How to improve walking, balance and social participation following stroke: a comparison of the long term effects of two walking aids--canes and an orthosis TheraTogs--on the recovery of gait following acute stroke. A study protocol for a multi-centre, single blind, randomised control trial

Stroke is the leading cause of acquired disability in adults [1]. In Switzerland approximately 9000 people a year suffer a first time stroke [2]. Of those patients who survive the acute phase, between 20 and 30% are unable to walk [3]. Those who can are often left with moderate to severe walking disability and reduced gait speeds [3, 4]. The risk of falling increases--a recent study indicating that at 3 months 28% of stroke patients have fallen [5]. The resulting disability has an enormous socioeconomic impact on patients, families and health service providers [6, 7].

Rehabilitation methods which improve balance and gait are crucial for the quality of life of stroke victims and to reduce the ongoing cost of long term care. Studies have shown that early rehabilitation in specialised settings e.g. stroke units results in better functional outcome than in non-specialised units [8, 9]. It has been suggested that these better results are achieved partly due to enhanced staff awareness of the importance of mobility thus preventing secondary complications such as loss of cardiovascular fitness or muscle weakness. The physical environment is also likely to be adapted enabling more independent movement [10, 11]. Current rehabilitation methods which aim to improve motor control during walking do not appear to deliver additional improvements [12, 13]. A recent meta-analysis showed that only cardiorespiratory physical fitness training provides robust evidence for a benefit to walking ability after stroke. Repetitive task training also appeared to have some effect. Motor and neurophysiological approaches did not demonstrate a positive effect on walking recovery [13]. This suggests that underlying mechanisms responsible for the recovery of motor co-ordination and control of walking following stroke are not significantly influenced by current therapy methods. These findings are further illustrated by studies which show that long term improvements in gait function occur in the absence of improvements in muscle co-ordination patterns [14] or improved kinematic or kinetic gait profiles [15]. Buurke et al. [14] concluded that "functional gait improvements may be more related to compensatory strategies than by restitution of muscle co-ordination patterns in the affected leg." Kautz et al. [16] concluded that "There is no evidence of improved locomotor co-ordination post intervention. The increased walking and pedalling speed were achieved by a more proficient use of the same impaired pattern without EMG timing changes, likely because of increased strength and endurance post intervention." We suggest that improvements in gait function due to current rehabilitation methods are predominantly achieved through more efficient use of abnormal movement patterns. This may be a reason for the generally low level of independence and function achieved following stroke.

Spontaneous recovery which occurs within the first weeks post stroke largely defines long term functional outcome [17]. Recent studies investigating changes in the cerebral cortex following focal injury have indicated that cortical plasticity and neuronal growth that occurs early after infarct may "underlie the brain's self-repair process" [18, 19]. Evidence shows that this early cortical plasticity is an important factor in the spontaneous recovery of motor control which predicts long term outcome [18–22]. We suggest that it is this process which is inadequately influenced by current rehabilitation methods. The recovery of motor skills following stroke relies on the brains ability to reorganise its neuronal control of movement [19, 20]. Reorganisation can occur within and between cortical networks both within the lesioned and non-lesioned hemisphere [18, 21]. Rossini et al. stated that "Reorganization phenomena following ischemic stroke observed so far, may be classified into two main groups: overactivation of areas belonging to the neural network for a specific task--or activation of unusual areas that attempt to replace the function of the damaged tissue" [18]. The overactivation of the original neural network is primarily activity in the lesioned hemisphere. Following stroke cortical activity has been recorded in both lesioned and non-lesioned hemispheres during movement [21, 23]. Increasingly findings show that increased activity in the lesioned hemisphere correlates with better recovery and improved motor performance [23–27]. Conversely increased activity in the contralesional hemisphere is associated with poorer motor recovery of the hemiplegic limb [26, 27]. It appears that improved motor recovery occurs when the brain is able to make use of the original neural network to control movement. When new networks are formed for example in the unaffected hemisphere, motor recovery is reduced. The authors suggest that activity in the original network represents "true" recovery. The recruitment of new networks may represent the learning of compensatory movement strategies associated with poorer functional outcome.

Findings in favour of this hypothesis have been demonstrated in both upper and lower limbs [25–27]. As post-injury behavioural experience is critical to the reassembly of adaptive networks and strongly influences cortical reorganisation [22, 28] these results indicate that treatment approaches which promote increased plasticity in the lesioned hemisphere will promote greater functional recovery. This assumption is supported by a recent meta-analysis [21] which examined whether participating in stroke rehabilitation which emphasised use of the paretic upper limb was associated with increased cortical activity of the lesioned hemisphere and consequently better function. The review concluded that neural changes in the sensorimotor cortex of the lesioned hemisphere were achieved with rehabilitation interventions that emphasized use of the paretic upper limb and resulted in improved functional motor gains.

Nudo et al. [28] demonstrated with squirrel monkeys that forced retraining of skilled hand use prevented loss of hand territory representation adjacent to the infarct and in some instances the representations expanded. This functional reorganization was accompanied by recovery of skilled hand movements. Control monkeys with identical lesions who did not receive therapy lost paretic hand representation und function.

