presented in the supplemental material. Post-hoc analysis revealed that LDS did not differ between PwMS and HC at the beginning of the 6MWT, but healthy controls became more stable (decrease in λS) over time, relative
Fig. 1. Median Local dynamic instability estimate, λS, for each participant at each minute block of time. Thick lines indicate group means. Thin lines depict individual subject data. Higher values of λS indicate less stability.
to PwMS (Fig. 1). Descriptively, all (but one) HC demonstrated an increase in the LDS (decline in λS) from first to last minute (DSI6-1: −4 to −20 %, without outlier). In contrast, changes in LDS from minute 1 to minute 6 were more heterogeneous in the MS group. About 8 PwMS exhibited an improvement in LDS (DSI6-1: −1 to −21 %), 3 with no change, 5 PwMS showed < 10 % decrease in LDS (DSI6-1: 4–5%), and 9 showed > 10 % decline in LDS (DSI6-1: 7–43%).
Table 2 Summary results of unadjusted linear mixed effects models for local dynamic stability (LDS) and distance walked. Bolded p values indicate significance at a 0.05 level. Beta Estimate Local Dynamic Stability Intercept Time Group Group*Time Distance Walked Intercept Time Group Group*Time
SE
p value
Lower 95% CI
Upper 95% CI
0.443 −0.008 −0.017 0.012
0.022 0.004 0.026 0.005
< 0.001 0.035 0.514 0.009
0.400 −0.016 −0.068 0.003
0.486 −0.001 0.034 0.021
94.7 0.4 −17.6 −0.8
4.4 0.3 5.2 0.3
< 0.001 0.197 < 0.001 0.017
86.0 −0.2 −27.9 −1.5
103.3 0.9 −7.4 −0.1
4.3. Walking distance Due to technical issues during the video recording of the 6MWT in 3 PwMS, walking distance for each minute could only be obtained in 22 PwMS. Both group effect and group*time interaction were significant, indicating that there were significant differences in walking distance between two groups and over time (Table 2). The group (p = 0.043) and group*time interaction (p = 0.016) remained significant after adjusting for covariates of age, gait speed, and total MFIS score. Covariates of age (p = 0.786) and total MFIS score (p = 0.439) did not have a significant effect on walking distance, but gait speed had a significant, positive effect on walking distance (p < 0.001). Fully adjusted model 124
Gait & Posture 76 (2020) 122–127
I. Arpan, et al.
Fig. 2. A. Walking distance covered by people with MS (PwMS) and Healthy Control (HC) groups during each minute of walking during the 6MWT. Thick lines indicate unadjusted group means; thin lines indicate individual subject data. B. Comparison of change in walking distance from minute 1 to minute 6 (DWI6-1) of the 6MWT among people with MS (PwMS) and Healthy Controls (HC). Greater the decline in walking distance from minute 1 to minute 6, the more likely the participants were subjected to walking-related motor fatigue. Data is displayed as mean ± SE for each group.
