Use of water- and land-based gait training to improve walking capacity in adults with complete spinal cord injury: A pilot study

Abstract

Objective

Little is known regarding the extent to which mobility can be improved using gait-based therapies in individuals with complete spinal cord injury (cSCI). Against this backdrop, the purpose of our study was to document changes in walking capacity following an extended period of underwater treadmill training (UTT) and supplemental overground walk training (OWT) in persons with cSCI.

Design

Longitudinal design.

Setting

University research center.

Participants

Five adults (mean age = 41.2 ± 5.9 years) with motor-complete (AIS A), chronic (mean years post-injury = 3.2 ± 1.6 years) cSCI who had not received epidural spinal cord stimulation (eSCS).

Intervention

Participants underwent one year of UTT (3 walking bouts per day; 2–3 days per week). Once independent stepping activity in the water was observed, OWT, as tolerated, was performed prior to UTT.

Outcome Measure

Walking capacity was evaluated using the Walking Index for Spinal Cord Injury (WISCI-II) prior to UTT (Time 1: T1), six months after the start of UTT (Time 2: T2), and following completion of UTT (Time 3: T3).

Results

Non-parametric analyses revealed a significant time effect (P < .05) for WISCI-II. Pre-planned comparisons revealed no difference in WISCI-II levels measured at T1 (0.20 ± 0.45) and T2 (4.80 ± 4.55) and at T2 (4.80 ± 4.55) and T3 (8.40 ± 1.34). However, the WISCI-II level obtained at T3 (8.40 ± 1.34) was significantly higher compared to the T1 value.

Conclusion

Our preliminary findings demonstrate that in the absence of eSCS, combined UTT and supplemental OWT can improve functional walking capacity in adults with cSCI.

Introduction

A treatment modality to improve walking ability that has gained popularity in the rehabilitation of individuals with spinal cord injury (SCI) is partial body weight-supported training (PBWSTT). In this land-based approach to retraining locomotion, body weight is partially supported by a harnessing system during treadmill walking, walking speed and duration are increased gradually in the context of task-oriented walking practice, and percentage of body weight support is reduced in a systematic fashion as locomotor performance improves. While evidence is limited regarding the effectiveness of ambulatory training in persons with SCI,^1–3 use of PBWST during treadmill and overground walking with manual assistance, robotic assistance, and functional electrical stimulation has shown some potential to aid in the recovery of walking following incomplete SCI (iSCI).^1–4 However, methodological constraints and challenges associated with these gait training strategies, including (1) the need for multiple skilled therapists and labor-intensive demands placed on trainers, (2) use of a harnessing system that can be uncomfortable and sometimes contraindicated in persons with SCI due to orthostatic hypotension and increases in diastolic blood pressure, (3) generation of abnormal patterns of muscle recruitment and reduced locomotor demands that can interfere with long-term recovery, and (4) difficulty in matching voluntary walking efforts with electrical stimulation patterns,^5–9 collectively highlight the importance of evaluating the potential of emerging, user-friendly, activity-related therapies to enhance mobility in persons with SCI.

A self-initiated gait intervention that has remained largely unexamined as a means of improving walking performance in individuals with SCI is underwater treadmill training (UTT). Primarily used in animal rehabilitation and sports medicine facilities, this water-based therapy to improve mobility and physical function in persons with neurological disease is new and innovative. In a typical overground treadmill harnessing system, a given percentage of body weight is supported, but the weight of the legs remains unchanged. Hence, if leg strength is inadequate, external assistance is required to move the lower extremities during walking. Conversely, use of water as an unloading medium during treadmill exercise reduces the weight of the legs and core weight, thus decreasing the strength needed for walking and body support while providing challenging, yet manageable, levels of resistance. In previous work,¹⁰ we have reported gains (P < .05) of moderate to large magnitude (partial η²= .51–.84) in leg strength (57%), balance (39%), preferred and rapid walking speed (34% and 61%, respectively), 6-min walk distance (82%), and daily step activity (121%) in 11 adults with chronic incomplete SCI (iSCI) following eight weeks of UTT featuring personalized levels of body weight support and incremental gains in walking speed and duration. In a companion paper,¹¹ we also demonstrated better cardiac performance during UTT, as shown by an average reduction of 12% in exercise heart rate over the course of three bi-weekly periods during which walking speed remained constant.

