JAAOS - Volume 05 - Issue 02 March 1997

Vol 5, No 2, March/April 1997 59

The extensor mechanism of thehand and digits has received con-siderable attention in the recon-struction of chronic deformities,but acute injuries have received farless attention. Although the exten-sor mechanism is relatively superfi-cial and therefore easy to expose,operative repair can be technicallychallenging because the extensortendons are thin and flat, are weak-er than flexor tendons, have lessgliding amplitude, and are difficultto suture well.

There has been a resurgence ofinterest concerning acute extensortendon repairs because outcomesafter traditional static splintinghave been shown to be less thanacceptable by contemporary stan-dards.1-5 Investigations haveshown that dynamic postoperativeprotocols can improve the out-come.6-10 Stronger suture repairtechniques and postoperative pro-tocols with active motion are cur-rently being developed.11,12 Thisarticle will review the diagnosisand treatment of acute open andclosed extensor injuries in zones I

through VIII, as well as recentadvances in treatment and rehabili-tation.

Anatomy

The extensor mechanism of thehand can be divided into eightzones (Fig. 1) to aid in the evalua-tion and treatment of acute in-juries.13 The even-numbered zonesare over bones, and the odd-num-bered zones are over joints. The dif-ferent numbering for the extensormechanism of the thumb reflects itssmaller number of phalanges.

Zone VIII, the most proximalzone, contains the musculotendi-nous junction. In zone VII, the ten-dons lie within an envelopingtenosynovium and the sheath ofthe extensor retinaculum over thewrist joint. Under normal circum-stances, diffusion is responsible formost of the tendon nutrition in thisregion.14 The vascular supply isderived from the mesotendon,which runs the entire length of theretinaculum. In the remaining

zones, the extensor tendons arecovered by paratenon, and nutri-tion is supplied primarily by perfu-sion. The intratendinous vasculararchitecture outside the retinacu-lum is similar throughout the ten-don, with small branches to thetendon from the surrounding fas-cia.15

In zone VI, the tendons of thelong, ring, and small fingers areconnected by juncturae tendinumthat course distally and obliquely.These interconnections must beconsidered when evaluating exten-sor tendon injuries. The extensorindicis proprius and extensor digitiquinti, which run ulnar to theirrespective extensor digitorum com-munis (EDC) tendons and allowindependent extension of the indexand small fingers, must also beconsidered in an evaluation.

In zone V, the extensor tendonsare centrally located over themetacarpal head, held in place bysagittal bands that run on each side

Dr. Newport is Associate Professor,Department of Orthopaedic Surgery,University of Connecticut Health Center,Farmington.

Reprint requests: Dr. Newport, Department ofOrthopaedic Surgery, University ofConnecticut Health Center, 10 Talcott NotchRoad, Farmington, CT 06034-4037.

Copyright 1997 by the American Academy ofOrthopaedic Surgeons.

Abstract

Until recently, extensor tendon injuries were often discounted as an importanthand problem. However, studies have shown that not all extensor lacerationsfare well and that loss of flexion can be problematic. Newer postoperative proto-cols emphasizing tendon gliding have improved results, and better repair tech-niques and postoperative rehabilitation regimens are under investigation. Thisarticle reviews the evaluation of acute open and closed extensor tendon injuries,their conservative and surgical treatment, and postoperative rehabilitationoptions.

J Am Acad Orthop Surg 1997;5:59-66

Extensor Tendon Injuries in the Hand

Mary Lynn Newport, MD

of the tendon and attach palmarlyto the volar plate of the metacarpo-phalangeal (MCP) joint. The ex-trinsic extensor insertion into theproximal phalanx itself is quiteweak.

The extensor mechanism be-comes more complicated over thedigit as the EDC becomes linked tothe intrinsic mechanism of the lum-bricals and the dorsal and volarinterossei. The EDC tendonsextend the MCP joints by pullingup on the sling formed by thesagittal bands. The EDC tendonsare also capable of extending theinterphalangeal joints if hyperex-tension of the MCP joints is pre-vented. Resistance to MCP hyper-extension is provided primarily bythe intrinsic musculature, as thevolar plates of the MCP joints donot prevent hyperextension. Be-

cause the intrinsic tendons aredirected volar to the axis of theMCP joint, they flex this joint. Theintrinsic tendons continue distallyas lateral bands, with contributionsinserting into the dorsum of themiddle phalanx along with thecentral slip of the EDC. The lastcontinuations of the lateral bandsmeet dorsally at the midportion ofthe middle phalanx, forming theterminal extensor tendon, whichinserts on the distal phalanx. Be-cause the intrinsics course dorsal tothe axis of the proximal interpha-langeal (PIP) and distal interpha-langeal (DIP) joints, they extendthese joints.

The intrinsic and extrinsic exten-sor systems are intricately coordi-nated during joint flexion andextension. Injury or adhesion for-mation between the extensor mech-anism and adjacent tissues will dis-turb this delicate balance andpotentially limit tendon excursionand joint motion.

History

Despite the work of Dargan1 andothers2,3 relating good and excel-lent results in only two thirds ofcases, many review articles andhandbooks indicate that extensortendon injuries do uniformly well.In a large series of extensor injuriestreated by static splinting, Newportet al5 reported good or excellentresults (using a rating systemdeveloped by Miller16 [Table 1]) inonly 52% of cases. They showedthat good or excellent resultsoccurred more frequently in thefour proximal extensor zones thanin more distal zones (65% vs 40%).5These results are consistent withthe anatomy, as the extensor mech-anism in the proximal zones is lesscomplex and less apposed to bone.The extensor mechanism in thedistal zones, as previously men-

tioned, is significantly more com-plex and more closely apposed tobone.

Hung et al9 and others17 havealso noted poorer results in injuriesover the digit. An unexpectedfinding in the series by Newport etal5 was that injuries within the reti-naculum of zone VII did not haveoutcomes different from those inzones VI and VIII; all three areashad approximately 66% excellentor good results. Others had pre-dicted a poorer result in zone VII,hypothesizing that the surroundingtenosynovium and the enclosingretinaculum would be a greatersource of adhesion formation.8,17

Two thirds of all extensor ten-don lacerations are associated withconcomitant injury to bone, skin, orjoint. In three large series,3-5

injuries associated with fracturehad a poorer outcome (50% goodor excellent results); joint capsulelaceration did not adversely affectoutcome (64% good or excellentresults). Overall loss of flexionexceeded overall loss of extension.5

Dynamic splinting for extensorinjuries has been shown to marked-ly improve results compared withstatic splinting, with 98% to 100%good or excellent results.6,7,11

Dynamic splinting typicallyinvolves a rubber-band outriggerapparatus (Fig. 2, A). It is based on


Journal of the American Academy of Orthopaedic Surgeons60

Fig. 1 The eight extensor tendon zones. T = thumb.

I

II

III

IV

V

VI

VII

VIII

T I

T II

T III

T IV

T V

Table 1Miller’s Classification of ResultAfter Extensor Tendon Repair

Total TotalExtensor Flexor

Result Lag Loss

Excellent 0° 0°Good ≤10° ≤20°Fair 11°–45° 21°–45°Poor >45° >45°

principles developed for flexor ten-don rehabilitation whereby protect-ed gliding of tendons decreasesadhesion formation without caus-ing undue stress at the repair site.Although Duran estimated that 3 to5 mm of excursion is required toprevent adhesion formation afterflexor tendon repair, no such crite-rion exists for extensor tendons.Indeed, there is disagreementabout how much extensor tendonexcursion occurs in the noninjuredhand.8,18-21 Dynamic extensorsplinting has also been used inmore distal zones, including zonesIII and IV, with improved resultscompared with injuries treated bystatic splinting.9,10

The relatively poor results withstatic splinting and the enthusiasmdirected toward dynamic splintingled to electrophysiologic investiga-tion of the dynamic splinting tech-nique. Unexpected and inappro-priate extensor musculature activityoccurred during both active flexionand passive extension and at rest.22

The addition of a dorsal blockhood, which holds the MCP jointsin 15 degrees of flexion (Fig. 2, B),halted this inappropriate activity.While inappropriate muscular

activity has not been shown tocause extensor repair rupture withstandard dynamic splinting, it doespoint to a peril peculiar to extensortendon repair.6-8 Many assump-tions based on flexor tendon repair,but without substantiation forextensor tendon repair, have beenmade in the past. These are nowbeing consistently challenged withlaboratory and outcomes analysis.

The clinical success of the newrehabilitation protocols has stimu-lated studies on repair tech-niques.11,18,23,24 Common suturetechniques (Fig. 3) used for repairin zones IV and VI have been eval-uated for their strength, tendencyto shorten the tendon, and poten-tial effect on digital range ofmotion.23,24 It has been shown thatthere may be an important iatro-genic component to fair and poorresults, at least in zone VI, wherecommon suture techniques typical-ly shorten the tendon approximate-ly 6 mm. In a cadaveric model, thisamount of shortening produces an18-degree loss of motion at both theMCP and PIP joints.

These biomechanical studieshave shown that the modifiedBunnell technique produces the

best repair in zone VI, providingthe optimal combination of tendonshortening, resultant MCP and PIPrange of motion, and repairstrength.23 The modified Bunnelland modified Kessler techniquesare equally effective in zone IV.24

However, a newly modifiedBunnell technique and a whipstitchtechnique have recently beenshown to be stronger and moreadvantageous in a cadavericmodel.12 These techniques are sim-ple to use and add approximatelyone third greater strength, asjudged by 2-mm gap formationand load to failure, than the usualmodified Bunnell technique.Clinical testing currently underway shows promising early results.

Other studies have investigatedthe effect of wrist position onextensor tendon excursion; howev-er, there continues to be disagree-ment about the amount of extensorexcursion that occurs, varying from2 mm to as much as 8 mm.8,18-20

Additional research has analyzedmathematically the forces acrossthe extensor system that occur withshort-arc active motion (0 to 30degrees), as is seen during postop-erative rehabilitation.11



Fig. 2 A, Traditional dynamic extension splint. B, Dynamic extension splint modified with dorsal hood to minimize extensor muscleactivity.

A B

Tendon Lacerations

EvaluationA careful evaluation of the

injured hand, including a thoroughneurovascular examination, isessential in the treatment of exten-sor tendon lacerations. Flexor ten-don function should be assessed,and appropriate radiographsobtained. The wound should bethoroughly inspected, with adjunc-tive use of local or regional blockanesthesia as necessary. Injuriesnear a joint must be carefullyinspected for violation of the cap-sule; sterile saline or methyleneblue should be injected into thejoint for verification if any doubtexists.

For lacerations in zones Vthrough VIII, careful testing ofMCP extension should be per-formed with the wrist held in neu-tral and the interphalangeal jointsextended. Extension of the inter-phalangeal joints, produced by theintrinsics, should not be interpret-ed as representing extrinsic exten-sor integrity. The patient should beasked to fully extend each finger atthe MCP joint against gentle resis-

tance. A complete laceration of theextensor tendon will prevent fullMCP extension.

A partial extensor lacerationmay be painful or may demon-strate incomplete MCP extension.Partial lacerations should be direct-ly visualized to determine whetherrepair is necessary. Although nostudies have been performed todetermine the amount of partiallaceration that requires repair, Ibelieve that a tendon lacerated over50% of its width should be repairedto ensure adequate balance and toprevent further disruption.

Injuries proximal to the junc-turae tendinum in zone VI requirespecial attention. A finger mayfully extend by way of a juncturaeven though its extensor tendon iscompletely lacerated. This can beevaluated by asking the patient toextend each finger individuallywhile holding all others flexed atthe MCP joints, thereby blockingthe pull of the juncturae. If theextensor tendon is intact, thepatient will be capable of at leastsome extension, which should becomparable to that of the contralat-eral finger of the noninjured hand.

TreatmentSeverely contaminated wounds,

open fractures, and joint capsulelacerations require emergent andthorough irrigation and debride-ment. Fractures and skin lossshould be treated in the initial pro-cedure when feasible. Fracturesshould be fixed rigidly enough toallow early dynamic splinting oractive motion.

For lacerations without associat-ed injury, the extensor tendon canbe repaired emergently or in adelayed primary fashion after irri-gation, debridement, and loose clo-sure of the wound. If the repair isdelayed, it should be performedwithin 7 days, before the tendonends retract or soften.

All repairs are best performedwith adequate anesthesia, lighting,and exposure. Repair should bedone with 3-0 or 4-0 nonabsorbablesuture material. For repairs inzones V through VIII, the modifiedKessler or Bunnell technique with4-0 nonabsorbable suture is effec-tive.

Injuries to the extensor retinacu-lum in zone VII should be repairedwith suture, taking care to avoidimpingement of the repair on theretinaculum. If impingementwould occur, a portion of the reti-naculum can be resected to allowunhindered tendon excursion.

For injuries in zones III and IV,emergent care should proceed asnecessary for open joints, fractures,or contaminated wounds. A modi-fied Kessler or Bunnell techniquewith 4-0 nonabsorbable suture iseffective if sufficient tendon thick-ness is present. These techniquesare significantly better (P<0.05)than a mattress or figure-of-eightrepair (Fig. 3).24 The lateral bands,if injured, should be repaired sepa-rately with 5-0 or 6-0 suture. Ex-tensor lacerations in these zonesoccur with a high rate of associatedinjuries (80%) and have the poorest



Mattress Figure-of-eight Modified Bunnell Modified Kessler

Fig. 3 Four commonly used repair techniques for extensor tendon lacerations.

outcome.11 Some advocate the useof percutaneous pinning of the PIPjoint in full extension to keep ten-sion off the repair. I perform thisonly when mandated by the boneinjury or when the patient is non-compliant.

The surgical techniques used forrepair in zones I and II are less welldefined than for the more proximalzones. Most often used are a run-ning or mattress technique or thetenodermodesis technique ofMcFarlane and Hampole,25 inwhich the skin and tendon aresutured as one layer. Because theDIP joint is extremely difficult tomaintain in full extension bysplints, pinning the joint in exten-sion best protects the repair andallows monitoring of the skin. ThePIP joint should remain free for fullactivity after the skin and nail bedhave healed.

RehabilitationIn recent years, research has

been directed toward postoperativerehabilitation for extensor injuriesin an effort to improve results.Static splinting has a long history,and the results are well docu-mented.

For injuries in zones V throughVIII, a static splint can be made ofplaster or molded plastic that is fit-ted to the volar aspect of the handand wrist. The fingers next to theinjured digit are also included toprotect against the pull of adjacenttendons through the juncturae.The wrist is placed in approximate-ly 30 degrees of extension, the MCPjoints in approximately 15 degreesof flexion, and the interphalangealjoints in full extension. Splinting ismaintained for 4 to 6 weeks, afterwhich active range of motion isbegun. Passive range-of-motionexercises are begun approximately2 weeks later.

The inherent disadvantages ofimmobilizing a lacerated tendon

are obvious. The tendon becomesadherent to underlying periosteumand overlying skin, producing apotential loss of flexion. Evans andThompson11 have shown that theforce of efforts to overcome adhe-sions can attenuate the repair andresult in increasing extensor lagduring the course of rehabilitation.Immobilized tendons also losestrength over time.11,26 Controlledstress has been shown to combatthis by improving tensile strength,improving gliding properties,increasing repair-site DNA, andaccelerating changes in peritendi-nous vessel density and configura-tion.26

Dynamic extension splintingoffers an alternative for minimizingadhesion problems by allowingseveral millimeters of extensor ten-don gliding without placing unduestress across the repair site. Al-though the optimal amount of ex-tensor excursion that is required tolimit adhesion is not known,dynamic splinting has been usefulin improving outcomes in proximalzones. The splint is applied dorsal-ly 3 to 5 days after injury and holdsthe wrist in 30 degrees of extensionand the MCP joints in 10 to 15degrees of flexion (Fig. 2). Theinterphalangeal joints are held in 0degrees of extension by rubberbands attached to slings. Whilemost reports have dealt with sim-ple injuries, this postoperative pro-tocol can also be beneficial in thetreatment of injuries that involvewell-stabilized fractures or othersoft-tissue damage.

Postoperative treatment is stillevolving for injuries in zones IIIand IV. Static splinting of only theaffected finger can be used.Although some would advocatesplinting of the wrist and MCPjoints for distal injuries, Dagumand Mahoney19 have shown thatlittle or no stress occurs at a zoneIII or IV repair site if the PIP joint

and wrist are splinted and the MCPjoint is left free. Evans andThompson11 recommend that onlythe interphalangeal joints beincluded. If a static rehabilitationprotocol is chosen, an aluminum-and-foam or molded-plastic splintis positioned on the dorsal or volaraspect of the digit, depending onthe soft-tissue injury and the sur-geon’s preference. Both the DIPand PIP joints are held in fullextension. If the wrist is included,it should be splinted in 30 degreesof extension. Static splintingshould be maintained for 4 to 6weeks, after which gentle activemotion is begun. Passive motion isdelayed another 2 to 4 weeks.

Others have advocated dynamicextension splinting for these distalinjuries.9,10 The splint is similar tothat used for more proximalinjuries, but does not include thewrist and need include only theaffected digit. The MCP joint isheld in slight flexion, and the out-rigger mechanism and sling holdthe interphalangeal joints in fullextension. Hung et al9 modify thisapparatus slightly to place theMCP joints in approximately 70degrees of flexion. In theory, thisposition rotates the sagittal bandsdistally, thus decreasing tensionwithin the distal segments of theextensor mechanism and limitingthe pull of the EDC tendon on therepair site. These techniques pro-duced improved outcomes com-pared with static splinting.

Another rehabilitation optionfor injuries in zones III and IV isearly controlled active motion asadvocated by Evans and Thomp-son,11 who demonstrated excellentclinical results. With this method,the finger is splinted in full exten-sion between exercise sessions.The MCP joint is not included inthe splint. During exercise, per-formed four to six times a day, thestatic splint is replaced by a splint



that allows 30 degrees of flexion.Active flexion to the block andactive extension to neutral areperformed several times duringeach session. After 2 weeks, thesplint is modified to allow ap-proximately 40 degrees of flexion.Splinting is discontinued after 6weeks, and full active range-of-motion exercises are begun. Theauthors have calculated that thisshort arc of motion producesapproximately 290 g of forceacross the extensor repair and thatlimiting flexion to 30 degrees pro-vides a safety factor. Corrobo-ration of these calculated forces inzone IV was shown in a laborato-ry study in which the force acrossthe repaired tendon produced byfull flexion was approximately 400 g.24 This compares favorablywith the initial repair strengthwith the modified Bunnell andKessler techniques, which was ap-proximately 2,150 g in a cadaverstudy.24

Static splinting of the DIP jointalone after repair of lacerations inzones I and II is generally recom-mended for 4 to 6 weeks, afterwhich active range-of-motion exer-cises can be begun. A pin transfix-ing the DIP joint is particularlyhelpful in dealing with any soft-tissue component of these injuries.

Gliding is my preferred methodof postoperative rehabilitationwhenever the clinical situationallows. For extensor tendon lacer-ations in zones V through VIII, Ihave used a dynamic splint with adorsal block hood in the past.With the use of stronger suturetechniques, active motion exerciseshave been incorporated into the re-habilitation program with promis-ing early results. For injuries inzones III and IV, I recommendshort-arc active-motion exercises.In zones I and II, I typically keepthe DIP joint pinned for 6 weeksafter tendon repair and allow only

active motion for another 6 weeksafter the pin is removed.

Closed Injuries

Closed injuries to the extensormechanism in zones III throughVIII are relatively rare, especiallywhen compared with open injuriesand mallet injuries. Closed injuriesin zones VI through VIII are almostnonexistent except when associatedwith a systemic disorder, such asrheumatoid arthritis.

While the central tendon is sel-dom injured in closed injuries tothe extensor mechanism in zone V,the sagittal bands that hold theEDC tendon centrally over themetacarpal head can rupture. Thiscan occur from a direct blow to thedorsal aspect of the metacarpalhead or from forced extension orflexion of the MCP joint. Swellingand ecchymosis are minimal. Thepatient usually notes modest painand is unable to initiate MCPextension from a flexed position,but will be able to maintain exten-sion after the finger has been pas-sively extended.

The differential diagnosisshould include extensor tendonrupture, tendon laceration, andradial nerve dysfunction. The longfinger is most often affected,because its radial sagittal fibers areweaker and its joint is the mostexposed to blunt trauma; however,any finger may be affected. Theextensor tendon usually dislocatesulnarly into the intermetacarpalvalley because of the ulnar pull ofthe flexor and extensor tendons,but ulnar sagittal fiber rupturewith radial dislocation has alsobeen described.27

If the injury is detected immedi-ately, a static splint or short-armcast can be applied with the MCPjoints of the affected finger andeach adjacent finger held in 0

degrees of flexion. The PIP jointsare left free, and range of motion isencouraged to promote gliding ofthe lateral bands. Incompleteinjuries can be treated with staticsplinting or with buddy-taping.Rayan and Murray28 recommendconservative splinting after sagittalband injury for up to 6 weeks afterinjury, even with complete dislo-cation of the extensor tendon.However, because the balance ofthe extensor mechanism over theMCP joint is delicate, open repairof complete injuries is generallyrecommended to ensure that theruptured sagittal fibers are appro-priately reapproximated and thetendon is well centered over theMCP joint. Repair can usually beaccomplished by simple reapproxi-mation of the sagittal fibers with 4-0 or 5-0 absorbable suture in amattress fashion. If some delay hasoccurred or the extensor tendonhas a tendency to subluxate afterreapproximation of the sagittalfibers, the opposite (usually ulnar)sagittal fibers should be partiallyreleased to allow accurate central-ization of the tendon.

Closed ruptures of the centralslip overlying the PIP joint canoccur with volar dislocation of thePIP joint, with forced flexion of thePIP joint, or with a severe contu-sion to the dorsum of the PIP. Thejoint dislocation itself is generallyrecognized from a careful historyand radiographs. A central sliprupture may not be recognized inthe immediate postinjury periodbecause of the pain and swellingassociated with the dislocation.17

Deformity may take 2 to 3 weeks tooccur as the untethered lateralbands slide volarly, producing atypical boutonniere posture. Be-cause patients frequently presentwith a presumed dislocation thatthey themselves have reduced, ahigh index of suspicion should bemaintained to identify a concomi-



tant injury to the central slip whenthere is dorsal tenderness andswelling about the PIP joint. Earlysimple treatment can prevent latedisabling deformity.

Detection of a central slip rup-ture is relatively straightforward.After the PIP joint has been relo-cated, extension of the joint is care-fully evaluated. Extension lag,pain with extension, or pain withresisted extension should raise thesuspicion of a central slip rup-ture.17 Full extension may still bepossible if the lateral bands havenot yet moved volarly, but it willbe painful and weak. If the diag-nosis remains unclear, a dilutesolution of arthrographic dye canbe injected into the joint. Extru-sion of dye dorsally into the softtissues indicates a central slip rup-ture along with rupture of the dor-sal capsule.

Central slip rupture is treatedwith immobilization of the PIPjoint in 0 degrees of extension witha dorsal or volar splint for 6 weeks;this encourages DIP joint motion tokeep the lateral bands mobile andgliding. These injuries can be treat-ed successfully up to 6 weeks afterrupture, although serial casting ordynamic splinting may be neces-sary to regain full passive exten-sion before continuous splinting infull extension. If the PIP joint can-not be passively brought into 0degrees of extension initially orafter serial splinting, capsular andvolar plate release may be neces-sary before performing a difficultextensor tendon reconstruction.

Closed rupture of the extensormechanism is quite common inzone I. The well-known mallet fin-ger is usually evident immediatelyafter injury, with an obvious droopof the finger at the DIP joint and alack of active extension. The mech-anism of injury is usually forcedflexion of the fingertip, often fromthe impact of a thrown ball. Pain,

swelling, ecchymosis, and fingerdroop are the usual presentingsigns and symptoms. Radiographsare used to assess the presence of afracture and the degree of jointsubluxation.

Doyle29 has described four typesof mallet injury. Type I is a typicalextensor tendon avulsion from thedistal phalanx. Type II is a lacera-tion of the tendon. Type III is adeep avulsion that injures tendonand skin. Type IV is a fracture ofthe distal phalanx; this type is fur-ther divided into three categories.Type IVA is a transepiphyseal frac-ture in a child. Type IVB involvesless than half the articular surfaceof the joint without joint subluxa-tion; the mechanism of injury isusually hyperflexion. Type IVCinvolves more than one half thearticular surface of the joint andcan involve volar subluxation ofthe joint; this injury may be causedby hyperextension of the joint.

Type I mallet finger is treated bysplinting the DIP joint, holding it in0 degrees of extension with a dor-sal or volar splint. Care must betaken not to hyperextend the jointas this can compromise the bloodsupply to the dorsal skin. I prefer touse a dorsally applied aluminum-and-foam splint, as this leaves thepatient’s touch pad free. Half thethickness of the foam is removed sothat the digit is not pulled intohyperextension. Splinting is con-tinuous for 6 weeks. If there isactive extension and little or nodroop at reevaluation, the splint isapplied at bedtime only for another6 weeks.

Type II injuries should be re-paired as discussed previously.Type III injuries may require skingrafting and other reconstruction.Type IVA fractures should bereduced if necessary and splintedfor 6 weeks. Type IVB fracturesshould also be reduced if necessaryand splinted.

Treatment of type IVC injuries isstill controversial. Many are re-ducible and can be treated conserv-atively. Wehbé and Schneider30

have shown that these injuries,even if associated with volar sub-luxation, can be treated with splint-ing alone, with results at least com-parable to those for injuries treatedwith open reduction and internalfixation. The difficulty of ap-proaching this area surgically isevident, as the skin has a tenuousblood supply, and the fracture frag-ment is quite small. Anatomicrestoration is often difficult, andcomplications are frequent.30

For type IVC injuries in whichthe volar subluxation of the joint isunacceptable after attempted re-duction and appropriate splinting,a more conservative option is toreduce the joint and pin it percuta-neously without direct exposure ofthe fracture. A Kirschner wire inthe distal fragment can serve as ajoystick to maneuver the fragmentinto position. The wire is then ad-vanced through the proximal frag-ment (if it is large enough) andacross the DIP joint. This allowsnear-anatomic restoration of jointand fracture and avoids the poten-tial complications of internal fixa-tion. Nevertheless, complications,such as articular incongruence,wire breakage, and infection, canstill occur.31

Summary

While extensor injuries often resultin loss of extension, this is seldomdisabling. However, these injuriesfrequently result in loss of flexion,which is greater in severity and fre-quency than loss of extension. As aconsequence, extensor tendoninjuries deserve careful attention indiagnosis, treatment, and rehabili-tation to offset potential loss offunction.



References

1. Dargan EL: Management of extensortendon injuries of the hand. SurgGynecol Obstet 1969;128:1269-1273.

2. Flynn JE: Problems with trauma to thehand. J Bone Joint Surg Am 1953;35:132-140.

3. Hauge MF: The results of tendonsuture of the hand: A review of 500patients. Acta Orthop Scand 1954;24:258-270.

4. Kelly AP Jr: Primary tendon repairs:A study of 789 consecutive tendonseverances. J Bone Joint Surg Am 1959;41:581-598.

5. Newport ML, Blair WF, Steyers CM Jr:Long-term results of extensor tendonrepair. J Hand Surg [Am] 1990;15:961-966.

6. Browne EZ Jr, Ribik CA: Early dy-namic splinting for extensor tendon in-juries. J Hand Surg [Am] 1989;14:72-76.

7. Chow JA, Dovelle S, Thomes LJ, et al:A comparison of results of extensortendon repair followed by early con-trolled mobilisation versus staticimmobilisation. J Hand Surg [Br] 1989;14:18-20.

