Muscle Tone

Physical examination of a child with CP should include muscle tone in the extremities. With the patient relaxed (even sitting on the lap of a parent), the extremities are brought through a full range of motion. Spasticity feels like tightness in the muscles, which become tighter the quicker the limbs are passively moved. Greater range of motion can be gained by slowly and gently stretching the joints in question. The Tardieu test is a measure of spasticity. For example, if the examiner is assessing hamstring spasticity, the angle at which a “grab” of resistance occurs when quickly extending the knee with the hip in flexion is compared with the amount of extension possible when the knee is stretched slowly.

Fine motor activities should be assessed. Passing the child a toy or a pen often reveals spastic hemiplegia in one extremity. Having the child clap the hands or wiggle the fingers may reveal difficulties in fine motor control.

Reflexes

Deep tendon reflexes are increased in patients with CP. Repetitive tapping of the deep tendons or quick passive dorsiflexion of the ankle may produce clonus, which establishes the presence of an upper motoneuron neurologic abnormality. In hemiparesis, reflexes will be asymmetric.

Infantile reflexes disappear in normal children by 3 to 6 months of age as the motor cortex matures; however, they are retained in children with CP. Bleck’s textbook on CP outlines these reflexes in clear detail. The startle reflex, or the Moro reflex, which should disappear by 4 months of age, is elicited by letting the infant’s head drop back into extension with the infant supine but slightly elevated. This causes the legs and arms to extend abruptly. A sudden loud noise can likewise cause an older child to extend and lurch from a wheelchair. The parachute reflex is tested by holding the child in the air and then lowering him quickly headfirst toward the examining table. Children older than 5 months will reach out with both arms to protect themselves. Children with CP cannot do so, and those with hemiplegia will reach out with only one arm.

The tonic neck reflex is elicited by turning the supine infant’s head to one side. The ipsilateral arm and leg will extend while the contralateral arm and leg flex. This reflex should disappear in infancy; persistence should raise suspicion for CP.

Balance, Sitting, and Gait

Balance, sitting, and gait are assessed by noting whether the child can sit unsupported without use of hands or get into a sitting position without assistance or whether balance is easily disturbed in the sitting position or as the child walks.

Clinical assessment of gait requires that the child’s joints be readily seen, so the child should be barefoot and in shorts or a short gown. The evaluation should be conducted with the examiner seated on a stool at the level of the child. Enough room should be available for the child to walk naturally. Heel-to-toe walking, hopping on either foot, and running are observed. A patient with mild hemiplegia may walk nearly normally but exhibit abnormal movement patterns while running; the affected upper extremity will draw upward and not have a normal arm swing.

Gait should be observed from the front of the child and then from the side, and the hips, knees, and ankles should be systematically assessed from each perspective. A crouched gait consisting of increased flexion at the hip and knee, toe-walking with genu recurvatum, or a footdrop in the swing phase of gait may all be indicative of CP. Disturbance in clearance of the swing-phase limb may be caused by footdrop or an inability to flex the knee.

Histopathologic and Imaging Findings

Two findings frequently described on histopathologic examination or imaging studies of the brain in children with CP are periventricular leukomalacia and intraventricular and periventricular hemorrhage. Periventricular leukomalacia is defined as patchy areas of necrosis in the periventricular white matter adjacent to the lateral ventricles. It results from an ischemic insult to the arterial watershed area close to the ventricular walls. Pyramidal tract fibers mapping to the lower extremities pass through this area and are therefore more susceptible to injury than fibers responsible for the upper extremities and face. The bigger the lesion, the more fibers that are injured and the greater the proportion of the body that is affected by CP.

The areas of the brain immediately adjacent to the ventricles are also most susceptible to hemorrhage. Hemorrhage may be seen on ultrasound, MRI, or CT. Mild hemorrhages involve the germinal matrix adjacent to the ventricles, whereas more severe hemorrhages extend into the ventricles themselves and into the parenchyma of the brain. Hypoxia is known to predispose to periventricular and intraventricular hemorrhage. Approximately one half of preterm infants with CP are found to have abnormalities on neuroimaging, such as echolucency in the periventricular white matter or ventricular enlargement on cranial ultrasound. In children with CP born at or near term, about two thirds have abnormalities on neuroimaging, including focal infarction, malformations, and periventricular leukomalacia.

Gait Analysis

Gait analysis has become popular in the assessment of movement disorders in children with CP, but the usefulness of such studies is controversial. Accurate documentation of dynamic range of motion may help in planning surgical treatment and assessing the results of orthopaedic operations, but it does not replace a good clinical examination. When gait analysis graphs are scrutinized together with information from the clinical examination and slow-motion videotape, better understanding of a patient’s gait can be gained. A detailed discussion of gait analysis can be found in Chapter 5 . The Functional Mobility Scale is a questionnaire that assesses the child’s ambulatory ability at 5 m (within the home), 50 m (short distances), and 500 m (community ambulation) and should be considered in conjunction with gait analysis in surgical decision making.

Kinematics.

Certain kinematic patterns are seen in patients with CP. In the hip, scissoring, which is caused by tightness in the adductor musculature and, in part, by weakness of the hip abductors, leads to a narrow base of gait and difficulty advancing the swing-phase limb past the stance-phase limb ( Fig. 35-5 ). In the sagittal plane, increased hip flexion and anterior pelvic tilt may be part of crouch gait as a result of increased tone in the iliopsoas ( Fig. 35-6 ). Increased femoral anteversion may be documented by gait analysis as increased internal rotation of the hips. Asymmetric pelvic rotation may be present, and gait analysis is particularly helpful when the examiner is trying to ascertain whether the abnormal rotation is originating from the pelvis, hips, or tibiae ( Fig. 35-7 ).

In normal gait ( black dotted curve ), the hip adducts slightly in stance phase as the contralateral hemipelvis drops, and it abducts slightly in swing phase. In patients with scissoring as a result of cerebral palsy ( red curve ), adduction of the hip is increased. Vertical lines designate divisions between the stance and swing phases for each leg.

Normal kinematics for pelvic tilt and hip sagittal-plane motion is represented by the black dotted curve . Patients with cerebral palsy ( red curve is the right side, blue dotted curve is the left side) may crouch at the hip joint, which is represented on gait analysis as increased anterior pelvic tilt and lack of hip extension at terminal stance phase. Vertical lines designate divisions between the stance and swing phases for each leg.

Transverse-plane kinematics is useful in determining the cause of intoeing or outtoeing in cerebral palsy. In this hemiplegic patient, the involved hemipelvis is characteristically externally rotated ( red curves ), excessive femoral anteversion is present, and the foot is internally rotated relative to the knee because of equinovarus deformity of the foot. This results in a foot progression angle of 40 degrees internally. The uninvolved side is represented by the blue dotted curves , and normal transverse-plane kinematics is represented by the black dotted curves . Vertical lines designate divisions between the stance and swing phases for each leg.

At the knee, sagittal-plane motion is usually abnormal. In patients with tight hamstrings, the knee remains flexed at initial contact and is unable to extend normally during stance phase. In swing phase, spasticity of the rectus femoris may inhibit the patient’s ability to flex the knee and clear the floor ( Fig. 35-8 ). Genu recurvatum during stance phase may be present in response to a tight Achilles tendon, which causes difficulty advancing the tibia forward over the foot.

Crouch at the knee in a diplegic patient ( red curve is the right knee, blue dotted curve is the left knee) is documented as increased knee flexion throughout stance phase. As the leg enters swing phase, the knee slowly flexes and reaches maximal flexion later than in the normal gait ( black dotted curve ) because of rectus femoris spasticity, which interferes with clearing of the foot as it swings forward. Vertical lines designate divisions between the stance and swing phases for each leg.

Ankle kinematics often shows disturbances in plantar flexion, dorsiflexion, and push-off. Plantar flexion on weight acceptance is generally abnormal in patients with equinus contractures. Dorsiflexion during midstance is diminished in the presence of a tight Achilles tendon, and push-off is reduced if the ankle is already plantar-flexed from the equinus. Swing-phase dorsiflexion may be absent as a result of weakness or inactivity of the tibialis anterior and lead to footdrop ( Fig. 35-9 ).

Sagittal-plane kinematics of a patient with cerebral palsy who has bilateral equinus ( red curve is the right side, blue dotted curve is the left side). The ankle remains in plantar flexion during stance phase rather than progressively dorsiflexing, as in the normal second rocker ( black dotted curve ). Vertical lines designate divisions between the stance and swing phases for each leg.

Gait analysis is particularly useful in assessing the cause of toe-walking. It is tempting to attribute all toe-walking to tight Achilles tendons; however, some children walk on their toes in response to crouch above the ankle and have neutral ankle dorsiflexion but increased flexion at the knee and hip. Lengthening of the Achilles tendon in these children would result in a calcaneus gait, with persistent crouch at the hip and knee but excessive dorsiflexion at the ankle leading to inefficient push-off.

Electromyography.

Spasticity leads to electrical overactivity on electromyography (EMG) during gait, and it seems that the more spastic the child, the greater the EMG signal and the less phasic the muscles in their contraction ( Fig. 35-10 ). Dynamic EMG data collected during gait analysis can be correlated with kinematic and kinetic graphs to gain a fuller understanding of the biomechanics of the child’s gait. For example, normally at initial contact, the ankle plantar-flexes while the anterior tibialis fires on the EMG. Kinetic plots show absorption of power as weight acceptance occurs. In stance phase a child with CP may exhibit early heel rise, seen as plantar flexion on kinematic plots, which correlates with gastrocsoleus overactivity on EMG. We find data from EMG most useful when evaluating a child with a stiff knee in swing phase for rectus femoris transfer and when assessing a child with an equinovarus foot for tendon transfer (either anterior tibialis or posterior tibialis split transfer).

Typical electromyographic pattern during gait in a child with cerebral palsy. The horizontal bars represent the situation in which a muscle is normally “on.” Stance phase is represented from 0 to 60 and swing phase from 60 to 100. Contraction of all muscles is inappropriate during gait.

Kinetics.

Kinetics is the force exerted across joints during gait. Each joint has well-described kinetic patterns, and the reader is referred to work by Gage for descriptions. Two particular forces are clinically interesting: hip pull-off power and ankle push-off power. Hip pull-off power, the force exerted by the iliopsoas and other hip flexors to lift the stance-phase limb off the ground and into swing phase, is often diminished in patients with CP and leads to decreased energy efficiency ( Fig. 35-11 ). Ankle push-off power, the force exerted by the gastrocsoleus at terminal stance to push the stance-phase limb off the ground, is diminished in patients with equinus or calcaneus gait ( Fig. 35-12 ).

A normal hip generates power at terminal stance phase as the iliopsoas pulls the leg off the ground ( black dotted curve ). In patients with cerebral palsy ( red curve is the right side, blue dotted curve is the left side), hip kinetics can be disturbed. This patient generates little power at terminal stance phase and is therefore less efficient. Vertical lines designate divisions between the stance and swing phases for each leg.

As the gastrocsoleus contracts at heel rise, the ankle generates a burst of power ( black dotted curve ). In patients with either equinus or calcaneus gait secondary to cerebral palsy, power generation is decreased. ( Red curve is the right side, blue dotted curve is the left side). Vertical lines designate divisions between the stance and swing phases for each leg.

Physical Therapy

Frequently, the first treatment rendered to a child with CP is physical therapy. Yet no controlled studies have confirmed that regular physical therapy improves the outcome of a child with CP. One of the first well-designed studies investigating the effect of physical therapy was performed by Wright and Nicholson in 1973. They found no difference in motor skills or reflexes after 12 months in children who had neurodevelopmental training or physical therapy and those who did not. Other studies followed and again showed no discernible improvement after different forms of therapy. In defense of physical therapy, these studies are difficult to carry out because they involve different age groups and children with varying severity of neurologic impairment and usually encompass just a brief period of therapy. However, as Bleck pointed out, “The burden of proof is on the proponents of the treatment. Critics need not prove ineffectiveness but can insist on positive data.” The efficacy of physical therapy can be proved or disproved only with a properly designed, collaborative, multicenter, randomized, controlled trial. Such a trial has yet to be undertaken.

Physical therapy, ranging from the medical model, with the aims of attaining ambulation, range of motion, or transfers, to neurodevelopmental training, sensory integration, and even craniosacral therapy, has been proposed. Electrical stimulation of the muscles has also been used in these patients. Families like physical therapy and attribute gains in their young child’s ability to interaction with the therapist. However, some of these gains are simply the result of neurologic maturation of the child.

