The priorities for reconstruction in the hand are: (l) wrist extension and finger flexion; and (2) recovery of sensation below the elbow in the C 6,7 dermatomes.
Neurolysis and nerve grafting are rarely effective in restoring hand function because the distance to the target muscles is long and reinnervation is rarely achieved.
The use of extraplexal nerve transfers has limited value in restoring hand motor function in total arm-type brachial plexus injury. The technique of intrathoracic harvesting of the phrenic nerve is promising as it allows direct repair to the median nerve in the distal arm.
Sensory reconstruction using a contralateral C 7 or the sensory rami of intercostal nerves (ICNs) can provide some sensation in the median nerve distribution of the hand.
Conventional tendon transfer procedures and tenodesis are useful in reconstruction of hand function, when reinnervation has occurred for select muscle groups.
The use of the pedicled latissimus dorsi is helpful in achieving finger flexion for hook grip.
The use of a single free-functioning muscle transfer (FFMT), combined with dorsal wrist tenodesis, may offer acceptable hand grip function.
Doi’s double free-muscle transfer is useful in restoring some prehension to the hand, especially if done early.
The combined gracilis adductor “triple FFMT” may offer another option in total avulsion brachial plexus injury and can be performed even in late cases.
Care of brachial plexus patients after reconstruction requires commitment by the patient and the family, as well as devotion to many years of dedicated therapy.
A brachial plexus injury (BPI) is a devastating injury, especially when it involves the lower roots with associated loss of hand function. In the developing world, these injuries are frequently caused by motorcycle accidents. BPI can be classified into an upper arm type (C 5,6 with or without C 7 ), a lower arm type (C 8 T 1 ), and a total arm type (C 5-8 T 1 ). Amongst these three types, hand function is mostly preserved in the upper arm type, except when there is involvement of the C 7 root. However, hand function is severely compromised in both the lower arm type and the total arm type. The restoration of hand function is challenging because of the great distance between the proximal plexus injury and the neuromuscular endplate of the forearm and hand muscles (approximately 40–60 cm). Even if axon regeneration proceeds smoothly (1 mm/day), it would take more than 18 months to reach the target muscle, by which time irreversible atrophy of the muscle would have occurred. The results of early nerve reconstructive procedures (neurolysis, nerve grafting, and nerve transfers) for restoration of hand function are poor when compared to the results obtained at the shoulder or the elbow. Therefore secondary procedures like tendon and muscle transfers are frequently used to restore hand function.
This chapter will briefly discuss the role of nerve reconstructive procedures and elaborate the technical details of the secondary procedures used to restore hand function.
Nerve reconstructive procedures
Microsurgical neurolysis of the brachial plexus is a controversial procedure with uncertain results. Lesions in continuity with external compression can be treated successfully by neurolysis. However, neurolysis is rarely effective in the recovery of hand function, except when a conduction block (neuropraxia) does not resolve spontaneously. When there is fibrosis of the perineurium and the epineurium, an incision over the nerve sheath can be performed. These surgical procedures can have a good result when the fascicular pattern and the endoneurial tissue are preserved ( Figure 20.1 ). If there is fibrosis in the fascicles or a loss of the fascicular pattern, neurolysis will not be successful and resection of the involved segment and restoring nerve continuity by nerve grafting is indicated. In many cases, it is not easy to differentiate between a salvageable nerve that can be preserved and a damaged nerve that should be resected. In such situations, it is better to resect and graft than to attempt neurolysis.
Direct neurorrhaphy is a successful treatment for a cleanly transected nerve, such as with a stab wound. In a typical brachial plexus avulsion lesion with loss of nerve substance, a direct neurorrhaphy is not possible and should not be attempted ( Figure 20.2 ). Narakas reported a large series of 800 BPI cases that were treated surgically. Primary suture was possible in only 21 cases (5 incised wounds and 16 iatrogenic injuries).
Most BPI cases today are caused by high-energy trauma with nerve avulsion injuries that are not suitable for primary repair. In a reported series of 386 cases of surgical procedures for BPI, only 5 primary repairs were performed. In patients with a partial nerve laceration after a stab or an incised wound, it is relatively easy to reconstitute the general fascicular organization of the nerve trunk by end-to-end suture, and the prognosis is usually good. The primary suture of a transected nerve trunk/cord should be performed immediately to avoid neuroma formation and the difficulty associated with intraneural dissection ( Figure 20.3 ). With regard to C 8 T 1 lesions of lower arm-type BPI, primary suturing has not resulted in satisfactory outcomes of hand function and additional procedures such as tendon grafting or free functional muscle transfers are always indicated.
