Effect of robotic-assisted gait training program on spatiotemporal gait parameters for ambulatory children with cerebral palsy: A randomized control trial

Abstract

BACKGROUND:

Gait training programs are commonly used to improve gait in children with cerebral palsy (CP).

OBJECTIVE:

To compared the effects of robotic-gait assistant training (RAGT) and conventional body weight support treadmill training (CBWSTT) on gait parameters among ambulatory children with CP.

METHODS:

The study is a randomized controlled trial of 36 children (17 in the RAGT group and 19 in the CBWSTT group) aged 5 to 14. Gait training involved 30—to 35-minute sessions three times per week over eight weeks.

RESULTS:

Mixed ANCOVA showed no main effect of time or group on all gait parameters (P > .05). Gross motor function measure dimensions D (GMFM D) and E (GMFM E) show main effects on step width. Stride length, step length, speed, swing phase, and double support phase interacted with GMFM D and E. There was a negative correlation between motor function level and the change from baseline. Children with lower motor function show a greater change from baseline.

CONCLUSION:

There were no significant differences between CBWSTT and RAGT for children with CP; however, with gait training interventions, the level of motor function should be considered.

Keywords

Robotic gait child cerebral palsy treadmill rehabilitation

1 Introduction

Cerebral palsy (CP) is the most common cause of childhood physical disability. It involves heterogeneous motor impairment caused by non-progressive damage to the developing brain. Children with CP can be classified according to their anatomical presentation into unilateral and bilateral CP, while the category classification includes spasticity, dyskinesia, or ataxic CP. CP leads to motor dysfunction, particularly in posture and movement (Patel, Neelakantan, Pandher, & Merrick, 2020).

One of the primary characteristics of cerebral palsy is abnormal gait patterns. A gait parameter is a specific measurement or characteristic that describes a gait pattern. Based on quantifiable variables, gait parameters, such as stride length, step length, step width, cadence, and stance and swing time, provide objective information about a person’s walking pattern. Healthcare professionals use these parameters to evaluate the effectiveness of treatments and interventions for individuals with cerebral palsy and monitor their progress over time (Gómez-Pérez, Font-Llagunes, Martori, & Vidal Samsó, 2019; Kharb, Saini, Jain, & Dhiman, 2011). The disabilities combined with CP produce abnormalities such as decreased gait velocity and stride length, step width, cadence, and other gait parameters (Corsi et al., 2021).

It is essential that physical therapy (PT) programs for children with CP include walking training. The focus of therapy for children with CP has shifted over the past decade from impairment-focused, such as muscle strengthening and range of motion exercise, to activity-based, with motor learning principles incorporated, such as repetitive and task-specific therapy. Repetitive practice of tasks may facilitate the integration of altered sensory and motor systems (Sananta, Mulia, Siahaan, & Huwae, 2022; Toovey, Bernie, Harvey, McGinley, & Spittle, 2017).

One of the approaches to task-specific training to improve walking is treadmill training (Han & Yun, 2020). Treadmill training maximizes the use of the lower limbs through a large number of steps and, consequently, a greater amount of load-bearing and activation of the weak muscles, particularly at faster speeds. Treadmills with a bodyweight support system may enable low-functioning individuals who cannot be safely supervised using traditional therapy methods to undertake early walking practice. During bodyweight support, walking can be practiced repetitively under controlled conditions (Han & Yun, 2020).

Conventional Body Weigh Supported Treadmill Training (CBWSTT) involves using a suspension system with a harness over a treadmill for gait training. The physical therapist assists with the gait dynamic by supporting the limb progression, heel strike, and knee control and providing tactile cues during swing phases (Han & Yun, 2020). It is cost-effective, easy to apply, and available in most clinical settings. However, this training methods need more than one physical therapist, exposes the therapist to more physical strain, and there is limited feedback for the patient.

To overcome the disadvantage of CBWSTT, treadmill-based Robotic-Assisted Gait Training (RAGT) was developed. Similar to the CBWSTT, the body weight is supported by a harness system over the treadmill. For this training method, a robotic-controlled exoskeleton is used to achieve accuracy as it is set considering the individual patient anthropometric and provides the reputation of movement with control of a range of motion and gait parameters. It required less effort from the physical therapist (Cherni & Ziane, 2022; Lefmann, Russo, & Hillier, 2017).

The CBWSTT and RAGT are effective gait training methods for pediatric gait disorders, including CP gait deviation (Aras et al., 2019; Cherni & Ziane, 2022; Drużbicki et al., 2013; Hilderley, Fehlings, Lee, & Wright, 2016; Lefmann et al., 2017). Both support the body weight and help the lower limbs and trunk to maintain proper alignment and patterns when initiating and performing the gait (Lefmann et al., 2017). Comparing the effect of using these two devices to maintain or improve gait parameters among children with CP is important to help physiotherapists determine which device is more convenient for rehabilitation clinics. The Lokomat^® Pro (Hocoma AG, Volketswil, Switzerland) is one of the robotic assistive devices used to train gait in children with CP (Cherni & Ziane, 2022).