Our approach to physical therapy is to establish therapy for monitoring the developmental milestones of very young children, around the age of 2 or 3 years. Therapy is continued if gains are being made in attaining ambulation. School-based programs are used in elementary school and often include adaptive physical education. Postoperative intensive physical therapy is essential to reestablish range of motion and strength after surgical intervention. Strength training of weak muscles has been successful in improving motor function. We also draw on the expertise of physical therapists in assessing orthotic needs and wheelchair seating when appropriate. No evidence supports the continued use of physical therapy for range of motion, particularly in a nonambulatory child. Physical therapists play an important consultative role in making treatment decisions for patients whom they treat on a regular basis and whom we examine on a relatively infrequent basis.

It is common for parents of children with CP to resist discontinuing physical therapy. We believe that goals must be set for therapy and, if progress toward these goals is not being made, either the goals need to be reassessed or therapy should be stopped because it is not useful. Setting measurable functional goals has led to greater success after physical therapy.

In an older child, transitioning from physical therapy to therapeutic recreation is desirable and generally met with enthusiasm by the patient. Adaptive sports or swimming allows the child to participate with peers and affords greater enjoyment than exercises in the therapy gym do. Time in school should be spent on education at this age and not on physical therapy.

Casting

Inhibitive casting has waxed and waned in popularity as a mode of treatment of spasticity in children with CP. It is based on the presence of areas on the feet that are target centers for increased tone at the ankle and, some believe, throughout the lower extremities in certain patients. Usually, short-leg casts are applied with extended toe plates, careful molding of the heel, and metatarsal head control. This has been used by physical therapists and by some physicians. The time spent in casts varies but is generally a minimum of 6 weeks and is followed by the use of orthoses.

In our experience, casting has a limited role in patients with CP. We have used casting in rare cases of very young children (<4 years) with equinus contractures to delay surgical heel cord lengthening. Because equinus recurs after casting, permanent resolution of the equinus is not a goal. Some studies show short-lived improvements in passive and dynamic ankle range of motion following casting, whereas others have reported longer resolution of equinus contractures in children younger than 6 years.

Orthoses

Orthoses can be helpful in improving gait in ambulatory patients with CP. Bracing is best prescribed when specific treatment goals for gait improvement are identified. Ankle-foot orthoses (AFOs) are helpful in positioning the ankle and foot during gait. In young children, dynamic equinus can be improved, with a plantigrade foot position obtained and the tendency for genu recurvatum decreased.

AFOs come in various designs ( Fig. 35-13 ). Solid ankle AFOs are used when minimal dorsiflexion of the ankle is possible. Of the various types, they afford the easiest fit. Solid ankle AFOs can also be used in patients with excessive dorsiflexion and knee flexion; by stabilizing the ankle, extension of the knee is encouraged. Ground reaction AFOs have an anterior shell of plastic across the proximal end of the tibia and are rear entry. The goal of a ground reaction AFO is to provide push-back on the knee during stance phase and help the knee extend. Patients who walk in mild crouch with knee flexion contractures of less than 10 degrees may benefit from this design. Hinged AFOs permit a more natural gait by allowing some dorsiflexion during stance but blocking plantar flexion and therefore eliminating equinus and footdrop. Greater power generation in preswing is seen with hinged AFOs than with solid ankle designs ; however, hinged AFOs are bulkier and break more frequently, and they offer no benefit if the patient does not have passive dorsiflexion of the ankle. If crouch gait is present, a hinged AFO will worsen knee flexion, thereby leading to ankle calcaneus and a less efficient gait. Posterior leaf spring AFOs are fabricated from more pliable polypropylene and have more severe posterior trim lines at the ankle. They are touted as providing a little push-off at terminal stance. They tend to break and may not provide enough hindfoot control in most patients with CP. Gait analysis has shown that posterior leaf spring AFOs function similar to hinged AFOs; that is, they allow some dorsiflexion in stance phase through deformation of the brace, and they control equinus in swing phase but do not return energy at push-off.

A to C, Different views of common ankle-foot orthoses (AFOs) used in patients with cerebral palsy. The orthosis on the left is a ground reaction AFO that extends across the anterior aspect of the tibia to prevent flexion of the knee in stance. The center orthosis is a conventional solid ankle AFO. The orthosis on the right is a hinged AFO that allows dorsiflexion of the ankle but prevents equinus.

Tone-reducing (or dynamic) AFOs incorporate a molded footplate, termed an inhibitive footplate, and have higher trim lines that extend across the dorsum of the foot. The goal of these orthoses is to apply pressure similar to that of an inhibitive cast and therefore reduce tone throughout the lower extremities. Little scientific evidence supports this design inasmuch as no significant functional changes in gait have been noted with the use of tone-reducing as opposed to standard AFOs. Nonetheless, they remain popular with many families.

Bracing above the knee with knee-ankle-foot orthoses (KAFOs) or hip-knee-ankle-foot orthoses (HKAFOs) is not generally done in those with CP. The weight and bulk of the braces usually interfere with rather than improve the child’s ability to walk. Short-term use of KAFOs or knee immobilizers after soft tissue surgery for crouch gait can be helpful in retraining children to walk.

Our indications for bracing are

• 1.
  

  To obtain a plantigrade foot position and reduce genu recurvatum in patients with dynamic equinus

  
  
• 2.
  

  To support the foot in dorsiflexion during swing phase when footdrop is present

  
  
• 3.
  

  To assist the foot postoperatively while weakness is being treated by physical therapy

  
  
• 4.
  

  To improve mild crouch

Some patients walk worse with AFOs than without them. In patients with significant crouch gait, flexion of the knee and hip with the foot supported in an AFO forces the child to either pull up out of the braces or toe-walk despite the orthosis ( Fig. 35-14 ). Toe-walking has not ever been found to be harmful, so it is often better to allow the child to walk on toes brace free or to proceed with surgery when appropriate. Valgus deformities of the hindfeet are often difficult to brace because braces can produce painful calluses and blisters over the prominence of the talar heads. Comfortable footwear without orthoses is preferable for these feet. Finally, as young children with CP become adolescents, they often refuse to wear orthoses for cosmetic reasons. Allowing the patient to make a personal choice in whether to wear a brace is particularly important in this age group.

A, Bilateral equinus in a child with spastic diplegia. B, The orthosis is unable to maintain a plantigrade foot position because of spasticity in the gastrocsoleus and flexion at the hip and knee.

Oral Medication

Oral pharmacotherapy, including diazepam and baclofen, has been used in an attempt to reduce tone in patients with CP. Side effects include somnolence. Oral tizanidine has been used to relieve spasticity with improved results. Patients with movement disorders such as dystonia or choreiform movements have been treated by various medications, with varying success.

Intrathecal Baclofen.

The inability to adequately reduce tone with oral baclofen because of the side effects with large dosages led to trials of intrathecal baclofen. When baclofen is injected into the intrathecal space, small doses can effectively reach the target neural tissue of the spinal cord and produce a greater reduction in tone. Baclofen, a γ-aminobutyric acid agonist, acts at the spinal cord level to impede release of the excitatory neurotransmitters that cause spasticity. After proof of efficacy with test injections of the drug, implantable pumps filled with baclofen are surgically inserted into the anterior abdominal wall and the dose of medication is titrated. Continuous infusion of the medication is delivered by the pump, which requires refilling approximately every 3 months. Studies have shown decreased upper and lower extremity tone and improvements in range of motion with continuous intrathecal baclofen infusion.

Complications occur in approximately 20% of patients. The complications are usually related to catheter dislodgment or fracture, but serious infection occurs in 5%. Overdose of intrathecal baclofen is very rare but serious and can produce somnolence and hypotonia, which leads to respiratory depression requiring mechanical ventilation.

Currently, no strict indications for the use of intrathecal baclofen have been determined, but most pumps are implanted in patients with severe spasticity that interferes with positioning and function of the upper and lower extremities. Pumps are currently being implanted in both nonambulatory and ambulatory patients. Spasticity, as graded by the Ashworth scale, has been shown to decrease 1 year following implantation of intrathecal baclofen pumps regardless of preoperative severity. Caregiver satisfaction in the nonambulatory group is high, with parents noting easier positioning, transfers, and personal care, as well as improved pain control. Pain relief and improved sleep are typically seen within 6 months of pump implantation. Many of these patients still need orthopaedic surgery for hip subluxation or contractures. Although progressive hip subluxation can occur, significant deterioration in migration percentage is uncommon. Close monitoring for scoliosis is also warranted because rapid progression of deformity has been described in some studies but refuted in others. Two studies comparing progression of scoliosis in patients who had intrathecal baclofen pumps implanted versus a well-matched control group of children with severe CP who were not treated with baclofen found no statistical difference in progression of scoliosis in the treated group. Both groups of patients experienced curve progression, which supports the premise that it is the severity of neurologic involvement and functional disability that predisposes the patient to scoliosis and not the medical management of spasticity ( Fig 35-15 ).

Eleven-year-old boy with 110-degree scoliosis, spastic quadriparesis, and an intrathecal baclofen pump. The patient had also undergone left hip reconstruction for subluxation.

General Considerations

When evaluating a child with CP for surgical intervention, a few general principles must be kept in mind. Speaking frankly with the family about the goals of surgery and the expected postoperative course is important. Frequently, the family will have unrealistic expectations and believe that surgery will “cure” the child and help him walk normally. It must be explained that CP is a brain disease and that orthopaedic surgery for CP will make the child walk differently, hopefully better, but will not make the child walk “normally.”

Timing of Surgery

Experts recommend combining multiple tendon surgeries and osteotomies into a single surgical event (SEMLS). Rang advised avoiding “birthday surgery,” or hospitalizing the child every year for yet another soft tissue surgery or osteotomy. Aside from avoiding repeated hospitalizations as one joint is released at a time, contractures present at one joint affect the position and movement of the rest of the extremity, so correcting all concomitant contractures simultaneously during one surgery is important to avoid recurrence or overcorrection of deformity. Because gait changes and matures until approximately 7 years of age, when feasible it is wise to avoid surgery until this time. At this age, multiple levels can be treated simultaneously to optimize the ambulatory skills of the patient, followed by a single aggressive course of physical therapy to maximize the benefits from surgery. This surgical approach has been termed “single-event multilevel surgery.” The literature supports the fact that well-executed surgery addressing all contractures can result in stabilization of patients’ GMFCS level, thereby preserving their function or in many cases resulting in a lower GMFCS level at follow-up (i.e., improving their gait closer toward normal).

In some patients surgery cannot be delayed until the age of 7 years. Young children with hip subluxation should undergo surgery when the problem is first recognized to improve coverage of the hip. At younger ages hip surgery is often less extensive and frequently consists of only soft tissue release, whereas in older children, additional osteotomies of the femur and pelvis are required.

For other children who are nearly ambulatory but whose progress has been halted by contractures in the lower extremities, earlier surgery may allow them additional range of motion to make walking less cumbersome. Adductor release for scissoring and gastrocsoleus surgery with or without hamstring lengthening in these younger children may be indicated. Parents should be forewarned about recurrent contractures requiring additional surgery in the future.

Studies have found that the natural history of gait disturbances in patients with CP is gradual deterioration beginning at 12 years of age. Decreased ankle, knee, and hip sagittal-plane motion, decreased cadence, and slower walking speed have been documented over time in patients who did not undergo intervening surgical procedures. It is common for crouch and contractures to worsen in the early teen years. Surgery may offer some benefit, but realistic goals must be discussed.

Anesthetic Concerns

Patients with CP experience latex allergy with increased frequency, and for those with suspected latex allergy, a latex-free surgical environment is necessary. Allergy testing is available for latex products. At-risk children are those who have undergone multiple previous operations or who have indwelling latex devices such as gastrostomy tubes or ventriculoperitoneal shunts.

Antiseizure medications may produce alterations in clotting. Increases in bleeding times and decreased platelet counts are known side effects of some antiepileptic agents. Preoperative evaluation of clotting parameters is recommended for major operations such as hip reconstruction or spinal fusion. A routine clotting profile consisting of a prothrombin time and partial thromboplastin time will miss bleeding time abnormalities. Platelet function assays can help identify these bleeding disorders.

Postoperative pain control is difficult in patients undergoing hip reconstruction, spinal surgery, and multilevel tendon surgery. We use continuous epidural infusions at our hospital and find that if the pain is well controlled, problems with muscle spasm are lessened. Oral medication often needs to include both pain control medications and muscle relaxants such as diazepam.

