arnal ofIGF-R y605602

606

6O8

617

619

619

620

622

623

624

ORIGINAL

THE JOURNAL OF

HANDSURGERY

Official journalAMERICAN SOCIETY FOR SURGERY OF THE HAND

COMMUNICATIONS

Internal topography of major nerves of the forearmand hand: A current view

Fresh cadaver nerves were examined by serial cross-sections and microdissection with the operatingmicroscope. The findings are compared with those of previous authors, primarily Sydney Sunderland. Ourstudy cot~rms and amplifies Sunderland’s findings: although it is true that funicular plexus formation and#tterchange takes place in the nerves of the human forearm, these connections are not of such a degree asto preclude operative procedures such as intraneural neurolysis, fascicular nerve repair, andinterfascicular nerve grafting, b~dividual branches and bundles can be identified and traced withht thema#t nerve trunk for considerable distances without significant trautna to conducting fibers. Thisarrangement lends itself to the application of modern microneurosurgical techniques. Clinical applicationsof these findings in the repair, lysis, and grafting of the major nerves of the forearm are described. Theflossibilit3’ of using such branches as the dorsal cntaneous branch of the ulnar nerve (if irreparablydamaged) as a donor nerve for grafting is noted.

Michael E. Jabaley, M.D., F.A.C.S., William H. Wallace, M.D., attdFrederick R. Heckler, M.D., Jackson, Miss.

"Honor those who go first, even if those who come later go further."

--A fabian proverb *

The unpredictable quality of functional recovery

which follows nerve repair has stimulated both clinicaland basic researchers to continue their investigations ofthe structure and function of peripheral nerves. In

This paper received the Emanuel Kaplan Award, presented by theNew York Hand Society, for the best paper on an anatomic subjectat the 1979 meeting of the American Society for Surgery of theHand.

From the Department of Surgery, Division of Plastic Surgery, Uni-versity of Mississippi: Medical Center, Jackson, Miss.

Received for pubhcat,~[arch 15, 1979.Reprint requests: Michael E. Jabaley, M.D., 1417 Lefia Dr.,

Jackson, MS 39216 [(601)362-6323].*Quoted by Sir Sydney Sunderla:ad. in his Sterling Bunnell Memorial

Lecture. San Francisco, N,~,. 3, :977.

1945, Sunderland reviewed 40 years of such researchand described his own work with the internal topog-raphy of the radial, median, and ulnar nerves.~-z Thiswell known work is often cited by "other authors toexplain their failure to attain a satisfactory level of

function following nerve repair and subsequent axonregeneration. Sunderland pointed out that the cross-sectional arrangement of nerve components changesfrom level to level and from millimeter to millimeter;histological sections of nerve ends which are examinedat different levels will appear to be different. This fact

has led some to adopt an attitude of resignation towardaccurate alignment of fascicles when gaps of more than

a few millimeters exist,The changing nature of ~he internal arrangement is

2 Jabaley, Wallace, HecklerThe Journal of

HAND SURGERY

Fig. 1. Sunderland’s three-dimensional drawing of thefunicular plexus formation in a 3 cm segment of a specimenof th~ musculocutaneous nerve of the arm. (Reproduced bykind permission of Churchill-Livingstone, Ltd., Edinburgh,London, and New York, and Sir Sydney Sunderland.)

unquestioned. Nevertheless, several clinical observa-tions led us to wonder if other factors known to beinvolved in peripheral nerve function might not out-weigh the handicap of an evolving cross-sectional rela-tionship in determining the quality of a final result. Forexample, bundles and funiculi of intact nerves can beseparated (lysed) with no loss of function and some-times even with clinical improvement. Similarly,in ulnar nerve transposition operations, the motorbranches can ’.~ safe],,.,’ separated from the m’:.:! ~ ,:~nk

for some dist~ance. Relatively long nerve gaps cansometimes be overcome with successful restoration ofboth motor and sensory function either by direct sutureor interpositional grafts, suggesting that some measureof effective funicular alignment has been accom-plished. Lastly, and most spectacular/y, Manktelowand McKee4 have reported transfer of the gracilis mus-cle to the forearm by microneurovascutar techniqueswith subsequent recovery of function. Clearly, nervesmay have a greater potential for functional recoverythan is generally assumed provided circumstances areotherwise favorable.5 :

Definition of terms

A peripheral nerve consists of (I) neural elementswhich are extensions of motor and sensory cells locatedin the dorsal ganglia and anterior horns of the spinalcord and (2) non-neural connective tissue which sup-ports, separates, nourishes, protects, and maintains therelative position of the various neural elements. Theneural elements, called axons (or fibers), may be affer-ent, efferent, or sympathetic (sympathetic fibers will beignored in this discussion). The non-neural elementsconsist of the epineurium, perineurium, and en-doneurium. The epineurium can be subdivided into theepineural sheath (which surrounds the whole nerve) andthe epineural connective tissue (which lies betweenfuniculi or bundles). Groups of axons enclosed byperineurium are calledfuniculi orfasciculi. They arethe smallest clinical units with which surgeons mightwork using current clinical techniques. Sunderlandchose to use the termfuniculus; we have elected to usethis term and its synonym,fascicle (orfasciculus), in-terchangably, since both are found frequently in theclinical literature. The term bundle is purely descriptiveand will be used to describe a group of fasciculi whichappear to be associated topographically within theepineurium, either in photomicrographs or upon clinicalor laboratory dissection. It is a clinical term and doesnot necessarily imply a functional association.

Sir Sydney Sunderland

Sunderland’s study of the intraneural topography ofthe radial, median, and ulnar nerves clarified the diver-gent opinions expressed in the literature prior to 1945.3

He later extended this investigation to other nerves5 andassembled all his studies in the monumental publica-tion,Nerves and Nerve Injuries. 6 Sunderland found "noconstant or characteristic funicular pattern for anynerve," and he stated that "the funicular pattern wascontinually modified along the entire len_~th of each

The Journal ofID SURGERY

ve gaps can:estoration ofdirect suture

ome measure~een accom-

Manktelowgracilis mus-tr techniques .early, nerveshal recoveryinstances are

ral elementscells located9f the spinal’ which sup-naintains the-~ments. Theaay be affer-~bers will beral elements’~, and en-ided into thee nerve) andies between.’nclosed by~i. They aregeons mightSunderland

ected to use:iculus), in-ently in the~ descriptive:iculi whichwithin the

lpon clinicalm and does/on.