The recent narrative review highlighted some of the limitations in the literature: there was a wide diversion of training procedures using the Lokomat for children with CP, training length per session ranged between 20 to 60 minutes, the reputation per week was between one to five repetition and the duration of treatment range between one to 12 weeks (Cherni & Ziane, 2022). Moreover, there are limited studies that used a control conventional intervention to assess its benefit on gait parameters (Aras et al., 2019; Drużbicki et al., 2013; Wallard, Dietrich, Kerlirzin, & Bredin, 2017).

In previous studies, children were categorized or compared based on Gross Motor Function Classifications (GMFCS) (McDowell, 2008); however, by utilizing more accurate outcome measures, one may be able to correlate changes in scores from baseline, regardless of whether significant improvements occurred or not, with impairment level. Additionally, it will indicate whether the level of impairment should be considered when selecting intervention methods for children regardless of their motor function classification. With GMFM-88, children with cerebral palsy can easily and objectively assess their motor function with a standardized scoring system in different dimensions. GMFM-88 provides a standardized measure of motor function that can be compared across a wide range of individuals, unlike GMFCS, which relies on clinical judgment.

Thus, the purpose of this study was; 1) to compare the effects of RAGT and CBWSTT on spatiotemporal gait parameters in ambulatory children with bilateral spastic CP, 2) to assess the effects of functional impairment on gait training interventions, and 3) to determine whether there is a correlation between changes in gait parameters over time and the level of impairment.

2 Methods

2.1 Study design and sitting

In this study, RAGT, using Lokomat^® Pro (Hocoma AG, Volketswil, Switzerland), was compared with CBWSTT with a ratio of (1 : 1) in a parallel prospective, double-blinded, randomized controlled trial. The study was conducted between January and April 2021. Participants were recruited from King Abdullah Children’s Hospital’s rehabilitation department in King Abdul-Aziz Medical City in Riyadh, Saudi Arabia.

2.2 Participants

This study included ambulatory children with bilateral diplegic spastic CP aged 5 to 14 who could follow instructions and accurately indicate pain or discomfort. The study included children with a passive range of motion within the minimum range required for Lokomat (hip and knee flexion contracture≤10° and knee valgus≤40°). Children were excluded if they had Botulinum toxin-A (BTX-A) injections to the lower limb within six months, orthopedic surgical intervention within nine months for muscles or 12 months for bons, used implanted infusion pump for Baclofen therapy, uncontrolled epilepsy, significant vision or hearing impairments, skin inflammation or open lesion around the limbs or trunk, hip joint instability, and any other contraindications to using Lokomat pro (Hocoma, 2022).

2.3 Study intervention

All participants received comprehensive physical therapy home program included the following: (1) Aerobic exercises involving throwing and catching a ball as well as step exercises on a stepping stool with forward, backward, and side-to-side movement; (2) Strengthening exercises using a resistive band from different positions (lying, sitting, or standing). The exercises involve placing a band over the foot while lying down and straightening the leg, abducting the leg while standing and holding on a chair, standing and flexing the knees while holding on a chair, sitting on a chair and extending the knee, in a standing or sitting position hold the band and abduct the arms, grasping a fixed band against the wall and flexing the shoulder, grasping a fixed band and extending the elbow, and Fix the band with the foot and flex the elbow while standing. Each exercise completes 3 sets of 10–15 reputations and; (3) Stretching exercises to be performed before aerobic exercise and strengthening exercises for the upper and lower limb muscles to prevent injury. These exercises involve Stretch and extend the knee from prone or side lying, extend the knee from supine position with hip flexed 90 degrees, extend the legs against a wall, flex the hips when lying down, hyperextend the hips when lying down, abduct the leg when lying down, reach with the hand the opposite shoulder, reach with the hand the back, extend the elbow and figure. Each exercise repeated 3 times with hold for 10–30 seconds. All participants’ parents or caregivers were given a checklist sheet to provide information about each exercise.

The gait training protocol was based on the previous studies that scored six or above on PEDro score (Armijo-Olivo et al., 2015). The gait training was 30–35 minutes (gait training began with 10 to 20 minutes in the first session and gradually increased to 30 to 35 minutes in subsequent sessions). Every 10 to 15 minutes, each child was given a rest period of 2 to 3 minutes, the frequency was three sessions per week, and the total duration of the intervention was eight weeks (Cherni & Ziane, 2022; Hilderley et al., 2016). After the equipment was set up and during the training, each participant was asked whether the apparatus was causing pain or discomfort; if so, the training was stopped, and adjustment was conducted; only then did the training continue. If any skin redness was visible at the end of the session, it was ensured that the next session would clear this up. In this study, no events were reported.

