- Email: [email protected]
Motor recovery and the breathing arm after brachial plexus surgical repairs, including re-implantation of avulsed spinal roots into the spinal cord
ARTICLE IN PRESS MOTOR RECOVERY AND THE BREATHING ARM AFTER BRACHIAL PLEXUS SURGICAL REPAIRS, INCLUDING RE-IMPLANTATION OF AVULSED SPINAL ROOTS INTO THE SPINAL CORD M. HTUT, V. P. MISRA, P. ANAND, R. BIRCH and T. CARLSTEDT From the Peripheral Nerve Injury Unit, Royal National Orthopaedic Hospital, Stanmore, UK and the Peripheral Neuropathy Unit, Imperial College, Hammersmith Hospital, London, UK
Forty-four patients with severe traction brachial plexus avulsion injuries were studied following surgical repairs. In eight patients, re-implanting avulsed spinal roots directly to the spinal cord was performed with other repairs and motor recovery in the proximal limb was similar to that achieved by conventional nerve grafts and transfers when assessed using the MRC clinical grades, Narakas scores, EMG and Transcranial Magnetic Stimulation (TMS). Thirty-four of the 37 patients had co-contractions of agonist and antagonist muscle groups. Spontaneous contractions of limb muscles in synchrony with respiration, the ‘‘breathing arm’’, were noted in 26 of 37 patients: in three patients, the source of the breathing arm was from spinal cord re-connection, providing evidence of regeneration from the CNS to the periphery. Our study shows that re-connection of avulsed spinal roots can produce good motor recovery and provides a clinical model for developing new treatments which may enhance nerve regeneration. Journal of Hand Surgery (European Volume, 2007) 32E: 2: 170–178 Keywords: brachial plexus injury, surgical repair, re-implantation, breathing arm, electrophysiology
Avulsion of the brachial plexus can be the most serious of all injuries to the peripheral nerves, with grave impairment of the quality of life (Bertelli and Ghizoni, 2003; Birch, 2003; Nagano, 1998; Terzis et al., 2001). Most patients are young men who have been involved in road traffic accidents and the incidence of these injuries is increasing (Birch et al., 1998). Usually both ventral and dorsal roots are involved and the patient is subjected to paralysis and sensory dysfunction, with numbness in the limb in conjunction with extreme, intractable pain (Berman et al., 1998; Birch et al., 1998). In general, the results of surgical repairs of brachial plexus injuries are influenced by the level and extent of injury, the type of surgical procedure (Samardzic et al., 2002) and, above all, by delay before repair. There have been significant recent advances in the surgical treatment of traction brachial plexus injury. The commonest means of repairing spinal cord root avulsion injuries in current practice is to transfer an intact neighbouring nerve to the distal stump of the damaged nerve, so restoring some motor and sensory function (Bentolila et al., 1999; Birch et al., 1998). The novel surgical strategy of re-implanting, or re-connecting, avulsed spinal roots, via nerve grafts, directly to the spinal cord (termed ‘‘re-implants’’) has been performed in patients with severe brachial plexus injury (Carlstedt et al., 1995, 2000, 2004). The main purpose of this cross-sectional study was to evaluate motor recovery after different surgical repairs, viz. conventional nerve grafts for plexus injuries distal to
the dorsal root ganglion and nerve transfers and/or re-implantation for spinal cord-root avulsion injuries. We have used the MRC (Medical Research Council) grades for muscle power, Narakas scores (Narakas and Hentz, 1988), Electromyography (EMG) and Transcranial Magnetic Stimulation (TMS). We also recorded phenomena related to motor recovery, viz. the cocontraction of agonist and antagonistic muscles and ‘‘the breathing arm’’. The ‘‘breathing arm’’ refers to contractions of arm muscles synchronously with respiration. This phenomenon has been described in patients who have had a variety of operations for the repair of brachial plexus injury, especially after intercostal nerve transfer (Swift, 1994; Takahashi, 1983). In this study, we looked, for the first time, for the ‘‘breathing arm’’ in patients with complete brachial plexus avulsion injuries repaired by re-implantation, which would provide evidence for regeneration of motoneurons from the central nervous system (CNS) to the peripheral nervous system (PNS), as, in some patients the re-implanted root or graft would be the sole conduit for the phrenic spinal motoneurons to the periphery. PATIENTS AND METHODS Forty-four patients who had sustained severe brachial plexus injury with spinal nerve root avulsions were studied after written informed consent and local ethics committee approval. Patients with associated spinal cord or brain injury, injury to the proximal major blood 170
ARTICLE IN PRESS MOTOR RECOVERY AND THE BREATHING ARM
vessels or double level lesions (as revealed during explorative surgery) were excluded. Not all patients had complete paralysis. The serratus anterior was functioning in 27 out of 44 patients, one patient had no lesion of C8 or T1 and three had lesions other than avulsions: two had rupture and one had a lesion-incontinuity distal to the dorsal root ganglion. Avulsion or intraspinal nerve root injury was confirmed by computerised tomography, (CT)-myelography, pre-operative electrophysiology, direct observation of the exposed brachial plexus during operation and inspection of the spinal cord (where applicable). In cases of complete avulsion, the ventral and dorsal roots, with dorsal root ganglia, were totally detached. When the pertinent spinal nerve at surgery was found to be ‘‘in situ’’, a diagnosis of avulsion was made from preoperative findings together with results of scanning and peri-operative electrophysiology. The operations were performed at the Peripheral Nerve Injury Unit, Royal National Orthopaedic Hospital. For surgical techniques, see Birch et al. (1998) and Carlstedt et al. (1995, 2000). Most patients had more than one type of operative repair. Conventional graft repair of ruptured spinal nerves was combined, if possible, with transfers of the accessory nerve, the accessory phrenic nerve, the supraclavicular nerves or the intercostal nerves. Spinal cord replantation of avulsed roots was performed in cases with four or more roots avulsed, no later than 1 month after injury, for neurobiological reasons – death of nerve cells with delayed repair (Bergerot et al., 2004). For the re-implantation procedure, the patient was positioned laterally on the operating table to allow a simultaneous approach to the extra and intraspinal parts of the brachial plexus (for details regarding surgical exploration, with illustrations, see Carlstedt et al., 2000). After a dissection through the neck muscles, respecting the accessory nerve, a C5 to C7 hemi-laminectomy was performed. After the dura had been opened, the intraspinal injury was verified. Nerve grafts were pulled through the intervertebral foraminae and positioned in the spinal canal. Stay sutures were applied to the denticulate ligament in order to gently rotate the spinal cord. Small slits were produced in the pia mater in the anterolateral aspect of the spinal cord to receive the ends of the nerve transplants, which were inserted deep to the surface. Tissue glue was applied to maintain the attachment of the nerve grafts to the spinal cord. The distal ends of the nerve grafts were connected to the motor parts of the avulsed spinal nerves. Patients were operated on up to 31 months after injury by the same two peripheral nerve surgeons (T.C. and R.B.). Patients who had sustained clinically complete brachial plexus injuries and had no surgical nerve repair, because of late referral or the wish of the patient, served as controls (n ¼ 10). A diagnosis of avulsion was made from spinal imaging, electrophysiology
171
Table 1—Mean numbers (SEM) of lesions in patients treated by different methods Method of repair
Avulsion
Rupture
Total no. of lesions
Graft and nerve transfers (n ¼ 26 patients) Re-implantation and other repairs (n ¼ 8 patients) No surgical repair (n ¼ 10 patients)
2.9 (0.2)
1.6 (0.1)
4.4 (0.2)
4.3 (0.3)
0.7 (0.4)
4.8 (0.2)
3.8 (0.4)
0.1 (0.1)
3.9 (0.4)
and preservation of dermatomal skin flare responses (Berman et al., 1998). The patients studied were all men. The age at which the injury was sustained ranged between 17 and 38 years, with a mean age of 24 years, for the graft and nerve transfer group; 18 and 36 years, with a mean age of 27 years, for the re-implantation and other repairs group; 17 and 45 years, with a mean age of 29 years, for the no repair group. The age of the patients at the time of surgery ranged between 17 and 38 years, with a mean age of 24 years, for the graft and transfer group and 18 and 36 years, with a mean age of 27 years, for the reimplantation and other repairs group. The mean numbers of lesions in patients treated by the different methods and in the patients with no repair are summarised in Table 1. The delay between injury and surgery is shown in Table 2. Clinical motor assessment All the patients underwent a detailed neurological examination. Power in the upper limb was assessed using the grading recommended by the Medical Research Council (Medical Research Council, 1976). The functional recovery of the shoulder and elbow was also assessed using Narakas’ functional scoring system (Narakas and Hentz, 1988), which evaluates global functional recovery, especially at the shoulder and elbow, rather than recovery of individual muscles. EMG examination Concentric needle (TECA disposable needle; Oxford Instruments Medical, Surrey, UK) EMG examination was performed with a Keypoints work station (Medtronic Functional Diagnostic A/S, Skovlunde, Denmark). Muscles were sampled using conventional electromyographic methods, including insertional and spontaneous activities and motor unit action potentials (MUPs). In addition, motor unit activity was also observed in relationship to respiration and coughing. Multi-channel simultaneous surface EMG recordings of
ARTICLE IN PRESS 172
THE JOURNAL OF HAND SURGERY VOL. 32E No. 2 APRIL
2007
Table 2—Delay between the injury, surgery and assessment Methods of repair
Graft and nerve transfers Re-implantation and other repairs
Delay between injury and surgery (weeks)
Duration of follow-up (years)
Mean (SEM)
Max
Min
Mean (SEM)
Max
Min
6 (1) 1 (0.2)
122 19
1 3
6 (1) 3 (0.4)
19 4
1 2
MAX ¼ maximum, MIN ¼ minimum.
