Sage Journals: Discover world-class research

Abstract

Background. Persons with stroke (PwS) are at increased risk of falls, especially toward the paretic side, increasing the probability of a hip fracture. The ability to recover from unexpected loss of balance is a critical factor in fall prevention. Objectives. We aimed to compare reactive balance capacity and step kinematics between PwS and healthy controls. Methods. Thirty subacute PwS and 15 healthy controls were exposed to forward, backward, right, and left unannounced surface translations in 6 increasing intensities while standing. Single step threshold, multiple step threshold, and fall threshold (ie, perturbation intensity leading to a fall into harness system) were recorded as well as reactive step initiation time, step length, and step velocity. Results. Twenty-five PwS fell into harness system during the experiment while healthy controls did not fall. Fourteen out of 31 falls occurred in response to surface translations toward the nonparetic side, that is, falling toward the paretic side. Compared with healthy controls, PwS demonstrated significantly lower fall threshold and multiple step threshold in response to forward, backward, and lateral surface translations. Impairments were more pronounced in response to forward surface translation and toward the nonparetic side (ie, loss of balance toward the paretic side). A trend toward significant shorter step length in response to lateral surface translations was found in PwS compared with healthy controls. Conclusions. Findings highlight the importance of assessing reactive balance capacity in response to perturbations in different directions and intensities in addition to the routine assessment in PwS.

Keywords

postural response compensatory stepping multiple step threshold fall threshold falls rehabilitation

Introduction

Falls are among the most common medical complications after stroke. About 70% of persons with stroke (PwS) fall during the first 6 months after discharge from hospital or rehabilitation setting and it remains a considerable health concern throughout their life span.^1-4 The proportion of falls leading to injury in PwS is up to 69%.⁴ Moreover, compared with the general population of older adults who fall, PwS who fall are twice as likely to sustain a hip fracture, mostly involving the affected side.^2,5-8 In addition, to physical consequences associated with fractures and related injuries, falls may have serious psychological and social consequences such as functional decline, poor quality of life, dependency, social isolation, and depression.⁸

The ability to execute a successful reactive response to a sudden loss of balance, that is, unexpected balance perturbation, is a critical factor that determines whether a fall occurs.⁹ Although many falls among PwS living in the community occur due to inability to recover from balance loss during walking,¹⁰ Mansfield et al¹¹ found that impaired reactive balance control during both standing and walking predicted increased risk of falls in PwS. Commonly used clinical instruments assessing balance and gait in PwS, such as the Berg Balance Scale,¹² Timed Up and Go,^13,14 Dynamic Gait Index,¹⁵ the Performance Oriented Mobility Assessment (POMA),¹⁶ and the miniBESTest¹⁷ can identify patients with balance impairments and risk for falls; however, they have limitations in their usage especially in respect to assessing the ability to recover from unexpected loss of balance and indicating specific impairments contributing to falls. Perturbations i.e., sternal pushes, are provided in the POMA, but a warning is given. In the miniBESTest the patient leans against the examiner’s resistance who then releases it, but here the direction and timing of perturbation is predicted. Thus, neither instrument assesses the ability to recover from unexpected loss of balance. Assessing balance recovery abilities using a horizontal platform translation system that moves unexpectedly may be more ecologically valid since it provides perturbation types, that is, slip or trip, similar to unexpected loss of balance that frequently precedes falls in real life.¹⁸ This may also resolve many of these issues by providing a high level of unpredictability.

Studies have documented characteristics of reactive stepping performances in stance that were associated with falls in older adults^19-23: initiating reactive step responses at lower levels of instability (ie, lower step threshold),^19,22,23 failure to recover equilibrium^19,24 and taking multiple balance recovery steps rather than a single step response.^19-21,25,26 Recent studies in PwS have shown impaired stepping reactions characterized by the need for assistance, inability to step with either lower limb, multiple step strategy, and falls.^11,27-29 Also, PwS demonstrated lower postural stability at step touch-down in response to slip-like forward perturbations (posterior center of mass position and larger backward center of mass velocity relative to base of support) and shorter step length compared with healthy controls.³⁰ However, these studies have focused on responses to large perturbations, provided at a single intensity in forward and backward directions. Providing unexpected multidirectional perturbations (forward, backward, and sideways) may be the closest approach to simulate a fall in real life, which is unexpected and may occur to all directions. There is relatively little information regarding characteristics of reactive balance strategies to lateral loss of balance in PwS. This is specifically important in PwS suffering from unilateral hemiparesis, since lateral perturbations involve different biomechanical demands than forward or backward perturbations. Also, in many falls a second lateral loss of balance that followed forward or backward step was frequently seen in older adults.⁷ The fact that PwS tend to fall toward the paretic side^10,31 highlights the clinical importance of studying these reactions in response to multidirectional perturbations.

Reactive balance assessment that includes increased perturbation intensities enables assessment of reactive balance capacity through measures such as single step threshold and multiple step threshold as was recently reported in chronic PwS.³² To our knowledge, fall threshold (ie, perturbation intensity that results in a fall into harness system) that might be a more direct measure of inability to prevent a fall in real life, has not yet been studied in PwS. Therefore, the main aim of the study was to compare single step threshold, multiple step threshold, and fall threshold in response to increasing intensities of unannounced random forward, backward, right, and left surface translations between PwS and healthy controls. Single step threshold is defined as the minimum perturbation intensity that consistently elicits a single compensatory step in at least 2 consecutive perturbations; multiple step threshold is defined as the minimum perturbation intensity that consistently elicits 2 or more compensatory steps in at least 2 consecutive perturbations and fall threshold is defined as the perturbation intensity that results in unsuccessful balance recovery, that is, fall into harness system.^33-35 We hypothesized that similarly to de Kam et al³² who tested recovery stepping in chronic PwS, subacute PwS in our study would demonstrate lower single step threshold, lower multiple step threshold as well as lower fall threshold compared with healthy controls. Our second aim was to explore differences in kinematics of reactive step responses between PwS and healthy controls. We hypothesized that similar to Salot et al³⁰ who reported shorter step length in response to forward perturbations, PwS in our study would exhibit impaired kinematics of step recovery responses, that is, delayed step initiation, shorter step length, and lower step velocity in all 4 directions. Based on previous studies conducted in persons with Parkinson’s disease^36,37 and multiple sclerosis,³⁸ we further hypothesized that reactive balance impairments would be more pronounced in response to forward surface translation than to backward surface translation and in response to translations toward the nonparetic side (ie, falling toward the paretic side) than toward the paretic side in PwS.