2.3.1 RAGT group

This group had gait training using the the Lokomat Pro, The Locomat is a robotic treadmill training system that uses a body weight support belt and driven gait orthoses (pediatric orthoses) for both legs. The RAGT is adjusted to provide the best possible fit for children. The child’s legs were fixed to each orthosis with three cuffs for each leg. Devices for hip and knee joints are also actuated. The robotic control uses an adjustable controller for hip and knee joints with an adjustable path. Elastic straps provide foot dorsiflexion support.

An initial fitting session was conducted prior to the intervention to determine the correct fit for straps and cuffs. This session also involved setting the alignment within a tolerable ROM and ensuring that each child was comfortable in the Lokomat Pro, able to follow instructions, and able to indicate any pain, dizziness, nausea, or discomfort. The therapist used virtual reality games during gait training to promote feedback and encourage the children. The RAGT allows therapists to control the walking speed, body weight support, and robotic leg assistance received by patients during training. RAGT uses an interactive display screen to monitor the patient’s movements during training. Additionally, the program has a balloon-kicking game that encourages patients and helps them address problems related to their step length. Moreover, the program asks patients to collect coins to improve their walking speed

2.3.2 CBWSTT group

The Biodex Unweighting System (Biodex Medical Systems, Inc., Shirley, NY, USA) was used to support the child’s body weight by an adjustable harness placed over the treadmill. Two therapists assisted the child’s gait; each guided one leg. This harness grips around the pelvis and lower trunk to hold the child’s weight during training. The harness is suspended by straps and clips attached to a bar over the child’s head. The amount of body weight support during the training varied from one child to another according to their impairment level. The treadmill started at a speed of 0.4 and gradually increased in the first 3 to 5 minutes until reaching the appropriate speed according to each child’s tolerance.

2.3.3 Outcome measures

Gait parameters were collected using the Zebris™ FDM System (Zebris Medical GmbH, Isny, Germany). This system used force sensors arranged in a 200×60 cm platform. The measuring plate is integrated into a level walking gait cycle that can be repeated up to four times. The gait parameters were automatically calculated in the software program installed in a computer connected directly to the system.

The gait parameters include Step Width (cm), Step Length (cm), Stride length (cm), Cadence (steps/min), Velocity (m/s), Swing phase (sec), Stance phase, Stride time (sec), Step time (sec), Foot’s rotation (degree), Total double support (%).

The participants were asked to walk at a speed at which they felt safe and comfortable on the platform. They started walking 2 meters before and after the end of the platform for acceleration and deceleration. Each child performed this test with one trial before the actual measurement. Children used their assistive devices and walking aids if needed.

2.4 Sample size

The sample size was estimated using the Sample Size Calculator for designing clinical research (https://sample-size.net/). A minimum of 32 participants was recommended to find a medium effect size, with a significance of.05 (one-tail) and a power level of.80. Considering a possible dropout rate of around 30%, the total sample size was raised to 44 children (20 per group).

2.5 Procedures

Demographic data were collected from patient medical records in the physiotherapy clinic at King Abdullah Specialist Children’s Hospital. Demographic data (age, gender, height, and weight) were collected from participant at baseline assessment session. In the first session, GMFM D (standing) and E (walking, running, and jumping) were assessed. Gait parameters were measured twice, pre and post the interventions.

2.5.1 Sampling method, randomization, and blindness

A convenient sample of 52 children were recruited. Only 40 children met the inclusion criteria. After the participants’ baseline assessments were completed, a block-stratified randomization method by age group (two levels: 5–9 years and 10–14 years) was used to balance the study’s two equal groups with an allocation ratio of 1 : 1 (20 participants each). Then, a simple randomized method (flipping a coin) was conducted between the two groups to determine the RAGT and CBWSTT groups. Outcome measurements were assessed by a physiotherapist blinded to the group allocations. Children and their families in each group were blinded to the intervention of the other group.

2.6 Statistical analysis

Data analyses were conducted using IBM SPSS Statistics for Windows, version 29 (IBM Corp., Armonk, NY, USA). For continuous variables, the Shapiro-Wilk test was used to determine normality. Statistical analysis was presented as mean and standard deviation if the data were normally distributed or as median and (1st – 3rd quartiles) if the data were non-normally distributed. Categorical data were presented using frequency and percentage. In order to compare the two groups at the baseline level, an independent sample t-test for normally distributed data, the Mann-Whitney test for non-normally distributed data, and the Chi-square test for categorical data were used.

