New Type of Cortical Neuroplasticity After Nerve Repair in Brachial Plexus Lesions

OBSERVATION New Type of Cortical Neuroplasticity After Nerve Repair in Brachial Plexus Lesions

Roland Beisteiner, MD, MA; Ilse Ho¨llinger, MSc; Jakob Rath, MD; Moritz Wurnig, MSc; Markus Hilbert; Nicolaus Klinger, MSc; Alexander Geißler, MSc, PhD; Florian Fischmeister, PhD; Christian Wo¨ber, MD; Gerhard Klo¨sch, MSc; Hanno Millesi, MD; Wolfgang Grisold, MD; Eduard Auff, MD; Robert Schmidhammer, MD

Background: In brachial plexus avulsion, a recent tech- Participants: Three healthy control subjects, 2 pa- nique connects the ending of the disrupted musculocu- tients with phrenic nerve end-to-side coaptation, and 1 taneous nerve to the side of the intact phrenic nerve to control patient with C7 end-to-end coaptation (same clini- regain elbow flexion. This requires the phrenic nerve to cal presentation but phrenic nerve unchanged). perform a new double function: independent control of breathing and elbow flexion. Neuroplastic changes as- Results: Clinical documentation showed that both patients sociated with acquisition of double nerve functions have with phrenic nerve end-to-side coaptation were able to con- not yet been investigated. trolthediaphragmandthebicepsindependentlyviathesame phrenic nerve. In contrast to all controls, both patients with Objective: To evaluate neuroplastic changes associ- phrenic nerve end-to-side coaptation activated the cortical ated with acquisition of double nerve functions in a mono- diaphragm areas with flexion of the diseased arm. functional nerve (phrenic nerve). Conclusion: Our functional magnetic resonance imaging Design: Clinical and functional magnetic resonance data indicate that the patient’s cortical diaphragm areas imaging investigations during arm movements, forced in- reorganize in such a way that independent control of spiration, and motor control tasks. breathing and elbow flexion is possible with the same neuronal population. Setting: Investigations at the Medical University of Vienna, Vienna, Austria. Arch Neurol. 2011;68(11):1467-1470

WING TO TOTAL PARESIS nerve targets: muscles originally inner- of the affected arm, vated by the contralateral C7 root or the complete brachial ipsilateral phrenic nerve are denervated. Author Affiliations: Study plexus lesions represent An alternative approach is the nerve fi- Group Clinical fMRI and MR a seriously disabling ber transfer using end-to-side nerve re- 4 Center of Excellence neurologicalO condition. In cases of root pair, where the ending of the disrupted (Drs Beisteiner, Rath, Geißler, avulsion, reconstruction of nerve conti- nerve is attached to the side of an intact and Fischmeister, Ms Ho¨llinger, nuity is impossible. To regain at least nerve via an epineurial window. Re- and Messrs Wurnig, Hilbert, partial arm functions, nerve fiber trans- cently, a modified end-to-side procedure and Klinger) and Department of fers that connect axon donors from out- has been suggested where a small motor Neurology (Drs Beisteiner, side the brachial plexus with the affected branch is used to restore motor func- Rath, Geißler, Fischmeister, plexus nerves are increasingly per- tion.5 The major advantage of end-to- Wo¨ber, and Auff, Ms Ho¨llinger, formed. However, the detailed neuro- side nerve repair is preservation of donor and Messrs Wurnig, Hilbert, plastic changes associated with clinical nerve function. In brachial plexus avul- Klinger, and Klo¨sch), Medical 1 University of Vienna, Millesi recovery are yet unknown. There are 2 sion, the ending of the disrupted muscu- Center for Brachial Plexus and major nerve transfer procedures for res- locutaneous nerve can be attached to the Peripheral Nerve Surgery (and toration of elbow flexion in cases of root side of the intact phrenic nerve, thereby Rehabilitation) and Ludwig avulsion. Nerve fibers from the contralat- preserving diaphragm innervation. Such Boltzmann Institute for eral healthy C7 root and nerve fibers patients represent an interesting new Experimental and Clinical from the ipsilateral healthy phrenic model for neuroplasticity: the same Traumatology, Austrian Cluster nerve can be used for reinnervation of phrenic nerve is required to control breath- for Tissue Regeneration the musculocutaneous nerve.2,3 This ing and elbow flexion independently. Al- (Drs Millesi and allows patients to regain elbow flexion of though neuroplastic changes with bra- Schmidhammer), and the paretic arm via contralateral C7 chial plexus injury have already been Department of Neurology and 6,7 Ludwig Boltzmann Institute for inputs or via the ipsilateral phrenic reported, the neuroplastic changes as- Neurooncology, Kaiser Franz nerve. This technique, called end-to-end sociated with acquisition of an additional Josef Hospital (Dr Grisold), nerve repair, has 1 major disadvantage. nerve function are unknown. We hypoth- Vienna, Austria. There is a loss of function in the donor esized that cortical phrenic nerve (dia-

