- Email: [email protected]
Reinnervation of motor units in intrinsic muscles of a transplanted hand
Neuroscience Letters 373 (2005) 138–143
Reinnervation of motor units in intrinsic muscles of a transplanted hand Marco Lanzettaa , Marco Pozzob , Andrea Bottinb , Roberto Merlettib , Dario Farinac,∗ b
a Hand Surgery and Reconstructive Microsurgery Unit, San Gerardo Hospital, University of Milan-Bicocca, Monza, Italy Laboratorio di Ingegneria del Sistema Neuromuscolare, Centro di Bioingegneria, Dip. di Elettronica, Politecnico di Torino, Torino, Italy c Center for Sensory-Motor Interaction (SMI), Department of Health Science and Technology, Aalborg University, Fredrik Bajers Vej 7 D-3, DK-9220 Aalborg, Denmark
Received 6 August 2004; received in revised form 27 September 2004; accepted 1 October 2004
Abstract Functional recovery of transplanted hand can be evaluated clinically but until now there has been no direct assessment of muscle control. In October 2000 we transplanted the right hand of a brain-dead man aged 43 onto a man aged 35 who had lost his right dominant hand 22 years before. Starting from day 205 after the transplant, multi-channel surface electromyographic (EMG) signals were recorded from intrinsic muscles of the transplanted hand in order to assess their degree of reinnervation. Eleven months post-operatively, the first motor unit action potential train was detected from the abductor digiti minimi. One month later, also the abductor pollicis brevis and the opponens pollicis muscles showed motor unit activity, while, after 15 and 24 months, the first dorsal interosseous and the first lumbricalis muscles, respectively, showed activation of their first motor units. An increase in the number of active motor units was observed after the first signs of reinnervation, although the process was rather slow. In sustained maximal contractions, the motor unit discharge rate decreased from (mean ± S.D.) 34.0 ± 6.7 pps to 23.4 ± 5.1 pps in 60 s for the abductor digiti minimi, although the subject was verbally encouraged to maintain a maximal activation. Moreover, the subject was able to perform basic control tasks involving voluntary modulation of motor unit discharge rate. With a visual feedback, he could increase discharge rate of the abductor digiti minimi approximately linearly over time, from 13.4 ± 6.7 pps to 32.5 ± 11.2 pps in 60 s. In conclusion, we showed reinnervation of single motor units in a transplanted hand after 22 years of denervation. Moreover, voluntary modulation of discharge rates of these motor units could be performed since the first sign of reinnervation. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Hand transplantation; Multi-channel surface electromyography; Motor unit; Discharge rate
More than 20 hand/digital transplantations have been performed worldwide [8]. All patients are currently healthy with viable grafts, except for the first, who experienced rejection after he arbitrarily stopped taking immunosuppressant drugs [8]. Sensory and motor recovery in these patients is contingent on nerve regeneration [18], shown to take place rapidly in transplanted hands even many years after the trauma. Sensory and motor nerve regeneration are usually assessed by clinical assessments such as strength testing of intrinsic hand muscles, Tinel’s sign, Semmes–Weinstein microfilaments, pinprick, hot and cold temperature, and light and deep pressure [18]. However, until now there has been no direct assessment of muscle control at the level of motor units. The analysis ∗
Corresponding author. Tel.: +45 963 58821; fax: +45 981 54008. E-mail address: [email protected] (D. Farina).
0304-3940/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2004.10.001
of motor unit properties in intrinsic muscles after transplant would provide further insight into the plasticity of the neuromuscular system. Electrophysiological techniques, such as intramuscular or surface electromyography (EMG), long used to assess segmental nerve damage, can be applied to better assess post-surgical nerve regeneration. Non-invasive approaches are preferred in hand-transplanted subjects, to minimize possible damage to the allograft and the risk of infections. The main objective of this study was to assess the reinnervation process in intrinsic muscles after hand transplantation at the motor unit level. This objective was achieved by noninvasive multi-channel surface EMG techniques. The Italian Hand Transplantation Program was approved by the Italian Health Ministry and the Ethical Committee of
M. Lanzetta et al. / Neuroscience Letters 373 (2005) 138–143
the University of Milan-Bicocca. Approximately 400 possible candidates were interviewed over a period of 2 years and studied under a preliminary protocol using strict inclusion criteria. The selected recipient was a 35-year-old male who had lost his right dominant hand in a farming accident when he was 13. The patient underwent a series of routine pretransplant investigations, including an angiogram, CT scan, muscle and nerve charts, magnetic resonance imaging of the stump, functional magnetic resonance imaging of the brain, and a comprehensive battery of psychological and clinical tests. The multi-organ cadaveric donor was a 43-year-old male who died from a stroke. He had the same blood group (O positive) as the recipient and there were six HLA mismatches; the cross-match was negative. The surgical procedure was similar to that followed during hand replantation, and included bone fixation, revascularization with microsurgical anastomoses of arteries and veins, tendon repair or tendon transfer when necessary, microsurgical nerve repair, and skin closure [7]. The transplant operation lasted 12 h with a total ischemia time of 11 h. The forearm bones were transversely cut so that the grafted limb would be of the same length as the contralateral. Bone fixation of the radius and ulna was achieved by means of compression plates and 4.5 mm screws, and autologous cancellous bone graft from the iliac crest was placed around the osteosynthesis site to improve healing. Most of the deep flexor and all extensor tendons were repaired. The hand was then revascularized by anastomosing the radial and ulnar arteries, as well as three veins. The median and ulnar nerves were repaired using a conventional epineurial technique. Lastly, the remaining more superficial flexor tendons were sutured. The skin was closed in layers. The patient weighed 73 kg and was given 250 ml of Dextran 40 before declamping and 20 ml/h for 7 days. Aspirin 150 was administered for 7 days, and wide-spectrum antibiotic therapy for 10 days. The induction immunosuppressive protocol consisted of 20 mg of monoclonal antibody antiCD25 (Baxilimab-Simulect® ) 2 h before the operation, on day 4 and on day 45 post-operatively, FK506 (TacrolimusPrograf® ), adjusted to maintain blood concentration between 15 and 20 mg/ml for the first month, mycophenolate mofetil (Cell Cept® 2 g/day), and steroid (prednisone 250 mg on day 1 and rapidly tapered to 20 mg/day). The maintenance therapy consisted of FK506 (blood levels between 5 and 10 mg/ml), mycophenolate mofetil 1 g/day and prednisone 10 mg/day [7,8,9,12]. The rehabilitation protocol started as soon as swelling subsided and was performed twice daily for 138 days, and once daily thereafter as the patient returned to work. It included a standard rehabilitation program for flexor and extensor tendons, sensory re-education and cortical re-integration [7]. In the early phases, hand movement was initiated passively, then with active-assisted exercises, and finally with active and against-resistance exercises. Electrical stimulation was started on day 60 and carried out twice daily since. Occupa-
139
Fig. 1. The electrode array used to acquire the multi-channel surface EMG signals. The silver dot electrodes are equally spaced by an inter-electrode distance of 2.5 mm. During EMG acquisitions, the array is located parallel to the muscle fiber direction and held in place by applying a gentle pressure on the skin.