8. Evans RB, Burkhalter WE: A study ofthe dynamic anatomy of extensor ten-dons and implications for treatment. J Hand Surg [Am] 1986;11:774-779.

9. Hung LK, Chan A, Chang J, et al:Early controlled active mobilizationwith dynamic splintage for treatmentof extensor tendon injuries. J HandSurg [Am] 1990;15:251-257.

10. Saldana MJ, Choban S, Westerbeck P,et al: Results of acute zone III extensortendon injuries treated with dynamicextension splinting. J Hand Surg [Am]1991;16:1145-1150.

11. Evans RB, Thompson DE: An analysisof factors that support early active

short arc motion of the repaired cen-tral slip. J Hand Ther 1992;5:187-201.

12. Newport ML, Waters SN: A newsuture technique for extensor tendonrepair. Presented at the 50th AnnualMeeting of the American Society forSurgery of the Hand, San Francisco,1995.

13. Kleinert HE, Verdan C: Report of theCommittee on Tendon Injuries. J HandSurg [Am] 1983;8(5 pt 2):794-798.

14. Manske PR, Lesker PA: Nutrientpathways to extensor tendons withinthe extensor retinacular compart-ments: An experimental study in dogs.Clin Orthop 1983;181:234-237.

15. Smith JW, Bellinger CG: The bloodsupply of tendons, in Tubiana R (ed):The Hand. Philadelphia: WB Saunders,1991, vol 1, pp 353-358.

16. Miller H: Repair of severed tendons ofthe hand and wrist: Statistical analysisof 300 cases. Surg Gynecol Obstet 1942;75:693-698.

17. Lovett WL, McCalla MA: Manage-ment and rehabilitation of extensortendon injuries. Orthop Clin North Am1983;14:811-826.

18. Minamikawa Y, Peimer CA, Yama-guchi T, et al: Wrist position and ex-tensor tendon amplitude followingrepair. J Hand Surg [Am] 1992;17:268-271.

19. Dagum AB, Mahoney JL: Effect ofwrist position on extensor mechanismafter disruption separation. J HandSurg [Am] 1994;19:584-589.

20. DeVoll JR, Saldana MJ: Excursion offinger extensor tendon elements inzone III. Orthop Trans 1989;13:220.

21. Duran RJ, Houser RG, Coleman CR, etal: Management of flexor tendon lac-erations in zone 2 using controlled

passive motion postoperatively, inHunter JM, Schneider LH, Mackin EJ,et al (eds): Rehabilitation of the Hand,2nd ed. St Louis: CV Mosby, 1984, pp273-276.

22. Newport ML, Shukla A: Electro-physiologic basis of dynamic extensorsplinting. J Hand Surg [Am] 1992;17:272-277.

23. Newport ML, Williams CD: Bio-mechanical characteristics of extensortendon suture techniques. J Hand Surg[Am] 1992;17:1117-1123.

24. Newport ML, Pollack GR, WilliamsCD: Biomechanical characteristics ofsuture techniques in extensor zone IV.J Hand Surg [Am] 1995;20:650-656.

25. McFarlane RM, Hampole MK: Treat-ment of extensor tendon injuries of thehand. Can J Surg 1973;16:366-375.

26. Feehan LM, Beauchene JG: Early ten-sile properties of healing chicken flex-or tendons: Early controlled passivemotion versus postoperative immobi-lization. J Hand Surg [Am] 1990;15:63-68.

27. Kettelkamp DB, Flatt AE, Moulds R:Traumatic dislocation of the long-finger extensor tendon: A clinical,anatomical, and biomechanical study.J Bone Joint Surg Am 1971;53:229-240.

28. Rayan GM, Murray D: Classificationand treatment of closed sagittal bandinjuries. J Hand Surg [Am] 1994;19:590-594.

29. Doyle JR: Extensor tendons: Acuteinjuries, in Greene DP, Hotchkiss RN(eds): Operative Hand Surgery, 3rd ed.New York: Churchill Livingston,1993, vol 2, pp 1925-1954.

30. Wehbé MA, Schneider LH: Malletfractures. J Bone Joint Surg Am 1984;6:658-669.



http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=4889425&Dopt=r





















Today’s orthopaedist must have abasic understanding of the causes ofback pain in children and adoles-cents in order to either evaluate andtreat the problem or refer the pa-tient appropriately. This reviewprovides a detailed discussion ofsome of the more common condi-tions associated with back pain,with special attention to incidence,history, physical findings, and diag-nostic tools, as well as a brief discus-sion of treatment modalities cur-rently recommended. Armed withthis information, the clinicianshould be able to construct an effi-cient and economical diagnostic andtreatment approach. More impor-tant, the orthopaedist should also beable to avoid missed or delayeddiagnoses of serious problems.

Epidemiology

Although the prevalence of lowback pain in children and adoles-

cents is far less than in the adultpopulation, as many as 36% ofschool-age children report experi-encing low back pain, and 7% seekmedical attention.1 It is importantto evaluate all patients carefully, asin many cases an organic cause fortheir pain will be found.2

History

A thorough history is often hard toobtain in pediatric patients. Thepresence of the parents or guardianis important when patients are tooyoung to give an accurate history.The physician should seek to deter-mine the nature of the onset of pain.Was the pain acute, with a clear his-tory of antecedent trauma, as seenin disk herniations and apophysealring fractures, or was it insidious,as in Scheuermann’s kyphosis andmalignant conditions? A history ofpain radiating to the buttock or pos-terior aspect of the thigh is more

characteristic of a lumbar herniatednucleus pulposus (HNP) than isradiation of pain below the knees,which is indicative of an epiduralabscess or an intraspinal tumor.3,4

Duration of pain greater than 4weeks is often a sign of a potentiallyserious problem.

The presence of neurologic signsand symptoms is very unusual forbenign conditions in children. Thephysician must ask about gaitabnormalities (e.g., a stiff-leggedgait may indicate spondylolisthe-sis), trunk list (a sign of HNP), orfoot deformities (which may signifyintraspinal anomalies or tetheredcord). Bowel or bladder dysfunc-tion should be noted. A history ofunexplained fever, weight loss, ormalaise should raise the suspicionof a systemic malignant condition,and recent bacterial or viral infec-tions may suggest a diagnosis ofdiskitis.

Dr. Ginsburg is Assistant Professor ofOrthopaedics, University of Nebraska MedicalCenter, Omaha. Dr. Bassett is AssociateProfessor of Orthopaedic Surgery and Chief ofPediatric Orthopaedics, Washington Univer-sity School of Medicine, St Louis.

Reprint requests: Dr. Ginsburg, University ofNebraska, 600 South 42nd Street, PO Box981080, Omaha, NE 68198-1080.


Abstract

Back pain in children and adolescents usually has a recognizable organic origin.The most common entities seen are spondylolysis, spondylolisthesis,Scheuermann’s kyphosis, disk herniations, infections, and tumors. Early recog-nition and treatment can provide patients the best chance at relief of symptomsand eradication of the underlying disease process. The goals of this review areto (1) familiarize the clinician with the various diagnoses associated with backpain in the skeletally immature patient and (2) to assist the clinician in makingthe appropriate diagnosis by providing a rational method of selecting diagnostictests that maximize specificity and minimize costs.


Back Pain in Children and Adolescents:Evaluation and Differential Diagnosis

Glen M. Ginsburg, MD, and George S. Bassett, MD

The physician should inquireabout aggravating or alleviatingfactors. Is the pain exacerbated bycertain activities (most typical ofScheuermann’s kyphosis, HNP,spondylolysis, and spondylolisthe-sis)? Is the pain worse at night orwhen the patient is supine or it isunrelieved by rest (suggestive ofmalignant neoplasm)? Is the painmore intense when the patient isprone (possibly a sign of epiduralabscess)? Is the pain relieved bynonsteroidal anti-inflammatorydrugs (NSAIDs), as may occur withosteoid osteoma? Has the patientbeen involved in any activities,such as football (as a lineman),gymnastics, or swimming, thathave been associated with an in-creased risk of spondylolysis? Hasthe patient been doing heavy farmwork, shoveling, or weightlifting(which may suggest Scheuer-mann’s kyphosis)? Has he or sherecently started a new sport oractivity or had to stop a previouslyenjoyed activity because of pain?Finally, has there been recent famil-ial unrest, increasing school diffi-culty, or a history of litigation? Theresponses to these few questions,combined with a complete physicalexamination, should provide theinterviewer with an accurate differ-ential diagnosis.

Physical Examination

The physical examination should beconducted with the patient com-pletely disrobed except for anexamination gown and underwear.The socks should be removed toexamine for foot abnormalities.The patient’s general habitus andaffect should be observed first. Isthere appropriate eye contactbetween the patient and the exam-iner? Cachexia or pallor should benoted as a potential sign of anunderlying malignant or nutritional

disorder. The skin should be exam-ined for the presence of heman-giomata, hair patches, or midlinedefects suggestive of an intraspinaldisorder. Gait should be assessed,with notation of standing posture,balance, the Trendelenburg sign,and pelvic obliquity. The presenceof a limp or ataxia may be the firstsign of an intraspinal disorder.

The spine should be examined forscoliosis and kyphosis. An L5-S1stepoff or heart-shaped buttocksmay be the clue to underlyingspondylolisthesis. The Adams for-ward-bending test should be used todocument any reversal of spinalrhythm, trunk deviation, rib or lum-bar prominence, or limitation of for-ward flexion, which is a sign of ham-string tightness. The back should bepalpated for posterior-element,paraspinal-muscle, or costovertebral-angle tenderness. The straight-leg-raising test will reveal HNP andapophyseal fractures if present andcan also be used to evaluate ham-string tightness. Also essential is acomplete neurologic examination,including motor and sensory func-tion, deep tendon and abdominalreflexes, and long-tract signs, such asthe Babinski and clonus tests.

Diagnostic Studies

The diagnostic tests used to evalu-ate the causes of back pain arebased on the level of severity ofsymptoms and physical findings.Effective and cost-efficient guide-lines for the use of diagnostic stud-ies have been proposed by Wenger5

(Table 1).

Differential Diagnosis

Although no disease entity isabsolutely specific to any one agegroup, some generalizations can bemade to help arrive at the most

likely diagnoses. Diskitis and ver-tebral osteomyelitis are more com-mon in children under the age of 10years. Neoplastic conditions in thisage group include eosinophilicgranuloma, leukemia, neuroblas-toma, and astrocytoma. Idiopathicjuvenile osteoporosis is also morecommon in this age group.

In patients over the age of 10,disorders involving repetitive load-ing or trauma, such as spondyloly-sis, spondylolisthesis, Scheuer-mann’s kyphosis, herniated lumbardisk, and apophyseal ring fracture,occur more frequently. Tumorsmore common to this age group areosteoblastoma, osteosarcoma, andlymphoma. Other tumors, such asosteoid osteoma and aneurysmalbone cysts, are not age-specific.Psychogenic back pain, conversionreactions, and overuse syndromesmay be seen in the preadolescentand teenager.

Spondylolysis andSpondylolisthesis

“Spondylolysis” is the term used todescribe a defect in the pars interar-ticularis, usually affecting the lum-bar spine. If defects are bilateral atthe same level, the upper vertebralsegment may slip forward on thelower segment, which is referred toas “spondylolisthesis.” This is oneof the most common causes of backpain in the adolescent (47% inci-dence in adolescent athletes com-plaining of back pain6). The overallprevalence has been reported torange from 4.4% in children aged 6to 6% in adults.7 Spondylolysis israrely symptomatic before the ado-lescent growth spurt.

The etiology of spondylolysisdepends on the type of defect seen.The isthmic defect, the most com-mon in this age group, has beenattributed to a stress fracture of thepars from repeated hyperextension,

Back Pain in Children and Adolescents


causing shear of the posterior ele-ments. The fifth lumbar vertebra isaffected most often, followed by thefourth and the third. There is anincreased incidence in participantsin sports associated with repetitiveflexion-extension activity, such asgymnastics.8 There is a strong association with thoracolumbarScheuermann’s disease, with up toa 50% coassociation with spondy-lolysis, thought to be secondary tothe compensatory hyperlordosis ofthe lumbar spine.9 The dysplastictype is much less common and iscaused by a congenital deficiency ofthe inferior facets of L5 and/or thesuperior facets of S1 and elongationof the pars, allowing forward trans-lation of L5 on S1.

The back pain in spondylolysisor spondylolisthesis is usually lum-bar and mild to moderate in severity.The pain may radiate to the poste-rior aspect of the thighs and but-

tocks, is aggravated by repetitiveflexion/extension maneuvers, andis relieved by rest. The classic gaitin this condition is stiff-legged,with a short stride length due totight hamstrings, often referred toas the “pelvic waddle.” Hamstringtightness also limits forward bend-ing. There may be pain on palpa-tion of the paraspinal muscles. Ifthe slip is severe, a stepoff can befelt at L5-S1. The buttocks have aheart-shaped configuration owingto the vertical position of thesacrum, and the abdomen is pro-truding, with transverse abdominalcreases.

The anteroposterior and lateralradiographs will establish the diag-nosis in most cases of spondylo-lysis when the defect is bilateral.Oblique films of the lumbosacralspine will usually demonstrate thepars abnormality, depicted as thecollar of the “Scotty dog” sign in

isthmic defects and the “greyhoundsign” indicating an elongated parsin dysplastic spondylolysis. Thispars defect is unilateral in 20% ofcases and occurs at more than onelevel in 4%.10

If standard radiographs are non-diagnostic, a technetium bone scanwill delineate an acute lesion within5 to 7 days of the onset of symp-toms. The bone scan can also beused to assess the healing activityof established lesions. In cases ofunilateral defects, it may demon-strate increased uptake in the con-tralateral pars, indicative of a stressreaction.

Single-photon emission comput-ed tomography (SPECT) has re-cently been useful in cases in whichpositive factors and symptoms(e.g., athletic patient, persistentback pain, tight hamstrings) existin the absence of radiographic andscintigraphic findings. This studyhas been shown to be the most sen-sitive method of diagnosing stressreaction or spondylolysis, makingpossible early intervention to pre-vent a stress reaction from becom-ing established lysis.11

The treatment of spondylolysisand type I spondylolisthesis (<25%slip) associated with a history ofrecent injury and short duration ofsymptoms involves restriction of theaggravating activity and a muscle-strengthening regimen for the backand abdomen. Healing can bemonitored by the resolution ofback pain and hamstring tightness.If this does not occur and the bonescan shows increased uptake in thearea of the fracture, a program ofbed rest, NSAID therapy, and im-mobilization in a thoracolumbo-sacral orthosis or cast should re-lieve symptoms.

A 73% rate of healing of early de-fects has been shown with the use oflumbosacral support only.10 Withan established symptomatic spondy-lolysis without spondylolisthesis



Permission to electronically reproduce this table was not available.

and a “cold” bone scan, primaryrepair of small defects (<7 mm) withinternal fixation across the parsdefect and bone grafting has beensuccessful in relieving back pain atlevels proximal to the lumbosacraljunction (Fig. 1).

Asymptomatic patients withslips ranging from 25% to 50%should be followed up with spotlateral radiographs until the end ofgrowth12 and cautioned that partic-ipation in contact sports may haveto be restricted if progression oc-curs. Although Seitsalo et al13

showed that 90% of slip progres-sion had already occurred at thetime of the first radiograph, mildslips in preadolescent patients mayprogress during the adolescentgrowth spurt and should be fol-lowed closely. These authors alsoshowed that the only radiologicfinding predictive of progressionwas a slip percentage of more than20% at initial presentation.

In situ posterolateral fusionwithout instrumentation is the pro-cedure of choice for patients whohave progressive slips, persistentback pain, neurologic deficits, orspondylolisthesis greater than 50%.Fusion from L5 to S1 is recom-mended for symptomatic slips lessthan 50%; fusion from L4 to S1, forslips greater than 50%. After suc-cessful arthrodesis, pain and ham-string spasm are reduced in 90% ofpatients.

Reduction of the slip remains acontroversial topic. Currently, anindication for reduction and fusionwould be a sagittal imbalance thatis functionally debilitating, a high-grade slip that is cosmetically unac-ceptable, or a neurologic deficitnecessitating laminectomy thatwould jeopardize the success offusion by accelerating translation.Reduction can be achieved in aclosed fashion by pelvic extensionfollowed by spica pantaloon cast-

ing after in situ fusion. This methodreduces the lumbosacral kyphosisand the risk of subsequent slip pro-gression.14

Reduction with modern instru-mentation techniques may be thefrontier of spondylolisthesis sur-gery. Transpedicular fixation al-lows distraction, sacral rotation,and posterior translation of thelumbar spine, but has been associ-ated with a higher frequency oflumbar nerve-root injuries thanclosed techniques. As experiencewith these techniques increases,complication rates should decline.

Scheuermann’s Kyphosis

Scheuermann’s kyphosis is a well-described disorder of endochon-dral ossification that affects the ver-tebral endplates and ring apophy-ses and results in intravertebraldisk herniation, anterior wedging



Fig. 1 A, Preoperative lateral radiograph of patient with bilateral spondylolysis at L4. Defect measures 3 mm. B, Postoperative lateralradiograph after internal fixation of bilateral defect and bone grafting. C, Postoperative anteroposterior radiograph.

A B C

of consecutive vertebrae, and afixed thoracolumbar kyphosis.This disorder may account for asmany as a third of all instances ofback pain as a presenting symptomin pediatric patients.2 The preva-lence has been estimated to be 0.4%to 8% of the general population.The condition mainly affects ado-lescents at puberty, with an equalmale-female distribution.

A familial predilection has beentheorized. Among the many causesthat have been proposed are disor-dered endochondral ossification ofthe endplates, vertebral osteoporo-sis, and avascular necrosis of thering apophysis. Increased heightand repetitive loading may be incit-ing factors.

The pain in Scheuermann’skyphosis generally occurs at theapex of the curve and is usuallyaggravated by prolonged sitting,standing, and activity. Poor pos-ture may be the reason for the ini-tial presentation; this disorder mustbe differentiated from posturalround back, which may also beassociated with back pain. Patientswith Scheuermann’s kyphosis willhave a sharp kyphotic deformity inthe thoracic spine that is exacerbatedby forward bending. The kyphosdoes not flatten when the patientlies supine or performs hyperexten-sion maneuvers, as it does withpostural round back. The patientwith Scheuermann’s kyphosis isoften well muscled and may havehamstring contractures.

Pain at the thoracolumbar junc-tion has been associated with lum-bar Scheuermann’s disease, whichaffects mainly young males who doheavy labor or are competitive ath-letes. The onset often occurs afterlifting heavy weights from a flexedposition. There may be no clinicaldeformity in lumbar Scheuer-mann’s disease.

The diagnosis of Scheuermann’skyphosis is confirmed by radio-

graphs (anteroposterior, lateral,and supine hyperextension lateralfilms) that show thoracic kyphosisgreater than 45 degrees, 5 degreesof anterior wedging of the threeadjacent vertebrae at the apex ofthe kyphos,15 Schmorl’s nodes, andendplate irregularities. In lumbarScheuermann’s disease involvingthe T10–L5 vertebrae, additionalfindings may include narrowing ofthe intervertebral disks, andanteroinferior or anterosuperior“scoop” defects of the vertebralbodies (Fig. 2). Bone scans are usu-ally not needed to diagnose tho-racic Scheuermann’s disease, exceptto rule out spondylolysis or diskitisas the cause of pain. In lumbarScheuermann’s disease, a SPECTstudy may show increased uptakeat one or two levels. This imagingmodality has been shown to bemore sensitive than bone scanningin evaluating lumbar disease.16

The treatment of postural roundback and mild kyphosis (<50 de-grees) involves thoracic-extensionand abdominal-strengthening exer-cises. For a patient with true Scheuer-mann’s kyphosis, an initial trial ofexercise, stretching, and cessationof heavy lifting should be imple-mented. The back pain in Scheuer-mann’s kyphosis usually resolvesby the end of growth.

If the kyphotic curve measuresmore than 70 degrees and thepatient is skeletally immature, athoracolumbosacral orthosis forlumbar disease or a Milwaukeebrace for thoracic disease shouldbe prescribed. Improvement inwedging has been shown after 18months of use. The orthosis isworn until skeletal maturity.According to a recent natural his-tory study of Scheuermann’skyphosis, patients treated conserv-atively may have more pain at theapex of the curve as adults, buttheir quality of life should not beaffected.17

Surgery may be indicated if thecurve is greater than 75 degreesand the patient is symptomatic.Restrictive lung disease is usuallynot seen with kyphosis of less than100 degrees. If the curve is rigidwith marked anterior wedging, thetreatment of choice is anteriordiskectomy and arthrodesis of theapical segment, followed by poste-rior fusion of the entire kyphoticsegment with instrumentation.

Diskitis or VertebralOsteomyelitis

In the past, a distinction was madebetween these two entities thatusually affect children under theage of 10 years. However, it may bemore accurate to think of diskitisand vertebral osteomyelitis as rep-resenting a temporal progression of



Fig. 2 Lateral radiograph of an adolescentboy shows typical findings associated withlumbar Scheuermann’s kyphosis, includinganterior wedging, disk-space narrowing,Schmorl’s nodes, and anterior scoopdefects (arrow).

the same disease process, that is, abacterial or viral infection involv-ing the vertebral body and spread-ing to the adjacent disk.18 In thechild, the blood supply traversesthe vertebral endplate from body todisk, establishing a plausible routefor the infection to travel. Thecause of the infection is probablybacterial, but milder cases of diski-tis may be viral.

Back pain is a common sign ofspine infection in the child, alongwith fever, irritability, anorexia, andmalaise. In the younger child (aged1 to 3 years), there may be a sudden,unexplained failure to walk, whichis seen in as many as a third ofpatients.18 In the older child (aged 3to 8 years), there may be associatedabdominal pain if the infectioninvolves the T8–L1 levels. Inteenagers, localized back pain,which may radiate to the buttocksand legs, is more common. Onphysical examination, there may be

generalized signs of irritability, lossof lumbar lordosis due to tight ham-strings, loss of spinal rhythm, a pos-itive straight-leg-raising test, andtenderness over the lumbar spine.

Radiographs of the spine areoften negative if the symptoms areof less than 3 weeks’ duration.Changes most often seen on radio-graphs include decreased disk-spaceheight at the involved levels anderosion or sclerosis of the endplate(Fig. 3, A and B). A complete bloodcell count with differential and anerythrocyte sedimentation rate(ESR) should be obtained if diskitisor osteomyelitis is suspected. Anelevated ESR has been noted in over90% of patients with establishedpyogenic infection of the spine. Thewhite blood cell count has been lessreliable; in the study by Ring et al,18

40% of patients with spine infectionshad a high-normal count, and only10% had an abnormal count. In astudy by Wenger et al,19 biopsy

specimens and blood cultures takenfrom patients with suspected bacter-ial spine infections were positive in67% and 41% of cases, respectively;Staphylococcus aureus was the mostcommon isolate.

In the presence of nondiagnosticradiographs and an increased ESR,a bone scan should be obtained.Scintigraphic findings of increaseduptake are seen in the adjacent ver-tebral endplates (Fig. 3, C). Mag-netic resonance (MR) imaging is theradiologic modality of choice fordifferentiating diskitis or osteo-myelitis from spinal epiduralabscesses. Patients with epiduralabscesses will have back pain andfever, but may also have markedroot symptoms, muscle weakness,and decreased reflexes.4 An MRimaging study can also be useful indistinguishing pyogenic osteo-myelitis from tuberculosis, whichwill rarely cause signal changes inthe disk.



Fig. 3 A, Lateral radiograph of the lumbar spine of a 2-year-old boy with low back pain for 1 week shows mild disk-space narrowing. B,Radiograph obtained 1 month later shows marked disk-space narrowing and endplate changes. C, Bone scan shows increased uptake inthe adjacent endplates of L2 and L3.

A B C

2

32

3

The treatment of diskitis inyoung children is controversial.Good results have been reportedwith immobilization alone withoutantibiotic therapy. However, con-current administration of parenteralantibiotics leads to more rapid reso-lution of symptoms and diminishesthe risk of long-term sequelae.Ring et al18 advocate the use of aparenteral antibiotic, such as naf-cillin or cefazolin, for 6 days fol-lowed by 4 to 6 weeks of oralantibiotics on an outpatient basis.“Excellent” relief of symptoms wasreported in 2 to 4 days by 73% oftheir patients and within 2 to 3weeks by 82%. The ESR can beused to monitor the effectiveness oftreatment; early termination ofantibiotic therapy can be consid-ered if the ESR returns to normal.

Lumbar Disk Herniation

Herniation of a lumbar disk is arare occurrence in children andadolescents; only 0.5% of all diskec-tomies are performed in patientsunder the age of 16. The peak inci-dence in the pediatric population isat 10 to 18 years, with a slight malepreponderance. Unlike lumbar diskherniations in the adult population,most of which are due to degenera-tive changes, herniations in adoles-cents are usually caused by trauma.In one report, the acute onset ofsymptoms was precipitated by atraumatic incident in 50% of ado-lescent patients.20 When an acutetraumatic event cannot be identi-fied, repeated microtrauma fromvigorous activity must be consid-ered. The herniated disk fragmentis usually composed of healthy,well-hydrated tissue and thustends to be larger than a herniatedfragment in an adult. A congenitaltransitional vertebra at L5 or S1 isalso considered to jeopardize thedisk.

Symptoms may be present forseveral months to a year before thepatient seeks medical treatment.Back pain with sciatica is the mostcommon complaint. Atypical pre-sentations may include pain in theposterior aspect of the thigh orknee or a “pulled” hamstring. Theonset may be acute or insidious.Coughing or sneezing will exacer-bate the pain in a third of patients,and the pain will be worse withstrenuous activity. The patientmay demonstrate an abnormal gaitbecause of lumbar spasm. Themost common physical finding inpatients with a herniated lumbardisk is the positive straight-leg-raising test (seen at <60 degrees),which will be present in 85% ofpatients. Paravertebral spasms arealso common. Neurologic findingswere seen in fewer than 40% ofpatients with HNP in a study byDeLuca et al21; these usually con-sisted of decreased reflexes andmild weakness, most commonly inthe great toe extensor. Sensorychanges and bowel and bladderdysfunction are rare.

Radiographs will be normal inabout half of patients with HNP.In the study by Grobler et al,22 themost common radiographic abnor-mality was nonstructural scoliosis,which was seen in 72% of cases.They found asymmetric facet orien-tation at the level of the herniationin 71% of their patients.

Magnetic resonance imaging isthe study of choice for evaluatingdisk herniation. It provides excel-lent visualization of disk materialand nerve roots and can differenti-ate epidural abscess, spinal cordtumors, and abnormalities of theconus or cauda. No myelographiccontrast medium is needed. A dis-advantage is that sedation may berequired for the very young orclaustrophobic patient.

Various nonoperative modalitieshave been used, including bed rest,

NSAIDs, muscle relaxants, andphysical therapy. Surgery is rec-ommended if there are persistentsymptoms or a progressive neuro-logic deficit or if repeated episodeshave not responded to conserva-tive care. In a report of surgicalintervention for herniated disks,DeLuca et al21 noted failure of con-servative treatment in 75% ofpatients. In that series, excellent orgood results were obtained in 91%of patients who underwent diskexcision through laminotomy. Theauthors stressed that confirmationof the appropriate level is essentialbecause of the common presence ofa transitional or extra lumbar verte-bra. Resolution of scoliosis andstraight-leg-raising symptoms, aswell as improvement in back painand mobility, can be expected.Deep tendon reflex improvementusually lags behind.