Because recovery of walking is not thought to be possible in persons with complete spinal cord injury SCI (cSCI) who display no voluntary motor control below the level of injury,^12–14 the extent to which walking capacity can be improved remains questionable. It is known that the human central nervous system is capable of plasticity after neurotrauma¹⁵ and cortical sensorimotor output maps undergo reorganization following locomotor training in humans who have experienced SCI or cerebral hemispherectomy.^16,17 In addition, recent evidence showing activation of the somatosensory cortex following somatosensory stimulation of below-injury insensate body regions in six of 11 adults with clinically complete traumatic SCI suggests that somatosensory conduction can be preserved in persons with a degree of injury severity between complete and incomplete SCI; i.e. discomplete SCI.¹⁸ From a functional perspective, recovery of intentional overground walking in two adults with chronic motor-complete SCI after receiving programmed epidural spinal cord stimulation (eSCS) and intense and prolonged land-based locomotor training¹⁹ also provides impetus for developing and implementing activity-based treatments to improve walking function in persons who have experienced severe neurological injury. Against this backdrop, the purpose of this exploratory study was to document changes in walking capacity among persons with cSCI who had not received eSCS following regular exposure to underwater treadmill training, a water-based exercise modality that makes stepping movements easier to accomplish, and supplemental overground walk training (OWT).

Methods

Participants

Five adults (males, n = 4; females, n = 1) with chronic (>1-year post-injury) and complete SCI [American Spinal Cord Injury Impairment Scale (AIS) A; absence of sensory and motor function below the level of neurological injury, including the lowest sacral segments (S4-S5)]²⁰ volunteered to participate in this investigation. Study participants were recruited through contacts with local hospitals, medical centers, rehabilitation clinics, and SCI support groups. Criteria for study enrollment included being 21 years of age or older, presence of traumatic or non-traumatic SCI for at least one year, and a medical diagnosis of paraplegia (T1 or below). Study exclusion criteria included presence of complex co-morbidities (e.g. health or neurological diagnoses that could limit involvement in training activities), surgical implantation or removal of a medical device six months prior to training, or use of electrical implants (e.g. eSCS) to promote ambulation. All study volunteers supplied medical documentation confirming their SCI and AIS A rating and obtained medical clearance This study was approved by a university institutional review board and participants provided written informed consent prior to data collection.

Assessment of overground walking capacity

Overground walking capacity was measured on a straight indoor track using the Walking Index for Spinal Cord Injury (WISCI – II).²¹ The WISCI – II, a clinically-useful instrument to assess improvements in gait function and ambulation following rehabilitation,²² is an ordinal scale that captures the degree and types of assistance (i.e. orthotic devices, supporting equipment, human helpers) needed for persons with SCI to walk 10 meters and ranges from 0 (unable to stand and/or participate in assisted walking) to 20 (ambulates with no devices, no braces, and no physical assistance). The WISCI-II has demonstrated excellent reproducibility and interrater and intrarater reliability and moderate to strong criterion and construct validity in adults with chronic SCI.^21,23,24 WISCI-II data were collected by a physical therapist with extensive clinical and research experience administering the instrument to persons with SCI.^10,11 WISCI-II values were obtained on study participants prior to UTT (Time 1: T1), six months after starting UTT (Time 2: T2), and immediately following completion of UTT (Time 3: T3).

Accommodation to underwater treadmill training

Before commencing UTT, study participants accommodated to underwater treadmill walking. The accommodation session was conducted on an underwater treadmill anchored in an enclosed therapy pool (Figure 1). Ankle weights ranging from 0.23 to 0.68 kg were attached to both ankles with Velcro straps to determine the minimum weight necessary to maintain foot contact during treadmill walking and counteract the tendency of the legs to float due to water buoyancy. Absolute levels of water height were established for each participant, based on the minimum water height necessary to provide adequate support while standing, and replicated during accommodation and training sessions. Additionally, water temperature was controlled to maximize comfort, aid in pain management, and assist in managing muscle tone. Once water height and water temperature were established, participants walked for one minute at the slowest speed setting on the underwater treadmill (0.20 m s⁻¹) by having a trainer stand behind each participant and use their feet to lift and position the left and right foot while the treadmill belt was moving. A variety of padded arm rests, fabricated to allow participants of different heights to walk in an upright position to the greatest extent possible, enabled them to rely on upper-extremity support, along with the partial buoyancy of the water environment, while walking. If the water height provided a sufficient level of support during the 1-min walk, this height was used for all subsequent training sessions. After completing the 1-min treadmill walk, participants completed three 5-min walks at 0.20 m s⁻¹, with external assistance provided by a trainer. A recovery period of at least five minutes separated walking bouts.