Methods

Participants

Thirty patients with first unilateral stroke and 15 healthy adults participated in the study (Table 1). PwS were recruited during their hospitalization at the Loewenstein Rehabilitation Hospital (LRH), Ra’anana, Israel as part of a baseline assessment in an ongoing randomized control trial (RCT) aiming to examine the effect of perturbation-based balance training on reactive balance responses in PwS (ClinicalTrials.gov #NCT02619175). Healthy participants were recruited through flyers (age range 40-75 years). Participants had to be able to stand for at least 2 minutes and to walk independently or under supervision with or without a walking aid. Exclusion criteria included other neurological disorders in addition to stroke, significant musculoskeletal conditions (eg, severe arthritis, joint replacement surgery) and significant visual impairment. Participants signed a written informed consent in accordance with approved procedures by the Helsinki Ethics Committee, LRH (#LOE-14-0021).

Table 1.

Subjects Characteristics.^a

Characteristics	Stroke (n = 30)	Controls (n = 15)
Age, years	61.7 ± 7.8	60.1 ± 6.9
Gender, male/female	25/5	13/2
Height, cm	168.5 ± 7.2	172 ± 5.7
Weight, kg	76.2 ± 11.1	78.4 ± 11.9
Type of stroke, ischemic/hemorrhagic	24/6	—
Stroke side, right/left	15/15	—
Time after stroke, days	51.7 ± 22.9	—
Fugl-Meyer score motor (0-34)	30 (15-36)	—
Fugl-Meyer score sensation (0-12)	12 (5-12)	—
Berg Balance Scale (0-56)	42 (12-56)	56 (56-56)***
6-minute walk test, meters	297.9 ± 139.5	572.4 ± 57.9***
Assistive device (n)
None	6	15
Walker	20
Crutches	1
Cane	1
Quadripode	2
Orthosis (n)
None	15	15
Eight band	12
Ankle foot orthosis (AFO)	3

Presented values are mean ± SD for continuous variables, median (range) for ordinal variables, and frequency (n) for categorial variables. Between-group comparisons were conducted using independent t-test for continuous variables, Mann-Whitney U test for ordinal variables, and Pearson’s chi-square test for independence for categorial variables.

***

P < .001.

Study Protocol

Participants stood on a computerized treadmill system with a horizontal movable platform (Balance Tutor, MediTouch, Israel), wearing a safety harness that prevented falls but did not restrict their movements. Walking aid usage was not allowed during the assessment. Participants were instructed to stand with feet placed together and to react naturally to prevent themselves from falling in response to random unannounced forward, backward, right, and left surface translations with time interval of 25 to 30 seconds between perturbations. Participants were informed neither of the direction nor the timing of perturbations. Surface translations were increased systematically in six intensities from low (intensity 1) to high (intensity 6) for a total of 24 perturbation trials (characteristics of perturbation intensities are described in Table 2). In this study, perturbation direction refers to the direction of the platform translation. In case of a fall into harness, determined by visual inspection of 2 physical therapists, the participant did not continue to higher intensity. Seated rest breaks were given between trials if needed. Participants performed all trials wearing their own sport shoes and foot orthosis in case they needed ankle support.

Table 2.

Perturbation Characteristics.

Perturbation Intensity	Forward/Backward Surface Translation			Lateral Surface Translation
Perturbation Intensity	Distance (cm)	Velocity (cm/s)	Acceleration (cm/s²)	Distance (cm)	Velocity (cm/s)	Acceleration (cm/s²)
1	10	17	126	5	20	73
2	14	24	150	8	27	97
3	17	32	173	10	35	121
4	21	39	197	12	42	145
5	25	46	220	15	50	169
6	29	54	244	17	58	193

Presented are peak values. The system controller receives from the PC program the required motion parameters, which are the target position, maximal velocity, acceleration, and deceleration. The controller has an internal motion profile generator that generates a trapezoidal velocity profile.

Prior to reactive balance control assessment, PwS underwent clinical assessment including the Fugl-Meyer Assessment, Berg Balance Scale (BBS), and 6-Minute Walk Test (6MWT). The clinical assessment for healthy participants included the 6MWT and BBS only.

Data Collection and Analysis

A 2-camera motion capture system with a sampling rate of 60-Hz was used to record body kinematics (APAS, Ariel Dynamics Inc, Trabuco Canyon, CA, USA). Cameras were mounted 4 meters in front of the perturbation system at an angle of 45° between each camera and the subject’s standing position, at a height of 2.5 m. Eight reflective markers were placed on the skin bilaterally on the following landmarks: midline of the anterior aspect of ankle joints, anterior superior iliac spines, shoulder acromion processes and radial styloid processes and 2 additional markers were placed on the moving platform to identify perturbation onset.³³ The raw marker data were mapped onto a 3D coordinate system using an internal direct linear transformation algorithm. Data were recorded, digitized, transformed, and low-pass filter smoothed (Butterworth second-order forward and backward passes) with a cutoff frequency of 5-Hz. This approach was shown to be valid and reliable, that is, mean point estimate error of <3.5 mm, 1.4 mm mean linear error, and 0.26° mean angular error.³⁹

Our primary outcome measures included the fall threshold and the multiple step threshold. Balance recovery strategies following perturbations were identified by visual inspection of the videos: (1) no step, (2) single step, (3) multiple step (≥2 steps), and (4) fall into harness system, that is, when a subject is unambiguously supported by the harness system.^34,35 In case a participant did not fall at perturbation intensity 6, his or her fall threshold was calculated as intensity 7.