Postoperative Management

Weakness is a frequent short-term sequela of lower extremity surgery in those with CP. Persistent diminished muscle strength has been documented 1 year following surgery. Plans should be made for frequent postoperative physical therapy beginning shortly after surgery with attention given to muscle strengthening. If at all possible, the patient should be kept ambulatory or weight bearing after the operation. Prolonged time in a wheelchair adds to the overall weakness of the limbs. Patients with osteotomies are the exception to this rule. In these children, weakness may be even more of a problem once casts are discontinued. A few centers have exhibited new enthusiasm for minimally invasive surgery to correct contractures in children with CP. Percutaneous releases with safe intramedullary fixation of rotational osteotomies may result in less loss of strength in the postoperative period. Because it is imperative, however, that safety be ensured, percutaneous releases are not applicable in all situations.

A trend toward less immobilization after soft tissue surgery has emerged. In our practice, short-leg casts are still used after Achilles tendon lengthening, but knee immobilizers rather than casts are becoming more frequently used after hamstring lengthening, and removable abduction bars are preferable to Petrie casts after adductor release. The goal of soft tissue surgery is greater range of motion, so avoiding overimmobilization and secondary stiffness is logical.

Differential Diagnosis of Equinus

Not all children who walk on their toes have CP. Idiopathic toe-walking caused by a congenitally short Achilles tendon is a condition in which otherwise neurologically normal children walk on their toes. Unlike children with CP, these children are not developmentally delayed and walk on time, have normal knee motion during gait, have no neurologic signs of spasticity, and have normal reflexes. Gait studies looking at kinematic and EMG data have attempted to differentiate the two diagnoses on these grounds, but a thorough history and physical examination by an experienced clinician are really all that is needed in most children.

Another condition that produces toe-walking is muscular dystrophy. The condition should be suspected in any young boy who walks on his toes and has a normal birth history. A delay in the age of walking is frequently seen in patients with Duchenne muscular dystrophy, so the developmental history may not differentiate it from CP. Testing for the Gower sign by having the child rise up rapidly from a sitting position on the floor will accentuate the presence of proximal muscle weakness. If the test for the Gower sign is suspicious, laboratory evaluation for serum muscle enzymes (creatine phosphokinase) is indicated.

Clinical Features

Clinical examination of a child with equinus associated with CP reveals an inability to fully dorsiflex the ankle. If the ankle can be passively dorsiflexed with the knee bent to 90 degrees but cannot be dorsiflexed with the knee extended, it is believed that the gastrocnemius is tight but the soleus is not contracted (Silfverskiöld test; Fig. 35-17 ). Some have used this test to determine which type of surgical lengthening to perform. If the ankle has sufficient dorsiflexion on passive range-of-motion examination, the equinus is termed dynamic , and surgical treatment is not usually needed.

A and B, Mild ankle equinus in a 17-year-old girl with spastic diplegia. The ankle cannot be dorsiflexed past neutral with the knee extended.

Equinus interferes with forward progression of the tibia over the foot during stance phase and therefore shortens stride length. Because the ankle is already plantar-flexed at terminal stance, little push-off power is generated and the gait is less efficient. If the tibialis anterior is unable to lift the foot to neutral during swing phase, footdrop results, with possible problems in clearing the foot in swing phase and tripping. Lack of normal swing-phase ankle dorsiflexion leads to forefoot contact on foot strike at the beginning of stance phase. Knee disturbances also result from ankle equinus. Genu recurvatum is seen when the tibia is unable to move forward over the plantigrade foot because of tightness in the gastrocsoleus, so the knee thrusts backward during midstance ( Fig. 35-18 ). The body and femur continue to move forward over the stationary tibia, and an extensor moment is produced at the knee. This aligns the ground reaction force anterior to the knee, thereby reducing demands on the quadriceps and improving the stability of the knee. Likewise, compensatory knee flexion can be seen during stance phase in patients who walk on their toes.

Hyperextension of the knees to compensate for fixed equinus deformity of the ankle. One method of aligning the trunk and bringing the center of gravity over the feet is by knee hyperextension.

Over time, ankle equinus leads to valgus positioning of the hindfoot and stretching of the plantar arch (midfoot break) . Although the foot may appear plantigrade, the talar head is very prominent in the longitudinal arch and the calcaneus is actually in equinus. Pain and calluses form over the head of the talus ( Fig. 35-19 ). Hallux valgus can develop in response to the valgus positioning of the foot.

Equinovalgus of the right foot of an adolescent girl with spastic quadriplegia. The talar head is very prominent, and a painful callus has developed.

Treatment

Surgical treatment of equinus is selective lengthening of the Achilles tendon or gastrocnemius fascial recession. Advocates of gastrocnemius recession state that this operation preserves or even increases push-off power more than Achilles tendon lengthening does by not involving the soleus muscle or tendon. Immobilization is minimized after gastrocnemius recession, whereas casting is necessary after Achilles tendon lengthening. The risk for overcorrection into a calcaneus gait is negligible. Perry and Hoffer believe that gastrocnemius recession should be performed when a Silfverskiöld test performed under anesthesia is positive and dynamic EMG shows more abnormality of the gastrocnemius than of the soleus during gait.

Those who prefer lengthening of the Achilles tendon recognize a greater rate of recurrence of equinus after gastrocnemius recession, which has been as high as 48% in some studies. They also state that Achilles tendon lengthening may be done percutaneously, unlike gastrocnemius recession. In one comparative gait analysis study, no significant difference was found between patients who had undergone gastrocnemius recession and those who had undergone lengthening of the Achilles tendon. In another study, if equinus was an isolated gait abnormality (i.e., the knee did not lack extension during stance phase), isolated percutaneous Achilles tendon lengthening to a plantigrade position was very successful in improving ankle kinematics and kinetics without resulting in knee abnormalities or a calcaneus gait. A higher risk for calcaneal overcorrection has been noted after percutaneous Achilles tendon lengthening than after Baker gastrocsoleus aponeurotic lengthening. At 9 years of follow-up, a calcaneus gait had developed in only 10% of 44 spastic diplegics treated by gastrocsoleus aponeurotic recession. In contrast, Dietz and colleagues studied 79 children who had undergone Achilles tendon lengthening. At an average of 7 years’ follow-up, they found that a calcaneus gait did not develop in the hemiplegic patients whereas 41% of the diplegic and 50% of the quadriplegic patients did walk with excessive knee flexion and ankle dorsiflexion. Vuillermin and colleagues side with Dietz. They studied the prevalence of crouch in two periods, the first when Achilles tendon lengthening was commonly performed and the second when gastrocnemius recession replaced Achilles tendon lengthening. They found a marked decrease in the development of crouch gait in the second cohort of patients. Confounding factors were the use of gait analysis for preoperative planning and adoption of the SEMLS approach in the group treated with gastrocnemius recession. Controversy on this subject continues, with surgeons split between those who believe that the adolescent crouch gait is the result of weakness from previous Achilles tendon–lengthening surgery and those who believe it to be due to growth and untreated hamstring spasticity. Increasing knee flexion (crouch gait) has been documented to develop over time in patients who have not previously undergone orthopaedic procedures. Meta-analysis of the existing literature failed to show a scientific preference for one procedure over another, although it was agreed that recurrence was most likely in younger patients and that overcorrection (or talipes calcaneus) was unlikely in hemiplegics when compared with diplegics.

Gastrocnemius Recession.

Gastrocnemius recession may be done with the techniques of Strayer, Baker, or Vulpius. In a Vulpius procedure, the aponeurosis of the gastrocsoleus is divided in chevron fashion and the midline fibrous septum of the soleus is transected, but the soleus muscle fibers are not disturbed ( Fig. 35-20 ). The fascia slides on the underlying soleus. The cut in the gastrocnemius is transverse and more proximal in the Strayer procedure, which does not lengthen the soleus whatsoever ( Fig. 35-21 ). In the Baker technique, the gastrocsoleus aponeurosis is cut in tongue-in-groove fashion and dissected free from the underlying soleus muscle. The fascia is allowed to slide on the underlying muscle, thereby increasing the overall length of the muscle, and the four corners of the aponeurosis are sutured in the lengthened position ( Fig. 35-22 ). These surgeries can be divided into groups based on the zone in which lengthening occurs. Zone 1 is from the gastrocnemius origin to the distal end of the medial gastrocnemius. Zone 2 is from the distal end of zone 1 to the end of the soleus muscle belly. Zone 3 represents the Achilles tendon. The Strayer lengthening occurs in zone 1 and the Vulpius and Baker lengthenings in zone 2.

Lengthening of the gastrocnemius by the Vulpius technique.

Distal recession of the gastrocnemius, Strayer technique.

Tongue-in-groove lengthening of the gastrocnemius aponeurosis in its middle third, the Baker technique.

Achilles Tendon Lengthening.

Achilles tendon lengthening may be performed by open or percutaneous techniques. In the open technique, a longitudinal incision is made just lateral to the Achilles tendon. The tendon is lengthened in Z -fashion and repaired with stout nonabsorbable suture ( Fig. 35-23 ). The tendon must be repaired with sufficient tension to avoid postoperative calcaneus gait. The ankle is then immobilized in a short-leg cast for 6 weeks. An open sliding lengthening of the Achilles tendon may also be performed; this procedure does not require suturing of the tendon.

Z -lengthening of the Achilles tendon.

The percutaneous technique may require two or three cuts in the tendon. We prefer the two-cut technique described by White. Just proximal to the calcaneal insertion, a little more than half the tendon is divided medially by inserting a scalpel longitudinally into the center of the tendon, turning the blade out medially, and ballotting the tendon down onto the blade. The lateral half of the tendon is then divided by using a similar maneuver more proximally. The heel is inverted into varus, and the ankle is dorsiflexed to the neutral position (see Plate 35-1 ). The tendon can be heard lengthening as the ankle is gently manipulated. It is not unusual to hear a pop as the tendon slides and lengthens, although we prefer more gradual and gentle lengthening to prevent overlengthening of the tendon.

It is important to check the integrity of the tendon after percutaneous lengthening. The calf is squeezed while the surgeon observes the ankle. If the ankle plantar-flexes when the calf is squeezed, the tendon can be presumed to be intact within the tendon sheath, and no sutures are needed. Steri-Strips are applied to the two stab wounds, followed by a cast. If no plantar flexion is noted, the tendon may have been completely separated by the lengthening procedure and open suture repair will be needed to reestablish continuity of the tendon. This is a very rare occurrence; usually, the percutaneous technique proceeds without complication. Again, a short-leg cast is applied and worn for 6 weeks, with weight bearing encouraged.

Hoke described a three-cut technique for percutaneous heel cord lengthening. After making one lateral and two medial cuts in the tendon, the ankle is dorsiflexed as the tendon slides on itself and lengthens. Postoperative care is identical to that after the two-cut technique.

Postoperative Care

After either heel cord lengthening or gastrocnemius recession, patients may tend to flex their knees after surgery. This is caused by muscle spasm in the hamstrings and gastrocnemius (which crosses the knee joint). We find that use of a knee immobilizer for the first few days to weeks can help reduce the amount of muscle spasm and prevent the development of a postoperative knee flexion contracture.

After the short-leg cast is removed, an AFO is prescribed postoperatively for most patients. Children who had footdrop in swing phase preoperatively will continue to have footdrop after heel cord lengthening, and patients and families should be forewarned of the continued need for orthoses after surgery to support the foot and improve toe clearance in swing phase. Many younger children benefit from the stability of an AFO postoperatively, whereas in most older children and teens, the goal of heel cord lengthening is to rid them of braces on their feet, so they are reluctant to use an AFO after surgery.

Complications

Complications are rare after heel cord lengthening or gastrocnemius recession. Recurrent equinus is the greatest risk; it occurs in approximately 15% to 35% and correlates strongly with the age of the patient at surgery. Children who undergo heel cord lengthening at 4 years or younger are particularly at risk for recurrence. It is impossible to know whether it is the greater amount of growth left in younger children after heel cord surgery that leads to the high rate of recurrence or whether it is preselection of young children with the greatest tone to undergo surgery at an early age after failure of nonoperative treatment that leads to recurrent contracture. In a long-term study by Rattey and colleagues, 26% of patients who had undergone Z -lengthening of the Achilles tendon required repeated lengthening for recurrent equinus at an average follow-up of 10 years. The risk for recurrent contracture was greater in patients with hemiplegia (41%) than in those with diplegia (18%). Grant and associates found that an inability to actively dorsiflex the foot preoperatively increased the risk for recurrence. Repeated surgery is possible and common. With repeated lengthening, the Achilles tendon becomes scarred and adherent, so repeated lengthening must usually be done with the open technique. Patients who have undergone gastrocnemius recession have a higher rate of recurrence than do those treated by tendon lengthening.