~ography ofd the diver-~r to 1945. z

nerves~ andtal publica-~ found "nom for anypattern was,~th of each

Vol. 5, No. IJanuary 1980 Internal topography of peripheral nerves 3

Fig. 2. The left median nerve in upper forearm, looking from proximal to distal. The radial side ison left. A, Discrete fascicles can be seen enclosed in perineurium and separated by epineurium. B,A line drawing of a nerve..(pt, pronator teres; t, terminal motor and sensory fascicles to the hand;fds, flexor digitorum superficialis; fcr, flexor carpi radialis; p, flexor digitorum profundus: ai,anterior interosseous.) Motor branches to forearm muscles are situated about the radial and ulnarsides; terminal branches to the hand are located in central core.

nerve by the repeated division, anastomosis, and mi-gration of the bundles." The longest section of anynerve with a constant pattern was 15 ram, althoughindividual bundles and bundle groups pursued longercourses without change; the average length of constantpattern was only 0.25 to 5.0 ram. Although he dis-cussed resection, repair, and regeneration in suturednerves, Sunderland made no attempt to correlate his

findings with function or clinical results. Nevertheless,his famous three-dimensional reconstruction of themusculocutaneous nerve (Fig. 1) is frequently associ-ated with explanations of suboptimal results followingnerve repair,r We believe that the hand sur~_’een’s in-

terpretation of Sundertand’s work is incomplete andunnecessarily pessimistic.

At the time of Sunderland’s study, standard nerverepair was performed by epineurial suture withoutmagnification. Intraneural topography was of someacademic interest, but had little clinical importance.Since the introduction of magnifying devices, sophisti-cated instruments, and microsurgical suture, more andmore surgeons have undertaken interfascicular dissec-tion for the purpose of neurolysis, interfascicular graft-ing, and repair. Because of the increasingly frequentuse of these techniques, an accurate understanding ofintrane:~ral architecture has become mandatory. To bet-

4 Jabaley, Wallace, HecklerThe Journal

HAND SURGERy

Ttfenar m.

Th~~ 2 enar sensory~N NERVE Radial index

nd web space

~ Flexor dig. superficialis

-""- Ant. interosseous%

Pronator teres

Fig. 3. Component micmdissection of the median nerve. The dark lines diagrammatically illustratethe dissection distance for each branch. The ring and arrow depict the anatomic origin of the nervebranch. The dark line distal to the ring and arrow represents the extraneural dissection distance; theline proximal represents the intraneural dissection distance.

ter understand the limits and the usefulness of thesetechnical advances, we have undertaken to perform se-rial nerve sections in the fashion described by Sunder-land and to compare these observations with thefindings of interfascicular microdissections of the majornerves of the human forearm.

Method

We examined the median~ ulnar, and radial nerves ofsix unfixed human arms by either serial section (threemedian nerves--one sectioned totally, two partially) ormicroscopic interfascicular dissection (four median,four ulnar, and two radial nerves). Serial sections weretaken between the medial epicondyle of the humerusand the terminal nerve branches in the palm; the mi-crodissections were extended 10 to 12 cm above themedial epicondyle.

Serial section study. Three median nerves werestudied, all from adult males. The median nerve and itsbranches were exposed throughout the forearm andhand and left in place. All branches were identified andlabeled. The volar (anterior) surface of the nerve wasmarked with a continuous No. 6-0 silk suture placedwithin its epineurium and India ink applied to its sur-face. The length of the nerve and the distance from theradial styloid reference point to each branch wa:~sured in n-liili.r-:~er.ers arid recorded. The icca~i~:branches on the perimeter of the main trer~k

noted. The entire median nerve and its branches wereremoved from the arm, pinned to a board to maintaincorrect length and orientation, and fixed in 10% forma~lin for 7 days.

After adequate tissue fixation the median nerve wascut into 44 separate blocks, either 5 or 10 mm in length.The proximal end of each block was labeled and thedistal 2 mm was removed as a separate subblock, pro-cessed, and embedded for microtome slicig’~. The 2mm block was serially cut into 20-micron-thick sec-tions, which were mounted unstained for examination.(Initially these sections were stained with hematoxylinand eosin, a step later omitted after we learned that wecould see adequate detail on unstained preparations.)Each microscopic section was projected and traced onthin paper to document the relative size and location offuniculi, the amount of epineurium present, and theorigin of branches. Branches which could be positivelyidentified were followed in retrograde fashion and theirfunicular makeup recorded. Sections were made im-mediately proximal and distal to all branches to confirmtheir identity. Additional sections were taken through-out each block when necessary to permit positive iden-tification of bundles and clarification of their changingpattern. Like Sunderland, we examined the nerves in adi~tat-~o-proximal directi:;~ :o detect change~ ;:~ num-ber, size, and location of fasciculi. Photograp::; of ap-propriate section:~ were ~ade for permanent

; Journal ofSURGERy

5, No. 11980 Internal topography of peripheral nerves 5

~ble I. Median nerve serial sections

maintainb forma-

!rve was

I length.andtheok, pro-

The 2ick sec-ination.~toxylinthat weations.)iced onation of~nd thesitivelyad their:de im-:on firmtrough-e iden-anginges in aI num-

of ap-

,rding.

Thenar motor branch.Sensory

1st web space2nd web space

~S): 3rd web space~2.~ Palmar cutaneous

Distali ~!.i_ Proximal

Sunderland

Firslinterconnection

Totalmeasurable distance

Authors

First

I

Totalfirst interconnection measurable distance

56 112 35 70

5O 95 2O 5558 97 25 5566 81 25 55

74 99 30 7037 62 25 65

Fle~er digitorum::.~icialis

Distal 2Proximal 16

Anterior interosseous 51Flexor carpi radialisPronator teres

Distal 5Proximal 95

4 I8 7532 22 2283 59,

53f 53~

41, 4I?27~ 27~

*Internal topography of the median nerve as measured in serial sections. All distances are in millimeters.~’Sites where bundle localization was arbitrarily stopped at the level of the medial epicondyle, even though more proximal identification was possible.

-,ve ultimately used a system where one 20-micron-thick section out of every five was mounted for viewingand the other four discarded. This was possiblewere

~i because no significant change patternin funicular was

ever noted within a 100-micron length of serially stud-ied nerve. Ten such sections could be mounted on eachslide so that one slide represented 1,000 microns or 1.0millimeter of nerve.

Two figures are listed for each branch: The first in-terconnection is the level where any connection, re-gardless of size, was seen. The total measurable dis-tance is the level to which the branch could be tracedbefore intermingling with other bundles.

Microscopic dissection was performed with the ZeissOpMi 7 operating microscope at magnifications of 4.,5

~"~. to 25 times using standard techniques with jeweler sS-’~!!I:=~:. forceps, scalpel, and scissors. Photographs of represen-~" tative fields were taken through the microscope with a~Yashica FL camera and automatic light meter. The~ films were Kodak Ektachrome, ASA 64, and Tungsten

Ektachrome, ASA 160. Three median, ulnar, and m-s were examined in fresh autopsy specimens.

:In addition, numerous observations were made in theroom on patients undergoing nerve repair or

grafting. These latter observations confirmed our opin-ions but were not included in the data drawn from mi-crodissection.

Terminal branches o( the metor and sensory di’d-skins of the median and ulnar nerves were identified in

Table II. Median nerve microdissection distances*

Sunderland [ Authors

Thenar motor branch 56 84Sensory branches

1st web space 50 114.. Thenar sensory 97

Radial side, index finger ’ 972nd web space 58 373rd web space 66 37

Palmar cutaneousDistal 61 119-Proximal 24

Flexor digitorum superficialis 58Distal 2-4Proximal 22

Anterior interosseous 51 147,Pronator teres 1

Distal 5Proximal 98?