ARCH NEUROL / VOL 68 (NO. 11), NOV 2011 WWW.ARCHNEUROL.COM 1467

©2011 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021 Patient 2 was a right-handed man aged 43 years at the time of Control: inspiration a traumatic left complete brachial plexus lesion. He had end-to- side coaptation of (1) the ipsilateral phrenic nerve to (2) the musculocutaneous nerve 5 months after trauma. Left elbow flexion was possible against medium resistance at the time of fMRI (7 years after surgery). A general case description including preliminary data analysis is included in the article by Beisteiner et al.8 In both patients with end-to-side nerve repair, the nerve fiber transfer from the phrenic nerve to the musculocutaneous nerve was done using 2 sural nerve grafts coapted end to side to the phrenic nerve and end to end to the musculocutaneous Patient 1: inspiration nerve. Every patient provided fully informed consent with a protocol approved by the local ethics committee.

HEALTHY CONTROL SUBJECTS

To be able to compare fMRI findings of the patients’ reorga- nized nervous systems with unchanged nervous systems, fMRI recordings were also performed in 3 male healthy control subjects aged 42, 27, and 22 years without any history of nervous system disease.

Patient 2: inspiration Expiration CLINICAL DOCUMENTATION

In the control patient with C7 end-to-end coaptation, chest radiography documented a normal bilateral diaphragm innervation with deep inspiration (Figure 1). Electromyography of the affected biceps muscle demonstrated independence of muscle innervation and breathing (Figure 2B). In patient 1 with phrenic nerve end-to-side coaptation, video recording showed a lack of biceps contractions with deep inspiration or coughing and no change of breathing patterns with elbow flexion. Chest radiography documented bilateral dia- Figure 1. Chest radiographs of participants (the left side of the images are phragm innervation with deep inspiration and no elevated dia- the right side of the participants). The control patient with C7 end-to-end phragm (Figure 1). Electromyography of the affected biceps coaptation shows normal bilateral diaphragm innervation with inspiration. Patient 1 with phrenic nerve end-to-side coaptation shows preserved bilateral muscle demonstrated independence of muscle innervation and diaphragm innervation. Patient 2 with phrenic nerve end-to-side coaptation breathing. shows preserved innervation but diaphragm elevation on the operated left In patient 2 with phrenic nerve end-to-side coaptation, video side. With inspiration, contraction of the left diaphragm produces a shift of recording and fluoroscopy of the thorax showed a lack of bi- about 2 intercostal spaces. ceps contractions with deep inspiration or coughing and a lack of diaphragm innervation with elbow flexion. Chest radiogra- phragm) representations reorganize in such a pattern that phy documented an elevated but innervated diaphragm on the independent control of breathing and elbow flexion is affected side (Figure 1). Electromyography of the affected bi- possible with the same neuronal population.8 Herein, we ceps muscle demonstrated independence of muscle innerva- present comprehensive functional magnetic resonance tion and breathing (Figure 2A). During forced inspiration and imaging (fMRI) studies of 2 patients with phrenic nerve coughing, spikes of motor activation appeared in very few parts end-to-side coaptation, 1 control patient with C7 end- of the electromyographic recordings. to-end coaptation, and 3 healthy control subjects to test this hypothesis. FUNCTIONAL MRI Investigations included 4 tasks: (1) elbow flexion of the dis- METHODS eased arm; (2) elbow flexion of the healthy arm; (3) forced abdominal inspiration; and (4) foot flexion on the side of the dis- PATIENTS eased arm. Patient 1 with phrenic nerve end-to-side coaptation per- The control patient was a right-handed boy aged 6 years at the formed tasks 1 through 4, patient 2 with phrenic nerve end- time of a traumatic left complete brachial plexus lesion. He had to-side coaptation and the healthy control subjects performed end-to-end coaptation of (1) contralateral root C7 to (2) the tasks 1 and 2, and the control patient with C7 end-to-end co- musculocutaneous nerve 5 months after trauma. Left elbow flex- aptation performed tasks 1 through 3. Repetitive investiga- ion was possible against light resistance at the time of fMRI (6.5 tions were performed with 3-T MRI and 7-T MRI (blood oxy- years after surgery). gen level–dependent gradient echo–echo planar imaging; 34 Patient 1 was a right-handed woman aged 29 years at the slices; 128ϫ128 matrix; 230ϫ230ϫ3-mm field of view; gen- time of a traumatic right complete brachial plexus lesion. She eralized autocalibrating partially parallel acquisition factor 2; had end-to-side coaptation of (1) the ipsilateral phrenic nerve and repetition time 2500 milliseconds; for 3-T MRI: echo time to (2) the musculocutaneous nerve 5 months after trauma. Right 35 milliseconds, bandwidth 2220 Hz; for 7-T MRI: echo time elbow flexion was possible against medium resistance at the 22 milliseconds, bandwidth 1396 Hz). Between 5 and 10 iden- time of fMRI (2.5 years after surgery). tical runs (blocked design; 4 rest and 3 activation phases; 20