Single pulse TMS was performed using the Magstims200 Monophasic stimulator (Novametrix Medical Systems Ltd., Whitland, UK). The output was connected to a Keypoint work station (Medtronic Functional Diagnostic A/S, Skovlunde, Denmark) to record responses. TMS is a technique in which the motor cortex is stimulated by a brief electromagnetic stimulus delivered via a hand held magnetic stimulation circular coil (a high power 90 mm coil). The coil is placed over the head on the area over the motor cortex. The evoked motor response is detected by electrodes placed over the relevant upper limb muscles and the motor action potentials recorded on an EMG machine. Facilitation was used whenever possible (i.e. the patient was asked to attempt to contract the muscle from which the recording was made, as this can improve the amplitude of the recorded motor response). Both the injured and the contra-lateral (intact) limb were studied. Statistics The Mann–Whitney U-test and the Kruskal–Wallis test were used. P-values o0.05 were considered to be significant. Data are presented as mean7SEM (standard error of mean) unless otherwise stated.
RESULTS The recovery of motor function from nerve grafts and transfers and re-implantations with other repairs are presented in Figs 1A and B and in Table 3. There was no statistically significant difference of pectorals, deltoid, biceps and triceps strength in patients who had nerve graft and transfer in comparison with reimplantation with other repairs, using the Mann–Whitney U-test. The results of deltoid and biceps motor recovery are shown in Figs 1A and B. In some patients,
Motor power in MRC grading
Magnetic stimulation
A
5
4
3
2
1
0 Pectorals
Deltoid
Biceps
Triceps
B 5 Motor power in MRC grading
the pectoral muscles, the deltoid, biceps and triceps muscles were made in order to determine the presence of co-contractions of limb muscles with voluntary movements and respiration.
4
Pectorals
Biceps
Deltoid
Triceps
3 2 1 0 Muscles
Fig 1 Recovery of motor function: (A) from graft and nerve transfer and (B) from re-implantation with other repairs.
co-contractions made it difficult to assess muscle power accurately, and, hence, some patient data points are not shown. In 22 of 26 of the graft patients, deltoid muscle power was assessed and 11 of these 22 patients regained power greater than 3 of the MRC scale. In 21 of 26 of the graft patients, biceps muscle power was assessed and seven of these 21 patients regained power greater than 3 of the MRC scale. In the re-implantation group, only
ARTICLE IN PRESS MOTOR RECOVERY AND THE BREATHING ARM
173
Table 3—The recovery of motor function after different surgical procedures Motor power (MRC grading)1
Methods of surgical repair
Graft and nerve transfers Re-implantation and other repairs P-values
Pectorals
Deltoid
Biceps
Triceps
3.4 (0.2) 2.7 (0.5)
2.4 (0.3) 1.5 (1.0)
2.0 (0.3) 2.0 (0.5)
2.1 (0.4) 2.2 (0.6)
0.2
0.3
1.0
1.0
MRC, Medical Research Council. 1 Values are mean7SEM.
Table 4—Breathing arm or co-contraction Number of patients (n ¼ 37) Breathing arm and co-contraction Breathing arm only Co-contraction only No breathing arm or co-contraction
25 1 9 2
one of four deltoid muscles regained power 3 of the MRC grading. Two of five patients regained biceps muscle power greater than 3 of the MRC grading. The Narakas’ functional scores for the shoulder and elbow for patients repaired by nerve graft and transfers were 3.4 (SEM 1.1) and 2.3 (SEM 0.5) and for patients repaired by means of re-implantation with other repairs were 3.5 (SEM 1.1) and 1.8 (SEM 0.7). There was no statistically significant difference in the Narakas’ functional score between the two groups for the shoulder or elbow. The duration after the injury to assessment and the duration of delay between injury and the operation are shown in Table 2. There were no statistically significant differences in the mean interval between the injury and surgery, or between the mean duration from injury to assessment, between the two groups. Thirty-seven patients were studied for the ‘‘breathing arm’’ and limb muscle co-contraction. As can be seen in Table 4, 25 patients were found to have both the breathing arm and co-contraction, while nine patients showed no evidence of the breathing arm but had limb muscle co-contractions. One patient showed evidence of the breathing arm, but not co-contractions. Two patients did not have either the breathing arm or cocontractions. In three patients, the source of the breathing arm could only have been from re-implanted ventral roots (see case study below). In the rest of the patients, the source of the breathing arm could have been from the C5 nerves, the intercostal nerves or the accessory phrenic nerve. The duration after the injury to follow-up assessments for the patients with and without the breathing arm was 6.7 (SEM 1.5) and 5.2 (SEM 1.5) years, respectively. This phenomenon was observed as early as 13 months
after surgery and was recorded up to 19 years after injury. EMG examination of the affected muscles generally showed evidence of partial denervation with reinnervation. Motor unit potentials (MUPs) firing synchronously with respiration (inspiration) in quiet, or voluntary, breathing were observed in several cases. The same motor units could usually be activated by volitional contraction. MUPs firing simultaneously were seen in groups of muscles (agonists and antagonists) that were co-contracting (Figs 2A and B). The amplitudes of the motor responses to TMS, where present, were smaller when compared to the contralateral arm, and the latencies of the responses were prolonged (Figs 3A and B). The absolute values of amplitude and latency of biceps response are shown in Figs 4A and B. There was no statistically significant difference of amplitude and latency of response on the injured side in the two groups of patients (P-values of 0.08 for amplitude and 0.81 for latency using the Mann–Whitney U-test). The differences of amplitude and latency between the injured and contralateral limbs are also shown (Fig 4C). There was no statistically significant difference between the two groups (P-values 0.75 for amplitude difference and 0.96 for the latency difference, using the Mann–Whitney U-test). There was no statistically significant difference of amplitude or latency of motor response of the biceps in limbs which did, or did not, show evidence of the breathing arm. In patients with clinically complete brachial plexus injuries who did not have any surgical nerve repairs, there was no evidence of recovery, i.e. motor power or gain in Narakas’ scores, except where considered as lesions-in-continuity. They were assessed 10.8 (SEM 4.0) years after injury. Case study A 28 year-old man sustained a complete right-sided brachial plexus injury in a motor cycle accident. He was operated on 4 days after injury. At surgery, spinal nerve roots with ganglia from C5 to T1 were found to be
ARTICLE IN PRESS 174
THE JOURNAL OF HAND SURGERY VOL. 32E No. 2 APRIL
2007
Fig 2 Electromyography studies: (A) multichannel EMG recording of co-contracting muscles (during elbow flexion) in the injured upper limb, (B) multi-channel EMG recording muscles in the injured upper limb showing bursts of motor units firing synchronously with inspiration.