*Median nerve microdissecfion distances, comparing Sunderland’s data withauthors’. All distances are in millimeters and represent the rno~t proximal dis-tance to which each branch could be isolated within main nerve trunk.*Levels in the upper arm where branches could still be identified but dissectionwas arbitrarily stopped.

the palm and traced proximally by incising theepineurium and separating the branches from the re-maining nerve trunk for as far as they could be iden-tified and easily separatec;. T~i~ ,vas done by gentle

C~nnee:ions between the branches and other bundles

6 Jabaley, Wallace, Heckler

The JournalHAND

Fig. 4. A photomicrograph of the median nerve at the distal transverse carpal ligament level. Thebundles and branches are clearly identifiable and are separated by loose epineurium. They can beeasily dissected from each other without injury to the fascicles. [a, two fascicles comprising thethenar motor branch; b, the bundle of sensory fibers to the thumb and the radial side of the indexfinger; c, bundle of sensory fibers to second web space (index-middle fingers); d. already separatebranch to the third web space (middle-ring fingers).]

within the nerve were noted and their size and numberrecorded. The superficial radial and posterior interosse-ous nerves were similarly dissected.

The number which is given in the tables for eachbranch’s microdissection distance is the most proximallevel to which the branch could be readily and posi-tively separated as a distinct anatomic unit within themain nerve trunk. This is the point where substantialconnections with other bundles could be seen andwhere composition became speculative.

¯ Results

Median nerve. The median nerves provided the

most information since we examined them both his-tologically and by microscopic dissection (Figs. 2 and3). We can compare these observations with each otheras well as with those of Sunderland (Tables I and II).The distance from the medial epicondyle, where thehighest histological sections were taken, to the radialstyloid varied from 26 to 30 cm. The median nervesdivided into their terminal motor and sensory branches

on the average 4 cm distal to the radial styloid, and all

of the distances described for these branches are re-ferred to their origin. Branches which arose in theforearm are described in millimeters proximal to theradial styloid. The distances cited in the serial sectionsare based o+: :~ .~:,;.;:-,inati,.>n of a single nerve: thefigures for (ae microcissec~ions are representative of

the five arms examined. All nerves in each armnot dissected.

The motor branch to the thenar muscles. Thebranch to the thenar muscles rose as a singleall nerves and consisted of two fascicles. In the serialsections it joined the main trunk on its vo40 mm distal to the radial styloid (Fig. 4). It could identified and traced proximally in this positionabout 70 mm (30 mm proximal to the radial styloid).Few interconnecting bundles were seen_in this span.We saw the first connection at 35 mm after it joined themain trunk, a small branch to an adjacent sensory fas.cicle. We then followed the motor branch for an addi-tional 35 mm before it became thoroughly mixedthe sensory bundles from the hand. Thus we couldidentify and follow the moto.r branch for a total of 70mm by serial section, considerably less than the 112mm measured by Sunderland, but still to a level well

above the transverse carpal ligament.Microdissection of the thenar mofor branch was

formed in four median nerves (Table II). It joinedmain trunk 31 mm distal to the radial styloid, and.first visible connection appeared 28 mm proximalthis bony landmark. We could clearly trace thebranch to 59 mm where we found largebranches to the adjacent Sensory bundles. Sincein.-_e..,:+:~:~nections were nearly as large as the motor

b~anch itself, we could go no furzher

The Journal 5, No. 1qD [i January 1980 Internal topography of peripheral ~,erves 7

’he [5? ~ ¯~. . . ~’. ~,,~

the g .... ~ "’ ¯

each ~ were

ties. ~e motor ; ’single branch in~,es. In the serial~[~

Fig. 5. Serial sections of a 20 mm span of median ne~e in the lower fore~ viewed from proximalvolar-radial side to distal ~d showing sep~ation of the palmar cutaneous ne~e (PCN) from the main trunk of the~. 4). It could be median ne~e. In this patient the PCN consisted 6f two branches which ~ose 20 and 25 mm:his position for~ ~. proximal to the radial styloid. A, ~e PCN is a sep~ate bundle located on the left (radial) side of the~ mdiM styloid). ~f main ne~e. It is loosely held by thin epineufum. One l~ge fascicle comprises the distal ~unk (D):en in this span. ~ and two small fascicles form the proximal trunk (P). B, Both ne~e branches ~e still present, butafter it joined the ~

cent sensory fas- [~:¯ off and ~e no longer seen. D, ~e l~ge fascicle (D) now has dep~ed as a second branch of the

inch for an ad~- ~; PCN.tghly mixed with~[[i

Thus we could~e of the nerves exa~ned, the motor branch could ditional 40 mm before merging with other sensory

for a total of 70"~ isolated for over 80 ~ before fusion with other fascicles.

less than the 112:~~undles occu~ed. In the fou~h a connection was seen In the microdissected specimens; this branch ~ose

ill to a level well~~tween the motor b~ch and common digital ne~e to singly at 88 mm and could be easily sep~ated to 190~[~e ~kd w ebspace. It was quite small and was the only mm without visible connections. At207 mm, consider-

r ~nnection seen for the first 45 ram. able mixing with adjacent bundles prevented fu~er,r branch was pe "~5.p ,- ~-~. altair cutan,t tr, e:~

t thermal sections o

d to~ linens confi:~S~g of a singl,

motor~~0n the radial sid

~ge~loid and coul,

~ m heft

Flexor digitorum superficialis. We saw two branchesin the serial sections, the distal one at 102 mm and theproximal one at 208 mm. The distal branch entered onthe radial dorsal aspect of the main trunk as a singlelarge funiculus. Within 20 mm it had separated intofour smaller funiculi. At this level one small funiculusc’:~’~:.:~-,~ wit~ a ~ensor’: bundle, but the other three

’, If). It joined cutaneous nerve. Histological analysis of the dissection.

d styloid, and sections of the palmar cutaneous branch of both

mm proximal to confirmed two separate branches, each con-

~y trace the of a single moderate-size funiculus. They enteredthe radial side 20 and 25 mm proximal to the radiallarge connectin~

r~dles. Since .. cl be folio.wed proximally for 30 and 253re any change in position or number of funiculirge as the motor ~could be ~een 5i7. 5~. They could be traced an ad-

with certain~y.

8 Jabaley, Wallace, Heckler

~FleDeep motor branch

xor d~

~ Flexorcapri ,,~’~

Dorsal cutaneous

Little finger, ulnar side

Ring finger, ulnar side

Little finger, radial side

Fig. 6. Component microdissection of the ulnar nerve. The dark lines diagrammatically illustratethe dissection distance for each branch. The ring and arrow depict the anatomic origin of the nervebranch. The dark line distal to the ring and arrow represents the extraneural dissection distance; theline proximal represents the intraneural dissection distance.