ARCH NEUROL / VOL 68 (NO. 11), NOV 2011 WWW.ARCHNEUROL.COM 1468

©2011 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021 A Patient 2B Control patient Forced respiration Forced respiration EMG: 100 µV EMG: 100 µV right biceps 1 s right biceps 1 s EMG: 100 µV EMG: 100 µV left biceps 1 s left biceps 1 s 10 µV 10 µV 1 s Respiration 1 s Respiration

Resting respiration Resting respiration EMG: 100 µV EMG: 100 µV right biceps 1 s right biceps 1 s

EMG: 100 µV EMG: 100 µV left biceps 1 s left biceps 1 s 10 µV 10 µV Respiration 1 s Respiration 1 s

Forced innervation of left biceps (diseased) Forced innervation of left biceps (diseased) 100 µV 100 µV EMG: EMG: right biceps 1 s right biceps 1 s EMG: 100 µV EMG: 100 µV left left biceps 1 s 1 s biceps 10 µV 10 µV Respiration 1 s Respiration 1 s

Forced innervation of right biceps Forced innervation of right biceps EMG: 100 µV EMG: 100 µV right right biceps 1 s biceps 1 s EMG: 100 µV EMG: 100 µV left biceps 1 s left biceps 1 s 10 µV 10 µV Respiration 1 s Respiration 1 s

Figure 2. Electromyographic (EMG) recordings of the biceps muscles of patient 2 with phrenic nerve end-to-side coaptation (A) and the control patient with C7 end-to-end coaptation (B) with simultaneous monitoring of respiration. Both patients demonstrate absence of muscle involvement during respiration and no correlation of EMG activity with breathing during forced biceps innervation. (Cyclic EMG activations correspond to electrocardiographic activity.)