avulsed and situated in the posterior triangle of the neck. After a hemi-laminectomy from C4 to C7 and opening of the dura, implantation of nerve grafts to the spinal cord segments of C5, C6 and C7 was performed. The nerve grafts were pulled through the intervertebral foramen of C7 and joined to the avulsed ventral roots of C5, C6 and C7. Return of motor function was observed about 10 months later in the proximal arm muscles. Three years after surgery, there was MRC grade 4 power in the pectoral and shoulder muscles. There was also grade 3 to 4 power of elbow flexion and extension and grade 1 power in the flexor carpi radialis. There were cocontractions between the biceps and triceps muscles. Inspiration-related contractions in the pectoral and biceps muscles were noted. EMG showed no muscle unit potentials (MUPs) in the deltoid or biceps muscles 6 months after surgery. However, at 3 years after surgery, reinnervated as well as normal looking MUPs were recorded on voluntary contraction in the biceps and
pectoral muscles and small reinnervated MUPs on voluntary contraction were recorded in the deltoid, triceps and infraspinatus muscles and in the long flexor muscles of the forearm. Co-contractions were recorded between the biceps, triceps and pectoral muscles. Inspiration elicited MUPs in the deltoid, triceps and biceps muscles. Electromagnetic stimulation (TMS)-produced muscle responses in the biceps and pectoral muscles were of reduced amplitude and, somewhat, longer latency than on the intact side. The significance of these findings is discussed below. DISCUSSION In this study, we have attempted to compare the results of different surgical repairs, including a novel surgical repair procedure. This was a cross-sectional study of severe brachial plexus avulsion injuries. It was not
ARTICLE IN PRESS MOTOR RECOVERY AND THE BREATHING ARM
Fig 3 Motor responses from Transcranial Magnetic Stimulation from the unaffected (A) and affected (B) side. Note the reduced amplitude and increased latency of the responses from muscles on the affected side.
possible, for obvious technical and ethical reasons, to perform a randomised controlled prospective study in this population. The patients in the two groups, (1) nerve graft and transfer and (2) re-implantation and other repairs, were of similar age and gender. While the two groups were generally comparable in other respects – severity of injury, delays between injury and operation and delays between injury/repairs and assessment – the re-implantation and other repairs group were operated on earlier and had more severe injury. Functional recovery in the proximal part of the limb after re-implantation of avulsed roots into the spinal cord and other repairs were found to be of the same magnitude to that achieved by nerve transfer and grafting. Successful outcome after re-implantation with other repairs seems to be dependent on swift intervention, as we have reported previously, although there are many other factors which could have influenced the regeneration of the motor neurons (Bentolila et al.,1999; Carlstedt et al., 1993, 2000; Cullheim et al., 1989; Kim et al., 2003). Re-implantation of avulsed spinal nerve roots is a new surgical procedure (Bertelli and Ghizoni, 2003; Carlstedt et al., 1995, 2000; Kim et al., 2003; Malessy and Thomeer, 1998). Nerve transfer has an important role in the management of the brachial plexus lesion (Narakas
175
1982, 1984, Narakas and Hentz, 1988), but this option is limited in some patients with multiple nerve root avulsions, because the number of nerves to be repaired is outnumbered by the available donor nerves. For these patients, the surgical strategy of re-implanting avulsed spinal roots or nerve grafts to the spinal cord has been applied. This approach followed a long series of laboratory investigations, including studies in nonhuman primates (Carlstedt, 1997; Carlstedt et al., 1986, 1993; Cullheim et al., 1989; Hallin et al., 1999; Smith and Kodama, 1991). A follow-up study of the first 10 re-implanted cases showed that three patients regained some useful function (Carlstedt et al., 2000). There were, in most cases, signs of regeneration of motor fibres from the spinal cord to the proximal arm muscles, and some recovery of muscle activity (Bertelli and Ghizoni, 2003; Carlstedt et al., 2000, 2004). In our series, the motor recovery and Narakas’ functional score in patients repaired by re-implantation with other repairs was comparable to the group with nerve grafts and transfers. Transcranial magnetic stimulation demonstrated connectivity from the motor cortex to the muscles, with no statistically significant differences in the amplitude and latency of the motor evoked responses between the two groups. In some re-implantation cases, this provided supporting evidence of regeneration from the CNS to the periphery, as illustrated by the case study reported above. Patients with intraspinal repair of brachial plexus avulsions are often hampered by muscle co-contractions (Carlstedt et al., 2004). We were able to demonstrate cocontractions between agonist and antagonist muscles using simultaneous multi-channel EMG recordings from arm muscles. Co-contractions between agonist and antagonist arm muscles on voluntary movement were seen in 34 of our patients. This could be due to aberrant muscle re-innervation (Carlstedt et al., 2000) or the failure of regeneration of Ia fibres, which are responsible for reciprocal inhibition. In patients who had repair by C5/C6 graft, it is likely to have been due to aberrant regeneration through the graft. Surgical and other strategies, such as botulinum toxin injection, have been used to ameliorate co-contractions. Our patients with and without the ‘‘breathing arm’’ had similar brachial plexus lesions and had undergone similar repair operations. There was similar recovery of motor function clinically as well as electrophysiologically in both groups of patients. In three patients, the source of the breathing arm could only have been from re-implanted ventral roots (as in the case study above). The origin of the inspiratory induced arm muscle contractions is the phrenic motoneuron pool, which is situated in spinal cord segments C3, C4 and C5. They occur as a discrete nucleus in the most medial part of the ventral horn next to the motoneurons for the shoulder and the upper part of the arm. Implantation of a peripheral nerve graft into the C5 spinal cord segment after complete brachial plexus avulsion could enable
ARTICLE IN PRESS 176
THE JOURNAL OF HAND SURGERY VOL. 32E No. 2 APRIL
A
7.5
Contralateral Injured
n=18 Amplitude (mV)
2007
5.0
n=5
2.5 n=19 n=5 0.0 Nerve graft and transfer
Re-implantation
B 20
n=5
Contralateral n=17 Injured
Latency (ms)
n=18 n=5 10
0 Nerve graft and transfer
C
Re-implantation
15 Nerve graft & transfer
Amp (mV) or Lat (ms)
Re-implantation
10
5
0 Amplitude difference
Latency difference
Fig 4 Transcranial Magnetic Stimulation studies of the biceps muscle: (A) absolute amplitude study, (B) absolute latency study and (C) amplitude and latency difference. Amp ¼ amplitude; Lat ¼ latency:
regeneration from the phrenic motoneuron pool to the arm instead of the diaphragm. In such a case, there are no other possibilities of reconnection other than through the implanted nerve graft between the upper arm muscles and the phrenic neurons. This phenomenon
after implantation demonstrates regeneration from the spinal cord into the PNS. In other patients, the source of the breathing arm could be from C5 nerves, intercostal nerves or the accessory phrenic nerve. In patients who had repair of
ARTICLE IN PRESS MOTOR RECOVERY AND THE BREATHING ARM
the upper trunk, the source of the breathing arm was presumably the phrenic nerve motor fibres. Fibres to the phrenic nerve sometimes travel with the C5 nerve before joining the phrenic nerve. Injuries to the upper trunk and consequent regeneration could, thus, lead to synkinesis between the muscles of the limb and diaphragm (Schwarz, 1965; Swift, 1994). It would be reasonable to assume that the source of the breathing arm in patients with lower plexus injury repaired by intercostal nerve transfers is through a similar mechanism. The breathing arm phenomenon was described by Erb more than a hundred years ago (Swift, 1994) and by others more recently (Carlstedt et al., 2004; Robinson, 1951; Schwarz, 1965; Swift, 1994). Swift reported that needle EMG examination of the limbs muscles showed the characteristic features of the MUPs pattern of the diaphragm (Swift, 1994). Robinson studied the movement of the diaphragm with fluoroscopy and found that the diaphragm and the limbs muscles were contracting in synchrony (Robinson, 1951). The breathing arm has commonly been described in patients with brachial plexus injury who have been subjected to intercostal nerve transfer to the musculocutaneous nerve (Malessy and Thomeer, 1998; Nagano et al., 1989). The muscle unit potentials (MUPs) associated with respiration have been reported to be present as early as 4 to 8 months after trauma or surgery, and then, often, to gradually wane (Takahashi, 1983), although the involuntary muscle contraction of the biceps still occurred when the patient sneezed or coughed (Nagano et al., 1989). In other studies, particularly in patients with obstetric brachial plexus palsy, the muscle contractions associated with respiration could still be observed 20 years later (Robinson, 1951; Schwarz, 1965). In our study, this phenomenon was observed many years following injury. This observation is in keeping with the finding in adult mammals that there is inability of motor programmes to change when innervating a new target and that motor circuits are usually maintained, even after the motor neurones innervate a foreign muscle (Gruart et al., 2003). Re-implantation of the avulsed spinal nerve root is an effective procedure in the context of the severe, or complete, brachial plexus injury. This study provides evidence of CNS to PNS regeneration and a clinical model to study new treatments which may enhance such regeneration. New strategies are needed, in combination with surgery, to improve the outcome of brachial plexus injury. These may derive from increasing knowledge of the molecular mechanisms of spinal cord injury and repair.
