Abstract
Background During the physical rehabilitation of individuals poststroke, therapists are challenged to provide sufficient amounts of task-specific practice in order to maximize outcomes of multiple functional skills within limited visits. Basic and applied studies have suggested that training of one motor task may affect performance of biomechanically separate tasks that utilize overlapping neural circuits. However, few studies have explicitly investigated the impact of training one functional task on separate, nonpracticed tasks.
Objective The purpose of this preliminary study was to investigate the potential gains in specific nonlocomotor assessments in individuals poststroke following only stepping training of variable, challenging tasks at high aerobic intensities.
Methods Individuals with locomotor deficits following subacute and chronic stroke (n=22) completed a locomotor training paradigm using a repeated-measures design. Practice of multiple stepping tasks was provided in variable environments or contexts at high aerobic intensities for ≥40 sessions over 10 weeks. The primary outcome was timed Five-Times Sit-to-Stand Test (5XSTS) performance, with secondary measures of sit-to-stand kinematics and kinetics, clinical assessment of balance, and isometric lower limb strength.
Results Participants improved their timed 5XSTS performance following stepping training, with changes in selected biomechanical measures. Statistical and clinically meaningful improvements in balance were observed, with more modest changes in paretic leg strength.
Conclusions The present data suggest that significant gains in selected nonlocomotor tasks can be achieved with high-intensity, variable stepping training. Improvements in nonpracticed tasks may minimize the need to practice multiple tasks within and across treatment sessions.
More than 80% of survivors of stroke have motor impairments that contribute to limitations in performance of many functional tasks.1 Therapists often utilize multiple interventions during clinical sessions to mitigate these deficits,2,3 which may account for the relatively small amounts of practice of any one specific task.3,4 Such limited practice may contribute to limited functional gains4 and represents a disconnect from literature suggesting substantial amounts of task-specific practice are often required to maximize outcomes.1,5 For example, previous data suggested that therapeutic sessions focused on only providing large amounts of stepping practice elicited significant gains in locomotor performance, particularly compared with interventions providing limited stepping practice.4,6–9 Similarly, repeated training of a separate task, such as repeated sit-to-stand (STS) performance, resulted in substantial gains in this task.10
A potential concern of providing substantial practice of only one functional task may be the reduced attention on impairments or other functional activities. For example, studies focused only on stepping training and without additional therapy often demonstrate minimal improvements in clinical assessment of balance (ie, Berg Balance Scale [BBS]),4,11,12 STS performance,11 or lower extremity strength.13 This lack of “transfer,” or improvements in untrained tasks following training of separate tasks, is consistent with data from animal models of spinal cord injury (SCI) detailing minimal improvements in static standing following only stepping practice.14 These findings also are consistent with the larger body of literature regarding task transfer,15 where improvements in an untrained task are thought to depend largely on their biomechanical similarities to trained behaviors.16
Many clinical and research interventions indirectly attempt to exploit the concept of task transfer to improve locomotor function, although biomechanical differences between trained tasks and tested behaviors may minimize gains. For example, lower extremity strengthening17 or standing balance activities18–20 both target primary determinants of upright locomotion21 and are often expected to improve walking ability. However, these studies demonstrated small and inconsistent changes in walking function. Similarly, practice of only STS performance resulted in minimal improvements in walking function.10 Providing sufficient task-specific practice for multiple behaviors despite limited therapeutic sessions represents a substantial challenge in stroke rehabilitation.22
Although the importance of task specificity has been described, more recent data in animal models of SCI suggest that strict specificity of motor task practice may not be warranted. For example, rodents with SCI that were provided stepping training in variable directions on a treadmill23 or overground over obstacles and stairs24 elicited greater improvements in forward treadmill stepping compared with rodents trained with forward treadmill training alone. The variable stepping paradigms often resulted in reduced amounts of practice secondary to the difficulty of the task, although the increased difficulty may have been critical to the observed improvements. Transfer of variable stepping tasks to forward treadmill stepping also may be due to the biomechanical similarities between tasks (ie, stepping) and could provide a rationale to structure therapeutic activities such that multiple tasks could be improved with practice of a few selective behaviors.