The Journal ofHAND SURGERy~,

Table III. Ulnar nerve microdissection distances*

Sunderland I Authors

Terminal motor 90Terminal sensory 90Individual sensory

Hypothenar eminence 52Ulnar little finger 484th web space 48

Dorsal cutaneous 171Flexor digitorum

profundusDistal 11Proximal 22

Flexor carpi ulnaris 7DistalProximal

8181

272626209t107

9774

* Expressed in millimeters.-~Levels in the upI~"r arm where branche~ could still be identified but dissectionwas arbitrarily stopped.

proceeded proximally for an additional 55 mm withoutfurther major change.

The proximal branch entered as a single large fasci-

cle on the ulnar dorsal surface at 208 mm. It was easilytraced for 22 mm without interconnection.

By microdissection of this nerve in a fresh autopsyspecimen, we isolated four separate branches to themuscle, at 121 mm (radial side), I24 mm (posterior

surface), 185 mm (ulnar side), and 196 mm (ulnarside). Although the first two branches were little morethan filaments, they wire clearly nerve branches. Wetr-~ced them for 15 mm to substantiate this poin;:

they were too small to dissect further. The two moreproximal branches were much larger and both origi-nated from the same large bundle within the maintrunk. The more proximal branch of the two came off196 mm above the radial styloid, whereas the remain-tier of the bundle proceeded down to 174 mm, where itexited and traveled parallel to the proximal branch. Thecommon trunk which contained all branches of theflexor digitorum superficialis could be separated for 58mm without injury to any large interconnections.

Anterior interosseous nerve and flexor carpi radi-alis. The anterior interosseous nerve is the largestmotor branch of the median nerve in the forearm (Fig.2). In the histological sections it could be seen on theulnar-dorsal surface, arising at the same level (208 mm)but separate from the proximal branch of the flexordigitorum superficialis. It contained 14 fascicles andtraveled proximally in the same position on the outerperimeter of the remainder of the median nerve. Westill could clearly identify the anterior interosseousnerve within the main nerve trunk at 267 ram, the levelof the medial epicondyle, and we could see no inter-connections with other portions of the median nerveduring this 59 mm span.

By microdissection we located the anterior interos-seous nerve arising on the dorsal surface at 239 mm. Itcould be easily separated to the level of the medialepicondyle (293 ram) and for an additional 93 beyond. Its only attachments to the median nerveepineurial and these were e~.sily divided. The branch tott~.e flexor carpi radiaiis ac~uatly arose from t,~.e anterior

anal ofGERY

t~ origi-~e mainame offremain-where ittch. The; of thed for 58

¯ pi radi-; largestrm (Fig.:n on the208 ram)ae flexoricies andthe outer~rve. We~rosseousthe levelno inter-

ian nerve

~r interos-39 mm. Ithe mediald 93 mm~erve werebranch to

ae anterior

Vol. 5, No. 1January 1980 l mernal topography of peripheral nerves 9

Fig. 7. Dissection of the ulnar nerve in the lower third of forearm and palm. The lower arrow pointsto deep motor branch. The upper arrow points to sensory component. The area to the right of thearrow is the common sensory bundle. The area to the left of the arrow has divided into sensorynerves to the ulnar side of the little finger and the little-ring finger web space.

interosseous nerve in the histological sections. It en-tered at 214 millimeters on the dorsal-ulnar aspect. Thebranch consisted of three medium-sized fascicles whichjoined as they proceeded proximally to become two andfinally one, large fascicle. This single fascicle began 24,mm proximal to the branch point of the flexor carpifadialis from the anterior interosseous nerve. Thebranch to the flexor carpi radialis was traced to themedial epicondyle, a distance of 53 mm, without inter-connection to other bundles.

By microdissection we also found the nerve to the

carpi radialis originating as a branch of the an-terior interosseous and departing some 3 mm distal to

nerve’s origin. No significant connections wereseen in the first 20 mm proximal to its branch point, andwe made no attempt to trace it further within the an-

interosseous nerve.Flexor digitorum profundus. The flexor digitorum

was located on histologic section at 230 mman independent branch from the radial side of the

median nerve. An additional contribution to this musclefrom the anterior interosseous nerve.

Two fascicles, one large-and one small, comprisedthe independent profundus branch at its point of depar-ture from the main trunk of the median nerve (230

mm). They were traced proximally for 37 mm to thelevel of the medial epicondyle, where they joined toform a single large funiculus. No sec.tions were madeproximal to this point.

By microdissection the branch to the flexor digi-torum profundus arose from the anterior interosseousnerve. Although it was dissected for only 10 mm prox-imal to its origin, the branch remained separatexrcerthis span.

Pronator teres nerve. The pronator teres nerve wasthe most proximal branch of the median nerve exam-ined histologically. It arose as two separate branches at226 and 240 mm from the radial-volar aspect. Thesebranches contained two and three f~aniculi, respec-tively, and could be easily identified at the level of themedial epicondyle (267 mm). No detectable connec-tions with other bundles were seen between.the originof these branches and the medial epicondyle, distancesof 41 and 27 mm, respectively.

By microdissection five separate branches to the pro-nator teres arose from a single large trunk whose originwas wel! proximal to the medial epicondyle (303 ram).This trunk was 12 mm long and was easily dissected for100 mm with no evidence of connecting branches too~he.r ?.ortions of the median nerve.

I0 Jabaley, Wallace, Heckler

The JournalHAND SURGERy

Fig. 8. Dissection of the dorsal sensory branch of ulnar nerve in the forearm. This branch (arrow)could be isolated from the main trunk, beginning at its normal anatomic origin and going to a pointwell above the elbow without transsecting any interconnections. This distance is approximately20 cm.

Ulnar nerve. We dissected four ulnar nerves intotheir component bundles with the aid of magnification(4.5 to 25 times) (Fig. 6), but we performed no tological examinations of this nerve (Table liD. Be-ginning in the palm at the terminal branches, the motorand sensory components (deep motor branch and common sensory trunk which became the sensorynerves to the ulnar side of the little finger and the fourthweb space) could be separated for a total length of 81mm (Fig. 7). In one nerve these two branches remainedciearly separate to the midforearm level. No connec-tions were noted before 21 ram; those seen were smallfilaments and did not make up a significant portion ofthe trunk. Within the sensory component the branchesto the ulnar side of the little finger and the fourth webspace were separable for 26 mm.

The level at which the dorsal cutaneous branchjoined the main trunk of the ulnar nerve was variable.In one specimen it entered as far distal as 48 mm prox-imal to the radial styloid; another joined at the 91 mmlevel. It traveled proximally as a separate bundle for209 mm, well above the epicondyle, without demon-strable connections to the main trunk (Fig. 8). It was asthough the dorsal cutaneous nerve and the ulnar nervewere two separate nerves traveling .within a common

epineural conduit while still retaining their autonomy.This arrangement was found in all of the ulnar nervesexamined.

