OrcrtptPpttai

Fwc

WE

Arthroscopic Side-to-Side Rotator Cuff Repair

Eugene M. Wolf, M.D., William T. Pennington, M.D., and Vivek Agrawal, M.D.

Purpose: To study the results of arthroscopic repair of full-thickness rotator cuff tears using aside-to-side suture technique without fixation to bone. Type of Study: Case series study of thelong-term results of patients who underwent purely arthroscopic rotator cuff repair with a side-to-sidesuturing technique. Methods: A retrospective review was performed of patients who underwentarthroscopic repair of full-thickness rotator cuff defects. Patients with full-thickness rotator cuff tearsrepaired in a side-to-side fashion without anchoring the repair to bone were selected. Patients wereevaluated using a modified UCLA shoulder scoring system. The data collected were analyzed todetermine the outcome in patients with a 4- to 10-year follow-up. Results: A total of 105 arthroscopicrotator cuff repairs were performed in 104 patients between February 1990 and February 1996.Forty-two patients had a full-thickness tear of the rotator cuff that was repaired using a purelyside-to-side suturing technique. The mean UCLA score of all patients in this group was 33; 23patients reported excellent results, 18 good results, and 1 poor result according to the UCLA scoringsystem. Conclusions: In this series, 98% of patients qualified as a good to excellent result accordingto the UCLA shoulder score. This study shows that patients with a full-thickness defect of the rotatorcuff tendon with anatomy amenable to side-to-side closure may be effectively treated with a purelyarthroscopic repair using only a side-to-side suturing technique with excellent long-term clinicalresults. Level of Evidence: Level IV. Key Words: Side-to-side suture repair—Rotator cuff tear.

fitvrcprhhwtdsbtt

ther surgeons have reported the results of arthro-scopic repair of full-thickness defects of the

otator cuff.1-7 These studies support the use of arthros-opy in the treatment of symptomatic rotator cuff tearsecalcitrant to nonsurgical management. As a result ofhese reports of successful postoperative outcomes,urely arthroscopic methods to repair full-thicknessears of the rotator cuff are being increasingly used.roponents of the arthroscopic treatment of rotator cuffathology cite less morbidity to the surrounding soft-issue envelope, no deltoid detachment, better visualiza-ion of the pathology of the rotator cuff, and an acceler-ted postoperative rehabilitation period with a lowerncidence of postoperative stiffness.6,7

From the California Pacific Medical Center (E.M.W.), Sanrancisco, California; St. Luke’s Medical Center (W.T.P.), Mil-aukee, Wisconsin; and Central Indiana Orthopedics (V.A.), Mun-ie, Indiana, U.S.A.Address correspondence and reprint requests to Eugene M.olf, M.D., 3000 California St, San Francisco, CA 94115, U.S.A.-mail: genewolfmd@aol.com© 2005 by the Arthroscopy Association of North America

c0749-8063/05/2107-2767$30.00/0doi:10.1016/j.arthro.2005.03.014

Arthroscopy: The Journal of Arthroscopic and Related

Surgical treatment of full-thickness rotator cuff de-ects has focused on recreating the anatomy of thentact rotator cuff with reinsertion and fixation of theendon to the greater tuberosity of the humerus. Pre-ious authors have presented open and arthroscopicepair techniques that involve recognition of the tearonfiguration and use a side-to-side repair as a com-onent of the rotator cuff repair.2,3,8,9 To date, alleports of purely arthroscopic repair of the rotator cuffave involved securing the repair to the proximalumerus with either suture through bone tunnels orith anchor-based fixation. This article is unique in

hat we report the results of repairs of full-thicknessefects of the rotator cuff with a purely arthroscopicide-to-side technique without anchoring the repair toone. This repair technique was performed in tearshat appeared amenable to this type of repair afterhorough arthroscopic visualization and evaluation.