Apophyseal Ring Fracture

Apophyseal ring fracture is theresult of a fracture through thejunction between the vertebralbody and the cartilaginous ringapophysis that occurs before com-plete fusion at approximately 18years. When a fracture occurs, theapophysis is displaced posteriorlywith the disk, attached by Shar-pey’s fibers. The prevalence ofthis fracture is similar to that ofdisk herniation. Most patients areteenaged boys involved in sportsor heavy lifting. Injury resultsfrom either acute trauma, such ashyperflexion of the loaded spine,or repetitive microtrauma. Takataet al23 identified 31 patients in a 3-year period, half of whom couldnot recall an inciting event. Rec-ognition of this condition ap-pears to have increased commen-surate with development of thetechnology of computed tomog-raphy (CT).



Back pain, usually bilateral, isthe hallmark of this injury. Thepain is associated with sciatica, butrarely radiates below the knee. Thepain, described as constant andburning, is aggravated by theValsalva maneuver. Patients usu-ally complain of back stiffness.Neurologic symptoms and bowelor bladder involvement are rare.The physical findings are similar to those of herniated disk exceptthat the contralateral straight-leg-raising test is usually positive andelicited at less than 40 degrees offlexion.24

Lateral radiographs may showa small triangular opacity poste-rior to the vertebral body. Theinferior apophysis of L4 is mostcommonly involved.23 A CT studywith metrizamide myelography isthe procedure of choice, providingexcellent bone visualization anddocumentation of anterior epiduralcompression. An MR imagingstudy does not distinguish corticalbone as well as a CT study, mak-ing differentiation of bone fromdisk material difficult. Most asso-ciated herniated disks are central(Fig. 4).24

The treatment of a slipped verte-bral apophysis without neurologicsymptoms is similar to the nonop-erative treatment of herniateddisks. If there is neurologic com-promise, excision of the bone frag-ment with attached cartilage anddisk is the treatment of choice. Abilateral laminotomy is recom-mended to safely remove the entirecentral fragment. Complete lami-nectomy may destabilize the spineand is not recommended.

Tumors

Spinal tumors are rare in all agegroups. Primary osseous neo-plasms of the spine accounted foronly 29 (1.5%) of 1,971 musculo-

skeletal neoplasms in a study byDelamarter et al.25 Of these, only 8(0.4%) were in children or adoles-cents. In 7 of 8 cases, back painwas the presenting symptom.

Most pediatric spinal neo-plasms are benign; osteoid osteo-ma and osteoblastoma are themost common types.25 Patientsusually present with a history ofback pain that is worse at nightand is relieved by NSAIDs. Pa-tients typically have a painfulnonstructural scoliosis with asym-metric forward bending and ten-derness to palpation over theparaspinal muscles. Althoughrare, neurologic symptoms due tocompression can be seen withlesions in the cervical or thoracicspine.

The most common radiographicfindings are sclerosis of the poste-rior elements with varying degreesof expansion. By definition, osteo-blastomas are greater than 2 cm indiameter. A bone scan is frequentlynecessary for localization (Fig. 5, A);

further localization of the nidus can be provided by CT scanning(Fig. 5, B).

Although Kneisl and Simon26

reported that relief of symptomswas achieved with an average of 33months of NSAID therapy, thetreatment of choice is excision ofthe nidus. Preoperative adminis-tration of radioactive tracer andintraoperative use of a radiosensi-tive bone probe can aid in localiza-tion and complete excision anddiminish the amount of uninvolvedbone excised. Drilling with CTguidance is another treatmentoption with proven success forlesions in the posterior elements ofthe spine in children.27

Eosinophilic granuloma (Langer-hans cell histiocytosis) of thespine usually affects younger chil-dren, the average age being 4.5years.28 Boys and girls are affect-ed equally. Nonspecific back painis the usual presenting symptom.Nonstructural scoliosis or torticol-lis may be seen. On plain radio-graphs, the classic finding of ver-tebra plana marks vertebral col-lapse. If one suspects eosinophilicgranuloma, a skeletal survey anda bone scan must be obtained tolook for other osseous involve-ment. Iliac-crest marrow aspira-tion is usually indicated in newcases to rule out systemic forms ofthe disease. A needle biopsy ofthe lesion may be necessary torule out a malignant neoplasm.Robert et al28 recommend conserv-ative treatment of these lesions,the majority of which show recon-stitution of vertebral body heightby endplate growth in the youngchild. The use of a spinal orthosismay be required to alleviatesymptoms. Low-dose radiother-apy is recommended for patientswho present with a neurologicdeficit. If radiotherapy fails, sur-gical decompression may be re-quired.



Fig. 4 Axial CT scan shows a slippedlumbar vertebral apophysis. Large centralosseous lesion impinges on canal.

Aneurysmal bone cysts consti-tute 10% of benign osseous le-sions.29 Most occur in personsunder the age of 20 years. As manyas 20% involve the spine, mostcommonly in the posterior ele-ments. Symptoms of pain maydevelop as a consequence of thelesion itself or may be due to apathologic fracture. Radiographsand CT scans show an expansilelesion with thinning of the sur-rounding cortex (Fig. 6). Intra-lesional resection and bone graftingis usually curative, with a recur-rence rate of up to 25%.29

Malignant tumors are very rarein the pediatric population. Backpain is a common presenting symp-tom. Suspicion should be high if achild under age 4 has back pain atnight when supine that is unre-sponsive to conservative therapy.Neurologic symptoms due tospinal cord compromise are com-mon with malignant lesions of thespine, which were present in 70%of patients in the series of Shives et al.30

Although osteosarcoma is thesecond most common primarymalignant neoplasm of bone, few

osteosarcomas (5%) involve thespine.30 Patients are usually in thesecond decade, and involvement ofthe lower vertebral bodies is mostcommon. Osteolytic, osteoblastic,or mixed lesions may be notedradiographically. Both CT scansand MR images are necessary forstaging. Osteosarcomas of thespine are difficult to treat becauseof the adjacent structures. Widesurgical resection is recommended,if possible. Adjuvant chemotherapyand local radiation therapy mayincrease long-term survival to45%,30 but the overall prognosis ispoor.

Ewing’s sarcoma occurs mostfrequently in the 5- to 15-year agegroup. Spinal lesions are seen in3.5% of all cases of Ewing’s sarcoma;the sacrum is the usual site.31 Mostpatients will have back pain, butonly 25% will have fever or a pal-pable mass. Radiographs show anexpansile lesion and at later stagesmay show vertebral collapse, simi-lar to what is seen in eosinophilicgranuloma. An MR imaging studyis recommended to visualize theentire lesion, including soft-tissueextension. Multiagent chemother-

apy has improved the 5-year sur-vival rate in these patients, anddecompressive laminectomy hasbeen shown to improve neurologicsymptoms in two thirds of pa-tients.31 Resection of the lesionprobably improves survival, butthere are no long-term studies todate.

Acute leukemia is the mostcommon malignant condition inthe pediatric population, generallyaffecting children under the age of10 years. Acute lymphoblasticleukemia is the most prevalenttype of pediatric leukemia (80% ofcases).32 The systemic symptomsare fever, lethargy, and anemia.Back pain has been reported as thepresenting symptom in 6% ofpatients with acute leukemia.33

Laboratory findings include anelevated peripheral white bloodcell count and an increased ESR.Vertebral compression fractureswere reported in 7% of the



A B

Fig. 5 A, Bone scan shows increased uptake representing osteoid osteoma in pedicle ofL1. B, Axial CT scan shows nidus (arrow) in pedicle and surrounding sclerosis.

Fig. 6 Aneurysmal bone cyst involving theposterior elements of T11 (arrow).

patients with acute lymphoblasticleukemia in the series by Heinrichet al32 (Fig. 7, A and B). Morecommonly, the vertebrae willshow generalized osteopenia.Other common radiographic find-ings are metaphyseal banding(Fig. 7, C), periosteal reaction, per-meative destruction, and osteo-lytic, osteosclerotic, and mixedlesions. The treatment involvestype-specific chemotherapy. The5-year survival rate is over 70% foracute lymphoblastic leukemia.Spinal bracing is used to amelio-rate symptoms, allow reconstitu-tion of collapsed vertebrae, andprevent future fractures due toosteopenia.

The spine is the most commonsite of skeletal metastasis (80%),34

but metastatic tumors are rarely thecause of nonspecific back pain.

Metastases generally affect childrenunder the age of 10 years. Pain atthe site of metastasis is the usualpresentation, accompanied by sys-temic symptoms of fever, weightloss, and malaise.

Neuroblastoma is the mostprevalent malignant condition thatproduces skeletal metastases. Theincidence of neuroblastoma is 8cases per 1,000,000 children. In asmany as 80% of cases there will bespinal metastases, most commonlyin the thoracic spine.34 Radio-graphs show diffuse permeativedestruction of the vertebrae.

Rhabdomyosarcoma, the mostprevalent soft-tissue sarcoma inchildhood, produces skeletal meta-stases in about one fifth of cases.34

Spine involvement is usual. Onradiographs, the lesions are osteo-lytic or radiolucent.

Spinal cord tumors in childrenusually appear in the first decade.Astrocytomas and ependymomasare the most common spinal cordtumors. Parents should be ques-tioned about signs of enuresis intoilet-trained children, gait abnor-malities, or delay in achievingmotor milestones. Back pain, sco-liosis, and lower-extremity weak-ness will be seen in about onethird of patients with cord tu-mors. Physical examination mayreveal asymmetric forward bend-ing and hamstring tightness.Radiographic findings includepedicle thinning or absence,foraminal widening on obliquefilms, and scoliosis without rota-tion. Gadolinium-enhanced MRimaging is the procedure ofchoice for diagnosing cord le-sions.



Fig. 7 Anteroposterior (A) and lateral (B) radiographs of the thoracolumbar spine of a 2-year-old child with new onset of back pain showmultiple compression fractures at T11–L2 and generalized osteopenia. C, Metaphyseal bands in the distal femurs.

A B C

10

10

3

Spinal Cord Abnormalities

Syringomyelias are spinal cordcysts that may or may not commu-nicate with the ventricles. Atypicalpainful scoliosis is seen in 25% ofpatients with syringomyelia. Otherassociated findings are headache,neck pain, cavus foot, and gaitabnormalities. There may be a lossof abdominal reflexes, and loss ofpain or temperature sensibility iscommon. Radiographs may showleft thoracic scoliosis, which hasbeen shown to have an associationwith intraspinal abnormality. Theprobability of there being anintraspinal lesion is increased if thepatient is less than 11 years old andhas any of the aforementionedsigns or symptoms. An MR imag-ing study of the entire spine is recommended. In the series ofSchwend et al,35 12 of 14 patientswith MR abnormalities had asyrinx.

Children with a tethered cordmay present with painful scoliosis,but this presentation is more com-mon in adults. Bladder dysfunc-tion is a frequent finding in thepediatric age group. The physi-cian must check for a cavovarusfoot, lumbosacral hair patch, he-mangioma, or dermal sinus. ABabinski sign or motor abnormali-ty may be present. Radiographsmay show spina bifida occulta ordiastematomyelia. An MR imagingstudy is necessary for evaluationand will typically demonstrate athickened filum or low-lyingconus (below L3 is always abnor-mal). Neurosurgical interventionis warranted for a documentedtethered cord.

Idiopathic JuvenileOsteoporosis

This uncommon entity usuallyaffects children under the age of 10years. Back pain was the present-ing symptom in 33% of 21 patientsin a recent study; another 33% pre-sented with long-bone pain due tometaphyseal compression frac-tures.36 The vertebrae are alwaysradiographically abnormal, with aprogressive loss of height, multiplegrowth-arrest lines, and extremesymmetric biconcavity. The find-ings from a routine laboratoryexamination are normal. Radio-density assays, such as dual-energyx-ray absorptiometry, show amarked decrease in spinal bone-mineral density.

The disease is usually self-limit-ing, and back pain should be treat-ed symptomatically. The spinalchanges are completely reversible.No benefit from hormonal or vita-min D therapy or administration ofantiresorptive agents has been doc-umented.

Diagnoses of Exclusion

Overuse syndromes are quite com-mon in children. Patients may pre-sent with back pain radiating to thebuttock, which is usually associat-ed with vigorous play or work.This pain should resolve promptlywith cessation of the aggravatingactivity. If it does not resolve with-in 2 to 3 weeks, another diagnosisshould be sought.

Compensation must be consid-ered in our litigious society if theright environment exists. A child

with back pain and a history of anautomobile accident must be ques-tioned carefully. The accident neednot be recent. The physician shouldbe wary when the parents do notallow the child to answer questions.The parents should be askedwhether litigation is pending. Thephysical examination may beequivocal for hard findings, and theradiographs are usually negative.The child’s symptoms may notresolve until the litigation is settled.

Conversion reactions are expres-sions of internal conflict as physicalsymptoms, such as back pain. Thephysician should determine whetheranyone else in the family is experi-encing or being treated for backproblems. Because back pain maybe a sign of familial unrest, the childshould be asked about a recentdeath or divorce in the family. If athorough workup for an organiccause of back pain is negative andconversion reaction seems likely,evaluation by a mental health pro-fessional is warranted.

Summary

Back pain in children and adoles-cents, although rare compared withback pain in adults, is cause for con-cern because of the high associationwith serious organic causes.Trivializing a child’s complaints as“growing pains” or “back strain” isa very dangerous practice. A carefulhistory and physical examination,along with a working knowledge ofthe differential diagnosis of pedi-atric back pain, should allow theclinician to treat young patientswith back pain safely and efficiently.



References

1. Olsen TL, Anderson RL, DearwaterSR, et al: The epidemiology of lowback pain in an adolescent population.Am J Public Health 1992;82:606-608.

2. Hensinger RN: Back pain in children,in Bradford DS, Hensinger RN (eds):The Pediatric Spine. New York: Thieme,1985, pp 41-60.

3. Conrad EU III, Olszewski AD, BergerM, et al: Pediatric spine tumors withspinal cord compromise. J PediatrOrthop 1992;12:454-460.

4. Jacobsen FS, Sullivan B: Spinal epi-dural abscesses in children. Orthopedics1994;17:1131-1138.

5. Wenger DR: Back pain in children, inWenger DR, Rang M (eds): The Art andPractice of Children’s Orthopaedics. NewYork: Raven Press, 1993, pp 455-486.

6. Micheli LJ, Wood R: Back pain inyoung athletes: Significant differencesfrom adults in causes and patterns.Arch Pediatr Adolesc Med 1995;149:15-18.

7. Fredrickson BE, Baker D, McHolick WJ,et al: The natural history of spondylol-ysis and spondylolisthesis. J Bone JointSurg Am 1984;66:699-707.

8. Jackson DW, Wiltse LL, Cirincoine RJ:Spondylolysis in the female gymnast.Clin Orthop 1976;117:68-73.

9. Ogilvie JW, Sherman J: Spondylolysisin Scheuermann’s disease. Spine1987;12:251-253.

10. Morita T, Ikata T, Katoh S, et al:Lumbar spondylolysis in children andadolescents. J Bone Joint Surg Br1995;77:620-625.

11. Bodner RJ, Heyman S, Drummond DS,et al: The use of single photon emis-sion computed tomography (SPECT)in the diagnosis of low-back pain inyoung patients. Spine 1988;13:1155-1160.

12. King HA: Back pain in children.Pediatr Clin North Am 1984;31:1083-1095.

13. Seitsalo S, Osterman K, Hyvarinen H,et al: Progression of spondylolisthesis

in children and adolescents: A long-term follow-up of 272 patients. Spine1991;16:417-421.

14. Burkus JK, Lonstein JE, Winter RB, etal: Long-term evaluation of adoles-cents treated operatively for spondy-lolisthesis: A comparison of in situarthrodesis only with in situ arthrode-sis and reduction followed by immobi-lization in a cast. J Bone Joint Surg Am1992;74:693-704.

15. Sorensen KH (ed): Scheuermann’sJuvenile Kyphosis: Clinical Appearances,Radiography, Aetiology and Prognosis.Copenhagen: Munksgaard, 1964.

16. Mandell GA, Morales RW, Harcke HT,et al: Bone scintigraphy in patientswith atypical lumbar Scheuermann dis-ease. J Pediatr Orthop 1993;13:622-627.

17. Murray PM, Weinstein SL, Spratt KF:The natural history and long-term fol-low-up of Scheuermann kyphosis. JBone Joint Surg Am 1993;75:236-248.

18. Ring D, Johnston CE II, Wenger DR:Pyogenic infectious spondylitis in chil-dren: The convergence of discitis andvertebral osteomyelitis. J PediatrOrthop 1995;15:652-660.

19. Wenger DR, Bobechko WP, Gilday DL:The spectrum of intervertebral disc-space infection in children. J Bone JointSurg Am 1978;60:100-108.

20. Epstein JA, Epstein NE, Marc J, et al:Lumbar intervertebral disk herniationin teenage children: Recognition andmanagement of associated anomalies.Spine 1984;9:427-432.

21. DeLuca PF, Mason DE, Weiand R, etal: Excision of herniated nucleus pul-posus in children and adolescents. JPediatr Orthop 1994;14:318-322.

22. Grobler LJ, Simmons EH, BarringtonTW: Intervertebral disc herniation inthe adolescent. Spine 1979;4:267-278.

23. Takata K, Inoue SI, Takahashi K, et al:Fracture of the posterior margin of alumbar vertebral body. J Bone JointSurg Am 1988;70:589-594.

24. Hashimoto K, Fujita K, Kojimoto H, et

al: Lumbar disc herniation in children.J Pediatr Orthop 1990;10:394-396.

25. Delamarter RB, Sachs BL, ThompsonGH, et al: Primary neoplasms of thethoracic and lumbar spine: An analysisof 29 consecutive cases. Clin Orthop1990;256:87-100.

26. Kneisl JS, Simon MA: Medical man-agement compared with operativetreatment for osteoid-osteoma. J BoneJoint Surg Am 1992;74:179-185.

27. Baunin C, Puget C, Assoun J, et al:Percutaneous resection of osteoidosteoma under CT guidance in eightchildren. Pediatr Radiol 1994;24:185-188.

28. Robert H, Dubousset J, Miladi L:Histiocytosis X in the juvenile spine.Spine 1987;12:167-172.

29. Hay MC, Paterson D, Taylor TKF:Aneurysmal bone cysts of the spine. JBone Joint Surg Br 1978;60:406-411.

30. Shives TC, Dahlin DC, Sim FH, et al:Osteosarcoma of the spine. J Bone JointSurg Am 1986;68:660-668.

31. Grubb MR, Currier BL, Pritchard DJ,et al: Primary Ewing’s sarcoma of thespine. Spine 1994;19:309-313.

32. Heinrich SD, Gallagher D, Warrior R,et al: The prognostic significance ofthe skeletal manifestations of acutelymphoblastic leukemia of childhood.J Pediatr Orthop 1994;14:105-111.

33. Rogalsky RJ, Black GB, Reed MH:Orthopaedic manifestations ofleukemia in children. J Bone Joint SurgAm 1986;68:494-501.

34. Leeson MC, Makley JT, Carter JR:Metastatic skeletal disease in the pedi-atric population. J Pediatr Orthop1985;5:261-267.

35. Schwend RM, Hennrikus W, Hall JE,et al: Childhood scoliosis: Clinicalindications for magnetic resonanceimaging. J Bone Joint Surg Am 1995;77:46-53.

36. Smith R: Idiopathic juvenile osteo-porosis: Experience of twenty-onepatients. Br J Rheumatol 1995;34:68-77.







http://www.ncbi.nlm.nih.gov/cgi-bin/Entrez/referer?/htbin-posdb=PubMed&list_uids=6373773&Dopt=r
































Pathologic fractures or impendingpathologic fractures in the region ofthe hip are problems that all ortho-paedic surgeons encounter. It is inthis anatomic region that the treat-ment of metastatic lesions has pro-gressed over the past 60 years frombenign neglect to aggressive inter-nal fixation.

In the late 19th and early 20thcenturies, pathologic fractures wereviewed as terminal events, una-menable to treatment by ortho-paedic surgeons. In reviewing theliterature from 1886 to 1904, Grunertstated, “in the true carcinomatousmetastasis, union of the fragmentscan never occur. . . . there has neverbeen a reported case of such arecovery.”1

By the early 1930s, reports hadbegun to appear in the literaturethat offered a somewhat more opti-mistic prognosis, with some clinicalinvestigators stating that as manyas 40% of pathologic fracturescould heal with treatment.2 Thetreatment alternatives then includ-ed disarticulation, immobilization

with traction, spica casting, andcaliper bracing combined withradiation.3

Reports of tumor resection andreplacement with large bulk allo-grafts first appeared in the litera-ture over 40 years ago. By the 1960sgreater efforts were being made notonly to alleviate pain but also torestore function in patients withmetastatic disease. More recently,advances in orthopaedic manage-ment, especially for patients whosetumors are amenable to chemother-apy and radiation therapy, havedramatically improved the outlook.

Epidemiology

Johnston4 reported that 32.5% of653 patients with malignant condi-tions had skeletal metastases atautopsy. Approximately 10% ofmetastatic lesions are located in thehip.5 Metastasis to bone most fre-quently arises from breast, pros-tate, lung, renal, and thyroid carci-nomas.6 The tumor that metasta-

sizes to the hip with the greatestfrequency is carcinoma of thebreast; 5% to 75% of metastaticlesions have been found to haveoriginated at that site.5-9 Ten per-cent of patients with disseminatedbreast cancer and 1.4% of all breastcancer patients will ultimately sus-tain a pathologic fracture of thehip.10 It is estimated that 40% ofpatients with pathologic fracturessurvive for at least 6 months aftertheir fracture, and 30% survive formore than 1 year.11

Pathophysiology

Bone metastasis leads to pathologicfracture due to the destruction ofnormal osseous architecture and

Dr. Berman is Professor and Chairman,Department of Orthopaedic Surgery andRehabilitation, Hahnemann University,Philadelphia. Dr. Hermantin is OrthopaedicResident, Hahnemann University. Dr.Horowitz is Associate Professor, Department ofOrthopaedic Surgery, and Co-Director,Regional Musculoskeletal Tumor Center,Medical College of Pennsylvania andHahnemann University Hospital System.

Reprint requests: Dr. Berman, Department ofOrthopaedic Surgery, Hahnemann University,Broad and Vine Streets, Mail Stop 420,Philadelphia, PA 19102.


Abstract

Lesions in the area of the hip secondary to metastatic disease present challeng-ing problems for the orthopaedic surgeon. With the advent of improved medicaltherapies for many types of cancer have come not only an increase in lifeexpectancy but also an increased likelihood that symptomatic metastatic bonelesions will appear. Advances in internal fixation have enabled the orthopaedicsurgeon to provide an increased level of comfort and mobility to many patientswith metastatic disease.


Metastatic Disease of the Hip:Evaluation and Treatment

Arnold T. Berman, MD, Frank U. Hermantin, MD, and Stephen M. Horowitz, MD

the replacement of bone with eithertumor or necrotic debris. Defectsin bone have been divided into twobroad categories depending ontheir size: those smaller than thediameter of the bone and thoselarger than the diameter of thebone. Defects of the former typeare considered to be stress risers,reducing the strength of the boneby causing an uneven distributionof stresses during load bearing,which in torsion can decrease bonestrength by 60%.12 The forces gen-erated by normal activities can then exceed the lowered strengththreshold, resulting in fracture.

The pathophysiology of meta-static bone destruction is mediatedby osteoclasts, which appear to be activated, perhaps indirectly,through osteoblasts, by tumorproducts that are not yet fullyunderstood. Various substances oftumoral origin have been proposedas mediators for this osteoclast activation, such as transforminggrowth factors, prostaglandins, cyto-kines, and parathyroid hormone–related peptide.13 It is thought thattumor cells directly destroy boneonly in the last stages of themetastatic process.

Evaluation

The preoperative evaluation beginswith a history and physical exami-nation. The patient should bequestioned about the severity, loca-tion, and characteristics of pain andwhich activities increase the levelof discomfort. Pain at rest may sig-nal an expanding osseous lesion.

It is important to identify the pri-mary tumor. If the primary site isunknown, a more extensive evalua-tion needs to be conducted, usuallyin cooperation with the primary-care physician and a medical oncol-ogist. A full description of the eval-uation of the patient with an

unknown primary neoplasm isbeyond the scope of this article, butmay include chest radiography,mammography, breast examinationin female patients, prostate exami-nation in male patients, renal ultra-sound, and serologic and other diag-nostic tests as indicated. Serologictests should at a minimum includeserum calcium, phosphorus, andelectrolyte determinations and acomplete blood cell count.

Most pathologic fractures orimpending fractures present with aknown primary malignant neo-plasm. It is the responsibility of thetreating orthopaedic surgeon toconfirm the diagnosis and in somecases to initiate an evaluation torule out a concurrent primary bonetumor. The need for assistivedevices, such as a cane or walker,and the distance the patient canwalk before having to rest shouldbe established.

The radiographic evaluationshould include plain radiographsthat visualize the hip, pelvis, andfemur (Fig. 1). A current bone scanis needed to assess other areas inthe skeleton that may also be at riskfor pathologic fractures. A bonescan that shows multiple lesions isalso evidence that the tumor in thehip is metastatic and not a new pri-mary neoplasm. It also should berecognized that in patients withmultiple myeloma, the bone scanmay not reveal lesions associatedwith skeletal destruction. If thepatient is too uncomfortable to tol-erate lying supine in the nuclearmedicine suite while bone scanningis being performed, a skeletal sur-vey may be more useful. Ad-ditional diagnostic studies, such ascomputed tomography (CT) andmagnetic resonance imaging, maybe required to evaluate the amountof bone destruction, particularly ifacetabular lesions are present.

A patient who presents with asingle lesion in the hip and a

known primary tumor but no otherevidence of metastatic diseaseshould be evaluated with caution.Many primary bone tumors, suchas chondrosarcoma, also typicallypresent in the hip area. It is notuncommon for patients initiallyassumed to have a metastatic lesionto undergo hip replacement orinsertion of an intramedullarydevice, only to have the pathologistsubsequently report that the tumorwas not metastatic but rather a pri-mary bone sarcoma. For this rea-son, it is recommended that a tis-sue diagnosis be obtained beforedefinitive internal stabilization in apatient with a solitary lesion aboutthe hip and a known primarytumor but no other evidence ofmetastatic disease. For lesions ofthe acetabulum, this can often beachieved with CT-guided needlebiopsy. For other lesions about thehip, CT-guided needle biopsy, anopen biopsy as a separate proce-dure, or an open biopsy and afrozen section may be appropriate.The important point is that thediagnosis should be confirmedbefore definitive treatment and sta-bilization.

Indications for Surgery

In 1970, Parrish and Murray7 pro-posed the following guidelines forselecting patients for operativeintervention: (1) The patient’s gen-eral condition must be sufficientlygood and life expectancy sufficient-ly long (more than 6 weeks) to jus-tify the procedure. (2) The surgeonmust be convinced that the opera-tion will be more beneficial thanclosed treatment. (3) The quality ofthe bone both proximal and distalto the fracture site must be ade-quate for stable fixation. (4) Theprocedure must expedite mobiliza-tion of the patient or facilitate gen-eral care.

Metastatic Disease of the Hip


In 1976, Harrington et al14 pro-posed additional guidelines basedon evaluation of plain radiographs.They considered lesions at risk forcausing a pathologic fracture to bethose that (1) were greater than 2.5cm in diameter, (2) destroyed 50%of the cortex, or (3) were painfuldespite treatment with radiation.An avulsion fracture of the lessertrochanter is also a risk factor.