Multiple motor branches to the two main forearmmuscles, the flexor digitorum profundus and the flexorcarpi ulnaris, were found in all instances. They some-times arose separately from the main trunk of the ulnarnerve and sometimes as branches of a common trunk,but always in the region of the elbow. Representativelevels of origin were 240 to 280 mm; representativedissection distances were 74 to 107 mm. These dis-tances permitted separation of the flexor digitorum pro-fundus and the flexor carpi ulnaris well into the upperarm.

Radial nerve. Technically, th~ radial nerve is not aforearm nerve, since it terminates at the level of themedial epicondyle. For this reason, we did not study itin detail (Fig. 9).

We dissected the posterior interosseous and super-.ficial radial nerves above their origins and found thai:we could easily separate them from each other for 90mm before encountering the first interconnection(Table IV). No attempt was made to separate individualmotor branches of either the radial or posterior interos-:

seous nerve. "

The Journal of~~ND SURGERY

5, No. 11980

Internal topography of peripheral nerves 11

~eir autonomy.te ulnar nerves

main forearmand the flexors. They some-nk of the ulnar:ommon trunk,Representativerepresentative

tm. These dis-digitorum pro-into the upper

[ nerve is not aae level of thedid not study it

~us and super-and found that:h other for 90interconnectionarate individual)sterior interos-

Posterior interosseous ~,~ ~

Fig. 9. A component microdissection of the radial nerve. The dark lines diagrammatically illustratethe dissection distance for each branch. The ring and arrow depict the anatomic origin of the nervebranch. The dark line distal to the ring and arrow represents the extraneural dissection distance; theline proximal represents the intraneural dissection distance.

Discussion: Studies conducted by numerous investigators over

the oast 70 years have demonstrated that the internalstru~:ure of peripheral nerves is characterized by nu-merous intercommunications, position changes, and

:the formation of funicular plexuses,a Even in the early1900s some investigators recognized the basic plexi-form character of the intraneural funicular pattern. Theuestion of functional significance was rarely raised,

~S:. however. Stimulation of exposed nerve by electrical::~~5currents suggested that the funcnona course of axons is~2.a straight one, even ff the fumcular course is not, but~2the methods employed made these studies suspect.

~:~}~ Now it has become important that we try to determine:~:~ in what av and to what de ree internal architecture"2~. ........ w_~ g

i~mgtters in clinical practice.~ .Sunderland performed the exhausting task of serial

~~l~istological section and examin.ation of the major~!,nerves of the extremities, catalogxng both the arrange

of individual fascicles and the distances overthey could be identified. Furthermore, he per-microdissections of the branches and recorded

distances over which they could be isolated fromthe main trunk of the nerve "without damaging or in-

with the fiber architecture." Although sometook note and recognized the significance of

i this latter work, Sunderland’s microdissections appearhave been largely unnoticed by most surgeons.

.... In his writings Sunderland emphasized that in theproximal portion of nerves, the majority of the

Table I¥. Radial nerve microdissection distances

I Sunderlandl Auth°r"

Posterior interosseous (ram) 72 90

Superficial radial (mm) 72 90

funiculi contain representative fibers of most if not.all of the peripheral branches. His composite model ofthe musculocutaneous.nerve illustrates this point (Fig...1). At more distal levels, however, a regrouping ofthe fibers is gradually effected, whereby individualbranches become identifiable as such and come to oc-cupy different funiculi and bundles. He conch.tded thatthe purpose of the intraneural plexuses seems to be "toassemble the requisite afferent and efferent fibers foreach branch from the appropriate segmental sources."He further thought that often a plexus is "the product ofmesenchymal condensations which occur during devel-opment, regardless of the destina.tion and function ofthe fibers." This is certainly possible, given the capri-cious nature and variation of developmental anatomy.

As we come to understand better the complex cir-

cuitry of the nervous system, we gain/in appreciationfor the large number of information interchanges re-quired. The refining process begun at the spinal cordlevel must continue through the larger nerves and theirbranches until individual fibers can be appropriatelypositioned at their target organs. The sorting probablyoccurs irrespective of whether the fibers are motor or

12 Jabaley, Wallace, HecklerThe Journal

HAND SURGER

Fig. 10. The current concept of the internal topography of a 3cm section of the median nerve in the midportion of theforearm. Since this representation is of the more distal seg-ments of the nerve, fascicular branching or mixing is in-frequently seen, either in cross-section or longitudinal mi,crodissection. This appearance differs from that seen in Fig.1, which represents the musculocutaneous nerve at a moreproximal location.

.sensory. In fact, Terzis8 has referred to her unpublishedwork which suggests that even a nerve such as the deepmotor branch of the ulnar nerve, heretofore thought tobe a "pure motor nerve," may contain as much as 50%sensory fibers. Despite these obligatory changes in fiberlocation, it now appears that the forearm and hand arecharacterized by connections within, not between,bundles, and that bundles are separable surgically.

Perhaps the most important aspects of Sunderland’swork are these: "Despite the changing plexiformcharacter of the funicular pattern, fibers from the pe-ripheral branches pursue a localized course in thenerve for variable, though often considerable dis-tances." The redistribution of bundles implied in theplexus principle "is local and not general." Fibers"scatter gradually" and, even with intermingling.

"remain in the same quadrant over long distances.Moreover, he points out that, "there is no proof that th~intercommunications effect a complete reorientation or!redistribution of fibers or that the intercommunicaterased localization." In fact, he even cautions that""conclusions based on such investigations have given~

very misleading results." Despite Sunderland’s clearrecognition of the plexiform nature of intraneural to-pography, these statements do not imply a constant~(and hopeless) intermixing of the components of ripheral nerve. Rather, they suggest an orderliness andpurpose which might be used to advantage in treating:mechanical ailments and injuries if it is recognized and::understood.

It is noteworthy that subsequent investigations byothers have confirmed these points,~ and we are una-:ware of any contradictory studies. Analysis of ourbased on serial sections of the median nerve furtheri:documents Sunderland’s interpretation of his mi-croscopic results. In the 33 years which have elapsedsince Sundertand’s publication, many surgeons have!!grasped the fact that funicular plexuses and intercon-~nections exist but they have failed to appreciate hisdemonstration that fitnctional units (and often discretebranches) remain localized in the same quadrants ofnerve trunks for considerable distances and are acces-:isible to surgical manipulation and repair.

Comparison of the absolute distances in ourand those of Sunderland show many discrepancies andthese distances deserve comment (Tables II, III, and~IV). Dissection distances are not important in and of:~themselves, for they vary from nerve to nerve and per~!!son to person. Moreover, landmarks such as bony~prominences cannot be strictly defined, and-actual mea-~suring techniques are imprecise. Finally, in fresh spec-~

imens the epineurium is loose and one can vary theorigin of a branch by a few mtlhmeters s~mply by tug-ffging on it. For all these reasons, the absolute distances~

except~that branches travel are probably unimportant,insofar as they relate to each. other. The point to be~

made is that, while some internal sectors of a nerve~may be changing, other portions may proceed for con-~siderable distances with no major change in position or ~.:

composition and thus may be isolatedsurgically. If one..~is to perform intraneuml dissections, an understanding~of the branch patterns and bundle arrangements is of the.~’~utmost importance.