Limited studies have explicitly evaluated whether providing high-intensity, variable locomotor training may enhance performance of nonlocomotor outcomes in individuals poststroke. For example, stepping training in multiple directions, around obstacles, or over uneven terrain may enhance locomotor function while challenging postural stability25,26 and may facilitate improvements with repeated training. Similarly, stepping up and down curbs or stairs is biomechanically similar to STS performance,27 and repeated stair/step climbing may result in improved STS performance and strength of muscles subserving these tasks.28 Whether providing stepping training exclusively, although in variable contexts (tasks and environments), can improve nonlocomotor behaviors has not been evaluated.
The purpose of this preliminary study was to investigate the potential impact of dynamic stepping training in variable contexts on nonlocomotor assessments in individuals with subacute and chronic stroke. Using a nonrandomized, repeated-measures design, we were specifically interested in the effects of dynamic stepping practice on STS performance, postural stability during upright balance tasks, and selected lower extremity impairments. Improved performance of nonlocomotor tasks following only stepping training may allow clinicians to prioritize treatment activities to maximize outcomes across multiple tasks.
Method
The present preliminary study represents an analysis of changes in nonlocomotor activities in individuals poststroke following an 8- to 10-week stepping training paradigm performed at high aerobic intensities and in variable contexts (ie, tasks and environments). The feasibility and preliminary efficacy of this intervention on locomotor outcomes have been described previously.29 Here we describe changes in nonlocomotor tasks and impairments following this intervention.
Participants
Inclusion criteria consisted of a history of unilateral, supratentorial stroke >1 month previously; 18 to 75 years of age; impaired locomotor function such that participants' walking function could range from requiring moderate physical assistance to walk30 (participant performing 50%–74% of the task) to requiring no physical assistance to ambulating 10 m overground with or without assistive devices and bracing with self-selected speed (SSS) <0.9 m/s; ability to sit unsupported for 30 seconds; Mini-Mental Status Examination score ≥23/30; and medical clearance to participate. Exclusion criteria included: cardiovascular instability, inability to ambulate >45.7 m (150 ft) prior to stroke, history of other central nervous system injury, and inability to adhere to study requirements. Participants could not be concurrently enrolled in a physical therapy treatment program. All participants provided written informed consent.
Of the initial 25 participants enrolled, data are presented only for the 22 participants who completed ≥4 weeks of training: 12 with subacute stroke (1–6 months) and 10 with chronic stroke (>6 months). The 3 individuals who were excluded (all with chronic stroke) could not complete the protocol due to exercise intolerance, relocation to another city, and a previously undiagnosed cardiovascular disorder that limited training. Demographics and baseline walking measures are provided in Table 1, and a CONSORT diagram is provided in Figure 1.
Demographic and Baseline Locomotor Assessments in Participants With Subacute and Chronic Strokea
CONSORT flow diagram of the study.
Experimental Protocol
A repeated-measures design was used to evaluate alterations in nonlocomotor behaviors following ≤40 one-hour sessions of stepping training provided at high aerobic intensities in variable contexts over 8 to 10 weeks (4–5 sessions/week). Preliminary baseline (pre-baseline) clinical assessments of scores obtained while performing the Five-Times Sit-to-Stand Test (5XSTS) as quickly as possible and of BBS scores were done on individuals with chronic hemiparesis 4 to 5 weeks prior to baseline testing to evaluate the stability of these measures. Participants with subacute stroke were not assessed pre-baseline due to anticipated neurological recovery. Clinical outcomes for the 5XSTS and BBS were assessed on all participants at baseline and repeated every 4 to 5 weeks (≤20 sessions) at mid-training and posttraining, with a 3-month follow-up. Additional testing of isometric strength and biomechanical measures of the 5XSTS were completed only at baseline and posttraining.
The intervention focused solely on reciprocal stepping in specific directions (forward, backward, or sideways versus stepping in place) with additional perturbations or challenges for up to 40 minutes and with rest breaks as needed. The average (±SD) number of steps performed per training session was 2,887±780 (subacute stroke group: 2,845±869; chronic stroke group: 2,967±722). The aerobic training intensity was targeted at 70% to 80% of their maximum heart rate reserve (HRR)31–33 and monitored continuously. For participants prescribed β-blockers, heart rate range was reduced by 10 to 15 bpm according to age,34 and scores of 15 to 17 on the Rating of Perceived Exertion (RPE) scale were targeted.31 Nineteen of 22 participants achieved the targeted intensity (either 70% HRR or a rating of perceived exertion of 15) within the first session; all others reached the targeted intensity by the sixth session and were able to reach this intensity at every session thereafter. The intervention required therapists to continually monitor physiologic and behavioral responses and to modify tasks and training demands to ensure intensity and variability of practice consistent with participants' ability.