Calcaneus deformity, defined as excessive dorsiflexion of the ankle during stance phase, may result for two reasons. First, the tendon may simply be overlengthened or may have lost continuity. When performing lengthening, the surgeon must be careful to bring the foot just to neutral. Overstretching of the tendon can lead to excessive length and a calcaneus gait, with poor push-off and a tendency for the development of progressive crouch of the knees.

The second reason that a calcaneus gait may occur after lengthening of the gastrocnemius or Achilles tendon is an unrecognized flexion contracture of the knee. If the knee remains crouched and the Achilles tendon becomes longer, the ankle sags into excessive dorsiflexion during stance phase and gait loses efficiency. A meticulous physical examination before surgery in which tone and range of motion of the knee are assessed is critical. If the popliteal angle is increased and the child exhibits excessive knee flexion during stance phase or an inability to straighten the leg as the heel makes contact with the ground, postoperative calcaneus will most likely develop if the knee is not surgically treated at the same time ( Fig. 35-24 ).

A, Child with a crouched knee and toe-toe gait. The ankle is actually in neutral position, but the child walks on the toes to compensate for the flexed knee. B, After inappropriate Achilles tendon lengthening, flexion at the knee remains unchanged and the ankle is now excessively dorsiflexed, which results in a calcaneus gait.

The incidence of calcaneus gait as a complication of heel cord lengthening averages approximately 5%. Segal and colleagues found calcaneus on kinematic graphs in 30% of patients. Although this figure seems high in comparison to previous studies, it does highlight the need to treat all levels of deformity during surgery and avoid overlengthening at all cost. Mathematic equations have been derived to calculate the exact amount of lengthening required for a particular patient, but most find the use of these equations somewhat cumbersome.

Although most patients in whom a calcaneus gait develops have previously undergone equinus surgery, in a group of patients calcaneus will develop de novo. An adolescent diplegic patient is at greatest risk for calcaneus. Patients who gain weight (as in the adolescent growth spurt) and who have plantar flexor weakness were found on serial gait analysis to be most prone to the development of a calcaneus gait.

Preferred Procedure

In our clinical practice we prefer percutaneous Achilles tendon lengthening as the surgical treatment of fixed equinus contractures. Postoperative cast immobilization is used for 6 weeks, and AFOs are prescribed on an individual basis. We agree with the philosophy that “a little equinus is better than any calcaneus” and avoid overlengthening. Most important, we thoroughly evaluate the child for other joint involvement, and we typically perform surgery on multiple levels of the lower extremity at the same time (SEMLS). We do see some recurrent contractures with growth, and we reoperate when these contractures interfere with gait. Gastrocnemius fascial lengthening is performed in our center for milder dynamic equinus.

Etiology and Clinical Features

Equinovarus deformity of the foot results from muscle imbalance in which the invertors of the foot, specifically the posterior and anterior tibialis muscles, overpower the evertors (the peroneals). The gastrocnemius contributes equinus to the deformity. Patients walk on the lateral border of the inverted foot, and painful calluses may develop laterally over the fifth metatarsal. The deformity is most prevalent in patients with spastic hemiplegia ( Fig. 35-25 ).

Equinovarus deformity in a 5-year-old boy with right spastic hemiplegia.

Gait analysis shows an internal foot progression angle as a result of inversion of the foot. Ankle equinus is present, and footdrop may be evident during swing phase. Pedobarographic assessment demonstrates decreased hindfoot pressure (secondary to equinus) and increased pressure beneath the lateral border of the forefoot and midfoot, specifically the fifth metatarsal ( Fig. 35-26 ).

Typical pedobarograph of a patient with equinovarus showing decreased hindfoot pressure and excessive pressure along the lateral border of the foot.

Treatment

Nonoperative treatment is usually unsuccessful. A supple deformity can be braced, but if the muscles are very spastic, the orthoses can exacerbate the blisters or calluses over the lateral border of the foot. Botulinum toxin injections have been tried, but no published studies have shown long-term efficacy for equinovarus.

Surgery is indicated to improve foot contact and relieve pain and skin changes. If the foot can be passively corrected with manipulation in the clinic to a neutral position, tendon surgery can be performed. Lengthening procedures and split transfers have been described, and more detail will be provided about these procedures. If the deformity is stiff and the foot cannot be manipulated into a plantigrade position, bone surgery will be necessary. In patients with spastic diplegia and quadriplegia, equinovalgus tends to develop later during childhood and adolescence. Young patients, specifically those younger than 8 years, who are diplegic or quadriplegic are most likely to have unsatisfactory long-term results after surgery for equinovarus. For this reason, nonoperative treatment options should be exhausted in these patients.

Surgical decision making for the soft tissue correction of spastic equinovarus focuses on distinguishing whether it is the anterior tibialis or the posterior tibialis that is the cause of the deformity. The evaluation should start in the clinic with a careful physical examination. The confusion test is helpful in this setting. The patient is asked to flex the hip against resistance while seated, and the ankle is inspected. In most children with CP the anterior tibialis fires while the hip is flexed; this is considered a positive confusion test ( Fig. 35-27 ). However, the reflex has been seen in some children without CP. The examiner should look for the position of the foot as the anterior tibialis fires. If supination of the forefoot is seen, the anterior tibialis is most likely contributing to the equinovarus deformity. If the foot dorsiflexes without supinating, the equinovarus is less likely to improve with surgery on the anterior tibialis.

Confusion test. Anterior tibialis function can be tested by having the patient flex the hip against resistance. The anterior tibialis normally dorsiflexes the ankle.

Next, the examiner feels for spasticity in the posterior tibialis muscle. Passive manipulation of the hindfoot into valgus while feeling the posterior tibialis tendon can help the physician appreciate tightness in the posterior tibialis. The examiner should look at where the varus appears to be located. Hindfoot varus is most probably caused by overpull of the posterior tibialis, whereas forefoot supination is more commonly caused by the anterior tibialis. The examiner should observe for tension in the anterior tibialis and posterior tibialis during gait; the spastic tendon may be visibly taut as the patient walks.

Gait analysis with dynamic EMG can help determine which muscle is acting inappropriately. Surface electrodes suffice for measuring anterior tibialis activity, but the posterior tibialis is deep, and a fine wire–needle electrode is necessary to measure its signal. The anterior tibialis should be quiet in midstance. Signal from the anterior tibialis throughout stance phase is indicative of overactivity ( Fig. 35-28 ). The posterior tibialis serves to stabilize the foot during stance phase. Early-onset signal and prolongation of the activity during swing phase indicate an abnormality in control of the posterior tibialis. Nearly one third of patients with equinovarus will show inappropriate activity of both the anterior and posterior tibialis muscles during gait.

Electromyographic (EMG) tracing showing activity of the tibialis anterior muscle during gait in a child with an equinovarus foot secondary to cerebral palsy. The tibialis anterior normally fires at foot contact to gradually lower the foot to the ground and again during swing phase to prevent footdrop. This EMG tracing shows activity throughout stance phase, occupying from 0% to 70% of this gait cycle, and lack of activity during swing phase.

Some patients with equinovarus feet have preexisting footdrop during swing phase. When this is seen, either in the clinic or during gait analysis, further weakening by anterior tibialis surgery may lead to significant footdrop postoperatively, and an AFO will be needed.

Posterior Tibialis Tendon Lengthening.

Surgical options for equinovarus deformity start with lengthening of the posterior tibialis tendon, usually done in conjunction with an Achilles tendon–lengthening procedure in young patients with mild varus in addition to equinus. The tendon may be Z -lengthened distally, or preferably intramuscular lengthening can be performed in the distal third of the calf ( Fig. 35-29 ). The patient is placed in a short-leg cast postoperatively for approximately 6 weeks. Complications consist of recurrence of the deformity and the development of postoperative valgus. The posterior tibialis is weakened by the lengthening, but rebalancing of forces around the foot does not occur to the same extent as with a tendon transfer. Yet in many milder cases this is sufficient to obtain a plantigrade foot.

A and B, Intramuscular lengthening of the posterior tibial tendon can be performed in the distal third of the leg in patients with mild varus deformity.

Split Posterior Tibialis Tendon Transfer.

Overcorrection with complete tendon transfers led to the popularization of split tendon transfers in children with CP, one of the most common being that of Kaufer and then Green. In this procedure the posterior half of the posterior tibialis tendon is detached from its insertion, split proximally to a level just proximal to the ankle, rerouted posterior to the tibia and fibula, and then woven into the tendon of the peroneus brevis ( Fig. 35-30 ). The remaining posterior tibialis tendon attached to the navicular is then balanced by the lateral half of the transferred tendon, which acts as an evertor. Four incisions were used in the original description of the operation. Thompson and colleagues described the use of a Cincinnati incision for this transfer. Usually, an Achilles tendon–lengthening procedure is also required to treat the equinus.

Surgical technique of split posterior tendon transfer. A, Line of the medial incision. Two smaller incisions may also be used. B and C, The posterior tibial tendon is split longitudinally and proximally to its musculotendinous junction; the dorsal portion of the tendon is left intact and attached to the navicular. The retinaculum of the ankle is not divided. Note that the neurovascular bundle and long toe flexors are gently retracted posteriorly. D, Lateral view of the foot and ankle showing the line of the lateral skin incision. Here also, two smaller incisions may be used. E, The peroneus brevis and longus tendons are identified and the tendon sheath is opened to expose the peroneal tendons. F and G, The split half of the posterior tibial tendon is transferred to the lateral side of the foot by passing anterior to the neurovascular bundle, long toe flexor tendons, and flexor digitorum longus. It is inferior to the lateral malleolus and deep to the peroneus brevis tendon. H and I, The split posterior tibial tendon is brought out into the peroneus brevis tendon sheath, the tension on it is adjusted, and it is sutured to the tendon as far distally as possible. It is best to weave it through the peroneus brevis tendon. J, Posterior view of the ankle and hindfoot showing the direction of tendon transfer. It is oblique from its musculotendinous junction above toward the tip of the lateral malleolus distally and laterally. Continuous contraction of the spastic posterior tibial tendon provides mechanical stability and control of the hindfoot in neutral position or 5 degrees of valgus inclination.

The prerequisite for a successful split posterior tibialis tendon transfer on gait analysis is swing-phase activity seen on dynamic EMG. Yet several authors have reported successful series of split posterior tibialis tendon transfers in which preoperative or postoperative EMG was not performed.

Usually, patients can walk without orthoses after this tendon transfer. As with any tendon transfer, the deformity must be passively correctable preoperatively for postoperative correction to be expected. Recurrent varus may be a complication after split posterior tibialis tendon transfer and generally results from inappropriate selection of patients with a deformity that is too rigid. Green stated that overcorrection into valgus has not occurred in their patients, but we have occasionally seen overcorrection in older patients in our outpatient clinics several years after tendon transfer.

Transferring the anterior half of the posterior tibialis tendon anteriorly through the interosseous membrane to the dorsum of the foot is a variation of the classic posterior tendon transfer. It is believed that this split transfer might assist dorsiflexion with less calcaneus deformity than is the case with transfer of the entire tendon because the posterior half of the tendon remains intact. Although early studies have shown promising results, this procedure has yet to be adopted universally.

Split Anterior Tibialis Tendon Transfer.

In split anterior tibialis tendon transfer (see Plate 35-2 ), the lateral half of the anterior tibialis is detached from the base of the first metatarsal and split up proximal to the level of the ankle. The tendon is then passed beneath the extensor retinaculum, inserted through a bone tunnel into the cuboid bone, and tied over a button on the sole of the foot under tension with the foot positioned in 5 to 10 degrees of dorsiflexion. Again, the procedure is usually combined with an Achilles tendon–lengthening procedure. When done in conjunction with posterior tibialis lengthening, it is known as the Rancho procedure ( Fig. 35-31 ).

A, Preoperative photo of a boy with right hemiplegia and equinovarus foot deformity that is passively correctable. B and C, Clinical appearance of a foot after a Rancho procedure. The patient is able to walk without the use of an orthosis.

The clinical prerequisite for the procedure is overactivity of the anterior tibialis causing supination rather than dorsiflexion of the foot during the confusion test. Gait analysis will show inappropriate signal on EMG during stance phase because of inappropriate activity of the anterior tibialis. Patients with profoundly weak anterior tibialis tendons and notable footdrop should not undergo split anterior tibialis tendon transfer.

