Cerebral palsy (CP) is the most common cause of physical disability affecting children in developed countries. The prevalence is about 2 per 1000 live births and is not decreasing. Children with CP have complex needs and are usually managed by a multidisciplinary team. The medical literature dealing with CP is extensive and varies from level I (Randomized Clinical Trials [RCTs]), to cohort studies and case reports (levels IV and V). In this chapter, the most recent and the highest level of evidence will be cited, where possible. However, randomized trials in CP are difficult to perform and relatively few have been published, especially on the orthopaedic aspects of CP. When RCTs are not available, cohort studies, with long-term follow up and objective outcome measures, will be referenced.

Cerebral palsy was described in 1861 by the English Physician William Little, who recognized a link between difficult births and the development of deformities (1). For many years, CP was known as “Little’s Disease.” Little popularized tenotomy to correct deformity in CP and was the first to bridge the gap between neurology and orthopaedics. Although his understanding of the link between brain injury and deformity has stood the test of time, his views on the causation of CP have been superseded. It is now accepted that only 10% to 20% of CP is related to perinatal events. The influence of “difficult births” may have been much greater in Little’s era, when maternal health was poor, maternal and infant mortality was high, and obstetric services were primitive.

The term cerebral palsy was also used by Sir William Osler in 1889 in a book titled “The Cerebral Palsies of Children” (2). Freud considered CP to be caused not just at parturition but also earlier in pregnancy because of “deeper effects that influenced the development of the foetus” (3). Many other definitions have been proposed and debated since then (4).

The revised definition of CP, published in 2007, is as follows:

Cerebral palsy (CP) describes a group of permanent disorders of the development of movement and posture, causing activity limitation, that are attributed to nonprogressive disturbances that occurred in the developing fetal or infant brain. The motor disorders of cerebral palsy are often accompanied by disturbances of sensation, perception, cognition, communication, and behavior, by epilepsy, and by secondary musculoskeletal problems (5).

The new definition has been widely accepted and is recommended as the most useful current operational definition of CP. However, it should be remembered that there is no test, genetic, metabolic, immunologic or otherwise, that demonstrates the existence or absence of CP. There is no specified cause such as cerebral pathology or even type of motor impairment, only that motor impairment exists resulting from nonprogressive cerebral pathology, acquired early in life (6).

During the first trimester of pregnancy, the growth of the brain is rapid and the brain differentiates into a recognizable cerebrum, cerebellum, brain stem, and spinal cord at a very early age of fetal development. During this time of explosive growth, the developing brain is highly susceptible to genetic influences, exogenous toxins, nutritional deficiencies and other insults, some of which can be characterized by meconium analysis (7). Neuronal development peaks in the second trimester. Neurons differentiate from neural stem cells around the periventricular regions and migrate centrifugally toward the surface of the cerebral cortex. This results in functional activity in neurons by 7 weeks of differentiation with reflex movements detectable in the fetus by the 15th week of gestation (8). By the end of the second trimester, the majority of neurons have been formed. Loss of neurons can be accommodated by neuronal plasticity but not by the generation of new neurons. The third trimester is characterized by extensive synaptogenesis and remodeling with glialization commencing in the second trimester and continuing at least until the age of 2 years. Myelination of neurons begins late in the third trimester, reaches a peak in the early years of childhood, and continues into adolescence, following a well-defined pattern (9). The myelination of complex pathways results in the progressive elimination of primitive reflexes, during the first 6 months of neonatal life as normal postural reflexes appear and the acquisition of gross motor skills occurs. In the typically developing infant, head control is achieved by age 3 months, independent sitting by 6 months, crawling by 8 months (usually accompanied by pulling to stand) and independent walking by the age of 12 months. However, even typically developing infants may take 3 to 6 months longer than these mean figures and still be considered to have typical development (10).

The development of gross motor function in children with CP can be described by a series of curves that were derived from longitudinal measurements of gross motor function, using the Gross Motor Function Measure (GMFM) (11, 12) (Fig. 14-1). The curves show rapid acquisition of gross motor function in infants with a progressive separation of the curves especially between the ages of 2 and 4 years. The curves plateau between the ages of 3 and 6 years. The five gross motor curves constitute the five levels of the Gross Motor Function Classification System (GMFCS) (11, 12, 13, 14 and 15).