Acknowledgements The authors thank the International Spinal Research Trust (ISRT) for financial support.
177
References Bentolila V, Nizard R, Bizot P et al. (1999). Complete traumatic brachial plexus palsy. Treatment and outcome after repair. Journal of Bone and Joint Surgery, 81A: 20–28. Bergerot A, Shortland PJ, Anand P et al. (2004). Co-treatment with riluzole and GDNF is necessary for functional recovery after ventral root avulsion injury. Experimental Neurology, 187: 359–366. Berman JS, Birch R, Anand P (1998). Pain following human brachial plexus injury with spinal cord root avulsion and the effect of surgery. Pain, 75: 199–207. Bertelli JA, Ghizoni MF (2003). Brachial plexus avulsion injury repairs with nerve transfers and nerve grafts directly implanted into the spinal cord yield partial recovery of shoulder and elbow movements. Neurosurgery, 52: 1385–1389 Discussion 1389–1390. Birch R. Management of brachial plexus injuries. In: Greenwood RJ, Barnes MP, McMillan TM, Ward CD (Eds) Handbook of neurological rehabilitation, Psychology Press, 2003: 663–694. Birch R, Bonney G, Wynn-Perry CB. Surgical disorders of the peripheral nerves, London, Churchill Livingstone, 1988: 157–207. Carlstedt T (1997). Nerve fibre regeneration across the peripheralcentral transitional zone. Journal of Anatomy, 190: 51–56. Carlstedt T, Linda H, Cullheim S et al. (1986). Reinnervation of hind limb muscles after ventral root avulsion and implantation in the lumbar spinal cord of the adult rat. Acta Physiologica Scandinavica, 128: 645–646. Carlstedt TP, Hallin RG, Hedstrom KG et al. (1993). Functional recovery in primates with brachial plexus injury after spinal cord implantation of avulsed ventral roots. Journal of Neurology, Neurosurgery and Psychiatry, 56: 649–654. Carlstedt T, Grane T, Hallin RG et al. (1995). Return of function after spinal cord implantation of avulsed spinal nerve roots. Lancet, 18346: 1323–1325. Carlstedt T, Anand P, Hallin R et al. (2000). Spinal nerve root repair and reimplantation of avulsed ventral roots into the spinal cord after brachial plexus injury. Journal of Neurosurgery (Spine), 93: 237–247. Carlstedt T, Anand P, Htut M et al. (2004). Restoration of hand function and so called ‘‘breathing arm’’ after intraspinal repair of C5-T1 brachial plexus avulsion injury. Case report. Neurosurgical Focus, 16: E7. Cullheim S, Carlstedt T, Linda H et al. (1989). Motoneurons reinnervate skeletal muscle after ventral root implantation into the spinal cord of the cat. Neuroscience, 29: 725–733. Gruart A, Streppel M, Guntinas-Lichius O et al. (2003). Motoneuron adaptability to new motor tasks following two types of facial–facial anastomosis in cats. Brain, 126: 115–133. Hallin RG, Carlstedt T, Nilsson-Remahl I et al. (1999). Spinal cord implantation of avulsed ventral roots in primates; correlation between restored motor function and morphology. Experimental Brain Research, 124: 304–310. Kim DH, Cho YJ, Tiel RL et al. (2003). Outcomes of surgery in 1019 brachial plexus lesions treated at Louisiana State University Health Sciences Center. Journal of Neurosurgery, 98: 1005–1016. Malessy MJ, Thomeer RT (1998). Evaluation of intercostal to musculocutaneous nerve transfer in reconstructive brachial plexus surgery. Journal of Neurosurgery, 88: 266–271. Medical Research Council (1976). Aids to the Examination of the Peripheral Nervous System. Memoradum No 45, London: HMSO. Nagano A (1998). Treatment of brachial plexus injury. Journal of Orthopaedic Sciences, 3: 71–80. Nagano A, Tsuyama N, Ochiai N et al. (1989). Direct nerve crossing with the intercostal nerve to treat avulsion injuries of the brachial plexus. Journal of Hand Surgery, 14A: 980–985. Narakas A (1982). Neurotization or nerve transfer for brachial plexus lesions. Annals de Chirurgie de la Main, 1: 101–118. Narakas AO (1984). Thoughts on neurotization or nerve transfers in irreparable nerve lesions. Clinics in Plastic Surgery, 11: 153–159. Narakas AO, Hentz VR (1988). Neurotization in brachial plexus injuries. Indication and results. Clinical Orthopaedics, 237: 43–56.