Secondary outcome measures included the single step threshold and spatiotemporal characteristics of the reactive step response. Single step threshold and multiple step threshold have been shown to demonstrate high test-retest reliability in young adults.⁴⁰ Also, high interobserver reliability for step threshold, multiple step threshold and fall threshold was found in older adults.³³ The spatiotemporal events extracted from the collected data were analyzed by customized semiautomated program written in C# (Microsoft, 2000) specifically for this study protocol³³: (1) The surface translation was detected as the first horizontal deviation of the marker placed on the treadmill greater than 3 standard deviations from the average baseline noise. (2) First step lift-off the ground was defined at the first deviation of the marker placed on the swinging leg ankle joint in the vertical direction, greater than 4 mm from the average baseline after the surface translation. (3) Step initiation (ms) was calculated as the time from surface translation to foot off the ground. 4) Foot-contact the ground poststepping was defined as the swinging leg ankle joint marker was placed on the ground completing the step, that is, same ground level as the vertical baseline marker placement. (5) Step duration (ms) was calculated as the time from step initiation, that is, foot off the ground to foot contact. (6) First step length was calculated as the Euclidian distance (cm) that the ankle markers displaced from step initiation to foot contact. (7) The average step velocity (m/s) was calculated as step length divided by step duration. During data analysis, foot off and foot contact were verified through visual inspection of the ankle marker. For the stroke group video recordings were viewed to determine the leg that initiated the reactive step response (ie, paretic leg or nonparetic leg).

Sample Size Estimation

Separate calculations were performed to determine sample size requirements based on multiple step threshold and kinematic parameters (ie, step length), that earlier found to be different between PwS and healthy controls.^30,32 Based on results by de Kam et al,³² the multiple step threshold was 2.2 m/s² in PwS and 4.0 m/s² in controls. Based on data reported by Salot et al,³⁰ the step length in response to forward perturbation was 0.05 m for PwS and 0.09 m for controls. Using net reduction values (1.8 m/s² and 0.04 m, respectively) in combination with the initial variance estimates (standard deviations of 1.3 m/s² and 0.045 m, respectively), it was determined that 9 and 14 participants per group would be required to be able to reject the null hypothesis with level of significance (α) of 0.05 and power (1 − β) of 0.8 (Power and Sample Size calculations, version 3.0).

Statistical Analysis

For statistical analysis, SPSS version 24.0 was used (IBM Corp). Normality of data was evaluated using the Shapiro-Wilk test. Subjects’ characteristics and clinical measures of balance and gait were compared between the 2 groups using independent t-test (age, weight, height, 6-minute walk test), Mann-Whitney U test (Berg Balance scale), and Pearson’s chi-square test for independence (gender and assistive devices) (Table 1).

The generalized linear models for ordinal variables (cumulative logit ordinal logistic regression) was conducted to test differences in fall threshold between groups. Differences in frequencies of falls by direction were tested using Pearson’s chi-square test for goodness of fit.

Multiple step threshold and single step threshold were assessed in response to all 4 surface translation directions. Therefore, group effect (PwS vs healthy controls), direction effect (forward vs backward; right vs left) and group-by-direction interaction (eg, whether performance in response to each direction of surface translation was similar between PwS and healthy controls) were tested using the generalized estimating equations for ordinal variables, with direction as within-subjects effect and group as between-subjects effect. Note: for the stroke group we compared between surface translations toward the paretic versus nonparetic side instead of toward right versus left. Significance levels were set at P < .0125 in order to adjust for multiple comparisons and trend levels were set at P < .05. Kinematic analysis for reactive step measures was conducted only for stepping responses at intensity 4, since stepping responses were not frequently induced at low perturbation intensities in healthy controls and most PwS were not able to recover balance loss at higher perturbation intensities. Since data were nonnormally distributed, group effect, direction effect, and group-by-direction interaction were tested using the generalized estimating equations for continuous variables, with Gamma link function, with direction as within-subjects effect and group as between-subjects effect. Significance levels were set at P < .008 in order to adjust for multiple comparisons.

All comparisons were repeated using the Mann-Whitney U test for group effects and the Wilcoxon signed-rank test for direction effects yielding similar results.

Cohen’s d effect size (ES) of group difference was calculated by taking the difference between the means of both groups divided by the average population standard deviation. The following guidelines were used when interpreting effect size magnitudes: 0.2 was considered small, 0.5 was considered medium, 0.8 was considered large, 1.2 was considered very large, and 2.0 considered huge.⁴¹

Results

A total of 478 balance recovery trials in PwS and 359 in healthy controls were analyzed. In the stroke group, 6 trials were not analyzed; 4 due to technical difficulties in the motion capture system and 2 as the participant asked to stop before test completion due to fatigue. In the healthy control group, one trial was unavailable for data analysis due to technical difficulties in the motion capture system. Twenty-five PwS fell into harness system during the experiment and according to protocol we did not continue to higher intensities than the intensity that resulted in a fall. Thus, they did not complete all 24 perturbations (Table 3).

Table 3.

Frequencies of Falls Into Harness, That Is, Failed Balance Recovery Responses and Number of Persons With Stroke Who Fell, by Perturbation Intensity and Direction.