Sit-to-stand performance was not explicitly practiced during training but was performed as necessary to initiate walking from a seated position using assistance and upper extremity support as necessary without verbal instructions. Conventional static balance training (ie, weight-shifting, pre-gait activities) and isolated (nonstepping) strengthening activities were never practiced. Rest breaks were allowed as needed throughout the session (approximately 4–8 per session) and were performed in a sitting or standing position, although no instructions or training on STS performance or standing balance were provided.
The training protocol has been described in detail previously.29 For the first 2 weeks (8–10 sessions), only forward treadmill stepping was performed, with a focus on the highest speeds tolerated within targeted aerobic intensities to improve exercise tolerance. Weight support was provided with a counterweight harness system to minimize knee buckling. Ankle-foot orthoses and posterior knee braces were allowed to minimize orthopedic concerns. Participants could use the handrail but were encouraged to minimize upper extremity use. Training over the remaining 6 weeks was divided in 10-minute increments between treadmill training with progressively increasing speeds (described above), training of variable walking skills on the treadmill, overground stepping training, and stair climbing. Treadmill training of variable walking skills was accomplished by altering tasks to challenge specific biomechanical gait components, including postural stability/weight support, limb swing, and propulsion.35 For example, stepping with reduced handrail use or in multiple directions, or both, or over obstacles placed on the treadmill was encouraged to challenge postural stability and balance. If participants could complete limb swing with a positive step length (swing limb advancing past the stance limb) without assistance, leg weights were applied or obstacles were placed on the treadmill, or both. Perturbations to challenge forward propulsion included use of a weighted vest, elastic resistance at the waist, or increased treadmill incline. Training tasks varied in difficulty according to each participant's specific abilities, with 2 to 5 tasks alternated and repeated within the 10-minute duration. Treadmill speeds were adjusted to ensure continuous stepping and maintain targeted aerobic intensities.
Similar challenges were provided during overground training and stair climbing, with assistance as needed for task completion and gradually reduced assistance as tolerated. Overground training included fast walking overground and in multiple directions, stepping on uneven, narrow, or compliant surfaces; negotiating around or over obstacles; and attempts to minimize use of assistive devices. Participants were guarded by therapists or utilized an overhead suspension system for safety. Stair climbing was conducted over static or rotating stairs (StepMill, StairMaster, Vancouver, Washington), with assistance for task completion and progression to higher speeds and reduced hand rail use.
Measures
The primary outcome measure in this study was timed performance during the 5XSTS; this task was chosen due to its functional importance36,37 and lack of biomechanical similarity to upright walking. A standard-height chair (44.5 cm [17.5 in]) was utilized for all assessments, with 7.6-cm (3-in) increments up to 59.7 cm (23.5 in) for participants who could not complete the task at lower heights (seat height was constant at all assessments for an individual participant). The chair was stabilized to prevent movement, and participants could not use their upper extremities.
Starting in the sitting position, participants were instructed to ascend to full standing and descend back to the seat 5 consecutive times as quickly as possible. A successful STS repetition was recorded if the greater trochanter reached ≥90% of standing height (determined visually with a measuring stick) and completed when the participant returned to the seat. Prior to each assessment, participants were asked to perform 1 or 2 STS practice trials and 1 trial of 2 or 3 consecutive attempts performed as rapidly as possible to ensure participants understood the task demands. Total 5XSTS duration was recorded with a digital stopwatch from the investigator command of “go” to the end of the final descent. Assessment of 5XSTS scores was done once on the level floor surface at all assessments, with additional trials at baseline and posttraining with simultaneous biomechanical data collection. Kinematic and kinetic data were collected at 100 Hz and were recorded using an 8-camera motion capture system (Motion Analysis Corp, Santa Rosa, California), 32 reflective markers on the lower limbs and pelvis (modified Cleveland Clinic marker set), and bilateral forceplates under a nonmoving spilt-belt treadmill (Bertec Corp, Columbus, Ohio). The chair was placed on the treadmill, and forces were normalized (“zeroed”) with the chair.