In 1989, Mirels15 proposed ascoring system for diagnosingimpending pathologic fractures inlong bones. This system considersthe site of the lesion (i.e., upper orlower extremity). It also takes intoaccount the presence and severityof pain; whether the lesion is blas-tic, lytic, or mixed; and the size ofthe lesion. The higher the score,

the greater the likelihood that thepatient will sustain a pathologicfracture if left untreated.

In 1995, Hipp et al16 also reportedon developing a framework for iden-tifying patients with bone defectssecondary to metastatic disease thatwould require prophylactic stabili-zation. Their work applied engi-neering principles to analysis of CTstudies to estimate the load-bearingcapacity, which, when comparedwith the load-bearing requirements,can be used to calculate a factor ofrisk. Their findings have yet to bevalidated for areas other than thespine but do suggest that the currentclinical guidelines of a 2.5-cm defector 50% cortical destruction are asso-ciated with large errors in estimatingthe load-bearing capacity of bone.

Patients who present with le-sions that do not meet the criteriafor internal fixation should be re-ferred to a radiation oncologist forconsideration of radiation therapy.Lesions that are small and minimal-ly symptomatic can often be treatedwith radiation therapy (if the tumoris radiosensitive) and protectedweight bearing with careful obser-vation. Follow-up by the ortho-paedic surgeon during and afterthe radiation treatments is impera-tive. If the amount of bone destruc-tion increases or if the lesion re-mains or becomes symptomatic,operative intervention may be in-dicated. In addition, the weight-bearing status of the patient mustbe monitored. In most cases, par-tial weight bearing for at least 6

Arnold T. Berman, MD, et al


Hip metastasis?

Operative treatment

Intertrochanteric Subtrochanteric Proximal femurAcetabulum

• THR• PMMA

No response

Observation

Favorableresponse

Femoral neck

• X-ray of hip, pelvis, femur• Bone scan• MRI or CT of hip

• Hip lesion > 2.5 cm• > 50% destruction of cortex• Fracture of lesser trochanter• Failure of XRT

• Bipolar hemi-arthroplasty

• PMMA

• Compressionscrew

• PMMA

• Compression screw• Long side plate• Reconstruction nail• PMMA

• Bipolar hemi-arthroplasty

• Proximal femoralreplacement

• Hip lesion < 2.5 cm• < 50% destruction of cortex• Lesser trochanter intact

Nonoperative treatment:• XRT• Protected weight bearing

Fig. 1 Algorithm for evaluation and treatment of metastatic disease to the hip. CT = computed tomography; MRI = magnetic resonanceimaging; PMMA = polymethylmethacrylate; THR = total hip replacement; XRT = radiation therapy.

weeks is advisable, usually untilthere is evidence that the lesionshave undergone healing.

If a patient sustains a fracturethrough a lesion during radiationtherapy, the treatment is internalfixation or prosthetic replacement,similar to the treatment used forimpending fractures. If the fractureoccurs through a lesion that hasbeen irradiated and the patient hasa relatively long projected survival,strong consideration should begiven to prosthetic replacement,because of the risk of nonunionand ultimate failure of internal fix-ation. However, it is always pref-erable to fix impending fracturesprophylactically, so the patient canavoid the discomfort and morbidi-ty associated with a pathologicfracture. A recent study by Alganand Horowitz17 demonstrated thatthe results of internal fixation forlesions about the hip were similarto those for the same operative pro-cedure performed for nonmeta-static conditions.

Treatment

Historical ReviewThe management of hip metasta-

sis has progressed considerablyover the past 60 years. At the turnof the century, the most commonform of therapy was the use oflight traction. The healing rate waspoor, with only 20% of patientsshowing evidence of fracture heal-ing.1 Emphasis was then placed onidentifying and treating sympto-matic lesions before fracture. Forexample, Coley and Higinbotham18

reported that fractures could beprevented by the use of calipersplints to decrease the load on theaffected bone.

By 1950 reports had appeared inthe literature supporting tumorresection and replacement with alarge bulk allograft. However, the

most favorable report of use of thismodality, by Higinbotham andColey,19 concerned patients whosetumors involved the upper extrem-ity or a minimal weight-bearingbone, such as the fibula. At thistime, internal fixation of impend-ing or actual fractures also becamepopular. In pathologic femoralneck fractures, the internal fixationnails often lost fixation in theweakened femoral head. This ledFrancis et al,20 in 1962, to advocateresection of the femoral head andneck as a primary treatment forlesions involving those structures.They believed that resection of thehead and neck, especially inpatients with early lesions, whohad the greatest life expectancy,would provide rapid relief of painand adequate ambulation, as wellas reduce the complications ofsurgery to a minimum. In theirseries, all 19 patients had satisfac-tory results as measured by reliefof pain, but only 5 of the patientswere able to partially bear weighton the surgically treated side with-in 3 weeks of surgery. Of the 4who survived for 2 years, only 2could fully bear weight withoutdiscomfort.

By 1976, Harrington had begunadvocating the use of polymethyl-methacrylate (PMMA) cement asan adjunct to internal fixation inpatients in whom a large amount ofbone had been lost as a result ofmetastatic disease. In 375 patients,Harrington et al14 excised the lesionand all inadequate bone stock andthen performed internal fixation orprosthetic replacement and rein-forcement with PMMA. A 94%ambulation rate was achieved. Thepresence of the PMMA did notseem to interfere with bone unionor radiation therapy.

Also in 1976, Zickel and Mou-radian8 reported on the use of anintramedullary fixation device theyhad devised for fractures in the

subtrochanteric region that did notrequire the use of PMMA. Thirty-five pathologic fractures and 11impending fractures were treatedsolely with the device; no attemptwas made to excise the tumor. Theauthors found that those patientswith an impending fracture ambu-lated sooner and survived longerthan those who presented with acomplete fracture. Operating timeand blood loss were not differentbetween groups. They further stat-ed that the criteria for operativeintervention should not be limitedto only those with 6 or more weeksto live.

In 1980, Lane et al10 reported onthe use of an endoprosthetic re-placement or a total hip prosthesisfor pathologic fractures or im-pending fractures of the hip. Theyconsidered that the combination ofan impending fracture (defined onthe basis of the size of the lesionand the amount of pain experi-enced by the patient) and a lifeexpectancy of more than 1 monthwas an indication for surgicalintervention. Good to excellentresults with regard to relief of painwere obtained in all of the patientstreated with either an Austin-Moore endoprosthesis (cementedor uncemented) or a total hipreplacement.

In 1981, Harrington21 reportedon the use of total hip prostheses inpatients with acetabular lesions. Anumber of their patients had lost somuch acetabular bone that conven-tional prostheses would not pro-vide sufficient support. Harring-ton designed a larger acetabularcomponent that would distributethe mechanical load to areas of lessinvolved bone.

Current TechniquesAfter a complete preoperative

evaluation, a surgical plan must becarefully designed. Patients whohave a highly vascular lesion, such



as metastatic renal carcinoma,should be treated with arterialembolization before surgery todecrease intraoperative blood loss.

AcetabulumIt is recommended that patients

with a pathologic fracture or im-pending fracture of the acetabulumbe evaluated with magnetic reso-nance imaging or CT so that theextent of bone destruction can beaccurately assessed.22 Patients withrelatively small to moderateamounts of bone destruction fre-quently can be treated with a protru-sio ring, either alone or in combina-tion with an acetabular cup (Fig. 2).

Harrington21 subdivided pa-tients with acetabular involvementinto four groups on the basis of thelocation of the lesion, the extent ofinvolvement, and the techniquerequired to accomplish the acetab-ular reconstruction. In class I, thelateral cortices and the superiorand medial portions of the acetabu-lum were structurally intact.Patients with class I lesions weretreated by conventional total hiparthroplasty, frequently with place-

ment of mesh along the medialportion of the acetabular wall forreinforcement. In class II, themedial portion of the wall wasdeficient. Patients with class IIlesions were treated with a protru-sio ring. In class III, the lateral cor-tices and the superior portion ofthe wall were deficient. Patients

with class III lesions were treatedwith Steinmann pins directed alongthe medullary canal of the ilium inaddition to the protrusio ring andacetabular prosthetic component(Fig. 3). Patients with class IVlesions had only a solitary metasta-sis in the acetabular area andunderwent an en bloc resection.



Fig. 2 A, Renal cell carcinoma metastatic to the acetabulum. B, Protrusio ring and acetab-ular cup in place.

A B

A

Fig. 3 Treatment of class III acetabular metastatic involvement. Anteroposterior (A) and anterolateral (B) views of the pelvis demon-strate the thinness of the ilium superior to the acetabulum, making that site unfavorable for attempting to anchor an acetabular prostheticcomponent. C, Tumor has destroyed the superior and medial portions of the acetabular bone, leaving only minimal intact cortex for fixa-tion of the acetabular component. D, Resection of tumor tissue leaves a large cavity as well as destruction of the acetabular roof, the medi-al portion of the wall, and most of the rim. Steinmann pins can be drilled into structurally sound bone of the superior part of the iliumand across the sacroiliac joint. E, The acetabular component positioned in the protrusio acetabuli shell. F, The combination of acetabularcup, protrusio acetabuli shell, and Steinmann pins incorporated into methylmethacrylate effectively transmits weight-bearing stresses intothe strong bone of the iliac wing and sacrum.

D E FCB

In patients who undergo hipreplacement with a protrusio cup,postoperative care is similar to thatafter a hip replacement for non-metastatic disease. This involvesdislocation precautions and partialweight bearing for 6 weeks postop-eratively. In cases of very exten-sive bone loss, some considerationshould be given to treating thepatient nonoperatively or with aGirdlestone procedure, because ofthe unreliability of internal fixation.

Femoral Head and NeckWe treat impending and com-

plete fractures of the femoral headand neck by cemented bipolarhemiarthroplasty because progres-sion of these lesions may result infailure of internal fixation (Fig. 4).A common error in patients withpathologic fractures in the femoralhead and neck is failure to appreci-ate distal lesions. This may resultin unrecognized perforation whilepreparing the femoral canal; thestem of the prosthesis often goesthrough this perforation. In addi-tion, a stem that ends just proximalto a missed distal lesion may causea stress riser, leading to later frac-ture. Therefore, it is recommendedthat radiographs be taken of theentire femur before this procedure.In patients who have only proximaldisease, a long-stem componentwill often bypass the lesion. Ifthere is a large lesion in the supra-condylar area, it may be necessaryto place a fixation device, such as asupracondylar screw and sideplate, to avoid a stress riser andpossibly a fracture about the tip ofthe prosthesis.

Postoperatively, patients whoundergo bipolar hemiarthroplastyfor metastatic disease are treated inmuch the same way as patients whoundergo this procedure for otherconditions. This involves disloca-tion precautions and partial weightbearing for 6 weeks after surgery.

Intertrochanteric FracturesImpending or complete fractures

in the intertrochanteric area with aminimal amount of bone destruc-tion can usually be treated with ascrew and side plate. Adjunctivebone cement is often needed toassist in fixation of the lag screw orproximal screw in the plate. Thistype of fixation is especially advan-tageous in patients who presentwith a solitary lesion in the inter-trochanteric area that is suspectedof being a metastasis but who haveno known primary tumor and noother lesions identified by bonescanning. The lesion can be partial-ly excised, usually with the directlateral approach, to obtain a biopsyspecimen for tissue diagnosis. If asarcoma is encountered on thefrozen section, the surgical proce-dure should be stopped unless thispossibility had been consideredpreoperatively and an en blocresection had been planned. Abiopsy specimen should beobtained as a separate procedure

from any lesion strongly consid-ered to be a sarcoma.

The disadvantage of using thescrew and side plate for these frac-tures is that stress is placed on thedevice during ambulation, whichmay cause it to eventually fail if thepatient becomes a long-term sur-vivor. Progression of the tumormay also compromise fixation,especially if it is relatively radio-resistant. In addition, in patientswho have received radiation thera-py in this area and survive for arelatively long period of time, theend of the plate may cause a stressriser on bone that is weakenedfrom the radiation and may even-tually cause a fracture at the distalaspect of the plate.

For patients with extensive de-struction in the intertrochantericarea and a complete or impendingfracture, use of a long-stem hipprosthesis or a proximal femoral re-placement is recommended (Fig. 5).The femoral component is usuallycombined with a bipolar head.



A B

Fig. 4 A, Metastatic lesion involving the femoral head and neck. B, After treatment withbipolar hemiarthroplasty.

Postoperatively, patients whoare treated with a compressionscrew progress at a rate dependingon the extent of bone loss and thestability of fixation. Patients whoundergo proximal femoral replace-ment are maintained in a hipabduction brace with a knee-foot-ankle orthosis extension for 6 to 8weeks postoperatively to decreasethe risk of dislocation.

Subtrochanteric FracturesIn patients with obvious meta-

static disease, we recommend in-tramedullary fixation with screwsplaced along the femoral neck.This is biomechanically superior to the screw and side plate, and itsuse is believed to be associated witha lower probability of mechanicalfailure. In the past, the Zickel nail(Howmedica, Rutherford, NJ) wasthe primary implant used for thistype of fixation. Currently, most ofthe manufacturers who producetrauma instrumentation have“reconstruction” nails that can beused for this purpose. In most of

these devices, two screws aredirected up the femoral neck, andthe nail can be locked distally.Recently, Synthes (Paoli, Pa) intro-duced an unreamed femoral nailthat utilizes a spiral blade ratherthan screw fixation in the femoralhead and neck. The potentialadvantage of this device is that itcan be inserted without reaming,which makes the surgical proce-dure faster. In addition, the angleof the blade to the nail can bechanged, which gives the surgeonmore flexibility.

In most cases, we recommendlocking the rod both proximallyand distally because of the low riskof complications associated withplacing the distal screws and thepotential for loss of stability if theyare not used. We use only bonecement when loss of bone makesthe screw fixation tenuous. Inthose instances, a 1⁄4-inch drill bit isused to make portals in the boneboth proximal and distal to thescrews. The area is first irrigatedwith saline, and the PMMA is then

inserted with a syringe so that itflows around the rod and thescrews.

Postoperatively, patients do notrequire dislocation precautions.Their weight-bearing status is pro-gressed depending on the extent ofbone loss and the stability of fixa-tion. For most patients, full weightbearing or ambulation with a caneis possible 6 to 12 weeks postopera-tively.

Radiation Therapy

In patients who have not receivedradiation preoperatively, radiationin the postoperative period may behelpful in slowing progression of alesion that may ultimately lead todisruption of the internal fixation.Townsend et al23 demonstratedthat the combination of postopera-tive radiation therapy and surgeryled to a better outcome than sur-gery alone, with a five timesgreater probability of attainingmaximal use of the extremity and adecreased need for a second surgi-cal procedure. We favor postoper-ative rather than preoperative radi-ation therapy whenever it is likelythat internal fixation will be need-ed. This is in part because of con-cern about impeded fracture heal-ing when preoperative radiationtherapy is used.

Summary

The outlook for patients withmetastatic disease about the hiphas improved dramatically, partic-ularly for patients whose tumorsare amenable to chemotherapyand/or radiation therapy. Ad-vances in orthopaedic managementfor patients with impending oroccult fractures may result in anincreased level of comfort andmobility.



A B

Fig. 5 A, Extensive metastatic disease involving the femoral neck and intertrochantericareas of the hip. B, After proximal femoral replacement.

References

1. Grunert D: Über PathologischeFrakturen (Spontan-Frakturen). DtschZ Chir 1905;76:254-289.

2. Eliason EL, Wright VWM: Pathologicfractures. Surg Clin North Am 1930;10:1335-1376.

3. Welch CE: Pathological fractures dueto malignant disease. Surg GynecolObstet 1936;62:735-744.

4. Johnston AD: Pathology of metastatictumors in bone. Clin Orthop 1970;73:8-32.

5. Bonarigo BC, Rubin P: Nonunion ofpathologic fracture after radiationtherapy. Radiology 1967;88:889-898.

6. Habermann ET, Sachs R, Stern RE, et al:The pathology and treatment of meta-static disease of the femur. Clin Orthop1982;169:70-82.

7. Parrish FF, Murray JA: Surgical treat-ment for secondary neoplastic frac-tures: A retrospective study of ninety-six patients. J Bone Joint Surg Am 1970;52:665-686.

8. Zickel RE, Mouradian WH: Intra-medullary fixation of pathologicalfractures and lesions of the sub-trochanteric region of the femur. J Bone Joint Surg Am 1976;58:1061-1066.

9. Levy RN, Sherry HS, Siffert RS: Sur-gical management of metastatic dis-

ease of bone at the hip. Clin Orthop1982;169:62-69.

10. Lane JM, Sculco TP, Zolan S: Treat-ment of pathological fractures of thehip by endoprosthetic replacement. J Bone Joint Surg Am 1980;62:954-959.

11. Marcove RC, Yang DJ: Survival timesafter treatment of pathologic fractures.Cancer 1967;20:2154-2158.

12. Pugh J, Sherry HS, Futterman B, et al:Biomechanics of pathologic fractures.Clin Orthop 1982;169:109-114.

13. Body JJ: Metastatic bone disease:Clinical and therapeutic aspects. Bone1992;13(suppl 1):S57-S62.

14. Harrington KD, Sim FH, Enis JE, et al:Methylmethacrylate as an adjunct ininternal fixation of pathological frac-tures: Experience with three hundredand seventy-five cases. J Bone JointSurg Am 1976;58:1047-1055.

15. Mirels H: Metastatic disease in longbones: A proposed scoring system fordiagnosing impending pathologic fractures. Clin Orthop 1989;249:256-264.

16. Hipp JA, Springfield DS, Hayes WC:Predicting pathologic fracture risk inthe management of metastatic bonedefects. Clin Orthop 1995;312:120-135.

17. Algan SM, Horowitz SM: Surgicaltreatment of pathologic hip lesions inpatients with metastatic disease. ClinOrthop 1996;332:223-231.

18. Coley BL, Higinbotham NL: Diag-nosis and treatment of metastaticlesions in bone. Instr Course Lect 1950;7:18-25.

19. Higinbotham NL, Coley BL: The treat-ment of bone tumors by resection andreplacement with massive grafts. InstrCourse Lect 1950;7:26-33.

20. Francis KC, Higinbotham NL, CarrollRF, et al: The treatment of pathologi-cal fractures of the femoral neck byresection. J Trauma 1962;2:465-473.

21. Harrington KD: The management ofacetabular insufficiency secondary tometastatic malignant disease. J BoneJoint Surg Am 1981;63:653-664.

22. Horowitz SM: The management ofpathological hip fractures. OperativeTechniques Orthop 1994;4:122-129.

23. Townsend PW, Rosenthal HG, SmalleySR, et al: Impact of postoperative radi-ation therapy and other perioperativefactors on outcome after orthopedicstabilization of impending or patholog-ic fractures due to metastatic disease. J Clin Oncol 1994;12:2345-2350.





















The number of female athletesincreased dramatically after thepassage of Title IX in 1972. Whenwomen were first admitted to theUS military academies in 1976, anincreased incidence of overuseinjuries was reported in femalecadets. As greater attention waspaid to the conditioning and train-ing of girls in the preadolescentand high-school age groups, theirinjuries were found to be generallysport-specific rather than gender-specific. However, a study of theWest Point class of 1992 found thatwomen engaging in sports hadmore lower-extremity injuriesthan men, particularly stress frac-tures and patellofemoral pain. Ahigher incidence of noncontactanterior cruciate ligament (ACL)

injuries has also been seen inwomen involved in a number ofsports. The problems that occurexclusively or more commonly inthe female athlete are the subjectof this article.

Spondylolysis andSpondylolisthesis

Spondylolysis is found in 6% of thegeneral population, but there is ahigher prevalence in persons whoparticipate in activities that placethe lumbar spine in hyperlordosis,such as diving, gymnastics, anddance.1 This association withhyperlordosis suggests that repeti-tive bending stresses in the parsinterarticularis can cause a stress

fracture. Progression of spondy-lolisthesis is more likely in femalesand is rare after adolescence.2

Among young patients, the mostcommon type of spondylolysis isisthmic and occurs at L5, withresultant slippage on the sacrum.Isthmic spondylolysis can be lytic,representing a fatigue fracture, orcan be associated with an elongatedbut intact pars.

History and PhysicalExamination

The athlete may present withintermittent low back pain associatedwith particular activities or acutesymptoms related to a specificevent. Radicular symptoms areuncommon in young patients un-less a high-grade slip is present.

On physical examination, somepatients demonstrate the classic

Dr. Teitz is Associate Professor, Departmentof Orthopaedic Surgery, University ofWashington, Seattle. Dr. Hu is AssistantProfessor, Department of Orthopaedic Surgery,University of California, San Francisco. Dr.Arendt is Associate Professor, Department ofOrthopaedic Surgery, University of Minnesota,Minneapolis.

Reprint requests: Dr. Teitz, Department ofOrthopaedic Surgery, Box 354060, Universityof Washington, Seattle, WA 98195-4060.


Abstract

Although many of the problems faced by the female athlete affect the male ath-lete as well, some occur exclusively or more commonly in women. Theseinclude spondylolisthesis, stress fractures in the pelvis and hip, and pelvic floordysfunction. Female athletes are also more likely to have patellofemoral prob-lems, noncontact anterior cruciate ligament tears, and bunions. For many ofthese conditions, the relative influences of osseous anatomy, ligamentous laxity,and the effect of sex hormones have not yet been established. There are alsoproblems related specifically to the menstrual cycle and pregnancy.Amenorrhea is present in up to 20% of vigorously exercising women. The term“female athlete triad” has been coined to describe the complex interplay of men-strual irregularity, disordered eating, and premature osteoporosis seen in thefemale athlete. Many of the concerns related to exercise during pregnancy focuson the safety of the fetus rather than the athlete herself. Musculoskeletal prob-lems in the physically active pregnant woman are related to weight gain, liga-mentous relaxation, lordosis, and change in the center of gravity.


The Female Athlete:Evaluation and Treatment of Sports-Related Problems

Carol C. Teitz, MD, Serena S. Hu, MD, and Elizabeth A. Arendt, MD

Phalen-Dickson sign: tight ham-strings and a bent-knee, flexed-hipgait.3 It is unclear whether thisstance is secondary to the verticalposition assumed by the sacrum,with resultant flexion of the pelvis,or whether nerve-root entrapmentleads to the hamstring tightness. Ifslippage has occurred, careful pal-pation of the spine may reveal astepoff, with prominence of the L5spinous process.

Radiologic EvaluationLateral plain radiographs may

not show spondylolysis but shouldreveal spondylolisthesis when it ispresent. The radiographs shouldbe taken with the patient standing.Spondylolisthesis is graded withthe Meyerding method accordingto the percentage of forward sliprelative to the top of the sacrum.The slip angle is a measure of lum-bosacral kyphosis and is an impor-tant prognostic indicator of slipprogression (Fig. 1).3 The slip angleis normally 0 degrees. Spondylo-listhesis is more likely to progressin patients with slip angles greaterthan 50 degrees.

In the case of patients who pre-sent with typical spondylolyticcomplaints but whose radiographsare normal, a bone scan may identi-fy an early stress fracture of thepars interarticularis. Scans can alsohelp determine whether a positiveradiographic finding represents arecent fracture or the lytic defect islong-standing.4 Computed tomog-raphy can occasionally help estab-lish whether the pars fracture iscomplete.

TreatmentRest can help decrease symp-

toms. Bracing is successful inreducing or eliminating symptomsin as many as 80% of patients with aless than 25% slip. The brace mustdecrease lumbar lordosis.5 In casesof acute lytic defects, prompt treat-

ment with bracing and rest also canpermit healing of the pars fracture.Physical therapy to stretch the lumbo-dorsal fascia and strengthen theabdominal musculature may helpdecrease hyperextension stress.

Patients with less than 25% slip-page can return to physical activityafter successful conservative treat-ment. Activities that do not hyper-extend the spine are permitted forthose whose slippage is between25% and 50%. Patients whosesymptoms are refractory to bracingand skeletally mature patients witha slip greater than 50% can be con-sidered for surgical stabilization.The standard treatment for slipsmeasuring as much as 50% at theL5-S1 level is bilateral lateral fusionof L5 to the sacrum with the use ofautologous bone grafts. When theslippage is greater than 50%, inclu-sion of L4 is advocated so that thefusion will be placed under com-pression.6 Fusion has been reportedto be successful in 83% to 95% ofpatients; the clinical results havebeen reported as good to excellentin 75% to 100% of cases.7 Althoughit may take as long as 18 months,hamstring tightness generally re-solves after fusion in situ and doesnot require decompression. Mostpatients who undergo successfulfusion for spondylolisthesis canreturn to sports as desired. Theymay lose a few degrees of motionin the lumbar spine, which may bea problem for athletes whose sportrequires extreme flexibility.

A young patient with a parsdefect at a level other than L5 andlittle slippage may be consideredfor a pars repair. This can beaccomplished by wiring the trans-verse processes to the spinousprocess or, if the transverse pro-cesses are small, by placing a screwin the pedicle and then wiring thescrew to the spinous process. Thepars defect must be cleaned of softtissue and bone grafted. After this

procedure, healing can be expectedin as many as 90% of patients un-der the age of 25 with normal disksand facets and less than 1 or 2 mmof slippage.8

Pubic Ramus StressFractures

Etiologic factors commonly seen inassociation with pubic ramus stressfractures are tight adductors, leg-length discrepancy (fracture in theshorter leg), a crossover runningstyle, and overstriding, either togain speed or to keep up with ataller running partner. All of thesefactors have in common chronicpull by the adductor musculatureattached to the pubic ramus.


Stress fractures of the pubicramus present with pain in thegroin. The site of fracture is tenderto palpation. Pain may be elicited bycompressing the pubic symphysis.

The Female Athlete


SA=26°

Fig. 1 The slip angle (SA) is the anglebetween the line perpendicular to a linedrawn along the posterior border of thesacrum and a line along the inferior borderof L5.

Occasionally, stretching of theadductor muscles or adduction ofthe thigh against resistance willproduce pain, making a stress frac-ture difficult to distinguish clinical-ly from an adductor strain.

Radiologic EvaluationRadiographs usually show a ver-

tical fracture in the pubic ramus,with or without callus (Fig. 2, A).When the radiographic appearanceis normal, a bone scan will revealincreased uptake in the area of frac-ture (Fig. 2, B).

TreatmentTreatment involves cessation of

participation in impact sports untilpubic tenderness is gone, whichusually takes 6 to 12 weeks. Anycontributing factors should beidentified and resolved throughphysical therapy and training oruse of a shoe lift if leg-length dis-crepancy is present. Subsequently,a progressive exercise programshould include stretching andstrengthening of the adductor mus-cles and running in the shallowend of a swimming pool. Whenthese activities are well tolerated,the patient can begin sport-specifictraining.

Femoral Neck StressFractures

Femoral neck fractures are typical-ly due to overuse, particularly inan amenorrheic athlete.


These fractures present with groinpain or anterior thigh pain in the dis-tribution of the obturator nerve.Pain may be present during weight-bearing activities as well as duringswimming, bicycling, and evenactivities of daily living. On physicalexamination, point tenderness may

be absent, except in the extremelythin individual. However, theTrendelenburg test is often positive,and an abductor lurch during gait is often present. Passive range ofmotion of the hip and abductionagainst gravity, with and withoutresistance, may be painful.