[27,28]. The unique aspect of this study was the quantification of changes in LDS at one-minute intervals during the 6MWT task in PwMS and HC. Though group* time results showed that, on a group level, PwMS became less stable over time compared to HC, individual participants’ trajectories over time were heterogeneous. Unlike HC, PwMS did not adapt their walking pattern to conserve LDS during the 6MWT, and about 60 % of PwMS became less stable over time. More importantly, 9 PwMS demonstrated > 10 % change in LDS, which has been estimated as a clinically significant difference for LDS between fall-prone and healthy elderly [29], vestibular sensory disruption [25], and concussion [26]. These findings suggest that PwMS may be at a higher risk of falls during long, sustained activities due to a decline in stability over time. Consistent with this hypothesis, at least three different studies have identified LDS as a pertinent marker for retrospective and future falls in MS population [14,30,31]. In contrast to our finding of no differences in stability among PwMS and HC for the first few minutes of the 6MWT (about 250 m), a previous study by Huisinga et al. reported worse average trunk stability in PwMS during a short walk of 30 s at a self-selected pace [32]. These differences can most likely be attributed to walking speed and analysis methods. In our study, participants were instructed to walk at a faster than normal speed and to maintain that speed over the whole sixminutes [17], while subjects in the previous study walked at their selfselected pace. Secondly, differences in sampling frequency and methods to calculate LDS likely also contributed to the discrepancy across studies. Here, we used methods that have been consistently used to estimate short bouts of gait in clinical tests [8,25,26]. Due to these methodological differences, our results demonstrating no immediate differences between PwMS and HC when walking fast cannot easily be compared to results from Huisinga et al. [32]. Furthermore, a very recent study by Tajali and colleagues found that LDS was able to predict future falls in a cohort of PwMS using two minutes of normal pace treadmill walking [14]. However, the subjects were instructed to walk for five minutes prior to the testing procedure to familiarize themselves with the treadmill and to select the appropriate belt speed. Our results complement this procedural decision and encourage future studies investigating the optimal duration of walking assessments in PwMS. Another key insight from the study was the investigation of stability changes over the 6MWT in PwMS and its relationship to changes in the walking performance. The modified 6MWT has been previously used in multiple MS studies to characterize the occurrence of motor fatigue [17,18]. The decline in performance during the 6MWT in PwMS can be attributed to both peripheral and central mechanisms [33,34]. Specifically, the inflammatory processes in MS result in demyelination or partial impairment of neural pathways; while neuronal functions may be partly preserved, axonal demyelination can slow down the neural conduction resulting in fatigue especially under high demand or highly repetitive requirements. In addition to affecting the rate of performance of a motor task, fatigue is also known to diminish the precision of motor control [19]. Since LDS quantifies the rate of divergence of movement trajectories, we speculate that the decline in LDS (i.e., faster divergence
results are presented in the supplemental material. PwMS walked significantly shorter distances compared to HC at each minute of the 6MWT (Fig. 2A). Further, compared to HC, PwMS demonstrated a greater decline in walking distance from minute 1 to minute 6 of the 6MWT (DWI6-1; Fig. 2B), potentially indicating a higher incidence of walking-related fatigue in the MS group. 4.4. Relationship between changes in dynamic stability and walking performance from minute 1 to minute 6 A significant relationship (r= -0.46, p = 0.03; Fig. 3) was observed between the change in dynamic stability (DSI6-1) and the change in distance (DWI6-1) from minute 1 to minute 6 of the 6MWT in PwMS, suggesting that a decline in dynamic stability was associated with a decline in walking performance due to motor fatigue in MS. 5. Discussion The primary aim of this study was to characterize minute-by-minute changes in dynamic stability during the 6MWT in PwMS and HC. The key finding was that PwMS did not have different stability than HC until minute 4 of the 6MWT and differences in LDS persisted at minute 6. This finding supports the use of long walks in MS as shorter walking tests may not capture changes in dynamic stability over time. In addition, we found a significant relationship between the decline in stability and decline in walking performance during the 6MWT in the MS group, suggesting that walking-related fatigue may be one of the contributing factors to worsening of balance control. Walking impairments are well documented in PwMS. PwMS are known to walk slower, with fewer, shorter steps and spend a greater percentage of a gait cycle on double support time compared to HC
Fig. 3. Correlation between the percent change in local dynamic stability (DSI6and the percent change in walking distance (DWI6-1) from minute 1 to minute 6 of the 6MWT in people with MS (PwMS). 1)
125
Gait & Posture 76 (2020) 122–127
I. Arpan, et al.
inappropriately influenced this work. The research in this proposal employs the technology of APDM, a company in which author, FH, has financial interest. This potential conflict has been reviewed and managed by OHSU.