tional therapy focused on sensory, visual and motor stimulation of the grafted hand. The functional outcome of the transplanted hand was assessed by clinical examination including sensory and motor tests. Monitoring of nerve recovery and regeneration was performed by observing the proximal to distal advance of Tinel’s sign. Motor recovery of the intrinsic muscles of the hand was evaluated using surface multi-channel EMG recordings starting on day 205 post-transplantation. A linear electrode array (16 dot electrodes, 2.5 mm interelectrode distance; Fig. 1) [6,16,19] was used in this study to assess the electrical activity of the intrinsic muscles of the transplanted hand. Surface EMG signals were amplified and band-pass filtered (−3 dB bandwidth, 10–500 Hz) by a multi-channel surface electromyograph (EMG-16, LISiNPrima Biomedical & Sport, Treviso, Italy), sampled at 2048 samples/s, converted to 12-bit digital data, displayed in real-time, and stored on a PC for further processing. The first set of multi-channel surface EMG recordings was obtained at day 205 (7 months) post-operatively, followed by a second evaluation at 11 months and then monthly thereafter, up to 10 sessions. An additional session was then performed 4 months after the 10th session. The muscles investigated were the abductor digiti minimi, abductor pollicis brevis, opponens pollicis, first dorsal interosseous, and first lumbricalis of the transplanted hand. The electrode array was held in place by an operator who demonstrated to the subject the specific movement to perform and provided an appropriate counter-resistance. If EMG activity was present, the final positioning of the electrode array was determined by asking the subject to perform short contractions while EMG signal was visually assessed. The array electrode was positioned in such a way that the motor unit ac-
140
M. Lanzetta et al. / Neuroscience Letters 373 (2005) 138–143
tion potentials were seen to propagate from the innervation zone to the tendons with minimal change in the waveform [16]. In the absence of EMG, the array was placed along the muscle fiber direction, as estimated by muscle palpation. For each muscle, the subject performed a 60-s long contraction at maximal level. The subject was verbally encouraged to maintain the maximal drive to the muscle during the 60-s long contractions. The same contraction was performed two other times with 5-min breaks in between. When motor unit action potentials could be identified, the subject performed three additional contractions linearly increasing force from zero to the maximum, as determined subjectively. When a single motor unit action potential train could be identified with potentials significantly larger than the background noise, the subject was provided with a visual feedback of the average discharge rate of this motor unit. A 5-min rest was given to the subject after each ramp contraction. For each innervated muscle, the location of the innervation zone of the active motor units was assessed by visual analysis of the surface EMG signals [15]. It corresponded to the point of inversion of propagation of the motor unit action potentials [17]. Single motor unit action potentials were identified from the surface EMG signals with an algorithm based on threshold and classification. The discharge rate of each motor unit was estimated as the inverse of the time interval between consecutive discharges. Three months after transplantation, the grafted hand began to show normal skin color and texture, as well as normal hair and nail growth. Arterial blood supply and venous outflow were normal. Nerve regeneration started immediately after transplantation and recovery of protective sensory functions (i.e., ability to detect pain, thermal stimuli and gross tactile sensation) was already present in the early post-operative period (3 months). There were no differences in protective sensory recovery in the radial or ulnar cutaneous areas of the hand, innervated by the median and ulnar nerves, respectively. Nerve regeneration progressed with time allowing a certain degree of discriminative sensation, although with some localization discrepancies. At 2 years post-transplantation the hand was evaluated for static two-point fingertip sensory discrimination, showing a satisfactory recovery according to the Highet scale as modified by Dellon [4] (grade S3; >15 mm). Functional magnetic resonance imaging 6 months postoperatively revealed activation in a previously quiescent region of sensorimotor cortex thought to represent hand movement. At 12 months post-transplantation, the patients showed the ability to perform a number of daily manual activities, including eating, driving, grasping objects, riding a bicycle or a motorbike, shaving, phoning and writing. Eleven months after transplant, the first clear motor unit potential train was detected from abductor digiti minimi muscle (Fig. 2A). Surface EMG analysis allowed detection of the point in which axon enters the motor end-plate (Fig. 2F) [16]. After 13 months, a second motor unit appeared during max-
imal contractions of the same muscle. Surface potentials of this second motor unit showed significantly smaller amplitudes than those of the first (Fig. 2A), indicating either a deeper or a smaller motor unit. No other motor units were observed in the abductor digiti minimi muscle until the last session reported in this study. Twelve months after transplant, abductor pollicis and opponens pollicis muscles showed surface EMG activity (Fig. 2B and C). A single clear motor unit action potential train was observed in opponens pollicis muscle, while three motor units were detected from abductor pollicis. After 15 months, the first dorsal interosseous muscle showed the first active motor unit (Fig. 2D). As for the case of the abductor digiti minimi and opponens pollicis, a single train of action potentials could be identified from this muscle. After 19 months, a motor unit activity was detected from the first lumbricalis, although, as for the case of the