Mixed analysis of covariance (ANCOVA) with two factors was used to analyze the effect of Gait training groups (RAGT group and CBWSTT group) and time (pre- and post-intervention) on the gait parameters, spatial gait parameter, stride length, stride velocity, step length; step width; and foot rotation, and temporal gait parameters; Cadence, Speed (Velocity), percentage of stance phase, percentage of swing phase and percentage of Total double support phase. To control the effect of age and height on gait parameters (Bell, Õunpuu, DeLuca, & Romness, 2002; Gieysztor, Kowal, & Paprocka-Borowicz, 2021), both were set as a covariance. Moreover, the level of motor impairment (measured by GMFM D and E) was added as an additional covariate to control the level of impairment at baseline (Malt, Aarli, Bogen, & Fevang, 2016). Levene’s test was used to indicate the homogeneity of variances between the groups (P > 0.05).

In order to control for the effect of patients’ initial status of the gait parameter on intervention outcomes and to account for the potential variability in their baseline values, Change-from-Baseline (the change between the baseline and after-intervention score of a gait parameter) was calculated. Then, Pearson’s (r) or Spearman’s (rho) correlation coefficient was used to examine the relationship between the change of gait parameter with baseline scores and with motor impairment level (GMFM D and E) for each group. Correlation coefficients were interpreted as follows: <.1; weak,.1–.3; moderate,.4–.6; and strong,.7–.8 (Akoglu, 2018).

Additional t-tests were conducted to assess the effect size if significant differences were present. The effect size was interpreted as <0.1 = trivial effect, 0.1 < 0.3 = small effect, 0.3 < 0.5 = moderate effect and≥0.5 = large difference effect (Cohen, 1988). All results were considered statistically significant at a p-value of < 0.05.

3 Results

Shapiro-Wilk test results showed that all continuous variables’ data of participant characteristics were normally distributed (P > .05), except for age, BMI, and GMFM D and E (P < 0.05). All gait parameters were normally distributed. There is a normal distribution for all changes from baseline scores, except for the speed and percentage of stance phase for CBWSTT and the foot rotation angle for RAGT groups.

Two children (one from each group) discontinued physical therapy sessions due to personal reasons. The final data analysis included 36 participants: 19 in the CBWSTT group and 17 in the RAGT group (Fig. 1).

Fig. 1

Participant flow diagram.

3.1 Participant characteristics

Table 1 presents the characteristics of the 36 participants. Their ages ranged from 5 to 14, and 21 were boys.

Table 1
Participants’ characteristics and differences at baseline

Characteristics CBWSTT group N = 17 RAGT group N = 19 Comparison at baseline

Age (years) 7.66 (6.00–13.13) 8.17(7.08–10.66) U = 151, P = .75

Age group 5–8 years 9 (52.9%) 13 (68.4%) X² = .34, df = 1, P = .34

9–14 years 8 (47.1%) 6 (31.6%)

Gender: Boys 10 (58.8%) 11(57.9%) X² = .003, df = 1, P = .96

Girls 7 (41.2%) 8 (42.1%)

Height (centimeter) 118.85±21.2 118.73±15.47 t = .02, df = 34, p = .99

Weight (kilogram) 23.97±10.74 23.88±7.67 t = .03, df = 34, p = .97

BMI score 15.68 (14.52–17.69) 15.45(14.15–18.34) U = 151, p = .75

GMFM D 66.66 (16.66–88.46) 69.23(61.53–76.92) U = 154, p = .82

GMFM E 72.22 (27.08–81.94) 65.27 (58.33–83.33) U = 160, p = .98

Characteristics	CBWSTT group N = 17	RAGT group N = 19	Comparison at baseline
Age (years)	7.66 (6.00–13.13)	8.17(7.08–10.66)	U = 151, P = .75
Age group 5–8 years	9 (52.9%)	13 (68.4%)	X² = .34, df = 1, P = .34
9–14 years	8 (47.1%)	6 (31.6%)
Gender: Boys	10 (58.8%)	11(57.9%)	X² = .003, df = 1, P = .96
Girls	7 (41.2%)	8 (42.1%)
Height (centimeter)	118.85±21.2	118.73±15.47	t = .02, df = 34, p = .99
Weight (kilogram)	23.97±10.74	23.88±7.67	t = .03, df = 34, p = .97
BMI score	15.68 (14.52–17.69)	15.45(14.15–18.34)	U = 151, p = .75
GMFM D	66.66 (16.66–88.46)	69.23(61.53–76.92)	U = 154, p = .82
GMFM E	72.22 (27.08–81.94)	65.27 (58.33–83.33)	U = 160, p = .98

For weight and height, data are presented as a mean±standard deviation and frequency and percentage for age group and gender. Otherwise, data is presented as median (1^st–3rd quartiles). N = number of participants, % =percent, U = Mann-Whitney, X² = chi-square, t = independent simple t-test, GMFM = Gross Motor Function measures. Significance P < .05.

3.1.1 Gait parameters at baseline

No significant difference was observed between the two gait training groups at baseline for most of the gait parameters examined, except for stride velocity (Table 2). As a result of this finding, there was no discernible difference in gait performance between the two groups at the beginning of the study.

