Gait assessment in cerebral palsy

Definition, causes and prevalence

Cerebral palsy is now defined as a group of permanent disorders of the development of movement and posture which can be attributed to brain damage to the fetus or infant (Rosenbaum et al., 2007). The condition is different from other forms of brain damage such as stroke or traumatic brain injury because it happens while the brain is still developing and this affects the subsequent neurological and musculoskeletal development of the child. The motor disorders of cerebral palsy are often accompanied by disturbances of sensation, perception, cognition, communication and behaviour, and by epilepsy. Cerebral palsy is the commonest cause of physical disability affecting children in the developed world with a prevalence of around 2 cases for every 1000 live births (Stanley et al., 2000). Figures for other parts of the world are difficult to ascertain but it is generally assumed to be at least as prevalent.

The brain damage is much more common in infants born prematurely but the precise cause is unknown in the majority of cases. The ultimate cause is a failure of the oxygen supply to an area of the developing brain which may be a consequence of damage to the blood vessels, such as haemorrhage or embolism, or a more general drop in fetal blood pressure. The brain damage, once it has occurred, is static and will not get any better or any worse. The clinical manifestations, however, will continue to develop and change as the child grows and matures. This is particularly true of the musculoskeletal manifestations of cerebral palsy, which are very important in determining whether and how people with the condition will walk.

Classification

Gross motor function classification system

The principal means of classifying children with cerebral palsy is with the Gross Motor Function Classification System GMFCS (Palisano et al., 1997 and 2000). This is, essentially a five-point scale reflecting the severity of the condition as it affects motor function. There are different definitions for different age groups but that for 6–12 year olds is the most relevant to gait analysts (Table 6.1). It can be seen that most children who are suitable for gait analysis will be of levels I–III. More recently descriptors for 12–18 year olds have been produced (Palisano et al., 2008). These are similar to those in Table 6.1 but allow for some deterioration in motor ability that occurs in late childhood. While gait analysis is most often used for children with cerebral palsy it is worth remembering that cerebral palsy is a life-long condition, with between 25% and 50% of adults reporting further deterioration in walking ability in early adulthood (Day et al., 2007; Jahnsen et al., 2004; Murphy et al., 1995).

Table 6.1 Gross Motor Function Classification System for children aged 6–12 years

Level I	Children walk indoors and outdoors and climb stairs without limitation. Children perform gross motor skills including running and jumping, but speed, balance and coordination are impaired
Level II	Children walk indoors and outdoors and climb stairs holding onto a railing but experience limitations walking on uneven surfaces and inclines and walking in crowds or confined spaces and with long distances
Level III	Children walk indoors and outdoors on a level surface with an assistive mobility device and may climb stairs holding onto a railing. Children may use a wheelchair for mobility when travelling for long distances or outdoors over uneven terrain.
Level IV	Children use methods of mobility that usually require adult assistance. They may continue to walk for short distances with physical assistance at home but rely on wheeled mobility (pushed by adult or operate a powered chair) outdoors, at school and in the community
Level V	Physical impairment restricts voluntary control of movement and the ability to maintain anti-gravity head and trunk posture. All areas of motor function are limited. Children have no means of independent mobility and are transported by adults

Classification by motor disorder and topography

The brain damage that causes cerebral palsy can affect the nervous system in several different ways. In about 85% of people the major limitation is spasticity, which leads to over activity in specific muscles. A further 7% are dyskinetic (dystonic-athetoid) having mixed muscle tone (high or low), which leads to slow sinuous movements superimposed on the intended motor pattern. Another 5% have ataxia which leads to rapid jerky movements and affects balance and depth perception. People with dyskinesia or ataxia tend to have very variable gait patterns, which can limit the usefulness of clinical gait analysis. Outcomes of orthopaedic surgery in people with dyskinesia can be unpredictable. Most children attending for clinical gait analysis are therefore those with a predominantly spastic motor type.