Figure 1.

Study participant with complete spinal cord injury engaged in underwater treadmill training.

Underwater treadmill training

Following accommodation to underwater treadmill walking, study participants embarked on a year-long training program consisting of two to three sessions of UTT per week performed on alternate days. During the initial phase of training, participants completed three 5-min walking bouts at 0.20 m s⁻¹ during each training session at the water height and water temperature established during treadmill accommodation. Treadmill speed was validated from measures of treadmill belt length and the time required for two treadmill belt revolutions. As noted earlier, stepping was facilitated by having trainers use their feet to lift and position the left and right foot of each participant. At the end of each walking bout, participants identified their overall rating of perceived exertion (RPE) using a modified 10-point Borg scale.²⁵ If RPE was greater than “6” (6 = tired) during a given walk, this was interpreted as an integrated signal of central and peripheral fatigue and participants ceased walking and rested until they were ready to resume and complete the exercise bout. Rest periods of at least five minutes were interspersed between daily walking bouts.

No increases in walking speed or duration were imposed upon study participants until independent stepping activity, defined as a minimum of five consecutive unassisted foot contacts occurring on two successive training sessions, was observed. Once independent stepping emerged during UTT, trainers gradually withdrew physical assistance as higher levels of walking independence were achieved and gradual increases in walking speed and duration were imposed on a weekly or bi-weekly basis (see Table 1), while maintaining previously-established water height levels, until participants were able to walk independently in the water for 45 min (three 15-min bouts) without experiencing fatigue (i.e. RPE ≤ 6), whereupon trainers no longer accompanied participants into the water tank. In limited instances, downward adjustments in walking speed were made to enable completion of a given exercise bout, with the goal of restoring training intensity to the desired setting as soon as possible. Once participants demonstrated the ability to complete three 15-min walking bouts during UTT in a non-fatigued state, overall training volume (intensity x duration) remained constant and emphasis was placed on diminishing reliance on upper-extremity support while walking.

Table 1.

Progression of walking duration and speed following attainment of independent stepping during underwater treadmill training (AIS-UTT).

Week after AIS-UTT	Walking duration	Walking speed (mּ s−1)	Walking speed (mph)
0	3 × 5 min	0.200	0.45
1	3 × 5 min	0.220	0.49
2	3 × 7 min	0.242	0.54
3	3 × 7 min	0.242	0.54
4	3 × 8 min	0.266	0.60
5	3 × 8 min	0.266	0.60
6	3 × 9 min	0.293	0.66
7	3 × 9 min	0.293	0.66
8	3 × 10 min	0.322	0.72
9	3 × 10 min	0.322	0.72
10	3 × 11 min	0.354	0.79
11	3 × 11 min	0.354	0.79
12	3 × 12 min	0.389	0.87
13	3 × 12 min	0.389	0.87
14	3 × 13 min	0.428	0.96
15	3 × 13 min	0.428	0.96
16	3 × 14 min	0.471	1.05
17	3 × 14 min	0.471	1.05
18	3 × 15 min	0.518	1.16
19	3 × 15 min	0.518	1.16
20+	3 × 15 min	0.518	1.16

Week 0 = Week that AIS-UTT was achieved; Week 20+ = protocol followed until study completion.