Table 2
Gait parameter at baseline

Gait parameters CBWSTT group N = 17 RAGT group N = 19 t df Sig

Stride length 78.06±20.07 76.74±19.818 .199 34 .844

Stride velocity 1.42±0.80 1.02±0.24 1.970 34 .032

Step length 37.41±11.62 37.15±13.53 .060 34 .952

Step width 10.12±5.19 11.68±4.16 –1.003 34 .323

Foot rotation –.28±11.58 2.40±10.85 –.715 34 .48

Cadence 103.35±38.93 124.68±27.32 –1.919 34 .063

Speed .67±0.32 .78±0.23 –1.212 34 .234

Percentage of stance phase 74.00±21.57 70.12±14.01 .684 34 .498

Percentage of swing phase 30.34±7.42 31.02±7.35 –.278 34 .783

Double support phase 37.71±13.24 36.60±11.15 .273 34 .786

Gait parameters	CBWSTT group N = 17	RAGT group N = 19	t	df	Sig
Stride length	78.06±20.07	76.74±19.818	.199	34	.844
Stride velocity	1.42±0.80	1.02±0.24	1.970	34	.032
Step length	37.41±11.62	37.15±13.53	.060	34	.952
Step width	10.12±5.19	11.68±4.16	–1.003	34	.323
Foot rotation	–.28±11.58	2.40±10.85	–.715	34	.48
Cadence	103.35±38.93	124.68±27.32	–1.919	34	.063
Speed	.67±0.32	.78±0.23	–1.212	34	.234
Percentage of stance phase	74.00±21.57	70.12±14.01	.684	34	.498
Percentage of swing phase	30.34±7.42	31.02±7.35	–.278	34	.783
Double support phase	37.71±13.24	36.60±11.15	.273	34	.786

Data are represented as mean±SD, t = independent simple t-test, df = degree of freedom, sig = Significance P < .05.

3.1.2 Effect of Gait training on children’s gait parameters

Mixed ANCOVA showed no main effect of time or group on all gait parameters (P > .05), indicating no significant differences after the intervention training for both groups, and there were no significant differences between the CBWSTT group and RAGT on gait parameters. moreover, no time×interaction observed (Table 2). Age and height show no effect on the outcome measured (P > .05). However, GMFM D and E show a main effect of step width (F = 7.57, p = .01, η ² = .20; F = 11.32, p = .002, η ² = .27, respectively). Additionally, there was the interaction of GMFM D and E and time for stride length, step length, speed, swing phase, and total double support phase (Table 3), which means that the level of impairment could be an important predictor of the outcomes for both interventions with small to moderate effects.

Table 3
Effect of time and intervention on gait parameters using Mixed ANCOVA

Source of variation Time (within subjects) Group (between subjects) Time x Group interaction