Perturbation direction (surface translation)	Intensity 1 (n = 30)	Intensity 2 (n = 27)	Intensity 3 (n = 23)	Intensity 4 (n = 20)	Intensity 5 (n = 13)	Intensity 6 (n = 8)	Total
Forward	2	2	1	2	1	1	9
Backward	—	—	—	1	—	—	1
Toward paretic side	—	1	2	4	—	—	7
Toward nonparetic side	1	1	3	2	4	3	14
No. of falls (No. of fallers)	3(3)	4(4)	6(3)	9(7)	5(5)	4(3)	31(25)

Persons with stroke (PwS) who exhibited a failed balance recovery response did not continue to the next intensity. A total of 20 PwS fell once, 4 PwS exhibited 2 falls, and 1 PwS exhibited 3 falls at the same perturbation intensity to different perturbation directions.

Fall Threshold

Thirty-one falls were observed in 25 of 30 PwS (83.3%), and none in healthy controls. Frequencies of falls in PwS by perturbation intensity and direction are presented in Table 3. Fall threshold was significantly lower in PwS compared with healthy controls, B = 4.3, Wald χ²(1) = 14.6, P < .001; ES = 1.63. Significant differences were found in falls direction in PwS, χ²(3) = 11.194, P = .011, demonstrating higher frequency of falls in response to surface translations toward the nonparetic side (ie, falling toward the paretic side) and lower frequency of falls in response to backward surface translations.

Multiple Step Threshold

In response to forward and backward surface translations, significant main effects of group, B = 2.8, Wald χ²(1) = 13.1, P < .001, and direction, B = −1.1, Wald χ²(1) = 6.9, P = .009, were found. PwS demonstrated significantly lower multiple step thresholds compared with healthy controls. Multiple step thresholds were lower in response to forward compared with backward surface translations. A trend toward significant group-by-direction interaction was found, B = −1.4, Wald χ²(1) = 5.9, P = .015 (Table 4). Simple effects analysis revealed that the difference between groups was larger in response to backward compared with forward surface translations (P < .001 and P = .042, respectively).

Table 4.

Comparisons of Multiple Step Thresholds and Single Step Thresholds by Group and Direction of Surface Translations.

Postural Characteristics	Direction of Surface Translation	Stroke	Control	Group	Direction	Group × Direction
Multiple step threshold (perturbation intensity)	Forward	2.39 ± 1.5	3.66 ± 2.1	P < .001*	P = .009*	P = .015^†
	Backward	3.26 ± 1.7	6.00 ± 1.8	P < .001*	P = .009*	P = .015^†
	Right/Paretic side	3.28 ± 1.5	5.53 ± 1.8	P = .011*	P = .866	P = .082
	Left/Nonparetic side	3.13 ± 1.3	4.80 ± 1.8	P = .011*	P = .866	P = .082
Step threshold (perturbation intensity)	Forward	1.50 ± 0.9	1.86 ± 1.2	P = .609	P = .009*	P = .574
	Backward	1.84 ± 0.8	2.26 ± 1.4	P = .609	P = .009*	P = .574
	Right/Paretic side	1.41 ± 0.7	1.86 ± 1.1	P = .096	P = .001*	P = .001*
	Left/Nonparetic side	2.20 ± 1.0	1.66 ± 0.7	P = .096	P = .001*	P = .001*

Values are mean +-SD. Comparisons were conducted using Generalized estimating equations for ordinal variables.

P < 0.0125.

†

Significant when not corrected for multiple comparisons.

In response to lateral surface translations, significant main effect of group was found, B = 1.6, Wald χ²(1) = 6.5, P = .011, demonstrating lower multiple step thresholds in PwS versus healthy controls regardless of direction (P = .001 and P = .006 in response to surface translations toward the paretic side vs right side and toward the nonparetic side vs left side, respectively). No significant main effect of direction or group-by-direction interaction were found, B = 0.09, Wald χ²(1) = 0.03, P = 0.866 and B = 1.4, Wald χ²(1) = 3.0, P = .082, respectively (Table 4).

Also, large effect sizes for multiple step threshold were found in response to forward (ES = 0.69), backward (ES = 1.26), to right/paretic side (ES = 1.07), and to left/nonparetic side (ES = 0.95).

Single Step Threshold

In response to forward and backward surface translations, significant main effect of direction was found, B = −1.0, Wald χ²(1) = 6.8, P = .009. Participants initiated a step response at lower perturbation intensities (ie, lower step threshold) in response to forward versus backward surface translations. No significant main effect of group and group-by-direction interaction were found, B = 0.3, Wald χ²(1) = 0.2, P = .609 and B = 0.3, Wald χ²(1) = 0.3, P = .574, respectively.

In response to lateral surface translations, significant main effect of direction, B = −1.8, Wald χ²(1) = 10.9, P = .001, and significant group-by-direction interaction, B = 2.047, Wald χ²(1) = 10.344, P = .001, were found. Simple effects analysis revealed that PwS demonstrated lower step threshold in response to surface translations toward the paretic side compared with the nonparetic side (P = .001) while step threshold did not differ between directions (right vs left) in healthy controls (P = .542).

Small to medium effect sizes were found for single step threshold in response to forward (ES = 0.35), backward (ES = 0.40), right/toward the paretic side (ES = 0.51), and left/toward the nonparetic side (ES = 0.58).

Spatiotemporal Parameters of Recovery Step

Step Initiation

A trend toward significant main effect for direction was found in response to lateral surface translations, B = −47.5, Wald χ²(1) = 5.5, P = .019 (Table 5). Participants initiated a reactive step response earlier in surface translations toward the paretic/right side versus nonparetic/left side. No significant main effect for group or group-by-direction interaction were found, B = 8.5, Wald χ²(1) = 0.1, P = .720 and B = −5.9, Wald χ²(1) = 0.03, P = .851, respectively. Effect sizes were small for forward (ES = 0.32), backward (ES = 0.05), toward right/paretic side (ES = 0.03), and toward left/nonparetic side (ES = 0.12).

