|Home | About | Journals | Submit | Contact Us | Français|
Obstetrical brachial plexus injuries are reported in the medical literature at a rate of 0.38 to 2.6 per thousand live births. Historically, the management of these lesions has been conservative, with observation and physical therapy as the primary modalities of treatment. However, experience has shown that a small majority of these devastating lesions have required more direct and invasive approaches. The experience gathered over a 15-year time span of managing these cervical nerve injuries has afforded the Texas Children's Hospital Brachial Plexus team the opportunity to come to several conclusions regarding the global treatment of these patients. The first is that diagnosis, observation, and therapy are the initial approaches to these injuries and should be initiated immediately. Second, early surgical intervention is essential to maximizing the long-term improvements in select patients by helping to prevent residual growth deformities and underdevelopment of the affected limbs. Third, the development of secondary residual deformities must be addressed with secondary reconstructive procedures to arrest the underdevelopment of affected limbs. These goals of reconstruction have been implemented over a period of 15 years and have been shown to provide marked improvements in the functionality and quality of life in patients affected with these physically disabling lesions.
Brachial plexopathy has been traced back to the early beginnings of civilized man, with the first diagnosis and treatment of cervical injuries with subsequent arm paralysis made by the great Roman clinician Galen in 138 AD1
Galen masterfully associated and treated a temporary paralysis and hypoesthesia of the upper extremity to a lateral traction injury of the neck. Galen also elucidated the existence of somatomotor and somatosensory units within the nerve structure.2
The complexity of the brachial plexus was then further deciphered by the exquisite and masterful drawings of Leonardo da Vinci in the 15th century, following his dissection of human cadavers.
Following these masterful findings, the true classification of obstetrical brachial plexus palsy then took inception from Smellie in 1746, with his description of transient arm palsy in a newborn.3 Duchenne, Erb, Klumpke, and Seeligmuller followed suit by isolating cervical nerve roots injuries associated with specific muscle groups in the upper extremity.4,5,6,7,8
Following the establishment of the diagnosis and classification of these injuries, history then became entangled in a quagmire of contradicting recommendations regarding the ideal surgical management of brachial plexus injuries. Thornburn in the early 1900s became the first to surgically repair a brachial plexus lesion. However, the high incidence of perioperative and postoperative morbidity and mortality lead to much controversy regarding the ideal surgical management of these patients in the following years. The world wars continued to produce a multitude of brachial plexus lesions; however, satisfactory results in the treatment of these lesions were scant.
It was not until the advent of microsurgery in the 1960s that surgical intervention of brachial plexus lesions truly became a mainstay of therapy. Developments and advances by Millesi in 1964 and Narakas in 1966 furthered the concept of primary nerve grafting and neurorrhaphy, revealing that surgical intervention afforded patients better long-term recovery and function of their disabling injuries.9,10,11
At present, the surgical philosophy of these lesions has now turned to early diagnosis and surgical intervention when indicated. The increasing sophistication of diagnostic equipment, microsurgical instruments and techniques, therapy, and the understanding of nerve injuries on a molecular and global nature has led to great advances in the magnitude of functional recovery of brachial plexus lesions. These modalities, coupled with the armamentarium of muscle transfers and releases pioneered in earlier years by surgeons such as Mayer (1927), Ober (1932), L'Episcopo (1934), and others has resulted in the renewed interest and global improvement of brachial plexus injuries.12,13,14
The Brachial Plexus Program at Texas Children's Hospital has been in existence for over 16 years, and over 5000 patients have been evaluated with obstetrical brachial plexus palsies. Over 650 brachial plexus operations were performed in 2002, and over 3000 have been performed over the last 12 years. Virtually all patients have shown improvement in upper extremity active range of motion, with no loss of function in our experience. A multidisciplinary approach including a dedicated pediatric neurologist, a pediatric neurosurgeon, two reconstructive microsurgeons, a physical medicine and rehabilitation specialist, an electrophysiologist, and a team of pediatric occupational therapists make up our staff. The continuing success of the program depends on basic science and clinical innovations only possible through research. All the physicians in the Texas Children's Hospital Brachial Plexus Team are full-time academic faculty members of the Baylor College of Medicine. Our current research endeavors range from outcome studies and epidemiology to nerve gene therapy studies supported by the National Institutes of Health.
The goal of brachial plexus surgery is to restore form and function. Reconstructive procedures may involve the nerves, bones, tendons, muscles, blood vessels, and skin. Over 1200 nerve grafts and over 350 nerve transfers have been performed, with over 90% improvement in function and over 3500 tendon/muscle transfers with over 95% improvement.