F Sig η ² F Sig η ² F Sig η ²

Stride length cm .055 .816 .002 .025 .875 .001 1.291 .265 .041

Stride velocity .406 .529 .013 2.490 .125 .077 4.037 .054 .119

Step length .302 .587 .010 .001 .980 .000 .134 .717 .004

Step width 1.311 .261 .042 .259 .615 .009 2.285 .141 .071

Cadence .418 .523 .014 1.282 .267 .041 2.043 .163 .064

Speed 1.093 .304 .035 .010 .922 .000 1.577 .219 .050

Foot rotation 2.06 .162 .064 .121 .731 .004 .956 .336 .031

Stance phase .001 .977 .000 .100 .754 .003 3.620 .067 .108

Swing phase 3.569 .069 .106 .010 .923 .000 .484 .492 .016

Double support phase 3.322 .078 .100 .281 .600 .009 1.246 .273 .040

Covariate Time * GMFM D score interaction Time * GMFM E score interaction

F Sig η ² F Sig η ²

Stride length 5.94 .021 .165 8.47 .01 .22

Stride velocity .03 .869 .001 1.75 .19 .06

Step length 11.67 .002 .280 14.22 <.001 .32

Step width 2.10 .158 .065 1.29 .26 .04

Cadence .21 .650 .007 .99 .33 .03

Speed (m/s) 6.02 .020 .167 5.55 .03 .16

Foot rotation .03 .863 .001 .003 .96 <.001

Stance phase (%) .01 .935 .000 .24 .63 .01

Swing phase (%) 4.67 .039 .135 8.18 .01 .21

Double support phase (%) 7.19 .012 .193 4.82 .04 .14

Source of variation	Time (within subjects)	Group (between subjects)	Time x Group interaction
Stride length cm	.055	.816	.002	.025	.875	.001	1.291	.265	.041
Stride velocity	.406	.529	.013	2.490	.125	.077	4.037	.054	.119
Step length	.302	.587	.010	.001	.980	.000	.134	.717	.004
Step width	1.311	.261	.042	.259	.615	.009	2.285	.141	.071
Cadence	.418	.523	.014	1.282	.267	.041	2.043	.163	.064
Speed	1.093	.304	.035	.010	.922	.000	1.577	.219	.050
Foot rotation	2.06	.162	.064	.121	.731	.004	.956	.336	.031
Stance phase	.001	.977	.000	.100	.754	.003	3.620	.067	.108
Swing phase	3.569	.069	.106	.010	.923	.000	.484	.492	.016
Double support phase	3.322	.078	.100	.281	.600	.009	1.246	.273	.040
Covariate	Time * GMFM D score interaction	Time * GMFM E score interaction
	F	Sig	η ²	F	Sig	η ²
Stride length	5.94	.021	.165	8.47	.01	.22
Stride velocity	.03	.869	.001	1.75	.19	.06
Step length	11.67	.002	.280	14.22	<.001	.32
Step width	2.10	.158	.065	1.29	.26	.04
Cadence	.21	.650	.007	.99	.33	.03
Speed (m/s)	6.02	.020	.167	5.55	.03	.16
Foot rotation	.03	.863	.001	.003	.96	<.001
Stance phase (%)	.01	.935	.000	.24	.63	.01
Swing phase (%)	4.67	.039	.135	8.18	.01	.21
Double support phase (%)	7.19	.012	.193	4.82	.04	.14

GMFM=Gross motor function measure, D = standing dimension, E = walking, running and jumping dimension, F = magnitude of effect, ηp² = partial eta square (effect size). Sig significance level p < .05.

Table 4

Relationship of baseline scores with the change from baseline for RAGT and CWBST groups

	RAGT group N = 17				CBWST group N = 19
	Correlation coefficient	sig	95% Confidence intervals		Correlation coefficient	sig	95% Confidence intervals
Stride length	–.62	.005	–.84	–.224	–.24	.35	–.65	.27
Stride velocity	–.55	.015	–.80	–.125	–.91	<.001	–.97	–.75
Step length	–.66	.002	–.85	–.29	–.21	.41	–.63	.29
Step width	–.27	.27	–.65	.21	–.69	.002	–.88	–.31
Cadence	–.77	<.001	–.91	–.49	–.572	.016	–.83	–.13
Speed	–.66	.002	–.86	–.28	–.32	.21	–.70	.20
Foot rotation	–.23	.345	–.63	.27	–.584	.014	–.832	–.144
Stance phase	–.27	.27	–.64	.22	–.68	.003	–.88	–.30
Swing phase	–.67	.002	–.86	–.32	–.580	.015	–.830	–.138
Double support phase	–.61	.006	–.83	–.21	–.693	.002	–.880	–.318

Correlations were presented by Pearson’s correlation coefficients, except for the speed and percentage of stance phase for CBWST and the foot rotation angle for RAGT groups, which were presented by Spearman correlation confidant. Sig significance level p < .05.

3.1.3 Relationship of baseline scores with the change from baseline for each group

All changes from baseline scores have a normal distribution, except for the speed and percentage of stance phase for CBWSTT and the foot rotation angle for RAGT groups.

Children in group RAGT showed a significant negative moderate to strong correlation between the baseline impairment level and gate parameters (except for step with and stance phase). On the other hand, Children in the CBWSTT group showed a significant negative moderate to very strong correlation with gait parameters (except stride length, step length, and speed). Children with lower motor functional levels generally show a greater change from baseline.

4 Discussion

The purpose of the current study was to compare the effect of RAGT and BWSTT on the spatio-temporal gait parameters and walking abilities in children with CP (5–14 years of age). The results revealed no significant differences among the two groups in all post-intervention measures.

This study confirmed previous findings. In a previous study by Druzbicki et al. (2013), 52 children with CP were equally divided into two groups: a RAGT group and a control group. All children received 20 sessions. This study noted non-significant improvement in participants’ walking speed. Another study conducted by Aras et al. (2019) in Turkey compared three interventions: RAGT, BWSTT, and anti-gravity treadmill exercises (ATE). In that study, 29 children with CP were divided into three groups: RAGT (n = 10), BWSTT (n = 10), and ATE (n = 9). All participants underwent 20 sessions (45 min duration) five days per week. The results indicated a non-statistically significant change in gait parameters among the groups after intervention (Aras et al., 2019).

These results were consistent with a study by Dobkin and Duncan (2012), which indicated that RAGT and BWSTT did not improve outcomes for children with CP. According to Dobkin and Duncan, —the conceptual bases for these promising rehabilitation interventions had once seemed quite plausible, but the results of well-designed, randomized clinical trials have been disappointing (Dobkin & Duncan, 2012). Thus, the results of the present study support this conclusion. Additionally, RAGT has not been proven superior to BWSTT for improving spatio-temporal gait parameters for children with CP.