Supplemental overground walk training

After independent stepping was recorded during UTT, a systematic and individualized program of overground walk training (OWT) was added to supplement UTT (Figure 2). The progression of training levels for OWT was as follows: (1) assisted squats, assisted sit-to-stand in a walker, and assisted stepping, with physical assistance for walker stabilization and blocking of the knees; (2) walking, with physical assistance (as needed) to (a) stabilize a walker, (b) facilitate stepping, and (c) block the knees during stance; (3) referral for lower-leg orthotics to provide support during stance when an independent swing phase was observed, followed by walking with braces and a rolling walker (with physical assistance, as needed) to achieve balance, stability, and positioning to sit in a wheelchair at the conclusion of walking; (4) gradual reduction of physical assistance until independent walking was achieved; and (5) exposure to additional walking challenges (e.g. land-based treadmill walking, ramp walking, walking and opening doors) as locomotor skill improved. Participants completed approximately 30 min of individualized OWT (with provision of ample rest periods) before commencing UTT. Prior to the start of the study, ambulatory goals were established that were personally relevant to each participant. As the study progressed, this goal-setting process evolved and led to higher-level physical activity and purposeful functional targets as improvement in water- and land-based walking ability occurred.

Figure 2.

Study participant engaged in community ambulation during Level 3 of overground walk training (walking with braces and a rolling walker (with physical assistance, as needed)) to achieve balance, stability, and positioning to sit in a wheelchair at the conclusion of walking.

Statistical analysis

Based on the non-normal distribution and ordinal nature of our data, the Friedman test was used to document the presence of overall training-related changes in WISCI-II levels. Pre-planned analyses using the Wilcoxon signed-rank test were employed to identify the location of pairwise differences in overground walking capacity across the three testing periods (T1, T2, T3). Statistical significance was established at P < .05.

Results

Participant characteristics are displayed in Table 2. An average of 29 training sessions (range = 9–62 sessions) was required for participants to register unassisted stepping activity during UTT (Table 3). A video recording of one study participant, documenting the transition from assisted UTT to self-initiated overground walking over the course of the 1-year training period, is included as Supplementary Material.

Table 2.

Descriptive characteristics of study participants.

Participant Number	Sex	Age (years)	Level of Lesion	AIS Grade	Years Post-Injury	Primary Mode of Locomotion
1	F	47	T10	A	4	Manual Wheelchair
2	M	39	T9	A	1.3	Manual Wheelchair
3	M	47	T4	A	5	Manual Wheelchair
4	M	33	T11	A	4	Manual Wheelchair
5	M	40	T11	A	1.6	Manual Wheelchair

M = Male; F = Female; T = thoracic; AIS = American Spinal Cord Injury Association Impairment Scale [Grade A = absence of sensory and motor function below the level of neurological injury, including the lowest sacral segments (S4-S5)].

Table 3.

Number of sessions needed to achieve independent stepping during UTT and WISCI-II scores at T1, T2, and T3.

Participant number	Number of sessions to achieve independent stepping during UTT	WISCI-II at T1	WISCI-II at T2	WISCI-II at T3
1	16	1	6	9
2	23	0	9	9
3	35	0	9	9
4	9	0	0	6
5	62	0	0	9

UTT = underwater treadmill training; T1 = prior to UTT; T2 = 6 months after the start of UTT; T3 = after completion of UTT; WISCI-II = Walking Index for Spinal Cord.

Measures of overground walking capacity are displayed in Table 3. Results from the Friedman test indicated a significant time effect (χ² = 8.000, P = .012) for WISCI-II. Planned follow-up analyses revealed no difference in WISCI-II levels (mean ± SD) obtained at T1 (0.20 ± 0.45) and T2 (4.80 ± 4.55) (Z = −1.633; P = .102) and at T2 (4.80 ± 4.55) and T3 (8.40 ± 1.34) (Z = −1.604; P = .109). However, the WISCI-II level documented at T3 (8.40 ± 1.34) was significantly higher compared to the T1 value (0.20 ± 0.45) (Z = −2.060; P = .039). Inspection of individual data revealed that all study participants registered a marked improvement in WISCI-II level over the 12-month training program (mean change = 8.2; range = 6–9) that corresponded with a large effect size (r = .65). Prior to UTT, four participants exhibited a WISCI-II level of 0 (unable to stand and/or participate in assisted walking) and one participant (a former gymnast who was unusually strong in his upper body and could use his trunk and momentum to swing one of his legs to initiate a step) displayed a WISCI-II level of 1 (able to ambulate in parallel bars, with braces and the physical assistance of two persons, for less than 10 meters). Following six months of UTT, two participants exhibited a WISCI-II level of 9 (ambulates with walker, with braces and no physical assistance, 10 meters), one participant displayed a WISCI-II level of 6 (ambulates with walker, with braces and physical assistance of one person, 10 meters), and two participants exhibited a WISCI-II level of 0. After completing 12 months of UTT, four participants displayed a WISCI-II level of 9 and one participant exhibited a WISCI-II level of 6.