first dorsal interosseous, the action potential amplitude was only slightly larger than the background noise. For the abductor digiti minimi, abductor pollicis, and opponens pollicis it was possible to identify the motor unit innervation zones (Fig. 2F). For all muscles, after the first motor unit reinnervation sign, motor unit activity could be observed in all subsequent measurement sessions. Only in abductor digiti minimi was it possible to use the surface detected action potentials as a feedback for the subject. In the other muscles, the action potentials were too small for effective feedback or there were 2–3 active motor units, which hindered the possibility of a selective feedback. Moreover, the abductor digiti minimi presented a clearly distinguishable activity of a single motor unit (Fig. 3), which allowed the detection of all the action potentials of this unit and accurate reconstruction of its entire discharge pattern. During the 60-s long contractions of the abductor digiti minimi, sustained with the maximal drive to the muscle, motor unit discharge rate decreased, despite the verbal encouragement given to the subject to keep it at the initial level (Fig. 4). This was observed in all experimental sessions since the beginning of the reinnervation process. On average (over the 10 experimental sessions in which motor unit activity was detected from the abductor digiti minimi muscle), the initial (first 5 s) discharge rate during maximal contraction (mean ± S.D.) was 34.0 ± 6.7 pulses per second (pps), significantly higher (Student’s t-test for depended samples, P < 0.001) than the discharge rate at the end of the contraction (23.4 ± 5.1 pps after 60 s). The subject was able to increase the frequency of activation of the reinnervated motor unit approximately linearly over time (Fig. 5). The subject could perform this simple motor control task since the beginning of the reinnervation with only a visual feedback on the discharge rate. The average discharge rate at the beginning (first 5 s), middle (between 30 and 35 s), and end (last 5 s) of the contractions with feedback (over all experimental sessions) was, respectively, 13.4 ± 6.7 pps, 25.4 ± 12.5 pps, and 32.5 ± 11.2 pps. The dis-
M. Lanzetta et al. / Neuroscience Letters 373 (2005) 138–143
141
Fig. 2. Multi-channel surface EMG signals acquired from intrinsic muscles of the transplanted hand during attempted maximal voluntary contractions against the resistance of the operator. For each muscle, the date when voluntary EMG activity was observed for the first time is indicated. A 16 channel, 2.5 mm inter-electrode distance array with silver dot electrodes (shown in Fig. 1) was used to record the EMG signals. Only the channels with high enough signal quality were plotted in each case. Note the different amplitude scale for each graph. The investigated muscles were: (A) abductor digiti minimi, (B) abductor pollicis brevis, (C) opponens pollicis, (D) first dorsalis interosseus, and (E) first lumbricalis. The orientation of the array for the investigated muscles (except the first dorsalis interosseus) is shown in (F). For each muscle, the two crosses (+) represent the location of electrodes 1 and 16 of the array, and the dashed line (- - -) indicates the array direction. For the muscles whose signal quality and number of propagating channels was high enough, the estimated position of the innervation zone (䊉 IZ) is also marked. SD stands for single differential.
charge rates at these three intervals of time were significantly different from each other (one-way repeated measures analysis of variance with Student–Newman–Keuls post hoc test, P < 0.05). The most important result in hand transplantation has been the finding that nerve regeneration takes place rapidly in
transplanted hands, even many years after the trauma. The stumps of the peripheral nerves to the hand have the ability to recommence their growth towards a more distal target [3,11]. In this study we report direct assessment of muscle properties during a reinnervation process following hand transplantation through the analysis of individual motor units. It is the first
142
M. Lanzetta et al. / Neuroscience Letters 373 (2005) 138–143
Fig. 3. Multi-channel surface EMG signals acquired from the abductor digiti minimi muscle of the transplanted hand during attempted maximum voluntary contraction. SD stands for single differential.
time that reinnervation and control properties of single motor units have been documented in a transplanted hand, 22 years after amputation. Anatomical information about the muscle were obtained by the localization of the innervation zones of the detected motor units (this was possible in three out of the five intrinsic muscles investigated), as proposed in previous work [15]. The detected action potentials varied in shape for the different muscles analyzed. In particular, the abductor digiti
Fig. 4. Surface EMG signals detected at the beginning (A) and end (B) (only one channels shown for clarity) of a contraction sustained at the maximal level by the subject with the abductor digiti minimi muscle. (C) Instantaneous discharge rate during the maximal contraction.
Fig. 5. Surface EMG signals acquired from the abductor digiti minimi muscle of the transplanted hand during a 60 s long voluntary ramp contraction. The subject was given a real-time feedback of the average discharge rate of its active motor unit and was instructed to follow a linearly increasing target. (A) Time course of one EMG channel. Note the constant amplitude due to the only active motor unit contributing to the signal. (B–D) Epochs of EMG signals (three channels shown), 1-s long, extracted from the signal at the beginning (11.0–12.0 s), middle (30.0–31.0 s) and end (52.0–53.0 s) of the ramp contraction. Note the increase in the discharge rate. (E) All the motor unit action potentials extracted from the signal (dark grey lines) on seven channels; the average motor unit action potential is shown superimposed (black lines). SD stands for single differential.