seconds/phase) were performed per task per patient. At least 2 coaptation activated the diaphragm areas with flexion of different MRI investigations were performed per patient on dif- the diseased arm. Flexion of the healthy arm and foot did ferent days. The healthy control subjects performed only 1 fMRI not activate diaphragm areas. Figure 3 shows the signifi- investigation at 3 T. Individual data analysis was performed with cant activation of diaphragm areas with diseased arm flex- SPM8 software (Wellcome Trust Centre for Neuroimaging, Lon- ion in comparison with healthy arm flexion. In addi- don, England) (general linear model; uncorrected PϽ.001; technique adapted for pathological brains9). tion, the primary arm areas were also activated (lateral from diaphragm areas). All patient findings could be rep- licated intraindividually and interindividually when re- RESULTS peating experiments on different days and at different magnetic field strengths (3 T, 7 T). Elbow flexion in the healthy Clinical documentation showed that both patients with control subjects only showed arm areas active—no ac- phrenic nerve end-to-side coaptation were able to con- tivity was found in diaphragm areas even when lower- trol the diaphragm and the biceps independently via the ing the threshold. same phrenic nerve: breathing did not change with arm movements (documented by video recording), the affected biceps muscle was not activated with normal breath- COMMENT ing (no regular biceps electromyographic activity), and both sides of the diaphragm were innervated during As demonstrated by the clinical investigations, the 2 pa- breathing and did not move with elbow flexion (chest tients with phrenic nerve end-to-side coaptation were able radiography, fluoroscopy of the thorax). to move their diseased arm independently from their dia- The fMRI studies showed bilateral superior activa- phragm via the same phrenic nerve. With respect to these tions of the primary motor cortex with forced inspira- tasks, they were clinically not distinguishable from the tion (Figure 3). In addition, lateralized midline activa- control patient with C7 end-to-end coaptation. How- tions were found in deeper slices.8 These activations ever, distinction was possible with the fMRI data, which correspond to earlier descriptions of diaphragm repre- showed activation of diaphragm areas with breathing as sentations and normal breathing networks.10,11 In con- well as diseased arm flexion in the patients with phrenic trast to the control patient with C7 end-to-end coapta- nerve end-to-side coaptation. As expected, there was no tion, both patients with phrenic nerve end-to-side diaphragm area activation in the healthy control sub-