The percentage of time spent in each phase of overground walk training is shown in Figure 3. As shown in this figure, all five participants progressed from the first to the fourth level of OWT and all but one progressed to the fifth level of OWT. Mean relative time engaged in OWT increased in a linear fashion across the first three stages of training (Level 1: 12%; Level 2: 20%; Level 3: 28%) and decreased in a similar manner over the remaining two training phases (Level 4: 24%; Level 5: 16%).

Figure 3.

Individual and group changes in the percentage of time spent in each phase of overground walk training.

Discussion

The enhancement in overground walking capacity we documented following water- and supplemental land-based gait training is consistent with recovery of intentional overground walking in two patients with chronic motor-complete cervical and thoracic SCI who completed 278 sessions (over 85 weeks) and 81 sessions (over 15 weeks) of programmed eSCS and weight-supported, land-based locomotor treadmill training.¹⁹ In our sample of participants with upper and lower thoracic chronic motor-complete SCI who had not undergone eSCS, independent stepping during UTT occurred in markedly fewer exercise sessions, a finding consistent with the use of an aquatic environment that provided sufficient levels of body weight support to initiate and maintain unassisted stepping movements.

In a report evaluating the reproducibility and validity of the WISCI-II in chronic spinal cord injury,²⁶ it was concluded that a change of one WISCI–II level in a person with chronic SCI can be interpreted as a real difference when quantifying activity- or exercise-related improvement in walking capacity. Hence, the extent to which overground walking capacity increased after UTT and supplemental OWT (mean change of eight levels) has positive clinical implications. The timing of improvement in walking capacity following long-term UTT and additional OWT varied among participants and likely reflected individualized responsiveness to prior training episodes; interaction among age, level of injury, and length of time post-injury; and ongoing medical management of secondary health conditions.

Although speculative, the improvement we observed in WISCI-II among persons with chronic cSCI who were classified as AIS A and did not receive eSCS may be linked to a number of potential mechanisms. Long-term exposure to underwater treadmill training may have led to the emergence of compensatory neuromuscular solutions that enabled locomotor recovery to occur with reduced impairment.^27–29 Another possible explanation for our findings is that an extended period of consistent exposure to rhythmic sensory input provided by treadmill walking in water, a supportive gait modality that alters sensory input and motor output, was sufficient to produce the training volume necessary to modulate spinal neural circuits and facilitate improved locomotor performance in the absence of supraspinal influence³⁰. The impact of activity-dependent plasticity in restoring sensorimotor function in individuals with chronic cSCI has also been highlighted in recent work showing enhancement of voluntary influence during stepping movements following application of transcutaneous eSCS and passive oscillation of the lower limbs in persons with motor complete paralysis.^31,32 In these experiments, the similarity in voluntary rhythmic stepping activity with and without transcutaneous eSCS in an unloaded condition provided evidence of reactivation of functional connectivity among neural networks in the brain and spinal cord.^31,32 Additionally, recent data have shown that somatosensory conduction was preserved in a subgroup of persons with clinically complete SCI following activation of the somatosensory cortex subsequent to somatosensory stimulation of insensate body regions below the level of injury.¹⁸ When this collective set of findings^18,31,32 is viewed in conjunction with results obtained in the present study and evidence showing that the adult human central nervous system is capable of plasticity in response to locomotor training,^16,17,33 a third possibility is that prolonged exposure to an enriched exercise setting featuring individually-tailored and progressive UTT and increasingly-challenging OWT tasks may have generated ample multisegmental proprioceptive input from lower-extremity load and joint receptors^32,34 to drive reorganization and entrainment of neuronal networks in the brain and spinal cord and produce volitional stepping on water and land in persons with chronic cSCI exhibiting some degree of anatomically preserved spinal motor connections across the level of injury.³⁵ Within the context of our multimodal training paradigm, the emergence and practice of independent stepping during UTT may have also served as a beneficial prerequisite for and accompaniment to OWT that incorporated higher physical and metabolic demands, increased weight-support requirements, and greater movement task complexity. Interestingly, once initiation of unassisted stepping activity was recorded during UTT, the transition to taking clusters of consecutive steps in subsequent training sessions occurred relatively quickly, implying an accelerated pace and persistence of sensory-motor learning.³¹