Additional secondary measures included assessment of standing postural stability using the BBS and isometric strength. The BBS was evaluated at all assessments using standardized protocols to ensure testing consistency (eg, guidelines for number of trials allowed and assistance as warranted), with specific emphasis with regard to paretic limb use during tandem and single-limb stance.38 Assessment of bilateral leg strength was performed at baseline and immediately following training and was evaluated under isometric conditions using a dynamometer (Biodex Medical Systems Inc, Shirley, New York) instrumented with a 6-degree-of-freedom load cell. Hip flexors and extensors were tested in the supine position with the hip flexed to 90 degrees. Knee flexors and extensors were tested in a sitting position (hip flexion approximately 80°) with the knee at 90 degrees during extensor testing and at 30 degrees during flexor testing. Moment arm lengths were consistent for each participant. For ankle strength, patients sat with the knee fully extended and foot secured in a footplate, with the ankle at 30 degrees of plantar flexion and at 0 degrees to assess dorsiflexor and plantar-flexor strength, respectively. Following multiple practice attempts, participants were asked to generate maximal-effort contractions for 3 to 5 seconds until torque decreased as visualized on an oscilloscope. Three trials were completed, with ≥1 minute between efforts. Torque signals were low-pass filtered at 200 Hz and acquired at 1,000 Hz using custom-designed software.
Five physical therapists were trained on the standardization of the outcomes measures used. In order to minimize bias, the therapists did not examine participants' performance from previous assessments prior to re-evaluation.
Data Analysis
For biomechanical 5XSTS measures, a bilateral 6-degree-of-freedom model of each participant's lower limbs (feet, shanks, thighs, and pelvis) was created from the reflective marker data during static standing using Visual3D (C-Motion Inc, Germantown, Maryland). Marker data were filtered (low-pass, second-order Butterworth filter; cutoff frequency=10 Hz), and the model for each testing day was applied to STS trials. Ankle, knee, and hip sagittal joint angles were calculated from the transformations between model segments. Sagittal-plane joint moments were calculated from inverse dynamics using filtered ground reaction force data (low-pass, second-order Butterworth filter; cutoff frequency=20 Hz) and joint angles. Sagittal joint powers were calculated as the product of joint moment and joint angular velocity. Lower limb inertial properties were estimated based on the participant model and anthropometric measurements of limb positions and joint centers.
Kinematic analyses focused on total time from initial ascent to end of the final descent and removed participant or therapist reaction time. Ascent time for each successful STS attempt was determined using the time from initial vertical rise of the sacral marker to peak vertical position at standing. Descent time was the duration from peak vertical position to return to the seated position. Sitting time was calculated as the duration between the end of each descent and the start of the next successful attempt and, therefore, included failed STS attempts. For kinetic analyses, moments (in newton-meters) and power (in watts) were normalized to body weight (in kilograms) and focused on hip and knee extensor moments and power separately for each ascent (STS) and descent (stand-to-sit) phase. Data were analyzed only when participants were not in contact with the chair and were averaged over the 5 STS trials, with mean peak power and moments presented. Ankle moments and power are not reported because of to the use of ankle orthoses by 10 participants and the minimal contributions of ankle kinetics to STS behaviors.39,40
For secondary measures, scores from the BBS were summed. For strength testing, torque signals were low-pass filtered (second-order Butterworth filter, cutoff frequency=20 Hz), and strength was determined as the average torque ± 50 milliseconds from the peak torque registered using custom-designed software. The 3 trials for each muscle group were averaged and normalized to body weight (in newton-meters/kilogram).
Data are presented as mean (±SD) in the text and tables, with standard error bars in the figures. Kolmogorov-Smirnov assessments verified parametric distribution of ratio data. No significant difference was observed between pre-baseline and baseline data in participants with chronic stroke, and subsequent analysis focused on comparisons with baseline data. Repeated-measures analysis of variance was used to determine differences in timed 5XSTS performance from baseline through follow-up, with Friedman assessment for BBS; Tukey post hoc or Wilcoxon comparisons were utilized to evaluate specific differences. Paired t tests were used to compare baseline and posttraining biomechanical data for 5XSTS and isometric strength. If significant differences were observed (α=.05), data from individuals with subacute and chronic stroke were analyzed separately.