METHODS

The results of 105 consecutive arthroscopic rotator

uff repairs in 104 patients performed by the senior

881Surgery, Vol 21, No 7 (July), 2005: pp 881-887

aaaeiw1wsbaopaw

aaprpesfftaa(

dppffda

S

ttairttftbt

trhbhthptsbn

882 E. M. WOLF ET AL.

uthor (E.M.W.) over a 6-year period between Febru-ry 1990 and February 1996 were evaluated. Initially,retrospective clinical chart review was performed forach case. Nine patients were lost to follow-up, leav-ng 96 shoulders in 95 patients available for evaluationith an average follow-up of 73 months (range, 48 to20 months). Forty-two of these patients who under-ent repair of a full-thickness rotator cuff tear in a

ide-to-side fashion without anchoring the repair toone were selected for this study. There were 24 malend 18 female patients, and the average age at the timef surgery was 59.8 years (range, 42 to 79 years). Allatients had been recalcitrant to conservative therapynd continued to experience unacceptable pain andeakness in the affected shoulder.All patients were clinically evaluated by the senior

uthor. Final outcome assessment was performed byn independent surgeon (W.T.P.) by telephone. Eachatient was contacted to assess for pain, function,ange of motion, strength, return to work date, anderceived success of the procedure. Outcome wasvaluated using a modified UCLA shoulder ratingcale10 (Table 1). This scale designates 10 points eachor pain and function and 5 points each for activeorward flexion, strength of forward flexion, and pa-ient satisfaction, for a total possible score of 35. Goodnd excellent results (total UCLA score 28-35 points)re considered satisfactory and fair and poor resultsless than 28 points) are considered unsatisfactory.

Eighty-one percent of patients had repairs of theirominant shoulder, with 33 right and 9 left repairs. Allatients also had arthroscopic subacromial decom-ressions. Seven patients had other procedures per-ormed concurrently, including 3 arthroscopic Mum-ord procedures, 1 SLAP lesion debridement, 1ebridement of a biceps tendon rupture, and 2 oscromiale excisions.

urgical Technique

Routine shoulder arthroscopy was performed withhe patient in the lateral decubitus position. Initially,he glenohumeral joint was inspected to evaluate forny significant intra-articular pathology. The cuff wasnspected from the articular side and the defect in theotator cuff tendon was debrided of all frayed, devi-alized tissue. The arthroscope was reconfigured intohe subacromial space and a decompression was per-ormed using a cutting-block technique. It is importanto remove all bursal tissue covering the rotator cuff toe able to evaluate the extent of the tear. The bursec-

omy was also necessary to provide enough visualiza- r

ion of the cuff and of the suture hooks used in theepair. The region of the greater tuberosity of theumerus was abraded with a full-radius shaver andurr to create a bed of bleeding bone to promoteealing of the cuff to the tuberosity. The mobility ofhe rotator cuff was evaluated with a grasper or nerveook. Each tear was assessed individually and re-aired with “L” or “V-Y” techniques. All repairs inhis report were performed arthroscopically using aide-to-side technique without fixation of the repair toone. All tears in this series were evaluated with aerve hook with an attempt to evaluate the anatomic

TABLE 1. Modified UCLA Shoulder Rating Scale

Patient satisfaction0 Patient feels procedure was not successful5 Patient feels procedure was a successActive forward flexion range of motion0 Less than 30°1 30°-45°2 45°-90°3 90°-120°4 120°-150°5 Greater than 150°Strength of forward flexion0 No active contraction1 Evidence of slight muscle contraction, no active elevation2 Complete active forward flexion with gravity eliminated3 Complete active forward flexion against gravity4 Complete active forward flexion against gravity with some

resistance5 Complete active forward flexion against gravity with full

resistancePain1 Present always and unbearable, strong medication frequently2 Present always but bearable, strong medication occasionally4 None or little at rest, present during light activities;

salicylates frequently6 Present during heavy or particular activities only, salicylates

occasionally8 Occasional and slight10 NoneFunction1 Unable to use limb2 Only light activities possible3 Able to do light housework or most activities of daily living6 Most housework, shopping, and driving possible; able to do

hair and to dress and undress, including fastening brassiere8 Slight restriction only, able to work above shoulder level10 Normal activitiesTotal

