ABSTRACT: The radical neck dissection is an operation for the managementof lymph node metastases from primary sites involving the oral cavity, larynx,and other areas of the head and neck. In this procedure, the spinal accessorynerve is removed along with other structures. In modified neck dissectionthe spinal accessory nerve is preserved. Patients undergoing the modifiedneck dissection have had variable functional outcomes from little or no painor disability, to significant muscle dysfunction. Our group hypothesized thatpatients with good functional outcomes following modified neck dissectionmay have had motor contributions from C2, C3, or C4 branches, while thosewith less favorable outcomes did not. To demonstrate the presence of motorinput and its significance both from the spinal accessory nerve and thebranches of the cervical plexus, we utilized intraoperative electroneurogra-phy. We find that although there is motor contribution from C2, C3, and C4to the trapezius muscle, it was not consistent or significant. 1997 JohnWiley & Sons, Inc. Muscle Nerve 20, 279–285, 1997.Key words: spinal accessory nerve; trapezius muscle; intraoperative electro-neurography; modified neck dissection

UTILIZATION OF INTRAOPERATIVEELECTRONEUROGRAPHY TOUNDERSTAND THE INNERVATION OFTHE TRAPEZIUS MUSCLE

SUBHADRA NORI, MD,1*, KEE CHEE SOO, MD,2 RONALD F. GREEN, MD,3

ELLIOT W. STRONG, MD,4 and SAUL MIODOWNIK, MEE, CCE5

1 Department of Rehabilitation Medicine, Montefiore Medical Center, 111 East210th Street, Bronx, New York 10467-2490, USA2 Department of Surgery, Head and Neck Service, Singapore General Hospital,Singapore3 Department of Rehabilitation Medicine, The New York Hospital–Cornell MedicalCenter, New York, New York, USA4 Department of Surgery, Head and Neck Service, Memorial Sloan-Kettering CancerCenter, New York, New York, USA5 Department of Bio-Engineering, Memorial Sloan-Kettering Cancer Center, NewYork, New York, USA

Received 11 December 1995; accepted 11 September 1996

The radical neck dissection is a classical operation volving pain, weakness, and deformity of the shoulderfor the management of suspected lymph node metas- girdle mechanism.4

tases from primary sites involving the oral cavity, lar- To avoid these complications, some surgeonsynx, and other areas of the head and neck. This have advocated grafting between the cut ends of theoperation involves the removal of the spinal accessory spinal accessory nerve.1 Others have performed vari-nerve along with the jugular vein and sternocleido- ous types of modified neck dissection in which bothmastoid muscle. the sternocleidomastoid and the spinal accessory

Resection of the spinal accessory nerve and the nerve are spared. The outcomes of the nerve sparingconsequent denervation of the trapezius muscle can procedures are not consistent.14,17,23

result in a well-recognized ‘‘shoulder syndrome’’ in- Our group was interested in trying to understandthe basis for these variations in outcome. Based onSoo’s27 work with cadavers, in which he observed thepresence of contributions from the C2, C3, and C4*Correspondence to: Dr. Subhadra Noribranches of the cervical plexus directly to the spinal

CCC 0148-639X/97/030279-07accessory nerve, we hypothesized that there might 1997 John Wiley & Sons, Inc.

Innervation of Trapezius Muscle MUSCLE & NERVE March 1997 279

be a significant motor contribution to the trapeziusmuscle from these branches.

There is some evidence that the spinal accessorynerve receives one or two contributions from C2 orC3 deep to the sternocleidomastoid muscle.27

Whether any of these give significant motor con-tribution to the trapezius muscle is as yet unclear.

Our study sought to demonstrate conclusively thepresence of motor input both from the spinal acces-sory nerve and the C2, C3, and C4 branches of thecervical plexus. We sought further to determinewhether the contributions from C2, C3, and C4 aresignificant and consistent enough to reassure sur-geons that, when doing a modified neck dissection,the shoulder will function without disability.