Before the advent of the GMFCS, the primary classification of children with cerebral palsy was with respect to the areas of the body that were most affected. Hemiplegia refers to one side of the body being affected including an arm and leg on the same side. Diplegia refers to the involvement of both legs and quadriplegia to involvement of all four limbs. Monoplegia (involvement of one limb) and triplegia (involvement of three limbs) are also used. Many children do not fall neatly into these categories and the distinction between quadriplegia and diplegia is rarely clear-cut; many diplegic individuals are affected much more on one side than the other. Such issues have led to recent advice that such terms be discontinued until they are defined more precisely. Unilateral and bilateral motor involvements are now the preferred terms for children who can walk with total body involvement being used for non-walkers. Having said this, the original terms are still in widespread use.

Classification by gait pattern

There have been several attempts to classify cerebral palsy on the basis of the gait pattern (Dobson et al., 2007). Some confusion has arisen because several authors have chosen to re-define terms used by previous authors and it is not always clear which classification system is being referred to by which term. Two classifications for spastic hemiplegic gait have been proposed (Hullin et al., 1996; Winters et al., 1987) of which Winters et al. is more commonly used. This is described in Figure 6.1 and Table 6.2. There is some confusion in how this is applied clinically. The original paper refers exclusively to the sagittal plane and is based purely on the gait pattern. It is not at all uncommon to hear clinicians incorporating consideration of the transverse plane (particularly at the hip) and inferences about the nature of the underlying pathology (rather than just the gait pattern) in classifying children. Agreement between clinicians in applying this classification is substantial (Dobson et al., 2006). A more recent population-based study suggested that groups I, II and IV are far more common than group III.

Fig. 6.1 • Common gait patterns in hemiplegic cerebral palsy

Table 6.2 Classification of gait patterns in hemiplegia *

Group I	Ankle equinus in the swing phase of gait due to underactivity of the ankle dorsiflexors in relation to the plantarflexors
Group II	Plantarflexion throughout stance and swing from either static or dynamic contracture of the triceps surae. Knee is often forced into slight hyperextension in middle or late stance
Group III	Findings of type II with reduced range of knee flexion/extension. Reference is now common to sub-types IIIa: reduced knee extension during stance, and IIIb: hyperextension in stance and reduced flexion in swing
Group IV	Findings of type III with involvement of the hip musculature†

^* Summary by Thompson et al. (2002) of original by Winters et al. (1987).

^† In the original paper, in which measurements were restricted to the sagittal plane, this was attributed to flexor and adductor involvement. A three-dimensional analysis would almost certainly have included increased internal rotation

Classifications of gait patterns in spastic diplegia are less clear cut. One example is the classification by Rodda et al. (2004) in which the whole gait pattern is described (Fig. 6.2 and Table 6.3). There is particular confusion with regard to a number of terms that different people have chosen to define differently. Sutherland and Davids (1993), for example defined ‘jump’ as referring to a pattern in which the knee is flexed in early stance but extends rapidly in a pattern reminiscent of jumping. Rodda et al. (2004) used the same term to refer to a posture of flexed knee and plantarflexed ankle in late stance. ‘Crouch’ gait is another term that has been defined differently by a range of authors. Many studies simply use the term to refer to a gait pattern exhibiting knee flexion of no less than a certain value (30° is commonly used) throughout stance. Gage et al. (2009) used a kinetic definition for patterns in which the internal knee moment is extensor throughout stance, whereas Rodda et al. (2004) required ankle dorsiflexon as well as knee flexion.

Fig. 6.2 • Common gait patterns in spastic diplegic cerebral palsy (Rodda et al., 2004).