Table 5.

Comparisons of Reactive Step Kinematics by Group and Direction of Surface Translations.^a

Postural Characteristics	Direction of Surface Translation	Stroke	Control	Group	Direction	Group × Direction
Step initiation (ms)	Forward	269.6 ± 72.0	250.5 ± 40.9	P = .888	P = .834	P = .678
	Backward	274.5 ± 93.5	269.6 ± 102.5	P = .888	P = .834	P = .678
	Right/Paretic side	229.0 ± 65.3	231.6 ± 68.2	P = .720	P = .019^†	P = .851
	Left/Nonparetic side	276.6 ± 67.3	285.1 ± 66.6	P = .720	P = .019^†	P = .851
Step length (cm)	Forward	11.65 ± 4.5	16.04 ± 6.8	P = .128	P = .046^†	P = .411
	Backward	14.66 ± 5.6	17.72 ± 5.8	P = .128	P = .046^†	P = .411
	Right/Paretic side	18.10 ± 11.1	22.98 ± 11.3	P = .056^†	P = .614	P = .729
	Left/Nonparetic side	16.13 ± 9.5	22.49 ± 8.8	P = .056^†	P = .614	P = .729
Step velocity (m/s)	Forward	0.47 ± 0.1	0.61 ± 0.2	P = .211	P = .016^†	P = .429
	Backward	0.58 ± 0.2	0.66 ± 0.2	P = .211	P = .016^†	P = .429
	Right/Paretic side	0.69 ± 0.3	0.69 ± 0.3	P = .120	P = .776	P = .579
	Left/Nonparetic side	0.46 ± 0.2	0.63 ± 0.4	P = .120	P = .776	P = .579

Values are mean ± SD. Comparisons were conducted using generalized estimating equations. Comparisons of reactive step kinematics were conducted at perturbation intensity 4, where most PwS and healthy controls performed reactive step responses.

P < .008.

†

Significant when not corrected for multiple comparisons.

Step Length

A trend toward significant main effect for direction was found in forward and backward surface translations, B = −0.230, Wald χ²(1) = 3.964, P = .046 (Table 5). Participants performed a shorter step in response to forward versus backward surface translations. No significant main effect for group or group-by-direction interaction were found, B = 0.1, Wald χ²(1) = 2.3, P = .128 and B = 0.1, Wald χ²(1) = 0.676, P = .411.

In response to lateral surface translations a trend toward significant group effect was found, B = 0.333, Wald χ²(1) = 3.641, P = .056. Step length was shorter in PwS compared with healthy controls. Analysis of group effect within each direction showed significant shorter step for PwS compared with healthy controls in response to surface translations toward the nonparetic side/left side (P = .048). No significant effect for direction or group-by-direction interaction were found, B = 0.11, Wald χ²(1) = 0.25, P = .614 and B = −0.09, Wald χ²(1) = 0.120, P = .729. Small to medium effect sizes were found in response to forward (ES = 0.72), backward (ES = 0.52), right/paretic (ES = 0.43), and left/nonparetic surface translations (ES = 0.66).

Step Velocity

A trend toward significant main effect for direction was found in response to surface translations to forward and backward directions, B = −0.111, Wald χ²(1) = 5.794, P = .016 (Table 5). Participants performed a slower step in response to forward versus backward surface translations. No significant main effect for group or group-by-direction interaction were found, B = 0.08, Wald χ²(1) = 1.5, P = .211; B = 0.06, Wald χ²(1) = 0.6, P = .429). Medium to large effect sizes were found in response to forward (ES = 0.82), backward (ES = 0.44), right/paretic (ES = 0.75), and left/nonparetic (ES = 0.56) surface translations.

Observations of the video recordings showed that in response to forward and backward surface translations most reactive step responses were initiated with the nonparetic leg (78.9% and 73.7%, respectively). In response to surface translations toward the nonparetic side (ie, the nonparetic leg was unloaded), 89.5% of recovery step responses were initiated with the nonparetic leg. However, in response to surface translations toward the paretic side (ie, the paretic leg was unloaded), 70% of reactive step responses were initiated by the paretic leg.

Discussion

Our results are in line with previous studies demonstrating lower ability to recover balance and prevent falls in response to laboratory-induced perturbations in PwS.^30,35,42-44 We have extended previous findings by (1) examining reactive balance in response to multidirectional perturbations, with an emphasis on lateral perturbations, which might be more challenging for PwS suffering from unilateral hemiparesis and (2) identifying perturbation intensity leading to insufficient balance recovery (ie, fall threshold). We found that subacute PwS demonstrated impaired reactive balance control as shown by lower multiple step thresholds and fall thresholds. Multiple step threshold is considered as a strong predictor for falls in older adults.²⁶ PwS in our study performed more than one step to recover balance loss in response to forward, backward, and lateral unexpected surface translations (Table 4). This is in line with de Kam et al³² who found similar results in chronic PwS.

In addition, we found significantly lower fall threshold in PwS. Assessing the perturbation intensity and direction that results in a failure to prevent a fall (ie, fall threshold) has not yet been studied in PwS. Lower fall threshold in PwS may serve as a more direct measure of the inability to prevent a fall during loss of balance in real-life. Eighty-three percent of PwS (25 of 30) experienced a fall into harness system during the examination, especially during loss of balance toward the paretic side (45% of falls, Table 2). No falls were observed in healthy controls (ES = 1.63). This is an important finding since Mansfield et al^11,44 found that measures of reactive balance control in the laboratory predict falls in subacute PwS.