As the drawing of Leonardo da Vinci initially demonstrated, the brachial plexus is a complex intertwining of somatosensory afferent and somatomotor efferent nerve structures from the fifth through the first thoracic spinal nerves. Contrary to an initial reaction, the complex relationship of the brachial plexus and its surrounding structures is somewhat constant with little variations (Fig. 1).
The brachial plexus originates as the C5–T1 spinal nerve roots exit the posterior triangle of the neck anterior to the middle scalene muscle forming the upper, middle, and lower trunks (Fig. 2). Variations in the origin of the plexus exist in situations when the plexus originates one nerve root above (prefixed) or one nerve root below (postfixed). These variations in anatomy are subject to increased risk of injury secondary to higher tensile forces and to higher susceptibility to crush injuries from surrounding structures.15
As the plexus continues its course, passing under the clavicle with the subclavian vessels, the three anterior and three posterior divisions of the plexus begin to form. The divisions of the plexus then position around the axillary artery and converge distal to the clavicle to form the lateral, medial, and posterior cords of the brachial plexus.
In the distal portion of the axilla, the lateral cord will then form the musculocutaneous nerve. The posterior cord will divide into the radial and axillary nerves, and the medial cord will form the ulnar nerve. Finally, as the cords course distally and form the terminal branches, the median nerve originates as the lateral and medial cords unite in an “M” pattern with both the lateral cord carrying somatoafferent fibers and the medial cord carrying the somatomotor fibers.
Having followed the pathway of the cervical nerve roots to their terminal motor end plates, it is now clear how each nerve root elicits an action potential in specific muscle groups of the upper extremity. The C5 nerve root elicits deltoid function as shoulder abduction, extension, and external rotation. The C6 nerve root elicits shoulder adduction and flexion, elbow flexion, and forearm supination. C6 and C7 provide internal rotation of the shoulder. C7 contributes to elbow and wrist and finger extension. Finally, C8 and T1 elicit wrist and finger flexion and intrinsic muscle function.
Reviewing the anatomy of the brachial plexus in such a cursory manner emphasizes the complexity of this structure. However, understanding the foundation of the nerve roots and their main pathways with the multiple divisions and subdivisions from the cervical region into the upper extremity is the basis for planning the reconstruction of injuries to the brachial plexus.
In addition to having a detailed knowledge of the anatomy of the brachial plexus, it is essential for the reconstructive surgeon to posses a thorough understanding of nerve injury at the cellular and molecular level. Insult to the plexus nerves may occur with varying degrees of severity and the location of the injury is paramount to developing a reconstructive plan.
When the nerve root has been avulsed from the spinal cord, a preganglionic disruption of ventral and dorsal horns has occurred. Experience has taught us that recovery from this lesion is not amenable to primary nerve grafting. Rupture of the nerve distal to the dorsal root ganglion, a postganglionic injury, is amenable to primary nerve grating if indicated. Determination of this is made by classification of peripheral nerve injuries according to the Seddon and Sunderland classifications (Table 1).
In the great majority of obstetrical brachial plexus lesions, a neuropraxia, or Sunderland first-degree injury, has occurred where the nerve has sustained a stretch injury causing an ionic conduction block. There has been no axonal disruption, and full recovery is expected. When the axon has sustained an injury—axonotmesis—recovery is dependent on the damage to the myelin sheath. If the endoneural tube is intact, a Sunderland second-degree injury has occurred, and full recovery is expected. However, if the endoneural tube is disrupted (third-degree injury), or if the epineurium is the only structure intact (fourth-degree injury), recovery is incomplete, leaving a neuroma-in-continuity. Neurotemesis, or Sunderland fifth-degree injury, is a complete disruption of the nerve. In 1989, Mackinnon introduced a sixth-degree injury to describe mixed injuries (first to fifth degrees) within different fascicles at the same anatomical level.16 Spontaneous regeneration of the nerve bud is disorganized and in a haphazard fashion, resulting in a neuroma.16 Functional recovery is not expected at this level, and surgical intervention is warranted for these grave injuries.