The results of nerve grafting to improve hand function in BPI patients are not satisfactory. In the total arm-type BPI, it is rare to obtain usable function in the hand by nerve graft surgery. In the majority of cases, patients merely obtain a “shovel hand” or “paperweight hand,” which is only useful to stabilize an object on a table. For patients with lower arm-type BPI, decisions regarding surgery should be based on the clinical examinations and additional diagnostic studies. When pseudomeningoceles appear on the preoperative magnetic resonance imaging or myelography studies, nerve graft procedures should not be performed because outcomes are invariably poor. In this situation, muscle/tendon transfer or a free functioning nerve transfer (FFMT) should be considered to reconstruct the hand. However, if the magnetic resonance imaging or myelography studies are normal in lower arm-type BPI cases, but spontaneous regeneration (by electromyogram and nerve conduction velocity) has not occurred, exploration surgery over the supraclavicular region is appropriate to assess the suitability of using nerve grafts. The early use of nerve grafts in such cases has restored extrinsic hand function in the M2–M3 range, but intrinsic hand function cannot be recovered ( Figure 20.4 ).
Neurotization for motor functions
The distal nerve stump may be reinnervated either by another root within the plexus (intraplexal neurotization) or by a healthy nerve that resides outside the plexus (extraplexal neurotization). It is almost impossible to gain useful hand function by neurotization because of the distance involved and limited donor proximal nerves. In addition, axon count in extraplexal motors is significantly lower when compared to the normal nerve roots. An intraplexal transfer, proposed by Chuang, uses a vascularized ulnar nerve graft to bridge the gap between an ipsilateral C 5 or C 6 root and the median nerve in patients with total arm-type BPI (C 5 and/or C 6 rupture associated with C 7,8 T 1 three-root or C 6-8 T 1 four-root avulsion). The transfer of ICN’s to the median nerve to restore hand function is an extraplexal transfer ( Figures 20.5 and 20.6 ).
To overcome the limited number of available donor nerves and insufficient myelinated axon fibers in the donor nerves, Gu from China proposed using the contralateral C 7 root to achieve useful hand function. This involves two stages. In the first stage, the pedicled ulnar nerve graft raised on the paralyzed side is coapted to the contralateral C 7 nerve root. The second stage is done 8–12 months later and the distal end of the reinnervated ulnar nerve graft is sutured to the avulsed plexus, giving priority to the musculocutaneous nerve, the median nerve, the axillary nerve, and then the other nerves ( Figures 20.7 and 20.8 ). Chuang prefers to transfer the ulnar nerve with the ulnar artery as a free flap, a procedure he calls the supercharged contralateral C 7 . In Chuang’s series, 8 out of 15 patients needed a secondary procedure with FFMT for hand function. Chuang also noted that sectioning of the C 7 root created a temporary weakness of the triceps and paresthesia in the superficial radial nerve territory over the dorsum of the hand in the donor upper limb.
A sound knowledge of the anatomy of the C 7 root, and muscles innervated by it is essential before considering this transfer. A transection of the posterior division of C 7 produces a weakness of the radial nerve-innervated muscles, especially the triceps and the wrist and finger extensors; however, there is no paralysis. These muscles regain their original strength by internal sprouting. The transection of the anterior division of C 7 slightly weakens the pectoralis muscle without visible loss of function and results in some sensory loss in the thumb and index finger. This sensory loss may not be tolerated well by some patients. In order to decrease the morbidity of C 7 transfer, Millesi emphasized that a contralateral C 7 transfer should never be performed immediately. Instead, Millesi recommends putting ligatures over the anterior and posterior divisions of the C 7 root, and then examining the patient carefully the next day. This would allow assessment of loss of function. A major loss of sensory function would indicate that the anterior division should not be cut and only the posterior division should be used. If there is significant loss of motor function, the C 7 transfer is not performed at all.
Waikakul et al., in their review of the results of using the anterior portion of the contralateral C 7 to the median nerve for total arm-type BPI, got wrist flexion (≧M3) in 29%, and finger flexion (≧M3) in 21% of cases. In Songcharoen’s series, only 29% of patients achieved M3 or M4 finger flexion by contralateral C 7 transfer. It is obvious that hand function in total arm-type BPI patients could not be reconstructed satisfactorily by only a contralateral C 7 to median nerve transfer ( Figure 20.9 ). The failure of contralateral C 7 transfers is primarily due to the long distance between the C 7 donor nerve coaptation site and the hand muscles innervated by the median nerve.