The high laboratory fall rates observed in our study may be a result of shorter step length (marginally significance for group effect, Table 5) and lower reactive step velocity (medium to large effect sizes) in PwS rather than delayed step initiation time (Table 5). This indicates that PwS are able to initiate recovery step rapidly, but fail to execute the step effectively, which may be the main reason for many of the falls in PwS. This is consistent with Salot et al³⁰ who found no significant difference in step initiation time, that is, the time from perturbation to foot lift, between PwS and age-matched controls. Other studies, however, have shown that PwS have delayed paretic leg muscle onset latencies following perturbations while standing compared with their nonparetic limb and compared with healthy controls.^32,45-48 These studies used EMG analysis while we and Salot et al³⁰ used kinematic analysis of limb movement. Thus, differences in measurement methods can explain these differences. Our results suggest that although motor nerve conduction velocities might be slower in PwS,^49,50 PwS initiated a step response comparable to healthy controls. Previous study showed that healthy individuals were able to “resist” stepping and thus demonstrated delay foot lift by decelerating the moving center of mass toward the base of support by a greater limb muscle strength and by performing arm and trunk movements.⁹ Thus, unlike PwS, healthy controls might have delayed the initiation of reactive step response. In the current study, 70% of steps in response to surface translations toward the paretic side were initiated with the paretic leg. The ability to swing the paretic leg rapidly (ie, step velocity) to the target location depends on neuro-motor mechanisms related to the build-up of muscle force and power generation. A decreased muscle force and power generation capacity of the paretic leg muscles was found previously.⁵¹ This may be a consequence of reduction in muscle cross sectional area and impaired ability to recruit large amount of motor units in the paretic muscles after stroke.^52-54

In response to forward, backward, and toward nonparetic side surface translations, most step responses were initiated with the nonparetic leg (78.9%, 73.7%, and 89.5%, respectively). The medium effect size that was found in the current study for step length, especially in response to forward and backward perturbations (ES = 0.72 and ES = 0.52, respectively) and the trend toward significant shorter step length in response to surface translations to nonparetic side in PwS compared with healthy controls (Table 5) suggests difficulty to fully bear weight on the paretic leg while executing a long recovery step with the nonparetic leg. Hsiao et al⁵⁵ demonstrated deficits in lateral weight transfer in PwS, which was related to propulsive lower limb muscle force capacity. This was supported by Kirker et al⁵⁶ who found weak gluteus medius hemiparetic muscle response during self-initiated sideways movement before stepping with the nonparetic leg. The higher fall rates in response to nonparetic surface translations, that is, when fall was initiated toward the paretic side and the higher single step threshold in response to nonparetic versus paretic surface translations further suggest difficulty to accept full body weight on the paretic leg when initiating a step with the nonparetic leg. This was supported by de Kam et al⁵⁷ demonstrating that weightbearing asymmetry in favor of the nonparetic leg in PwS, increased the step threshold in response to lateral surface translation toward the nonparetic side.

Balance recovery impairments in PwS were more pronounced in response to forward surface translations (ie, backward loss of balance) compared with backward surface translations (ie, forward loss of balance). Multiple step thresholds and single step thresholds were significantly lower in response to forward versus backward surface translations in PwS and healthy controls (Table 4). PwS also experienced higher fall rates in response to forward versus backward surface translations (9 falls vs 1 fall, Table 3). This was also found previously in patients with Parkinson’s disease,^36,37 multiple sclerosis,³⁸ and healthy subjects⁵⁸ and might be a result of less effective visual feedback and/or due to a smaller distance of the COM from base of support provided by the feet in backward than forward loss of balance.^38,58 The high frequency of laboratory falls among PwS found in our study together with data indicating high fall rates especially during the first 6 months after discharge from hospitalization^3-6 highlights the importance of reactive balance control assessment during this particular period. Assessing and training reactive balance abilities, especially in response to lateral perturbation is important since a sizeable proportion of falls involves falling toward the paretic side, which may result in severe injury such as hip fracture.^4,6

This study has several limitations. First, our sample was fairly small and restricted to subacute PwS hospitalized in a rehabilitation setting who are able to walk independently, so results cannot be generalized to very high functioning community dwelling PwS or to PwS that need assistance during standing and walking. Also, it should be noted that the interpersonal variance of PwS in our study was relatively high in terms of brain lesion characteristics, severity of sensory-motor impairments and clinical measures of balance and gait, yet all PwS in our study were able to walk independently or under supervision (ie, wheelchair-bound patients were excluded) and so these are the patients who are more prone to falls and they are the ones with whom it is important to assess reactive balance abilities. Second, differences in spatiotemporal parameters of the step response were analyzed by perturbation direction, but we did not further classify responses by stepping leg or stepping strategy, that is, crossover step, loaded leg side step, and so on, since there were only few of each of these stepping strategies so a larger sample of PwS is required. Fourth, in the current study we did not examine the association between reactive balance capacity and clinical tests that also provide perturbations of balance such as the POMA and the miniBEST.

In conclusion, PwS demonstrated impairments in balance recovery capacity, that is, low multiple step thresholds and fall thresholds. Fall threshold is an important parameter that may serve as a more direct measure of inability to prevent a fall in real life in PwS. Also, reactive balance impairments were more pronounced when the unexpected loss of balance was toward the paretic side (ie, surface translation toward the nonparetic side) and toward backward (forward surface translation). Results highlight the importance of assessing reactive balance in response to different unexpected perturbation directions and intensities, which may better identify reactive balance impairments and may help target interventions aiming to improve reactive balance response and reduce falls in PwS. Further study is needed to explore the ability of reactive balance capacity assessment (ie, fall threshold and multiple step threshold) to predict future falls in PwS and to examine whether reactive balance capacity and reactive step length and velocity can be improved by specific training in subacute PwS.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was partially supported by a grant from the Ben-Gurion University, by the Helmsley Charitable Trust through the Agricultural, Biological and Cognitive Robotics Initiative of Ben-Gurion University of the Negev, by a trust from the Loewenstein Rehabilitation Hospital for the doctoral program (Shirley Handelzalts) and by Raphael Rozin prize for excellent study in rehabilitation.