Understanding the position of a child’s development in relation to their gross motor curve provides a rational basis for the understanding of management strategies, goal setting, and long-term gross motor function. For example, a 2-year-old child GMFCS level II with signs of spastic diplegia is treated with a physical therapy program, ankle-foot orthoses (AFOs), and injections of Botulinum toxin A (BoNT-A) to the gastrocsoleus and hamstring muscles. Within 3 months, the child is noted to have progressed from standing with support to independent walking. While the intervention may well have contributed to these gains in gross motor function, the child is at the stage of rapid acquisition of gross motor function with or without intervention (12). This underlines the need for intervention studies in the first 6 years of life to be controlled. The popularity of many forms of intervention in early childhood in children with CP is the mistaken attribution of improvements in gross motor function to the intervention, when natural history has an undoubtedly much greater effect. Association is not causation.

In the majority of children aged 6 to 12 years, gross motor function has reached a plateau (Fig. 14-1). At the same time, gait parameters are noted to show deterioration as contractures and bony deformities increase (16, 17, 18 and 19). Changes in gross
motor function and in gait during this plateau can be more realistically attributed to intervention, and longitudinal cohort studies are less liable to misinterpretation than in the birth to 6 years age group.

The incidence of CP varies from 1 to 7 children per 1000 live births, according to maternal health, prenatal and perinatal maternal, and child health care services (6, 20). Prevalence rates are accurately reported in countries with well-developed health care services and are most reliable in countries with national CP registers including some European countries and Australia. In these countries, prevalence rates are comparable at around 2 per 1000 live births (20, 21). In most countries, prevalence rates are either static or increasing.

There is a paradoxical relationship between prevalence rates and the provision of neonatal intensive care. Sophisticated neonatal intensive care for premature and low birth weight infants may reduce the risk of brain injury in some and eliminate brain injury in other high-risk neonates. However, the lives of very premature and very low birth weight neonates with a severity of health problems, which would previously have resulted in premature mortality, are saved. These infants may survive with an increased risk of moderate and severe CP (6, 20).

Males are at higher risk of CP, perhaps due to gender-specific neuronal vulnerabilities (22). The risk of CP increases with decreasing gestational age. However, because births before 32 weeks contribute <2% of neonatal survivors, they contribute a minority (20% to 25%) of all CP in developed countries (6, 20, 23). The majority of CP cases are born at term.

The risk of CP increases 4-fold in twins and 18-fold in triplets (24, 25, 26 and 27). The widespread use of in vitro fertilization has greatly increased the rate of multiple births, which has resulted in an increase in CP rates (6, 28). The high CP rates in multiple births are in part explained by shorter gestation and low birth weight, but these are not the only factors. Any factor causing preterm birth may lie on a causal pathway to CP (6).

The introduction of statewide and national registers is extremely important in monitoring the prevalence of CP and detecting changes. In this way, causative factors and causal pathways may be identified that in turn may lead to primary and secondary preventive strategies (6, 21, 24).

CP is the most common cause of the upper motor neuron (UMN) syndrome in childhood, a syndrome characterized by positive features (spasticity, hyperreflexia, and co-contraction) and negative features (weakness, loss of selective motor control, sensory deficits, and poor balance) (Fig. 14-2). Clinicians have traditionally focused more on the positive features because it is possible to treat spasticity. However, it is the negative features, which determine the locomotor prognosis. Weakness and loss of selective motor control determine when or if a child will walk. Balance deficits may dictate long-term dependence on a walking aid (13, 15).

The brain lesion, which results in CP, is a “static encephalopathy.” In other words, the brain lesion is not progressive and is unchanging. This is in contrast to musculoskeletal pathology in the limbs, which is progressive and constantly changing during growth and development (6, 29, 30).

At least 70% of cases have antecedents during pregnancy and only 10% to 20% have any relation to the child’s delivery (6, 20, 31). The mechanism of causal pathways suggests that in any one case of established CP a number of factors may have contributed to the brain lesion resulting in the specific clinical phenotype. There are a number of genetic predispositions to CP that may require an addition of an environmental trigger such as a maternal infection to be expressed as a brain lesion and CP. The genotype may load the gun and the environment pulls the trigger. A large number of major brain malformations have a genetic basis, and subtle genetic polymorphisms may also play a role (31, 32).