Excellent: 34-35Good: 28-33Fair: 21-27Poor: 0-20

elationship between the margins of the torn cuff

eaucpttap

UaCwatntrpsitIsbohaTirstaetmohsos

aswsEptf

eApe

c1U(sa

ss(2(t

rrs

Fdthcca

883ARTHROSCOPIC SIDE-TO-SIDE ROTATOR CUFF REPAIR

dges. The nerve hook was believed to be better thangrasper because of its smaller size. Its blunt tip wassed to penetrate various points on the edge of theuff, which was then mobilized to determine the ap-ropriate configuration of repair. This evaluation ofhe cuff tear with a nerve hook is of foremost impor-ance to ensure the optimal restoration of the normalnatomy without producing any undue tension on anyart of the repair.Burkhart9 has suggested that the deepest point of a-shaped tear does not represent tear retraction but is

ctually an L-shaped tear under physiologic load.hronic rotator cuff tears have tapered edges and mayell have a “U” or “V” configuration, but this can be

ppreciated by trying to approximate the deep point ofhe U-shaped tear to the greater tuberosity with aerve hook and noticing the creation of “dog ears” inhe remainder of the cuff, showing the nonanatomicelationship of this configuration. The repair is thenerformed with the appropriate suture passer to passutures that will approximate the tendon edges, clos-ng the entire defect over the bleeding trough of bonehat was previously created on the proximal humerus.t is important to note that during the process of theide-to-side repair, a suture is often placed in what iselieved to be the anterior corner of the posterior leaff the tear followed by passage beneath the transverseumeral ligament and through the coracohumeral lig-ment to help advance the posterior leaf anteriorly.his tendon-to-tendon stitch provides secure approx-

mation of the retracted posterior leaf of the tear to theotator interval while holding the entire repairedtump over the prepared bleeding trough in the greateruberosity. There were no partial repairs performed;ll tears in this series were closed completely and werevaluated arthroscopically from the bursal side of theear as the shoulder was placed through a full range ofotion to ensure the maintenance of the relationship

f the tendon repair to the greater tuberosity of theumerus. An average of 4 sutures (range, 1 to 7utures) were used per cuff repair. A clinical examplef a side-to-side repair of a large U-shaped tear ishown in Figs 1-4.

Repairs were performed exclusively with absorb-ble PDS suture in 88% of repairs and nonabsorbableuture in 12%. This includes 36 shoulders repairedith No. 1 PDS and 1 shoulder with No. 0 PDS. Four

houlders were repaired with nonabsorbable No. 2thibond (Ethicon, Somerville, NJ). One repair waserformed using No. 2 Mersilene suture. Postopera-ively, patients were placed in a simple immobilizer

or 6 weeks but allowed to begin pulley and pendulum s

xercises after their first visit 5 days postoperatively.ctive exercises began 6 weeks postoperatively. Noatients were treated with an abduction brace postop-ratively.

RESULTS

Ninety-eight percent of patients had good and ex-ellent postoperative scores with 23 excellent (55%),8 good (43%), and 1 poor result (2%). The averageCLA score was 33. Forty-one of the 42 patients

98%) rated their surgery as being successful and wereatisfied with the repair. One patient rated his surgerys unsuccessful.

The UCLA shoulder scoring system evaluated fortrength, pain, and function (Table 1). The mean re-ponse in all patients grading their strength was 4.6range 2-5), mean response for pain was 9.0 (range-10), and mean perceived function grade was 9.3range 1-10). The average grade for forward flexion ofhe shoulder was 4.9 (range 1-5).