Examination of nerve regeneration at the molecular level has also shown that this process is as complex as its cellular counterpart. Nerve regeneration has been shown to occur and proceed at blistering speeds of ~1 mm/day, or more globally at 1 inch/month.17 This information is important because terminal muscle atrophy and loss of motor end-plate density occurs between 12 and 18 months following denervation. It is for these two reasons that if a neurotemetic obstetrical brachial plexus lesion occurs, it is paramount to perform nerve grafting as soon as the diagnosis is established. Surgical intervention at this stage will produce far superior functional repair than will attempting secondary reconstructive procedures once secondary deformities have occurred.18,19
The incidence of obstetrical brachial plexus lesions is reported to be 0.38–2.6 per 1000 term infants in United States.20,21,22 History has shown that the great majority (95.7%) of obstetrical brachial plexus palsies do resolve spontaneously, with 92% of the recovery taking place during the first 3 months.23
In our center, we have experienced an incidence of these lesions of 1 per 1000 live births. Cases presenting predominate as upper cervical root injuries, or Erb-Duchenne palsy (73%); of the remainder, 25% present as total or global plexus injuries, and 2% present as lower or Klumpke's palsy. Bilateral lesions are seen ~4% of the time, and they are associated with breech presentations.24 In patients with unilateral injuries, right-sided lesions predominate in 54.1% of patients, versus left-sided lesions in 45.9% of patients. Further breakdown of the figures has shown that one to two patients per 10,000 live births require surgery.25
The underlying mechanism of obstetrical brachial plexus injuries has been studied and is the source of much controversy and recent litigious debate.26 Recent studies in the obstetrical literature have alluded to alternative causes of obstetrical brachial plexus lesions.27,28,29,30 These studies determine that certain cases of brachial plexus injuries occur secondary to shoulder dystocia associated with high intrauterine forces, not traction injuries. Findings such as disuse osteoporosis immediately after birth, hypoplastic glenohumeral heads at birth, and immediate postnatal electromyogram (EMG) revealing “old injury” pattern have been put forth to denounce the notion that lateral neck traction during the birthing period caused arm paralysis. The obstetrical literature also explains that elective cesarean sections had no reduction in the rate of obstetrical brachial plexus injuries.29,30
In the review of our center's experience and of the literature available, we believe that traction is the predominant force creating these devastating injuries. The pathogenesis of injury, as explained by Server and Taylor, is excessive lateral traction on the head of the infant away from the shoulder during labor.31,32 Experiments from Metaizieu corroborated the above observations and attempted to quantify the forces needed to injure the brachial plexus.33 Although the above-mentioned studies from the obstetrical literature may help to explain certain cases in which brachial plexus lesions occurred in the absence of lateral neck traction, we believe that in the great majority of lesions, lateral traction during birth is the culprit. These beliefs are further supported by a myriad of other compounding factors that have been associated with increased risks of sustaining obstetrical brachial plexus lesions.
Studies performed by Server determined that prolonged labor, forceps delivery, and shoulder dystocia were compounding risk factors in obstetrical brachial palsy injuries.31 Gordon and Gilbert corroborated these findings, and Gilbert further noted that factors such as gestational diabetes, forceps delivery, vacuum extraction, and shoulder dystocia all correlated highly with brachial plexus lesion.19,34,35 In addition, other investigators have found breech delivery, multiparity, macrosomia, and history of a previous child with brachial plexus injury to be independent risk factors for injury.20,21,36
On presentation at The Texas Children's Hospital Brachial Plexus Clinic, patients have already had an established diagnosis of obstetrical brachial plexus palsy. The initial presentation can vary from the classic Erb palsy posture consistent with C5–C6 injury (internal rotation and adduction of the shoulder, elbow extension, forearm pronation, and wrist flexion), to the modified Erb palsy involving C5–C7 (elbow flexion in addition to Erb palsy), to the flail limb and claw hand of a complete C5–T1 lesion.
These patients are placed under the care of a multidisciplinary team including the departments of neurology, neurosurgery, plastic and reconstructive surgery, physical medicine and rehabilitation, and physical/occupational therapy, which have collaborated for over 15 years. This multidisciplinary team is essential to the optimization of our patients' care as has been corroborated by Curtis.37
Patients undergo a meticulous history and physical examination. The obstetrical history is documented, paying particular attention to the perinatal period.38 Parents are asked about the use of any instrumentation; presence of clavicular or humeral bone fractures; phrenic, recurrent laryngeal, or facial nerve paralysis; and Horner syndrome. The history of initial arm function is then obtained, along with any initial treatment modalities and imaging studies obtained. A very detailed physical and neurologic examination is then performed, attempting to grade and document the affected upper extremity function. Concomitant injuries such as Horner syndrome, facial nerve palsy, hoarseness, and contralateral arm function are also documented.
At Texas Children's Hospital, we have employed the use of the British Medical Research Council muscle grading system for the perinatal period (Table 2). The classification system of Mallet is also used to grade muscle function in older patients and in those with residual secondary deformities (Table 3). Numerous other grading modalities have been developed such as those by Gilbert, Tassin, Curtis, and Bashher to overcome the shortcomings of the Medical Research Council muscle grading system and the classification system of Mallet.19,34,37,39,40 However, in our experience, we have had success with the aforementioned scales.