MATERIALS AND METHODS

To demonstrate conclusively the presence of motorinput from the spinal accessory nerve and the C2, C3,and C4 branches of the cervical plexus, we performeddirect stimulation of each of these nerves intraopera-tively in patients undergoing modified radical neckdissection and supraomohyoid neck dissections, oneof the variations of the modified neck dissection. Inthese procedures, the spinal accessory nerve and allthe branches of the cervical plexus are exposed andpreserved, thus affording the opportunity for thestudy and objective documentation of the innerva-

FIGURE 1. Dissection of the posterior triangle of the neck showingtion of the trapezius muscle in the intact living subjectthe anatomy during surgery. Arrows point to: A, spinal accessory

under direct vision (Fig. 1). nerve; B, C2; C, C3; D, C4; E, trapezius; F, phrenic nerve.The intraoperative stimulation technique has

been utilized in the past for the median nerve, thebrachial plexus,29,30 sciatic nerve,22 ulnar nerve,3 andother peripheral nerves, but not for the study of the Test Clip #4724, Pomona, CA), which was not only

effective but also more economical than existingcervical plexus.Twelve patients—10 males and 2 females—were commercial products (Fig. 2). With this tool, we were

able to make reliable and repeatable monopolar con-studied on completion of the operative procedure.The median age was 50 years, with a range of 25–78 nections to the nerve to be stimulated without fear

of damage or trauma due to excessive compression.years. In 10 patients, the operation performed was thesupraomohyoid neck dissection; in 2, the modified We could also apply the stimulation in isolation

along a length of dissected nerve with little or noradical neck dissection. In each of these patients,the entire length of the spinal accessory nerve was energy being transmitted to adjacent conductive

fluid or structures. These probes were gas-sterilizedpreserved as well as the contributions from the uppercervical plexus that join the sternocleidomastoid, and prior to each use.

An electrode switch box was fabricated (Fig. 3),some cervical plexus branches running to the trape-zius independently. making it possible to select a pair of EMG recording

electrodes with a turn of a knob. Disposable electro-Electromyographic (EMG) recordings were per-formed using a TECA 5 TD5 electromyograph with cardiographic (EKG) electrodes (Medtronic/An-

dover Medical #1600-001, Haverhill, MA) and con-storage scope. Latency was measured to the onset ofnegative deflections of the motor response. Stimulus ventional monitoring lead wires were used at the sites

being studied. The switch box directed the outputof 0.1 ms duration was delivered at 1/s to the exposednerves, medially, close to the carotid sheath. of a pair of these electrodes to the EMG preamplifier.

Three channels were used for the upper, middle,Special electrode hooks were fashioned from aset of grabber probes (Pomona ITT Maxigrabber and lower portions of the trapezius, respectively. In

280 Innervation of Trapezius Muscle MUSCLE & NERVE March 1997

FIGURE 3. Electrode switch box with three channels, selectorFIGURE 2. Hook electrode, in comparison to the regular stimulat-knob, and connected grabber probes.ing electrodes.

in the upper trapezius, 11 out of 12 in the middle,this manner, three sets of traces could be recorded and 10 out of 12 in the lower trapezius (Table 1).in rapid succession. The latencies ranged 2.2–5.6 ms (mean 3.8, median

The nerve conduction study was performed ac- 3.6) to the upper trapezius, 2.6–6.0 ms (mean 5.9,cording to the method outlined by Green and Brian.10

For the upper trapezius, the active recording elec-trode was placed 5 cm lateral to the spinous processof the seventh cervical vertebra. For the middle trape-zius, it was placed halfway between the midpoint ofthe ipsilateral scapular spine and the spinous processof the thoracic vertebra at the same level. For thelower trapezius, it was placed two finger’s breadthfrom the spinal column on a line perpendicular tothe spinal column at the level of the inferior angleof the ipsilateral scapula6 (Fig. 4).

The nerves were stimulated under direct visualiza-tion by the operating surgeon. Stimulating voltagewas 50 V. The low filter setting was 8 Hz and the highfilter setting was 8 kHz. Sweep speed was 2 ms perdivision, and gain was 500 mV per cm.