minimi showed motor unit action potentials with longer duration than the other muscles. Opponens pollicis presented potentials with the smallest duration. This difference may be due to many factors, whose further investigation may provide important information on the reinnervation process. Short action potentials may be generated by motor units with muscle fibers innervated in almost the same location; in this case the spread of single fiber potentials is minimal. Other factors which affect the potential duration are the subcutaneous layer thickness, the depth of the motor unit, and its conduction velocity. Focal muscle injury (i.e., focal myopathy) or focal injury of the motor end-plates may also affect action potential duration. The control properties of the detected motor units were investigated by the analysis of their discharge patterns, which were similar to those of normal subjects [1] with discharge frequencies in the range 13–34 pps. During the maximal contractions, motor unit discharge rate decreased over time, despite verbal encouragement to maintain maximum effort. A decrease in central activation may explain this phenomenon. Another possibility is that the decrease in discharge rate was mediated by reflex mechanisms due to the feedback from muscle afferents. Decrease in motor unit discharge rates with sustained contraction is observed also in normal subjects, both with decreasing [2] and constant [5,14] force. The phenomenon is probably mediated by a reflex inhibition of the small diameter afferents (groups III and IV) [2] and by a decreased spindle support to alpha-motoneurones [14]. Indeed, in deafferented muscles of healthy subjects, the discharge fre-
M. Lanzetta et al. / Neuroscience Letters 373 (2005) 138–143
quencies of motor axons do not progressively decline as occurs with normally-innervated motor units during sustained contractions [13]. Thus, feedback from muscle afferents was probably intact in the transplanted hand at the time of reinnervation, in agreement with the early recovery of sensory inputs. From the earliest detectable stages of reinnervation, the subject was able to modulate the discharge rate of single motor units with a visual feedback. The control of motorneuron discharge rates is possible in normal subjects with no proprioceptive input other than knowledge of the motor commands [10]. Thus, the motor unit control does not necessarily imply intact muscle afferent feedback. However, without proprioceptive input, during attempted maximal efforts, the motor axon discharge frequencies are significantly lower than those of normally-innervated motor units [10]. Since the average discharge rates at the beginning of maximal contractions were higher in the investigated patient than those observed from healthy subjects during acute deafferentation of the muscles [10,13], the mechanisms involving muscle afferent facilitation to the motor unit discharge were probably intact. In conclusion, we showed, by advanced non-invasive EMG techniques, that reinnervation of single motor units occurs in a transplanted hand after 22 years of denervation. Selective assessment at the motor unit level of intrinsic muscles in the transplanted hand is, thus, feasible. This provides important clinical and neurophysiological information and suggests new directions for research in the areas of limb transplantation and motor control studies. Acknowledgements The authors are sincerely grateful to Enrico Merlo (Laboratorio di Ingegneria del Sistema Neuromuscolare (LISiN), Torino, Italy) Luisa Stroppa (Unit`a di Rieducazione della Mano, Ospedale S. Gerardo, Monza, Italy), and Velio Macellari (Istituto Superiore di Sanit`a, Roma, Italy).
References [1] F. Bellemare, J.J. Woods, R. Johansson, B. Bigland-Ritchie, Motorunit discharge rates in maximal voluntary contractions of three human muscles, J. Neurophysiol. 50 (1983) 1380–1392. [2] B. Bigland-Ritchie, N.J. Dawson, R.S. Johansson, O.C. Lippold, Reflex origin for the slowing of motoneurone firing rates in fatigue of human voluntary contractions, J. Physiol. 379 (1986) 451–459.
143
[3] R. Chunasuwankul, C. Ayrout, Z. Dereli, A. Gal, M. Lanzetta, E. Owen, Low dose discontinued FK506 treatment enhances peripheral nerve regeneration, Int. Surg. 87 (2002) 274–278. [4] A.L. Dellon, A numerical grading scale for peripheral nerve function, J. Hand Ther. 6 (1993) 152–160. [5] C.J. De Luca, P.J. Foley, Z. Erim, Motor unit control properties in constant-force isometric contractions, J. Neurophysiol. 76 (1996) 1503–1516. [6] G. Drost, J.H. Blok, D.F. Stegeman, J.P. van Dijk, B.G. van Engelen, M.J. Zwarts, Propagation disturbance of motor unit action potentials during transient paresis in generalized myotonia: a high-density surface EMG study, Brain 124 (2001) 352–360. [7] J.M. Dubernard, E. Owen, G. Herzberg, M. Lanzetta, X. Martin, H. Kapila, M. Dawahra, N.S. Hakim, Human hand allograft: report on first 6 months, Lancet 353 (1999) 1315–1320. [8] J.M. Dubernard, E. Owen, M. Lanzetta, N. Hakim, What is happening with hand transplants? Lancet 357 (2001) 1711–1712. [9] J.M. Dubernard, E. Owen, N. Lefrancois, P. Petruzzo, X. Martin, M. Dawahra, D. Jullien, J. Kanitakis, C. Frances, X. Preville, L. Gebuhrer, N. Hakim, M. Lanzetta, H. Kapila, G. Herzberg, J.P. Revillard, First human hand transplantation. Case report, Transplant. Int. 13 (2000) 521–524. [10] S.C. Gandevia, G. Macefield, D. Burke, D.K. McKenzie, Voluntary activation of human motor axons in the absence of muscle afferent feedback. The control of the deafferented hand, Brain 113 (1990) 1563–1581. [11] B.G. Gold, T. Storm-Dickerson, D.R. Austin, The immunosuppressant FK506 increases functional recovery and nerve regeneration following peripheral nerve injury, Rest. Neurol. Neurosci. 6 (1993) 287–296. [12] M. Lanzetta, R. Nolli, A. Borgonovo, E. Owen, J.M. Dubernard, H. Kapila, X. Martin, N. Hakim, M. Dawahra, Hand transplantation: immunosuppression, ethics and indications, J. Hand Surg. 26 (2001) 511–516. [13] G. Macefield, S.C. Gandevia, B. Bigland-Ritchie, R.B. Gorman, D. Burke, The firing rates of human motoneurones voluntarily activated in the absence of muscle afferent feedback, J. Physiol. 471 (1993) 429–443. [14] G. Macefield, K.E. Hagbarth, R. Gorman, S.C. Gandevia, D. Burke, Decline in spindle support to alpha-motoneurones during sustained voluntary contractions, J. Physiol. 440 (1991) 497–512. [15] T. Masuda, H. Miyano, T. Sadoyama, The position of innervation zones in the biceps brachii investigated by surface electromyography, IEEE Trans. Biomed. Eng. 32 (1985) 36–42. [16] R. Merletti, D. Farina, M. Gazzoni, The linear electrode array: a useful tool with many applications, J. Electromyogr. Kinesiol. 13 (2003) 37–47. [17] R. Merletti, D. Farina, A. Granata, Non-invasive assessment of motor unit properties with linear electrode arrays, Electroencephalogr. Clin. Neurophysiol. 50 (1999) 293–300. [18] E. Owen, J.M. Dubernard, M. Lanzetta, H. Kapila, X. Martin, M. Dawahra, N.S. Hakim, Peripheral nerve regeneration in human hand transplantation, Transplant. Proc. 33 (2001) 1720– 1721. [19] M.J. Zwarts, D.F. Stegeman, Multichannel surface EMG: basic aspects and clinical utility, Muscle Nerve 28 (2003) 1–17.