Correlation analyses were performed to evaluate potential contributions of improvements in balance and paretic leg strength to changes in STS performance and locomotor function (reported previously29; see “Results” section). Pearson or Spearman coefficients were calculated, as appropriate, between change scores from baseline to posttraining for each of the measures identified above and previously for locomotor outcomes of SSS, fastest possible speed, 6-minute walk distance, and daily stepping activity (reported in the “Results” section).
Role of the Funding Source
Funding for the study was provided by National Institute on Disability and Rehabilitation Research grant H133B031127.
Results
Twenty-two participants (12 with subacute stroke, 10 with chronic stroke) completed at least 4 weeks of training, averaging 36 sessions (range=17–40) over 8 to 10 weeks. All but 3 participants (all in the subacute stroke group) walked overground without therapist assistance at baseline. Improvements in locomotor function from baseline to posttraining were described in detail previously29 and included improvements in SSS (subacute stroke group: 0.33±0.20; chronic stroke group: 0.23±0.17 m/s), fastest speed (subacute stroke group: 0.54±0.35 m/s; chronic stroke group: 0.39±0.23 m/s), 6-minute walk distance (subacute stroke group: 144±98 m; chronic stroke group: 89±60 m), and daily stepping (subacute stroke group: 1,216±454 steps/day; chronic stroke group: 658±1,367 steps/day). All improvements were statistically significant (P<.001) and maintained at follow-up.29
5XSTS Performance
Of the 22 participants who completed ≥4 weeks of training, 3 (2 with subacute stroke, 1 with chronic stroke) were not able to initially perform the 5XSTS task at baseline at any seat height without use of their upper extremities, and their data were excluded from analysis. For the remaining 19 participants, average timed 5XSTS performance at baseline was 25±17 seconds (subacute stroke group: 27±12 seconds; chronic stroke group: 22±16 seconds) at the lowest possible seat height (12 participants at 44.5 cm, 4 at 52.1 cm [20.5 in], and 3 at 59.7 cm). Timed 5XSTS performance at pre-baseline in participants with chronic stroke was not significantly different from baseline performance (23±16 seconds, P>.05).
Changes in 5XSTS performance were significant throughout and following training. Repeated-measures analysis of variance revealed improvements in timed performance across all participants (P<.0001), with post hoc assessments indicating differences from baseline to mid-training (decrease of 6.0±12 seconds), posttraining (9.3±8.6 seconds), and follow-up (11±10 seconds) but no differences among the latter 3 assessments (Fig. 2A). Changes from baseline to posttraining in participants with subacute stroke (12±3.8 seconds [41%±19%]) and those with chronic stroke (5.9±5.8 seconds [24%±16%]) were significant (P<.01) in both groups. All 3 participants who could not perform the 5XSTS at baseline were successful at the posttraining and follow-up assessments.
Changes in clinical assessment of (A) Five-Times-Sit-to-Stand Test (5XSTS) scores and (B) Berg Balance Scale (BBS) scores at each assessment (baseline [BSL], mid-training [MID], posttraining [POST], and 3-month follow-up [F/U]), with an additional assessment of participants with chronic stroke 1 month prior to baseline (pre-baseline [PRE-BSL]). Improvements were significant from BSL assessment to MID, POST, and F/U, with no differences among the latter groups. **P<.01.
For kinematic STS measures, Figure 3 (parts A and B) depicts the vertical excursion of the sacral marker of 2 participants during 5XSTS at baseline and posttraining. Improvements in both individuals were characterized by faster ascent and descent times and decreased duration between repeated efforts (“sitting time”), with baseline testing characterized by multiple failed STS attempts (Fig. 3A, ie, rise of sacral marker without successful STS). Decreases in total 5XSTS time (8.4±8.7 seconds), total ascent time (1.4±2.3 seconds), and total descent time (2.1±2.5 seconds) were significant across the study sample (all P<.01), with variable but significant changes in sitting time (5.5±8.8 seconds, P<.05, Fig. 3C).
Five-Times-Sit-to-Stand Test (5XSTS) representative trial for baseline (BSL) to posttraining assessment (POST) for 2 individuals classified as (A) severely impaired (self-selected speed <0.50 m/s) and (B) moderately impaired (self-selected speed=0.50–0.90 m/s). Tracings represents sacral marker vertical position (y-axis) versus time (x-axis), with gray arrows indicating start of successful ascent, black arrows indicating termination of ascent, and gray circles representing end descent. (C) Grouped data for total time, ascent time, descent time, and sitting time. *P<.05, **P<.01.