About 10% of infants with CP weigh <1500 g at birth. In this low birth weight group, the risk of CP is 90 per 1000, compared to 3 per 1000 in infants born at term and weighing more than 2500 g. Maternal risk factors include viral infections, urinary tract infections in late pregnancy, dietary deficiency, some prescription drugs, drug or alcohol abuse, maternal epilepsy, mental retardation, hypothyroidism, preeclamptic toxemia, cervical incompetence, and third-trimester bleeding. Obstetric risk factors include multiple births, placental abruption, premature rupture of membranes, chorioamnionitis, and prolonged labor. Other obstetric factors include the administration of oxytocin, cord prolapse, and breech presentation, when accompanied by low Apgar scores (6, 28). Neither routine use of fetal monitoring during labor nor increasing rate of Cesarean section has resulted in a decrease in the prevalence of CP (6).

The older the child at the time of the acquired brain lesion, the more the clinical syndrome is likely to differ from classical CP. Age 2 to 3 years is an important watershed (5).

CP was formerly a clinical diagnosis with occasional confirmation by central nervous system (CNS) imaging. With the availability of MRI and safer anesthesia for children, the majority of children suspected of having a CP will have brain imaging (31). A recent practice parameter from the American Neurological Association recommended that the diagnosis of CP be confirmed by imaging (33). In a large recently published multicenter study, the brain lesions identified by MRI are given in Table 14.1.

On the basis of their MRI scans, only 20% of cases of CP were considered as possibly being secondary to some type
of obstetric mishap. Twelve percent of cases, with a clinical diagnosis of CP, had a normal MRI scan (31).

By definition CP is a static encephalopathy, but the musculoskeletal pathology is progressive (6, 30). Chronic neurologic impairment affects the development of bones and muscles (Fig. 14-3). In spastic hemiplegia, the affected side demonstrates muscle atrophy and limb shortening, compared to the unaffected side. Thus, CP is a neuromusculoskeletal disorder (29, 30).

The key feature of the musculoskeletal pathology in CP is failure of longitudinal growth of skeletal muscle. An apt synonym for CP is “short muscle disease.” The conditions for normal muscle growth are regular stretching of relaxed muscle, under physiologic loading conditions and normal levels of activity (29). In children with CP, skeletal muscle does not relax during activity because of spasticity and the
children have greatly reduced levels of activity because of weakness and poor balance (30). In animal models of CP, such as the hereditary spastic mouse, there is failure of longitudinal muscle growth, in relation to bone growth. Affected mice develop equinus deformities because of a failure of longitudinal gastrocsoleus muscle growth when compared to tibial growth (34). However, muscle growth can be enhanced by the injection of BoNT-A soon after birth (35). The newborn child with CP does not have contractures or lower limb deformities, and most do not show signs of spasticity (29, 30). With time, spasticity develops, activity levels remain low, the growth of muscle-tendon units (MTUs) lags behind bone growth and contractures develop. Although the MTU is short in CP, muscle fibers may not be short (36, 37). Because of disordered growth and abnormal biomechanics, torsional abnormalities persist or develop in the long bones and instability of joints including the hip and subtalar joint develop. Eventually, premature degenerative arthritis may develop (6, 30, 38, 39, 40 and 41).

An important therapeutic window exists for spasticity management before the development of fixed contractures (38, 39) (Fig. 14-3). A second therapeutic window exists for the correction of fixed musculoskeletal deformities, before the onset of decompensation (39). There are three important longitudinal studies of gait in children with spastic diplegia that confirm that the musculoskeletal pathology and the attendant gait disorder are progressive during childhood. These studies provide an important insight into natural history and a framework to interpret the results of surgical intervention (16, 17 and 18).

CP may be classified by the cause (when known) and the brain lesion as determined by MRI. Classification by movement disorder, topographical distribution, and gross motor function may be relevant to management, including orthopaedic surgery (13, 31, 42, 43) (Figs. 14-4 and 14-5A,B).

The spinal cord is not involved in CP and the most common neurologic impairment is epilepsy, affecting about 30% of children. Epilepsy is most commonly seen in hemiplegia, especially when it is acquired postnatally and in children with quadriplegia (48, 51). Intellectual disability is variable and more severe in children at GMFCS levels IV and V. However, many children at GMFCS levels I to III exhibit learning difficulties, autism spectrum disorders, and behavioral and emotional difficulties. Impairments of hearing, speech, and vision are also common and may adversely impact schooling and learning. The prevalence of visual problems is so high (about 50%) that routine screening is advised (52).

In the majority of children who are referred to an orthopaedic surgeon, the diagnosis of CP has already been established by a pediatrician or a pediatric neurologist. The exception is a small number of children with mild hemiplegia and monoplegia who present after walking age with toe walking or some other subtle disturbance of gait. The postural abnormalities associated with monoplegia and hemiplegia are best observed by asking the child to walk and then run in a sufficiently long, well-lit corridor (Fig. 14-8).