Table 6.3 Common gait patterns in spastic diplegic cerebral palsy (Rodda et al., 2004)

Group I	The ankle is in equinus. The knee extends fully or goes into mild recurvatum. The hip extends fully and the pelvis is within the normal range or tilted anteriorly
Group II	The ankle is in equinus, particularly in late stance. The knee and hip are excessively flexed in early stance and then extend to a variable degree in late stance, but never reach full extension. The pelvis is either within the normal range or tilted anteriorly
Group III	The ankle has a normal range but the knee and hip are excessively flexed throughout stance. The pelvis is normal or tilted anteriorly
Group IV	The ankle is excessively dorsiflexed throughout stance and the knee and hip are excessively flexed. The pelvis is in the normal range or tilted posteriorly

Classification by gait pattern does not give the whole picture. The two most commonly used schemes are based on sagittal plane data only (Rodda et al., 2004; Winters et al., 1987). Winters et al. proposed their classification system for people who had had no previous orthopaedic intervention and it is not clear how much value the scheme has for people who have had surgery. Rodda et al. did not discuss how surgery, or other interventions, would affect classification by gait pattern. There is considerable variability within grades and some people clearly have patterns on the boundary between one grade and the next. Indeed two studies which include data that can be used to infer whether natural groupings exist (Dobson, 2007; Rozumalski and Schwartz, 2009) would suggest, on the contrary, that gait patterns vary continuously across a multi-dimensional spectrum and groupings will only ever be a rather gross guide to the general pattern of movement. They can, however, still be useful in giving an overall impression of the person’s gait pattern particularly if full instrumented gait analysis is not available. A full gait analysis is still essential if we want to be specific about how a person is walking and what is causing them to do so in a particular manner.

Impairments

An impairment is something that is wrong with a person’s body structures or the way they function (World Health Organization (WHO), 2001). Examples of impairments include the loss of a limb or loss of vision. These can be complex in people with cerebral palsy. Primary impairments are direct consequences of the brain damage, and thus affect basic neurological function, whereas secondary impairments are indirect consequences arising from the effects of altered neurology on other structures over a long period of time. Thus muscles may become contracted or bones misaligned. Impairments generally fall into four broad categories: spasticity, weakness, muscle contracture and bony malalignment.

Spasticity

The term ‘spasticity’ means different things to different people. Most clinicians working with people with cerebral palsy ¹ tend to use it to refer to a very specific ‘velocity-dependent increase in tonic stretch reflexes, with exaggerated tendon jerks resulting from hyperexcitability of the stretch reflex’ (Lance, 1980). In other words if a tendon is stretched rapidly then its muscle will become active. Depending on the severity of the spasticity this might be for a short period and the muscle may thus show phasic, but inappropriate activity throughout the gait cycle, or for a longer period in which case the muscle may be inappropriately active throughout the gait cycle. This stretch reflex is present in all of us but in most people is suppressed by signals originating in the brain and passing down the spinal column (descending control). In people with cerebral palsy the brain damage reduces the capacity to send such signals and the reflexes are effectively out of control. Thus, although it is common to hear reference to ‘spastic muscles’, spasticity is really a property of the nervous system.

In cerebral palsy spasticity commonly affects some muscles more than others. In particular the bi- and multiarticular muscles appear to be susceptible (Gage et al., 2009). Thus spasticity is commonly found in the gastrocnemius, hamstrings, rectus femoris and psoas muscles. In more severely affected children, spasticity of the monoarticular hip adductors is a particular issue.

Spasticity, so tightly defined, is not the only alteration to neurological function. In some people with cerebral palsy, particularly those most severely affected, there can be inappropriate muscle activity in the absence of movement as well, and this is generally referred to as resting tone. Dystonia and ataxia are also essentially the manifestation of neurological impairments.

Muscle contracture

Spasticity occurs in a wide range of conditions and in most of these the muscles affected are susceptible to develop contractures. In these the passive length of the muscle and its tendon are reduced and this leads to a restriction in the range of movement that is available at the joints. Most children with cerebral palsy are relatively free from contractures at birth and in infancy and then develop them through childhood. They are commonly named by the position they are held in: for example, a flexion contracture of the knee means that the knee is held in some degree of flexion and cannot fully extend, with the most commonly seen contractures including knee flexion, ankle plantarflexion and hip flexion.