Radiologic EvaluationTo avoid mistakenly treating a

stress fracture as a muscle injury,thereby risking complete fracture,the patient should be instructed touse crutches, and imaging studiesshould be obtained immediatelywhen a femoral neck stress fractureis suspected. The initial plain radio-graphs may not reveal the fracture.However, after 2 or 3 weeks, callusmay be found at the fracture site.A bone scan will reveal the fracturein the first 48 hours after occur-

rence. If the diagnosis is question-able, magnetic resonance (MR)imaging may be useful in differen-tiating a stress fracture from otherlesions that produce increaseduptake on a bone scan.9

TreatmentExtreme caution should be exer-

cised when a stress fracture of thefemoral neck is suspected. Becausea frank femoral neck fracture canhave a poor outcome, such as avas-cular necrosis of the femoral head,stress fractures in this area in youngpatients must be treated aggressive-ly. Not all of these injuries needinternal fixation. It is generallyaccepted that fractures on the ten-sion (lateral) side of the neck aremore likely to progress to completefracture and should be internallyfixed in situ with either multiple

Carol C. Teitz, MD, et al


Fig. 2 A, Radiograph shows a stress fracture in the inferior pubic ramus. B, Bone scanreveals the stress fracture seen on the radiograph as well as an additional fracture in thesuperior pubic ramus.

A B

compression screws or a slidingscrew and side plate.10 Fractures onthe compression (medial) side of theneck may heal without any internalfixation. However, they must bewatched closely, and the patientmust be reliable. If the patient cannotwalk without a lurch, she shoulduse crutches for 3 to 4 weeks or untilshe can walk normally and withoutpain. The risk of frank fractureshould be discussed with the pa-tient at the outset. Stress fracturesmay take 3 months to heal.

Pelvic Floor Dysfunction

The muscles of the pelvic floor, orpelvic diaphragm, include the leva-tor ani group (pubococcygeus, pubo-rectalis, pubovaginalis, and iliococ-cygeus) and the coccygeus. Thepiriformis and obturator internusmuscles are also continuous withthe pelvic diaphragm. The endo-pelvic fascial floor and the smooth-muscle diaphragm in the base ofthe broad ligament also contributeto support of the pelvic floor.

There are four types of pelvicfloor dysfunction: disuse, support-ive, hypertonic, and incoordina-tion.11 Disuse dysfunction implieslack of awareness of the pelvicfloor muscles due to lack of train-ing, modesty, or muscle imbalance.This presents as stress incontinenceor urge incontinence.

Supportive dysfunction is sec-ondary to loss of nerve (pudendal),muscle, ligament, or fascial integrity.This can be congenital, hormonal,or secondary to connective tissuedisease or, most commonly, can bedue to trauma from childbirth orsurgery. It presents as back painand a sensation of suprapubic pres-sure.

Hypertonic dysfunction is char-acterized by excessive tone in thepelvic floor muscles. This can bepsychogenic, iatrogenic, or muscu-

loskeletal in origin. The primarysymptom is pain, which is poorlylocalized in the perivaginal, peri-rectal, or suprapubic regions.Because the pain can radiate downthe posterior aspect of the thigh,hypertonic dysfunction is easilyconfused with sciatica. Symptomsmay be reproduced by a vaginal orrectal examination. Associated sex-ual dysfunction and dyspareuniaare common.

Incoordination dysfunction isdifficulty in contracting or relaxingthe pelvic floor muscles. Imbalanceof the gluteal, adductor, andabdominal muscles can mask weakcontraction of the pelvic floor mus-cles. Myofascial or scar-tissue for-mation can restrict the contractilityof the pelvic floor muscles. Thiscan also be caused by neural dam-age and motor dysynergia.11


Pelvic floor dysfunction usuallypresents as urinary incontinencebut can also produce pain thatmimics sciatica.11 Patients presentwith pain localized to the upperouter quadrant of the buttock orpain in the posterior aspect of thethigh that originates from the areaof the ischial tuberosity. The painis not exacerbated by activities orthe Valsalva maneuver. A historyof urinary incontinence is common,and dyspareunia may be present.The patient is likely to be parousand to have had one or more trau-matic vaginal deliveries.

The physical examination ismore remarkable for pertinent neg-ative findings than for positivefindings. Patients with sciatica sec-ondary to pelvic floor dysfunctionhave no restriction of back motionand no neurologic findings. Theirhamstrings are not particularlytight, and a straight-leg-raising testdoes not reproduce the sciatic pain.Occasionally, piriformis muscle

tightness will be noted, as manifest-ed by decreased internal rotation onthe involved side. Both hips nor-mally have a similar total arc ofmotion, although one hip may havemore or less internal rotation thanthe other. In patients with a unilat-erally tight piriformis muscle, thetotal arc of motion of the involvedhip will be less than that of the nor-mal hip, and the loss will be ininternal rotation. A vaginal or rec-tal digital examination may revealhypertonic or flaccid muscles.When muscles are hypertonic, itwill be difficult to perform theexamination. When muscles areflaccid, the patient will not be ableto squeeze the examining digit.

Additional StudiesThe tone of the pelvic floor mus-

cles can be determined with use ofperineometers or surface elec-tromyography. These can alsoreveal whether the patient can vol-untarily contract her perineal mus-cles in a coordinated fashion orwhether they are constantly flaccidor tonic.

TreatmentThe initial treatment for these

problems is pelvic floor exercises.12

Many women do not perform theseexercises correctly. Improper exer-cise technique causes bearing downinstead of lifting up of the pelvicfloor. The principles of training arethe same as for any other muscle:specificity, overload, and progres-sion. Biofeedback is very useful ineducating women about the use oftheir pelvic floor muscles. This canbe done with pressure-sensing per-ineometers that can be insertedvaginally (generally by a physicaltherapist) to obtain numerical val-ues and to provide visual feedbackto the patient during pelvic floormuscle contractions. Surface elec-tromyography can also be used forbiofeedback in the training mode.

The Female Athlete


A home program utilizingweighted vaginal cones can beinstituted. Vaginal cones encour-age proper lifting of the pelvic floorduring contractions. These exercisesare begun in positions that elimi-nate gravity and are advanced toupright and functional positions.They can be done concentrically,eccentrically, and isometrically andshould include work on coordina-tion as well as strengthening.12

Anterior CruciateLigament Injury

A high incidence of ACL injurieshas been reported in women’sgymnastics, team handball, volley-ball, and alpine skiing. A recentreview in which the injury surveil-lance system of the NationalCollegiate Athletic Association(NCAA) was used showed thatfemale basketball players experi-enced four times as many ACLinjuries as their male counterpartsand that female soccer players hadtwice the injury rate of their malecounterparts.13 No conclusive evi-dence exists to explain the genderdifferences in this injury. Inter-condylar notch anatomy has beenimplicated as an etiologic factor,14

as have skill and experience level,muscle strength and coordination,limb alignment, joint laxity, andshoe-surface friction.


The proportion of ACL tearsthat are due to a noncontact mecha-nism is higher in the female athlete.The physical examination in theacute setting reveals a hemarthro-sis. Anterior knee laxity may ormay not be demonstrable, depend-ing on the status of the secondarystatic restraints, the degree ofswelling, guarding by the patient,and the experience of the examiner.

Additional StudiesRadiographs occasionally dem-

onstrate a small fracture off theposterolateral tibial plateau or atibial eminence fracture. If the diag-nosis of additional intra-articular in-juries would change the immediatetreatment plan, MR imaging is rec-ommended to look for associatedtorn menisci and osteochondralinjuries.

TreatmentTreatment is the same for male

and female athletes. However,there are certain features of thefemale knee that may need specialconsideration when contemplatingACL reconstruction. In a smallknee, a large graft carries with itthe potential complication of patel-lar donor-site fracture, graft im-pingement, and diminished integrityof the remaining patellar tendon.When patellofemoral pain predatesinjury to the ACL, one should con-sider the use of hamstring auto-graft or patellar tendon allograft.The extensor mechanism disrup-tion that occurs from harvesting apatellar tendon autograft may pro-duce deficits in quadriceps strengthand increased patellofemoral dys-function,15 although these prob-lems are usually reversible and ofshort duration with adequate post-operative rehabilitation. Finally,avoiding postoperative stretchingof the graft in a knee with a preex-isting recurvatum deformity greaterthan 15 degrees is important for asuccessful outcome.

Patellofemoral Pain

Patellofemoral pain syndrome isoften associated with variations inlimb alignment, including in-creased anteversion of the femoralneck, external tibial torsion, andpronation of the foot. This limbalignment is also associated with

an increased Q angle, patella alta,and generalized ligamentous laxity.Although these anatomic variationsare not limited to women, they areseen more commonly in the femalepopulation.


Patients generally complain ofanterior peripatellar knee pain thatis aggravated by squatting. Sittingfor long periods of time with theknees flexed may be uncomfortable(theatre sign). In addition to look-ing for the variations in limb align-ment noted above, one should tryto differentiate the pain caused byperiretinacular tightness from thatcaused by a mobile patella thatmay have an irregular articularsurface. Patients with retinaculartightness demonstrate less than onequadrant of medial mobility or lessthan 15 degrees of lateral patellartilt on passive testing (raising thelateral border of the patella to thehorizontal plane or slightlybeyond). Excessive lateral patellarcompression can lead to overloadof the lateral bone, cartilage, or softtissues. Patients with this syn-drome often present with lateralknee discomfort near the lateralfemoral condyle and the distalinsertion of the iliotibial band. Thissyndrome is frequently associatedwith radiographically demonstratedpatellar tilt.16

Additional StudiesAxial views of the patellofemoral

joint may reveal normal alignment,lateral patellar tilt, or lateral patellarsubluxation. The diagnosis of patel-lar malalignment requires docu-mentation of the changing relation-ships of the patella to the trochleargroove in all planes of kneemotion.17 Kinematic MR imagingstudies can now reveal patellartracking patterns in real time.18

However, we still do not have good



correlation between physical exami-nation features, images, and patientpain and function.

TreatmentQuadriceps strengthening is the

cornerstone of treatment for patel-lofemoral disorders and shouldemphasize closed-kinetic-chainexercises. These approximate thetibial rotation and subtalar motionassociated with functional activities.Eccentric quadriceps activity isimportant in the decelerator func-tion of the extensor mechanism andis important to emphasize duringrehabilitation. However, it mayincrease symptoms in patients withpatellar articular-surface abnormali-ties. In a young woman, musclestrength and coordination may needto “catch up” to bone maturation.

Because patellofemoral functionis an integral component of limbfunction, any important anatomicvariations in the joints proximaland distal to the knee should beaddressed with the use of flexibilityand strengthening exercises andpossibly orthotic devices. Al-though orthoses are commonlyused to control the overpronatedfoot in patients with patellofemoralpain, few controlled studies ontheir effectiveness have been per-formed with the use of standard-ized outcome measures.

McConnell taping alters glide,tilt, and rotation of the patella whenaugmented with muscle strength-ening.19 Although McConnell’ssuccess rate of 96% has not beenduplicated, taping can help relievepain during strengthening exercises.Use of knee sleeves is controversial,but may be an adjunct to quadriceps-strengthening exercises when sub-luxation is present.

Electrical muscle stimulation isuseful as an early adjunct to muscle-strengthening exercises when theseproduce patellar pain. The electric-ity produces muscle contractions

with a resultant load across thepatellofemoral joint that is lowerthan that produced by quadricepsexercises.20

Educating the patient is critical.Important recommendations are tofrequently alter positions of theflexed knee while sitting, to takebreaks to straighten the knee whendriving, and to avoid full squats.

The indications for surgicalintervention for patellofemoral painare narrow. Surgical realignmenttechniques are most likely to benefitpatients with patellofemoral paindue to lateral subluxation and/orlateral tilt. Patellar shaving isappropriate when combined withsurgical intervention to unloadstressed areas of patellar articularcartilage. Lateral retinacular releaseis helpful in relieving tilt but incon-sistent in controlling subluxation.Surgical correction of subluxationrequires a medial procedure, suchas imbrication of the medial retinac-ulum to realign the patella.

Medial advancement of the tib-ial tubercle is most appropriate inthe patient with recurrent patellarinstability when the Q angle isincreased, whether or not malalign-ment is demonstrated radiographi-cally.16 Reduction of the Q angle to10 degrees is advocated. Thetubercle sulcus angle (sitting Qangle) should approach 0 degrees.

The role of tibial or femoralosteotomy in the treatment ofpatellofemoral pain is less clear. Itis recognized that genu valgumand increased femoral anteversioncontribute to patellar trackingabnormalities and instability.When the relationship between thepatella and the trochlear groove isnormal and the patella is stable butlimb malalignment is present, anextra-articular long-bone osteoto-my may decrease patellofemoralpain. When malalignment and in-stability are both present, a patella-stabilizing operation and an oste-

otomy should be considered. Thesecan be done simultaneously or asseparate procedures.

Bunions

Bunions are nine times more com-mon in women than in men, due toa combination of hereditary predis-position, poorly fitting shoes, liga-mentous laxity, and overpronation.The last can be the result of liga-mentous laxity or compensation forincreased femoral anteversion.Overpronation causes increasedvalgus stress on the great toe dur-ing the push-off phase of gait.


The pain from a bunion is ini-tially limited to the prominent tis-sue at the medial side of the firstmetatarsal head. Subsequently, asthe great toe drifts into more val-gus deviation and bears lessweight, pain can shift to the headsof the second and third metatarsals.Dorsal subluxation of the secondtoe can cause corns and deformitiesof the interphalangeal joint as well.

Additional StudiesStanding anteroposterior and

lateral radiographs of both feetshould be taken. One should notethe first and second intermetatarsalangle as well as the relative lengthsof these metatarsals. The degree ofhallux valgus should also be noted.Any metatarsophalangeal arthritisor subluxation of the sesamoidsshould be recorded.

TreatmentIn the dancer or the athlete, par-

ticularly the running athlete or gym-nast, the treatment of bunionsrequires attention to certain princi-ples. The goal of treatment is to pro-duce a stable correction while main-taining adequate range of motion of

The Female Athlete


the first metatarsophalangeal joint.It is particularly important in ath-letes not to interfere with the sesa-moids and to avoid dorsal or plantardisplacement of the first metatarsalhead, so that the first metatarsal con-tinues to bear weight appropriately.A chevron osteotomy and bunion-ectomy have been reported to besuccessful in elite female runners.21

However, bunionectomies shouldnot be performed during the activecareer of dancers or athletes whorequire extreme dorsiflexion of thefirst metatarsophalangeal joint, suchas sprinters and gymnasts. Theyshould be treated symptomaticallyuntil their performing or sportscareers are over.

The importance of properly fit-ting shoes cannot be overempha-sized. Many women have a wideforefoot and a narrow hindfoot andas a result have problems with fit.Women tend to buy shoes so as toavoid slipping at the heel and mayend up with a shoe that is too nar-row in the forefoot. Instead, a shoewith a combination last or a shoewith a variable eyelet pattern thatallows adjustable lacing for theforefoot and hindfoot will providethe necessary width in the forefootand a snug fit in the hindfoot.22

The forefoot should be wellpadded. Orthoses including medialsupport to decrease the valgus stressand a metatarsal pad to decrease theweight borne by the second andthird metatarsal heads may be help-ful. A deep toe box may alleviatesymptoms due to subluxation of thesecond toe on top of the first. Ashoe with a straighter last will alsohelp to control pronation.

Problems Associated Withthe Menstrual Cycle

Relative Iron DeficiencyThere are no known negative

effects of the menstrual cycle on the

female athlete. Medals have beenwon and records broken by womenin every phase of the menstrualcycle. However, primarily becauseof menstrual blood loss, manyfemale athletes with normal hemat-ocrits are relatively iron-deficient,as manifested by low ferritin levels.Restoring iron stores to normal hasbeen associated with improvedperformance.

Athletic AmenorrheaExcluding pregnant women,

amenorrhea is present in up to 20%of vigorously exercising women. Itsprevalence may reach as high as50% in elite runners and profession-al ballet dancers.23 Amenorrheicathletes are more likely to havebegun training at an earlier age thannormally menstruating athletes.Although “athletic amenorrhea”was first thought to be due to aninsufficient amount of body fat, ath-letes who have normal body fat butwho consume less than the numberof calories needed for intensivesports training also become amenor-rheic. Women participating in en-durance sports in particular find itdifficult to consume sufficient calo-ries. Some athletes try to controlweight by excessive vigorous exer-cising; others starve themselves pur-posefully. Attempting to impose auniversal body shape stereotypic fora specific sport encourages eatingdisorders, ranging from a preoccu-pation with food and body image tomore severe problems, such asanorexia nervosa and bulimia.

Amenorrhea can be classified aseither primary or secondary. Pri-mary amenorrhea is defined as nomenstrual bleeding by the age of16. Secondary amenorrhea is de-fined as no menstrual cycles in a 6-month period in a woman who hashad at least one episode of men-strual bleeding. The most commoncause of amenorrhea is pregnancy.Amenorrhea also can be due to

structural abnormalities of thereproductive tract or hormonalabnormalities.24 Athletic amenor-rhea is thought to be a form ofhypothalamic amenorrhea in whichpulsatile gonadotropin-releasinghormone (GnRH) is deficient,absent, or inappropriately secreted.The neurohormones that modulateGnRH can be affected by psycho-logical or physical stress, tumors,and congenital anomalies. Par-ticipation in endurance sportsresults in increases in serum levelsof endogenous opioids, cortisol,melatonin, and dopamine, which inturn suppress the frequency andamplitude of GnRH pulses.25,26

Resumption of normal menstrualcycles may take months or yearsafter the psychological or physicalstress is relieved. Prolonged amen-orrhea can result in osteoporosis.

The bone mineral loss seen inathletes who have had amenorrheafor more than 6 months resemblesthat seen after menopause. Al-though high-intensity exercise mayincrease bone mineral density atmaximally stressed skeletal siteseven in amenorrheic and oligomen-orrheic athletes,27 whole-body bonemineral density is significantlylower in amenorrheic athletes thanin control subjects (P<0.05).28

Sixty percent to 70% of peakbone mass in women is acquiredbefore the age of 20. If a youngfemale athlete is amenorrheic anddoes not lay down a normalamount of bone during adoles-cence, she may always have de-creased bone mass. Restoration of normal menses may retard therate of further bone loss, but thebone already lost cannot be re-placed.29

The term “female athlete triad”was coined in 1991 to describe thecomplex interplay of menstrualirregularity, disordered eating, andpremature osteoporosis seen in thefemale athlete. The factors in this



triad are interdependent and canoccur as a result of intense athleticparticipation.

History and Physical ExaminationIt is critical when evaluating the

young female athlete with a stressfracture to consider the possibilityof early osteoporosis related toamenorrhea. This is particularly sowhen a history of overuse is notforthcoming or when this is not thefirst stress fracture for the patient.

Questions about menstrual his-tory, nutritional history, and body-weight history are important forscreening. The menstrual historyshould include the age at menar-che, the frequency and duration ofmenstrual periods, the date of thelast menstrual period, and the useof hormonal therapy. The nutri-tional history should include a 24-hour recall of food intake, the usualdaily number of meals and snacks,and a list of forbidden foods (e.g.,meat or sweets). The body-weighthistory should include the highestand lowest weights since menarcheand the athlete’s satisfaction withher present weight. What does shefeel her ideal weight should be?Has she ever tried to control herweight by using vomiting, laxa-tives, or diuretics?

Additional StudiesIn amenorrheic women, it is

important to rule out pregnancyand medical problems, such as thy-roid or pituitary disorders, beforeascribing amenorrhea solely to awoman’s exercise program. Fur-ther workup is usually not in thepurview of the orthopaedic sur-geon, but may include pregnancytesting, determination of thyroidhormone and prolactin levels, anda progesterone challenge.

Bone mineral density is mostprecisely measured by dual-energyx-ray absorptiometry. This studysubjects the patient to less than 5

mrem of radiation per scan, com-pared with 20 to 60 mrem for achest radiograph and 300 mrem fordental x-ray films.

TreatmentThe team physician is in the ideal

position to screen for eating disor-ders and abnormal menses duringthe preparticipation physical exami-nation. Treatment of the female ath-lete triad requires a multidiscipli-nary approach involving physi-cians, a nutrition specialist, andoften a psychologist or psychiatrist.In the case of a high school or inter-collegiate athlete, one should alsoinvolve the athletic trainer, thecoach, and the patient’s parents. Anadequate diet must include not onlythe appropriate caloric intake butalso at least 1,500 mg of calcium perday. Treatment of more seriouslydisordered eating may require con-tracts between the physician and thepatient, specifying that the level ofsports participation is decreased orcompetition is prohibited until cer-tain goals, such as a specific weightgain, are reached. Treatment withreplacement hormones may be nec-essary to prevent further bone loss.Typically, oral contraceptives areused initially. Additional educa-tional materials about these prob-lems can be obtained from theAmerican College of Sports Medi-cine and the NCAA.

Exercise DuringPregnancy

The recommendations of theAmerican College of Obstetrics andGynecology30 regarding exerciseduring pregnancy are shown inTable 1. In summary, the pregnantathlete should try to maintain acore body temperature less than38°C. Diabetes, hypertension,multiple gestation, cervical defects,and a history of miscarriage are

contraindications to exercise dur-ing pregnancy. The pregnant ath-lete may need to omit contactsports and diving from her exercisechoices. Vigorous exercise pro-grams should not be undertakenduring pregnancy by women whoare not fit, especially during thefirst trimester. The types of exer-cises recommended should mini-mize the risk of injury whileacknowledging the patient’s prefer-ences, and should be coordinatedwith the physician managing thepatient’s pregnancy.

Researching the effects of exer-cise in pregnancy is difficult be-cause of the great variety of typesof exercise, exercise intensities, anddurations, which may have verydifferent maternal and fetal effects.There is contradictory evidenceconcerning the influence of exerciseon the onset of labor, the course oflabor, and fetal growth. Neverthe-less, most fit women with normalpregnancies may continue theirregular program of exercise with-out having an adverse effect onmost aspects of labor and fetalgrowth.

Many of the concerns related toexercise during pregnancy focus onthe safety of the fetus rather thanproblems affecting the athlete her-self. The primary concerns are fetalhyperthermia and the risk of neural-tube defects, insufficient placentalblood flow, and inadequate glucoseavailability for the fetus.31

Temperature elevation is propor-tional to exercise intensity. A well-conditioned athlete can dissipateheat through sweating, but a poorlytrained athlete is more likely tobecome hyperthermic. Dehydrationand hot, humid environmental con-ditions will increase the likelihoodof elevated body temperature dur-ing exercise. Intense training canraise rectal temperature above thelevel found to be teratogenic insheep (39.2°C). However, no pro-

The Female Athlete


spective studies have demonstratedtemperature elevation to be a terato-gen in humans.

Although plasma volume ex-pansion occurs as a result of bothexercise and pregnancy and mayhelp maintain uterine blood flow,prolonged exercise also decreasessplanchnic blood flow to 40% to50% of resting levels. Since theuterine circulation is part of thesplanchnic bed, there are concerns,again based on observations insheep, about decreased blood flowto the placenta during exercise.

Glucose is a major fetal fuel, andits availability is a function of glu-cose levels in maternal blood.Glucose utilization by the musclesduring exercise will decrease levelsof circulating glucose and maylimit fetal glucose availability.

Musculoskeletal problems in thephysically active pregnant womanare related to weight gain, ligamen-tous relaxation, lordosis, andchange in the center of gravity.Near term, running sports become

more difficult as the weight andbulk of the abdomen increase. Inthe third trimester, exercise inwater is advocated because thebuoyant effect reduces the stress ofweight bearing. There is concernthat impact sports and sportsrequiring a lot of torque may causemembrane rupture, placental sepa-ration, umbilical cord entangle-ment, or direct fetal injury. Theabsolute intensity of weight-bearingexercise also should decrease aspregnancy progresses because theoxygen demands during such exer-cise increase during pregnancy. Itis difficult to judge the intensity ofmaternal exercise with the usualcriterion of heart rate because preg-nancy increases maternal blood vol-ume, heart rate, and cardiac output.


One of the most common com-plaints of pregnant women, partic-ularly in the second or third tri-mester, is back pain. It may or may

not have a radicular componentand is normally aggravated byactivities in the standing position.Changes in sensory, motor, anddeep tendon reflexes are rarely pres-ent. Most commonly, a Patrick(flexion, abduction, external rota-tion, and extension, or “fabere”) testwill produce pain consistent withstrain in the sacroiliac ligaments.

TreatmentBack pain can be decreased by

switching from standing activities,such as running and dancing, tositting activities, such as rowingand bicycling. Abdominal supportstraps may provide symptomaticrelief, as will pelvic tilts and “angrycat” exercises (performed whilefacing down on hands and knees,producing lumbar kyphosis).

Summary

Female athletes are now highly visi-ble and very successful. More girlsand women are becoming and stay-ing active than ever before.Appropriate conditioning programsare important, particularly indecreasing the incidence of overuseinjuries. These programs are bestaccepted if they are part of a sport-specific training program. A goodconditioning program shouldinclude both anaerobic and aerobicexercise and both strengthening andstretching exercises. Special empha-sis should be placed on adequatenutritional intake. Also important isabdominal strengthening, with par-ticular attention to decreasing exces-sive lumbar lordosis. An aspect ofconditioning that should be modi-fied for the female athlete is quadri-ceps strength training. Becausepatellofemoral problems are soprevalent in women, female athletesshould avoid squats and full-arc iso-tonic knee-extension exercises inwhich the tibia is loaded.



Table 1Abridged ACOG Recommendations for Exercise During Pregnancy30

• Regular exercise (at least three times per week) is preferable to intermittent activity.

• Women should avoid exercise in the supine position after the first trimester.

• Pregnant women should stop exercising when fatigued and should not exercise to exhaustion.

• Exercise in which loss of balance could be detrimental to maternal or fetal well-being is contraindicated. Any type of exercise involving the potential for even mild abdominal trauma should be avoided.

• Pregnancy requires an additional 300 kcal/d to maintain metabolic homeo-stasis. Women who exercise during pregnancy should be particularly careful to ensure an adequate diet.

• Pregnant women who exercise in the first trimester should allow for heat dissipation by ensuring adequate hydration, wearing appropriate clothing, and maintaining optimal environmental surroundings during exercise.

• Many of the physiologic and morphologic changes of pregnancy persist 4 to 6 weeks post partum. Prepregnancy exercise routines should be resumed gradually.

References

1. Bradford DS, Hu SS: Spondylolysisand spondylolisthesis, in Weinstein SL(ed): The Pediatric Spine: Principles andPractice. New York: Raven Press,1994, vol 1, pp 585-601.

2. Seitsalo S, Osterman K, Hyvarinen H,et al: Progression of spondylolisthesisin children and adolescents: A long-term follow-up of 272 patients. Spine1991;16:417-421.

3. Boxall D, Bradford DS, Winter RB, etal: Management of severe spondy-lolisthesis in children and adolescents.J Bone Joint Surg Am 1979;61:479-495.

4. van den Oever M, Merrick MV, ScottJH: Bone scintigraphy in symptomaticspondylolysis. J Bone Joint Surg Br1987;69:453-456.

5. Pizzutillo PD, Hummer CD III: Non-operative treatment for painful adoles-cent spondylolysis or spondylolisthe-sis. J Pediatr Orthop 1989;9:538-540.

6. Pizzutillo PD, Mirenda W, MacEwenGD: Posterolateral fusion for spondy-lolisthesis in adolescence. J PediatrOrthop 1986;6:311-316.

7. Frennered AK, Danielson BI, Nachem-son AL, et al: Midterm follow-up ofyoung patients fused in situ forspondylolisthesis. Spine 1991;16:409-416.

8. Winter M, Jani L: Results of screwosteosynthesis in spondylolysis andlow-grade spondylolisthesis. ArchOrthop Trauma Surg 1989;108:96-99.