The published results of split anterior tibialis tendon transfer and split posterior tibialis tendon transfer are similar and encouraging, with nearly all patients doing well, brace free, and without overcorrection. Pedobarographs can document the normalization of plantar pressure after the Rancho procedure ( Fig. 35-32 ). As in posterior tibialis tendon surgery, if the deformity is not flexible preoperatively, the split anterior tibialis tendon transfer will not be successful in correcting the deformity.

A, Pedobarograph of a child with equinovarus showing lack of heel contact and excessive pressure over the fifth metatarsal. B, Postoperative pedobarograph showing evidence of heel strike and a more normal pressure distribution with unweighting of the lateral border of the foot.

Incidence and Etiology

Valgus deformity of the foot occurs in up to 25% of patients with CP and is most common in older diplegic and quadriplegic patients. Equinovalgus deformity develops in up to 42% of patients with spastic diplegia and up to 68% with spastic quadriplegia. It is usually present bilaterally. Complaints consist of abnormal shoe wear and pain from calluses and blisters in the area of the talar head, which becomes very prominent in the arch of the foot ( Fig. 35-33 ). Parents note that the ankles appear to roll in as the hindfoot valgus increases. With time, hallux valgus develops in response to the everted foot position, which may be painful ( Fig. 35-34 ).

Patient with spastic diplegia and pes valgus. The prominent talar head was painful.

Clinical photo of the feet of a 17-year-old boy with spastic diplegia. Note the valgus of the left heel, the severe hallux valgus, and abnormalities in plantar pressure with excessive medial midfoot pressure.

Pes valgus can be caused by spastic peroneal muscles, weakness of the posterior tibialis, or a tight gastrocsoleus, in any combination. EMG studies have shown that the peroneals may be continuous or may be phasic but inappropriate in pes valgus. The same studies found some children in whom the posterior tibialis muscle was silent during stance phase, further enabling the peroneals to pull the foot out into abduction and valgus.

Clinical Features and Radiographic Findings

Physical examination should assess for coexisting equinus or calcaneus deformities. Frequently, the gastrocsoleus complex will not appear tight on initial examination. The examiner must be certain to maintain the hindfoot in varus and then passively dorsiflex the ankle. Valgus will mask equinus unless this is done. The foot may appear flat in the standing position, yet the hindfoot may be positioned in significant equinus, the talar head plantar-flexed, and the midfoot overly mobile to maintain the plantigrade position. Midfoot break is the term used to describe plantar flexion of the talus and calcaneus, a collapsed longitudinal arch, and dorsiflexion and pronation of the forefoot.

Radiographs should be obtained in the standing position. Lateral radiographs are most helpful. The position of the hindfoot can be assessed for equinus or calcaneus deformity, and plantar flexion of the talus can be appreciated ( Fig. 35-35 ). The navicular moves laterally and is seen as uncovering of the talar head on an anteroposterior radiograph. Standing radiographs of the ankle should also be obtained. It is not uncommon to find ankle valgus on an anteroposterior radiograph of the ankle coexisting with hindfoot valgus. In such cases the fibular physis will migrate proximally and lie superior to its normal position opposite the joint line of the distal end of the tibia ( Fig. 35-36 ).

Lateral radiograph of a patient with pes valgus. The talus is excessively plantar-flexed.

A, Radiographic appearance of an adolescent boy with a hindfoot valgus deformity and pain in the medial arch over the prominent talar head. B, Standing anteroposterior radiograph of the ankle demonstrating coexistent ankle valgus that is due to the proximal location of the distal part of the fibula. The fibular physis lies opposite the distal tibial physis rather than at the joint line. C, Lateral radiograph obtained 5 months after subtalar arthrodesis for correction of the hindfoot valgus. D, Medial malleolar epiphysiodesis with a screw was performed concomitantly to address the valgus of the ankle.

Treatment

Treatment is controversial. Conservative treatment should be vigorously pursued because shoe inserts and orthosis modifications may be adequate to relieve the pain in some patients, and therefore surgery can be avoided. As long as the foot is painless, even orthotic support of the valgus foot may not be necessary, with some children doing well in athletic shoes, and in fact symptoms may develop only after rigid orthoses are placed on their feet.

After conservative measures have been exhausted, surgical treatment may be considered. If the valgus is thought to be secondary to contracture of the Achilles tendon, heel cord lengthening may offer improvement in a young child. Clinically, these patients will have a normal longitudinal arch without medial prominence when the ankle is held in plantar flexion, and valgus becomes apparent as the foot is dorsiflexed to neutral. With further dorsiflexion, midfoot break occurs. Aggressive realignment of the valgus foot may not be required in this group of young patients. Once the equinus is corrected surgically, postoperative support with orthoses is needed to control the hindfoot. This treatment does not work in older children because the valgus becomes fixed with age.

Tendon lengthening of the peroneals has been studied as treatment of pes valgus but has not been found to be successful in obtaining and maintaining a plantigrade foot, with some feet being undercorrected and others drifting into varus over time. Not only can tenotomy of the peroneals lead to a varus deformity, but a dorsal bunion of the first metatarsophalangeal joint can also occur because of the absence of plantar flexion of the first metatarsal head by the peroneus longus. Transfer of the peroneus brevis to the posterior tibialis has also been performed, with resultant overcorrection. We do not advocate these procedures.

In valgus deformity, bone surgery is the only predictable alternative for full and lasting correction. Surgical options are (1) the Grice extraarticular arthrodesis, (2) lateral column lengthening of the calcaneal neck, (3) calcaneal osteotomy, and (4) triple arthrodesis.

Grice Extraarticular Arthrodesis (see Plate 35-3 ).

The Grice procedure is defined as placement of a structural graft, such as from the fibula or tricortical iliac crest, into a shallow trough in the sinus tarsi laterally to elevate the plantar-flexed talus and correct the valgus of the subtalar joint. The articular surfaces of the subtalar joint are not resected. The advantage of this operation is that it does not interfere with growth of the tarsal bones because it is not a formal arthrodesis. Growth disturbance results from arthrodesis because the articular surfaces of the tarsals are growth centers. Grice arthrodesis is often combined with lengthening of the peroneal tendons or the Achilles tendon, or both. Grice reported satisfactory results in 79% of patients ( Fig. 35-37 ).

A, Weight-bearing lateral radiograph of a patient with pes equinovalgus. B, Radiograph obtained 8 months after Achilles tendon lengthening and Grice extraarticular arthrodesis. C, Lateral radiograph obtained 3 years, 6 months postoperatively.

Although excellent outcomes after Grice arthrodesis have been reported, the results of the procedure are unpredictable. The graft is not inherently stable in the Grice procedure, and loss of correction because of dislodgment of the graft is well documented. Bleck pointed out that the orientation of the graft must be vertical for the forces through it to be compression rather than rotation. If the graft is aligned obliquely, dislodgment and fracture of the graft are likely. Failure of the graft to unite and undercorrection of valgus at the time of surgery also occur. Unsatisfactory outcomes have been reported in 30% to 70% of patients as a result of uncorrected contracture of the Achilles tendon, overcorrection into varus, ankle valgus, and graft nonunion. Although maintenance of good results and high patient satisfaction at 20-year follow-up in patients with CP who have undergone Grice arthrodesis have been reported, a tendency toward ankle valgus was noted. However, it may have been present at the time of surgery or may have developed as a result of harvesting the ipsilateral fibula as graft material. At an average 22.6-year follow-up, Leidinger and co-workers found excellent or good results in 78% of 51 feet but warned that overcorrection into varus is poorly tolerated.

Because of lack of predictable success with the Grice arthrodesis, several modifications of the original procedure were proposed. First, the site from which the bone graft is taken was changed from the fibula to the iliac crest as symptomatic fibular nonunion and progressive ankle valgus were recognized. Some groups began inserting bone plugs or dowels across the sinus tarsi to stabilize the subtalar joint. Without internal fixation, graft fracture and loss of correction are still possible risks because the bone dowel is placed obliquely to the axis of weight bearing.

Internal fixation using an iliac crest graft helps maintain the position of the subtalar joint. Local bone graft from the calcaneus has also been used. Less loss of correction occurs with the addition of internal fixation (70% to 95% satisfactory short-term outcomes), and problems with screw breakage are very rare. Cancellous bone grafting accelerates bony fusion. Again, coexisting contractures must be corrected to improve results. Alman and associates used K-wires rather than a screw to maintain subtalar joint position while the graft healed, with good results in 48 of 53 feet. Bioabsorbable screws were used in an extraarticular subtalar arthrodesis with comparable results in one small series, but this technique must be considered investigational at present.

Lateral Column (Calcaneal) Lengthening.

Lateral column lengthening was first described by Evans and has enjoyed recent popularity after a series by Mosca. Mosca performed the procedure to correct pes valgus in 31 feet, including 26 procedures done for valgus secondary to CP and myelomeningocele. Correction is achieved by lengthening the neck of the calcaneus to tighten the plantar fascia and reduce the lapsed talonavicular joint. The procedure is summarized in Plate 35-4 and the following text.

The calcaneus is approached through an oblique incision laterally that follows the skin lines. The peroneal sheath is incised and the tendons are retracted. The lateral dorsal cutaneous nerve is identified and protected. The extensor digitorum brevis is reflected from the lateral surface of the calcaneus and the sinus tarsi. Subperiosteal exposure of the distal portion of the calcaneus is achieved, and the calcaneocuboid joint is identified but its capsule is left intact. An osteotomy is created 1.5 cm proximal to the calcaneocuboid joint, in the area between the anterior and middle facets. The osteotomy is opened laterally, and a tricortical iliac crest graft is inserted into the osteotomy. Care is taken to prevent dorsal subluxation of the distal calcaneus or calcaneocuboid joint, and pinning of the osteotomy and the calcaneocuboid joint is performed when needed. Reefing of the posterior tibialis and medial talonavicular capsule is done when laxity persists. Nearly always, the Achilles and peroneal tendons must be lengthened to attain a plantigrade foot. If the forefoot is supinated following lengthening of the calcaneus, a plantar-based closing wedge osteotomy of the medial cuneiform may be necessary to obtain a plantigrade foot. The foot is then immobilized in a non–weight-bearing, short-leg cast.

Published series have reported very good results in patients with flatfoot deformity and valgus of different origins, including CP, myelomeningocele, and idiopathic pes planovalgus ( Fig. 35-38 ). The preoperative rigidity of the foot does not always correlate with the postoperative result. Several patients continued to use orthoses after surgery, but pain relief and resolution of the talar head callus or blisters were common. Objective results obtained via pedobarography show improvement in plantar pressure following lateral column lengthening in comparison to extraarticular subtalar arthrodesis in patients with CP. Complications consisted of graft dislodgment and dorsal subluxation of the calcaneocuboid joint. Contraindications to the procedure are advanced osteoarthrosis and the presence of other bony deformities of the foot. Poor results are unusual but are generally caused by recurrence of planovalgus, which is less likely in patients who are independent walkers and have mild to moderate deformity. Overcorrection into varus because of spasticity of the tibialis posterior has been described.

A and B, Preoperative anteroposterior (AP) and lateral radiograph of the foot of a 17-year-old girl with right hemiplegia and painful pes valgus. C and D, AP and lateral radiographs after lateral column lengthening.

Subtalar Arthrodesis.

Subtalar arthrodesis has been proposed to provide long-lasting correction of symptomatic pes equinovalgus. The articular surfaces of the subtalar joint are resected. Tricortical autograft from the iliac crest or allograft can be used to achieve fusion and improve hindfoot position, followed by screw fixation of the subtalar joint ( Fig. 35-39 ). A large series from the Dupont Institute found good results in 96% following subtalar arthrodesis at an average 4.8-year follow-up.

A, Lateral foot radiograph of a 13-year-old, Gross Motor and Functional Classification System level 2 girl with spastic diplegia and symptomatic equinovalgus. Note the prominence caused by the plantar flexion. B, Lateral radiograph after subtalar fusion. The position of the foot is improved and the pain resolved.

Triple Arthrodesis.

Triple arthrodesis may be necessary for severe rigid symptomatic pes valgus in an adolescent with CP ( Fig. 35-40 ). By resecting the subtalar, calcaneocuboid, and talonavicular joints, the growth of these bones is disturbed, which leads to a small, shortened foot in younger patients. In adolescents, however, a well-corrected triple arthrodesis can yield a stable plantigrade foot for future ambulation. Indications for triple arthrodesis are pain, skin ulcerations over the talar head, and deformity interfering with shoe wear or ambulation in a child with a severe deformity not amenable to osteotomy. Valgus alone without disabling symptoms does not merit triple arthrodesis.