Kinetic analyses focused on changes in power and moments of the paretic and nonparetic hip and knee, with single-subject kinetic data averaged across 5 repeated STS maneuvers at baseline and posttraining (Fig. 4). Small increases in joint power were demonstrated in the nonparetic limb from baseline to posttraining, with little change in the paretic limb. These data are consistent across the study sample (eTable), with significant changes only in averaged nonparetic knee eccentric power and extensor moment during descent.
Knee and hip power profiles for paretic and nonparetic lower extremities for a representative participant: average over the 5 repeated, successful sit-to-stand attempts. (A) paretic knee, (B) nonparetic knee, (C) paretic hip, (D) nonparetic hip. BSL=baseline, POST=postintervention.
Secondary Measures
Assessment of balance demonstrated baseline BBS scores of 34±14 (subacute stroke group: 31±16; chronic stroke group: 37±12), with no changes from pre-baseline in the chronic stroke group (38±10, P>.05). Following initiation of the stepping training, significant differences were observed at mid-training (change of 5.0±5.5), posttraining (change of 7.5±7.4), and follow-up (change of 7.9±9.0), with no differences among the 3 assessments (Fig. 2B). Improvements from baseline to posttraining were significant in both groups (chronic stroke group: 6.2±5.8; subacute stroke group: 8.7±8.5; both P<.01). In the 3 participants with subacute stroke who were nonambulatory at baseline, average baseline scores of 9.3±5.9 improved to 31±9.8 at posttraining and 35±11 at follow-up (changes of 21±4.2 and 25±12, respectively).
Improvements in isometric strength also were observed (eTable), with significant improvements in most paretic muscle groups. The observed changes, however, were strongly biased toward participants with subacute stroke, with average increases from 60% to 100% (all P<.05 except dorsiflexors and hip extensors). In contrast, improvements ranged from 7% to 27% in participants with chronic stroke, with only the plantar flexors approaching significance (P=.08). Changes in nonparetic leg strength were not significant.
Correlation analyses evaluated potential contributions of changes in balance and strength to changes in timed 5XSTS performance and locomotor outcomes. The eFigure demonstrates selected associations between primary and secondary outcomes, with all potential correlations provided in Table 2. There were no significant associations between changes in BBS scores and changes in 5XSTS scores (eFigure, graph A) or any locomotor outcome (BBS versus SSS shown in the eFigure, graph B). Conversely, changes in strength of selected paretic muscle groups were moderately correlated to primary outcomes, although not statistically significant with the adjusted α=.0014 for multiple correlations. For example, improvements in 5XSTS scores were moderately correlated to changes in paretic hip extensor strength (r=.47, P=.04; eFigure, graph C), with few correlations between other paretic muscles. For locomotor outcomes, changes in all walking outcomes were moderately correlated with changes in paretic knee extensor and flexor strength (range of correlation coefficients=.46–.63, P=.002–.04; see eFigure, graph D, for knee extensors).
Correlation Coefficients of Changes in Five-Times-Sit-to-Stand Test (5XSTS) Scores and Primary Locomotor and Nonlocomotor Outcomes Reported Previously With Secondary Measures of Balance (Berg Balance Scale [BBS]) and Paretic Limb Strengtha
Discussion
The present study investigated changes in nonlocomotor outcomes in participants with hemiparesis poststroke following high-intensity stepping training in variable contexts. Our previous data suggest that this training paradigm is feasible and tolerated by most patients, with significant improvements in locomotor function.29 The secondary analyses presented here detail improvements in nonlocomotor tasks of the 5XSTS and BBS and in lower limb strength without explicit focus on these activities during training.
STS Performance
Rising from a seated to a standing position represents an important functional task often required prior to walking. Specific STS parameters are correlated with selected locomotor behaviors37 and risk of falling in the home and community,36 and STS training has been shown to improve balance41 and decrease fall incidence.42 Accordingly, practice of STS and similar transition behaviors are often given priority in the clinical setting3 in an effort to improve the patient's performance.10,43 The data presented here suggest that clinically important STS improvements may be achieved with this training paradigm without focused STS practice.