There were 2 patients in whom this arthroscopicepair was a revision of a previous open rotator cuffepair and 3 patients had previously undergone arthro-copic assisted mini-open repair on the contralateral

IGURE 1. This view from the midlateral subacromial portal isirected caudally and shows the bursal side of a large U-shapedear of the supraspinatus tendon. A bed of cancellous bone in theumeral head (H) is seen beneath the large defect in the rotatoruff. The anterior (A) and posterior (P) margins of the torn rotatoruff can be appreciated from this viewing portal. This large tearppears to be amenable to a side-to-side technique of closure.

houlder. All 3 of these patients stated that they were

matf

wa

tagqwssflaoiatcaaacipa

Fmph

Fvtase

Fit

884 E. M. WOLF ET AL.

ore satisfied with their side in which the purelyrthroscopic repair was performed and had a percep-ion of a quicker period of recovery and return tounction than with their open repair. The 1 patient

IGURE 2. This arthroscopic view through the midlateral subacro-ial viewing portal shows the passage of a No. 1 PDS suture from

osterior to anterior through a long crescent-shaped suture passingook (Linvatec, Largo, FL) through the apex of the U-shaped tear.

IGURE 3. This image again through the midlateral subacromialiewing portal shows the effect of this initial suture placed throughhe apex of the previously large-appearing U-shaped tear. Therthroscopic knot pusher is shown securing the initial Duncan

cliding knot of the single suture that was seen in Fig 2. Note theffect of this single suture on closing the remaining defect.

ith a poor result has failed subsequent open repairss well.

DISCUSSION

Rotator cuff tears are often attritional in nature andhe defect present often involves more than just anvulsion of the musculotendinous cuff from thereater tuberosity of the humerus. Burkhart9 has elo-uently described a broad classification scheme tohich rotator cuff tears can be classified: crescent-

haped or U-shaped tears. He describes the crescent-haped tear as a disruption of the tendinous insertionrom the greater tuberosity of the humerus without aarge element of retraction. The U-shaped tear usuallyppears on initial inspection to be a large retracted tearften medial to the level of the glenoid (Fig 1). Plac-ng the nerve hook in the base of such a tear andttempting to approximate the base to the greateruberosity usually yields 1 of 2 results with this tearonfiguration: approximation of the tendon is notchievable to the tuberosity secondary to tension, orpproximating the base of the U-shaped tear results in“dog-eared” appearance of the repaired cuff, indi-

ating tension mismatch and a nonanatomic repair. Its important for the surgeon to recognize this tearattern and use margin convergence as the primarypproach to repair. Repairing a U-shaped tear by an-

IGURE 4. This final arthroscopic view from the midlateral view-ng portal shows complete closure of the rotator cuff defect withhe placement of 3 side-to-side No. 1 PDS sutures.

horing the apex of the tear to the tuberosity will result

is

rMrldttrtrsctgifttstst

esesfhbtrbmcr

orBdottcaede

cw

oawwbwbccbrdrbss

fatodi

toipitnvtmtpscitsi

sfnwo

885ARTHROSCOPIC SIDE-TO-SIDE ROTATOR CUFF REPAIR

n tension overload of the repair, which has beenhown by Burkhart et al. to be doomed to failure.11,12

The use of side-to-side suturing as an element ofotator cuff repair has been described previously by

cLaughlin8 in his open approach to treating largeetracted tears of the rotator cuff. Although McLaugh-in advocated the use of this method to help close largeefects, he also was a proponent of final fixation of theendinous disruption of the cuff to bone in the head ofhe humerus at the point on the tuberosity that it wouldeach without undue tension with the arm at the pa-ient’s side.8 This early description is echoed in theepair techniques employed today with the arthro-copic approach. That is, anatomic restoration of theuff without the introduction of tension at the site ofhe repair. Burkhart et al. have coined the term “mar-in convergence” to describe the observation that dur-ng side-to-side repair the surgeon can visualize theree margin of the tear converging toward the greateruberosity with each suture being placed. They agreehat using margin convergence in the repair of U-haped tears decreases the amount of strain at theendon bone interface of the repair and thereforehould be protective to the tendon bone interface ofhe repair.9,13