It is impossible, of course, to study the courses oindividual axon fibers by the techniques employedthis study. Sunderland assumed that constituent fibers~’i:’.termin~!ed whenever there was fusion of f:~’niculi and

me 1oumal ofD SURGERY

Vot. 5, No. IJanuary 1980 lnternal topography of peripheral nerves 13

~ distances."proof that the3rientation ormmunicationcautions that~s have given:rland’s clearatraneural to-ly a constantaents of a pe-rderliness andge in treating~cognized and

.~stigations by~ we are una-sis of our datanerve fuaherof his mi-

t have elapsedmrgeons haveand intercon-

appreciate his’ Qften discrete~ quadrants of"~nd are acces-

s in our studyc~pancies andles II, m, anda~t in and ofne~e and per-such as bony

and actual mea-,, in fresh spec-,e can va~ the~ simply by tug-solute distancestpo~ant, except"he point to betors of a ne~e,roceed for con-~e in position or~trgically. If onen understandinggements is of the

the courses o~es employed in:onstituent fibers

of funiculi and

Fig. 11. A, A partially dissected median nerve, seen through the operating microscope. Theundissected portion is at the bottom of the photograph. The upper portion shows several parallelbundles and fasciculi which were separated by incising their epineurium. Only an occasionalconnection is seen. B, A high power view of the mixed area. There are severn1 large parallelfasciculi with -no intercoanection. A medium-sized connection is seen in the center of the field and asmaller one is located at the-lower right.

14 Jabaley, Wallace, HecklerThe Journal of:

HAND SURGERY

that subsequent separations left representatives of allthe daughter fibers in each newly formed fascicle. Hefurther classified plexuses as (I) those where all thefascicles are derived from one branch and wherechanges are confined to bundles of like origin. (2)Those where the fascicles of a bundle or branchblended with fascicles of a nearby but separate bundle,i.e., one of a different origin. This implies a blendingof unlike fibers. (3) Those occurring at the most proxi-mal levels of a nerve, where bundles which alreadycontain fibers of diverse sources divide and recombinerepeatedly. These three types of union probably occurin a distal-to-proximal direction and result in progres-sively greater fiber, mixing as nerves become nearer totheir central origin. Since our study was limited to theforearm and hand, many of the changes we noted infunicular pattern were similar to Sunderland’s firsttype: as two adjacent funiculi moved closer to eachother, the intraneural epineurium between them thinnedand then disappeared; the position and appearance ofthe bundles changed, but there may have been no actualmixing of fibers (Fig. I0). If this supposition is true,then bundles which behave in this fashion can probablybe separated surgically until a level is reached wherefibers actually pass from one bundle to another. Suchfiber mixing rarely occurs when branches and bundlesfirst fuse, but only at more proximal levels in the nerve.

In our microdissections of the median, ulnar, andradial nerves, we found that major branches could beseparated from the main trunk for considerable dis-tances by incising the epineural sheath and teasing theintraneural epineurium apart (Table II, III, and IV).These data are comparable to similar observations bySunderland3 and Seddon.1° Further corroboration canbe found in those clinical studies which describe theseparation of the motor branches for transposition ofthe ulnar nerve at the elbow and separation of the motorand sensory branches of the ulnar nerve at its termina-tion to permit rerouting and individual repair of thebranches in the palm. In addition, neurolyses of themedian nerve at the carpal tunnel are performed suc-cessfully without functional loss. More recently, Mil-lesix~ has reported separating the palmar branches of themedian nerve "to the middle or proximal half of theforearm" in preparation for specific funicular grafts tothese branches. These observations all lend credibilityto the conclusion that it is possible to perform suchseparations in the clinical setting as well as in the lab-oratory. They suggest that many of the interconnectionsfound below the elbow in the median and ulnar nervesmay be of less surgical importance, than has beenthought heretofore.

In our microdissections of the fresh autopsy materialwe were abletto identify large interconnecting branches"iwith ease (Fig. 11). There were relatively few suchconnections in the distal portion of the forearm.though it is possible that smaller branches were presentand could not be seen through the microscope, it ismore likely that microdissection is an effective way ofseparating bundles and funiculi with very little damageto whatever connections exist. It is even possible thatthe plane of clinical dissection was between funiculiwhich might have appeared to be associated in a singlebundle by histological examination. Since limited dis-sections were performed above the medial epicondyleof the humerus, we cannot say whether such separa-tions would also have been possible in the more proxi-mal portions of the nerve where considerable inter-mingling of fibers is thought to occur. Sunderland’sdissections at this level suggest that such separationwould have been less likely.

An understanding of the character of the non-neuralconnective tissue layers of peripheral nerves suggests apossible explanation for the ease of dissection of indi-vidual bundles’-° (Fig. 12). The epineurium invests theentire nerve trunk as well as groups or single funiculi.The intraneural epineurium is loose and areolar, it tearseasily, and it possesses relatively little strength, but it isthe stuff which separates the bundles. Perineurium, onthe other hand, surrounds each funiculus, has a tightlamellar composition, and is quite strong. It is the ma-terial one sews when making an interfascicular stitch,and it is the natural plane of dissection upon which oneseparates fasciculi. In performing interfascicular dis-section, the tissue which gives way most easily isepineural connective tissue. Thus one would expectthat in the relatively distal segments of p-6"ripheralnerve, the bundles, might be separated into their com-ponent funiculi with little or no loss of function becausethere are relatively few interconnecting fibers. Ourfindings suggest that this is indeed the case.

It is unclear why a funicutus would travel first withone bundle and then with another’. Sunderland’s expla-nation that it is part of the sorting process which beginswith spinal segments and evolves to branches is mostacceptable. Furthermore, his idea of fortuitous conden-sation of mesoderm also may be true. The impressionwe have formed is that fiber and funicular behavior, atleast in the more distal portions of nerves, is ratherpurposeful and that it is the random wandering ofepineurium which is responsible for much of the changeobserved. Although this may seem a subtle distinctionits importance must be emphasized. It is quite a differ-ent ma~er ’,o suggest that the epineurium shifts and net

e Journal of . :~:;.%, Vol. 5, No. 1

SURGERY ~ January 1980

tg branchest few suchrearm. A1- ~’ere present;cope, it is:ire way of:tie damage9ssible that

ep.icondyleach separa-

table inter-anderland’sseparation

non-neural~ suggests aion of indi-invests the

~te funiculi.3lar, it tears..,th, but it isleurium, onhas a tightt is the ma-:ular stitch,~ which onecicular dis-st easily is3uld expect’ peripheral) their corn-:ion becausefibers. Our

el first withrod’s expla-~hich begins:hes is mostous conden-~ impression’~ehavior, ats, is rather~ndering ofCthe changedistinction,rite a differ-rifts and not

Internal topography of peripheral nerves 15

Fig. 12. A scanning electron micmgraph of a dog’s nerve. About the outer portion, the epine’uriumis loosely and randomly arranged. The tighter and more orderly perineurium encloses the crescent-shaped mass of individual axon fibers. Each axon occupies a separate area and is surrounded byendoneurium. In interfascicular dissections the perineurium is the natural plane of dissection.