A to C, Triple arthrodesis for a severe fixed painful valgus deformity in a nonambulatory girl with cerebral palsy.

The technique used is identical to that performed in the nonneuromuscular population. Wedges of bone are resected with the articular surfaces, and internal fixation is performed when the bones are sufficiently strong. Screws or staples can be used. A short-leg, non–weight-bearing cast is applied, followed by a walking cast and then an orthosis until the fusion is mature.

Satisfactory outcomes are achieved when the deformity is well corrected. Postoperative malalignment usually results from undercorrection of the valgus deformity at the time of surgery. It is technically more challenging to perform a perfect triple arthrodesis for a valgus foot than for a varus foot, and it is particularly difficult to achieve fusion at the talonavicular joint. When visualization of the talonavicular joint is compromised, it is advised that a second medial incision be made to complete the joint resection.

Patients are generally satisfied with the results of triple arthrodesis, and rarely is talonavicular joint pseudarthrosis sufficiently symptomatic to interfere with function or necessitate further surgery. Degenerative changes have been documented in the ankle joint at an average of 18 years after triple arthrodesis in 43% of the pediatric population, but functional limitations in this group of patients are unusual.

Triple arthrodesis has also been used at our institution for the treatment of severe fixed deformity in nonambulatory patients who cannot wear shoes. The bony wedges resected in such cases are quite large, but patients and parents have been pleased with the improvement in the position of the feet and the ability to wear shoes in public.

Frost and associates have described triple arthrodesis combined with lateral column lengthening in patients with rigid planovalgus, but the authors have no experience with this procedure.

Clinical Features

The hamstrings are nearly always affected in patients with CP. Spasticity or contracture in the hamstrings is generally the cause of a crouch knee gait, but quadriceps and gastrocsoleus weakness may also lead to excessive knee flexion in stance phase. Findings during gait analysis include greater than normal knee flexion during midstance and an inability to extend the knee fully at the end of swing phase and continuing into initial contact ( Fig. 35-43 ). Step length then decreases as the knee loses excursion. Increasing demand is placed on the quadriceps to resist the progressive crouch, and energy expenditure during gait rises. The quadriceps and patellar tendon stretch, and patella alta and anterior knee pain may result. In severe cases, knee flexion contracture leads to failure of the extensor mechanism and fracture of the inferior pole of the patella ( Fig. 35-44 ).

Clinical appearance of a 13-year-old boy with spastic diplegia in crouch gait. He walks with increased hip flexion, knee flexion, and calcaneus at the ankles, primarily because of spasticity of the hamstrings.

Lateral radiograph of the knee of a teenager with spastic diplegia showing fracture of the inferior pole of the patella. No traumatic incident was noted.

It is important to note that the hamstrings cross two joints, the hip and the knee. At the hip the hamstrings serve as hip extensors, whereas at the knee they serve as knee flexors. The medial hamstrings also produce some dynamic internal rotation of the hip during gait.

Clinically, hamstring spasticity can be measured via the popliteal angle. The patient is positioned supine on an examining table and the hip is flexed to 90 degrees. The ipsilateral flexed knee is then extended, and the angle between the vertical and the position to which the tibia may be extended is the popliteal angle ( Fig. 35-45 ). Normal popliteal angles are variable, with a mean value of 26 degrees in normal children 4 years and older. Values greater than 50 degrees in this age range are considered abnormal. A decrease in the angle of straight-leg raising is also seen in those with tight hamstrings ( Fig. 35-46 ).

The Holt method of determining hamstring contracture. A, With the contralateral hip in extension, the tested limb is flexed to 90 degrees at the hip and the knee is extended passively. The angle between the anterior aspect of the leg and the axis of the thigh determines the degree of hamstring contracture. Bleck measures the angle on the popliteal surface between the leg and thigh. B, The popliteal angle of the left leg measures 70 degrees.

Method of determining hamstring tautness by straight-leg raising. The knee should be in complete extension and the pelvis should be stabilized. The angle between the lower limb and the examination table is measured.

In severe hamstring contracture, a fixed knee flexion contracture develops ( Fig. 35-47 ). It is important to assess for a fixed contracture because the presence of a contracture may lead to disappointing results after hamstring lengthening. When the contracture is most severe, the patient becomes unable to flex the hips, and a poor sitting posture with lumbar kyphosis and a slumped position results ( Fig. 35-48 ). Lack of lumbar lordosis can be seen radiographically in patients with increased popliteal angles.

Fixed knee flexion contracture in a teenage boy with spastic diplegia. Serial casts were required to correct the contracture after hamstring lengthening.

Hamstring tightness can lead to an inability to flex the hips sufficiently to sit. The patient then thrusts forward in the chair and sits with lumbar kyphosis.

As discussed earlier in the section on equinus, the examining physician must carefully assess other joints for spasticity and contracture. A bent-knee gait may be a compensation for equinus and toe-walking if the popliteal angle is normal. The hip must also be examined because correction of hamstring contractures without treating concomitant hip flexion contractures leads to increased hip flexion and anterior pelvic tilt during gait. Rotational malalignment such as excessive internal hip rotation or external tibial torsion should also be evaluated.

Treatment

Surgical Technique.

Historically, transfer of the medial and lateral hamstring tendons to the posterior femoral condyle has been performed for correction of crouched gait but has fallen out of favor because genu recurvatum was a frequent complication ( Fig. 35-49 ). Release of the proximal hamstring off the ischial origin has also been described, but hyperlordosis of the spine and anterior pelvic tilt occur frequently. Patients at greatest risk had hip flexion contractures of 25 degrees or greater. Proximal release is not recommended in patients who are able to walk. Proximal hamstring release in nonambulatory patients is discussed further in the later section “Soft Tissue Release for Subluxation or a Hip at Risk.”

Eggers transfer of hamstrings to the femoral condyles. a, Quadriceps femoris muscle; b, hamstring muscles; c, soleus muscle. A, Before the procedure. B, After the procedure.

(Redrawn from Eggers GNM: Transplantation of hamstring tendons to femoral condyles in order to improve hip extension and to decrease knee flexion in cerebral spastic paralysis. J Bone Joint Surg Am 34:827, 1952, with permission from The Journal of Bone and Joint Surgery, Inc.)

Surgical lengthening of the distal hamstrings is the preferred surgical treatment of crouched-knee gait and is usually performed in combination with other procedures. The technique of hamstring lengthening varies among surgeons. We prefer an open approach to perform intramuscular aponeurotic lengthening of the semimembranosus, Z -lengthening of the semitendinosus, and either tenotomy or Z -lengthening of the gracilis at a level proximal to the knee. When the lateral hamstrings are included in the procedure, intramuscular aponeurotic lengthening of the biceps femoris is done (see Plate 35-5 ). Usually, two cuts are needed in the fascia of the semimembranosus and biceps femoris for adequate lengthening. Medial hamstring lengthening suffices in ambulatory patients with mild to moderate crouch and increased popliteal angles. The addition of lateral hamstring lengthening further improves maximum knee extension during stance phase and is therefore helpful with more severe crouch and in patients with knee flexion contractures. The addition of lateral hamstring lengthening does increase the risk for knee hyperextension, however, especially in patients with spastic gastrocsoleus muscles. The popliteal angle is gently rechecked, and adequate lengthening has been accomplished when the angle is reduced to around 20 to 30 degrees. Percutaneous techniques to release the semitendinosus and gracilis have been described; lengthening of the semimembranosus and biceps femoris was performed via an open approach in this study.

Patients with severe crouch and fixed knee flexion contractures present a surgical challenge. Although improvement can be expected after medial and lateral hamstring lengthening, rarely can the crouch be eliminated. Posterior capsulotomy has been described, but newer alternatives may provide simpler and safer correction. Great care should be taken to prevent sciatic nerve palsy in these cases. Serial casts can be used to progressively extend the knees in an awake child after hamstring release.

Bone surgery has become accepted as a safer procedure for patients with CP who have fixed flexion contractures or those who have previously undergone hamstring lengthening. Shortening extension osteotomy of the distal end of the femur is useful in improving crouch at the knee in older children and teens who have previously undergone hamstring lengthening and who have knee flexion contractures. Pediatric implants such as a pediatric condylar blade plate provide stable fixation and enable earlier mobilization. Stout and colleagues studied a group of 74 patients who underwent distal femoral extension osteotomy with or without patellar tendon reefing and a group that underwent patellar tendon reefing without osteotomy. They found that patients who did not undergo patellar tendon reefing continued to walk in excessive knee flexion and advocated for patellar tendon advancement as treatment of the quadriceps insufficiency. Furthermore, they advocated that hamstring lengthening is not generally necessary when performing distal femoral extension osteotomy with patellar tendon advancement. Satisfactory results following shortening femoral osteotomy combined with patellar tendon reefing and transfer of the hamstrings to the femur have also been reported in older children and adolescents who were ambulatory but had severe crouch gait. These publications have in common the need to shorten the femur to gain knee extension but protect the sciatic nerve, which is at risk for postoperative palsy in patients with significant crouch gait. Although extension osteotomy can immediately correct a fixed flexion contracture of the knee, recurrence of deformity in time has been seen even following osteotomy.

A preliminary report on the use of anterior distal femoral staples or plating to gradually correct knee flexion contractures via guided growth principles shows possible utility. Guided growth may be preferable to extension osteotomy in skeletally immature patients with fixed knee flexion contractures. Tension band plates are placed medial and lateral to the patella, centered at the physis ( Fig. 35-50 ).

A and B, Growth modulation plates applied for residual flexion contracture of the knee following revision hamstring lengthening in a 10-year-old boy with spastic diplegia. C, One year following implantation, the distal end of the femur has grown into extension as evidenced by change in screw trajectory.

Complications.

Frequently, anterior pelvic tilt increases after hamstring lengthening as a result of weakening of the hamstrings. Because the hamstrings are also hip extensors, weakness leads to more relative hip flexion as a result of muscle imbalance and to forward tilt of the pelvis and trunk. Medial hamstring transfer to the femur was performed in a small series of children undergoing SEMLS procedures to preserve the hamstrings’ function as hip extensors. Knee hyperextension was seen in 12.5% following this procedure but was addressed by orthotic modifications. DeLuca and colleagues found that if just the medial hamstrings were lengthened, anterior pelvic tilt did not occur, but if the medial and lateral hamstrings were lengthened without psoas surgery, the pelvis did tip anteriorly. Muscle length modeling has been performed for the hamstrings and psoas in crouch gait and has shown that the hamstrings are not usually particularly short because they also cross the hip whereas the psoas is often shortened. If a hip flexion contracture is present, it must be surgically lengthened as well to minimize the postoperative tendency toward more hip flexion.

Postoperative gait analyses have also shown that extension at the knee may be improved, but rarely normalized, following hamstring with or without rectus surgery but that residual crouch at the knee will contribute to the development of calcaneus at the ankle with loss of push-off power during walking. In children with significant knee flexion preoperatively in whom it is predicted that knee position will be improved but full extension not achieved, extreme caution should be taken with any Achilles tendon–lengthening procedure because further loss of ankle power will worsen the tendency toward a postoperative calcaneus gait.

Yet another complication of hamstring lengthening may be stance-phase hyperextension of the knee. Patients with a preoperative “jump knee” pattern, described as increased knee flexion at initial contact but extension of the knee in midstance because of the ankle plantar flexor–knee extensor couple, are particularly prone to this outcome. It can occur with medial hamstring lengthening even after rectus femoris transfer in this group at risk. In such patients, postoperative bracing with an AFO to maintain ankle dorsiflexion can be helpful in minimizing the knee hyperextension.

On occasion, hamstring lengthening can lead to palsy of all or part of the sciatic nerve ( Fig. 35-51 ). A mathematic equation is available to predict how much straightening can be performed safely, but clinical judgment in not aggressively stretching the posterior structures of the knee in patients with significant contractures is required. Intraoperative EMG shows diminution of amplitude with progressive extension of the knee, particularly when the hip is flexed during surgery. Hip flexion (i.e., long sitting at 90 degrees) with the knees extended in casts can further stretch the sciatic nerve. In our experience, partial or complete sciatic nerve palsy occurred in 9.6% of patients following hamstring lengthening (as part of the SEMLS approach). Patients most at risk were older and had impairment in communication. Nonambulatory patients were most at risk. Postoperative nerve palsy is extremely difficult to manage. Casts must be removed and the knee allowed to flex as soon as the nerve palsy is recognized. Simply splitting the cast does nothing to relieve stretch on the nerve. Medical management of the painful paresthesias is often necessary with medications such as amitriptyline (Elavil) or gabapentin (Neurontin). Prevention of excessive stretch is important in older children with knee flexion contractures, which has led to popularization of shortening extension osteotomy of the distal end of the femur in such cases.