and slower control) may be an indication of the slowing of motor unit firing patterns in fatigued PwMS. The fact that decline in walking performance in this study was significantly related to the decline in LDS supports the hypothesis that there is a diminished central reserve for sustained muscle contractions and coordination during the long walking tasks in PwMS. However, it is unlikely that a single mechanism is responsible for the trends that we observed over time. Even within our cohort of MS participants, we saw between-subject variation in the trends of LDS over time – it is possible there were mechanistic subgroups within our MS cohort. Future work should further explore the mechanistic causes of the decline in motor task performance and its potential relationships with LDS in a larger sample of PwMS. Our findings have important implications for rehabilitation in today’s health care environment. Falls can result in severe injuries resulting in loss of independence, institutionalization and even death. As most falls occur during walking, assessing fall-risk based on gait parameters may be a promising approach. In addition, traditional fatigue and balance interventions programs should include gait stability retraining for fall prevention in patients that demonstrate decline in stability during the motor task performance, so that PwMS do not have to prioritize fatigue management over LDS or vice versa while long continuous activities. And finally, this study supports the use of long walks by clinicians and researchers to study stability changes during walking tasks in PwMS. Though 6-minute bout of continuous walking is uncommon, it is equally important to test and train PwMS to confidently cover large distances during activities such as recreation, shopping, etc. without experiencing a decline in physical resources.
Declaration of Competing Interest The authors (IA, PCF, & BWF) declare that there are no financial or personal relation with people or organizations that have inappropriately influenced this work. The research in this proposal employs the technology of APDM, a company in which author, FH, has financial interest. This potential conflict has been reviewed and managed by OHSU. Acknowledgements The authors would like to thank research assistants, Spencer Smith, BS and Maddy Dunn, BS, for their support in data collection. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.gaitpost.2019.10.032. References [1] J.A. Knight, Physical inactivity: associated diseases and disorders, Ann. Clin. Lab. Sci. 42 (2012) 320–337. [2] R.W. Motl, P.L. Sandroff, BM, The importance of physical fitness in multiple sclerosis, J. Nov. Physiother. 3 (2013) 7, https://doi.org/10.4172/2165-7025. 1000141. [3] C.L. Martin, et al., Gait and balance impairment in early multiple sclerosis in the absence of clinical disability, Mult. Sclerosis (Houndmills, Basingstoke, England) 12 (2006) 620–628, https://doi.org/10.1177/1352458506070658. [4] M.H. Cameron, J.M. Wagner, Gait abnormalities in multiple sclerosis: pathogenesis, evaluation, and advances in treatment, Curr. Neurol. Neurosci. Rep. 11 (2011) 507–515, https://doi.org/10.1007/s11910-011-0214-y. [5] S. Coote, J.J. Sosnoff, H. Gunn, Fall incidence as the primary outcome in multiple sclerosis falls-prevention trials: recommendation from the international MS falls prevention research network, Int. J. MS Care 16 (2014) 178–184, https://doi.org/ 10.7224/1537-2073.2014-059. [6] M. McCarthy, Falls are leading cause of injury deaths among older people, US study finds, BMJ 354 (2016) i5190, https://doi.org/10.1136/bmj.i5190. [7] J.B. Dingwell, J.P. Cusumano, P.R. Cavanagh, D. Sternad, Local dynamic stability versus kinematic variability of continuous overground and treadmill walking, J. Biomech. Eng. 123 (2001) 27–32. [8] K.S. van Schooten, et al., Ambulatory fall-risk assessment: amount and quality of daily-life gait predict falls in older adults, J. Gerontol. Ser. A 70 (2015) 608–615, https://doi.org/10.1093/gerona/glu225. [9] J.G. Nutt, F.B. Horak, B.R. Bloem, Milestones in gait, balance, and falling, Mov. Disord. 