Physiotherapy is the most popular and widely used management strategy in children with CP (80, 81, 82 and 83). Some have access to therapy services integrated within the school program and others outside of the school program. The frequency of such programs may be related to the severity of the individual child’s involvement but rarely reflect the family and the child’s real needs. This type of background physical therapy support is valued by the parents of children with CP. These programs provide education and emotional support to the parents coping with the diagnosis of CP and starting the journey through the uncharted waters of raising a child with a lifelong physical disability. Roles that may be assumed by therapists may include education, counseling, coordination of access to other services including physical medicine and rehabilitation, orthotics, and orthopaedic surgery. The parents will often seek the view of the child’s therapist on recommendations for spasticity management, the type of orthosis, and the timing and type of surgical intervention (80). The need for clear communication and teamwork within the multidisciplinary team is obvious. These “background” programs of physical therapy are rarely adequate to ensure an optimum rehabilitation from major interventions such as SDR or SEMLS. Therapy around such episodes needs to be carefully planned and is often best as a team approach involving the child’s community therapist as well as providing the additional therapy in the tertiary hospital or rehabilitation center.

The physical therapy management of children with CP has been based on a variety of theoretical perspectives. The theoretical frameworks most commonly applied can be broadly categorized into biomechanical (splints, orthoses, stretching, and strength training), neurodevelopmental, cognitive (including conductive education [CE] and motor learning), and constraint-induced movement therapy (CIMT) (80, 81, 82 and 83). These approaches are not mutually exclusive. They can have shared components and most therapists use a combination of approaches. Given that the surgical management of CP has a biomechanical basis, surgeons find it easier to communicate with therapists who share this approach (80).

The “spasticity compass” can be used to classify and compare interventions for spasticity management, as focal or generalized in their effect and as temporary or permanent (84). Each intervention can then be located in the appropriate quadrant. For example, oral medications are general (all nerves or all muscles in all body areas) but temporary in effect (86). SDR is a neurosurgical procedure in which 30% to 50% of the dorsal rootlets between L1 and S1 are transected for the permanent relief of spasticity in a highly selected group of children with
spastic diplegia. The principal effects are on the lower limbs although there may be minor effects on the upper limbs. The position on the grid is therefore permanent and half way between general and focal (87, 88) (Fig. 14-10).

Oral medications used for the management of spasticity in children with CP include diazepam, baclofen, dantrolene sodium, and tizanidine. Artane and L-dopa are used in dystonia. All are limited in usefulness by a combination of limited benefits and side effects. They have been extensively reviewed in several recent publications (84, 89).

The limited lipid solubility of baclofen when administered orally can be overcome by intrathecal administration using a programmable, battery-operated implantable pump connected to a catheter and delivery system to the intrathecal space (86). This is an invasive procedure with associated morbidity and mortality (38, 39, 88). However, it is the most effective current method available for the management of severe spasticity, dystonia, and mixed movement disorders in CP and a number of other conditions including spasticity of spinal origin and acquired brain injury. The role of ITB has been reviewed extensively in several recent publications (86, 88).

Chemodenervation is useful in the management of focal spasticity and dystonia. Phenol neurolysis was much more widely utilized before the introduction of Botulinum neurotoxin A (BoNT-A). The principal limitation on its use is pain at the site of injection and post injection dysesthesia. Phenol is not selective and has the same effect on sensory nerve fibers as motor fibers. The principal indications are neurolysis of the musculocutaneous nerve for elbow flexor spasticity and the obturator nerve for adductor spasticity. These nerves have limited sensory distribution (84, 90).

Injection of skeletal muscle with BoNT-A results in a dosedependent, reversible chemodenervation, by blocking presynaptic release of acetylcholine at the neuromuscular junctions (84). Because of the toxin’s rapid and high-affinity binding to receptors at the neuromuscular junctions of the target muscle, little systemic spread occurs. Neurotransmission is restored first by sprouting of new nerve endings, followed by the original nerve endings regaining their ability to release acetylcholine (91, 92). BoNT-A may be useful in children with CP to manage dynamic gait problems and to delay the need for orthopaedic surgery until the child is older (Fig. 14-11).

BoNT-A is generally safe in children with CP (93, 94, 98, 99). Most adverse events are localized, minor, and self-limiting. Systemic side effects including temporary incontinence and dysphagia have been reported (100). Dysphagia, aspiration, and chest infection are the most serious complications after injection of BoNT-A and if unrecognized or inadequately treated could lead to death from asphyxia (100, 122).

