Exp Brain Res (2014) 232:31–42 DOI 10.1007/s00221-013-3717-8


Effect of a robotic restraint gait training versus robotic conventional gait training on gait parameters in stroke patients Céline Bonnyaud · Raphael Zory · Julien Boudarham · Didier Pradon · Djamel Bensmail · Nicolas Roche 

Received: 8 January 2013 / Accepted: 22 September 2013 / Published online: 10 November 2013 © Springer-Verlag Berlin Heidelberg 2013

Abstract  Kinematic and kinetic gait parameters have never been assessed following robotic-assisted gait training in hemiparetic patients. Previous studies suggest that restraint of the non-paretic lower limb during gait training could be a useful rehabilitation approach for hemiparetic patients. The aim of this study is to compare a new Lokomat® asymmetrical restraint paradigm (with a negative kinematic constraint on the non-paretic limb and a positive kinematic constraint on the paretic limb) with a conventional symmetrical Lokomat® training in hemiparetic subjects. We hypothesized that hip and knee kinematics on paretic side would be more improved after the asymmetrical Lokomat® training than after the conventional training. In a prospective observational controlled study, 26 hemiparetic subjects were randomized to one of the two groups Lokomat® experimental gait training (LE) or Lokomat® conventional gait training (LC). They were assessed using 3D gait analysis before, immediately after the 20 min of gait training and following a 20-min rest period. There was a greater increase in peak knee flexion on the paretic side following LE than LC (p  = 0.04), and each type of training induced different changes in vertical GRF during single-support phase on the paretic side. Several other spatiotemporal, kinematic and kinetic gait parameters were similarly improved after both types of training. Lokomat® restrained gait training with a negative kinematic constraint on the non-paretic limb and a positive kinematic constraint on the paretic limb appears to be an effective approach to C. Bonnyaud (*) · R. Zory · J. Boudarham · D. Pradon · D. Bensmail · N. Roche  Groupe de Recherche Clinique et Technologique sur le Handicap, Laboratoire d’analyse du mouvement EA 4497, CIC‑IT 805, Hôpital Raymond Poincaré, Université Versailles Saint Quentin en Yvelines, 104 Bld Raymond Poincaré, 92380 Garches, France e-mail: [email protected]

specifically improve knee flexion in the paretic lower limb in hemiparetic patients. This study also highlights spatiotemporal, kinematic and kinetic improvements after Lokomat® training, in hemiparetic subjects, rarely investigated before. Keywords  Stroke · Lokomat® · Restraint · Asymmetry · Gait training · Biomechanical gait parameters

Introduction The majority of stroke patients with hemiparesis have residual gait impairments (Bohannon 1987; Olney et al. 1994; Von Schroeber et al. 1995). Hemiparetic gait following stroke is characterized by spatiotemporal impairment such as a decrease in cadence, stride length and speed (Bohannon 1987; Olney et al. 1994; Von Schroeber et al. 1995; Brandstater et al. 1983; Pinzur et al. 1987). It is also characterized by kinematic alterations (i.e. active range of motion around lower limbs joints) such as decreased peak hip flexion (Olney et al. 1994), peak knee flexion (Kerrigan et al. 1991; Robertson et al. 2009; Hutin et al. 2010) and peak ankle dorsiflexion (Pittock et al. 2003) on paretic side. These kinematic alterations are associated with changes in kinetic parameters (i.e. forces involved) such as a decrease in ground reaction forces (GRF) under the paretic limb (Morita et al. 1995). A rehabilitation programme involving gait training is the main treatment commonly used to improve gait kinematics and kinetics in these patients. In recent years, robotic-assisted gait training based on the concepts of repetitive, intensive and task-oriented training has been developed. Among these systems, the Lokomat® (which provides mechanical assistance to