A new method of neurotization to restore hand function has been proposed. This requires a harvest of the full length of the phrenic nerve (via a thoracotomy) that is transferred to the median nerve in the distal arm. Zhao et al. performed an anatomical study on 17 cadavers and reported clinical application in 1 patient. They found that the branches of the median nerve in the distal arm were consistently organized into three fascicular groups located in the anterior, middle, and posterior parts of the median nerve trunk. The anterior fascicular group was composed of the branches to the pronator teres and the flexor carpi radialis. The posterior fascicular group was composed of the anterior interosseous nerve and branches to the palmaris longus, whereas the middle fascicular group was made up of the branches to the hand and the flexor digitorum superficialis (FDS). They reported a patient with total arm-type BPI where they transferred the phrenic nerve to the posterior fascicular group of the median nerve. The patient obtained M4–M5 recovery of flexor pollicis longus and the flexor digitorum profundus (FDP) of all four digits at 16 months after surgery. The harvest of the full length of the phrenic nerve allowed transfer to the level of the distal arm, thus shortening the distance and the time needed for reinnervation of the forearm muscles. This technique, although quite promising, has not gained popularity in BPI reconstruction. Further studies and additional clinical cases are needed to assess hand motor recovery and donor site morbidity, such as pulmonary function deterioration after phrenic nerve harvest.
The use of extraplexal transfers to neurotize lower trunk roots (C 8 T 1 ) to regain intrinsic function has not worked. The neurotization of the ulnar or the median nerve using ICNs may result in some wrist function and protective hand sensation, but patients do not regain useful hand function.
Neurotization for sensory function
Most reconstructive procedures for total arm-type BPI have focused on the restoration of motor function, and few studies have addressed restoration of sensory function. The lack of cutaneous perception on the reconstructed hand will inevitably affect the prehensile function of the hand, such as lifting or holding objects in daily activities, because of a lack of tactile sensory feedback. Furthermore, a reconstructed hand with motor recovery, but without sensory restoration, will not give the patient protective sensation. Several types of sensory reconstructive procedures have been reported in the literature, such as primary nerve grafting on the trunk level, and intraplexal or extraplexal neurotization.
Three different types of nerve transfer techniques have been described using C 5 nerve root, ICN, or the supraclavicular nerve as the donor nerve. When C 5 was used as the donor nerve, a free vascularized nerve graft from the sural nerve or from the ulnar nerve in the paralyzed forearm was harvested, and then transferred to the lateral root of the median nerve at the infraclavicular level of the affected side. When the ICN was used as the donor nerve, the lateral cutaneous branches of the third through sixth ICN were assembled with fibrin glue, and then directly transferred to the median nerve around the axillary region. Kotani et al. reported the results of 15 cases in which the ICN had been used for sensory reconstruction. Limited sensation of touch and pain with paresthesia were obtained in 11 cases. In 1977, Millesi reported the recovery of protective sensation in 15 of 18 cases who received ICN nerve transfers to the median nerve for total arm-type BPI. Hattori et al. reported 17 cases receiving ICN transfer to either the median nerve or the ulnar nerve for sensory reconstruction of total avulsion BPI. All of their patients obtained protective sensation of the hands, and perceived at least the 6.65 filament in the median or ulnar nerve territories. Eight patients perceived heat sensation and 13 patients perceived cold sensation, but none regained two-point discrimination.
When the supraclavicular nerve was used as the donor nerve, two or three branches of this nerve were assembled with fibrin glue, and then directly sutured to the median nerve component of the lateral cord at the clavicular level without nerve graft. They reported an acceptable sensory recovery in 12 of 15 patients, and concluded that even the limited sensory recovery was useful for completely anesthetic BPI patients. The contralateral C 7 transfer has been reported to have satisfactory sensory recovery. Although this method was originally designed for reconstructing motor functions, reports in the literature suggest that sensory recovery was better than motor recovery. The reason for poor recovery of motor function following contralateral C 7 transfer is muscle atrophy. However the sensory end-organs are preserved for much longer and, despite the long distance between the C 7 donor nerve coaptation site and the docking site of the vascularized ulnar nerve coaptation with the median nerve, sensory reinnervation is usually satisfactory.
We recommended using either ICNs or supraclavicular nerves as donor nerves to transfer to the median nerve without the need for nerve grafts. The C 5 nerve root is often used with nerve grafting to the axillary, suprascapular, or musculocutaneous nerve for motor function rather than connecting it to the median nerve for sensory function.
Tendon and muscle transfers
Tendon transfer techniques are useful for restoring hand function in patients with incomplete injuries or those who have partial recovery of function following nerve grafting or nerve transfer procedures. It should be borne in mind that any tendon that is transferred for another function will lose one grade of power after the transfer. For example, a M4 muscle will become a M3 muscle after transfer.