Few studies have investigated the possibility of improved STS performance following training of only stepping activities. The findings that changes in 5XSTS scores were significant by mid-training assessment suggest that neural activation strategies versus muscular changes likely contributed to the improve performance, although the training tasks appear to be important. In a study by Globas et al,11 for example, patients with chronic hemiparesis poststroke who completed 6 months of forward treadmill training demonstrated negligible STS improvements from baseline assessments or compared with control interventions. This lack of improvement may have been due to the different biomechanical constraints and neural activation patterns required for successful STS performance versus treadmill stepping. Conversely, studies utilizing closed chain activities emphasizing hip and knee extension (eg, step-ups43 or repeated STS10) showed improved STS performance. Exposure to stepping tasks requiring greater hip or knee extensor activity than level walking (ie, stair climbing,44 stepping on inclined45 or compliant46 surfaces) may have contributed to our findings.
Other factors also may contribute to improved STS performance. For example, repeated testing every 4 to 5 weeks and practice trials performed prior to assessments could result in testing effects across the patients. However, stability of 5XSTS performance in the participants with chronic stroke partially refutes this notion. Patient exposure to STS practice during or outside of training sessions also may have contributed, although there was no focus on STS performance throughout any of the training and no instructions of techniques to improve STS during or outside of sessions. Importantly, the aforementioned treadmill training and STS training studies suggest that the training activities strongly influence the observed changes. Nonetheless, evaluation of outcomes following this training paradigm compared with other therapeutic strategies is certainly warranted.
Despite consistent changes in summed ascent and descent 5XSTS time, nonsignificant changes in paretic hip and knee kinetics were demonstrated. Rather, the largest changes were due to decreased sitting time. Such changes may be due to improved muscle timing and coordination47 to allow more rapid transitions between ascent and descent strategies. The lack of kinetic changes during STS was surprising given the improvements in isometric paretic knee strength and focus only on reciprocal stepping training. These limited gains may be a result of instructions provided during testing, which were only to complete the test as rapidly as possible rather than emphasize paretic limb use. Altered testing instructions48,49 or practicing STS performance with a focus on the paretic limb49 may improve kinetic performance, although dedicated practice directed toward STS may lessen gains in locomotor outcomes.10 In general, however, further studies evaluating biomechanical performance during STS are necessary, particularly following training paradigms, as information regarding altered movement strategies posttraining is extremely limited.
Static Balance and Lower Limb Strength
Analysis of secondary measures revealed improvements in BBS scores and lower limb impairments, with selected differences in participants with subacute and chronic stroke. Although average gains in BBS scores in the subacute stroke group were at or above minimal detectable changes (MDC=6.3 for independent ambulators, 8.1 for those requiring assistance50), improved standing balance following only stepping training may not be surprising given the expected neurological recovery early following injury. Interestingly, however, very few studies have demonstrated improvements in BBS scores in patients with subacute stroke with a focus only on stepping training and without concurrent clinical physical therapy (for details, see Hidler et al51). Without interventions beyond those described here, the mean improvement in BBS scores (ie, 21±4.2 points) in the 3 participants who initially were nonambulatory with practice of only walking activities was striking. Provision of only stepping practice in patients who are nonambulatory varies substantially from traditional interventions that focus on static balance or pre-gait activities prior to gait training.52
As we expected, the improvements in BBS scores in participants with chronic stroke were more modest, with average changes still above MDCs (2.5–4.6).53,54 Conversely, published changes in BBS scores in individuals with chronic stroke following only forward treadmill stepping were consistently smaller (range=1–3 points4,12,55,56). The limited improvements may be reflective of higher functioning in patients in other studies or of a lack of scale precision at higher ranges of balance abilities.38 Alternatively, postural stability may not be sufficiently challenged during forward treadmill training compared with more dynamic stepping activities. Other static and dynamic balance training studies that do not explicitly focus on walking have demonstrated similar or larger improvements in BBS scores (eg, changes of 5–12 points57–59), although improvements in locomotor function reported in these studies were more modest. Direct comparison of other intervention strategies with the present training paradigm is warranted, although these preliminary results suggest that balance improvements are possible following stepping activities that challenge postural stability.