Cadaveric dissection as well as arthroscopic clinicalvaluation has led to an appreciation of the relation-hip of the rotator crescent and rotator cable. Burkhartt al.14 described a consistently identifiable crescent-haped insertion of the distal supraspinatus and in-raspinatus tendons into the greater tuberosity of theumerus bordered on its medial margin by a thickenedundle of fibers oriented perpendicular to the axis ofhe supraspinatus and infraspinatus tendons.14 Theotator cable was grossly and histologically confirmedy a cadaveric study performed by Clark and Harry-an.15 It has been theorized that the thick rotator

able, when intact, provides stress shielding of theotator cuff crescent much like a suspension bridge.14

We believe that arthroscopic evaluation of the anat-my of the rotator cuff tear is an essential step inestoring the anatomy of the disrupted rotator cuff.urkhart9 suggests that visualization of the tear fromifferent arthroscopic portals allows the surgeon tobtain a 3-dimensional understanding of the tear pat-ern superior to that obtained by open means. We echohis sentiment in that arthroscopic repair of the rotatoruff allows a thorough evaluation of the completenatomy of the cuff disruption. Furthermore, withach suture passed, the effect may be evaluated byirect visualization of the impact of the suture on the

ntire cuff. The creation of flaps or “dog ears” indi- t

ates a nonanatomic repair and tension mismatch thatill likely fail under cyclic loading.While other researchers have reported their results

f purely arthroscopic repair, this is the first report ofseries of patients undergoing arthroscopic repair

ith purely tendon-to-tendon sutures being placedithout secure anchor or transosseous fixation toone. This technique of repair is used only in cases forhich, after thorough arthroscopic evaluation, it iselieved that a configuration exists that tear marginonvergence can be accomplished with side-to-sidelosure. All patients in this series had a trough ofleeding bone created in the proximal humerus; whenepair was complete, the repaired tendon stump layirectly over this bleeding bed of bone through a fullange of motion. All of these repairs were believed toe complete repairs because visualization of the bursalide of the tear after completion of the repair failed tohow any remaining defect in the rotator cuff.

The 98% good to excellent results compares quiteavorably with results previously reported in the liter-ture for surgical treatment of full-thickness defects ofhe rotator cuff. The minimum duration of follow-upf 4 years with an average of more than 6 yearsemonstrates an excellent long-term clinical outcomen this subset of patients.

As alluded to earlier and described by other inves-igators, there is an inherent balance of forces through-ut the musculotendinous insertion of the rotator cuffnto the greater tuberosity of the humerus. Our hy-othesis to explain such a positive long-term outcomen these patients without secure fixation of the repairo bone is that with the recreation of the anatomy, theatural balance of the rotator cuff musculature pro-ides an environment that is relatively stress free athe tendon-to-bone interface. This allows the distal-ost end of the repair that is overlying the bleeding

rough of bone to heal to the tuberosity during theeriod of postoperative convalescence. This hypothe-is is supported by Burkhart’s suspension bridge con-ept of stress shielding of the rotator crescent by anntact rotator cuff cable. If the cable is intact and ableo transfer the stress away from the crescent, thistress-free environment should be conducive to heal-ng of the tendon-to-bone interface.