A and B, Fourteen-year-old boy who is minimally ambulatory with severe crouch at the hip and knee and fixed knee flexion contractures. Hamstring lengthenings led to dysesthesias in his feet postoperatively, which resolved over time.

It is common to also see a decrease in flexion of the knee in swing phase as a result of spasticity of the rectus femoris muscle after hamstring lengthening. Normally, the knee should flex at least 60 degrees in swing phase, and this flexion occurs early in swing phase. As a muscle that crosses two joints, the rectus acts to flex the hip at initial swing and to extend the knee. Preoperative EMG will often show inappropriate electrical activity in the rectus femoris during midswing. Gait analysis in affected patients shows a decrease in the amount of swing-phase knee flexion and a delay in when the peak swing-phase knee flexion occurs ( Fig. 35-52 ). When severe, it leads to problems clearing the foot in swing phase and thereby results in tripping and dragging the toe. Patients may complain of difficulty climbing stairs or stepping up onto a street curb or in moving from a standing position to a seated position and vice versa, known as transitional movements.

Hamstring spasticity leads to an inability to extend the knee and accept weight at initial contact (which occurs at 0% of the gait cycle). When the rectus femoris is spastic, the knee is unable to flex rapidly at initial swing phase, which is to the right of the vertical lines . The amount of swing-phase knee flexion is decreased, and its timing is delayed. Normal kinematics is depicted by the black dotted curve and that of a child with spastic diplegia by the red curve . Vertical lines designate divisions between the stance and swing phases.

Spasticity in the rectus femoris can also be tested for during physical examination. The Duncan-Ely test is performed by positioning the patient prone and then flexing the knee to 90 degrees. If the rectus femoris is spastic, the ipsilateral buttock will rise from the table as a result of the hip flexion caused by the rectus ( Fig. 35-53 ). Patients with normal Duncan-Ely test results are unlikely to benefit from rectus transfer. Unfortunately, this test is not specific for the rectus because a patient with a hip flexion contracture secondary to tightness of the iliopsoas will also have a positive Duncan-Ely test. Another clinical measure of rectus spasticity is rectus grab. With the patient supine on the examining table, the knee is rapidly flexed. If resistance is felt, the rectus is spastic.

A and B, The Duncan-Ely test. With the patient prone, the knee is passively flexed. A positive result occurs when the ipsilateral buttock rises, which may indicate rectus femoris spasticity.

A symptomatic stiff-knee gait from overactivity of the rectus femoris during swing phase does not develop in all patients after isolated hamstring lengthening. Although Damron and colleagues found that 71% of patients lost some knee flexion in swing phase after hamstring lengthening, only 13% of ambulatory patients required rectus transfer for correction of stiff-knee gait. Dhawlikar and colleagues described recurvatum after distal hamstring lengthening and a need for subsequent rectus femoris transfer in 17% of their patients.

Indications

Surgical treatment of stiff-knee gait and inability to flex the knee in swing phase consists of rectus femoris transfer. This procedure is often performed simultaneously with hamstring lengthening and other soft tissue procedures but has also been performed in isolation in children without crouch but with stiff-knee gait. The principle behind rectus transfer is to preserve the role of the rectus femoris as a hip flexor but to move the distal rectus insertion posterior to the axis of the knee to eliminate its role as an inappropriate knee extensor during swing phase. Release of the proximal rectus femoris was studied but found to increase swing-phase knee flexion less than after distal transfer of the rectus tendon. Release of the rectus from the patella with mobilization of the muscle was also determined to be ineffective in treating loss of knee flexion in swing, probably because of readherence to the underlying quadriceps postoperatively. Neither of these procedures physically moves the distal part of the rectus posterior to the knee joint, which may be the reason why they do not work as well as rectus transfer. Riewald and Delp measured knee moments after rectus transfer and did not see that the rectus generated a knee flexor moment after surgery. MRI of the trajectory of the transferred tendon likewise does not support change of the rectus to a knee flexor. Regardless, rectus transfer has been found to increase swing-phase knee flexion by an average of 16 degrees. When rectus transfer is combined with a hamstring-lengthening procedure, dynamic range of motion and crouch improve without loss of swing-phase knee flexion.

Surgical Technique

(see Plate 35-6 and Video 35-2 )

An incision is made superior to the proximal pole of the patella. Many incisions have been described, but we prefer to use a short transverse incision two to three fingerbreadths above the patella. Through this cosmetic incision the rectus femoris is dissected off the underlying vastus intermedius. Distally, the two muscles and their tendons are adherent, so it is easier to start the dissection more proximally, where the tissue plane can be identified. The vastus lateralis and medialis also converge distally at the patellar insertion of the quadriceps tendon. Care must be taken to preserve these two muscles as well. Once the rectus femoris is dissected from the other parts of the quadriceps, the tendon is divided transversely just proximal to the superior pole of the patella, again taking great care to leave the tendon of the rest of the quadriceps undisturbed. A sturdy stitch is woven into the tendon of the rectus femoris, and a subcutaneous tunnel is made to the site of transfer. The tendon is then passed medially, usually through the posterior wound used for concomitant hamstring lengthening, and the rectus is sewn into the stump of the gracilis tendon, the sartorius muscle, or the lengthened semitendinosus. The remainder of the quadriceps tendon is then repaired by suturing the vastus lateralis to the medialis over the intermedius.

Postoperative Care

Postoperative care consists of either a long-leg cast or knee immobilizer, and early weight bearing and ambulation are encouraged, as for hamstring lengthening.

Results

Abundant research investigating the outcome of rectus femoris transfer has been published. First, the preferred site for transfer was studied by Ounpuu, Gage, and others. Although it was hypothesized that rotation of the hip would become more external if the tendon were transferred medially and more internal if the tendon were transferred laterally, they found that rotation of the femur did not change, regardless of where the tendon was transferred. The site for tendon transfer is based on the surgeon’s preference and the existence of wounds from other concomitant surgeries such as simultaneous hamstring lengthening. Miller transfers the rectus to the sartorius, whereas Gage prefers to transfer it to the gracilis. Aiona and Sussman transferred the rectus to the iliotibial band in a group of patients and found the results to be identical to those from a group in which the rectus was reattached to the medial hamstring tendons. It appears that the results of transfer of the rectus tendon do not depend on the anchor site.

Results on the role of EMG in determining whether a stiff-knee gait will occur after hamstring surgery and in predicting the outcome of rectus femoris transfer are conflicting. Preoperative EMG of the rectus femoris and vastus lateralis has not been shown to be predictive of the amount of peak swing-phase knee flexion after rectus release or transfer. Miller and co-workers found that the best results were achieved in patients who had phasic but inappropriate rectus activity in swing phase on EMG. Patients with less than 80% of normal preoperative dynamic range of knee motion (i.e., stiff knees preoperatively) benefit from rectus transfer more so than do those with nearly normal motion.

Other predictive variables used to study rectus femoris transfer include walking speed, dynamic range of motion, and joint kinetics. Patients whose walking speed is at least 80% that of age-matched normal subjects walk better after rectus transfer than do their slower counterparts. It is therefore logical that independent ambulators would have better results than walker-dependent or exercise ambulators. This has been substantiated in a study by Rethlefsen and colleagues, who found poor results following rectus transfer in children who were GMFCS level 4 before surgery because of worsened crouch. Patients with 80% or more of normal dynamic range of motion of the knee on preoperative gait analysis do not appear to benefit from rectus transfer, whereas patients with good power generation at the ankle and hip do best with rectus transfer. If a patient has difficulty initiating swing phase and cannot powerfully flex the hip to lift it from the ground, little momentum is available to produce swing-phase knee flexion. If power is satisfactory, however, sufficient momentum is present to allow knee flexion if the rectus femoris spasticity does not interfere—hence the better results for transfers in the presence of good joint power. Inferior outcomes have been described in patients who underwent rectus transfer but had rotational abnormalities exceeding 8 degrees. If the knee and feet are not pointing straight ahead, swing-phase knee flexion does not occur in the sagittal plane and rectus transfer is not optimal.

Cruz and coauthors in 2011 published the results of 42 patients treated by intramuscular lengthening of the rectus femoris. Using gait analysis, they found similar results as seen following rectus transfer after this simpler procedure. Further study comparing rectus lengthening versus transfer is needed to clarify which patients will benefit most from which procedure.

Derotational Osteotomy

Correction of rotational malalignment of the lower extremity is best achieved through derotational osteotomies. Femoral anteversion is treated by femoral osteotomy, either proximally at the intertrochanteric or subtrochanteric level or distally at the supracondylar level ( Fig. 35-55 ).

Distal femoral rotational osteotomy for the treatment of excessive anteversion in an 11-year-old girl with spastic diplegia.

Those who advocate proximal osteotomy believe that rotation of the knee extensor mechanism with the distal osteotomy is undesirable, although comparative studies have not been conducted. Computer simulation of intertrochanteric, subtrochanteric, and supracondylar osteotomies has shown minimal effect on the length of the hamstring and adductor muscles. If the osteotomy is performed proximally, the patient is usually positioned prone on the operating table. Rotation to allow twice as much external rotation of the hip as internal rotation is the goal—for example, 30 degrees of internal rotation and 60 degrees of external rotation. Mathematic models have been devised to quantify the amount of rotation needed intraoperatively but have not been widely used. Fixation with a blade plate, a standard fixation plate and screws, or a locking plate is performed, and postoperative immobilization is used only when the surgeon thinks that loss of fixation may result because of osteopenia.

Distal osteotomy is performed at the supracondylar level through a lateral approach with the patient supine and the legs draped free. The benefits of performing the osteotomy distally are ease of the procedure and the ability to use a tourniquet. The femur is exposed by elevating the vastus lateralis anteriorly off the intramuscular septum. K-wires are used to quantify the amount of rotation intraoperatively. Hoffer and colleagues used Steinmann pins to quantify the rotation and then used the pins for fixation by incorporating them into the cast. They did encounter pin tract infections, so those who perform distal osteotomy now generally use stable internal fixation with a plate and screws. Hip rotation can be assessed in flexion after provisional fixation so that symmetry in internal and external rotation can be achieved. Immobilization with long-leg casts in knee extension allows standing and early weight bearing.

Finally, fixation of femoral rotational osteotomies can be achieved with flexible intramedullary nails. This technique may result in less postoperative weakness because of lack of dissection in the zone of the osteotomy. Stable internal fixation is emphasized when performing rotational osteotomies on children with CP. Stable fixation, whether of the proximal or distal end of the femur, can allow early weight bearing and more timely resumption of ambulation.

Postoperative gait laboratory studies have shown improvement in hip rotation after femoral osteotomy for intoeing in patients with CP. These changes are appreciated by the parents, who voice high satisfaction with the procedure.

Recurrence or persistence of some degree of internal rotation occurs in up to a third of patients with CP who undergo femoral rotational osteotomy. Patients who are younger than 10 years at the time of surgery are at higher risk for recurrence. Gait studies have shown that the mean change in dynamic and static hip rotation after either proximal or distal femoral osteotomy is approximately 40% less than what was reported at surgery. Changes in pelvic rotation can be unpredictable and may compromise accurate correction of the intoeing. For example, patients with spastic hemiplegia or asymmetric diplegia typically walk with external pelvic rotation and internal hip rotation on the more neurologically affected side. Pelvic rotation often becomes more neutral after femoral osteotomy. If not planned for, the net result of femoral osteotomy with spontaneous correction of pelvic rotation would be persistent intoeing.

In patients with tibial rotational deformities, surgical correction should be performed at the distal level. Proximal osteotomies are associated with a higher risk for neurovascular injury. Satisfactory results have been published for tibial osteotomy without concomitant fibular osteotomy, but most surgeons continue to cut the fibula in cases in which larger amounts of rotation are desired. Internal fixation with a plate and screws or crossed K-wire fixation can be performed with a low complication rate ( Fig. 35-56 ). Alternatively, osteoclasis of the distal end of the tibia through percutaneous drill holes and intramedullary nail fixation in skeletally mature teens has been used. Gait studies have shown that realignment of the tibia tends to normalize the forces working at the ankle and foot.