Improvements in lower limb strength also may be expected in patients with subacute stroke, although the limited gains in participants with chronic stroke were surprising considering the neuromuscular demands associated with training. The small changes could suggest greater volitional neuromuscular activity only during functional tasks, although the lack of kinetic changes during the 5XSTS refutes this notion. Previous treadmill training studies have demonstrated small but significant improvements in strength,13 whereas more directed strength training paradigms have shown much greater gains (eg, Clark and Patten60). As with STS and balance training studies, however, gains in walking recovery following strength training are limited.17,60,61 The combined findings again suggest that the biomechanical demands of practiced tasks may be essential to improved STS performance. Practice of stepping tasks that challenge postural stability and increase the demands of specific motor pools (ie, hip/knee extensors) as described here may elicit improvements in multiple tasks (stepping, BBS, 5XSTS) secondary to their similar biomechanical demands.
Limitations and Clinical Significance
Specific limitations include the lack of a control group or blinded assessors for outcome measures detailed here and previously.29 Future blinded, randomized studies may provide more substantive information regarding the utility of this training protocol compared with other interventions. Calculated effect sizes of 1.0 to 1.2 for the 5XSTS and BBS throughout the duration of training, with more variable changes for isometric strength, allow determination of adequate sample sizes necessary for future studies. Although the total number of training sessions (N=36) was relatively large, the significant improvements in both 5XSTS and BBS scores at mid-training assessment (≤5 weeks or 20 sessions) suggest that reduced training sessions may be required to elicit similar gains. Unfortunately, whether further training up to 10 weeks (40 sessions) is necessary to maintain these gains at follow-up is not well established.
In addition, further identification of activities during stepping training that can “transfer” to nonlocomotor behaviors may be significant for future clinical interventions. More directly, clinical application of this or similar training strategies could streamline rehabilitation sessions, as repeated practice of a few behaviors may improve performance of multiple functional tasks.62 Previous data from cross-sectional investigations of inpatient rehabilitation support this hypothesis, where patients poststroke who received aggressive early physical therapy interventions, including greater amount of walking training, tended to “leapfrog” over lower-level rehabilitation activities and demonstrated larger improvements in mobility and performance of selected transitional skills (ie, toilet transfers).63,64 Although the amount and type of practiced tasks could be critical and require further study, such training varies substantially from traditional rehabilitation, where therapists focus on multiple impairments prior to initiating walking and progress treatment incrementally above the patient's previous performance.63,64
In summary, this preliminary study revealed gains in STS performance, BBS scores, and selected lower limb impairments following focused, high-intensity stepping training in variable contexts. Whether such findings are superior to those obtained with other walking or nonwalking interventions warrants additional testing, although the possibility of amelioration of multiple functional deficits and impairments with similarly structured practice sessions may contribute to the development of more efficient rehabilitation strategies.
The Bottom Line
What do we already know about this topic?
Patients poststroke present with multiple impairments and functional deficits that are targeted during physical therapist interventions. Physical therapists often provide many different interventions to improve these deficits, although practice of multiple behaviors during therapy may limit the attention and effort directed towards to any one specific task.
What new information does this study offer?
This study suggests that structuring interventions may facilitate gains in multiple tasks. Specifically, although previous data indicate that stepping practice in variable contexts can improve walking outcomes, the present data suggest that such practice also may contribute to improvements in nonstepping tasks, such as balance and transfers. These findings often are not observed following provision of stepping training in more controlled environments, such as forward treadmill training.
If you're a patient, what might these findings mean for you?
Improvements in nonpracticed tasks could reduce the need to practice multiple behaviors during therapy sessions.
Footnotes
Dr Straube, Ms Kinnaird, and Dr Hornby provided concept/idea/research design. Dr Straube, Ms Holleran, Dr Leddy, Dr Hennessy, and Dr Hornby provided writing. Dr Straube, Ms Holleran, Ms Kinnaird, Dr Leddy, and Dr Hornby provided data collection and analysis. Ms Kinnaird and Dr Hornby provided project management. Dr Hornby provided fund procurement, facilities/equipment, and institutional liaisons. Dr Leddy and Dr Hornby provided study participants. Ms Kinnaird provided clerical support. Dr Straube, Ms Kinnaird, and Dr Hennessy provided consultation (including review of manuscript before submission).
The project was approved by the Northwestern University Institutional Review Board.
Portions of this research were presented at the Combined Sections Meeting of the American Physical Therapy Association; February 8–11, 2012; Chicago, Illinois.
Funding for the study was provided by National Institute on Disability and Rehabilitation Research grant H133B031127.
- Received November 12, 2013.
- Accepted March 7, 2014.
- © 2014 American Physical Therapy Association
References
Issue highlights