In general these results suggest that therapies which emphasise use of the hemiplegic side promote plasticity in the lesioned hemisphere through emphasising the use of previously established neural networks and are associated with improved function. It may be that therapies which promote compensatory movement or reduce activity of the hemiplegic limb inhibit plasticity in the lesioned hemisphere and promote the development of new neural networks associated with reduced function.

To date therapies which emphasise use of the hemiplegic limb have been confined to the upper extremity [29, 30] and have not been applied during gait rehabilitation following stroke. Canes are very commonly used post stroke although studies have consistently shown a significant reduction in surface electromyography (EMG) activity in all muscle groups on the side contralateral to cane use in both stroke and non-stroke patients [14, 31–33]. In light of the factors which increase plastic reorganisation in the lesioned hemisphere, namely interventions which emphasise use of the paretic limb, the effect of canes which reduce activity in the hemiplegic side may be to inhibit activity in the original neural networks responsible for lower limb control resulting in poorer walking function.

In gait rehabilitation, attention should also be paid to the optimal restoration of balance. In relation to balance, evidence shows that balance control does not occur automatically at spinal cord and brainstem level as has previously been thought, but rather is highly influenced by cortical activity and cognitive control [34–36]. Two main types of balance strategies are recognised--fixed support or change of support strategies [34, 35]. Fixed support strategies are used when no stepping or reaching activities are needed to maintain balance. Rotatory torques are generated through muscle activity primarily around the hip and ankle. Change of support strategies are used in challenging conditions when stepping or reaching reactions are necessary to maintain equilibrium. Fixed support strategies use less cognitive resources than change in support [34, 35]. Elderly people or subjects with poor balance use more change in support strategies and therefore more cognitive resources than younger, healthy subjects to maintain equilibrium under the same conditions.

The authors suggest that balance rehabilitation should attempt to restore balance strategies used by healthy individuals--namely fixed support responses requiring fewer cognitive resources in unchallenging situations. However cane use increases the base of support when walking through use of the arms. This strategy may emphasise use of cortical networks used for reaching "change of support" reactions. It emphasises use of cognitive resources on safe level ground and reduces the use of "fixed support" strategies. The authors suggest this leads to a long term reduction of automatic balance responses. It may be that because cognitive resources are needed in safe, level environments fewer additional resources are available for more difficult conditions such as walking outside or on public transport. The long term effects of canes on balance recovery and functional gait following stroke has to our knowledge been investigated in one study [37]. Balance recovery and community participation were shown to be reduced.

Taking all of these considerations together, the authors suggest that an optimal walking aid for post stroke gait rehabilitation would provide enough support to enable independent early walking without reducing hemiplegic muscle activity or inhibiting the use of balance reactions. The immediate effect of an elasticised orthotic walking aid (TheraTogs) on hemiplegic hip abductor activity has been previously investigated by the authors [38]. Activity in gluteus medius was increased by 16.5% compared to walking without aids when walking with TheraTogs (with cane use activity in gluteus medius was reduced by 22% compared to walking without aids). The increased activity with TheraTogs may be due to increased proprioceptive input provided by the orthosis or to the physical shortening of the muscle leading to increased overlap between the actin and myosin filaments and consequently a stronger contraction.

The authors hypothesize that cane use will inhibit activity on the hemiplegic side leading to reduced ipsilesional cortical activity and poorer long term functional gait outcomes. In contrast we hypothesize that TheraTogs will increase activity on the hemiplegic side during walking leading to increased ipsilesional cortical activity and improved long term functional gait outcomes.

We further hypothesize that cane use will inhibit the use of normal balance reactions leading to reduced balance recovery and poorer social participation. As no external support is provided with the TheraTogs orthosis the use of automatic balance responses will not be inhibited during walking. This will result in improved balance recovery and social participation.

The aim of this study is to investigate the long term effects of canes and TheraTogs on the recovery of motor control and co-ordination, gait, daily activity, balance and social participation when used in early gait rehabilitation following stroke.

Various forms of walking aids are commonly used in post stroke rehabilitation although the long term effects have been sparsely researched. In order to ensure the positive effect of these treatments and to enable evidence based practice, these interventions should be clinically researched.

Early walking is an important aim of stroke rehabilitation for many reasons including the psychological well-being of the patient, to prevent loss of cardiovascular fitness or the development of secondary musculoskeletal problems. This study questions whether canes are the optimal walking aid to enable early independent walking as their use inhibits rather than stimulates activity of the hemiplegic musculature and reduces the use of automatic balance responses contrary to the aims of rehabilitation. A possible alternative to cane use was recently tested by the authors in whom the immediate effects of TheraTogs were to significantly improve hip abductor muscle activity compared to cane walking or walking without walking aids [38].

The rate of recovery of all impairments after stroke is greatest in the first few weeks and slows down after two to three months. Kwakkel et al. discussed an early window during which rehabilitation has the most long term impact [17]. The potential negative or positive effects of walking aids may therefore be particularly significant in this early stage. To our knowledge no studies have investigated this question before. The results of this study may have important clinical significance and could be used in the development of guidelines for gait rehabilitation following stroke.