~.the fasciculi, but this seems to be the case over theencountered in our lower forearm dissections.

What is certain and what must be stressed is that thebetween the cell body in the dorsal or ven-

ganglia and the end organ, whether sensory oris continuous. Over the distances commonly

encountered m peripheral nerve repair, lysis, or graft, it

appears that the fibers, collected in small funiculi,travel in identifiable bundles and can be separated sur-gically with little or no damage to the component parts.Since most injuries occur in the more distal portion of anerve, the funicular arrangement and relationship atthis level is critical and appears to favor the surgeon.

We have speculated ~e,:bre that the nervous system

16 Jabaley, Wallace, Heckler

is blessed with many times the number of sensory fi-bers necessary to perform basic functions, providedtheir central connections are appropriate and undis-torted. 13, 14 Thus a reduction in functioning axons ofsome measurable percentage, if it could be accom-plished in an orderly fashion, still might permit rela-tively normal sensation which is simply less dis-criminating. This may well happen when one performsa neurolysis or separates branches of a nerve. It almostsurely occurs in a repair or graft where regeneratingaxons are fewer in number and may reach similar butdifferent end organs. Useful function returns, but thestrength or ability to discern is reduced. This phe-nomenon occurs naturally in peripheral neuropathiesand affects both motor components, where it is easilymeasured, and sensory components, where it is not.

Clinical applications

Carpal tunnel and other entrapment syndromes.Curtis and Eversman15 suggested, and others have cor-roborated, that internal neurolysis, when indicated, canbe safely carried out in patients with carpal tunnel syn-drome and that functional recovery will be more com-plete than in those patients in whom only division of thetransverse carpal ligament is performed. We now knowthat axons can be compressed by an epineural sheathwhich is scarred or thickened, as well as by a tightcarpal tunnel, and that only incision of this epineuriumwill relieve symptoms of intraneural compression. Ourstudies suggest that in such cases, whether at the elbowor wrist, the bundles or individual funiculi can be safelyseparated and that little or no functional loss is likely tooccur.

Primary or secondary repair of transseetednerves. In cases where a nerve is sharply transsected orwhere the zone of injury is limited, gaps of up to 5 cmmay be encountered, and primary or secondary repair,including excision of neuroma and glioma, can be per-formed. It is not our purpose here to discuss the role oftension in the outcome of such cases, and whether suchgaps should be managed by direct suture or by inter-positional graft. We only wish to emphasize that it ispossible to recognizd anatomically similar bundles inthe proximal and distal stumps at these distances andto connect them correctly by whichever means onewishes. In the forearm and wrist, distances of a fewcentimeters have the anatomic potential for recoveryof motor and sensory function, other factors notwith-standing.

Interpositional grafts. As the gap between nerveends increases, there occurs a point ~vhere an interposi-tional graft becomes necessary. The technique of inter-

The JournalHAND SURGERY~

fascicular grating seems most appropriate in such in-stances.’6 A technical disadvantage of this approachthe difficulty one encounters in matching seemindisparate portions of the proximal and distal ends whena large gap exists. We should emphasize that, despitethe disparity in fascicular patterns, fibers remain lo-calized to the same quadrant of a nerve over a distanceof several centimeters, and it is possible to selectmatching motor and sensory components on a quadrantbasis. Both Sunderland and Millesi lr spoke to thispoint, and Sunderland observed that in such instancesthe surgeon "must use his best clinical judgement tomatch groups of funiculi." Millesi has adopted Sunder-land’s term "’group fascicular matching" and has usedit to describe his technique for choosing bundles to beconnected by grafts. As used by Millesi, "group fas-cicular matching" means just that: bundles or groups ofbundles seen at the cut end of a nerve are selected forgrafting on the basis of size, number, and quadranticlocation. It is impossible to match each individual fas-cicle and it is unnecessary and probably unwise to try.

Sunderland’s initial publication on the topography ofthe nerves of the upper extremity contains tables whichlist the components of each quadrant of each nerve atseveral levels and compares them with the findings ofother investigators. These tables can be invaluable tothe surgeon who undertakes to bridge a large gap by anerve graft.

Under ideal circumstances, it appears that sufficientnumbers of regenerating axons will cross two suturelines, enter the distal stumps, and grow to their termi-nations to make nerve grafting a successful procedure.Although it is unlikely that any fibers reach their origi-nal destination, many may at least find ttmir way tosimilar funiculi, so that motor fibers are directed tomotor fasciculi and sensory fibers to sensory fasciculi.The level of functional recovery in such patients, ~though not ideal, is better than has been possible previ-ously and more than justifies the procedure.

We can now make some ge.neral statements aboutspecific segments of each forearm nerve which may beof some value in accomplishing alignment of bundlesor quadrants. These comments are based on both the ~.~microdissections and analysis of bundl~ location in the ~serial sections.

In the median nerve in the upper third of the forearm ~i

the motor branches to the extrinsic muscles lie about ~

the periphery, primarily on the radial and ulnar sides ~

(mF, ig. 2).ruThc~ensory branches to the hand, the thenar :~OlvO~ ~era~ ~ s, and the palmar cutaneous branch oc-i~c~ "h~ en ral core and dorsal quadrant. Thus, if one

v~~shes to bridge a gap at this level by grafting, the

real of Vol. 5, No. 1GERY 1980 Internal topography of peripheral nerves 17

core of the proximal stump should connect dis-with the motor and sensory components of the

i.hand, and the large identifiable forearm motor branches~ould be attached about the periphery of the proximal

end.. In the middle third of the forearm, there are very few

!: i branches and no major quadrant changes. Here corre-sponding quadrants can be connected directly or with

i:~.~ sections of graft as necessary. If appropriate quadrantsthe likelihood of a correct match-up is good.

In the distal third of the forearm, the thenar motor,~mbrical, and sensory components are well segregated

i and each can be isolated for repair. The distal compo-

nen:s, of course, can be positively identified by dissec-tic.:~, and attached to proximal bundles chosen on thebasis of quadrantic location. Specific bundles can be

: identified and expected to conform closely for about 5cm above the wrist.

:-~i: In a cooperative patient, one might use the stimula-

tion technique of Hakstianis to further identify proxi-:- real elements. Although the cross-sectional appearance

of the median nerve changes in all of these zones, thechanges are relatively subtle and occur over relativelygreat, distances. They are usually between fascicles of

¯ fi,c ~ame or similar branches; these bundles lend them-selves to connection with reasonable accuracy..