Photographing and videotaping of the upper extremity are also performed at this time to evaluate and record functionality at time of presentation. The patients are recorded attempting to use the affected extremities while placed through several exercises. Incentives such as lollipops, cookies, and toys are all employed to elicit functional movements designed to isolate deltoid, biceps, triceps, forearm, wrist, and hand function. Any specific functionality is then graded and recorded.
Imaging studies in the neonatal period are used to evaluate concomitant injuries.41 A chest roentgenogram is used to examine the position of the diaphragm and elucidate the possibility of a phrenic nerve injury. Presence of a phrenic nerve injury is used to extrapolate the severity of injury, as done by Doi et al.42 The clavicle is also examined to ascertain whether a fracture existed or compounded the injury. Shoulder films and computed tomography are used to evaluate the glenohumeral joint and determine the degree of subluxation or acromion impingement.
Additional radiographic modalities are not initially employed at our center. Other investigators have used computed tomography and magnetic resonance imaging to elucidate the extent of injury. Gilbert, in addition, has employed the myelogram and the computed tomography myelogram to predict the state of the lesions with high certainty.19
Although the technology is improving and several authors do advocate the use of these modalities, we have seen high false-positive and false-negative rates for root avulsion diagnoses in our center (computed tomography false positive 39%, false negative 12%; magnetic resonance imaging false positive 11%, false negative 35%).42,43 The use of these modalities is advocated before secondary reconstructive procedures, when the higher resolution of the glenohumeral joint is needed to determine the ideal therapy required.
Following the initial evaluation, patients are then started with aggressive physical therapy consisting of daily passive range-of-motion exercises implementing the assistance of the family along with professional guidance from therapists. All joints are addressed to maintain the full range of motion and prevent muscular contractures. Shoulder abduction, adduction, flexion, and extension; elbow flexion and extension; forearm supination and pronation; and wrist and hand flexion and extension are essential.44 Patients with flail wrists and elbows must have protective splints to prevent dislocation. However, special care must be taken to prevent pressure sores and contractures with the application of such splints.
Once physical therapy has been implemented, electrophysiologic studies are ordered to document and assist with the evaluation of root avulsion and possible nerve regeneration. The true contribution of implementing this diagnostic modality is a matter of controversy. Literature review has shown that some authors believe the EMG will predict reinnervation of a muscle before any other modality.45,46,47 However, other investigators raise the true validity of the EMG in correlating the results to prognosis.34,39,48,49,50 At our center, this modality is used in situations in which the physical exam has led us to believe that nerve root avulsions have occurred or in situations in which improvement in muscle function has reached a plateau and no further progress is noted.
The true progression of the obstetrical brachial plexus palsy has been a topic of much review and experimentation. Literature review alludes to the fact that no true data have been established to ascertain the recovery of such plexus lesion. However, as history has shown, the great majority of obstetrical brachial plexus lesions encountered do recover spontaneously. Works by Clarke, Michelow, Gordon, and Greenwald have corroborated findings from the Collaborative Perinatal Study in which 95% infants with obstetrical brachial plexus palsy have spontaneous return of function.22,35,51,52 Of these patients, 93% had some function at 4 months.52 However, as is also seen, 5% of infants with obstetrical brachial plexus palsy required surgical treatment. If surgical intervention is performed early, 90% of the patients have significant improvement in function. When treatment was delayed, only a 50–70% rate of improvement was encountered. These findings have led to the continual process of attempting to identify and separate those patients who will have the greatest possibility of recovery with conservative management from those patients who will require surgical intervention.
Attempts at diagnosing the level of injury and extrapolating that information to prognostic factors has been fruitful. Narakas developed a classification of plexus injury linking the Sunderland type of nerve injury with the clinically seen root injuries to predict functional recovery.11 Review of the classification projects that type I lesions, which are the least severe, have the greatest return of function seen early. Type IV and V lesions, correlating with severe global plexus root ruptures and avulsions, as expected, had the poorest prognosis.
Work by Gilbert and Tassin further supported the above findings with their series of 185 patients, in which 36% of lesions were type I according to the Narakas classification, 34% were type II, and 30% were a combination of type III and IV.11,19,34,39 They noted that 100% of type I lesions that were allowed to recover spontaneously resulted in grade 3, by Mallet classification, recovery of shoulder function. None of these patients recovered to Mallet grade 4. Type II lesions fared slightly worse. Of those allowed to recover spontaneously, 70% resulted in grade 3 with 30% showing evidence of poorer grade 2 function.