BoNT-A can be used to treat muscle spasm following operative procedures, such as adductor-release surgery (123). Injection of BoNT-A can be useful for short-term relief of pain associated with hip displacement (124). Target muscles include the hip adductors, medial hamstrings, and hip flexors. Pain relief is associated with a decrease in spastic adduction and scissoring postures (123, 124). It is not clear if short-term pain relief can be sustained by repeat injections and the need for salvage surgery avoided. Some children with neglected hip displacement have limited life expectancy and may not survive salvage surgery. BoNT-A may provide useful palliation in such circumstances (124). Better still is to prevent painful hip displacement.

In 1987, a four-group classification of sagittal gait patterns in spastic hemiplegia was developed by Winters et al. (125). Their four-group classification has been extensively used by
clinicians as a template for clinical management including prescription of orthoses, spasticity management by injection of BoNT-A, and musculoskeletal surgery (Fig. 14-12).

Knee patterns in spastic diplegia have been classified as recurvatum knee, jump knee, stiff knee, and crouch (128). The knee classification has been extended to the sagittal plane as true equinus, jump gait, apparent equinus, and crouch gait (97) (Fig. 14-12).

In a study of muscle excursion and cross-sectional area, it was demonstrated that the gastrocsoleus and ankle dorsiflexors have such different physical characteristics that they cannot
be considered to be “in balance,” in either normal subjects or in children with CP (129). The plantarflexors are six times as strong as the dorsiflexors. The plantarflexors of the ankle must be balanced against the GRF not the dorsiflexors. Lifting the foot and ankle during swing phase (dorsiflexor function) requires a very small muscle moment. Push-off in terminal stance (plantarflexor function) requires a large muscle moment. The concept of muscle balance should be redefined as a requirement for balance between the three anatomical levels, hip, knee and ankle, in the sagittal plane, not at a single level (129).

Lower limb muscles have different sensitivities to surgical lengthening, related to their gross anatomy and morphology. The soleus is exquisitely sensitive to lengthening, but the iliopsoas and semitendinosus are relatively resistant. A 1-cm lengthening of the soleus reduces its moment-generating ability by 30% and a 2-cm lengthening reduces its moment by 85% (130). A small error in terms of overlengthening the soleus may be disastrous. A 4-cm lengthening of the psoas is required to reduce its moment by 50% (130). The surgical implications are to lengthen the gastrocnemius only, when there is no contracture of the soleus. When the soleus requires lengthening, a precise and stable technique should be used, with careful control of the position postoperatively in a cast. By contrast, it is difficult to overlengthen the psoas. Intramuscular lengthening at the pelvic brim without immobilization postoperatively is safe and effective.

Many children who walk with flexed knee gait have hamstrings that are of normal length. It is the psoas that is shortened and requires lengthening, not the hamstrings. It is easy to do too much hamstring lengthening and not enough psoas lengthening (59, 131, 132).

The GMFCS gives an accurate summary of a child’s current gross motor function and long-term prognosis (13, 15). It is difficult to frame a logical discussion of musculoskeletal management outside of the context of the GMFCS. Most disagreements in the literature are apparent rather than real because those taking opposing views are often considering a child in a different GMFCS level. The terms mild, moderate, and severe diplegia along with mild, moderate, and severe quadriplegia are not meaningful or useful. Age-appropriate GMFCS descriptors are valid, reliable, stable, and clinically meaningful (47, 49, 50). Recommendations for both gait correction surgery and surgery for hip displacement are much more easily understood when the child’s GMFCS level is known. In the following sections, musculoskeletal management will be discussed by GMFCS level, in conjunction with both topographical classification and sagittal gait patterns. Gait correction surgery is an option for many children with CP at GMFCS levels I to III but with differences at each level. SEMLS will be discussed primarily in the GMFCS level II section. Preventive hip surgery will be discussed in the GMFCS level III section, reconstructive hip surgery in the GMFCS level IV section, and salvage surgery in the GMFCS level V section. Obviously, reconstructive surgery may be appropriate at GMFCS levels III to V and salvage surgery is sometimes needed at both GMFCS IV and V but is best avoided.

1. Between 6th and 12th birthday

2. Between 12th and 18th birthday

3. Risk of hip displacement: Developmental dislocation of the hip occurs at the same rate as in the normally developing population. Spastic hip displacement is not seen (49, 50).

4. Mean femoral neck anteversion (FNA): 30 degrees

5. Mean neck shaft angle (NSA): 136 degrees