RESULTS

When the spinal accessory nerve was stimulated, FIGURE 4. Placement of active electrodes (x) for determination ofaccessory nerve latencies to upper, middle, and lower trapezius.evoked responses were obtained from all 12 patients

Innervation of Trapezius Muscle MUSCLE & NERVE March 1997 281

Table 3. Evoked responses of C3 nerve.Table 1. Latency of evoked response of the spinalaccessory nerve.

Upper Middle Lowertrapezius trapezius trapeziusUpper Middle Lower

trapezius trapezius trapeziusNo. of patients

studied 12 12 12No. of patientsstudied 13 13 13 No. of responses

elicited 7 7 6No. of responseselicited 13 12 12 Average latencies 3.7 ms 4.2 ms 6.0 ms

Average amplitudes 1.35 mV 1.2 mV 0.6 mVAverage latencies 3.8 ms 5.9 ms 6.0 msAverage amplitudes 3.2 mV 2.60 mV 1.81 mV

median 4.4) to the middle, and 2.6–8.6 ms (mean 6.0, The configuration of the evoked responses corre-median 5.9) to the lower trapezius. The amplitudes sponded to the previous studies of Petrera and Troja-ranged 2–4 mV to the upper trapezius, 2–4.2 mV to borg.21 The latencies obtained showed a slightly widerthe middle trapezius, and 1.5–4.0 mV to the lower tra- range—from 2.2 to 6.3 ms—and were slightly pro-pezius. longed compared with values obtained in normal

C2 contributions were seen in 5 out of the 12 controls.10 This discrepancy can be explained by thepatients, supplying all three parts of the muscle in 2 varying degree of cooling caused by devascularization(Table 2). C2 latencies ranged 2.5–3.6 ms (mean and traction injury to these nerves during surgery. No3.1, median 3.2) to the upper trapezius, 3.4–3.6 ms immediate or delayed postoperative complications(mean 3.5, median 3.5) to the middle, and 3.6–5.2 were seen. All patients tolerated the procedure well.ms (mean 4.4, median 4.4) to the lower trapezius.The amplitudes ranged 1.75–3 mV to the upper tra-

DISCUSSIONpezius, 2–2.5 mV to the middle trapezius, and 2–3Loss of shoulder function due to resection of themV to the lower portion of the trapezius muscle.spinal accessory nerve and denervation of the trape-Evoked responses from C3 were seen in 9 out ofzius muscle is the most disturbing sequela of the12 patients, supplying all three parts of the muscletraditional neck dissection. The trapezius muscle isin 5 (Table 3). The latencies ranged 2.8–4.8 msresponsible for primary rotatory and supportive ac-(mean 3.7, median 3.6) to the upper trapezius, 3.0–tions of the shoulder girdle mechanism.12,28 Paralysis5.2 ms (mean 4.2, median 4.4) to the middle, andof the trapezius muscle results in a well-recognized3.6–10.4 ms (mean 6.0, median 5.5) to the lower‘‘shoulder syndrome,’’ in which the patient experi-trapezius. The amplitudes ranged 1.75–3.0 mV to theences pain, weakness, and deformity of the shoulderupper trapezius, 2–3.0 mV to the middle, and 1–1.5girdle mechanism.20mV to the lower trapezius.

The trapezius is a large muscle that anatomicallyC4 contributions were seen in 4 out of 12 patients,consists of three distinct parts—upper, middle,supplying all three parts of the muscle in 2 (Tablelower—each serving complementary and yet differ-4). The range was 2.6–5.2 ms (mean 3.55, medianent roles. The simultaneous actions of the upper and3.2) to the upper trapezius, 2.6–8.8 ms (mean 6.7,lower parts of the trapezius cooperate to exert a rota-median 6.8) to the middle, and 2.8–10.8 ms (meantory action of the scapula.19,31 This upward rotation7.53, median 9.0) to the lower trapezius. The ampli-of the scapula accompanies abduction of the arm attudes ranged 1–1.5 mV to the upper trapezius, 1–1.25the glenohumeral joint and produces elevation ofmV to the middle trapezius, and 1.5–2.0 mV to thethe arm above shoulder level.12,13,28,31lower part of the trapezius muscle.