26 (2011) 1166–1174, https://doi.org/10.1002/mds.23588. [10] K. Takakusaki, Functional neuroanatomy for posture and gait control, J. Mov. Disord. 10 (2017) 1–17, https://doi.org/10.14802/jmd.16062. [11] J. McLoughlin, C. Barr, M. Crotty, S.R. Lord, D.L. Sturnieks, Association of postural sway with disability status and cerebellar dysfunction in people with multiple sclerosis: a preliminary study, Int. J. MS Care 17 (2015) 146–151, https://doi.org/ 10.7224/1537-2073.2014-003. [12] A.J. Solomon, J.V. Jacobs, K.V. Lomond, S.M. Henry, Detection of postural sway abnormalities by wireless inertial sensors in minimally disabled patients with multiple sclerosis: a case-control study, J. Neuroeng. Rehabil. 12 (2015) 74, https:// doi.org/10.1186/s12984-015-0066-9. [13] A. Alizadeh, S.M. Dyck, S. Karimi-Abdolrezaee, Myelin damage and repair in pathologic CNS: challenges and prospects, Front. Mol. Neurosci. 8 (2015), https://doi. org/10.3389/fnmol.2015.00035 35-35. [14] S. Tajali, et al., Impaired local dynamic stability during treadmill walking predicts future falls in patients with multiple sclerosis: a prospective cohort study, Clin. Biomech. (2019), https://doi.org/10.1016/j.clinbiomech.2019.05.013. [15] I. Cogliati Dezza, et al., Functional and structural balances of homologous sensorimotor regions in multiple sclerosis fatigue, J. Neurol. 262 (2015) 614–622, https:// doi.org/10.1007/s00415-014-7590-6. [16] S.R. Schwid, et al., Quantitative assessment of motor fatigue and strength in MS, Neurology 53 (1999), https://doi.org/10.1212/wnl.53.4.743 743-743. [17] M.D. Goldman, R.A. Marrie, J.A. Cohen, Evaluation of the six-minute walk in multiple sclerosis subjects and healthy controls, Mult. Scler. 14 (2008) 383–390, https://doi.org/10.1177/1352458507082607. [18] C. Leone, et al., Prevalence of walking-related motor fatigue in persons with multiple sclerosis: decline in walking distance induced by the 6-Minute walk test, Neurorehabil. Neural Repair 30 (2016) 373–383, https://doi.org/10.1177/
6. Limitations One of the major limitations of the study was the small sample size of both HC and PwMS. Therefore, findings from this exploratory analysis should be rigorously tested and replicated in future studies. In addition to the issues surrounding the lower sample size, another limitation was the lack of falls data. Since not everybody in the MS group demonstrated the decline in dynamic stability during walking, it would have been interesting to see if PwMS, who maintained LDS during the duration of the 6MWT, were at a lower risk of falling. Further, by separating the 6MWT into one-minute bouts, we retained only an average of 5 bouts of straight gait in each minute. This small number of gait bouts within each minute likely increased the minute-to-minute variability in our LDS estimates [35]. However, we repeated our analysis using two-minute long segments and obtained similar results – between-group differences were not present initially, but emerged after at least two minutes of walking. These results using two-minute long segments are presented in the supplemental material. While our findings were robust, future studies using different segmentations of the 6MWT should confirm our results. 7. Conclusions PwMS decreased stability over time relative to controls during longwalking tasks. Further, the decline in stability and motor fatigue were related to each other, suggesting a possible role of fatigue in worsening of walking balance in PwMS. Still, the underlying mechanisms are not completely resolved and further research incorporating measurements of cortical gait control and its relationship to worsening of the motor task performance might be promising. Funding This research was supported by the Collins Medical Trust (Portland, OR), the Medical Research Foundation (Portland, OR), and the National MS Society’s Mentored Postdoctoral Fellowship Award. The authors (IA, PCF, & BWF) declare that there are no financial or personal relation with people or organizations that have 126