Reinnervation of motor units in intrinsic muscles of a transplanted hand Marco Lanzettaa , Marco Pozzob , Andrea Bottinb , Roberto Merlettib , Dario Farinac,∗ b
a Hand Surgery and Reconstructive Microsurgery Unit, San Gerardo Hospital, University of Milan-Bicocca, Monza, Italy Laboratorio di Ingegneria del Sistema Neuromuscolare, Centro di Bioingegneria, Dip. di Elettronica, Politecnico di Torino, Torino, Italy c Center for Sensory-Motor Interaction (SMI), Department of Health Science and Technology, Aalborg University, Fredrik Bajers Vej 7 D-3, DK-9220 Aalborg, Denmark
Received 6 August 2004; received in revised form 27 September 2004; accepted 1 October 2004
Abstract Functional recovery of transplanted hand can be evaluated clinically but until now there has been no direct assessment of muscle control. In October 2000 we transplanted the right hand of a brain-dead man aged 43 onto a man aged 35 who had lost his right dominant hand 22 years before. Starting from day 205 after the transplant, multi-channel surface electromyographic (EMG) signals were recorded from intrinsic muscles of the transplanted hand in order to assess their degree of reinnervation. Eleven months post-operatively, the first motor unit action potential train was detected from the abductor digiti minimi. One month later, also the abductor pollicis brevis and the opponens pollicis muscles showed motor unit activity, while, after 15 and 24 months, the first dorsal interosseous and the first lumbricalis muscles, respectively, showed activation of their first motor units. An increase in the number of active motor units was observed after the first signs of reinnervation, although the process was rather slow. In sustained maximal contractions, the motor unit discharge rate decreased from (mean ± S.D.) 34.0 ± 6.7 pps to 23.4 ± 5.1 pps in 60 s for the abductor digiti minimi, although the subject was verbally encouraged to maintain a maximal activation. Moreover, the subject was able to perform basic control tasks involving voluntary modulation of motor unit discharge rate. With a visual feedback, he could increase discharge rate of the abductor digiti minimi approximately linearly over time, from 13.4 ± 6.7 pps to 32.5 ± 11.2 pps in 60 s. In conclusion, we showed reinnervation of single motor units in a transplanted hand after 22 years of denervation. Moreover, voluntary modulation of discharge rates of these motor units could be performed since the first sign of reinnervation. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Hand transplantation; Multi-channel surface electromyography; Motor unit; Discharge rate
More than 20 hand/digital transplantations have been performed worldwide [8]. All patients are currently healthy with viable grafts, except for the first, who experienced rejection after he arbitrarily stopped taking immunosuppressant drugs [8]. Sensory and motor recovery in these patients is contingent on nerve regeneration [18], shown to take place rapidly in transplanted hands even many years after the trauma. Sensory and motor nerve regeneration are usually assessed by clinical assessments such as strength testing of intrinsic hand muscles, Tinel’s sign, Semmes–Weinstein microfilaments, pinprick, hot and cold temperature, and light and deep pressure [18]. However, until now there has been no direct assessment of muscle control at the level of motor units. The analysis ∗
Corresponding author. Tel.: +45 963 58821; fax: +45 981 54008. E-mail address: [email protected] (D. Farina).
0304-3940/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2004.10.001
of motor unit properties in intrinsic muscles after transplant would provide further insight into the plasticity of the neuromuscular system. Electrophysiological techniques, such as intramuscular or surface electromyography (EMG), long used to assess segmental nerve damage, can be applied to better assess post-surgical nerve regeneration. Non-invasive approaches are preferred in hand-transplanted subjects, to minimize possible damage to the allograft and the risk of infections. The main objective of this study was to assess the reinnervation process in intrinsic muscles after hand transplantation at the motor unit level. This objective was achieved by noninvasive multi-channel surface EMG techniques. The Italian Hand Transplantation Program was approved by the Italian Health Ministry and the Ethical Committee of
M. Lanzetta et al. / Neuroscience Letters 373 (2005) 138–143
the University of Milan-Bicocca. Approximately 400 possible candidates were interviewed over a period of 2 years and studied under a preliminary protocol using strict inclusion criteria. The selected recipient was a 35-year-old male who had lost his right dominant hand in a farming accident when he was 13. The patient underwent a series of routine pretransplant investigations, including an angiogram, CT scan, muscle and nerve charts, magnetic resonance imaging of the stump, functional magnetic resonance imaging of the brain, and a comprehensive battery of psychological and clinical tests. The multi-organ cadaveric donor was a 43-year-old male who died from a stroke. He had the same blood group (O positive) as the recipient and there were six HLA mismatches; the cross-match was negative. The surgical procedure was similar to that followed during hand replantation, and included bone fixation, revascularization with microsurgical anastomoses of arteries and veins, tendon repair or tendon transfer when necessary, microsurgical nerve repair, and skin closure [7]. The transplant operation lasted 12 h with a total ischemia time of 11 h. The forearm bones were transversely cut so that the grafted limb would be of the same length as the contralateral. Bone fixation of the radius and ulna was achieved by means of compression plates and 4.5 mm screws, and autologous cancellous bone graft from the iliac crest was placed around the osteosynthesis site to improve healing. Most of the deep flexor and all extensor tendons were repaired. The hand was then revascularized by anastomosing the radial and ulnar arteries, as well as three veins. The median and ulnar nerves were repaired using a conventional epineurial technique. Lastly, the remaining more superficial flexor tendons were sutured. The skin was closed in layers. The patient weighed 73 kg and was given 250 ml of Dextran 40 before declamping and 20 ml/h for 7 days. Aspirin 150 was administered for 7 days, and wide-spectrum antibiotic therapy for 10 days. The induction immunosuppressive protocol consisted of 20 mg of monoclonal antibody antiCD25 (Baxilimab-Simulect® ) 2 h before the operation, on day 4 and on day 45 post-operatively, FK506 (TacrolimusPrograf® ), adjusted to maintain blood concentration between 15 and 20 mg/ml for the first month, mycophenolate mofetil (Cell Cept® 2 g/day), and steroid (prednisone 250 mg on day 1 and rapidly tapered to 20 mg/day). The maintenance therapy consisted of FK506 (blood levels between 5 and 10 mg/ml), mycophenolate mofetil 1 g/day and prednisone 10 mg/day [7,8,9,12]. The rehabilitation protocol started as soon as swelling subsided and was performed twice daily for 138 days, and once daily thereafter as the patient returned to work. It included a standard rehabilitation program for flexor and extensor tendons, sensory re-education and cortical re-integration [7]. In the early phases, hand movement was initiated passively, then with active-assisted exercises, and finally with active and against-resistance exercises. Electrical stimulation was started on day 60 and carried out twice daily since. Occupa-
139
Fig. 1. The electrode array used to acquire the multi-channel surface EMG signals. The silver dot electrodes are equally spaced by an inter-electrode distance of 2.5 mm. During EMG acquisitions, the array is located parallel to the muscle fiber direction and held in place by applying a gentle pressure on the skin.