The senior author has previously reported the re-ults of the first 54 purely arthroscopic repairs ofull-thickness defects of the rotator cuff. As a compo-ent of this previous study, second-look arthroscopyas performed on 23 patients to evaluate the integrityf the repair at a minimum of 6 months postopera-

ively. Nine of these patients had purely side-to-side

raeca2wWtatficgturwpp

lwttaAptrstha

Urfiprpecpidds

tis

tcrbmstticapaahsmumtosasrtpsi

886 E. M. WOLF ET AL.

epair of their rotator cuff and the second-lookrthroscopy was performed at 5 to 16 months postop-ratively. When examined, all 9 of these tears wereompletely healed to the greater tuberosity withoutny evidence of residual defect. Overall, including all3 arthroscopies performed, 70% of the repaired cuffsere intact at the time of second-look arthroscopy.1

e believe that these findings support our hypothesishat tendon-to-bone healing does occur despite thebsence of secure tendon-to-bone fixation with thisechnique. Reviewing the literature one would alsond ample evidence that the concept of “watertight”losure of the rotator cuff is difficult to achieve re-ardless of the method of fixation of the repairedendon to the greater tuberosity. Previous studies eval-ating the integrity of the rotator cuff have foundesidual rotator cuff defects in 34% to 90% of patientsho had previously undergone open rotator cuff re-air, despite secure intraoperative fixation of the re-air to the greater tuberosity.16-20

As discussed in the surgical technique section ear-ier, a suture that is often used by the senior authorhen performing repairs with this technique is passed

hrough the anterior corner of the posterior leaf of theear and then beneath the transverse humeral ligamentnd through the coracohumeral ligament anteriorly.fter this suture is tied, there is usually secure ap-roximation of the anterior corner of the posterior leafo the remaining intact rotator cuff anterior to theotator cuff interval. This suture is often useful inmall tears involving only the supraspinatus with ex-ension anterior into the rotator cuff interval. It is alsoelpful to aid in advancing the retracted posterior leafnteriorly when closing large L-shaped tears.

This study has admitted shortcomings. Although theCLA shoulder scores are available for the time pe-

iod of 4 to 10 years postoperatively, this only signi-es wellness at this moment in time after the treatmentrovided. Ideally, scores during the preoperative pe-iod with sequential scores during the perioperativeeriod would provide conclusive evidence of directffect of treatment on function of the shoulder. Be-ause some lived at a distance, not all patients couldresent to the office for physical examination. In thesenstances, the final physical examination scores at theate of their last follow-up were used along with aetailed telephone interview to confirm the currenttrength and range of motion of the shoulder.

We believe that this study supports the use of ar-hroscopy in the treatment of rotator cuff tears. Moremportantly, this study shows the excellent clinical

uccess that can be achieved with the correct restora-

ion of the rotator cuff anatomy by recognizing theonfiguration of the tear and performing an anatomicepair of the tendon over a prepared bed of bleedingone. As a multitude of studies continue to appear ateetings and in the literature evaluating the pullout

trengths of different types of fixation techniques ofhe musculotendinous unit to bone, this study enforceshe importance of methodically approaching each tearndividually to first recognize and then restore theorrect anatomic relationship of the rotator cuff. Justs the orthopaedic surgeon is trained to recognize theersonality of a fracture and use this personality to andvantage when performing stabilization, one maylso keep in mind that, similarly, each rotator cuff tearas a personality of its own with its own subtleties thathould be appreciated and used to help achieve opti-al end results. Although this report is focused on the

se of the side-to-side repair technique as a sole treat-ent of certain rotator cuff tears, the use of tendon-

o-bone fixation is obviously still used in the treatmentf the majority of the tears that we encounter. In thiseries, 98% of patients had a good to excellent resultccording to the UCLA shoulder score. This studyhows that patients with a full-thickness defect of theotator cuff tendon, with anatomy amenable to side-o-side closure, may be effectively treated with aurely arthroscopic repair using solely a side-to-sideuturing technique, and have excellent long-term clin-cal results.

REFERENCES

1. Baylis RV, Wolf EM. Arthroscopic rotator cuff repair: Clinicaland second look assessment. Presented at the Annual Meetingof the Arthroscopy Association of North America, San Fran-cisco, CA, May 1995.

2. Gartsman GM, Khan M, Hammerman SM. Arthroscopic repairof full-thickness tears of the rotator cuff. J Bone Joint Surg Am1998;80:832-840.