Moreover, different devices used by researchers may yield different results. For example, a wearable exoskeleton device study showed improvements in different spatiotemporal gait parameters for children with CP (Lerner, Damiano, & Bulea, 2017). A study conducted by Cherni showed significant improvements in gait parameters that run contrary to the results of the present study (Cherni, Ballaz, Lemaire, Dal Maso, & Begon, 2020). This may be due to the age of the target sample since the present study targeted children aged 5 to 14 years, whereas the study by Cherni targeted individuals aged 7 to 20 years with a bilateral CP. Also, the results of the current study were inconsistent with those of a study conducted by Sarhan, Chevidikunnan, and Gaowgzeh (2014). A study aimed to compare RAGT and manual treadmill programs among children (n = 12) with CP. The results showed a significant improvement in gait parameters by using RAGT among children with cerebral palsy under five years old (Sarhan, Chevidikunnan, & Gaowgzeh, 2014).

Several limitations exist in the present study. First, it was conducted in only one center, so the results cannot be generalized. Second, the absence of follow-up for children may limit the application of the results to the short-term effects of RAGT and CBWSTT only. Considering the possibility of bias in these responses is essential.

Prior studies have focused on measuring the post-intervention effects with less focus on the magnitude of change; this study’s findings provide valuable insights into the effects of functional impairment level on improving gait parameters in children with CP. The results demonstrate a negative correlation between functional impairment level and the magnitude of change in some gait parameters for both groups. This suggests that children with more severe functional impairments may benefit more from gait training interventions than those with a lower functional impairment level. Similarly, Beretta et al. (2020) demonstrated that improvements in gait were only significant for low motor function levels (GMFCS III) in children with CP after RAGT (Beretta et al., 2020).

5 Conclusion

This study found that RAGT using the Lokomat and CBWST had similar effects on spatiotemporal gait parameters among ambulatory children with bilateral cerebral palsy. However, the GMFM D and E dimensions were found to be significant moderators of gait changes, with children with lower levels of motor function showing greater improvements. This study highlights the importance of considering motor function levels when implementing gait training programs for children with cerebral palsy. Overall, more investigation is needed for both RAGT and CBWSWTT on the effectiveness of gait outcomes in this population; further research is needed to establish their comparative effectiveness.

Footnotes

Acknowledgments

The study was supported by the Researchers Supporting Project, King Saud University, Riyadh, Saudi Arabia (number RSPD2024R659).

Conflict of interest

The authors report no competing interests.

Ethics statement

In accord with the Declaration of Helsinki of 1964, ethics approval was attained from the Institutional Review Board (IRB) of the College of Medicine at King Saud University (No. 20/0233) on the 12^th of April, 2020, and from King Abdullah International Medical Research Centre (KAIMRC) at King Abdul-Aziz Medical City (No. RC1913631 R) on the first of March, 2020. This study was retrospectively registered at clinicaltrials.gov (identifier: NCT06368180).

Informed consent

The parents of all participants signed a formal consent form after receiving a full explanation of the study’s objectives and methods before the study began. All participants aged 5 to 14 give assents as appropriate to their age.

Funding

No funding was received.

Data availability

The data that support the findings of this study are available upon request from the corresponding author. The data are not publicly available due to their containing information that could compromise the privacy of the research participants.

References

Akoglu,

(2018). User’s guide to correlation coefficients. Turkish Journal of Emergency Medicine, 18(3), 91–93.

Aras,

, Yaşar,

, Kesikburun,

, Türker,

, Tok,

, & Yılmaz,

(2019). Comparison of the effectiveness of partial body weight-supported treadmill exercises, robotic-assisted treadmill exercises, and anti-gravity treadmill exercises in spastic cerebral palsy. Turkish Journal of Physical Medicine and Rehabilitation, 65(4), 361.

Armijo-Olivo,

, da Costa,

B. R.

, Cummings,

G. G.

, Ha,

, Fuentes,

, Saltaji,

, & Egger,

(2015). PEDro or Cochrane to assess the quality of clinical trials? A meta-epidemiological study. PloS one, 10(7), e0132634.

Bell,

K. J.

, Õunpuu,

, DeLuca,

P. A.

, & Romness,

M. J.

(2002). Natural progression of gait in children with cerebral palsy. Journal of Pediatric Orthopaedics, 22(5), 677–682.

Beretta,

, Storm,

F. A.

, Strazzer,

, Frascarelli,

, Petrarca,

, Colazza,

, ... Maghini,

(2020). Effect of robot-assisted gait training in a large population of children with motor impairment due to cerebral palsy or acquired brain injury. Archives of Physical Medicine and Rehabilitation, 101(1), 106–112.

Cherni,

, Ballaz,

, Lemaire,

, Dal Maso,

, & Begon,

(2020). Effect of low dose robotic-gait training on walking capacity in children and adolescents with cerebral palsy. Neurophysiologie Clinique, 50(6), 507–519.