9. Shin AY, Morin WD, Gorman JD, et al:The superiority of magnetic resonanceimaging in differentiating the cause ofhip pain in endurance athletes. Am JSports Med 1996;24:168-176.

10. Hajek MR, Noble HB: Stress fracturesof the femoral neck in joggers: Casereports and review of the literatureAm J Sports Med 1982;10:112-116.

11. Wallace K: Female pelvic floor func-tions, dysfunctions, and behavioral

approaches to treatment. Clin SportsMed 1994:13:459-481.

12. Nygaard I: Prevention of exerciseincontinence with mechanical devices.J Reprod Med 1995;40:89-94.

13. Arendt E, Dick R: Knee injury pat-terns among men and women in colle-giate basketball and soccer: NCAAdata and review of the literature. Am JSports Med 1995;23:694-701.

14. LaPrade RF, Burnett QM II: Femoralintercondylar notch stenosis and cor-relation to anterior cruciate ligamentinjuries: A prospective study. Am JSports Med 1994;22:198-203.

15. Sachs RA, Daniel DM, Stone ML, et al:Patellofemoral problems after anteriorcruciate ligament reconstruction. Am JSports Med 1989;17:760-765.

16. Fulkerson JP: Patellofemoral pain dis-orders: Evaluation and management. JAm Acad Orthop Surg 1994;2:124-132.

17. Nagamine R, Otani T, White SE, et al:Patellar tracking measurement in thenormal knee. J Orthop Res 1995;13:115-122.

18. Shellock FG, Mink JH, Deutsch AL, etal: Patellar tracking abnormalities:Clinical experience with kinematic MRimaging in 130 patients. Radiology1989;172:799-804.

19. McConnell J: The management ofchondromalacia patellae: A long termsolution. Aust J Physiother 1986;32:215-223.

20. Laughman RK, Youdas JW, GarrettTR, et al: Strength changes in the nor-mal quadriceps femoris muscle as aresult of electrical stimulation. PhysTher 1983;63:494-499.

21. Lillich JS, Baxter DE: Bunionectomiesand related surgery in the elite femalemiddle-distance and marathon runner.Am J Sports Med 1986;14:491-493.

22. Frey C: Shoes, in Teitz CC (ed): The Fe-male Athlete. Rosemont, Ill: American

Academy of Orthopaedic Surgeons,1997, pp 63-73.

23. Marshall LA: Clinical evaluation ofamenorrhea, in Agostini R, Titus S(eds): Medical and Orthopedic Issues ofActive and Athletic Women. Phila-delphia: Hanley & Belfus, 1994, pp152-163.

24. Loucks AB, Laughlin GA, Mortola JF, et al: Hypothalamic-pituitary-thyroidal function in eumenorrheicand amenorrheic athletes. J ClinEndocrinol Metab 1992;75:514-518.

25. Samuels MH, Sanborn CF, Hofeldt F,et al: The role of endogenous opiatesin athletic amenorrhea. Fertil Steril1991;55:507-512.

26. Laughlin GA, Loucks AB, Yen SS:Marked augmentation of nocturnalmelatonin secretion in amenorrheicathletes, but not in cycling athletes:Unaltered by opioidergic or dopamin-ergic blockade. J Clin Endocrinol Metab1991;73:1321-1326.

27. Slemenda CW, Johnston CC: Highintensity activities in young women:Site specific bone mass effects amongfemale figure skaters. Bone Miner1993;20:125-132.

28. Myburgh KH, Bachrach LK, Lewis B,et al: Low bone mineral density ataxial and appendicular sites in amen-orrheic athletes. Med Sci Sports Exerc1993;25:1197-1202.

29. Weltman A, Snead DB, Weltman JY, et al: Effects of calcium supplementa-tion on bone mineral density (BMD) inpremenopausal women runners [ab-stract]. Med Sci Sports Exerc 1992;24:S12.

30. Exercise during pregnancy and thepostpartum period. Am Coll ObstetGynecol Technical Bull 1994, No. 189.

31. Clapp JF III: A clinical approach toexercise during pregnancy. Clin SportsMed 1994:13:443-458.

The Female Athlete





























Rotator cuff disease is a commoncause of shoulder disability, partic-ularly in patients beyond the fourthdecade of life. Anterior acromio-plasty, combined with rotator cuffrepair when indicated, generallyprovides predictable pain reliefand improved function.1 However,when pain continues in spite ofsurgery for rotator cuff disease,patient management becomes morecomplicated and less predictable.It is important to recognize thatpersistent rotator cuff disease isonly one of the many potentialcauses for such pain (Table 1).Possible extrinsic causes includecervical radiculopathy; supra-scapular, long-thoracic, or spinal-accessory neuropathy; and adjacentor metastatic neoplastic disease.Potentially causative intrinsicshoulder disorders may be intra-

articular, such as osteoarthritis,adhesive capsulitis, recurrent ante-rior subluxation, and labral orbicipital tendon abnormalities, orextra-articular, such as subacromialimpingement, persistent or recur-rent rotator cuff defect, acromio-clavicular joint arthropathy, anddeltoid insufficiency. Successfulmanagement begins with an accu-rate identification of the underly-ing pathologic process responsiblefor the pain.

Evaluation

In most cases, an initial diagnosticimpression can be formulated onthe basis of the history, physicalexamination, and routine radiogra-phy. Additional studies that maybe useful include arthrography,

ultrasonography, magnetic reso-nance (MR) imaging, electromyog-raphy, and scintigraphy. Selectiveinjections into the subacromialspace and the acromioclavicularjoint can help localize the pain orquantitate how much pain is attrib-utable to each area when both areinvolved. Diagnostic arthroscopymay be useful, especially whenextrinsic disorders have beenexcluded, the previously per-formed acromioplasty has beenjudged adequate by radiographiccriteria, and the rotator cuff isintact.

Extrinsic ShoulderDisorders

It is important to recognize thatpersistent pain after rotator cuffsurgery may be the result of patho-logic processes extrinsic to the

Dr. Williams is Assistant Professor, Universityof Pennsylvania School of Medicine, andAttending Surgeon, Shoulder and ElbowService, Hospital of the University ofPennsylvania, Philadelphia.

Reprint requests: Dr. Williams, Department ofOrthopaedic Surgery, University of Pennsyl-vania, Shoulder and Elbow Service, PennMusculoskeletal Institute, 1 Cupp Pavilion,Presbyterian Medical Center, 39th and MarketStreets, Philadelphia, PA 19104.


Abstract

Persistent shoulder pain after surgery for rotator cuff disease may be caused byconditions that are either extrinsic or intrinsic to the shoulder. Extrinsic causesof persistent shoulder pain include cervical radiculopathy, suprascapular neu-ropathy, abnormalities of scapular rotation (due to long-thoracic or spinal-accessory neuropathy), and adjacent or metastatic neoplasms. Causes of persis-tent pain that are intrinsic to the shoulder include both intra-articular condi-tions (e.g., glenohumeral osteoarthritis, adhesive capsulitis, recurrent anteriorsubluxation, and labral and bicipital tendon abnormalities) and extra-articularconditions (e.g., persistent subacromial impingement, persistent or recurrentrotator cuff defects, acromioclavicular arthropathy, and deltoid muscle deficien-cy). Successful management requires an accurate diagnosis, maximal rehabili-tation, judicious use of surgical intervention, and a well-motivated patient. Theresults of revision surgery in patients with persistent subacromial impinge-ment, with or without an intact cuff, are inferior to reported results after prima-ry acromioplasty or rotator cuff repair.


Painful Shoulder After Surgery for Rotator Cuff Disease

Gerald R. Williams, Jr, MD

shoulder. In addition, an extrinsiccause of persistent pain (e.g., cervi-cal radiculopathy) may coexistwith an intrinsic cause (e.g., recur-rent rotator cuff defect), in whichcase diagnostic injection into thesubacromial space may help distin-guish between the intrinsic andextrinsic components of the pain.

When an extrinsic cause for thepersistent pain has been identified,treatment should be directedaccordingly.

Of the extrinsic causes of persis-tent shoulder pain, cervical radicu-lopathy involving the fifth or sixthcervical root is perhaps the mostcommon. The symptoms of neckpain accompanied by radiation intothe upper extremity, numbness, orparesthesias suggest this diagnosis.Routine radiography may revealcervical spondylosis or neuralforaminal encroachment. If indi-cated, MR imaging of the cervicalspine and electromyography mayconfirm the diagnosis.

Long-thoracic and spinal-acces-sory neuropathies result in scapularwinging and poor scapular rotationduring overhead elevation.Secondary impingement symptomsmay develop as scapular rotationlags behind glenohumeral eleva-tion. Although true scapular wing-ing is an uncommon cause of per-sistent pain after rotator cuff sur-gery, many patients will exhibitvarying degrees of scapulothoracicdysfunction. Scapulothoracic andscapulohumeral rhythm should beobserved in all patients with persis-tent symptoms after acromioplastyor cuff repair. In patients with se-vere scapular dysfunction associat-ed with winging, electromyographymay confirm the neurologic lesion.

Suprascapular neuropathy mayalso result in impingement-likesymptoms because of the posteriorcuff weakness that results fromchronic nerve compression. Patientspresent with severe atrophy of eitherthe supraspinatus and infraspinatusor the infraspinatus alone. This isassociated with weakness of externalrotation with the arm at the side.Electromyography is helpful in con-firming the diagnosis and localizingthe site of compression to the infra-spinatus alone or to both thesupraspinatus and the infraspinatus.Magnetic resonance imaging mayreveal a ganglion cyst compressingthe suprascapular nerve (Fig. 1).

Neoplastic processes are a veryrare but devastating cause of per-sistent shoulder pain after rotatorcuff surgery. The apical lung fieldsshould always be inspected onshoulder radiographs, because api-cal lung tumors (i.e., Pancoasttumors) cause referred shoulderpain through extension to thebrachial plexus or cervical roots. Ifa lung mass is suspected, appropri-ate chest radiographs and medicalconsultation are indicated. Persis-tent pain may also be caused bydirect involvement of the shoulderby a neoplastic process. Magneticresonance imaging may be used tofurther characterize masses orunusual prominences discoveredon physical examination (Fig. 2).

Painful Shoulder After Rotator Cuff Surgery


Table 1Causes of Persistent ShoulderPain After Rotator Cuff Surgery

Extrinsic shoulder pathologyBrachial plexopathyCervical radiculopathyLong-thoracic neuropathyNeoplasmReflex sympathetic dystrophySpinal-accessory neuropathySuprascapular neuropathyThoracic outlet syndrome

Intrinsic shoulder pathologyIntra-articular

Adhesive capsulitisArticular cartilage defectBicipital tendinitisInstabilityLabral tearsOsteoarthritis

Extra-articularAcromioclavicular arthropathyDeltoid insufficiencyRotator cuff defectSubacromial impingement

Fig. 1 Left, Severe atrophyof the supraspinatus andinfraspinatus muscles in apatient with continued painafter arthroscopic acromio-plasty. Right, MR imagedepicts a ganglion cystcompressing the supra-scapular nerve.

Intrinsic ShoulderDisorders

Causes of persistent pain that areintrinsic to the shoulder includeboth intra-articular conditions(e.g., glenohumeral osteoarthritis,adhesive capsulitis, recurrent ante-rior subluxation, and labral andbicipital tendon abnormalities) andextra-articular conditions (e.g.,persistent subacromial impinge-ment, persistent or recurrent rota-tor cuff defects, acromioclaviculararthropathy, and deltoid muscledeficiency).

Intra-articular Causes ofPersistent Pain

Unrecognized glenohumeral disor-ders may be responsible for persis-tent postsurgical shoulder pain.Intra-articular causation should besuspected when postoperativeradiographs reveal adequate de-compression of the supraspinatusoutlet, and the acromioclavicularjoint is asymptomatic.

Articular CartilageAbnormalities

Glenohumeral osteoarticulardisease may be a cause of persis-tent pain in at least two circum-

stances: (1) unrecognized orunderappreciated preoperativeosteoarthritis and (2) cuff teararthropathy, or Milwaukee shoul-der syndrome. Primary gleno-humeral osteoarthritis is character-ized by subchondral sclerosis andcyst formation, glenohumeral joint-space narrowing and osteophyteformation, asymmetric posteriorglenoid wear, and an intact orrepairable rotator cuff.2 The man-agement of primary osteoarthritisdoes not differ substantiallywhether or not there has been priorimpingement or rotator cuff sur-gery.

Cuff tear arthropathy is charac-terized by destruction of the gleno-humeral articular surfaces, accom-panied by chronic, massive rotatorcuff insufficiency and proximalhumeral migration, that persists orrecurs in spite of one or more pre-vious attempts at cuff repair.3

Persistent pain may be improvedby humeral hemiarthroplasty.4,5

Functional improvement is lesspredictable than pain relief, espe-cially if the coracoacromial liga-ment was sacrificed during previ-ous cuff repair.

Traumatic articular cartilagedefects of the humerus and glenoidmay cause persistent shoulder painin the absence of generalized artic-

ular degeneration. A history of asingle traumatic event is oftenelicited. Examination may revealpainful glenohumeral crepitus dur-ing glenohumeral rotation. Radio-graphs and MR images are oftennormal. In this circumstance, diag-nostic arthroscopy may be neces-sary to confirm a humeral or gle-noid articular defect (Fig. 3).

Adhesive CapsulitisThe hallmark of capsular con-

tracture or adhesive capsulitis is asymmetric decrease in both activeand passive range of motion, whichcan be localized or can involve allplanes of motion. Localized poste-rior capsular contracture is com-mon with subacromial impinge-ment syndrome and is character-ized not only by limited elevationbut also by decreased cross-bodyadduction and internal rotation,both of which are more pronouncedwith the arm at 90 degrees of eleva-tion in or anterior to the scapularplane. The presence of localizedposterior capsular contracture post-operatively is a sign of an incom-pletely rehabilitated shoulder andcan be a factor contributing to con-tinued pain and disability. Gener-alized capsular contracture is lesscommon with primary rotator cuffdisease or subacromial impinge-ment syndrome than localized pos-terior contracture. It is character-ized by loss of motion in all planes(especially passive external rotationwith the arm at the side) and is animportant source of persistent painand disability after surgery for rota-tor cuff disease.

The initial management of adhe-sive capsulitis consists of physio-therapy for joint mobilization andcapsular stretching. If motion can-not be restored through the use ofnonoperative joint-mobilizationtechniques, then closed manipula-tion or surgical capsular release isindicated. Postoperative frozen



Fig. 2 Patient had persis-tent pain associated with atender mass in the region ofthe trapezius after arthro-scopic acromioplasty. MRimaging revealed a soft-tissue mass that proved tobe metastatic carcinomafrom the lung.

shoulder is often unresponsive toclosed manipulation. Traditionally,surgical capsular release was per-formed through an anterior del-topectoral approach in combinationwith subscapularis lengthening.6

Arthroscopic capsular release hasrecently been reported as an alterna-tive,7,8 but this procedure requiresadvanced arthroscopic surgicalskills and may be contraindicated inthe presence of extra-articular adhe-sions.

Recurrent Anterior SubluxationIn patients less than 40 years of

age, particularly those who engagein sports involving overheadmotion, there is an overlap be-tween rotator cuff overuse andrecurrent anterior subluxation.9

Young patients with persistentshoulder pain after acromioplastymay be experiencing secondary im-pingement symptoms as a result ofsubtle anterior subluxation. Theymay report a forceful abduction–external rotation injury, a distaltraction injury, or “dead arm”symptoms while throwing.

Examination may reveal in-creased passive external rotationwith the arm at 90 degrees of ele-vation in the scapular plane, un-derlying multidirectional laxity orgeneralized ligamentous laxity, or

a positive relocation test. Radio-graphic evaluation should includespecialized views such as the api-cal oblique or Garth view,10 theWest Point view,11 and the Strykernotch view.12 These may demon-strate small Hill-Sachs defects andcalcification or fracture of the gle-noid rim consistent with recurrentposttraumatic anterior subluxation(Fig. 4).

Treatment includes activitymodification and strengtheningexercises for the rotator cuff, del-toid, and scapular stabilizers. If

this treatment fails, surgical stabi-lization may be considered.

Labral or Bicipital TendonAbnormalities

The tendon of the long head ofthe biceps traverses the bicipitalgroove, enters the glenohumeraljoint slightly anterior to the supra-spinatus insertion, becomes conflu-ent with the superior labrum, andattaches to the supraglenoid tuber-cle. Because of its course, thebiceps tendon may become in-volved in the subacromial impinge-



Fig. 3 Arthroscopic imagesof patients with continuedpain after open acromio-plasty and rotator cuffrepair. Left, One patienthad an articular defect ofthe anterior glenoid. Right,Other patient had an articu-lar defect of the humeralhead.

A B

Fig. 4 A, Standing anteroposterior 30-degree tilt radiograph of a patient with continuedpain after two arthroscopic acromioplasties and one distal clavicle excision. Physicalexamination findings were consistent with anterior subluxation. B, Stryker notch viewrevealed calcification at the inferior glenoid margin.

ment process.1,13 In addition, attri-tional changes to the tendon withinthe groove, primary biceps tendini-tis, and anterior-to-posteriorlesions of the superior labrum(“SLAP” lesions) may result in per-sistent symptoms after surgery forimpingement syndrome.

The physical findings are non-specific but may include painfulresisted forearm supination withthe elbow at 90 degrees of flexion.Diagnostic arthroscopy allowsvisualization of the superior la-brum and the biceps tendon. Theextra-articular portion of the ten-don within the bicipital groove canbe visualized by advancing the ten-don into the joint with the assis-tance of a probe or other instru-ment placed through an anteriorportal (Fig. 5). Treatment optionsinclude labral repair, labral de-bridement, and biceps tenodesis.

Extra-articular Causes ofPersistent Pain

Persistent SubacromialImpingement

Insufficient supraspinatus outletdecompression may result fromresidual anterior acromial spur-ring,14-17 regrowth of bone or sub-acromial calcification,18 inferiorprojecting acromioclavicular osteo-phytes,13 and persistence orregrowth of the coracoacromial lig-ament.16,17 Persistent impingementsyndrome related to residualsupraspinatus outlet narrowing is acommon cause of continued shoul-der pain after surgery for rotatorcuff disease and has been reportedin 18% to 79% of patients withfailed acromioplasty.14-17

Physical examination reveals apositive impingement sign and theimpingement reinforcement sign(i.e., Hawkins, or abduction inter-nal rotation [“ABIR”], sign).Substantial reduction in the pain

associated with these maneuversafter subacromial injection of lido-caine (i.e., a positive impingementtest) helps to confirm the presenceof continued subacromial impinge-ment.1 Radiography should in-clude a supraspinatus outlet view19

and a 30-degree caudal tilt view20

to evaluate for continued anterioracromial spurring and a Zancaview21 (standing anteroposteriorview with 15- to 30-degree cephalictilt) to visualize any inferiorly pro-jecting acromioclavicular osteo-phytes (Fig. 6).

The results of revision acromio-plasty are less reliable than the re-sults of primary acromioplasty.14-17

Flugstad et al14 reported the casesof 13 patients who underwent revi-sion acromioplasty with an intactcuff. Six patients described theirshoulders as “much better”; theother 7, as “better.” Hawkins etal15 reported the cases of 51patients in whom acromioplastyhad failed. Twelve of these pa-tients underwent repeat acromio-plasty, one with a rotator cuffrepair. All 12 patients were receiv-ing workmen’s compensation.

Only 1 achieved a satisfactoryresult. Ogilvie-Harris et al16 evalu-ated 67 shoulders in 65 patientsmore than 2 years after an initialacromioplasty for impingementsyndrome without a cuff tear.Eighteen of the 65 patients under-went revision rotator cuff surgery(6 rotator cuff repairs and 12 revi-sion acromioplasties). There was agood result in 9 of the 12 patients(75%). Rockwood and Williams17

reported 67% good or excellentresults in 27 patients who under-went revision acromioplasty withan intact or repairable cuff.

Because of the inconsistent re-sults of revision acromioplasty,successful management of patientswith persistent subacromial outletnarrowing requires careful patientselection. Nonoperative manage-ment should be maximized in allcases. Repeat surgery is reservedfor patients with radiographic evi-dence of continued impingementwho obtain pain relief with sub-acromial lidocaine. In spite ofthese stringent selection criteria,the results of revision acromioplas-ty will likely not approach those ofprimary acromioplasty.

Persistent or Recurrent RotatorCuff Defect

EvaluationThe presence of a full-thickness

rotator cuff defect can be compati-ble with asymptomatic shoulderfunction.22 Furthermore, someauthors have reported high per-centages of patients with good orexcellent results after acromioplas-ty and cuff repair in spite of arthro-graphically and ultrasonographi-cally proven persistent or recurrentrotator cuff defects.23-25 Therefore,when evaluating patients with continued pain and a persistent orrecurrent rotator cuff defect afterrotator cuff repair, it is importantto eliminate other causes of persis-



Fig. 5 Arthroscopic image of severe par-tial tearing of the long head of the biceps ina patient with continued pain after openacromioplasty and cuff repair followed byopen distal clavicle excision.

tent pain before focusing on theresidual rotator cuff defect.

Physical findings are variableand depend on the size of therecurrent rotator cuff defect. Smalldefects, which primarily affect thesupraspinatus tendon, are charac-terized by an intact anterior (i.e.,subscapularis) and posterior (i.e.,infraspinatus and teres minor) rota-tor cuff force couple. The impinge-ment and impingement-reinforce-ment signs may be positive andaccompanied by subacromial crepi-tus. However, range of overheadelevation, shoulder strength, andfunction are relatively normal.

Large defects extend anteriorlyand/or posteriorly into the sub-scapularis and infraspinatus–teresminor, respectively. Posteriorextension results in weakness ofexternal rotation with the arm atthe side and the humerus in neutralrotation. If the posterior rotatorcuff insufficiency is severe enough,the patient will be unable to raisethe arm overhead, in spite of fullpassive motion.

The signs of anterior (i.e., sub-scapularis) rotator cuff insufficien-cy can be more subtle than thesigns of posterior rotator cuff insuf-

ficiency. Increased passive externalrotation with the arm at the side issuggestive of subscapularis in-volvement. Subscapularis insuffi-ciency is verified by a positive “liftoff” test.26 This test is performedby passively resting the back of thepatient’s hand against the ipsilater-al buttock and then asking thepatient to actively lift the hand offthe back and away from the bodywithout simultaneously extendingthe shoulder or the elbow (Fig. 7).This requires maximal internalrotation with the subscapularis.Inability to perform this test isindicative of subscapularis insuffi-ciency. However, pain and limita-tion of passive internal rotationmay make interpretation of this testdifficult.

Ultrasonography, arthrography,and MR imaging have all beenused to evaluate rotator cuffpathology.27-29 When there hasbeen prior surgery, the presence ofsubacromial scarring, subacromialbursal thickening, and postsurgicaltendon irregularities may compli-cate the interpretation of theimages obtained with these modali-ties. Therefore, imaging studiesmust be interpreted with caution

and correlated carefully with theoverall clinical impression. In par-ticular, MR imaging of the rotatorcuff is not as sensitive or specific asin the shoulder that has not beentreated surgically.30

Abnormalities of tendon signalintensity in the absence of alter-ations in signal morphology may



A B

Fig. 6 The 30-degree caudal-tilt radiograph (A) and the Zanca view (B) are useful adjuncts to the supraspinatus outlet and axillary viewswhen evaluating patients with continued pain after rotator cuff surgery.

Fig. 7 A patient with an intact subscapu-laris is able to lift a hand placed on the but-tock off the back and away from the bodyby maximal internal rotation withoutsimultaneously extending the shoulder orelbow.

have no clinical relevance andshould be interpreted with caution(Fig. 8, A). However, the presenceof a well-defined gap in the tendonwith synovial fluid traversing theentire thickness of the tendon intothe subacromial space is definitiveevidence of a persistent or recur-rent defect (Fig. 8, B-D). When afull-thickness defect is present, MRimaging can accurately quantitatethe size of the defect in both theanteroposterior and medial-lateralplanes and can estimate atrophy ineach of the four rotator cuff mus-cles.

TreatmentIn many patients, a persistent

cuff defect is accompanied by con-tinued supraspinatus outlet nar-rowing. DeOrio and Cofield31

reported the data on 27 patients (27shoulders) who underwent a sec-ond attempt at repair of a rotatorcuff tear. Seven patients had phys-ical findings consistent with contin-ued subacromial impingement, andonly 12 of the 27 shoulders hadundergone an anterior acromio-plasty at the time of the initialrepair. Neviaser and Neviaser32

reported on 46 cases of revisioncuff repair, in all of which repeatacromioplasty was necessary, pre-sumably because of persistentsupraspinatus outlet narrowing.Bigliani et al33 documented a 90%incidence of inadequate prioracromioplasty in their 31 patientswho underwent a repeat repair.

The reported results of revisionrotator cuff repair are inconsistentand, in general, inferior to theresults of primary cuff repair.31-33

In the study by DeOrio andCofield,31 7 of the 27 patients (26%)who underwent revision rotatorcuff repair required a third opera-tive procedure before study com-pletion and were not, therefore,included in the final results. Noneof the remaining 20 patients had

excellent results, and only 42% hadgood results. Bigliani et al33 report-ed satisfactory results in 52% of 31patients who underwent repeatrotator cuff repair. Neviaser andNeviaser32 reported on 46 revisionrotator cuff repairs and criticallyevaluated return of range ofmotion in their outcome analysis.Twenty-two patients gainedmotion (mean, 45 degrees), 22 hadno change, and 2 lost motion.

Given the relatively disappoint-ing results of revision acromioplasty

and rotator cuff repair, the mer-its of nonoperative managementshould not be overlooked. Animportant component is activitymodification, which should involveemployment, daily-living, andrecreational activities. Physio-therapy, including capsular stretch-ing and strengthening exercises forthe remaining portions of the rota-tor cuff, the deltoid, and the scapu-lar rotators, should be maximized.Revision rotator cuff repair shouldbe considered if nonoperative man-



Fig. 8 In the postoperative setting, MR imaging criteria for rotator cuff tears must bemore stringent. A, Isolated abnormal signal intensity may have no clinical relevance andshould be interpreted with caution. The size of a recurrent defect can be quantitated inboth the anteroposterior direction (B) and the medial-lateral direction (C). The presence ofa tendon signal defect traversed by synovial fluid is indicative of a recurrent defect. D,Atrophy of individual muscles can be assessed.

A B

C D

agement has failed and the patientis willing to accept the reality ofinconsistent results.

The goal of all revision rotatorcuff procedures is to achieve a sur-gical repair that ultimately heals tobone at the operative site andremains intact over the long term.Patients who achieve this goal aremost likely to experience the bestresults with regard to pain,strength, and function.34 Withsmaller, more mobile cuff tears,this goal is often attainable. Revi-sion acromioplasty and/or removalof inferior acromioclavicular osteo-phytes is performed in conjunctionwith rotator cuff repair when resid-ual supraspinatus outlet narrowingfrom anterior acromial or inferioracromioclavicular spurring exists.

The rotator cuff tears most likelyto rerupture after repair are the largetears with two- or three-tendoninvolvement, particularly in olderpatients.34 In addition, large initialtears are most likely to be difficult torepair, primarily because of poor tis-sue quality. Therefore, revision offailed repairs of large rotator cufftears is technically difficult andwould be expected to be less likelyto result in a permanently healedtendon.