tional therapy focused on sensory, visual and motor stimulation of the grafted hand. The functional outcome of the transplanted hand was assessed by clinical examination including sensory and motor tests. Monitoring of nerve recovery and regeneration was performed by observing the proximal to distal advance of Tinel’s sign. Motor recovery of the intrinsic muscles of the hand was evaluated using surface multi-channel EMG recordings starting on day 205 post-transplantation. A linear electrode array (16 dot electrodes, 2.5 mm interelectrode distance; Fig. 1) [6,16,19] was used in this study to assess the electrical activity of the intrinsic muscles of the transplanted hand. Surface EMG signals were amplified and band-pass filtered (−3 dB bandwidth, 10–500 Hz) by a multi-channel surface electromyograph (EMG-16, LISiNPrima Biomedical & Sport, Treviso, Italy), sampled at 2048 samples/s, converted to 12-bit digital data, displayed in real-time, and stored on a PC for further processing. The first set of multi-channel surface EMG recordings was obtained at day 205 (7 months) post-operatively, followed by a second evaluation at 11 months and then monthly thereafter, up to 10 sessions. An additional session was then performed 4 months after the 10th session. The muscles investigated were the abductor digiti minimi, abductor pollicis brevis, opponens pollicis, first dorsal interosseous, and first lumbricalis of the transplanted hand. The electrode array was held in place by an operator who demonstrated to the subject the specific movement to perform and provided an appropriate counter-resistance. If EMG activity was present, the final positioning of the electrode array was determined by asking the subject to perform short contractions while EMG signal was visually assessed. The array electrode was positioned in such a way that the motor unit ac-
140
M. Lanzetta et al. / Neuroscience Letters 373 (2005) 138–143
tion potentials were seen to propagate from the innervation zone to the tendons with minimal change in the waveform [16]. In the absence of EMG, the array was placed along the muscle fiber direction, as estimated by muscle palpation. For each muscle, the subject performed a 60-s long contraction at maximal level. The subject was verbally encouraged to maintain the maximal drive to the muscle during the 60-s long contractions. The same contraction was performed two other times with 5-min breaks in between. When motor unit action potentials could be identified, the subject performed three additional contractions linearly increasing force from zero to the maximum, as determined subjectively. When a single motor unit action potential train could be identified with potentials significantly larger than the background noise, the subject was provided with a visual feedback of the average discharge rate of this motor unit. A 5-min rest was given to the subject after each ramp contraction. For each innervated muscle, the location of the innervation zone of the active motor units was assessed by visual analysis of the surface EMG signals [15]. It corresponded to the point of inversion of propagation of the motor unit action potentials [17]. Single motor unit action potentials were identified from the surface EMG signals with an algorithm based on threshold and classification. The discharge rate of each motor unit was estimated as the inverse of the time interval between consecutive discharges. Three months after transplantation, the grafted hand began to show normal skin color and texture, as well as normal hair and nail growth. Arterial blood supply and venous outflow were normal. Nerve regeneration started immediately after transplantation and recovery of protective sensory functions (i.e., ability to detect pain, thermal stimuli and gross tactile sensation) was already present in the early post-operative period (3 months). There were no differences in protective sensory recovery in the radial or ulnar cutaneous areas of the hand, innervated by the median and ulnar nerves, respectively. Nerve regeneration progressed with time allowing a certain degree of discriminative sensation, although with some localization discrepancies. At 2 years post-transplantation the hand was evaluated for static two-point fingertip sensory discrimination, showing a satisfactory recovery according to the Highet scale as modified by Dellon [4] (grade S3; >15 mm). Functional magnetic resonance imaging 6 months postoperatively revealed activation in a previously quiescent region of sensorimotor cortex thought to represent hand movement. At 12 months post-transplantation, the patients showed the ability to perform a number of daily manual activities, including eating, driving, grasping objects, riding a bicycle or a motorbike, shaving, phoning and writing. Eleven months after transplant, the first clear motor unit potential train was detected from abductor digiti minimi muscle (Fig. 2A). Surface EMG analysis allowed detection of the point in which axon enters the motor end-plate (Fig. 2F) [16]. After 13 months, a second motor unit appeared during max-
imal contractions of the same muscle. Surface potentials of this second motor unit showed significantly smaller amplitudes than those of the first (Fig. 2A), indicating either a deeper or a smaller motor unit. No other motor units were observed in the abductor digiti minimi muscle until the last session reported in this study. Twelve months after transplant, abductor pollicis and opponens pollicis muscles showed surface EMG activity (Fig. 2B and C). A single clear motor unit action potential train was observed in opponens pollicis muscle, while three motor units were detected from abductor pollicis. After 15 months, the first dorsal interosseous muscle showed the first active motor unit (Fig. 2D). As for the case of the abductor digiti minimi and opponens pollicis, a single train of action potentials could be identified from this muscle. After 19 months, a motor unit activity was detected from the first lumbricalis, although, as for the case of the