From a clinical perspective, improvement in WISCI-II levels following UTT and supplemental OWT is consistent with the progression in assisted and unassisted overground walking observed during supplemental OWT (Figure 3) and the realization of self-selected goals, gains in functional status, and an enhanced ability to perform daily living tasks. Examples of personal goals that were achieved by study participants included (1) standing and walking during worship services, (2) walking across grass to stand at a fence to watch a baseball game, (3) walking into and out of a bathroom in a community building, (4) walking while completing household tasks and standing to retrieve an object from a shelf, (5) walking up and down ADA ramps, (6) stepping up onto a standard curb, (7) releasing one hand from a walker to open a car door, (8) walking at a pace that allowed for social interaction with people, (9) working in a garden, (10) retiling a kitchen backsplash, and (11) having the confidence to walk, if needed, in any situation. Interestingly, the mobility goals of most participants at the beginning of the study were modest (e.g. take a few steps), with some indicating no or limited functional goals or a belief that overground walking was likely not possible. However, once independent walking was achieved during UTT and progression to OWT occurred, participants updated their mobility and activity goals on an ongoing basis, reflecting gains in walking status and a greater likelihood of being able to perform purposeful physical activities. With respect to performance-based outcomes, two study participants (who were athletes prior to injury) exceeded the maximal prescribed speed standard during UTT (1.16 mph) [with one walking at 1.6 mph (for all three 15-min walks) following eight months of training and another walking at 1.78 mph (for one 15-min walk) after six months of training] and a third participant walked a horizontal distance of 400 feet overground in approximately 30 min, while requiring only one 5-min rest period.

Limitations

In considering the enhancement in walking capacity observed in our sample of adults with cSCI, it is important to acknowledge the presence of limitations in our study. While each participant’s personal physician supplied medical documentation confirming the existence of SCI and AIS status, access to primary-source evaluations would have reduced bias resulting from potential classification errors. It should also be noted that the absence of a control group and the small sample size of our patient group limit both the extent to which gains in walking capacity can be solely attributable to UTT and supplemental OWT and the ability to discern trends linking time-related changes in WISCI-II levels to the onset of independent stepping in the water and performance achievements measured during water- and land-based training. Because few published studies exist chronicling the impact of locomotor training in persons with cSCI, the focus of our investigation was to quantify the effects of combined UTT and OWT on WISCI-II, a valid and reliable research measure of walking capacity in persons with chronic SCI^21,23,24. However, given the magnitude of change in WISCI-II that occurred, assessment of kinematic, kinetic, electromyographic, and strength-based variables related to walking function should be included in future studies to provide a more comprehensive picture of functional changes in mobility and better understand stepping strategies and solutions leading to the recovery of walking in persons with cSCI in response to the progressive demands of a multimodal gait rehabilitation program.

Conclusion

The significant and clinically meaningful increase in overground walking capacity we observed following an extended period of UTT and supplemental OWT lends credence to the notion that locomotor function can be enhanced in persons with cSCI who have not undergone eSCS. Clinically, our findings have potential therapeutic applications in limiting maladaptive plasticity and relearning and retaining locomotor skills following severe neurological trauma. The use of smaller, portable underwater treadmills may also extend the accessibility of underwater treadmill training, a technologically-advanced therapy option, beyond the research environment and into public fitness settings to enhance ambulatory status and physical function in persons with SCI and other severe neuromuscular disorders.

Disclaimer statements

Contributions Both authors participated in the design of the study and SS conducted the study. DM wrote the manuscript and both authors reviewed the final manuscript for submission.

Funding The authors received no funding to conduct the research described in this manuscript.

Conflicts of interest The authors declare that they have no conflicts of interest to disclose in this manuscript.