ARTICLE IN PRESS 178
Robinson PK (1951). Associate movements between limb and respiratory muscles as a sequel to brachial plexus birth injury. Bulletin of the Johns Hopkins Hospital, 89: 21–29. Samardzic M, Grujicic D, Rasulic L et al. (2002). Restoration of upper arm function in traction injuries to the brachial plexus. Acta Neurochirurgica (Wien), 144: 327–334 Discussion 334–325. Schwarz GA (1965). Posttraumatic respiratory-biceps brachii synkinesis. Case report. Archives of Neurology, 13: 427–431. Smith KJ, Kodama RT (1991). Reinnervation of denervated skeletal muscle by central neurons regenerating via ventral roots implanted into the spinal cord. Brain Research, 551: 221–229. Swift TR (1994). The breathing arm. Muscle and Nerve, 17: 125–129. Takahashi M (1983). Studies on conversion of motor function in intercostal nerves crossing for complete brachial plexus injuries of
THE JOURNAL OF HAND SURGERY VOL. 32E No. 2 APRIL
2007
root avulsion type. Nippon Seikeigeka Gakkai Zasshi, 57: 1799–1807. Terzis JK, Vekris MD, Soucacos PN (2001). Brachial plexus root avulsions. World Journal of Surgery, 25: 1049–1061. Received: 20 November 2005 Accepted after revision: 15 November 2006 Prof Thomas Carlstedt, The PNI Unit, The Royal National Orthopaedic Hospital, Brockley Hill, Stanmore, Middlesex, HA 7 4LP. Tel.: +44 0 208 9095803; fax: +44 0 208 4206582. E-mail: [email protected]
r 2006 The British Society for Surgery of the Hand. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jhsb.2006.11.011 available online at http://www.sciencedirect.com
Forty-four patients with severe traction brachial plexus avulsion injuries were studied following surgical repairs. In eight patients, re-implanting avulsed spinal roots directly to the spinal cord was performed with other repairs and motor recovery in the proximal limb was similar to that achieved by conventional nerve grafts and transfers when assessed using the MRC clinical grades, Narakas scores, EMG and Transcranial Magnetic Stimulation (TMS). Thirty-four of the 37 patients had co-contractions of agonist and antagonist muscle groups. Spontaneous contractions of limb muscles in synchrony with respiration, the ‘‘breathing arm’’, were noted in 26 of 37 patients: in three patients, the source of the breathing arm was from spinal cord re-connection, providing evidence of regeneration from the CNS to the periphery. Our study shows that re-connection of avulsed spinal roots can produce good motor recovery and provides a clinical model for developing new treatments which may enhance nerve regeneration. Journal of Hand Surgery (European Volume, 2007) 32E: 2: 170–178 Keywords: brachial plexus injury, surgical repair, re-implantation, breathing arm, electrophysiology
Avulsion of the brachial plexus can be the most serious of all injuries to the peripheral nerves, with grave impairment of the quality of life (Bertelli and Ghizoni, 2003; Birch, 2003; Nagano, 1998; Terzis et al., 2001). Most patients are young men who have been involved in road traffic accidents and the incidence of these injuries is increasing (Birch et al., 1998). Usually both ventral and dorsal roots are involved and the patient is subjected to paralysis and sensory dysfunction, with numbness in the limb in conjunction with extreme, intractable pain (Berman et al., 1998; Birch et al., 1998). In general, the results of surgical repairs of brachial plexus injuries are influenced by the level and extent of injury, the type of surgical procedure (Samardzic et al., 2002) and, above all, by delay before repair. There have been significant recent advances in the surgical treatment of traction brachial plexus injury. The commonest means of repairing spinal cord root avulsion injuries in current practice is to transfer an intact neighbouring nerve to the distal stump of the damaged nerve, so restoring some motor and sensory function (Bentolila et al., 1999; Birch et al., 1998). The novel surgical strategy of re-implanting, or re-connecting, avulsed spinal roots, via nerve grafts, directly to the spinal cord (termed ‘‘re-implants’’) has been performed in patients with severe brachial plexus injury (Carlstedt et al., 1995, 2000, 2004). The main purpose of this cross-sectional study was to evaluate motor recovery after different surgical repairs, viz. conventional nerve grafts for plexus injuries distal to
the dorsal root ganglion and nerve transfers and/or re-implantation for spinal cord-root avulsion injuries. We have used the MRC (Medical Research Council) grades for muscle power, Narakas scores (Narakas and Hentz, 1988), Electromyography (EMG) and Transcranial Magnetic Stimulation (TMS). We also recorded phenomena related to motor recovery, viz. the cocontraction of agonist and antagonistic muscles and ‘‘the breathing arm’’. The ‘‘breathing arm’’ refers to contractions of arm muscles synchronously with respiration. This phenomenon has been described in patients who have had a variety of operations for the repair of brachial plexus injury, especially after intercostal nerve transfer (Swift, 1994; Takahashi, 1983). In this study, we looked, for the first time, for the ‘‘breathing arm’’ in patients with complete brachial plexus avulsion injuries repaired by re-implantation, which would provide evidence for regeneration of motoneurons from the central nervous system (CNS) to the peripheral nervous system (PNS), as, in some patients the re-implanted root or graft would be the sole conduit for the phrenic spinal motoneurons to the periphery. PATIENTS AND METHODS Forty-four patients who had sustained severe brachial plexus injury with spinal nerve root avulsions were studied after written informed consent and local ethics committee approval. Patients with associated spinal cord or brain injury, injury to the proximal major blood 170
ARTICLE IN PRESS MOTOR RECOVERY AND THE BREATHING ARM
vessels or double level lesions (as revealed during explorative surgery) were excluded. Not all patients had complete paralysis. The serratus anterior was functioning in 27 out of 44 patients, one patient had no lesion of C8 or T1 and three had lesions other than avulsions: two had rupture and one had a lesion-incontinuity distal to the dorsal root ganglion. Avulsion or intraspinal nerve root injury was confirmed by computerised tomography, (CT)-myelography, pre-operative electrophysiology, direct observation of the exposed brachial plexus during operation and inspection of the spinal cord (where applicable). In cases of complete avulsion, the ventral and dorsal roots, with dorsal root ganglia, were totally detached. When the pertinent spinal nerve at surgery was found to be ‘‘in situ’’, a diagnosis of avulsion was made from preoperative findings together with results of scanning and peri-operative electrophysiology. The operations were performed at the Peripheral Nerve Injury Unit, Royal National Orthopaedic Hospital. For surgical techniques, see Birch et al. (1998) and Carlstedt et al. (1995, 2000). Most patients had more than one type of operative repair. Conventional graft repair of ruptured spinal nerves was combined, if possible, with transfers of the accessory nerve, the accessory phrenic nerve, the supraclavicular nerves or the intercostal nerves. Spinal cord replantation of avulsed roots was performed in cases with four or more roots avulsed, no later than 1 month after injury, for neurobiological reasons – death of nerve cells with delayed repair (Bergerot et al., 2004). For the re-implantation procedure, the patient was positioned laterally on the operating table to allow a simultaneous approach to the extra and intraspinal parts of the brachial plexus (for details regarding surgical exploration, with illustrations, see Carlstedt et al., 2000). After a dissection through the neck muscles, respecting the accessory nerve, a C5 to C7 hemi-laminectomy was performed. After the dura had been opened, the intraspinal injury was verified. Nerve grafts were pulled through the intervertebral foraminae and positioned in the spinal canal. Stay sutures were applied to the denticulate ligament in order to gently rotate the spinal cord. Small slits were produced in the pia mater in the anterolateral aspect of the spinal cord to receive the ends of the nerve transplants, which were inserted deep to the surface. Tissue glue was applied to maintain the attachment of the nerve grafts to the spinal cord. The distal ends of the nerve grafts were connected to the motor parts of the avulsed spinal nerves. Patients were operated on up to 31 months after injury by the same two peripheral nerve surgeons (T.C. and R.B.). Patients who had sustained clinically complete brachial plexus injuries and had no surgical nerve repair, because of late referral or the wish of the patient, served as controls (n ¼ 10). A diagnosis of avulsion was made from spinal imaging, electrophysiology
171
Table 1—Mean numbers (SEM) of lesions in patients treated by different methods Method of repair
Avulsion
Rupture
Total no. of lesions
Graft and nerve transfers (n ¼ 26 patients) Re-implantation and other repairs (n ¼ 8 patients) No surgical repair (n ¼ 10 patients)
2.9 (0.2)
1.6 (0.1)
4.4 (0.2)
4.3 (0.3)
0.7 (0.4)
4.8 (0.2)
3.8 (0.4)
0.1 (0.1)
3.9 (0.4)
and preservation of dermatomal skin flare responses (Berman et al., 1998). The patients studied were all men. The age at which the injury was sustained ranged between 17 and 38 years, with a mean age of 24 years, for the graft and nerve transfer group; 18 and 36 years, with a mean age of 27 years, for the re-implantation and other repairs group; 17 and 45 years, with a mean age of 29 years, for the no repair group. The age of the patients at the time of surgery ranged between 17 and 38 years, with a mean age of 24 years, for the graft and transfer group and 18 and 36 years, with a mean age of 27 years, for the reimplantation and other repairs group. The mean numbers of lesions in patients treated by the different methods and in the patients with no repair are summarised in Table 1. The delay between injury and surgery is shown in Table 2. Clinical motor assessment All the patients underwent a detailed neurological examination. Power in the upper limb was assessed using the grading recommended by the Medical Research Council (Medical Research Council, 1976). The functional recovery of the shoulder and elbow was also assessed using Narakas’ functional scoring system (Narakas and Hentz, 1988), which evaluates global functional recovery, especially at the shoulder and elbow, rather than recovery of individual muscles. EMG examination Concentric needle (TECA disposable needle; Oxford Instruments Medical, Surrey, UK) EMG examination was performed with a Keypoints work station (Medtronic Functional Diagnostic A/S, Skovlunde, Denmark). Muscles were sampled using conventional electromyographic methods, including insertional and spontaneous activities and motor unit action potentials (MUPs). In addition, motor unit activity was also observed in relationship to respiration and coughing. Multi-channel simultaneous surface EMG recordings of
ARTICLE IN PRESS 172
THE JOURNAL OF HAND SURGERY VOL. 32E No. 2 APRIL
2007
Table 2—Delay between the injury, surgery and assessment Methods of repair
Graft and nerve transfers Re-implantation and other repairs
Delay between injury and surgery (weeks)
Duration of follow-up (years)
Mean (SEM)
Max
Min
Mean (SEM)
Max
Min
6 (1) 1 (0.2)
122 19
1 3
6 (1) 3 (0.4)
19 4
1 2
MAX ¼ maximum, MIN ¼ minimum.