3. Tauro JC. Arthroscopic rotator cuff repair: Analysis of tech-nique and results at a 2- and 3-year follow-up. Arthroscopy1998;14:45-51.

4. Gazielly DF, Gleyze P, Montagnon C, Thomas T. Arthro-scopic repair of distal supraspinatus tears with Revo screw andpermanent mattress sutures—A preliminary report. Presentedat the Annual Meeting of the American Shoulder and ElbowSurgeons, Amelia Island, FL, March 1996.

5. Snyder SJ, Mileski RA, Karzel RP. Results of arthroscopicrepair. Presented at the Annual Meeting of the AmericanShoulder and Elbow Surgeons, Amelia Island, FL, March1996.

6. Weber SC. Arthroscopic versus mini-open rotator cuff repair.Presented at the Arthroscopy Association of North AmericaFall Course, San Diego, CA, October 1999.

7. Wolf EM, Pennington WT, Agrawal V. Arthroscopic rotatorcuff repair: 4- to 10-year results. Arthroscopy 2004;20:5-12.

1

1

1

1

1

1

1

1

1

1

2

887ARTHROSCOPIC SIDE-TO-SIDE ROTATOR CUFF REPAIR

8. McLaughlin HL. Lesions of the musculotendinous cuff of theshoulder: The exposure and treatment of tears with retraction.J Bone Joint Surg 1944;26:31-51.

9. Burkhart SS. A stepwise approach to arthroscopic rotatorcuff repair based on biomechanical principles. Arthroscopy2000;16:82-90.

0. Ellman H, Hanker G, Bayer M. Repair of the rotator cuff:End-result study of factors influencing reconstruction. J BoneJoint Surg Am 1986;68:1136-1144.

1. Burkhart SS, Johnson TC, Wirth MA, Athanasiou KA. Cyclicloading of transosseous rotator cuff repairs: “Tension overload”as a possible cause of failure. Arthroscopy 1997;13:172-176.

2. Burkhart SS, Diaz Pagan JL, Wirth MA, Athanasiou KA.Cyclic loading of anchor-based rotator cuff repairs: Confirma-tion of the tension overload phenomenon and comparison ofsuture anchor fixation with transosseous fixation. Arthroscopy1997;13:720-724.

3. Burkhart SS, Athanasiou KA, Wirth MA. Margin conver-gence: A method of reducing strain in massive rotator cufftears. Arthroscopy 1996;13:335-338.

4. Burkhart SS, Esch JC, Jolson SJ. The rotator crescent and

rotator cable: An anatomic description of the shoulder’s“suspension bridge.” Arthroscopy 1993;9:611-616.

5. Clark JM, Harryman DT. Tendons, ligaments and capsule ofthe rotator cuff. Gross and microscopic anatomy. J Bone JointSurg Am 1992;74:713-725.

6. Liu SH, Baker CL. Arthroscopically assisted rotator cuff re-pair: Correlation of functional results with integrity of the cuff.Arthroscopy 1994;10:54-60.

7. Lundberg BJ. The correlation of clinical evaluation with op-erative findings and prognosis in rotator cuff rupture, In:Bayley JI, Kessel L, eds. Shoulder surgery. Berlin: Springer,1982;35-38.

8. Calvert PT, Packer NP, Stoker DJ, Bayley JI, Kessel L. Ar-thrography of the shoulder after operative repair of the tornrotator cuff. J Bone Joint Surg Br 1986;68:147-150.

9. Harryman DT, Mack LA, Wang KY, Jackins SE, RichardsonML, Matsen FA. Repairs of the rotator cuff. Correlation offunctional results with integrity of the cuff. J Bone Joint SurgAm 1991;73:982-989.

0. Gazielly DF, Gleyze P, Montagnon C. Functional and anatom-

ical results after rotator cuff repair. Clin Orthop 1994;304:43-53.