The most important aspects ofsurgical technique in these difficultcases are tendon identification andmobilization. The subacromialbursa may be abnormally thick-ened and must not be mistaken forthe torn rotator cuff tendon edge.Once the retracted tendon edge hasbeen identified, it is systematicallymobilized laterally. First, thesuperficial surface of the retractedtendon is freed from any overlyingadhesions to the bursa, the spine ofthe scapula, and the deep surfaceof the posterior deltoid and trape-zius. Second, the retracted tendonedge is pulled laterally in order toidentify any contracture of thecoracohumeral ligament, which is

released if present. Finally, if nec-essary, any tenodesis effect of theunderlying capsule is addressed bystretching the posterior capsulewith an intra-articular “metal fin-ger” or by releasing the capsulesharply slightly distal to the la-brum. The mobilized tendon isthen repaired to bone on the greatertuberosity or at the anatomic neck,slightly medial to the anatomicinsertion site.

The subscapularis tendon shouldroutinely be inspected for partial orcomplete avulsion, especially inpatients with a positive preopera-tive lift-off test. This can be accom-plished through a standard superi-or incision by flexing the humerusto bring the subscapularis into thewound. Alternatively, if preopera-tive evaluation indicates an isolat-ed subscapularis injury, an anteriordeltopectoral approach can be uti-lized. In either case, the subscapu-laris tendon is mobilized laterallyand repaired to bone. Sufficientmobilization to allow repair mayrequire release of the underlyinganterior capsule.

Continued shoulder pain associ-ated with a failed previous cuffrepair in an irreparable persistentrotator cuff defect is a potentiallydifficult problem, which may nothave a good solution. The interac-tion between the deltoid, the rota-tor cuff, and the coracoacromialarch (anterior acromion, distal clav-icle, and coracoacromial ligament)during elevation of the arm is com-plex and not completely under-stood. In the presence of an intactand normally functioning rotatorcuff mechanism, the potential prox-imal humeral migration generatedby deltoid contraction is resisted bythe rotator cuff; the humerusremains relatively centered on theglenoid fossa, and normal over-head elevation is accomplished.35,36

Under these circumstances, the rel-ative role of the coracoacromial

arch as a humeral-head contain-ment mechanism is minor.

In some cases involving irrep-arable rotator cuff tears, enoughanterior and posterior rotator cufffunction remains to effectivelyresist proximal humeral migrationduring deltoid contraction. Thehumeral head again remains rela-tively centered, and overhead ele-vation is normal or near normal inrange but may be weak. The rotatorcuff function lost to the irreparablecuff defect is “compensated” for bythe remaining balanced anteriorand posterior rotator cuff forcecouple.37 The degree to which thecoracoacromial arch functions as ahumeral-head containment mecha-nism is variable and is probablydependent on the amount of anteri-or and posterior rotator cuff re-maining.

If the persistent rotator cuffdefect is too large, the associatedloss of rotator cuff function cannotbe compensated for. In this rela-tively “uncompensated” shoulder,the remaining anterior and posteri-or rotator cuff mechanism is unableto effectively resist the proximalhumeral migration associated withdeltoid contraction. Consequently,the coracoacromial arch becomesmore important as a humeral-headcontainment mechanism.38,39 In-competence of the coracoacromialarch due to prior acromioplastyand coracoacromial ligament resec-tion combined with a poorly com-pensated or uncompensated rotatorcuff defect may result in severecompromise of overhead shoulderfunction.39

Surgical treatment of a patientwith persistent pain and an irrep-arable rotator cuff defect is poten-tially difficult and is dependent onthe supposed cause of the contin-ued pain as well as the size of thedefect. In the presence of contin-ued supraspinatus outlet narrow-ing, as documented on supraspina-



tus outlet and 30-degree caudal-tiltradiographs, persistent pain is like-ly to be the result of continued sub-acromial impingement. If pain isrelieved with subacromial lido-caine and the irreparable rotatorcuff defect is compensated for, asevidenced by intact overhead func-tion and relative preservation ofthe acromiohumeral interval (i.e.,an acromiohumeral interval of 7mm or greater), repeat subacromialdecompression without repairshould provide acceptable painrelief while preserving overheadfunction.40

Rockwood et al40 have reportedsatisfactory results with subacromialdecompression and partial cuffdebridement in patients with sub-acromial impingement syndromeassociated with chronic irreparablerotator cuff defects. The resultswere less satisfactory in patientswho had undergone prior rotatorcuff surgery. However, many ofthese patients also had iatrogenicdeltoid insufficiency. Althoughmore complicated surgical optionsfor management of the irreparablecuff defect have been reported,41-47

none has been demonstrated to besuperior to debridement alone whenthe defect is well compensated.

Debridement alone for patientswith persistent pain associatedwith an uncompensated irrepara-ble rotator cuff defect is unlikely toeither alleviate pain or improvefunction. If the patient is unable toactively raise the arm overheadpreoperatively, even when pain isrelieved with subacromial lido-caine, it is unlikely the ability toraise the arm overhead postopera-tively will be regained unless someof the lost anterior or, more com-monly, posterior rotator cuff func-tion can be reestablished. In fact,repeat subacromial decompressionand partial rotator cuff debride-ment may further compromiseshoulder function by removing the

humeral-head containment provid-ed by any remaining portions ofthe acromion and coracoacromialligament.39

The painful shoulder with anuncompensated irreparable rotatorcuff defect and an incompetentcoracoacromial arch is currently aproblem without a solution. Manytechniques have been described toreconstruct massive irreparablerotator cuff defects.41-47 However,few of them have the potential torestore lost rotator cuff function, asopposed to merely filling thedefect. Reconstruction of the supe-rior defect with autograft fascialata, allograft fascia lata or rotatorcuff, or prosthetic material mayprovide a tenodesis effect, but isnot likely to restore function toseverely atrophic rotator cuff mus-culature.44,45,47 Superior transposi-tion of the teres minor and/or thesubscapularis has the potentialadvantage of improving headdepression but has the potentialdisadvantage of destabilizing theanterior-posterior force couple.41,46

From a conceptual point of view,transfer of the latissimus dorsiinsertion into the posterosuperiorhumeral head is appealing.48 Itprovides a functional musculo-tendinous unit without sacrificingany remaining anterior or posteriorrotator cuff function. In addition,the resultant line of action providespotential head depression. Theindications for unipolar latissimusdorsi transfer continue to bedefined. The reported results havebeen variable and seem to be bestwhen the subscapularis is not alsodeficient.

The role of coracoacromial archreconstruction in this setting hasyet to be established. Wiley39

described the use of a coracoacro-mial interpositional iliac-crest auto-graft in five patients with persistentsymptoms associated with irrep-arable rotator cuff defects and defi-

cient coracoacromial arches after afailed acromioplasty and rotatorcuff repair. The results were disap-pointing, and useful overheadfunction could not be restored. Atleast three of these patients hadanterior deltoid deficiency, whichmay have contributed to the poorpostoperative elevation. The im-portance of a functional coraco-acromial arch in patients with anuncompensated irreparable rotatorcuff defect seems clear. However,additional work is required todefine surgical techniques andindications for coracoacromial archreconstruction or repair.

Acromioclavicular JointArthropathy

Acromioclavicular arthropathyis a relatively common cause ofpersistent pain after acromioplastywith or without cuff repair.Resectional arthroplasty or distalclavicle excision is indicated if thefollowing criteria are met: (1) theacromioclavicular joint is tender topalpation and painful during cross-body adduction, (2) there is radio-graphic evidence of arthritis, and(3) temporary pain relief follows alocal intra-articular injection oflidocaine.

The optimal amount of bone tobe resected from the distal clavicleremains somewhat controversial.Displacement of the clavicle alongits longitudinal axis, toward theacromion, is primarily controlledby the trapezoid portion of thecoracoclavicular ligament.49 Withlarge displacements, the acromio-clavicular ligaments primarilyresist anteroposterior displacementof the clavicle, and the coracocla-vicular ligament (especially theconoid portion) resists superoinfe-rior displacement.49 Results of dis-tal clavicle excision may be nega-tively affected by excessive transla-tion of the distal clavicle in boththe anteroposterior and superoin-



ferior planes. Therefore, theamount of bone resected should besufficient to prevent axial compres-sion or contact between the resid-ual clavicle and the acromion, butnot so much as to compromise thecapsular and coracoclavicular liga-ments.

Resection can be performedarthroscopically or by traditionalopen techniques. Our current prac-tice in most cases is to arthroscopi-cally remove 1.0 cm of distal clavi-cle, which results in a final gap dis-tance of 1.2 to 1.5 cm.

Deltoid InsufficiencyDenervation or postoperative

detachment of the deltoid afteracromioplasty and cuff repair is a devastating complication, which is best managed by prevention (Fig. 9). The axillary nerve exits thequadrilateral space and dividesinto a posterior branch, whichinnervates the teres minor and theposterior portion of the deltoid,and an anterior branch, whichinnervates the middle and anteriordeltoid. As the anterior branchcourses from posterior to anterior,it lies approximately 4 to 5 cm dis-tal to the lateral edge of theacromion. In this position, thenerve is vulnerable to injury if thesurgical incision splits the deltoidbeyond the 4- to 5-cm “safe zone.”1

If this occurs, all portions of thedeltoid anterior to the deltoid inci-sion can be denervated, whichresults in substantial functionalimpairment. Therefore, extremecaution should be used when split-ting the deltoid in line with itsfibers, so that the length of the splitdoes not exceed 4 to 5 cm.

Postoperative deltoid detach-ment can be minimized by using adeltoid-preserving approach dur-ing acromioplasty and cuff repair.50

Once the interval between the ante-rior and middle deltoid fibers hasbeen identified, the deltoid split is

extended proximally into the delto-trapezius aponeurosis, at the ante-rior edge of the acromion. Theincision in the deltotrapeziusaponeurosis should be carefullyplaced so that it leaves a strongtendinous edge on the anterior del-toid to allow secure reattachment.Deltoid reattachment is accom-plished by intratendinous repair ofthe deltotrapezius aponeurosis,which can be supplemented bytransosseous sutures through theacromion.

If detachment of the deltoid isrecognized early in the postopera-tive period, repair is much easierand more likely to yield a satisfac-tory result than if the postoperativedetachment is discovered late,when the tendon has retracted andthe muscle has atrophied. There-fore, the deltoid repair should beroutinely inspected at each postop-erative visit. The findings associat-ed with deltoid dehiscence can besubtle. If the patient is requested togently contract the deltoid whilethe arm is supported by the exam-iner, the integrity of the deltoid ori-gin can be verified. Early postoper-ative failure of the deltoid repair is

often associated with large hema-toma formation, which shouldalways raise the index of suspicionfor possible deltoid disruption.When deltoid detachment is sus-pected, operative repair is warrant-ed. If the initial repair was nottransosseous, attempting reattach-ment to bone should be considered.Because the tissue quality is oftensuboptimal, an abduction brace orpillow may be used for protection.

The surgical management ofchronic postoperative deltoiddetachment or denervation in-cludes primary repair, local muscletransposition, and distant muscletransfer.51,52 When the defect issmall to moderate in size, primaryrepair is attempted. Complete clo-sure of larger defects may requireanterior transposition of a portionof the middle deltoid. Loss of theentire anterior deltoid due to de-nervation is a very difficult prob-lem. If the deltoid deficiency isaccompanied by a massive, poten-tially irreparable rotator cuff defectand coracoacromial arch incompe-tence, arthrodesis may be the mostprudent option. If rotator cuffintegrity has been maintained,however, bipolar transfer of thelatissimus dorsi may be indicated.52

Patients who have undergoneradical or complete acromionecto-my represent a specific subgroupof patients with postoperative del-toid insufficiency that is even moredifficult to treat than the group as awhole.53 Satisfactory results withradical acromionectomy have beenreported.54 However, when deltoiddehiscence occurs after radical orcomplete acromionectomy, absenceof the acromion makes reattach-ment of the deltoid technically dif-ficult, if not impossible. In addi-tion, radical acromionectomy, bydefinition, results in coracoacro-mial arch insufficiency. Postoper-ative deltoid detachment after radi-cal acromionectomy combined



Fig. 9 Deltoid detachment is an operativedisaster, as in this patient who underwentradical acromionectomy and sustainedpostoperative deltoid disruption.

with a persistent uncompensatedrotator cuff defect results in severefunctional disability, which is prob-ably not salvageable withoutarthrodesis. For these reasons, rad-ical acromionectomy is unpopular.

Summary

Shoulder pain that persists afterrotator cuff surgery may be the

result of many causes, both intrinsicand extrinsic to the shoulder. Ap-propriate evaluation may identify asubset of patients with intrinsicshoulder disorders amenable to sur-gical correction. When continuedpain is the result of persistent sub-acromial impingement or a persis-tent rotator cuff defect, the results ofrevision surgery are inferior to thereported results of primary acromio-plasty and cuff repair. The goals of

revision rotator cuff repair are a de-compressed supraspinatus outletand a permanently healed tendon. Ifthe rotator cuff defect is irreparablebut compensated, satisfactory resultscan be obtained with repeat subacro-mial decompression and partialrotator cuff debridement. The com-bination of an irreparable uncom-pensated rotator cuff defect andcoracoacromial arch incompetence iscurrently an unsolved problem.



References

1. Neer CS II: Anterior acromioplasty forthe chronic impingement syndrome inthe shoulder: A preliminary report. JBone Joint Surg Am 1972;54:41-50.

2. Neer CS II: Replacement arthroplastyfor glenohumeral osteoarthritis. J BoneJoint Surg Am 1974;56:1-13.

3. Neer CS II, Craig EV, Fukuda H: Cuff-tear arthropathy. J Bone Joint Surg Am1983;65:1232-1244.

4. Arntz CT, Jackins S, Matsen FA III:Prosthetic replacement of the shoulderfor the treatment of defects in the rota-tor cuff and the surface of the gleno-humeral joint. J Bone Joint Surg Am1993;75:485-491.

5. Pollock RG, Deliz ED, McIlveen SJ, etal: Prosthetic replacement in rotatorcuff deficient shoulders. J ShoulderElbow Surg 1992;1:173-186.

6. McLaughlin HL: The “frozen shoul-der.” Clin Orthop 1961;20:126-131.

7. Pollock RG, Duralde XA, Flatow EL, etal: The use of arthroscopy in the treat-ment of resistant frozen shoulder. ClinOrthop 1994;304:30-36.

8. Harryman DT II: Shoulders: Frozen andstiff. Instr Course Lect 1993;42:247-257.

9. Jobe FW, Tibone JE, Jobe CM, et al:The shoulder in sports, in RockwoodCA Jr, Matsen FA III (eds): TheShoulder. Philadelphia: WB Saunders,1990, vol 2, pp 961-990.

10. Garth WP Jr, Slappey CE, Ochs CW:Roentgenographic demonstration ofinstability of the shoulder: The apicaloblique projection—A technical note. JBone Joint Surg Am 1984;66:1450-1453.

11. Rokous JR, Feagin JA, Abbott HG:Modified axillary roentgenogram: Auseful adjunct in the diagnosis ofrecurrent instability of the shoulder.Clin Orthop 1972;82:84-86.

12. Hall RH, Isaac F, Booth CR: Dislo-cations of the shoulder with specialreference to accompanying small frac-tures. J Bone Joint Surg Am 1959;41:489-494.

13. Neer CS II: Impingement lesions. ClinOrthop 1983;173:70-77.

14. Flugstad D, Matsen FA, Larry I, et al:Failed acromioplasty: Etiology andprevention. Orthop Trans 1986;10:229.

15. Hawkins RJ, Chris T, Bokor D, et al:Failed anterior acromioplasty: Areview of 51 cases. Clin Orthop 1989;243:106-111.

16. Ogilvie-Harris DJ, Wiley AM,Sattarian J: Failed acromioplasty forimpingement syndrome. J Bone JointSurg Br 1990;72:1070-1072.

17. Rockwood CA Jr, Williams GR: Theshoulder impingement syndrome:Management of surgical treatmentfailures. Orthop Trans 1992/93;16:739-740.

18. Lazarus MD, Chansky HA, Misra S, etal: Comparison of open and arthro-scopic subacromial decompression. J Shoulder Elbow Surg 1994;3:1-11.

19. Neer CS II, Poppen NK: Supra-spinatus outlet. Orthop Trans 1987;11:234.

20. Cone RO II, Resnick D, Danzig L:Shoulder impingement syndrome:Radiographic evaluation. Radiology1984;150:29-33.

21. Zanca P: Shoulder pain: Involvementof the acromioclavicular joint (analysisof 1,000 cases.) AJR Am J Roentgenol1971;112:493-506.

22. Sher JS, Uribe JW, Posada A, et al:Abnormal findings on magnetic reso-nance images of asymptomatic shoul-ders. J Bone Joint Surg Am 1995;77:10-15.

23. Blauth W, Gartner J: Ergebnisse post-operativer Arthrographien nach Nahtrupturierter Rotatorenmanschetten.Orthopade 1991;20:262-265.

24. Calvert PT, Packer NP, Stoker DJ, et al:Arthrography of the shoulder afteroperative repair of the torn rotatorcuff. J Bone Joint Surg Br 1986;68:147-150.

25. Wulker N, Melzer C, Wirth CJ:Shoulder surgery for rotator cuff tears:Ultrasonographic 3-year follow-up of97 cases. Acta Orthop Scand 1991;62:142-147.

26. Gerber C, Krushell RJ: Isolated rup-ture of the tendon of the subscapularismuscle: Clinical features in 16 cases. JBone Joint Surg Br 1991;73:389-394.

27. Ghelman B, Goldman AB: The doublecontrast shoulder arthrogram:Evaluation of rotary cuff tears.Radiology 1977;124:251-254.

28. Mack LA, Gannon MK, Kilcoyne RF,et al: Sonographic evaluation of therotator cuff: Accuracy in patients with-out prior surgery. Clin Orthop1988;234:21-28.

29. Iannotti JP, Zlatkin MB, Esterhai JL, etal: Magnetic resonance imaging of theshoulder: Sensitivity, specificity, andpredictive value. J Bone Joint Surg Am1991;73:17-29.

30. Owen R, Iannotti JP, Kneeland B, et al:Shoulder after surgery: MR imagingwith surgical validation. Radiology1993;186:443-447.

31. DeOrio JK, Cofield RH: Results of asecond attempt at surgical repair of afailed initial rotator-cuff repair. J BoneJoint Surg Am 1984;66:563-567.

32. Neviaser RJ, Neviaser TJ: Reoperationfor failed rotator cuff repair: Analysisof 46 cases. Orthop Trans 1989;13:241.
























33. Bigliani LU, Cordasco FA, McIlveenSJ, et al: Operative treatment of failedrepairs of the rotator cuff. J Bone JointSurg Am 1992;74:1505-1515.

34. Harryman DT, Mack LA, Wang KW,et al: Repairs of the rotator cuff:Correlation of functional results withintegrity of the cuff. J Bone Joint SurgAm 1991;73:982-989.

35. Poppen NK, Walker PS: Normal andabnormal motion of the shoulder. J Bone Joint Surg Am 1976;58:195-201.

36. Kelkar R, Newton PM, Armengol J, etal: Three-dimensional kinematics ofthe glenohumeral joint during abduc-tion in the scapular plane. TransOrthop Res Soc 1993;18:136.

37. Burkhart SS: Arthroscopic treatmentof massive rotator cuff tears: Clinicalresults and biomechanical rationale.Clin Orthop 1991;267:45-56.

38. Flatow EL, Wang VM, Kelkar R, et al:The coracoacromial ligament passivelyrestrains anterosuperior humeral sub-luxation in the rotator cuff deficientshoulder. Trans Orthop Res Soc 1996;21:229.

39. Wiley AM: Superior humeral disloca-tion: A complication following decom-pression and debridement for rotatorcuff tears. Clin Orthop 1991;263:135-141.

40. Rockwood CA Jr, Williams GR Jr,Burkhead WZ Jr: Debridement of

degenerative, irreparable lesions of therotator cuff. J Bone Joint Surg Am 1995;77:857-866.

41. Cofield RH: Subscapular muscletransposition for repair of chronicrotator cuff tears. Surg Gynecol Obstet1982;154:667-672.

42. Debeyre J, Patte D, Elmelik E: Repairof ruptures of the rotator cuff of theshoulder with a note on advancementof the supraspinatus muscle. J BoneJoint Surg Br 1965;47:36-42.

43. Ha’eri GB, Wiley AM: Advancementof the supraspinatus muscle in therepair of ruptures of the rotator cuff. JBone Joint Surg Am 1981;63:232-238.

44. Nasca RJ: The use of freeze-dried allo-grafts in the management of globalrotator cuff tears. Clin Orthop 1988;228:218-226.

45. Neviaser JS, Neviaser RJ, Neviaser TJ:The repair of chronic massive rupturesof the rotator cuff of the shoulder byuse of a freeze-dried rotator cuff. JBone Joint Surg Am 1978;60:681-684.

46. Neviaser RJ, Neviaser TJ: Transfer ofthe subscapularis and teres minor formassive defects of the rotator, inBayley I, Kessel L (eds): Shoulder Sur-gery. Berlin: Springer, 1982, pp 60-63.

47. Parrish FF, Murray JA, Urquhart BA:The use of polyethylene mesh (Marlex)as an adjunct in reconstructive surgery

of the extremities. Clin Orthop 1978;137:276-286.

48. Gerber C: Latissimus dorsi transferfor the treatment of irreparable tears ofthe rotator cuff. Clin Orthop 1992;275:152-160.

49. Fukuda K, Craig EV, An KN, et al:Biomechanical study of the ligamen-tous system of the acromioclavicularjoint. J Bone Joint Surg Am 1986;68:434-440.

50. Matsen FA III, Arntz CT: Subacromialimpingement, in Rockwood CA Jr,Matsen FA III (eds): The Shoulder.Philadelphia: WB Saunders, 1990, pp623-646.

51. Groh GI, Simoni M, Rolla P, et al: Lossof the deltoid after shoulder opera-tions: An operative disaster. J ShoulderElbow Surg 1994;3:243-253.

52. Itoh Y, Sasaki T, Ishiguro T, et al:Transfer of latissimus dorsi to replacea paralysed anterior deltoid: A newtechnique using an inverted pedicledgraft. J Bone Joint Surg Br 1987;69:647-651.

53. Neer CS II, Marberry TA: On the dis-advantages of radical acromionecto-my. J Bone Joint Surg Am 1981;63:416-419.

54. Bosley RC: Total acromionectomy: Atwenty-year review. J Bone Joint SurgAm 1991;73:961-968.






















Plantar fasciitis is a common clini-cal problem. Despite this, therehas been remarkably little ad-vancement in our understandingand treatment of this annoyingcondition. The old adage, “Themore treatments available for acondition, the less effective any ofthem is,” certainly applies to plan-tar fasciitis. Although there is con-sensus that conservative treatmentis effective most of the time, thereis no agreement as to whichmodality is the most effective.Furthermore, comparison of treat-ment regimens is difficult, as manypublications deal with only a sin-gle method, which varies fromstudy to study.1-4

Anatomy andBiomechanics

The plantar fascia extends longitu-dinally along the plantar surface ofthe foot deep to the fibrofatty sub-cutaneous tissue and covers theintrinsic musculature and neuro-vascular structures. It extendsfrom the tubercles of the calcaneusproximally to the plantar aspect ofthe metatarsophalangeal joints andthe bases of the toes distally (Fig. 1).When the metatarsophalangealjoints are passively dorsiflexedduring the toe-off phase of gait, theinelastic plantar fascia is placedunder tension, stabilizing and ele-vating the arch of the foot in a

mechanism that has been com-pared to a windlass.

Most of the weight-bearing sup-port in the foot occurs in the staticstructures.5 The plantar fascia playsa dominant role, contributing a largerproportion of maintenance of thearch than the spring ligament orplantar ligaments.6 The calcanealattachment is subject to tensilestress with weight bearing and loco-motion. The proximal attachmentsite at the calcaneus is in an area ofspecialized fibrocartilaginous tissuesometimes termed an “enthesis.”7,8

This tissue has longitudinal fibersthat are strong in tension and havebeen described as vascular, inner-vated, and metabolically active.9

Tensile forces are concentrated atthis attachment site, particularly onthe medial tubercle of the calcaneus.

Etiology

The term “plantar fasciitis” impliesan inflammatory process. Wood

Dr. Gill is in private practice with MillerOrthopaedic Clinic, Charlotte, NC.

Reprint requests: Dr. Gill, Miller OrthopaedicClinic, 1001 Blythe Boulevard, Suite 200,Charlotte, NC 28203.


Abstract

Plantar fasciitis is a common cause of heel pain, which frustrates patients andpractitioners alike because of its resistance to treatment. It has been associatedwith obesity, middle age, and biomechanical abnormalities in the foot, such astight Achilles tendon, pes cavus, and pes planus. It is considered to be mostoften the result of a degenerative process at the origin of the plantar fascia at thecalcaneus. However, neurogenic and other causes of subcalcaneal pain are fre-quently cited. A combination of causative factors may be present, or the truecause may remain obscure. Although normally managed with conservativetreatment, plantar fasciitis is frequently resistant to the wide variety of treat-ments commonly used, such as nonsteroidal anti-inflammatory drugs, rest,pads, cups, splints, orthotics, corticosteroid injections, casts, physical therapy,ice, and heat. Although there is no consensus on the efficacy of any particularconservative treatment regimen, there is agreement that nonsurgical treatmentis ultimately effective in approximately 90% of patients. Since the natural his-tory of plantar fasciitis has not been established, it is unclear how much ofsymptom resolution is in fact due to the wide variety of commonly used treat-ments.


Plantar Fasciitis:Diagnosis and Conservative Management

Lowell H. Gill, MD

originally described the entity in1812, attributing the inflammationto tuberculosis.10 Pathologic stud-ies done more recently on surgical-ly removed specimens demonstratemicrotears of the fascia, collagennecrosis, angiofibroblastic hyper-plasia, and chondroid metaplasia.7These changes are consistent with achronic degenerative/reparativeprocess secondary to repetitivestress. Positive bone scans of thecalcaneus at the attachment sitereflect this chronic stress pattern(Fig. 2).11

Differential Diagnosis

There are numerous causes of sub-calcaneal heel pain.8,12 Inflam-matory arthropathies, tumors,infections, and stress fractures ofthe calcaneus may all be associatedwith pain beneath the heel. Al-

though plantar fasciitis is bilateralin as many as 20% to 30% ofpatients,5 this presentation raisesthe index of suspicion for a sys-temic cause, such as a seronegativespondyloarthropathy (e.g., anky-

losing spondylitis) or Reiter’s syn-drome. A systemic arthritic disor-der subsequently develops in up to16% of patients who present withsubcalcaneal pain syndrome.8

Neuropathies, such as those sec-ondary to diabetes and alcoholism,and lumbar spine disorders occa-sionally cause pain in the foot,including the heel, as can vascularinsufficiency. Metabolic condi-tions, such as osteomalacia, andother conditions, such as Paget’sdisease and sickle cell disease, mayalso be associated with inferior heelpain.12

Bordelon8 considers calcanealapophysitis to be a distinct cause ofsubcalcaneal pain, even in adults.Plantar fat pad pain has also beendescribed as a separate entity.13

The pain is said to be in the fat paditself posterior to the plantar fasciainsertion. Pacinian corpuscles havebeen noted in the plantar fat pad,lending credence to this less oftenconsidered cause of subcalcanealpain.12 However, the existence ofthese two entities, apophysitis andfat pad pain, as distinct from plan-tar fasciitis, is conjectural at pres-ent. Heel pain that is not subcal-caneal, such as that due to Achillestendinitis, retrocalcaneal bursitis,

Plantar Fasciitis


Fig. 1 Location of nerves in proximity to the heel. Arrow indicates force of dorsiflexion.Inset, Windlass mechanism involving fascial attachment at base of proximal phalanges.