Single pulse TMS was performed using the Magstims200 Monophasic stimulator (Novametrix Medical Systems Ltd., Whitland, UK). The output was connected to a Keypoint work station (Medtronic Functional Diagnostic A/S, Skovlunde, Denmark) to record responses. TMS is a technique in which the motor cortex is stimulated by a brief electromagnetic stimulus delivered via a hand held magnetic stimulation circular coil (a high power 90 mm coil). The coil is placed over the head on the area over the motor cortex. The evoked motor response is detected by electrodes placed over the relevant upper limb muscles and the motor action potentials recorded on an EMG machine. Facilitation was used whenever possible (i.e. the patient was asked to attempt to contract the muscle from which the recording was made, as this can improve the amplitude of the recorded motor response). Both the injured and the contra-lateral (intact) limb were studied. Statistics The Mann–Whitney U-test and the Kruskal–Wallis test were used. P-values o0.05 were considered to be significant. Data are presented as mean7SEM (standard error of mean) unless otherwise stated.
RESULTS The recovery of motor function from nerve grafts and transfers and re-implantations with other repairs are presented in Figs 1A and B and in Table 3. There was no statistically significant difference of pectorals, deltoid, biceps and triceps strength in patients who had nerve graft and transfer in comparison with reimplantation with other repairs, using the Mann–Whitney U-test. The results of deltoid and biceps motor recovery are shown in Figs 1A and B. In some patients,
Motor power in MRC grading
Magnetic stimulation
A
5
4
3
2
1
0 Pectorals
Deltoid
Biceps
Triceps
B 5 Motor power in MRC grading
the pectoral muscles, the deltoid, biceps and triceps muscles were made in order to determine the presence of co-contractions of limb muscles with voluntary movements and respiration.
4
Pectorals
Biceps
Deltoid
Triceps
3 2 1 0 Muscles
Fig 1 Recovery of motor function: (A) from graft and nerve transfer and (B) from re-implantation with other repairs.
co-contractions made it difficult to assess muscle power accurately, and, hence, some patient data points are not shown. In 22 of 26 of the graft patients, deltoid muscle power was assessed and 11 of these 22 patients regained power greater than 3 of the MRC scale. In 21 of 26 of the graft patients, biceps muscle power was assessed and seven of these 21 patients regained power greater than 3 of the MRC scale. In the re-implantation group, only
ARTICLE IN PRESS MOTOR RECOVERY AND THE BREATHING ARM
173
Table 3—The recovery of motor function after different surgical procedures Motor power (MRC grading)1
Methods of surgical repair
Graft and nerve transfers Re-implantation and other repairs P-values
Pectorals
Deltoid
Biceps
Triceps
3.4 (0.2) 2.7 (0.5)
2.4 (0.3) 1.5 (1.0)
2.0 (0.3) 2.0 (0.5)
2.1 (0.4) 2.2 (0.6)
0.2
0.3
1.0
1.0
MRC, Medical Research Council. 1 Values are mean7SEM.
Table 4—Breathing arm or co-contraction Number of patients (n ¼ 37) Breathing arm and co-contraction Breathing arm only Co-contraction only No breathing arm or co-contraction
25 1 9 2
one of four deltoid muscles regained power 3 of the MRC grading. Two of five patients regained biceps muscle power greater than 3 of the MRC grading. The Narakas’ functional scores for the shoulder and elbow for patients repaired by nerve graft and transfers were 3.4 (SEM 1.1) and 2.3 (SEM 0.5) and for patients repaired by means of re-implantation with other repairs were 3.5 (SEM 1.1) and 1.8 (SEM 0.7). There was no statistically significant difference in the Narakas’ functional score between the two groups for the shoulder or elbow. The duration after the injury to assessment and the duration of delay between injury and the operation are shown in Table 2. There were no statistically significant differences in the mean interval between the injury and surgery, or between the mean duration from injury to assessment, between the two groups. Thirty-seven patients were studied for the ‘‘breathing arm’’ and limb muscle co-contraction. As can be seen in Table 4, 25 patients were found to have both the breathing arm and co-contraction, while nine patients showed no evidence of the breathing arm but had limb muscle co-contractions. One patient showed evidence of the breathing arm, but not co-contractions. Two patients did not have either the breathing arm or cocontractions. In three patients, the source of the breathing arm could only have been from re-implanted ventral roots (see case study below). In the rest of the patients, the source of the breathing arm could have been from the C5 nerves, the intercostal nerves or the accessory phrenic nerve. The duration after the injury to follow-up assessments for the patients with and without the breathing arm was 6.7 (SEM 1.5) and 5.2 (SEM 1.5) years, respectively. This phenomenon was observed as early as 13 months
after surgery and was recorded up to 19 years after injury. EMG examination of the affected muscles generally showed evidence of partial denervation with reinnervation. Motor unit potentials (MUPs) firing synchronously with respiration (inspiration) in quiet, or voluntary, breathing were observed in several cases. The same motor units could usually be activated by volitional contraction. MUPs firing simultaneously were seen in groups of muscles (agonists and antagonists) that were co-contracting (Figs 2A and B). The amplitudes of the motor responses to TMS, where present, were smaller when compared to the contralateral arm, and the latencies of the responses were prolonged (Figs 3A and B). The absolute values of amplitude and latency of biceps response are shown in Figs 4A and B. There was no statistically significant difference of amplitude and latency of response on the injured side in the two groups of patients (P-values of 0.08 for amplitude and 0.81 for latency using the Mann–Whitney U-test). The differences of amplitude and latency between the injured and contralateral limbs are also shown (Fig 4C). There was no statistically significant difference between the two groups (P-values 0.75 for amplitude difference and 0.96 for the latency difference, using the Mann–Whitney U-test). There was no statistically significant difference of amplitude or latency of motor response of the biceps in limbs which did, or did not, show evidence of the breathing arm. In patients with clinically complete brachial plexus injuries who did not have any surgical nerve repairs, there was no evidence of recovery, i.e. motor power or gain in Narakas’ scores, except where considered as lesions-in-continuity. They were assessed 10.8 (SEM 4.0) years after injury. Case study A 28 year-old man sustained a complete right-sided brachial plexus injury in a motor cycle accident. He was operated on 4 days after injury. At surgery, spinal nerve roots with ganglia from C5 to T1 were found to be
ARTICLE IN PRESS 174
THE JOURNAL OF HAND SURGERY VOL. 32E No. 2 APRIL
2007
Fig 2 Electromyography studies: (A) multichannel EMG recording of co-contracting muscles (during elbow flexion) in the injured upper limb, (B) multi-channel EMG recording muscles in the injured upper limb showing bursts of motor units firing synchronously with inspiration.