Table 4. Evoked responses of C4 nerve.Table 2. Evoked responses of C2 nerve.

Upper Middle LowerUpper Middle Lowertrapezius trapezius trapezius trapezius trapezius trapezius

No. of patientsNo. of patientsstudied 12 12 12 studied 12 12 12

No. of responsesNo. of responseselicited 5 2 2 elicited 4 3 3

Average latencies 3.55 ms 6.7 ms 7.53 msAverage latencies 3.1 ms 3.5 ms 4.4 msAverage amplitudes 0.8 mV 0.3 mV 0.4 mV Average amplitudes 0.4 mV 0.2 mV 0.4 mV

282 Innervation of Trapezius Muscle MUSCLE & NERVE March 1997

Paralysis of the spinal accessory nerve will result middle, and lower parts of the trapezius; however,they state that ‘‘it cannot be decided from the presentin downward displacement of the scapula, with a

downward and forward drop of the shoulder.7,18,20,24,25 study whether in all cases one is actually stimulatingthe accessory nerve or one of the cervical nerves, inA dull aching discomfort is experienced along the

upper margin of the scapula, frequently described cases where the former ends in the sternocleidomas-toid muscle.’’21as resembling a toothache. This pain is due to the

stretching of scapular retractors (rhomboids) and The anatomy of the spinal accessory nerve andthe elevator (levator scapulae), initiated by the the cervical plexus contributions to the trapeziusunbalanced pull of the serratus anterior. In addi- have been studied and described from the cadavertion, arm abduction is limited to 908, causing diffi- dissections performed by Soo.11,27 In these dissections,culty with even simple tasks such as combing the accessory nerve received one or two contributionsone’s hair. from C2 or C3 deep to the sternocleidomastoid mus-

To avoid this disabling shoulder syndrome, vari- cle. In addition, large branches from C3 and C4 wereous modified radical neck dissections have evolved seen running into the trapezius independent of theattempting to salvage the spinal accessory nerve, with accessory nerve (Fig. 5). Whether any of these gavevarying results. Although these nerve-sparing proce- significant motor contributions to the muscle hasdures have been successful in some cases, their out- been unclear.comes have not been consistent;14,17,23 varying degrees The objective of our study was (1) to demonstrateof shoulder dysfunction have been reported.4,9,16 The conclusively the presence of motor input to the trape-literature further shows that the patient’s age or zius from the C2, C3, and C4 nerves; (2) to determinehealth do not seem predictive of ultimate symptom with what consistency; (3) to confirm motor inputseverity or disability; patients in the same age groupor with the same health status have had varying out-comes.

Short and associates found that patients under-going nerve-sparing procedures have less disabilitythan those who undergo radical neck dissections.26

However, Nahum and Brodal have reported varyingdysfunction despite anatomic preservation of thespinal accessory nerve.2,20 Could this variability bedue to additional innervation from the C3 or C4level? There is much debate in the literature onthe efferent nerve supply to the trapezius muscle.Brodal has stated that the upper and middle por-tions of the muscle are innervated by the spinalaccessory nerve and the lower portion by the thirdand fourth cervical rami, C3 and C4.2 Others believethat the contributions from C3 and C4 are purelyproprioceptive.1 Although various anatomical stud-ies in animals and human embryos have been donein an attempt to resolve this question, there hasbeen no evidence to date to suggest that the C3 andC4 branches to the trapezius have motor input.5,17,26

Gluckman drew no definite conclusions regardingthe existence of this innervation.8 Anderson andFlowers believe that, on the basis of electromyo-graphic and nerve conduction studies, the contribu-tions from C3 and C4 are purely proprioceptive.1

There has been no electrophysiological evidenceto date to suggest that the C3 and C4 branchesthat run to the trapezius have motor input.5,19,28