In upper-arm-type BPI with associated C 7 injury, patients usually present with loss of wrist and finger extension. The functional needs of the hand after a C 7 injury (radial nerve palsy) are: (1) wrist extension; (2) finger and thumb extension; and (3) thumb proximal stability. For wrist extension, a transfer of the pronator teres to the wrist extensors (extensor carpi radialis longus and extensor carpi radialis brevis) can be considered if there has been adequate recovery of pronator teres function. For finger extension and thumb extension, the traditional transfers are the flexor carpi ulnaris (FCU) to the extensor digitorum communis (EDC), and the palmaris longus to the extensor pollicis longus (EPL) are recommended. This FCU transfer often results in a slight radial deviation of the hand at the wrist. If the patient has significant radial deviation at the wrist, the insertion of the extensor carpi radialis longus should be transferred to the extensor carpi ulnaris, or the FCU tendon transfer should not be done. The finger and thumb extension can also be reconstructed by transferring the flexor FDS tendons to the long and ring fingers. The ring-finger FDS is attached to the EDC, similar to the FCU transfer, and the long-finger FDS is attached to the EPL. An alternative method is to transfer the palmaris longus to EPL, and the flexor carpi radialis to the EDC ( Figure 20.10 ).
Tenodesis procedures have been used for restoring thumb stability and correcting flexion contractures seen after a single FFMT. For proximal stability of the thumb, the extensor pollicis brevis is mobilized from the first dorsal compartment and tenodesed to the palmaris longus. If the palmaris longus has been used for thumb extension reconstruction, split flexor carpi radialis tenodesis to both abductor pollicis longus and EPL tendons can be considered.
When a single FFMT is used to restore elbow and digital flexion, wrist and digital extension is usually not reconstructed and most of the reconstructed fingers are in a state of flexion contracture ( Figure 20.11 ). Although some dynamic splintage devices have been developed to prevent flexion contractures of the fingers and wrist, tenodesis of wrist and finger extensor tendons is a useful procedure in these cases. The tenodesis procedures are always performed after the FFMT has recovered motor function. The position and angle of the wrist and hand in the tenodesis procedure depends on the degree of flexion deformity and the tightness of the soft tissue. A tenodesis on antagonist tendons may also enhance the efficiency of a previously transferred muscle.
Muscle transfers can be pedicled or free. The latissimus dorsi is the most frequently used pedicled muscle transfer. Three variations of free muscle transfers have been described: (1) single gracilis FFMT; (2) two-stage double gracilis FFMT; and (3) two-stage triple muscle FFMT.
Pedicled latissimus dorsi muscle flap transfer
For the past 30 years, the latissimus dorsi muscle has been used to restore elbow flexion, but it is not often used to restore digital flexion. The restriction in the use of the latissimus dorsi muscle may be related to inadequate perfusion of the distal third of the muscle by the dominant vessel – the thoracodorsal artery. The perfusion of this portion of the muscle relies on the perforating branches of the intercostal and other collateral arteries. However, in the series published by Gousheh et al., the thoracodorsal artery was sufficient in nourishing the entire length of the latissimus dorsi muscle, and the pedicled latissimus dorsi muscle could be harvested for more than 40 cm, which is long enough to reach the target tendons over the forearm and wrist.
The indications for pedicled latissimus dorsi muscle transfer in BPI include:
lower arm-type BPI, with preserved latissimus dorsi muscle function
permanent paralysis of the hand due to irreparable lesions to the median, radial, and ulnar nerves above the elbow (Millesi type 4, infraclavicle lesion)
combined large soft-tissue loss with BPI that requires simultaneous wound coverage and tendon reconstruction
total arm-type BPI: although the latissimus dorsi motor function is lost in total arm-type BPI, hand function can be reconstructed by a pedicled latissimus dorsi transfer followed by immediate neurotization by ICN transfer to reinnervate the paralyzed latissimus dorsi.
Under general anesthesia, the patient is placed in the lateral decubitus position with the aid of a well-padded Mayo stand. The ipsilateral arm and lateral chest and back are prepared and draped. The length of latissimus muscle needed is determined by the distance from the tendinous insertion of the latissimus dorsi muscle at the upper humerus to the flexor/extensor tendon of hand over mid-forearm level. The route for passing the muscle is marked. It is better to pass the muscle through a subcutaneous tunnel in the arm and forearm for smooth gliding of the latissimus dorsi muscle. A bulky flap with excessive flap width should be avoided, as tunneling will be difficult and skin grafting will impede muscle gliding.