ORCID iD

Itshak Melzer

References

Davenport

Dennis

Wellwood

Warlow

CP.

Complications after acute stroke. Stroke. 1996;27:415-420.

Forster

Young

Incidence and consequences of falls due to stroke: a systematic inquiry. BMJ. 1995;311:83-86.

Batchelor

Mackintosh

Said

Hill

KD.

Falls after stroke. Int J Stroke. 2012;7:482-490.

Weerdesteyn

de Niet

van Duijnhoven

Geurts

AC.

Falls in individuals with stroke. J Rehabil Res Dev. 2008;45:1195-1213.

Langhorne

Stott

Robertson

, et al. Medical complications after stroke: a multicenter study. Stroke. 2000;31:1223-1229.

Ramnemark

Nilsson

Borssén

Gustafson

Stroke, a major and increasing risk factor for femoral neck fracture. Stroke. 2000;31:1572-1577.

Neuman

Silber

Magaziner

Passarella

Mehta

Werner

RM.

Survival and functional outcomes after hip fracture among nursing home residents. JAMA Intern Med. 2014;174:1273-1280.

Winstein

Stein

Arena

, et al. Guidelines for adult stroke rehabilitation and recovery: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2016;47:e98-e169.

Maki

McIlroy

WE.

The role of limb movements in maintaining upright stance: the “change-in-support” strategy. Phys Ther. 1997;77:488-507.

10.

Hyndman

Ashburn

Stack

Fall events among people with stroke living in the community: circumstances of falls and characteristics of fallers. Arch Phys Med Rehabil. 2002;83:165-170.

11.

Mansfield

Wong

McIlroy

, et al. Do measures of reactive balance control predict falls in people with stroke returning to the community? Physiotherapy. 2015;101:373-380.

12.

Maeda

Kato

Shimada

Predicting the probability for fall incidence in stroke patients using the Berg Balance Scale. J Int Med Res. 2009;37:697-704.

13.

Podsiadlo

Richardson

The timed “Up & Go”: a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc. 1991;39:142-148.

14.

Flansbjer

Holmback

Downham

Patten

Lexell

Reliability of gait performance tests in men and women with hemiparesis after stroke. J Rehabil Med. 2005;37:75-82.

15.

Jonsdottir

Cattaneo

Reliability and validity of the dynamic gait index in persons with chronic stroke. Arch Phys Med Rehabil. 2007;88:1410-1415.

16.

Tinetti

ME.

Performance-oriented assessment of mobility problems in elderly patients. J Am Geriatr Soc. 1986;34:119-126.

17.

Leddy

Crowner

Earhart

GM.

Utility of the Mini-BESTest, BESTest, and BESTest sections for balance assessments in individuals with Parkinson disease. J Neurol Phys Ther. 2011;35:90-97.

18.

Robinovitch

Feldman

Yang

, et al. Video capture of the circumstances of falls in elderly people residing in long-term care: an observational study. Lancet. 2013;381:47-54.

19.

Wolfson

Whipple

Amerman

Kleinberg

Stressing the postural response. A quantitative method for testing balance. J Am Geriatr Soc. 1986;34:845-850.

20.

Chandler

Duncan

Studenski

SA.

Balance performance on the postural stress test: comparison of young adults, healthy elderly, and fallers. Phys Ther. 1990;70:410-415.

21.

Rogers

Hedman

Johnson

Cain

Hanke

TA.

Lateral stability during forward induced stepping for dynamic balance recovery in young and older adults. J Gerontol A Biol Sci Med Sci. 2001;56:M589-M594.

22.

Jensen

Brown

Woollacott

MA.

Compensatory stepping: the biomechanics of a preferred response among older adults. Exp Aging Res. 2001;27:361-376.

23.

Mille

Rogers

Martinez

, et al. Thresholds for inducing protective stepping responses to external perturbations of human standing. J Neurophysiol. 2003;90:666-674.

24.

Pavol

Runtz

Edwards

Pai

YC.

Age influences the outcome of a slipping perturbation during initial but not repeated exposures. J Gerontol A Biol Sci Med Sci. 2002;57:M496-M503.

25.

Maki

Edmondstone

McIlroy

WE.

Age-related differences in laterally directed compensatory stepping behavior. J Gerontol A Biol Sci Med Sci. 2000;55:M270-M277.

26.

Hilliard

Martinez

Janssen

Edwards

Mille

Zhang

Rogers

MW.

Lateral balance factors predict future falls in community-living older adults. Arch Phys Med Rehabil. 2008;89:1708-1713.

27.

Inness

Mansfield

Lakhani

Bayley

McIlroy

WE.

Impaired reactive stepping among patients ready for discharge from inpatient stroke rehabilitation. Phys Ther. 2014;94:1755-1764.

28.

Martinez

Mille

Zhang

Rogers

MW.

Stepping in persons poststroke: comparison of voluntary and perturbation-induced responses. Arch Phys Med Rehabil. 2013;94:2425-2432.

29.

Patel

Bhatt

Does aging with a cortical lesion increase fall-risk: examining effect of age versus stroke on intensity modulation of reactive balance responses from slip-like perturbations. Neuroscience. 2016;333:252-263.

30.

Salot

Patel

Bhatt

Reactive balance in individuals with chronic stroke: biomechanical factors related to perturbation-induced backward falling. Phys Ther. 2016;96:338-347.

31.

Mackintosh

Hill

Dodd

Goldie

Culham

Falls and injury prevention should be part of every stroke rehabilitation plan. Clin Rehabil. 2005;19:441-451.

32.

de Kam

Roelofs

JMB

Bruijnes

AKBD

Geurts

ACH

Weerdesteyn

The next step in understanding impaired reactive balance control in people with stroke: the role of defective early automatic postural responses. Neurorehabil Neural Repair. 2017;31:708-716.