movement and body weight support) aims to improve gait in patients with neurological pathologies. In stroke patients, the Lokomat® is commonly used to impose a symmetrical pattern in order to reduce the asymmetry of the spatiotemporal, kinematic and kinetic parameters described above. Studies evaluating the effects of gait training with the Lokomat® on stroke patients have found contradictory results. Indeed, Schwartz et al. (2009) showed that gait velocity and performance on the 6-min walk test were not improved after Lokomat® training in patients less than 3 months post-stroke (Schwartz et al. 2009). Similarly, Hidler et al. (2009) found that the use of the Lokomat® did not improve spatiotemporal gait parameters, while conventional gait rehabilitation did (Hidler et al. 2009). In contrast, Mayr et al. (2007) found a greater increase in performance on the 6-min walk test following Lokomat® training than following conventional rehabilitation (Mayr et al. 2007). Husemann et al. (2007) found that training on the Lokomat® increased the percentage of the gait cycle spent in single-support phase of the gait cycle on the hemiparetic lower limb (Husemann et al. 2007). This difference in these results might be explained by the different degrees of chronicity of the patients included, differences in the study designs and also by the settings used during the Lokomat® gait training session (range of motion, amount of guidance, speed). Mehrholz et al. (2007) published a review on robotic-assisted gait training devices used in rehabilitation to improve walking after stroke. They concluded that robotic-assisted gait training increases the odds of becoming independent in walking, but does not increase gait velocity (Mehrholz et al. 2007). The effects of Lokomat® training have been evaluated in stroke patients using outcomes such as functional tests and spatiotemporal parameters, but, as far are we aware, no studies have evaluated the effect of Lokomat® gait training on kinematic and kinetic gait parameters in hemiparetic patients. To this end, we will investigate these gait parameters. Over the past ten years, the concept of restraining the non-affected part of the body such as the upper limb (Kunkel et al. 1999; Sirtori et al. 2009), lower limb (Marklund and Klassbo 2006; Regnaux et al. 2008) or trunk (Michaelsen et al. 2001) to force the use of the affected part has been developed in the rehabilitation of stroke patients. Marklund and Klassbo (2006) and Regnaux et al. (2008) found gait improvements following restraining the non-paretic lower limb (with an orthosis limiting knee movements or with a mass on the ankle, respectively) in hemiparetic patients (Marklund and Klassbo 2006; Regnaux et al. 2008). These studies suggest that the use of a negative constraint on the nonparetic limb of hemiparetic patients during a gait training session could be a useful rehabilitation technique for


Exp Brain Res (2014) 232:31–42

the improvement of gait. However, the absence of control groups in these studies does not allow the impact of such a rehabilitation technique to be clearly determined. Recently, our team found no superiority of a constraint gait training session with a mass applied on the nonparetic ankle in comparison with a control overground gait training on gait parameters of stroke patients (Bonnyaud et al. 2013). Further studies are therefore suitable to investigate the other types of lower limb constraint on stroke patients’ gait improvement. The first aim of robotic devices used in rehabilitation is to assist and favour movements which are close to physiological movements. The Lokomat® is currently used to impose physiological movements meaning positive constraint for the paretic lower limb (i.e. an increase in range of motion to facilitate movement). Most studies assessing the effects of Lokomat® training used this paradigm. As well as imposing a positive constraint, it is also possible to impose a negative constraint (i.e. a reduction in range of motion to restrain the movement) using the Lokomat® robotic system. A highly original gait training paradigm would be to impose a positive kinematic constraint on the paretic lower limb and a negative kinematic constraint on the non-paretic lower limb. A constraint gait training like this one might provide proprioceptive inputs in the range of motion not usually used by stroke patients. This hypothesis is in accordance with recent results of Houldin et al. (2011) who suggested that proprioceptive feedback, induced by resistive movements, could enhance locomotor strategies in incomplete spinal cord injury. The aim of this randomized controlled study was thus to compare the effects of a gait training session using the Lokomat® set in order to provide a negative constraint on the non-paretic limb and a positive constraint on the paretic limb, with conventional Lokomat® gait training inducing a symmetrical gait pattern. We evaluated spatiotemporal, kinematic and kinetic gait parameters using 3D gait analysis after the gait training session. We hypothesized that a gait training session with this asymmetrical Lokomat® gait pattern would improve hip and knee kinematics on paretic side, which were our primary outcomes, more than conventional symmetrical Lokomat® training.

Methods Subjects Twenty-six hemiparetic subjects voluntarily participated in this study (14 with right-sided hemiparesis, 12 with left-sided hemiparesis, average age 50.7 ± 11.8 years, time since stroke 6.7 ± 8.9 years). Patients’ characteristics


Exp Brain Res (2014) 232:31–42 Table 1  Patients characteristics Age (years)

Side of hemiparesis

Time post-stroke (years)

Sex (M/W)