In some respects the ulnar nerve provided the most<-~ interesting and potentially valuable information of the

entire study. In the elbow region, the ulnar motor.... branches are well defined and can be separated for long

~distances before encountering interconnections. Such aseparation facilitates anterior transposition of the maintrunk, a procedure which is still indicated at times for:relief of compression or entrapment symptoms. In ourdissections the terminal motor and sensory branches at

.~ Guyon’s canal had only a few fine filamentous connec-.tions and could be easily separated for a full 8 cm. Forthis reason an injury in this area offers several treatment

Among them is a separation of the motor corn-for some distance proximally and rerouting it

~ through the carpal tunnel. This maneuver gains lengthpermits a tension-free juncture of the motor branch

having to resort to a graft. If one wishes todirect suturing or to interpose a graft, the ease

of separation and identification of the terminal motorbranches of the ulnar nerve suits them well

for these procedures.The dorsal cutaneous branch of the ulnar nerve is

.unique among major forearm branches. Although itjoins the main trunk in .the distal third of the

forearm, it progresses independently along-side the~:. main nerve trunk tt-~;ough the forearm and well proxi-

mal to the medial epicondyle. In reality it is a separatenerve travehng in the same ep~neural sheath with theulnar nerve. The situation is similar to the arrangementof the tibial and common peroneal nerves. It offerssome intriguing opportunities. In an acute ulnar nerveinjury at the midorearm level proximal to the separationof the dorsal cutaneous nerve, this branch can be dis-sected out and repaired separately, thereby increasingthe likelihood of accurate alignment of the remainingmotor and sensory components. If an injury at thissame level requires a nerve graft, yet another possibil-ity exists. In some instances, the dorsal cutaneous.nerve may be considered expendable. Should this bethe case, it might be possible to dissect it for somedistance and to use it as a graft. This maneuver couldprovide as much as 25 cm of graft and have three majoradvantages: (I) a graft from the same operative site, (2)a reduction in the cross-sectional size of ulnar nerverequiring a graft, and (3), again, a corresponding in-crease in the accuracy of alignment.

Little can be said about the radial nerve, since solittle of this study pertains to it. It is not primarily aforearm nerve, but it is clear that its major sensorybranch and the posterior interosseous nerve can beseparated for several centimeters above the medialepicondyle without difficulty and that the principlespreviously outlined for the median and ulnar nervesalso apply to the radial nerve. Again, the intriguingpossibility exists for using the radial sensory compo-nent for a donor graft in irreparable injuries above theelbow, by dissecting it out and using it to bridge themotor deficit in the posterior interosseous nerve. Al-though there is a possibility of painful neuroma forma-tion in the superficial radial nerve when it is injured__more distally, the rarity of this problem in the upperarm makes its use as a donor nerve at least a possibility.This suggestion is speculative and we have not usedthis nerve as a graft.

Finally, a word of caution: The study of Rydevik,Lundborg, and Nordborg suggests that intraneural dis-section in animals may induce fibrosis witfiin the nerveand aggravate rather than relieve intraneural constric-tion which has resulted from epineural scarring. ~Inci-sion of the epineurium and interfascicular dissection isa procedure which should be reserved for the specificsituations where it is likely that epineurium is constrict-ing the fasciculi and producing symptoms which cannotbe relieved by external decompression alone. Our datasuggest that this procedure, if performed carefully, canbe accomplished without significant injury to theperineurium and without further loss of function. Simi-larly, if intraneural dissection is indicated to facilitate

18 Jabaley, Wallace, Heckler

repair, rerouting, or grafting of peripheral nerves, it canbe performed, but one must always weigh the potentialgain against the risk of further damage.

A project of this magnitude has many contributors and wegratefully acknowledge their contributions: Mr. ThomasWhite, an aspirant medical student, prepared all the mi-croscopic sections. Mr. James Goodman created the originalart work, and Mr. William DeVeer produced the photo-graphs. Ms. Cheryl Lyon expertly prepared the manuscript,and Ms. Mary Lou Percy provided the scanning electron mi-crograph. Dr. Douglas Gorman assisted with the mi-crodissections.

REFERENCES

1. Langley JN. Hashimoto M: On the suture of separatebundles in a nerve trunk and on internal nerve plexuses. JPhysiol 51:318-45, 1917

2. Lewis D: Principles of peripheral nerve surgery. JAMA75:73-7, 1920

3. Sunderland S: The intraneural topography of the radial,median, and ulnar nerves. Brain 68:243-98, 1945

4. Manktelow RT. McKee NH: Free muscle transplantationto provide active finger flexion. J H,xrqo SURG 3:416-26.1978

5. Sunderland S: Factors influencing the course of regener-ation and the quality of the recovery after nerve suture.Brain 75:19-54, 1952

6. Sunderland S: Nerves and nerve injuries, ed 1. Edin-burgh, London, 1968, Churchill-Livingstone, Ltd

7. Sunderland S: Nerves and nerve injuries, ed. 2. Edin-burgh, London, 1978, Churchill-Livingstone, Ltd, p 32

8. Terzis JK: Principles, practices, and techniques of pe-ripheral nerve surgery, in Daniels and Terzis JK, edi2"

HAND SURGERy

tots: Reconstructive microsurgery. Boston, 1977, LittleBrown g Co. p 387

9. Tamura K: The funicular pattern of Japanese peripheralnerves. Arch Jap Chir 38:35-58, 1969

10. Seddon H J: Surgical disorders of the peripheral nerves,~Baltimore, 1972, Williams & Wilkins Co, p 267

11. Millesi H: Surgical management of bmchial plexus in- ~.~.’juries. J HAteD SURG 2:367-79, 1977

12. Kuczynski K: Functional microanatomy of the peripheral :~;-~.nerve trunks. Hand 6:1-10, 1974

13. Jabaley ME, Bums JA, Orcutt BS, Bryant WM: Com-~parison of histologic and functional recovery after pe-ripheral nerve repair. J HANt~ StJRG 1:119-30, 1976

14. Jabaley ME: Recovery of sensation following peripheralnerve repair, in Fredericks S, Brody GS. editors: Sym-posium on the neurologic aspects of plastic surgery, vol17. St. Louis. 1978. The CV Mosby Co, pp 36-46

15. Curtis RM, Eversman WW: Internal neurolysis as anadjunct to the treatment of the carpal tunnel syndrome. JBone Joint Surg [Am] 55:733, 1973

16. Millesi H, Meissl G, Berger A: The interfascicular nervegrafting of the median and ulnar nerve. J Bone Joint Surg54:727-50, 1972

17. Sunderland S. Millesi H: Nerve repair: its clinical andexperimental basis. A symposium in honor of Sir SydneySunderland. Nov 3-5, 1977, University of California,San Francisco. Calif (to be published)

18. Hakstian RW: Funicular orientation by direct stimula-tion: an aid to peripheral nerve repair. J Bone Joint Surg[Am] 50:1178-86, 1968

19. Ryderik B, Lundborg G, Nordborg C: Intraneural tissuereactions induced by internal neurolysis. Scand J PlastReconstr Surg 10:3-8, 1976