first dorsal interosseous, the action potential amplitude was only slightly larger than the background noise. For the abductor digiti minimi, abductor pollicis, and opponens pollicis it was possible to identify the motor unit innervation zones (Fig. 2F). For all muscles, after the first motor unit reinnervation sign, motor unit activity could be observed in all subsequent measurement sessions. Only in abductor digiti minimi was it possible to use the surface detected action potentials as a feedback for the subject. In the other muscles, the action potentials were too small for effective feedback or there were 2–3 active motor units, which hindered the possibility of a selective feedback. Moreover, the abductor digiti minimi presented a clearly distinguishable activity of a single motor unit (Fig. 3), which allowed the detection of all the action potentials of this unit and accurate reconstruction of its entire discharge pattern. During the 60-s long contractions of the abductor digiti minimi, sustained with the maximal drive to the muscle, motor unit discharge rate decreased, despite the verbal encouragement given to the subject to keep it at the initial level (Fig. 4). This was observed in all experimental sessions since the beginning of the reinnervation process. On average (over the 10 experimental sessions in which motor unit activity was detected from the abductor digiti minimi muscle), the initial (first 5 s) discharge rate during maximal contraction (mean ± S.D.) was 34.0 ± 6.7 pulses per second (pps), significantly higher (Student’s t-test for depended samples, P < 0.001) than the discharge rate at the end of the contraction (23.4 ± 5.1 pps after 60 s). The subject was able to increase the frequency of activation of the reinnervated motor unit approximately linearly over time (Fig. 5). The subject could perform this simple motor control task since the beginning of the reinnervation with only a visual feedback on the discharge rate. The average discharge rate at the beginning (first 5 s), middle (between 30 and 35 s), and end (last 5 s) of the contractions with feedback (over all experimental sessions) was, respectively, 13.4 ± 6.7 pps, 25.4 ± 12.5 pps, and 32.5 ± 11.2 pps. The dis-
M. Lanzetta et al. / Neuroscience Letters 373 (2005) 138–143
141
Fig. 2. Multi-channel surface EMG signals acquired from intrinsic muscles of the transplanted hand during attempted maximal voluntary contractions against the resistance of the operator. For each muscle, the date when voluntary EMG activity was observed for the first time is indicated. A 16 channel, 2.5 mm inter-electrode distance array with silver dot electrodes (shown in Fig. 1) was used to record the EMG signals. Only the channels with high enough signal quality were plotted in each case. Note the different amplitude scale for each graph. The investigated muscles were: (A) abductor digiti minimi, (B) abductor pollicis brevis, (C) opponens pollicis, (D) first dorsalis interosseus, and (E) first lumbricalis. The orientation of the array for the investigated muscles (except the first dorsalis interosseus) is shown in (F). For each muscle, the two crosses (+) represent the location of electrodes 1 and 16 of the array, and the dashed line (- - -) indicates the array direction. For the muscles whose signal quality and number of propagating channels was high enough, the estimated position of the innervation zone (䊉 IZ) is also marked. SD stands for single differential.
charge rates at these three intervals of time were significantly different from each other (one-way repeated measures analysis of variance with Student–Newman–Keuls post hoc test, P < 0.05). The most important result in hand transplantation has been the finding that nerve regeneration takes place rapidly in
transplanted hands, even many years after the trauma. The stumps of the peripheral nerves to the hand have the ability to recommence their growth towards a more distal target [3,11]. In this study we report direct assessment of muscle properties during a reinnervation process following hand transplantation through the analysis of individual motor units. It is the first
142
M. Lanzetta et al. / Neuroscience Letters 373 (2005) 138–143
Fig. 3. Multi-channel surface EMG signals acquired from the abductor digiti minimi muscle of the transplanted hand during attempted maximum voluntary contraction. SD stands for single differential.
time that reinnervation and control properties of single motor units have been documented in a transplanted hand, 22 years after amputation. Anatomical information about the muscle were obtained by the localization of the innervation zones of the detected motor units (this was possible in three out of the five intrinsic muscles investigated), as proposed in previous work [15]. The detected action potentials varied in shape for the different muscles analyzed. In particular, the abductor digiti
Fig. 4. Surface EMG signals detected at the beginning (A) and end (B) (only one channels shown for clarity) of a contraction sustained at the maximal level by the subject with the abductor digiti minimi muscle. (C) Instantaneous discharge rate during the maximal contraction.
Fig. 5. Surface EMG signals acquired from the abductor digiti minimi muscle of the transplanted hand during a 60 s long voluntary ramp contraction. The subject was given a real-time feedback of the average discharge rate of its active motor unit and was instructed to follow a linearly increasing target. (A) Time course of one EMG channel. Note the constant amplitude due to the only active motor unit contributing to the signal. (B–D) Epochs of EMG signals (three channels shown), 1-s long, extracted from the signal at the beginning (11.0–12.0 s), middle (30.0–31.0 s) and end (52.0–53.0 s) of the ramp contraction. Note the increase in the discharge rate. (E) All the motor unit action potentials extracted from the signal (dark grey lines) on seven channels; the average motor unit action potential is shown superimposed (black lines). SD stands for single differential.