Although the eventual recovery of spontaneous function has been the rule in a majority of patients, another correlate has been used to predict overall global improvement. In patients who do recover spontaneously, the timing of functional return has been an important prognostic factor. In our center, patients who began to exhibit functional muscle activity in deltoid, biceps, or triceps by 4 months of age had the greatest possibility of near normal recovery by the first year of life.53 These findings are further supported by studies from Boome and Kaye.54,55 Conversely, patients who lacked functional muscle activity in deltoid, biceps, or triceps by 3 months of age had poor functional recovery without surgical intervention, according to Wyeth and Sharpe in 1917 and Taylor in 1920.32
The protocol for timing of obstetrical brachial plexus reconstructions at our center has followed the general recommendations of our current times (Fig. 3). Gilbert and other corroborating investigators have established parameters for surgical exploration and primary nerve grafting at 3 months of age if there is lack of biceps recovery by 3 months, if there is complete palsy with flail arm at 1 month, if there is C5–C6 palsy after breech presentation with absence of function at 3 months, or if there is absent biceps function with upper root injury by 3 months.19
The recommendations from Texas Children's Hospital are to perform primary reconstruction if there is a total root injury without recovery of significant function by 3 months of age; absent motor function in one or more muscle units, especially deltoid, biceps, and triceps, function at 3–6 months of age; or if the patient has muscle grade 1–2 in Mallet classification with no progress at 6 months of age. The limit to primary nerve grafting at our center has been for cases presenting after 12–18 months.25 Prognosis for this group of patients is poor, and the residual deformities are then addressed with secondary reconstructions.56
Approach to the brachial plexus is performed by the neurosurgical service through a supraclavicular incision with extension to the infraclavicular region, if needed, in severe cases or in cases of older patients. The roots, trunks, and early branches of the plexus are exposed above the clavicle, using an incision on the posterior border of the sternocleidomastoid muscle, with a lateral extension superior to the clavicle.
Meticulous care must be taken during the dissection secondary to the fact that the region has had much inflammatory reaction, creating thick, dense adhesions and making dissection difficult. The neurosurgical service then performs an anatomical drawing of the exposed plexus, including roots, trunks, divisions, and any identified neuromas, avulsions, or ruptures, down to the clavicular area, where the plexus then dives into the thoracic inlet (Fig. 4).
Having the exposure needed, direct visualization of the plexus is then performed to assess the potential injury. The general appearance of the plexus is appreciated, taking specific note of the color, size, and caliber of each root, trunk, and distal division. The size of the neuroma is then appreciated to plan for appropriate reconstructive plans. In the great majority of cases, the length of the neuroma is ~3–4 cm in the upper trunk. In our center, the neurorrhaphy is performed in the shortest distance possible, with tension-free coaptation. This procedure is based on findings that postoperative outcomes are inversely related to graft length.57
The physical medicine and rehabilitation service, in conjunction with the neurosurgery service, then performs an intraoperative electrophysiologic assessment. This assessment helps to elucidate and document a root avulsion versus a rupture. Somatic evoked potentials are measured, and distal motor nerve conduction across the area of injury is then measured and graded.
Having assessed the location of injury as well as the extent and the percentage of nerve conduction across the injured segments, the reconstructive strategy formulated intraoperatively is then implemented. Throughout our 15 years of experience at Texas Children's Hospital, we have encountered different case scenarios that have led to the formulation of different strategies for primary nerve reconstructions. These strategies are based on accepted sound surgical principles: minimization of tension on nerve or graft repair site; nerve grafts to maximize fascicular regeneration; functional restoration of elbow first followed by shoulder and hand.
In a preliminary survey of 415 patients who had primary brachial plexus explorations performed at our clinic, we have seen the following profile: C5 and C6 lesions occur at a rate of 43%; C5, C6, and C7 lesions occur 37% of the time; and C5–T1 or isolated C8, T1 lesions occur in 22% of primary explorations.25 The findings have been corroborated by works from Gilbert and Hentz, who noted in their series that the most common anatomic lesion is rupture of C5, C6, and C7, with avulsion of C8 and T1 (64%), followed by rupture of C5 and C6 with avulsion of C7–T1 (35%).19,34,36,39,45
The final reconstructive plan is then made in conjunction with the family, taking into consideration three determining factors: the preoperative clinical condition of the patient (clinical function, age of primary exploration), the electrophysiologic conduction studies, and the level and extent of root involvement (number of roots involved and presence of rupture or avulsion).
In this clinical scenario, the patient has sustained an upper trunk lesion involving the C5, C6, and/or C7 nerve roots with a clinical picture of Erb palsy with the arm adducted and internally rotated, elbow extended, and wrist flexed. Exploration reveals a neuroma in continuity with greater than 50% conduction through the neuroma. The ideology in these patients is that recovery will eventually occur. Intraoperative neurolysis relieving the compression from perineural fibrosis is performed, followed by end-to-side fascicular grafts attempting to augment peripheral regeneration around the neuroma. Nerve grafts are obtained by harvesting the cervical plexus sensory branches, including the great auricular, lesser occipital, transverse cervical, and supraclavicular nerves.