The mechanisms by which contractures develop are not fully understood. In relation to children with cerebral palsy it is quite common to hear contractures talked about as a failure of normal growth, where the bone keeps on growing but the muscle does not. This cannot be the full picture, as some children, particularly those with hemiplegia, can develop contractures over a short period of time which are too severe to be explained simply by failure of growth. Contractures are also common in many adults who have spasticity (in conditions such as stroke). Another common belief is that contractures are a result of immobilisation, which is supported by some early clinical and animal work suggesting that this reduces the length of the muscle fibres (Shortland et al., 2002). Again though, there are many other conditions in which immobilisation of muscles is much more complete than in cerebral palsy, but such severe contractures do not develop. More recent work suggests that a reduction in muscle belly length through shortening of the aponeuroses is much more significant than any reduction in fibre length (Shortland et al., 2002), but no mechanism has been proposed for this. In summary, the development of contractures is probably related to failure of growth, immobilisation and the consequences of spasticity but the relative contributions of these factors remain unknown.

Weakness

It is only relatively recently that the effect of muscle weakness on the gait of people with cerebral palsy has been fully acknowledged. Wiley and Damiano (1998) surveyed the strength of children with cerebral palsy and demonstrated that muscle weakness is also important with the gluteal muscles with the plantarflexors also being particularly weak with respect to those of age-matched able-bodied children. Muscle weakness can arise either because of the anatomy and physiology of the muscles or through reduced neural activity to stimulate contraction. Despite cerebral palsy being an essentially neurological problem there have been few investigations to establish the relative proportions of the muscular and neurological components of weakness.

The weakness has generally been attributed to a muscle’s reduced physiological cross-sectional area attributed to short muscle belly length and may therefore be strongly related to the development of contracture in both spastic hemiplegia and diplegia (Elder et al., 2003; Fry et al., 2007). The clinical picture does not always agree with this, with some muscles appearing to be quite weak without evidence of contracture. It is also known, however, that there is more extra-cellular material of lower quality than in normal muscle. Recent evidence suggests that these changes are quite different from those that arise simply through either chronic stimulation or disuse (Foran et al., 2005) but, as with the development of contracture, the underlying mechanisms are not understood. During movement there is a further factor in that contractures or specific walking patterns may result in muscle fibres functioning away from their optimal length to generate contractile force. The clinical picture appears to be that the muscles get weaker with age with respect to those of children without cerebral palsy. A particular issue arises in late childhood and early adolescent when weight increases rapidly and relative weakness becomes significant.

Bony malalignment (and capsular contracture)

Children with cerebral palsy are also susceptible to developing malalignment of the bones. Anteversion of the femur is the most common example of this. It is essentially a twist somewhere along the shaft of the femur which results in the femoral neck pointing too far forwards in relation to the knee joint axis. Although it is commonly referred to as a developmental deformity it is actually a result of persistence of the original bony alignment. At birth most children have anteversion of about 40° but this reduces in normal growth to about 10° at skeletal maturity. This reduction does not occur in many children with cerebral palsy. It is not uncommon for the tibia to be twisted externally along its shaft, tibial torsion, and malalignment of the bones of the foot, particularly the calcaneus is also common. In more severely affected children spinal, thoracic and upper limb deformities are also common.

As with the other impairments the precise mechanisms are unclear but at least three factors need to be considered. Bones grow in length at the growth plates. Most long bones have a growth plate proximally and distally. The rate at which the bones grow is determined by the stress exerted across the growth plate. In areas of the growth plate across which large compressive stresses are exerted, longitudinal growth will be suppressed and elsewhere growth may be stimulated. Thus normal bone growth requires normal stress distributions. Walking is believed to be a major contributor to such stresses and thus if children do not walk normally the bones will not grow normally. Remodelling is another factor. Once bones have grown the bone continues to be replaced on an ongoing basis and this can lead to a change in shape of the bone. Interestingly the opposite law applies here. High compressive stresses tend to stimulate the production of more bone whereas lower stresses suppress this. The third factor is that cartilaginous bone can actually be deformed mechanically and this may be particularly important in the feet where some bones are not fully calcified until around the age of 10 years.