minimi showed motor unit action potentials with longer duration than the other muscles. Opponens pollicis presented potentials with the smallest duration. This difference may be due to many factors, whose further investigation may provide important information on the reinnervation process. Short action potentials may be generated by motor units with muscle fibers innervated in almost the same location; in this case the spread of single fiber potentials is minimal. Other factors which affect the potential duration are the subcutaneous layer thickness, the depth of the motor unit, and its conduction velocity. Focal muscle injury (i.e., focal myopathy) or focal injury of the motor end-plates may also affect action potential duration. The control properties of the detected motor units were investigated by the analysis of their discharge patterns, which were similar to those of normal subjects [1] with discharge frequencies in the range 13–34 pps. During the maximal contractions, motor unit discharge rate decreased over time, despite verbal encouragement to maintain maximum effort. A decrease in central activation may explain this phenomenon. Another possibility is that the decrease in discharge rate was mediated by reflex mechanisms due to the feedback from muscle afferents. Decrease in motor unit discharge rates with sustained contraction is observed also in normal subjects, both with decreasing [2] and constant [5,14] force. The phenomenon is probably mediated by a reflex inhibition of the small diameter afferents (groups III and IV) [2] and by a decreased spindle support to alpha-motoneurones [14]. Indeed, in deafferented muscles of healthy subjects, the discharge fre-
M. Lanzetta et al. / Neuroscience Letters 373 (2005) 138–143
quencies of motor axons do not progressively decline as occurs with normally-innervated motor units during sustained contractions [13]. Thus, feedback from muscle afferents was probably intact in the transplanted hand at the time of reinnervation, in agreement with the early recovery of sensory inputs. From the earliest detectable stages of reinnervation, the subject was able to modulate the discharge rate of single motor units with a visual feedback. The control of motorneuron discharge rates is possible in normal subjects with no proprioceptive input other than knowledge of the motor commands [10]. Thus, the motor unit control does not necessarily imply intact muscle afferent feedback. However, without proprioceptive input, during attempted maximal efforts, the motor axon discharge frequencies are significantly lower than those of normally-innervated motor units [10]. Since the average discharge rates at the beginning of maximal contractions were higher in the investigated patient than those observed from healthy subjects during acute deafferentation of the muscles [10,13], the mechanisms involving muscle afferent facilitation to the motor unit discharge were probably intact. In conclusion, we showed, by advanced non-invasive EMG techniques, that reinnervation of single motor units occurs in a transplanted hand after 22 years of denervation. Selective assessment at the motor unit level of intrinsic muscles in the transplanted hand is, thus, feasible. This provides important clinical and neurophysiological information and suggests new directions for research in the areas of limb transplantation and motor control studies. Acknowledgements The authors are sincerely grateful to Enrico Merlo (Laboratorio di Ingegneria del Sistema Neuromuscolare (LISiN), Torino, Italy) Luisa Stroppa (Unit`a di Rieducazione della Mano, Ospedale S. Gerardo, Monza, Italy), and Velio Macellari (Istituto Superiore di Sanit`a, Roma, Italy).
References [1] F. Bellemare, J.J. Woods, R. Johansson, B. Bigland-Ritchie, Motorunit discharge rates in maximal voluntary contractions of three human muscles, J. Neurophysiol. 50 (1983) 1380–1392. [2] B. Bigland-Ritchie, N.J. Dawson, R.S. Johansson, O.C. Lippold, Reflex origin for the slowing of motoneurone firing rates in fatigue of human voluntary contractions, J. Physiol. 379 (1986) 451–459.
143
[3] R. Chunasuwankul, C. Ayrout, Z. Dereli, A. Gal, M. Lanzetta, E. Owen, Low dose discontinued FK506 treatment enhances peripheral nerve regeneration, Int. Surg. 87 (2002) 274–278. [4] A.L. Dellon, A numerical grading scale for peripheral nerve function, J. Hand Ther. 6 (1993) 152–160. [5] C.J. De Luca, P.J. Foley, Z. Erim, Motor unit control properties in constant-force isometric contractions, J. Neurophysiol. 76 (1996) 1503–1516. [6] G. Drost, J.H. Blok, D.F. Stegeman, J.P. van Dijk, B.G. van Engelen, M.J. Zwarts, Propagation disturbance of motor unit action potentials during transient paresis in generalized myotonia: a high-density surface EMG study, Brain 124 (2001) 352–360. [7] J.M. Dubernard, E. Owen, G. Herzberg, M. Lanzetta, X. Martin, H. Kapila, M. Dawahra, N.S. Hakim, Human hand allograft: report on first 6 months, Lancet 353 (1999) 1315–1320. [8] J.M. Dubernard, E. Owen, M. Lanzetta, N. Hakim, What is happening with hand transplants? Lancet 357 (2001) 1711–1712. [9] J.M. Dubernard, E. Owen, N. Lefrancois, P. Petruzzo, X. Martin, M. Dawahra, D. Jullien, J. Kanitakis, C. Frances, X. Preville, L. Gebuhrer, N. Hakim, M. Lanzetta, H. Kapila, G. Herzberg, J.P. Revillard, First human hand transplantation. Case report, Transplant. Int. 13 (2000) 521–524. [10] S.C. Gandevia, G. Macefield, D. Burke, D.K. McKenzie, Voluntary activation of human motor axons in the absence of muscle afferent feedback. The control of the deafferented hand, Brain 113 (1990) 1563–1581. [11] B.G. Gold, T. Storm-Dickerson, D.R. Austin, The immunosuppressant FK506 increases functional recovery and nerve regeneration following peripheral nerve injury, Rest. Neurol. Neurosci. 6 (1993) 287–296. [12] M. Lanzetta, R. Nolli, A. Borgonovo, E. Owen, J.M. Dubernard, H. Kapila, X. Martin, N. Hakim, M. Dawahra, Hand transplantation: immunosuppression, ethics and indications, J. Hand Surg. 26 (2001) 511–516. [13] G. Macefield, S.C. Gandevia, B. Bigland-Ritchie, R.B. Gorman, D. Burke, The firing rates of human motoneurones voluntarily activated in the absence of muscle afferent feedback, J. Physiol. 471 (1993) 429–443. [14] G. Macefield, K.E. Hagbarth, R. Gorman, S.C. Gandevia, D. Burke, Decline in spindle support to alpha-motoneurones during sustained voluntary contractions, J. Physiol. 440 (1991) 497–512. [15] T. Masuda, H. Miyano, T. Sadoyama, The position of innervation zones in the biceps brachii investigated by surface electromyography, IEEE Trans. Biomed. Eng. 32 (1985) 36–42. [16] R. Merletti, D. Farina, M. Gazzoni, The linear electrode array: a useful tool with many applications, J. Electromyogr. Kinesiol. 13 (2003) 37–47. [17] R. Merletti, D. Farina, A. Granata, Non-invasive assessment of motor unit properties with linear electrode arrays, Electroencephalogr. Clin. Neurophysiol. 50 (1999) 293–300. [18] E. Owen, J.M. Dubernard, M. Lanzetta, H. Kapila, X. Martin, M. Dawahra, N.S. Hakim, Peripheral nerve regeneration in human hand transplantation, Transplant. Proc. 33 (2001) 1720– 1721. [19] M.J. Zwarts, D.F. Stegeman, Multichannel surface EMG: basic aspects and clinical utility, Muscle Nerve 28 (2003) 1–17.