Gait & Posture 76 (2020) 122–127
I. Arpan, et al.
1016/j.gaitpost.2012.10.012. [28] J.J. Sosnoff, B.M. Sandroff, R.W. Motl, Quantifying gait abnormalities in persons with multiple sclerosis with minimal disability, Gait Posture 36 (2012) 154–156, https://doi.org/10.1016/j.gaitpost.2011.11.027. [29] T.E. Lockhart, J. Liu, Differentiating fall-prone and healthy adults using local dynamic stability, Ergonomics 51 (2008) 1860–1872, https://doi.org/10.1080/ 00140130802567079. [30] R. Hilfiker, et al., Local dynamic stability as a responsive index for the evaluation of rehabilitation effect on fall risk in patients with multiple sclerosis: a longitudinal study, BMC Res. Notes 6 (2013), https://doi.org/10.1186/1756-0500-6-260 260260. [31] A.T. Peebles, A.P. Bruetsch, S.G. Lynch, J.M. Huisinga, Dynamic balance is related to physiological impairments in persons with multiple sclerosis, Arch. Phys. Med. Rehabil. 99 (2018) 2030–2037, https://doi.org/10.1016/j.apmr.2017.11.010. [32] J.M. Huisinga, M. Mancini, R.J. St George, F.B. Horak, Accelerometry reveals differences in gait variability between patients with multiple sclerosis and healthy controls, Ann. Biomed. Eng. 41 (2013) 1670–1679, https://doi.org/10.1007/ s10439-012-0697-y. [33] K.R. Sharma, J. Kent-Braun, M.A. Mynhier, M.W. Weiner, R.G. Miller, Evidence of an abnormal intramuscular component of fatigue in multiple sclerosis, Muscle Nerve 18 (1995) 1403–1411, https://doi.org/10.1002/mus.880181210. [34] A.T. White, J.N. Lee, A.R. Light, K.C. Light, Brain activation in multiple sclerosis: a BOLD fMRI study of the effects of fatiguing hand exercise, Mult. Scler. 15 (2009) 580–586, https://doi.org/10.1177/1352458508100034. [35] K.S. van Schooten, S.M. Rispens, P.J. Elders, J.H. van Dieen, M. Pijnappels, Toward ambulatory balance assessment: estimating variability and stability from short bouts of gait, Gait Posture 39 (2014) 695–699.
1545968315597070. [19] P.-C. Kao, M.A. Pierro, K. Booras, Effects of motor fatigue on walking stability and variability during concurrent cognitive challenges, PLoS One 13 (2018), https:// doi.org/10.1371/journal.pone.0201433 e0201433-e0201433. [20] J. Bowen, L. Gibbons, A. Gianas, G.H. Kraft, Self-administered Expanded Disability Status Scale with functional system scores correlates well with a physician-administered test, Mult. Sclerosis (Houndmills, Basingstoke, England) 7 (2001) 201–206, https://doi.org/10.1177/135245850100700311. [21] J. Hobart, D. Lamping, R. Fitzpatrick, A. Riazi, A. Thompson, The Multiple Sclerosis Impact Scale (MSIS-29): a new patient-based outcome measure, Brain 124 (2001) 962–973, https://doi.org/10.1093/brain/124.5.962. [22] M. El-Gohary, et al., Continuous monitoring of turning in patients with movement disability, Sensors (Basel, Switzerland) 14 (2013) 356–369, https://doi.org/10. 3390/s140100356. [23] S.M. Cain, M.V. Porter, L. Ojeda, N.C. Perkins, Accurate and robust gait event detection using foot-mounted inertial measurement units, Proceedings of the American Society of Biomechanics, (2019). [24] H. Kantz, A robust method to estimate the maximal Lyapunov exponent of a time series, Phys. Lett. A 185 (1994) 77–87. [25] L.H. Sloot, et al., Sensitivity of local dynamic stability of over-ground walking to balance impairment due to galvanic vestibular stimulation, Ann. Biomed. Eng. 39 (2011) 1563–1569, https://doi.org/10.1007/s10439-010-0240-y. [26] P.C. Fino, A preliminary study of longitudinal differences in local dynamic stability between recently concussed and healthy athletes during single and dual-task gait, J. Biomech. 49 (2016) 1983–1988, https://doi.org/10.1016/j.jbiomech.2016.05.004. [27] M.J. Socie, R.W. Motl, J.H. Pula, B.M. Sandroff, J.J. Sosnoff, Gait variability and disability in multiple sclerosis, Gait Posture 38 (2013) 51–55, https://doi.org/10.
127