avulsed and situated in the posterior triangle of the neck. After a hemi-laminectomy from C4 to C7 and opening of the dura, implantation of nerve grafts to the spinal cord segments of C5, C6 and C7 was performed. The nerve grafts were pulled through the intervertebral foramen of C7 and joined to the avulsed ventral roots of C5, C6 and C7. Return of motor function was observed about 10 months later in the proximal arm muscles. Three years after surgery, there was MRC grade 4 power in the pectoral and shoulder muscles. There was also grade 3 to 4 power of elbow flexion and extension and grade 1 power in the flexor carpi radialis. There were cocontractions between the biceps and triceps muscles. Inspiration-related contractions in the pectoral and biceps muscles were noted. EMG showed no muscle unit potentials (MUPs) in the deltoid or biceps muscles 6 months after surgery. However, at 3 years after surgery, reinnervated as well as normal looking MUPs were recorded on voluntary contraction in the biceps and
pectoral muscles and small reinnervated MUPs on voluntary contraction were recorded in the deltoid, triceps and infraspinatus muscles and in the long flexor muscles of the forearm. Co-contractions were recorded between the biceps, triceps and pectoral muscles. Inspiration elicited MUPs in the deltoid, triceps and biceps muscles. Electromagnetic stimulation (TMS)-produced muscle responses in the biceps and pectoral muscles were of reduced amplitude and, somewhat, longer latency than on the intact side. The significance of these findings is discussed below. DISCUSSION In this study, we have attempted to compare the results of different surgical repairs, including a novel surgical repair procedure. This was a cross-sectional study of severe brachial plexus avulsion injuries. It was not
ARTICLE IN PRESS MOTOR RECOVERY AND THE BREATHING ARM
Fig 3 Motor responses from Transcranial Magnetic Stimulation from the unaffected (A) and affected (B) side. Note the reduced amplitude and increased latency of the responses from muscles on the affected side.
possible, for obvious technical and ethical reasons, to perform a randomised controlled prospective study in this population. The patients in the two groups, (1) nerve graft and transfer and (2) re-implantation and other repairs, were of similar age and gender. While the two groups were generally comparable in other respects – severity of injury, delays between injury and operation and delays between injury/repairs and assessment – the re-implantation and other repairs group were operated on earlier and had more severe injury. Functional recovery in the proximal part of the limb after re-implantation of avulsed roots into the spinal cord and other repairs were found to be of the same magnitude to that achieved by nerve transfer and grafting. Successful outcome after re-implantation with other repairs seems to be dependent on swift intervention, as we have reported previously, although there are many other factors which could have influenced the regeneration of the motor neurons (Bentolila et al.,1999; Carlstedt et al., 1993, 2000; Cullheim et al., 1989; Kim et al., 2003). Re-implantation of avulsed spinal nerve roots is a new surgical procedure (Bertelli and Ghizoni, 2003; Carlstedt et al., 1995, 2000; Kim et al., 2003; Malessy and Thomeer, 1998). Nerve transfer has an important role in the management of the brachial plexus lesion (Narakas
175
1982, 1984, Narakas and Hentz, 1988), but this option is limited in some patients with multiple nerve root avulsions, because the number of nerves to be repaired is outnumbered by the available donor nerves. For these patients, the surgical strategy of re-implanting avulsed spinal roots or nerve grafts to the spinal cord has been applied. This approach followed a long series of laboratory investigations, including studies in nonhuman primates (Carlstedt, 1997; Carlstedt et al., 1986, 1993; Cullheim et al., 1989; Hallin et al., 1999; Smith and Kodama, 1991). A follow-up study of the first 10 re-implanted cases showed that three patients regained some useful function (Carlstedt et al., 2000). There were, in most cases, signs of regeneration of motor fibres from the spinal cord to the proximal arm muscles, and some recovery of muscle activity (Bertelli and Ghizoni, 2003; Carlstedt et al., 2000, 2004). In our series, the motor recovery and Narakas’ functional score in patients repaired by re-implantation with other repairs was comparable to the group with nerve grafts and transfers. Transcranial magnetic stimulation demonstrated connectivity from the motor cortex to the muscles, with no statistically significant differences in the amplitude and latency of the motor evoked responses between the two groups. In some re-implantation cases, this provided supporting evidence of regeneration from the CNS to the periphery, as illustrated by the case study reported above. Patients with intraspinal repair of brachial plexus avulsions are often hampered by muscle co-contractions (Carlstedt et al., 2004). We were able to demonstrate cocontractions between agonist and antagonist muscles using simultaneous multi-channel EMG recordings from arm muscles. Co-contractions between agonist and antagonist arm muscles on voluntary movement were seen in 34 of our patients. This could be due to aberrant muscle re-innervation (Carlstedt et al., 2000) or the failure of regeneration of Ia fibres, which are responsible for reciprocal inhibition. In patients who had repair by C5/C6 graft, it is likely to have been due to aberrant regeneration through the graft. Surgical and other strategies, such as botulinum toxin injection, have been used to ameliorate co-contractions. Our patients with and without the ‘‘breathing arm’’ had similar brachial plexus lesions and had undergone similar repair operations. There was similar recovery of motor function clinically as well as electrophysiologically in both groups of patients. In three patients, the source of the breathing arm could only have been from re-implanted ventral roots (as in the case study above). The origin of the inspiratory induced arm muscle contractions is the phrenic motoneuron pool, which is situated in spinal cord segments C3, C4 and C5. They occur as a discrete nucleus in the most medial part of the ventral horn next to the motoneurons for the shoulder and the upper part of the arm. Implantation of a peripheral nerve graft into the C5 spinal cord segment after complete brachial plexus avulsion could enable
ARTICLE IN PRESS 176
THE JOURNAL OF HAND SURGERY VOL. 32E No. 2 APRIL
A
7.5
Contralateral Injured
n=18 Amplitude (mV)
2007
5.0
n=5
2.5 n=19 n=5 0.0 Nerve graft and transfer
Re-implantation
B 20
n=5
Contralateral n=17 Injured
Latency (ms)
n=18 n=5 10
0 Nerve graft and transfer
C
Re-implantation
15 Nerve graft & transfer
Amp (mV) or Lat (ms)
Re-implantation
10
5
0 Amplitude difference
Latency difference
Fig 4 Transcranial Magnetic Stimulation studies of the biceps muscle: (A) absolute amplitude study, (B) absolute latency study and (C) amplitude and latency difference. Amp ¼ amplitude; Lat ¼ latency:
regeneration from the phrenic motoneuron pool to the arm instead of the diaphragm. In such a case, there are no other possibilities of reconnection other than through the implanted nerve graft between the upper arm muscles and the phrenic neurons. This phenomenon
after implantation demonstrates regeneration from the spinal cord into the PNS. In other patients, the source of the breathing arm could be from C5 nerves, intercostal nerves or the accessory phrenic nerve. In patients who had repair of
ARTICLE IN PRESS MOTOR RECOVERY AND THE BREATHING ARM
the upper trunk, the source of the breathing arm was presumably the phrenic nerve motor fibres. Fibres to the phrenic nerve sometimes travel with the C5 nerve before joining the phrenic nerve. Injuries to the upper trunk and consequent regeneration could, thus, lead to synkinesis between the muscles of the limb and diaphragm (Schwarz, 1965; Swift, 1994). It would be reasonable to assume that the source of the breathing arm in patients with lower plexus injury repaired by intercostal nerve transfers is through a similar mechanism. The breathing arm phenomenon was described by Erb more than a hundred years ago (Swift, 1994) and by others more recently (Carlstedt et al., 2004; Robinson, 1951; Schwarz, 1965; Swift, 1994). Swift reported that needle EMG examination of the limbs muscles showed the characteristic features of the MUPs pattern of the diaphragm (Swift, 1994). Robinson studied the movement of the diaphragm with fluoroscopy and found that the diaphragm and the limbs muscles were contracting in synchrony (Robinson, 1951). The breathing arm has commonly been described in patients with brachial plexus injury who have been subjected to intercostal nerve transfer to the musculocutaneous nerve (Malessy and Thomeer, 1998; Nagano et al., 1989). The muscle unit potentials (MUPs) associated with respiration have been reported to be present as early as 4 to 8 months after trauma or surgery, and then, often, to gradually wane (Takahashi, 1983), although the involuntary muscle contraction of the biceps still occurred when the patient sneezed or coughed (Nagano et al., 1989). In other studies, particularly in patients with obstetric brachial plexus palsy, the muscle contractions associated with respiration could still be observed 20 years later (Robinson, 1951; Schwarz, 1965). In our study, this phenomenon was observed many years following injury. This observation is in keeping with the finding in adult mammals that there is inability of motor programmes to change when innervating a new target and that motor circuits are usually maintained, even after the motor neurones innervate a foreign muscle (Gruart et al., 2003). Re-implantation of the avulsed spinal nerve root is an effective procedure in the context of the severe, or complete, brachial plexus injury. This study provides evidence of CNS to PNS regeneration and a clinical model to study new treatments which may enhance such regeneration. New strategies are needed, in combination with surgery, to improve the outcome of brachial plexus injury. These may derive from increasing knowledge of the molecular mechanisms of spinal cord injury and repair.