A C5, C6, ±C7 neuroma with less than 50% conduction on electrophysiologic studies in patients who lack deltoid and biceps function by 3 months of age but have preservation of hand function. In this case, the neuroma is sectioned back until healthy nerve fascicles are visualized in the proximal root stumps and the distal trunks (Fig. 4). Nerve grafts are obtained from the cervical plexus and the sural nerve. These nerve grafts are then cut to the appropriate length, 3–4 cm, and secured in place, using 5.5× loupe magnification, 9.0 microsuture, and meticulous microsurgical technique. In this scenario, the C5 nerve root is then grafted to the suprascapular nerve and posterior division of the upper trunk, which is the contribution to the deltoid. The C6 nerve root is then grafted to the anterior division of the upper trunk, which is the contribution to the divisions producing the lateral cords to the musculocutaneous nerve. If the C7 nerve root is involved, the root is then grafted to the middle trunk.
The clinical picture is one of a total plexus injury with a flail arm. Conduction studies failed to show conduction through the neuromas. The reconstructive plan, similar to that outlined in strategy II, is expanded to perform neurorrhaphies from the C8 nerve root to the lower trunk, reconstructing the motor element to the wrist and hand.
The treatment of patients in this category and beyond presents more complicated scenarios secondary to the fact that root avulsion come into play. In these scenarios the number of intraplexal donors becomes limited and in some cases sparse, placing a greater burden on the remaining roots to reinnervate a greater number of muscle units.
The first case is that of a C5 nerve root avulsion with a rupture of the C6 nerve root. The clinical scenario is similar to the second strategy in clinical presentation. However, on a cellular level, the reconstructive plan is much different. The C6 nerve root must now be grafted to the upper trunk in addition to having the spinal accessory nerve neurotized to the suprascapular nerve.
This patient also presents with a clinical scenario of Erb palsy. The upper trunk and middle trunk lack conduction through the neuromas, and therefore grafting is performed. The C5 nerve root is grafted to the suprascapular nerve, the posterior division of the upper trunk, and the middle trunk. This arrangement of neurorrhaphies is performed to enforce synergistic muscle groups: the supraspinatus, infraspinatus, teres minor, deltoid, triceps, and hand extensors. The C6 nerve root is grafted to the anterior division of the upper trunk, attempting to restore elbow flexion.
This scenario is one of lower root, C8–T1, injuries. As was previously documented by Gilbert and Laurent, these rare lesions, when they occur, are most commonly root avulsions.19,34,39,58 The clinical finding is Klumpke palsy with isolated hand paralysis. Patients with avulsion of the C8 and T1 roots must be counseled early about the poor prognosis of lower root injuries. Attempted restoration of hand function may necessitate multiple operative procedures and extended rehabilitation not possible with one operative intervention.
The anterior division of the C7 nerve root is used for neurorrhaphy to the lower trunk. The basis of this transfer is the contralateral C7 nerve transfer performed in cases of total nerve root avulsions. Our experience and that of other surgeons has shown that there are no residual deformities in such a transfer.
These patients are then given an initial recovery period. If no significant hand function, grade 2 or less, is noted, then intercostals are used to neurotize the medial cord—the motor portion of the median and ulnar nerves.
In this particular lesion, the upper roots have been avulsed from the spinal cord. The clinical picture is that of a classic Erb palsy. The difficulty in treating this lesion is that the upper roots are not available as donors, and the C7–T1 roots are uninvolved. The theory then is to ideally restore shoulder and elbow function. The suprascapular nerve is neurotized by the ipsilateral spinal accessory nerve. The axillary nerve is neurotized by redundant branches to the triceps. The thoracodorsal or the subscapularis branches can also be neurotized to the axillary nerve to provide better shoulder function. The biceps function is then addressed with an Oberlin procedure.
Of special note, it has been our experience at Texas Children's Hospital that isolated avulsions of C5–C7 nerve roots without lower trunk lesions is rare. The etiology behind this finding is believed to be that when the plexus sustains enough force to avulse the upper plexus roots, there is sufficient force to avulse the lower roots.
The clinical picture of this lesion is one of a global plexus lesion. Electrodiagnostic studies further elucidate the injury by revealing a rupture of the C5 nerve root with avulsions of the remaining plexus roots. In this scenario, the presence of the C5 nerve root allows neurorrhaphies to the entirety of the upper, middle, and lower trunks. The suprascapular nerve is neurotized by the ipsilateral spinal accessory nerve. The severity of the injury may also require that additional neurotization be performed at a later time, once determination of what functional recovery will occur over the next few months.