ARCH NEUROL / VOL 68 (NO. 11), NOV 2011 WWW.ARCHNEUROL.COM 1469

©2011 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021 Author Contributions: Dr Beisteiner takes responsibil- Healthy control subjects Patients ity for the integrity of the data and the accuracy of the data analysis. Study concept and design: Beisteiner, Millesi, Right arm Right arm Forced vs left arm vs left arm inspiration Grisold, Auff, and Schmidhammer. Acquisition of data: RightLeft Right Left Beisteiner, Ho¨llinger, Rath, Wurnig, Hilbert, Klinger, Geißler, Wo¨ber, Klo¨sch, and Schmidhammer. Analysis Control Control 1 and interpretation of data: Beisteiner, Ho¨llinger, Fischmeis- patient ter, and Schmidhammer. Drafting of the manuscript: Beisteiner, Ho¨llinger, Rath, Wurnig, Hilbert, Klinger, Superior representation of diaphragm Geißler, Fischmeister, Wo¨ber, Klo¨sch, Millesi, Auff, and Schmidhammer. Critical revision of the manuscript for im- Patient 1, portant intellectual content: Beisteiner, Grisold, and Control 2 right arm injured Schmidhammer. Statistical analysis: Beisteiner and Fischmeister. Obtained funding: Beisteiner, Auff, and Schmidhammer. Administrative, technical, and material t value support: Beisteiner, Ho¨llinger, Rath, Wurnig, Hilbert, Maximum Klinger, Geißler, Fischmeister, Wo¨ber, Klo¨sch, Grisold, Patient 2, Control 3 left arm Auff, and Schmidhammer. Study supervision: Beisteiner, injured Millesi, Grisold, and Schmidhammer. Minimum Financial Disclosure: None reported. Funding/Support: This study was supported by grants Figure 3. Original functional magnetic resonance images showing brain P18057 and P23611 from the Austrian Science Fund and activations in the diaphragm area (circles) in the superior primary motor by the Ludwig Boltzmann Institute for Experimental and cortex and in the adjacent arm areas. The right hemisphere is shown on the Clinical Traumatology, Vienna, Austria. left side. Only primary motor cortex activations are depicted, and colors indicate t value distributions within depicted activation clusters (relative scaling according to local maxima and minima). All images are thresholded at PϽ.001 uncorrected. Representative slices are shown for the 3 healthy control subjects, the control patient with C7 end-to-end coaptation, and the 2 patients with phrenic nerve end-to-side coaptation. The columns labeled REFERENCES “right arm vs left arm” show significant activation differences of rightϾleft arm (red) and leftϾright arm (blue). In the patients with phrenic nerve 1. Anastakis DJ, Malessy MJ, Chen R, Davis KD, Mikulis D. Cortical plasticity fol- end-to-side coaptation, the diaphragm area is activated only with flexion of lowing nerve transfer in the upper extremity. Hand Clin. 2008;24(4):425-444, the injured arm, not with flexion of the healthy arm. In all controls, arm vi-vii. flexions did not activate the diaphragm area. Forced abdominal inspiration 2. Gu YD, Wu MM, Zhen YL, et al. Phrenic nerve transfer for treatment of root avul- selectively activated the diaphragm area in all subjects (rightmost column, sion of the brachial plexus. Chin Med J (Engl). 1990;103(4):267-270. blue activation clusters). 3. Chuang DC. Neurotization procedures for brachial plexus injuries. Hand Clin. 1995; 11(4):633-645. jects performing identical elbow flexions. The fMRI re- 4. Viterbo F, Trindade JC, Hoshino K, Mazzoni Neto A. Latero-terminal neurorrha- sults indicate that the human brain can adapt to periph- phy without removal of the epineural sheath: experimental study in rats. Rev Paul eral nervous system changes with a hitherto unknown Med. 1992;110(6):267-275. 5. Schmidhammer R, No´gra´di A, Szabo´ A, et al. Synergistic motor nerve fiber trans- type of cortical neuroplasticity: the diaphragm areas re- fer between different nerves through the use of end-to-side coaptation. Exp Neurol. organize such that independent control of breathing and 2009;217(2):388-394. elbow flexion is now possible with the same neuronal 6. Hsieh JC, Cheng H, Hsieh HM, et al. Loss of interhemispheric inhibition on the population. ipsilateral primary sensorimotor cortex in patients with brachial plexus injury: fMRI study. Ann Neurol. 2002;51(3):381-385. We conclude that specific cortical neuroplasticity pro- 7. Malessy MJ, Bakker D, Dekker AJ, Van Duk JG, Thomeer RT. Functional mag- vides the neurophysiological basis for rehabilitation af- netic resonance imaging and control over the biceps muscle after intercostal- ter peripheral end-to-side repair. Data from fMRI indi- musculocutaneous nerve transfer. J Neurosurg. 2003;98(2):261-268. cate effectiveness of this surgical procedure. Neurological 8. Beisteiner R, Ho¨llinger I, Schmidhammer R. Neurological picture: FMRI evi- practice should consider this option for therapeutic han- dence for a new therapeutic option for deafferentiated muscles. J Neurol Neu- rosurg Psychiatry. 2010;81(11):1209-1210. dling of complete plexus lesions. 9. Beisteiner R, Klinger N, Ho¨llinger I, et al. How much are clinical fMRI reports influenced by standard postprocessing methods? an investigation of normaliza- Accepted for Publication: March 25, 2011. tion and region of interest effects in the medial temporal lobe. Hum Brain Mapp. Correspondence: Roland Beisteiner, MD, MA, Depart- 2010;31(12):1951-1966. 10. McKay LC, Evans KC, Frackowiak RS, Corfield DR. Neural correlates of volun- ment of Neurology, Medical University of Vienna, tary breathing in humans. J Appl Physiol. 2003;95(3):1170-1178. Wa¨hringer Gu¨ rtel 18-20, A-1090 Vienna, Austria (roland 11. Foerster O. Motorische Felder und Bahnen. Berlin, Germany: Springer Verlag; [email protected]). 1936.

ARCH NEUROL / VOL 68 (NO. 11), NOV 2011 WWW.ARCHNEUROL.COM 1470