Acknowledgements The authors thank the International Spinal Research Trust (ISRT) for financial support.
177
References Bentolila V, Nizard R, Bizot P et al. (1999). Complete traumatic brachial plexus palsy. Treatment and outcome after repair. Journal of Bone and Joint Surgery, 81A: 20–28. Bergerot A, Shortland PJ, Anand P et al. (2004). Co-treatment with riluzole and GDNF is necessary for functional recovery after ventral root avulsion injury. Experimental Neurology, 187: 359–366. Berman JS, Birch R, Anand P (1998). Pain following human brachial plexus injury with spinal cord root avulsion and the effect of surgery. Pain, 75: 199–207. Bertelli JA, Ghizoni MF (2003). Brachial plexus avulsion injury repairs with nerve transfers and nerve grafts directly implanted into the spinal cord yield partial recovery of shoulder and elbow movements. Neurosurgery, 52: 1385–1389 Discussion 1389–1390. Birch R. Management of brachial plexus injuries. In: Greenwood RJ, Barnes MP, McMillan TM, Ward CD (Eds) Handbook of neurological rehabilitation, Psychology Press, 2003: 663–694. Birch R, Bonney G, Wynn-Perry CB. Surgical disorders of the peripheral nerves, London, Churchill Livingstone, 1988: 157–207. Carlstedt T (1997). Nerve fibre regeneration across the peripheralcentral transitional zone. Journal of Anatomy, 190: 51–56. Carlstedt T, Linda H, Cullheim S et al. (1986). Reinnervation of hind limb muscles after ventral root avulsion and implantation in the lumbar spinal cord of the adult rat. Acta Physiologica Scandinavica, 128: 645–646. Carlstedt TP, Hallin RG, Hedstrom KG et al. (1993). Functional recovery in primates with brachial plexus injury after spinal cord implantation of avulsed ventral roots. Journal of Neurology, Neurosurgery and Psychiatry, 56: 649–654. Carlstedt T, Grane T, Hallin RG et al. (1995). Return of function after spinal cord implantation of avulsed spinal nerve roots. Lancet, 18346: 1323–1325. Carlstedt T, Anand P, Hallin R et al. (2000). Spinal nerve root repair and reimplantation of avulsed ventral roots into the spinal cord after brachial plexus injury. Journal of Neurosurgery (Spine), 93: 237–247. Carlstedt T, Anand P, Htut M et al. (2004). Restoration of hand function and so called ‘‘breathing arm’’ after intraspinal repair of C5-T1 brachial plexus avulsion injury. Case report. Neurosurgical Focus, 16: E7. Cullheim S, Carlstedt T, Linda H et al. (1989). Motoneurons reinnervate skeletal muscle after ventral root implantation into the spinal cord of the cat. Neuroscience, 29: 725–733. Gruart A, Streppel M, Guntinas-Lichius O et al. (2003). Motoneuron adaptability to new motor tasks following two types of facial–facial anastomosis in cats. Brain, 126: 115–133. Hallin RG, Carlstedt T, Nilsson-Remahl I et al. (1999). Spinal cord implantation of avulsed ventral roots in primates; correlation between restored motor function and morphology. Experimental Brain Research, 124: 304–310. Kim DH, Cho YJ, Tiel RL et al. (2003). Outcomes of surgery in 1019 brachial plexus lesions treated at Louisiana State University Health Sciences Center. Journal of Neurosurgery, 98: 1005–1016. Malessy MJ, Thomeer RT (1998). Evaluation of intercostal to musculocutaneous nerve transfer in reconstructive brachial plexus surgery. Journal of Neurosurgery, 88: 266–271. Medical Research Council (1976). Aids to the Examination of the Peripheral Nervous System. Memoradum No 45, London: HMSO. Nagano A (1998). Treatment of brachial plexus injury. Journal of Orthopaedic Sciences, 3: 71–80. Nagano A, Tsuyama N, Ochiai N et al. (1989). Direct nerve crossing with the intercostal nerve to treat avulsion injuries of the brachial plexus. Journal of Hand Surgery, 14A: 980–985. Narakas A (1982). Neurotization or nerve transfer for brachial plexus lesions. Annals de Chirurgie de la Main, 1: 101–118. Narakas AO (1984). Thoughts on neurotization or nerve transfers in irreparable nerve lesions. Clinics in Plastic Surgery, 11: 153–159. Narakas AO, Hentz VR (1988). Neurotization in brachial plexus injuries. Indication and results. Clinical Orthopaedics, 237: 43–56.
ARTICLE IN PRESS 178
Robinson PK (1951). Associate movements between limb and respiratory muscles as a sequel to brachial plexus birth injury. Bulletin of the Johns Hopkins Hospital, 89: 21–29. Samardzic M, Grujicic D, Rasulic L et al. (2002). Restoration of upper arm function in traction injuries to the brachial plexus. Acta Neurochirurgica (Wien), 144: 327–334 Discussion 334–325. Schwarz GA (1965). Posttraumatic respiratory-biceps brachii synkinesis. Case report. Archives of Neurology, 13: 427–431. Smith KJ, Kodama RT (1991). Reinnervation of denervated skeletal muscle by central neurons regenerating via ventral roots implanted into the spinal cord. Brain Research, 551: 221–229. Swift TR (1994). The breathing arm. Muscle and Nerve, 17: 125–129. Takahashi M (1983). Studies on conversion of motor function in intercostal nerves crossing for complete brachial plexus injuries of
THE JOURNAL OF HAND SURGERY VOL. 32E No. 2 APRIL
2007
root avulsion type. Nippon Seikeigeka Gakkai Zasshi, 57: 1799–1807. Terzis JK, Vekris MD, Soucacos PN (2001). Brachial plexus root avulsions. World Journal of Surgery, 25: 1049–1061. Received: 20 November 2005 Accepted after revision: 15 November 2006 Prof Thomas Carlstedt, The PNI Unit, The Royal National Orthopaedic Hospital, Brockley Hill, Stanmore, Middlesex, HA 7 4LP. Tel.: +44 0 208 9095803; fax: +44 0 208 4206582. E-mail: [email protected]
r 2006 The British Society for Surgery of the Hand. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jhsb.2006.11.011 available online at http://www.sciencedirect.com