The global avulsion of the C5–T1 nerve roots proves to be a complex reconstructive challenge. The approach to these lesions depends on the use of any and all available donor nerves. The experience of our group has relied primarily on intraplexal neurotizations, with augmentation from extraplexal neurotizations.
The goals of reconstruction should be the restoration of at least two muscle groups that will provide functional ability for carrying objects, protecting the limb, and providing stability to shoulder, elbow, and wrist joints. On the basis of these principles, restoration of muscle function should be prioritized. The highest priority is preservation of shoulder function, followed by that of elbow function. This is followed by reconstruction of median-innervated motor and sensory function. Because recovery of the intrinsic function of the hand is usually poor, repair of the ulnar nerve is of low priority.
To stabilize the shoulder and allow elbow flexion, the supraspinatus, deltoids, and biceps take priority in repair. This is accomplished with direct neurotization of the suprascapular nerve by the spinal accessory nerve and transfer of three intercostal nerves to the anterior division of the upper trunk to allow elbow flexion and protective sensation of the hand via the musculocutaneous nerve and the lateral portion of median nerve. Alternatives include direct intercostal neurotization of the musculocutaneous nerve and utilization of contralateral C7 or ipsilateral cervical plexus motor donors to neurotize the axillary and medial nerves.
Following the primary reconstructions, patients remain in the hospital for a period of 23–48 hours. The arm is placed in a splint immobilized to the chest. During this time, the family undergoes training by the occupational therapy and nursing component of the Brachial plexus team to optimally manage postoperative care. Follow up is then arranged at 3 weeks, with removal of the splint. The patient is started on an aggressive physical therapy regimen concentrating on keeping joints supple and preventing contractures. Patients are then evaluated at 3 months' time, given that nerve regeneration should have occurred at least to the level of the deltoid, biceps, and triceps.
Patients who do not show improvement in function undergo EMG studies to elicit any nerve regeneration in muscle function. If activity is noted, patients are continued on physical therapy with the possibility of undergoing electrical stimulation to further augment muscular strength.
In those cases that fail to reveal electrical activity on EMG studies and that have failed to improve clinically, patients then undergo an exploration of the brachial plexus through an axillary approach. The musculocutaneous, median, ulnar, radial, and axillary nerve are all identified. A neurolysis is then performed, and intraoperative nerve conduction studies are obtained.
In our treatment of these patients, a great majority of patients requiring a secondary exploration respond to the neurolysis and require no additional therapy. These patients will then continue with aggressive physical therapy and electrical stimulation when tolerated.
The dilemma, then, are those patients who do not respond to neurolysis. The highest priority is once again placed on stabilization of the shoulder. These goals are accomplished by performing neurotizations from available nerve donors, identified with a handheld electrical stimulation unit. Experience has shown that the thoracodorsal nerve or nerve to the teres major respond to stimulation and are therefore neurotized to the axillary nerve. The intercostal nerves are used for recovery of biceps function through utilization of the remaining sural nerve, if available, or the ulnar nerve can be used as a nerve graft, transposing the distal end of the ulnar nerve to the median nerve. This is followed by reconstruction of median-innervated motor and sensory function through a contralateral C7 nerve transfer.
These patients then are continued on a similar postoperative plan as was employed with the postprimary procedure. Follow up is arranged at 3 months, 6 months, 9 months, and 1 year. Patients are followed in our clinic for as long at patients tolerate the visits.
Review of the literature has shown that implementation of early surgical intervention in patients fulfilling the initial inclusion criteria previously set forth resulted in improved functional recovery.45,59 Gilbert reported on 436 patients who underwent surgical treatment for obstetric brachial plexus palsy, and he reported the following results at 2+years follow up: for C5–C6 lesions, 52% have a grade IV–V Mallet upper extremity function; for C5–C7 lesions, 36% had grade IV–V Mallet; for C5–T1 lesions, 25% recovered useful hand function, with 49% evincing a grade IV Mallet at 4 years' follow up.19
In our experience at Texas Children's Hospital, initial studies revealed that for C5–C6 lesions, primary microsurgical reconstruction resulted in improved deltoid function and biceps function relative to preoperative assessments (Table 3). Deltoid improved from 1% grade IV–V British Medical Research Council (MRC) to 26% grade IV–V. Biceps also increased from 5 to 45% grade IV–V MRC. For C5–C7 lesions, deltoid function went from 0 to 18% grade IV–V MRC; biceps went from 10 to 35% grade IV–V MRC; and triceps went from 20 to 57% grade IV–V MRC. In complete C5–T1 palsy, deltoid function went from 0 to 7% grade IV–V MRC, biceps from 0 to 15% grade IV–V MRC, triceps from 7 to 29% grade IV–V MRC, and hand from 1 to 17% grade IV–V MRC (Table 4).