33.

Batcir

Sharon

Shani

, et al. The inter-observer reliability and agreement of lateral balance recovery responses in older and younger adults. J Electromyogr Kinesiol. 2018;40:39-47.

34.

Marigold

Eng

Dawson

Inglis

Harris

Gylfadottir

Exercise leads to faster postural reflexes, improved balance and mobility, and fewer falls in older persons with chronic stroke. J Am Geriatr Soc. 2005;53:416-423.

35.

Honeycutt

Nevisipour

Grabiner

MD.

Characteristics and adaptive strategies linked with falls in stroke survivors from analysis of laboratory-induced falls. J Biomech. 2016;49:3313-3319.

36.

Carpenter

Allum

Honegger

Adkin

Bloem

BR.

Postural abnormalities to multidirectional stance perturbations in Parkinson’s disease. J Neurol Neurosurg Psychiatry. 2004;75:1245-1254.

37.

Dimitrova

Horak

Nutt

JG.

Postural muscle responses to multidirectional translations in patients with Parkinson’s disease. J Neurophysiol. 2004;91:489-501.

38.

Peterson

Huisinga

Spain

Horak

FB.

Characterization of compensatory stepping in people with multiple sclerosis. Arch Phys Med Rehabil. 2016;97:513-521.

39.

Klein

DeHaven

JJ.

Accuracy of three-dimensional linear and angular estimates obtained with the Ariel Performance Analysis System. Arch Phys Med Rehabil. 1995;76:183-189.

40.

Crenshaw

Kaufman

KR.

The intra-rater reliability and agreement of compensatory stepping thresholds of healthy subjects. Gait Posture. 2014;39:810-815.

41.

Sawilowsky

New effect size rules of thumb. J Mod Appl Stat Methods. 2009;8;597-599.

42.

Marigold

Eng

JJ.

Altered timing of postural reflexes contributes to falling in persons with chronic stroke. Exp Brain Res. 2006;171:459-468.

43.

Kajrolkar

Yang

Pai

Bhatt

Dynamic stability and compensatory stepping responses during anterior gait-slip perturbations in people with chronic hemiparetic stroke. J Biomech. 2014;47:2751-2758.

44.

Mansfield

Inness

Wong

Fraser

McIlroy

WE.

Is impaired control of reactive stepping related to falls during inpatient stroke rehabilitation?

Neurorehabil Neural Repair. 2013;27:526-533.

45.

Di Fabio

. Lower extremity antagonist muscle response following standing perturbation in subjects with cerebrovascular disease. Brain Res. 1987;406:43-51.

46.

Di Fabio

Badke

MB.

Influence of cerebrovascular accident on elongated and passively shortened muscle responses after forward sway. Phys Ther. 1988;68:1215-1220.

47.

Di Fabio

Badke

Duncan

Adapting human postural reflexes following localized cerebrovascular lesion: analysis of bilateral long latency responses. Brain Res. 1986;363:257-264.

48.

Marigold

Eng

Inglis

JT.

Modulation of ankle muscle postural reflexes in stroke: influence of weight-bearing load. Clin Neurophysiol. 2004;115:2789-2797.

49.

Martinez

AC.

Electrophysiological study in hemiparetic patients. Electromyography, motor conduction velocity, and response to repetitive nerve stimulation. Electromyogr Clin Neurophysiol. 1983;23:139-148.

50.

Chokroverty

Medina

Electrophysiological study of hemiplegia. Motor nerve conduction velocity, brachial plexus latency, and electromyography. Arch Neurol 1978;35:360-363.

51.

Kristensen

Stenager

Dalgas

Muscle strength and poststroke hemiplegia: a systematic review of muscle strength assessment and muscle strength impairment. Arch Phys Med Rehabil. 2017;98:368-380.

52.

Ryan

Dobrovolny

Smith

Silver

Macko

RF.

Hemiparetic muscle trophy and increased intramuscular fat in stroke patients. Arch Phys Med Rehabil. 2002;83:1703-1707.

53.

Hafer-Macko

Ryan

Ivey

Macko

RF.

Skeletal muscle changes after hemiparetic stroke and potential beneficial effects of exercise intervention strategies. J Rehabil Res Dev. 2008;45:261-272.

54.

Suresh

Rymer

Suresh

NL.

Assessing altered motor unit recruitment patterns in paretic muscles of stroke survivors using surface electromyography. J Neural Eng. 2015;12:066001.

55.

Hsiao

Gray

Creath

Binder-Macleod

Rogers

MW.

Control of lateral weight transfer is associated with walking speed in individuals post-stroke. J Biomech. 2017;26;72-78.

56.

Kirker

Simpson

Jenner

Wing

AM.

Stepping before standing: hip muscle function in stepping and standing balance after stroke. J Neurol Neurosurg Psychiatry. 2000;68:458-464.

57.

de Kam

Kamphuis

Weerdesteyn

Geurts

ACH.

The effect of weight-bearing asymmetry on dynamic postural stability in people with chronic stroke. Gait Posture. 2017;53:5-10.

58.

Oude Nijhuis

Allum

Borm

Honegger

Overeem

Bloem

BR.

Directional sensitivity of “first trial” reactions in human balance control. J Neurophysiol. 2009;101:2802-2814.

Insufficient Balance Recovery Following Unannounced External Perturbations in Persons With Stroke

Abstract

Keywords

Introduction

Methods

Participants

Study Protocol

Data Collection and Analysis

Sample Size Estimation

Statistical Analysis

Results

Fall Threshold

Multiple Step Threshold

Single Step Threshold

Spatiotemporal Parameters of Recovery Step

Step Initiation

Step Length

Step Velocity

Discussion

Footnotes

Declaration of Conflicting Interests

Funding

ORCID iD

References