A further common impairment is contracture of the joint capsule. This is particularly common at the knee but may also restrict extension and internal/external rotation of the hip. It results from contracture of the ligaments which form the joint capsule and is thus quite distinct from contracture of the muscle. It is included with bony deformities because the capsule is a deep structure and quite difficult to approach surgically. Management of joint contractures is thus often achieved through bony surgery similar to those used to manage bony malalignment.

Clinical management

Strengthening

Now that the importance of weakness in cerebral palsy has been acknowledged there has been considerable research into physiotherapy programmes to actively strengthen muscles (Damiano and Abel, 1998; Damiano et al., 1995, 2002; Dodd and Taylor, 2005; Dodd et al., 2002, 2003). These generally used the principles of progressive resistive strength training, which have now been demonstrated to result in increases in muscle strength in children with cerebral palsy of up to 25% (which is consistent with their use in many other conditions). Such programmes are now being used more and more routinely.

Gait analysis

Clinical gait analysis

The most obvious use of clinical gait analysis is for planning complex, multilevel orthopaedic surgery. In most cases the surgeon already knows that surgery is required, and the aim of the gait analysis is to determine exactly which combination of surgical procedures will be of most benefit to the individual child. Many centres will also perform follow-up analysis, often between 1 and 2 years after surgery to assess outcomes as part of clinical audit. This allows the clinical team to benefit from the experience in managing each child and is important to maintain and improve levels of service provision. Gait analysis can also be useful in planning botulinum toxin injections, physiotherapy, orthotic interventions and more general monitoring of progress. Unfortunately the cost, availability and time required for the analysis often precludes its use for routine clinical purposes.

We will now focus on how clinical gait analysis is able to support surgical decision making. Gait analysis is only part of this process which includes capturing gait data, performing a comprehensive physical examination and providing a biomechanical analysis of the results. The actual decision as to whether surgery is required and which procedures should be included requires consideration of a number of other issues such as medical imaging, patient’s history and psychosocial background and the surgeon’s competences and level of support, which are outside the scope of this book.

Clinical gait analysis works within a framework of clinical governance that ensures that the services delivered to patients are both safe and of the highest quality. This includes a commitment to evidence-based practice, which requires that all techniques used should be well established and have been the subject of rigorous research. This is different from gait analysis in an academic environment, where innovation and experimentation are encouraged.

Data capture

Most clinical services focus on obtaining good-quality kinematic and kinetic data based on the measured positions of retro-reflective markers using techniques documented elsewhere in this book. The conventional gait model (CGM) (Baker and Rodda, 2003; Davis et al., 1991; Kadaba et al., 1990; Ounpuu et al., 1996) is by far the best documented and most common approach (see Chapter 4). Other models are used but it is questionable whether they have yet been sufficiently well validated in clinical practice to satisfy the strict demands of clinical governance. Many centres will also capture EMG data either concurrently with full kinematic data or with a reduced marker set.

Children are generally asked to walk up and down in bare feet first and with the minimum walking aids to achieve a reasonable gait pattern. This information gives the best indication of what their body is capable of. They may then be asked to walk when wearing their ankle foot orthoses and any other usual walking aids to give an indication of how they usually walk. As well as capturing full three-dimensional gait analysis data, it is important to capture good-quality standardised video recordings of the child walking.

Clinical examination

In addition to the gait analysis, a full clinical examination is also completed. The range of movement of various joints is assessed which helps identify muscle and joint contractures. Muscle testing is generally carried out, grading muscles on the conventional five-point scale (Kendall and Kendall, 1949) and this is often accompanied by a simple assessment of the degree of selectivity with which the child can control muscle activation. The Tardieu test (Boyd and Graham, 1999) is used to measure spasticity and the modified Ashworth scale (Bohannon and Smith, 1987), while often referred to as a measure of spasticity, should probably be regarded as a measure of resting tone. Measures of bony malalignment such as femoral anteversion and tibial torsion are also recorded. An example of the results of such a clinical examination are included in Figure 6.3. The results of the clinical examination give additional information which aids the process of interpreting the gait analysis data to elucidate exactly which impairments are most affecting the gait pattern.