LE group

52.1 ± 13.8

7.8 ± 11.8

LC group

49.1 ± 9.5

8 right 5 left 6 right 7 left

9 M 4 W 8 M 5 W

All patients

50.7 ± 11.8

14 right 12 left

5.5 ± 4.7 6.7 ± 8.9

are presented in Table 1. The two groups were closely similar for clinical parameters and gait speed. The inclusion criteria were as follows: a single-hemispheric cerebral vascular lesion more than 6 months previously, able to walk 10 meters with no assistance or assistive device and for 20 min non-stop. All subjects gave written informed consent. The study was carried out according to “The ethical codes of the World Medical Association” (Declaration of Helsinki) and was approved by the local ethics committee. Experimental procedure After inclusion, patients were randomized to one of two groups, each corresponding to a specific gait training condition: • Lokomat® constraint training (which we will call Lokomat® experimental training, LE) with a negative kinematic constraint applied to the non-paretic limb (smallest range of motion as possible) and a positive kinematic constraint applied to the paretic limb (largest range of motion as possible). The Lokomat® was set in order to impose the largest as possible degree of hip and knee flexion/extension asymmetry between the paretic and non-paretic limbs during the bipedal gait training. Ranges of motion were set to induce the maximum kinematics asymmetry that the patients could follow without triggering the safety system of the Lokomat® (which would interrupt the training). The individualized settings in this group are listed in Table 2. • Lokomat® conventional training (LC) based on a symmetrical kinematic pattern. Range of motion was set to induce a pattern based on physiological range of motion (45° around hip joint and 60° around knee joint). It can be noticed that the kinematic settings of the Lokomat® were not equivalent in both groups with a larger kinematics pattern on the paretic side in the LE group than in the LC group, and a smaller kinematics pattern on the nonparetic side in the LE group due to the negative constraint applied.

17 M 9 W

Barthel (median)

NFAC (median)

Comfortable gait speed (cm/s)

95 ± 4.3

7 ± 0.7

75.5 ± 22.8

100 ± 2.4

7 ± 0.6

81.3 ± 17.2

100 ± 3.6

7 ± 0.6

78.4 ± 20.0

Table 2  Percentage of asymmetry induced by the Lokomat® in LE group Subjects

Hip asymmetry

Knee asymmetry

1 2 3 4 5 6 7 8 9 10 11 12

27.3 ± 3.5 20.7 ± 0.0 30.7 ± 8.1 29.3 ± 7.2 40.4 ± 6.2 35.3 ± 4.4 25.7 ± 4.3 46.5 ± 15.6 38.7 ± 8.9 47.3 ± 5.9 18.7 ± 6.2 15.7 ± 3.7

31.6 ± 5.5 40.1 ± 3.9 61.4 ± 15.9 34.1 ± 9.3 46.4 ± 8.7 33.5 ± 10.4 28.2 ± 3.4 54.2 ± 26.6 45.0 ± 12.3 36.4 ± 13.7 45.9 ± 8.5 21.2 ± 4.9


50.8 ± 5.8

62.1 ± 16.3

Percentage of asymmetry induced by the Lokomat®  is calculated with the Symmetry Index (SI) defined by Robinson et al. (1987) The Symmetry Index (SI) was calculated following Equation 1: |(Vparetic−Vnonparetic)| × 100 SI = 0.5|(Vparetic+Vnonparetic)| Equation 1. Symmetry Index SI: percentage symmetry index; V paretic: value of the hip (or knee) range of motion of the paretic lower limb; V non-paretic: value of the hip (or knee) range of motion of the non-paretic lower limb A value of zero indicates perfect symmetry. Higher values indicate a higher level of asymmetry on the paretic side relative to the nonparetic side

Gait training session parameters Each subject underwent 20 min of gait training using the basic version of the Lokomat® (Hocoma, Volketswil, Switzerland). Gait speed was set to 1.5 km/h for each patient. Guidance was set to 100 % for both groups in order to ensure that the full range of the imposed pattern (symmetry in the conventional group or asymmetry in the experimental group) was carried out. Patients were instructed to participate as actively as possible without stopping. Peurala et al. (2004) demonstrated the importance of therapist instructions during gait training to optimize the effects of rehabilitation (Peurala et al. 2004). The 3D motion analysis



system was located in a different room to that where the training sessions were performed; therefore, subjects were moved in a wheelchair pushed by the therapist between the laboratory and training room. Assessments Clinical Before the training session, the patients underwent a clinical evaluation of their level of independence in activities of daily living (Barthel Index) and their gait level (New Functional Ambulation Classification NFAC) (Mahoney and Barthel 1965; Brun et al. 2000).

Exp Brain Res (2014) 232:31–42

Statistical analysis The average value of each gait parameter for each group and the standard error of the mean are reported in Tables 3, 4 and 5. Since the variables analysed were normally distributed (test Shapiro–Wilk), a repeated measures of variance ANOVA with condition as the first factor (experimental training or conventional training) and time as the second factor [before training (Baseline), immediately after training (Post0) and after 20 min of rest (Post20)] was carried out. All analyses were performed using Statistica (version 7.1), and a significance level of p 

Effect of a robotic restraint gait training versus robotic conventional gait training on gait parameters in stroke patients.

Kinematic and kinetic gait parameters have never been assessed following robotic-assisted gait training in hemiparetic patients. Previous studies sugg...
891KB Sizes 0 Downloads 0 Views