FIGURE 5. Cadaver dissection, showing the accessory nerve andPetrera and Trojaborg stimulated the spinal ac- cervical contributions. Arrows point to: A, XI nerve; B, sternocleido-cessory nerve transcutaneously in the posterior trian- mastoid muscle; C, C2 branches; D, trapezius; E, C3 branch; and

F, phrenic nerve.gle and obtained evoked responses from the upper,

Innervation of Trapezius Muscle MUSCLE & NERVE March 1997 283

from the spinal accessory nerve to the trapezius mus-REFERENCEScle; and (4) to describe a safe and reproducible intra-

1. Anderson R, Flowers RS: Free grafts of the spinal accessoryoperative electroneurography technique to stimulateduring radical neck dissection. Am J Surg 1969;118:796.the cervical plexus.

2. Brodal A: Neurological Anatomy, vol 457, Oxford, Oxford Uni-We utilized the technique of intraoperative elec- versity Press, 1981.

troneurography,15 in which nerve function is mea- 3. Campbell MD, Pridgeon, Sahani KS: Entrapment neuropathyof the ulnar nerve at its point of exit from the flexor carpisured by direct stimulation of the exposed nerve dur-ulnaris muscle. Muscle Nerve 1986;9:662.ing surgery, a technique that had not previously been

4. Carenfelt C, Elaisson K: Occurrence, duration and prognosisused on the spinal accessory nerve and/or cervical of unexpected accessory nerve paresis in radical neck dissec-

tion. Acta Otolaryngol 1980;90:470.plexus. Evoked potentials were measured to each part5. Cihak R: Muscularis trapezius and the changes of its formationof the trapezius muscle on stimulation of the spinal

in human ontogenesis. Acta Univ, Carol [Med] (Praha)accessory, C2, C3, and C4 nerves.1974;20:45.

Latencies were measured to the onset of negative 6. Delagi EF, Perotto A, Iazetti J, Morrison D: Anatomic Guide forElectromyographers, 2nd ed. Springfield, IL, Charles C.deflections of the motor response: Technical modi-Thomas, 1975.fications such as the use of fine probes for stimulation

7. Ewing MR, Martin H: Disability following radical neck dissec-and a three-channel switch box allowed rapid and tion. Cancer 1952;5:873.accurate identification of responses. 8. Gluckman JL, Myer CM, Asseff JN: Rehabilitation following

radical neck dissection. Laryngoscope 1983;93:1083–1085.The latencies obtained showed a slightly wider9. Gordon SL, Graham WP, Black JT, Miller SH: Accessory nerverange and were slightly prolonged compared with

function after surgical procedures in the posterior triangle.normal controls.10 This can be explained by varying Arch Surg 1977;112:262.degrees of neurapraxia during surgery. 10. Green RF, Brian M: Accessory nerve latency to the middle

and lower trapezius. Arch Phys Med Rehabil 1985;66:23.Our study had a significant technical advantage11. Hamlyn PJ, Pegginton J, Soo KC: The cervical plexus and theover that of Petrera and Trojaborg, whose study uti-

motor supply to the trapezius. Abstract, Scientific Program.lized surface stimulation,21 because we stimulated the British Association of Clinical Anatomist’s annual meeting,nerves under direct vision. June, 1986.

12. Hollinshead WH, Rosse C: Pectoral region, axilla and shoul-der, in Textbook of Anatomy, 6th ed. Hagerstown, MD, Harper

Conclusion. Our study confirms our previous find- and Row, 1985, pp 193–194.ing that the spinal accessory nerve provides the most 13. Inman BT, Sanders M, Abott LC: Observations on the function

of the shoulder joint. J Bone Joint Surg 1944;26:1.important and consistent motor input to the trape-14. Jesse, RH, Ballantyne AJ, Larson DL: Radical or modified neckzius muscle.

dissection: a therapeutic dilemma. Am J Surg 1978;136:The study demonstrates the presence of motor 516.

input from the C2, C3, and C4 branches of the cervi- 15. Kline DG, Dejonge BR: Evoked potentials to evaluate periph-eral nerve injuries. Surg Gynecol Obstet 1968;127:1239.cal plexus but finds that they were either not consis-