Calcanealnerves

Lateralplantarnerve

Heel pad

First branch oflateral plantarnerve

Heel pad

Plantarfascia

Lateral plantarnerve

Abductor hallucis

Fig. 2 Bone scans show bilateral plantar fasciitis. Increased uptake can be seen at theattachment site of the medial calcaneal tubercle and at the fascial enthesis attachment site.

subtalar arthritis, or posterior tibialand other tendinitis, should be lessfrequently confused because of thedifferent location.

Differentiating between plantarfasciitis and calcaneal stress frac-ture is important. Calcaneal stressfractures usually present with cal-caneal swelling, increased warmth,and tenderness to touch. There isoften a positive “squeeze test”when the patient’s calcaneus issqueezed between the volar surfaceof the examiner’s fingers and thethenar eminence of his or her hand.These findings are commonlyabsent in the patient with plantarfasciitis.

Neurogenic causes of heel painare so frequently cited in conjunc-tion with plantar fasciitis that thedistinction between them and thevarious possible causes of plantarfasciitis may be blurred. Someauthors feel that neurogenic factorsare part of the true etiology of thesyndrome of subcalcaneal pain atthe plantar fascia origin8,14,15 andthat the degenerative/inflammato-ry process at the origin of the plan-tar fascia may lead to secondaryneuropathy.8,15

The first branch of the lateralplantar nerve is a mixed motor-sensory nerve to the abductor digitiquinti minimi, which passes supe-rior to the attachment of the plantarfascia (Fig. 3). Several authors, par-ticularly Baxter, have drawn atten-tion to an impingement syndromethat can occur in several areasalong the course of this nerve. Onthe medial hindfoot, after exitingthe tarsal tunnel, the nerve runsdeep to the abductor muscle fasciaand through its fascial leash, wherethe inferior edge close to the plan-tar fascia attachment on the calca-neus can be thick and unyield-ing.14,16 The nerve then turns later-ally as it courses across the hind-foot superior to the attachment ofthe plantar fascia (Fig. 3). A heelspur just dorsal to the plantar fas-cia may add to the nerve impinge-ment farther along the nerve distal-ly (Fig. 4).8,16,17 Unfortunately, it isdifficult to obtain reliable electrodi-agnostic studies to support this eti-ology.15 Therefore, the best evi-dence at present remains the physi-cal examination findings and thepain relief noted after decompres-sion.

The medial calcaneal nerve is asensory nerve that is considered bysome authors to contribute to sub-calcaneal pain.8,18,19 Savastano18

has described an operation forresection of this nerve in cases ofintractable heel pain. Compressionof the posterior tibial nerve, a tarsaltunnel syndrome, has also beenassociated with heel pain,12 al-though this commonly causes amore widespread pain distributionradiating distally to the forefoot orproximally into the tunnel.

When evaluating for nerve com-pression, the precise location ofpain should be assessed. Themedial calcaneal nerve is the mostposterior and the most superficialbeneath the skin and subcutaneoustissue. The nerve to the abductordigiti quinti, which is the firstbranch of the lateral plantar nerve(Baxter’s nerve), is deep to theabductor hallucis muscle andcourses just superior and medial tothe plantar fascia insertion. Thelateral and medial plantar nervesare more anterior in the foot afterleaving the tarsal tunnel.

Diagnosis

Despite some disagreement as tothe true source of pain in subcal-caneal pain syndromes and a longlist of possibilities in the differen-tial diagnosis, the diagnosis ofplantar fasciitis is usually straight-forward. Several factors in the his-tory and examination are so char-acteristic that in most cases thediagnosis is not difficult. Pain thatis worse on first arising in themorning or after a period of rest ishighly suggestive of plantar fasci-itis specifically. The pain oftenimproves after more ambulationbut may recur after prolonged, con-tinued, or more stressful activity.When severe, the pain may have athrobbing, searing quality. A delay

Lowell H. Gill, MD


Fig. 3 Axial (A) and sagittal (B) T1-weighted magnetic resonance images. Arrows indi-cate the first branch of the lateral plantar nerve.

A B

in athletes,17,20,21 obesity,3,4,9,19,22-26

and middle age9-12,26 (the most com-mon age for presentation).

Abnormal foot biomechanicsmay predispose to this condition.The cavus foot accommodatespoorly to variable stresses. Withless of the normal hindfoot andmidfoot motion, especially prona-tion, to diffuse stress in the cavusfoot, the fascia is subject to in-creased stress. The excessive laxityof an overly pronated pes planusfoot also places excessive stress onthe plantar fascia. With less liga-mentous support in the flat foot, agreater burden is placed on the fas-cia. The association of plantarfasciitis with both cavus andpronated foot is supported by clini-cal observations.17 A tight heelcord also contributes to excessivestress in the plantar fascia.4,12,27,28

Relationships have also beenproposed, but not established, forother factors, such as acute injury,the presence of a heel spur, theshoe type, the walking surface, andemployment or chronic repetitiveactivity other than athletics. Somestudies suggest a correlation withprolonged standing or walk-ing.1,4,12,24 The British eponymousterm “policeman’s heel” implies a

in the appearance of symptoms,such as when the pain occurs themorning after a day of increasedactivities, is common and maycause the patient and hence thephysician to overlook a relation-ship to the increased activity.

The second highly characteristicfeature is the location of the pain,which is usually at the origin of theplantar fascia from the medialtubercle of the calcaneus. The painmay be aggravated by passive dorsi-flexion of the toes in more severecases. Infrequently, the pain alsoradiates distally along the plantarfascia. The abductor hallucis originfrom the calcaneus just superior andmedial to the plantar fascia may alsobe tender. A bone scan frequentlyshows increased uptake in the areaof the fascial attachment at themedial calcaneal tubercle (Fig. 2),11

mirroring the pathologic changes inthe nearby enthesis. Bone scanningcan be of benefit in patients with anatypical clinical presentation.

Risk Factors

Specific risk factors are known to beclearly associated with plantar fasci-itis. These include repetitive stress

work-related overuse syndromesecondary to prolonged standing.9A correlation with acute injury isless clear; however, patients oftenmention stepping on a rock oranother hard object as an initiatingevent, hence the common name“stone bruise.”

The association of plantar fasci-itis with a radiographically visual-ized heel spur has caused consider-able confusion. It is well acceptedthat even though a spur may beseen coincidentally or even associ-ated with the clinical condition, it isnot by itself the etiologic factor.However, some studies suggestthat there may be some associationbetween the presence of a spur and the clinical syndrome.9,11,19,23

Baxter indicates that the spur mayadd to neurogenic pain with com-pression of the first branch of thelateral plantar nerve.7,16,17

Heel spurs have been found inapproximately 50% of patients withplantar fasciitis. This exceeds the15% prevalence of radiographicallyvisualized spurs in normal asymp-tomatic patients noted by Tanz.19

However, middle age itself is awell-recognized risk factor, andspurs are more common as peopleage. Therefore, the associationbetween spurs and plantar fasciitismay be coincidental. Furthermore,it has been shown that the spursoccur in the short toe flexors justsuperior to the fascia, rather thanthe plantar fascia itself (Fig. 4).16,19

Some believe the spur is the result,not the cause, of plantar fasciitis11;others believe the spur is unre-lated.4,12,22

Treatment

Treatment regimens for plantarfasciitis vary widely. With the pos-sible exception of casting, no singlemethod stands out as clearly superi-or. Moreover, the orthopaedic liter-

Plantar Fasciitis


Fig. 4 Gradient-echo mag-netic resonance image with20-degree flip angle. Noterelationship of spur in shorttoe flexors dorsal to theplantar fascia and in prox-imity to the neurovascularbundle (arrow), includingthe first branch of the later-al plantar nerve.

ature offers little guidance. Com-parison of results is difficult becausemost reports reflect only a singlemethod of treatment or because sev-eral treatments are used simultane-ously. Furthermore, statisticalanalysis of treatments is problemat-ic because patient numbers areoften small, and there are many dif-ferent treatments to be considered.

Physical TherapyPhysical therapy modalities are

frequently employed. Baxter andThigpen17 have noted, however,that ultrasound and whirlpool arenot helpful. In another study,24

application of ice provided moder-ate benefit to 23% of patients butexcellent results to only 4.5%.Likewise, heat provided some ben-efit to 15.1% of patients but excel-lent results to only 1.7%.

Stretching exercises are pre-ferred by many practitioners.Wolgin et al26 had successfulresults in 83% of their patients, andDavis et al27 reported that stretch-ing was their most effective conser-vative treatment. Stretching exer-cises may also benefit patients witha tight Achilles tendon, a groupwho are known to be at risk forplantar fasciitis. The advantages ofthese modalities are relative easeand minimal expense when self-administered by the patient.

Nonsteroidal Anti-inflammatoryDrugs

Furey25 reported a 71% successrate for phenylbutazone treatmentof 78 patients. Wolgin et al26 foundthat 39 (76%) of 51 patients whoused nonsteroidal anti-inflammato-ry drugs (NSAIDs) had a successfuloutcome. In contrast, Williams9

found NSAIDs generally ineffec-tive. In a study of 283 patientstreated with NSAIDs, 202 patients(71%) reported positive results, butthe improvement was most oftenrated as being only slight.24 Al-

though 17 (6%) of the patientsthought the results with NSAIDswere excellent and 75 (27%) report-ed considerable improvement, 81patients (28%) thought the NSAIDswere ineffective.

Heel CupsIn theory, plastic heel cups offer

protection by supporting the fibro-fatty tissue beneath the heel, pro-viding a better cushion. Thesecups are available in only twosizes: adult and pediatric. Snookand Chrisman23 reported that thecups dramatically relieved pain in13 of 22 patients and consideredplastic cups their most successfultreatment. Leach et al10 noted thatin their clinical experience manypatients obtained relief with heelcups. Schepsis et al5 reported thatheel cups were sometimes helpful.

With the exception of the reportby Snook and Chrisman,23 howev-er, there appear to be no data in theorthopaedic literature to verify thebenefit of these devices. In an out-come study of the use of heel cupsin 131 patients,24 8 patients (6%)ranked the treatment as excellent,and 59 (45%) ranked it as poor.Furthermore, Katoh et al13 showedworsening of an abnormal gait pat-tern with use of heel cups. Usingforce plates, these authors studiedvertical reaction force and timespent during gait for the hind-,mid-, and forefoot. In patients withplantar fasciitis, heel cups actuallyaggravated an abnormal gait pat-tern by decreasing the amount oftime spent on the heel and increas-ing the amount of time spent on themidfoot and forefoot.

Tuli CupsTuli cups are made of natural

latex rubber and have a ribbeddesign. The ribs are crushed onimpact and then rebound, whichdissipates the force of heel strike.Thus, the cup acts like both a cush-

ion and a cup, which is an attrac-tive concept. Leach and Schepsis21

noted that the Tuli cup is particu-larly helpful for patients who wishto continue their athletic activities.However, the only study presentlyavailable in the orthopaedic litera-ture shows that no patients rankedit as excellent, and 21 of 38 rankedit as poor.24 Therefore, despiteclaims of benefit, there are noreports to date documenting thesuccess of Tuli cups.

PadsWolgin et al26 compared various

conservative modalities in 100patients and found that pads weresuccessful in 83%. Davis et al27

reported that viscoelastic polymerheel cushions were often helpful.Another study,24 however, foundthat only 4 (2%) of 184 patientsranked foam pads as excellent, and62 (34%) reported no improvement.

OrthoticsOrthotics are more frequently

used for subcalcaneal heel painwhen there is a coexistent biome-chanical variation, such as pesplanus or pes cavus. In pes planus,a medial arch support or medialwedge may be employed. In pescavus, an attempt is made to dissi-pate stress over a broad area, as witha diabetic foot insert. Bordelon8

prefers an orthotic designed to cush-ion the heel while relieving pressureon the tender area. A molded Plasta-zote insert with a medial elevationmay be employed in resistant cases.Baxter and Thigpen17 note that heelpain in an athlete with a cavus footcan usually be controlled with aflexible support or orthosis. Camp-bell and Inman2 have reported suc-cess in 31 of 33 cases with use of aUC-BL orthosis. With the exceptionof their study, however, there hasbeen no orthopaedic study that eval-uates the benefits of any orthotic inthe treatment of plantar fasciitis.

Lowell H. Gill, MD


Corticosteroid InjectionsPlantar fascia ruptures after

corticosteroid injections have beenreported,28 which can lead to gaitabnormalities.29 Many practitionershave reported using injections occa-sionally.1,4,5,9,11,12,19,24,25,27 Injectionsare generally limited to two perside; in very rare instances a thirdmay be given if a long time haspassed since the last injection, andthe first one or two were of benefit.If three have not been successful, itseems fruitless to continue thisform of treatment.

The use of an injection wasfound by Blockey1 to have cured 10of 13 painful heels, but he alsonoted that pain was relieved in 5 of9 heels with placebo saline injec-tions alone. Davis et al27 foundinjections to be effective in 26 (52%)of 50 symptomatic heels in 41patients. Wolgin et al26 cited a 35%success rate (11 of 31 patients).Miller et al30 noted that pain reliefwas rated good or better in 27 heelsin 24 patients; however, the effectwas only temporary for most pa-tients. In a study in which patientsranked treatments,24 the group of171 patients who received an injec-tion rated that form of treatmenthigher than ten nonsurgical treat-ments other than casting. Of the171 patients, 41 reported no benefitfrom the injections, 31 rated theirresults excellent, and the remaining99 ranked the treatment betweenthose extremes.

Night SplintsWapner and Sharkey3 have

reported success with use of a 5-degree dorsiflexion night splint,which holds the plantar fascia in acontinuously tensed state. This ispostulated to minimize the changeof tension that occurs with each newday’s activities and thereby to mini-mize the chronic repetitive micro-trauma seen in plantar fasciitis.Night splints may also be used con-

veniently as an adjunct when dis-continuing cast treatment. The pos-terior half of the bivalve cast issaved and reapplied by the patienton a nightly basis after discontinua-tion of casting. This method doesnot incorporate the 5 degrees of dor-siflexion recommended by Wapner.3

Miscellaneous TreatmentsThe use of special types of foot-

wear (e.g., running shoes or soft-soled shoes) may be beneficial insome cases.24 Taping is mentionedin the podiatry literature but has notbeen evaluated scientifically withregard to plantar fasciitis. Shoemodification, such as use of a steelshank to limit metatarsophalangealdorsiflexion during toe-off or a heellift, and a change to wearing high-heel shoes to decrease heel impacthave also been tried. Radiationtherapy has been reported to pro-vide disappointing results.12

Combination TreatmentMany patients receive various

combinations of nonsurgical treat-ments. All 323 patients (364 heels)of Lapidus and Guidotti4 werecured with the combination ofphenylbutazone, corticosteroidinjection, and rest. Furey25 reportedthat the use of phenylbutazone inconjunction with mechanical aids,such as heel pads and arch sup-ports, yielded good results in 71%of 78 patients at 5 years. Davis etal27 reported an 89.5% success ratefor the combination of NSAIDs, rel-ative rest, heel cushions, Achilles-stretching exercises, and occasionalinjections. Clancy31 reported on theuse of a medial heel wedge, flexibleleather support, heel-cord stretch-ing, and rest for 6 to 12 weeks.Unfortunately, there are no studiescomparing different conservativetreatments used independently,which makes it difficult to ascertainwhich of the many available treat-ments really make a difference.

RestRest helps ameliorate symp-

toms.24,27 However, the recommen-dation to rest is often poorlyaccepted by patients, particularlyas resolution of symptoms maytake months or longer. Poor pa-tient compliance, especially bypatients who consider their prob-lem a minor one, may account forthe high percentage of failures withthe various conservative treatmentregimens.

CastingSchepsis et al5 reported that cast-

ing is not helpful. Other studies,however, suggest that a cast can beeffective even in recalcitrantcases.19,24,32 McBryde20 recom-mends casting in long-standingcases. Tisdel and Harper32 foundcasting to be satisfactory in over50% of their most recalcitrant cases,in which numerous treatments foran average of 1 year had failed toprovide relief (the 13 patientswould therefore be considered sur-gical candidates). In anotherstudy,24 casting was also used inthe most difficult cases and wasfound to be the most successful of11 nonsurgical treatments assessed.

A cast may work by providingcontinuous unchanging tension onthe plantar fascia, thus minimizingmicrotrauma with each new day’sstretching (similar to the mecha-nism postulated by Wapner andSharkey3 for use of a night splint).A cast may also relieve tension onthe plantar arch in much the sameway that Campbell and Inman2

postulated for the UC-BL insert.Cast immobilization also undoubt-edly enforces rest. A combinationof all three mechanisms may be thereason for the apparent success ofcasting. Campbell and Inman haveeven conjectured that surgeonswho used operative treatment ofplantar fasciitis might “have beenequally successful if they had omit-

Plantar Fasciitis


ted the surgical procedure and sim-ply used the walking cast.”

Author’s Preferred Method ofTreatment

If the clinical condition is rela-tively mild, as in the athlete who isable to run through the first fewpainful steps and then continuetraining with minimal discomfort,the more convenient therapies (e.g.,stretching, ice, heat, NSAIDs, oruse of a foam or viscoelastic pad,athletic shoe, or crepe-soled shoewith soft heel pad) are acceptable.However, patients should bewarned that plantar fasciitis isoften a recalcitrant condition, dueto repetitive stress, and that de-creased activity is essential. It isimportant to point out that in-creased pain may not occur untilthe morning after a particularlylong run or active day. With thetime delay in onset of symptoms,many patients will overlook thecorrelation with increased activity.

If the clinical condition is ofmoderate severity and the initialtreatments have failed, which iscommon, an injection is recom-mended because of its convenience.Patients are warned, however, thatnot more than two or, in rareinstances, three injections will begiven and that rest must accompa-ny the treatment. A night splint isan alternative to an injection at thisstage. Patients are also advisedthat if they want to do the most tohelp their problem, a cast is recom-mended.

If the injection and/or numer-ous other treatments have failed orthe patient is very obese or has hadprolonged symptoms (6 to 12months or longer), the patient isstrongly urged to consider casting.A period of rest from work may beadded if the situation warrants orthe cast precludes work. The castis left in place for 5 to 6 weeks. Ifthe patient is intolerant of casting, a

zipper cast (Neofrakt-Motion Med-ical Distributors, Birmingham, Ala)or a night splint is offered. Itshould be explained that zippercasts may be less effective and aresubject to cracking. If casting isineffective, there is increased suspi-cion of other diagnoses in the dif-ferential, such as nerve entrapmentand systemic disease.

SurgeryWith the increased popularity of

endoscopic plantar fascia release,there is concern regarding the over-zealous and inappropriate use of atechnique with known risks fornerve damage. For this reason, theAmerican Orthopaedic Foot andAnkle Society has recently devel-

oped a position statement regard-ing heel surgery (Table 1).

At the time of planning surgery,the differential diagnosis is re-viewed, looking for other possiblecauses for subcalcaneal pain mas-querading as plantar fasciitis.Appropriate laboratory tests areordered, and medical evaluation isdone if indicated. A bone scan isfrequently obtained before surgeryto substantiate the diagnosis. Inmy practice, patients are also re-quired to have a period of castingbefore consideration of surgery; onoccasion, this is repeated.

A large number of surgical tech-niques have been described,29 butthere is as yet little consensus re-garding the optimal procedure. In

Lowell H. Gill, MD


Table 1American Orthopaedic Foot and Ankle Society Position Statement onEndoscopic and Open Heel Surgery*

1. Nonsurgical treatment is recommended for a minimum of 6 months andpreferably 12 months.

2. More than 90% of patients respond to nonsurgical treatment within 6 to10 months.

3. When surgery is considered for the remaining patients, then a medicalevaluation should be considered prior to surgery.

4. Patients should be advised of complications and risks if an endoscopic oropen procedure is not indicated.

5. If nerve compression is coexistent with fascial or bone pain, then an endo-scopic or closed procedure should not be attempted.

6. The AOFAS does not recommend surgical procedures before nonopera-tive methods have been utilized.

7. The AOFAS supports responsible, carefully planned surgical interventionwhen nonsurgical treatment fails and workup is complete.

8. The AOFAS supports cost constraints in the treatment of heel pain whenthe outcome is not altered.

9. The AOFAS recommends heel padding, medications, and stretching priorto prescribing custom orthoses and extended physical therapy.

10. This position statement is intended as a guide to the orthopaedist and isnot intended to dictate a treatment plan.

* Reproduced with permission from the American Orthopaedic Foot and Ankle Society.

References

1. Blockey NJ: The painful heel: A con-trolled trial of the value of hydrocorti-sone. BMJ 1956;1:1277-1278.

2. Campbell JW, Inman VT: Treatmentof plantar fasciitis and calcaneal spurswith the UC-BL shoe insert. ClinOrthop 1974;103:57-62.

3. Wapner KL, Sharkey PF: The use ofnight splints for treatment of recalci-trant plantar fasciitis. Foot Ankle1991;12:135-137.

4. Lapidus PW, Guidotti FP: Painful heel:Report of 323 patients with 364 painfulheels. Clin Orthop 1965;39:178-186.

5. Schepsis AA, Leach RE, Gorzyca J:Plantar fasciitis: Etiology, treatment,surgical results, and review of the lit-erature. Clin Orthop 1991;266:185-196.

6. Huang CK, Kitaoka HB, An KN, et al:Biomechanical evaluation of longitudi-nal arch stability. Foot Ankle 1993;14:353-357.

7. Schon LC: Plantar fasciitis/heel pain,in Pfeffer GB, Frey CC, Anderson RB,et al (eds): Current Practice in Foot andAnkle Surgery. New York: McGraw-Hill, 1993, vol 1, pp 243-261.

8. Bordelon RL: Heel pain, in Mann RA,Coughlin MJ (eds): Surgery of the Footand Ankle, 6th ed. St Louis: Mosby-Year Book, vol 2, pp 837-857.

9. Williams PL: The painful heel. Br JHosp Med 1987;38:562-563.

10. Leach RE, Seavey MS, Salter DK:Results of surgery in athletes withplantar fasciitis. Foot Ankle 1986;7:156-161.

11. Graham CE: Painful heel syndrome:Rationale of diagnosis and treatment.Foot Ankle 1983;3:261-267.

12. Karr SD: Subcalcaneal heel pain.Orthop Clin North Am 1994;25:161-175.

13. Katoh Y, Chao EYS, Morrey BF, et al:Objective technique for evaluatingpainful heel syndrome and its treat-ment. Foot Ankle 1983;3:227-237.

14. Kenzora JE: The painful heel syn-drome: An entrapment neuropathy.Bull Hosp Jt Dis Orthop Inst 1987;47:178-189

15. Schon LC, Glennon TP, Baxter DE:Heel pain syndrome: Electrodiagnosticsupport for nerve entrapment. FootAnkle 1993;14:129-135.

16. Baxter DE, Pfeffer GB, Thigpen M:Chronic heel pain: Treatment ratio-nale. Orthop Clin North Am 1989;20:563-569.

17. Baxter DE, Thigpen CM: Heel pain:Operative results. Foot Ankle 1984;5:16-25.

18. Savastano AA: Surgical neurectomy forthe treatment of resistant painful heel.Rhode Island Med J 1985;68:371-372.

19. Tanz SS: Heel pain. Clin Orthop 1963;28:169-178.

20. McBryde AM Jr: Plantar fasciitis.Instr Course Lect 1984;33:278-282.

21. Leach RE, Schepsis A: Hindfoot painin athletes: Why, and what can bedone? J Musculoskel Med 1985;2:16-25.

22. Hill JJ Jr, Cutting PJ: Heel pain andbody weight. Foot Ankle 1989;9:254-256.

23. Snook GA, Chrisman OD: The man-agement of subcalcaneal pain. ClinOrthop 1972;82:163-168.

24. Gill LH, Kiebzak GM: Outcome ofnonsurgical treatment for plantar fasci-itis. Foot Ankle Int 1996;17:527-532.

25. Furey JG: Plantar fasciitis: The painful

most cases, I perform subtotal plan-tar fascia release. Decompressionof the first branch of the lateralplantar nerve as reported by Baxterand Thigpen17 has merit in appro-priate cases.

These procedures are not with-out risk. Sellman28 has reportedmidfoot pain after plantar fasciarupture. Foot biomechanics areknown to be altered after release.29

There have been an increasingnumber of reports, as yet unpub-lished, of lateral foot pain afteroverzealous plantar fascia release.This observation may be the clini-cal result of the biomechanicalchanges that follow plantar fasciarelease.

Summary

The plantar fascia plays an impor-tant role in support and is subject

to chronic repetitive tensile stressat its calcaneal origin. There are anumber of potential causes of sub-calcaneal heel pain. Plantar fasci-itis, which is a degenerativeprocess in the fascial enthesis, isone of the most common. Therehas also been considerable interestin the role of neurogenic causes ofheel pain in conjunction with plan-tar fasciitis. The first branch of thelateral plantar nerve, in particular,lies close to the plantar fascial ori-gin as well as to a spur, if present,and may be affected by localizedtissue changes and swelling or bytethering and tight fascial struc-tures.

The diagnosis is most oftenmade on the basis of the location ofpain at the medial tubercle of thecalcaneus and the typical history ofpain after a period of rest. Risk fac-tors include biomechanical abnor-malities of the foot, increased body

weight, middle age, and repetitivestress. A heel spur seen on radio-graphs is most commonly consid-ered to be unrelated, although insome cases it may contribute tonerve impingement.

Conservative treatment regi-mens vary, and there is no clearconsensus on the most effectivemodalities. Recent reports indicatethat a trial of casting is worthwhilebefore consideration of surgery.There is a consensus that nonsurgi-cal treatment is effective approxi-mately 90% of the time.

Acknowledgments: The author would liketo thank Gary M. Kiebzak, PhD, for hisassistance in editing; Peggy A. Vain andCharlene Polson for their assistance in thepreparation of this manuscript; BrianHoward, MD, for his assistance in compila-tion and interpretation of magnetic reso-nance imaging studies; and Cynthia Ceronfor her assistance with illustrations.

Plantar Fasciitis






















heel syndrome. J Bone Joint Surg Am1975;57:672-673.

26. Wolgin M, Cook C, Graham C, et al:Conservative treatment of plantar heelpain: Long-term follow-up. Foot AnkleInt 1994;15:97-102.

27. Davis PF, Severud E, Baxter DE:Painful heel syndrome: Results of non-operative treatment. Foot Ankle Int1994;15:531-535.

28. Sellman JR: Plantar fascia ruptureassociated with corticosteroid injec-tion. Foot Ankle Int 1994;15:376-381.

29. Daly PJ, Kitaoka, HB, Chao EYS:Plantar fasciotomy for intractableplantar fasciitis: Clinical results andbiomechanical evaluation. Foot Ankle1992;13:188-195.

30. Miller RA, Torres J, McGuire M:Efficacy of first-time steroid injection

for painful heel syndrome. Foot AnkleInt 1995;16:610-612.

31. Clancy WG: Runner’s injuries: Part II.Evaluation and treatment of specificinjuries. Am J Sports Med 1980;8:287-289.

32. Tisdel CL, Harper MC: Chronic plan-tar heel pain: Treatment with a shortleg walking cast. Foot Ankle Int 1996;17:41-42.

Lowell H. Gill, MD











JAAOS - Volume 05 - Issue 02 March 1997

Documents

Transcript of JAAOS - Volume 05 - Issue 02 March 1997