Bilateral rotational osteotomies of the femora and tibiae for the treatment of femoral anteversion and external tibial torsion.

Clinical Features

Hip flexion contractures are found most commonly in patients with spastic diplegia and spastic quadriplegia and are one component of the patient’s overall crouched-gait pattern. Hip flexion contractures are nearly always seen in combination with increased hip adduction and internal rotation, knee flexion secondary to hamstring spasticity, and equinus, calcaneus, or valgus deformities of the feet. Hence, surgery to improve hip flexion contractures is part of the SEMLS approach in conjunction with other soft tissue or bony procedures in patients with CP.

Diagnosis

The flexion contracture is caused by increased tone in the hip flexors, primarily the iliopsoas, and relative weakness of the hip extensors, such as the gluteal muscles. The contracture is identified during physical examination by performing the Thomas and Staheli maneuvers. In the Thomas test, the patient is positioned supine on the examining table. The opposite hip is fully flexed to flatten the lordosis of the lumbar spine and lock the pelvis against moving. The angle between the table and the hip in question is then measured because the hip will rise up in flexion off the table in the presence of a contracture ( Fig. 35-57 ). The Staheli test is performed by placing the upper part of the body of the patient prone on the table with the hips dangling off the edge of the table. The angle formed by the horizontal and the thigh, at the point at which further hip extension causes the pelvis to move, is the hip flexion contracture ( Fig. 35-58 ).

The Thomas test reveals a 30-degree flexion contracture of the right hip. The opposite hip is fully flexed to flatten the lumbar lordosis.

The Staheli test, used to determine hip flexion deformity with the patient prone. A, The pelvis is stabilized, the patient’s thigh is raised toward the ceiling, and the tested hip is extended. Normal extension is 30 degrees. B, The degree by which the hip fails to reach neutral position is the degree of deformity.

During gait, a hip flexion contracture is apparent either as increased flexion of the hip during the middle of stance phase (when the hip should be extended) or as increased anterior pelvic tilt. The anterior pelvic tilt produces either forward lean of the upper part of the body during gait or increased lumbar lordosis as the spine extends above the flexed pelvis.

On radiographs, the sacrofemoral angle can be used to objectively quantify the hip flexion contracture. A standing lateral radiograph that includes the proximal femoral shaft and the lumbar spine is taken. A line is drawn along the superior surface of the sacrum and another along the femoral shaft. The intersection of these lines is the sacrofemoral angle, which should normally be between 45 and 65 degrees. In the presence of a hip flexion contracture, the sacrofemoral angle decreases ( Fig. 35-59 ).

The sacrofemoral angle. With increasing hip flexion contracture the pelvis tips forward and the sacrum becomes more vertical. The angle formed between a line drawn along the superior aspect of the sacrum and the femoral shaft decreases with flexion of the hip.

Hip flexor surgery in a walking child is done to improve the hip flexion contracture, but more often than not the goal of the surgery is to prevent increasing hip flexion and anterior pelvic tilt when hamstring lengthening is performed. As noted earlier in the discussion on knee surgery, not only are the hamstrings knee flexors, but they are also hip extensors. The hamstrings lose strength after lengthening, so any preexisting hip flexion contracture will be exacerbated after hamstring surgery. Therefore, hip flexor lengthening may be part of the overall surgical correction of crouch gait. A study by Truong found that surgical management of crouch gait led to greater improvement (in gait analysis) in GMFCS 3 and 4 children with maximum stance-phase hip extension no greater than 8 degrees of flexion, anterior pelvic tilt greater than 24 degrees, and excessive range of sagittal-plane pelvic motion when psoas lengthening was performed.

Surgical Technique

The recommended procedure to correct increased hip flexion is a psoas tenotomy performed over the pelvic brim. The surgical approach is anterior, through an oblique incision just distal to the anterior superior iliac spine. The psoas is located between the sartorius and the femoral sheath. The femoral nerve nearly overlies the psoas tendon. The tendon of the psoas is identified deep within the iliacus muscle, which is not lengthened. The tendon is then transected and slid within the iliacus, thereby increasing the overall length of the iliopsoas. This is similar to the technique described by Salter as part of his innominate osteotomy. Gait analysis studies have shown improvement in hip extension during stance and in hip moments and power and no loss of strength after lengthening either at the pelvic brim or over it.

Release of the iliopsoas tendon off the lesser trochanter of the femur should not be done in ambulatory patients because it results in loss of hip power and inability to forcibly flex the hip against gravity. Climbing stairs becomes extremely difficult, and gait deteriorates. The gait of children who have undergone iliopsoas release is characterized by increased pelvic motion and a decreased arc of hip flexion and extension as the trunk tries to substitute for the weak hip flexors in pulling the leg forward off the ground. Others circumduct to advance the leg.

Bleck advised against simply releasing the iliopsoas as well and suggested attaching the distal iliopsoas tendon anteriorly into the hip capsule. This would allow additional length, yet hip flexion would be preserved. Because psoas tenotomy over the pelvic brim is technically easier, transferring the tendon to the capsule is rarely performed at present.

Proximal rectus femoris release has been described for correction of hip flexion contracture in patients with CP. Gait studies have failed to show efficacy, however.

Clinical Features

Spasticity in the adductor muscles in CP results in a narrow base of gait and scissoring. The patient has difficulty advancing one limb in front of the other as the swing limb contacts the ground in front of the other leg. Young children may be unable to progress in their ability to ambulate because of the scissoring ( Fig. 35-60 ). Over time, the untreated adduction contractures, when combined with a hip flexion contracture, lead to progressive hip subluxation and possible dislocation. Surgery for the unstable hip in patients with CP is discussed in the later section “Soft Tissue Release for Subluxation or a Hip at Risk.”

Seven-year-old boy with cerebral palsy and scissoring of the hips.

The muscles leading to the adduction contracture are the adductor longus, adductor brevis, adductor magnus, gracilis, and occasionally the pectineus.

Diagnosis

Clinical examination reveals an inability to abduct the hips in flexion and extension. The tendon of the adductor longus is palpable and visibly tight in the groin. The child walks with knees knocking, and one foot scissors over the other in stance phase. The feet may appear locked together as the child has difficulty initiating swing phase. EMG has shown abnormal swing-phase electrical activity in the adductor muscles in patients with scissoring.

A word of caution is needed here. Increased femoral anteversion when combined with crouch at the knee can produce the appearance of scissoring. This clinical scenario has been termed pseudoadduction . Careful observation of the patellae during gait will alert the surgeon to the internal rotation.

Surgical Treatment

Bracing has not been shown to improve adduction contractures, and although botulinum toxin injections may relieve dynamic adduction, this treatment modality is still under investigation. Surgery to improve adduction contractures is limited to adductor release, with or without obturator neurectomy, and posterior adductor transfer.

Posterior Adductor Transfer.

The rate of recurrent contracture led Perry in the early 1960s to devise a procedure in which the adductor longus, adductor brevis, and gracilis tendons are transferred from the pubic ramus to the ischium ( Fig. 35-61 ). The new origin of the muscle converts the adductors to hip extensors, thereby lowering the risk for recurrent contractures and further stabilizing the hips. The surgery was designed for patients with poliomyelitis, but it soon began to be used in the CP population. Many studies of adductor transfer then followed and found improved abduction, extension, hip stability, scissoring, and standing balance, with better results achieved in ambulatory patients.

Posterior transfer of the hip adductors to the ischium. A and B, Anteroposterior and lateral views showing the line of division of the gracilis and adductor longus muscles at their origin and the line of myotomy of the adductor brevis. C and D, The adductor longus and gracilis muscles are transferred to the ischium, and the adductor brevis is divided.

Pelvic obliquity and unilateral hip subluxation have been reported in patients 10 years after posterior transfer of the adductor tendons, presumably because of unilateral loss of fixation of the tendons to the ischium. Loder and colleagues found that only 19 of 33 transfers to the ischium remained attached and that asymmetry of the hip and pelvis occurred if only one side of the tendon remained attached. To avoid this, Beals sutured the adductor longus and brevis into a lengthened gracilis, which remains attached to its origin, thereby increasing the integrity of the transfers and improving abduction. Although adductor transfer has been purported to maintain abduction better than release does, adductor tenotomy is simpler.

We currently do not perform adductor transfer at our institution for the treatment of adduction contractures in CP. We continue to perform adductor tenotomy but no longer recommend obturator neurectomy because of problems with abduction contractures. We mobilize patients earlier than previously, and we use removable abduction bars more frequently now to decrease postoperative stiffness.

Etiology and Epidemiology

Before an extensive discussion about surgical reconstruction of a subluxated or dislocated hip in a patient with CP, it is important to understand the epidemiology and etiology of hip instability in this condition.

Hip dysplasia or instability is a common problem in patients with CP and occurred in approximately 21% of 1450 patients at the Hospital for Special Surgery. Other series report a prevalence of subluxation or dislocation ranging from 3% to 47%. The incidence of hip dysplasia varies with the severity of neurologic involvement. Hip dysplasia and dislocation may rarely develop in patients with spastic hemiplegia. Patients with spastic diplegia are at increased risk. Patients with spastic quadriplegia who have total body involvement have the highest rate of hip instability, with hip subluxation or dislocation developing in almost 50%. The incidence of hip subluxation and dislocation is also linked to the ability of the patient to walk. Nonambulatory patients are at much higher risk than those who can walk and account for 89% of those with hip instability and CP. Soo and associates correlated the incidence of hip dysplasia with the GMFCS level. They found that level 1 children (community ambulators with minimal disability) had a 0% incidence of hip dysplasia; level 2, a 15% incidence; level 3, a 41% incidence; level 4, a 69% incidence; and level 5 (totally involved wheelchair-bound children without head or trunk control), a 90% incidence of dysplasia. The relationship between the prevalence of hip dysplasia and more severe GMFCS level has been verified in other studies as well. The mean age at which patients with CP are initially seen with subluxation or dislocation is 7 years, although radiographic changes consistent with subluxation can be found as early as 18 months of age in some patients. Hip subluxation develops in 60% of children with CP who are unable to walk by 5 years of age.

Hip subluxation develops in response to muscle imbalance. Spasticity and contracture of the adductors and flexors of the hip overpower the weaker and noncontracted hip extensors and abductors. Subluxation develops gradually, with increasing lateralization and proximal migration of the femoral head with respect to the acetabulum. This is completely different from developmental dislocation of the hip, in which soft tissue laxity leads to instability of the hip ( Fig. 35-62 ). The hip in CP is not grossly unstable on clinical examination; it is an extremely rare case (usually a hypotonic child) in which an Ortolani maneuver is positive with reduction of the hip. Rather than laxity, the hip is pried from the acetabulum over time by spastic muscles. It has been found through computer modeling that the forces exerted across a spastic hip in CP are up to six times greater than normal.

Mechanism of superior and posterior displacement of the femoral head out of the acetabulum. A, A normal hip. B, In cerebral palsy the hip adductor and iliopsoas muscles are spastic and shortened, and the gluteus maximus and medius muscles are weak. The center of movement of the hip is translated from the center of the femoral head distally to the level of the lesser trochanter. The hip joint capsule is elongated superoposteriorly, with gradual dislocation of the hip.

(Redrawn from Sharrard WJW: Paediatric orthopaedics and fractures , ed 2, Oxford, 1979, Blackwell.)

Bony deformity, then, occurs in response to the spasticity. The normal remodeling of the femoral anteversion seen in a neurologically normal young child does not occur in patients with CP, and anteversion persists into adulthood. The increased anteversion has been shown to correlate strongly with the development of hip dysplasia, particularly in nonwalking patients. The neck–shaft angle becomes increased as coxa valga develops. The anteversion worsens the radiographic appearance of the valgus neck. The lesser trochanter becomes elongated because of pull from the iliopsoas. Acetabular changes consisting of an increased acetabular angle and erosion of the lateral lip of the acetabulum by the subluxating femoral head occur as the hip subluxates. Finally, the shape of the femoral head changes, with superolateral and then superomedial notching as a result of pressure from the capsule, the rim of the acetabulum, the abductors, and the ligamentum teres ( Fig. 35-63 ).

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

Back Pain Other Anatomic Disorders of the Spine Slipped Capital Femoral Epiphysis Legg-Calvé-Perthes Disease Disorders of the Spinal Cord Poliomyelitis

Stay updated, free articles. Join our Telegram channel

Join