Expanding our collection of data over the 15 years of experience of the Texas Children's Hospital Brachial plexus team, a total of 223 primary brachial plexus procedures were performed from 1987 to 1995. An additional 335 primary procedures were done from 1995 to 2002. The outcomes of the primary reconstructions were graded using the MRC grading scale. Percentage of patients who recovered good to excellent function (grade IV–V) in deltoid, biceps, and external rotators was 83%. For triceps, functional improvement to a good to excellent classification was 70%. Hand function was noted to improve to grade IV–V for 60% of the patients.
Global functional recovery of the upper extremity is related to the histopathologic type of injury, age, and type of repair.1,3,5,6,9,12 Functional assessments of patients are then performed at every follow-up visit. Patients are photographed, and video is used to grade muscular function. The Mallet and the MRC are once again used to grade functional recovery. Patients are then placed into one of four groups, depending on the initial level of injury. Group I are those patients who sustained upper plexus injuries (C5 and C6). This group is further subdivided to grade shoulder function, triceps, and hand. The second group comprises patients who sustained upper and middle plexus lesions (C5–C7). The third are lower plexus lesions (C8–T1), and the fourth are the global injuries (C5–T1).
The author's experience with obstetric brachial plexus repairs at Texas Children's Hospital has shown the greatest improvement to be in the upper Erb/Duschenne palsy of deltoid, biceps, and triceps, with primary brachial plexus repairs (Table 5). Upper plexus (C5 and C6) injuries had better prognosis than lower plexus injuries (C7, C8, and T1), and the number of avulsions was inversely correlated to the amount of improvement after surgery (Table 6). Similar results have been reported by Gilbert, Terzis, Alnot, and Millesi.19,34,45,60,61 Millesi found useful recovery of the arm (shoulder stability and elbow flexion) in 80% of patients who underwent upper brachial plexus repair.60,61 Alnot reported 50–73% success in return of shoulder function and 74% recuperation of elbow flexion with microsurgical nerve reconstruction in 810 patients.62 In Terzis's experience with 204 patients with brachial plexus injuries, return of function was successful in up to 75% of the suprascapluar nerve, 40% of the deltoid, and 48% of the biceps reconstructions with primary repair of the brachial plexus.45 Worldwide, Merrell and Wolfe3,63 recorded 71% success (≥M3 function) in 965 reported reconstructions of elbow flexion and 73% of 123 restorations of shoulder abduction.
In our experience of secondary reconstructions, we noted a marked improvement in abduction and external rotation of the shoulder after surgery for obstetric brachial plexus injuries (Table 3). The success with these procedures was independent of whether previous therapy was instituted, including primary reconstruction. For patients who developed secondary deformities after a primary reconstruction, the mean passive range of motion for abduction climbed from 10 degrees preoperatively to 91 degrees postoperatively, and external rotation improved from 0 to 49 degrees; 49% had >90 degrees abduction, and 36% had >45 degrees external rotation (versus 5 and 0% preoperatively). For those who had been treated solely with physical therapy before the development of residual deformities, abduction improved from 48 to 94 degrees, and external rotation improved from 7 to 46 degrees; 59% had abduction >90 degrees, and 48% had external rotation >45 degrees, versus 7 and 4% preoperatively.
In summary, the management of obstetrical brachial plexus lesions at Texas Children's Hospital has evolved over our experience to optimize the functional recovery of our patients. This evolution has taught us certain key points that cannot be overemphasized enough: the diagnosis of obstetrical brachial plexus palsy should be established early, and therapeutic options should be implemented immediately; physical therapy is the first tool in the armamentarium available to the treating facility; follow up and reevaluation of functional recovery must be performed continually throughout childhood and adolescence to identify cases that will benefit from early surgical intervention; early surgical intervention, be it primary nerve grafting or primary and secondary reconstructive procedures in conjunction, could prevent the long-term sequelae of residual deformities seen in obstetrical brachial plexus palsy; these residual deformities should not be overlooked, as patients have been shown to suffer psychologically from their disabilities64; and finally, the care of these patients requires a vast amount of support and expertise in fields of neurosurgery, plastic and reconstructive surgery, orthopedic surgery, neurology, physical medicine, and rehabilitation. A multidisciplinary management is highly recommended and is used at our center.