Cherni,

, & Ziane,

(2022). A Narrative Review on Robotic-Assisted Gait Training in Children and Adolescents with Cerebral Palsy: Training Parameters, Choice of Settings, and Perspectives. Disabilities, 2(2), 293–303.

Cohen,

(1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, N.J.: L. Erlbaum Associates.

Corsi,

, Santos,

M. M.

, Moreira,

R. F.

, Dos Santos,

A. N.

, de Campos,

A. C.

, Galli,

, & Rocha,

N. A.

(2021). Effect of physical therapy interventions on spatiotemporal gait parameters in children with cerebral palsy: A systematic review. Disability and Rehabilitation, 43(11), 1507–1516.

10.

Dobkin,

B. H.

, & Duncan,

P. W.

(2012). Should body weight–supported treadmill training and robotic-assistive steppers for locomotor training trot back to the starting gate? Neurorehabilitation and Neural Repair, 26(4), 308–317.

11.

Drużbicki,

, Rusek,

, Snela,

, Dudek,

, Szczepanik,

, Zak,

, ... Sobota,

(2013). Functional effects of robotic-assisted locomotor treadmill thearapy in children with cerebral palsy. Journal of Rehabilitation Medicine, 45(4), 358–363.

12.

Gieysztor,

, Kowal,

, & Paprocka-Borowicz,

(2021). Gait parameters in healthy preschool and school children assessed using wireless inertial sensor. Sensors, 21(19), 6423.

13.

Gómez-Pérez,

, Font-Llagunes,

J. M.

, Martori,

J. C.

, & Vidal Samsó,

(2019). Gait parameters in children with bilateral spastic cerebral palsy: A systematic review of randomized controlled trials. Developmental Medicine & Child Neurology, 61(7), 770–782.

14.

Han,

Y.-G.

, & Yun,

C.-K.

(2020). Effectiveness of treadmill training on gait function in children with cerebral palsy: Meta-analysis. Journal of Exercise Rehabilitation, 16(1), 10.

15.

Hilderley,

A. J.

, Fehlings,

, Lee,

G. W.

, & Wright,

(2016). Comparison of a robotic-assisted gait training program with a program of functional gait training for children with cerebral palsy: Design and methods of a two group randomized controlled cross-over trial. Springerplus, 5(1), 1–14.

16.

Hocoma. (2022). Legal notes. Retrieved from https://www.hocoma.com/legal-notes/#lokomatpro_lokomatnanos

17.

Kharb,

, Saini,

, Jain,

, & Dhiman,

(2011). A review of gait cycle and its parameters. IJCEM International Journal of Computational Engineering & Management, 13, 78–83.

18.

Lefmann,

, Russo,

, & Hillier,

(2017). The effectiveness of robotic-assisted gait training for paediatric gait disorders: Systematic review. Journal of Neuroengineering and Rehabilitation, 14(1), 1–10.

19.

Lerner,

Z. F.

, Damiano,

D. L.

, & Bulea,

T. C.

(2017). A lower-extremity exoskeleton improves knee extension in children with crouch gait from cerebral palsy. eaam. Science Translational Medicine, 9(404), 9145.

20.

Malt,

M. A.

, Aarli,

Å.

, Bogen,

, & Fevang,

J. M.

(2016). Correlation between the Gait Deviation Index and gross motor function (GMFCS level) in children with cerebral palsy. Journal of Children’s Orthopaedics, 10(3), 261–266.

21.

McDowell,

(2008). The gross motor function classification system-Expanded and revised. Developmental Medicine and Child Neurology, 50(10), 725.

22.

Patel,

D. R.

, Neelakantan,

, Pandher,

, & Merrick,

(2020). Cerebral palsy in children: A clinical overview. Translational Pediatrics, 9(Suppl 1), S125.

23.

Sananta,

, Mulia,

E. R.

, Siahaan,

L. D.

, & Huwae,

T. E. C.

(2022). Robot-Assisted Gait Training for Children with Cerebral Palsy: A Literature Review. International Journal of Medical Reviews and Case Reports, 6(7), 9.

24.

Sarhan,

R. S.

, Chevidikunnan,

M. F.

, & Gaowgzeh,

R. A. M.

(2014). Locomotor treadmill training program using driven gait orthosis versus manual treadmill therapy on motor output in spastic diplegic cerebral palsy children. Journal of Health and Allied Sciences NU, 4(04), 010–017.

25.

Toovey,

, Bernie,

, Harvey,

A. R.

, McGinley,

J. L.

, & Spittle,

A. J.

(2017). Task-specific gross motor skills training for ambulant school-aged children with cerebral palsy: A systematic review. BMJ Paediatrics Open, 1(1), e000078.

26.

Wallard,

, Dietrich,

, Kerlirzin,

, & Bredin,

(2017). Robotic-assisted gait training improves walking abilities in diplegic children with cerebral palsy. European Journal of Paediatric Neurology, 21(3), 557–564.