An incision running from the posterior axillary fold to the midpoint of the iliac crest is made ( Figures 20.12 and 20.13 ). The skin is sutured to the muscle to prevent possible shearing of the cutaneous component of the flap because the skin is supplied by the perforating vessels coming through the muscle ( Figure 20.14 ). The muscle is freed from its lateral, distal, and medial borders. A longer length of the muscle can be obtained by including a small portion of the gluteal aponeurosis or the fascia attaching the muscle to the spinous process along with the distalmost portion of the muscle. The flap is folded over, and its anterior surface is separated from the chest wall. The neurovascular pedicle (thoracodorsal artery, vein, and nerve) is identified and carefully dissected as far proximally as the circumflex scapular vessels to maximize the reach of the muscle flap. The branches to the serratus anterior muscle are ligated. This will enable the muscle to reach the forearm ( Figure 20.15 ). At this point, the entire muscle flap can be lifted and the neurovascular pedicle on the upper part of this muscle is revealed. It is not necessary to divide the muscle insertion on the humerus as the limiting factor to movement is the vascular pedicle. However, dividing the muscle insertion from the humeral intertubercular sulcus and reattaching it to the acromion (a bipolar transfer) will give it a better biomechanical advantage as it places the muscle in a straight line over the arm and forearm. If this latissimus dorsi flap is being used for functional reconstruction of total arm-type BPI, two ICNs can be harvested at the same time, and neurotization is performed by coapting the two ICNs to the thoracodorsal nerve.
A subcutaneous tunnel is created between the axilla and the forearm and the muscle is carefully delivered through the tunnel, taking care not to twist the vascular pedicle ( Figure 20.16 ). Pronator teres or the wrist flexors may serve as a pulley for finger flexion reconstruction, whereas brachioradialis or wrist extensors may serve as a pulley when the latissimus dorsi is used to reconst the finger extensors ( Figure 20.17 ). The use of a pulley prevents the development of a bowstring deformity at the elbow. The distal fascial portion of the latissimus dorsi muscle is sutured to the FDP ( Figure 20.18 ) or the EDC. The fingers are maintained either in a flexed position (for digital flexor reconstruction) by temporary, multiple Kirschner wire fixation or in an extended posture (for digital extensor reconstruction) by postoperative splinting/casting. It is recommended to keep the elbow at 90° flexion, when suturing the latissimus dorsi to the tendons. The wound over the donor site is sutured in layers and closed suction drainage is inserted. Active and passive range-of-motion finger exercises are started after K-wire or cast removal 3–4 weeks after surgery.
Most patients with a pedicled latissimus dorsi transfer for hand function may obtain immediate finger flexion or extension. The only exceptions are patients with total arm-type BPI where the latissimus dorsi was neurotized using ICNs. These patients need 6–12 months to achieve M3–M4 motor recovery, similar to the FFMT recovery time. Gousheh et al. reported 28 cases who received this type of reinnervated latissimus dorsi flap transfers; all obtained active finger motor function. The mean flexion of the metacarpal joint was 38°, of the proximal interphalangeal joint was 85°, and of the distal interphalangeal joint was 30°. The mean metacarpal extension was 33°, and all patients could manage their daily activities. In our series of 35 patients (32 males and 3 females) using the latissimus dorsi pedicle flap for finger motor function, 2 patients had neglected lower-arm-type BPI, 5 patients had Millesi type 4 infraclavicle lesions, and 28 patients had total arm avulsed BPI. All the latissimus dorsi muscle flaps were transferred for the purpose of obtaining finger flexion. All flaps survived and obtained at least M3–M4 grip function of the hand. In the 7 patients who had preserved motor function of the latissimus dorsi, the average recovery of metacarpal flexion was 42°, proximal interphalangeal flexion was 75°, and distal interphalangeal flexion was 35° ( Figure 20.19 ). Twenty-eight patients who received 2 ICN neurotization (using the third and fourth ICN, or fourth and fifth ICN) had finger flexion recovery M3–M4 at an average of 9.5 months (5.5–14 months) after flap transfer. The average recovery of metacarpal flexion was 36°, proximal interphalangeal flexion was 61°, and distal interphalangeal flexion was 20°. It was obvious that the functional recovery was slightly better in latissimus dorsi flap transfers that had intact thoracodorsal nerve function than in those who needed neurotizations, although there was no statistical difference in the final motion. Doi et al. reported his series of 6 patients who received latissimus dorsi flap transfer for finger flexion, and the spinal accessory nerve was used to neurotize the thoracodorsal nerve. Finger flexion was achieved in all of these patients, and the grasp power and finger control improved continuously. Doi et al. concluded that the functioning latissimus dorsi flap transfer appeared to be promising for the restoration of hand function in BPI patients.