16. Leipzig G, Suen JY, English J, Barnes J, Hooper M: Functionaltently present or, when present, did not consistentlyevaluation of the spinal accessory nerve after neck dissection.

innervate all three parts of the trapezius. Our study Am J Surg 1983;146:526.does not fully support the assumption that there may 17. Lingerman RE, Stephens R, Helmus C, Ulm J: Neck dissection:

radical or conservative. Ann Otolaryngol 1977;86:737.be definitive motor input from the upper cervical18. Mead S: Posterior triangle operations and trapezius paralysis.plexus to the trapezius muscle. Therefore, we cannot

Arch Surg 1952;64:752.conclude that, in cases of modified neck dissection, 19. Meckenzie J: The morphology of the sternocleidomastoid andthere would be preservation of motor function sup- the trapezius muscle. J Anat 1955;89:526.

20. Nahum AM, Mullally W, Marnior L: A syndrome resultingplemented by the cervical plexus. We find the resultsfrom radical neck dissection. Arch Otolaryngol 1961;74:424.sufficiently encouraging, however, to warrant addi-

21. Petrera JE, Trojaborg W: Conduction studies along the acces-tional studies with a larger number of patients. sory nerve and followup of patients with trapezius palsy. JThis study utilized the existing principles of intra- Neurol Neurosurg Psychiatry 1984;47:630.

22. Phillips LH, Park TS: Electrophysiological studies of selectiveoperative electroneurography to stimulate the spinalposterior rhizotomy patients, in Neurosurgery: State of the Artaccessory nerve and the branches of the cervicalReviews, vol 4. Philadelphia, Hanley & Velfus Inc., 1989, ppplexus. This technique is reliable, safe, and reproduc- 459–469.

ible, and can be utilized in future studies. 23. Remmler D, Byers R, Scheetz J, Shell B, White G, ZimmermanS, Goepfert H: A prospective study of shoulder disability re-sulting from radical and modified neck dissections. Head Neck

The authors would like to thank the following: the Memorial Sloan- Surg 1986;8:280–286.Kettering Cancer Center and the surgeons of its Head and Neck 24. Saunders WH, Johnson EW: Rehabilitation of the shoulderService for providing the subjects for this study; Kathryn K. Wheeler after radical neck dissection. Ann Otol 1975;84:812.

25. Schuller DE, Reiches NA, Hamaker RC, Lingeman RE, Weis-for editorial assistance; and Cathy B. Wong for preparation ofberger EC, Suen JY, et al.: Analysis of total disability resultingthe manuscript.

284 Innervation of Trapezius Muscle MUSCLE & NERVE March 1997

from treatment programs including radical neck dissections, surement of motor action potentials. Head Neck 1993;15:216–221.and modified neck dissection. Head Neck Surg 1983;6:

29. Strauss NL, Brazier HA: The spinal accessory nerve and its551–558.musculature. Q Rev Biol 1936;2:387.26. Short SO, Kaplan JN, Laramore GE, Cummings CW: Shoulder

30. Terzis J, Daniel R, Williams H: Intraoperative assessment ofpain and function after neck dissection with or without preser-nerve lesions with fasicular dissection and electrophysiologicalvation of the spinal accessory nerve. Am J Surg 1984;148:478.recordings, in Omar G, Spinners M, (eds): Management of27. Soo KC, Hamlyn TJ, Guiloff RJ, Oh A: Anatomy of the accessoryPeripheral Nerve Problems. Philadelphia, W.B. Saunders, 1980,nerve and its cervical contributions in the neck. Head Neckpp 462–472.Surg 1986;9:111.

31. Warwick, R (ed): Gray’s Anatomy, 35th ed. Philadelphia, W.B.28. Soo KC, Strong EW, Spiro RH, Shah JP, Nori S, Green R:Saunders, 1971.Innervation of the trapezius muscle by the intra-operative mea-

Innervation of Trapezius Muscle MUSCLE & NERVE March 1997 285