Arthroscopic RotatorCuff Repair

AbstractArthroscopic rotator cuff repair is being performed by an increasingnumber of orthopaedic surgeons. The principles, techniques, andinstrumentation have evolved to the extent that all patterns andsizes of rotator cuff tear, including massive tears, can now berepaired arthroscopically. Achieving a biomechanically stableconstruct is critical to biologic healing. The ideal repair constructmust optimize suture-to-bone fixation, suture-to-tendon fixation,abrasion resistance of suture, suture strength, knot security, loopsecurity, and restoration of the anatomic rotator cuff footprint (thesurface area of bone to which the cuff tendons attach). By achievingoptimized repair constructs, experienced arthroscopic surgeons arereporting results equal to those of open rotator cuff repair. Assurgeons’ arthroscopic skill levels increase through attendance atsurgical skills courses and greater experience gained in theoperating room, there will be an increasing trend towardarthroscopic repair of most rotator cuff pathology.

The past decade has seen a majorchange in the way that rotator

cuff surgery is performed. Thischange is the direct result of a para-digm shift in cuff repair. The para-digm shift itself is based on soundbiomechanical principles coupledwith a technological shift in thedevelopment of reliable, procedure-specific arthroscopic instrumenta-tion. In a 2003 survey of 908 mem-bers of the Arthroscopy Associationof North America (AANA) and theAmerican Orthopaedic Society forSports Medicine, the 700 surgeonswho responded reported that theyperformed 24% of their rotator cuffrepairs via an all-arthroscopic tech-nique, whereas 5 years previously,they had performed only 5% of cuffrepairs arthroscopically.1 More re-cently, at the 2005 Annual Meetingof the American Academy of Ortho-

paedic Surgeons, 167 orthopaedicsurgeons attending a rotator cuff sym-posium session were polled electron-ically. In response to the question ofhow they would repair a mobile 3-cmrotator cuff tear, 62% answered thatthey would repair it arthroscopical-ly.2 These surveys demonstrate a sub-stantial increase in the number of ar-throscopic cuff repairs in just 7 years,and that rate is expected to acceler-ate during the next several years.

Thus, surgeons today are in themidst of a philosophical and tech-nological paradigm shift that holdsgreat promise for improving thewell-being of patients who need ro-tator cuff surgery. To fully appreciatethe scope of the changes requires un-derstanding the current state of theart in arthroscopic rotator cuff re-pair, examining its development,and imagining its future.

Stephen S. Burkhart, MD

Ian K. Y. Lo, MD

Dr. Burkhart is Director of OrthopaedicEducation, The San AntonioOrthopaedic Group, San Antonio, TX.Dr. Lo is Associate Professor,Department of Surgery, The University ofCalgary, Calgary, AB, Canada.

Dr. Burkhart or the department withwhich he is affiliated has receivedroyalties from Arthrex. Dr. Burkhart orthe department with which he isaffiliated serves as a consultant to or isan employee of Arthrex. Neither Dr. Lonor the department with which he isaffiliated has received anything of valuefrom or owns stock in a commercialcompany or institution related directly orindirectly to the subject of this article.

Reprint requests: Dr. Burkhart, The SanAntonio Orthopaedic Group, Suite 300,400 Concord Plaza Drive, San Antonio,TX 78216.

J Am Acad Orthop Surg 2006;14:333-346

Copyright 2006 by the AmericanAcademy of Orthopaedic Surgeons.

Volume 14, Number 6, June 2006 333

Biomechanical Functionas a Predictor ofAnatomic Structure

Form follows function. The humanbody is a prime example of this tenetof nature. When function is impaired,there is a derangement in form, oranatomy. Therefore, restoration offunction requires solving the puzzleof how to restore the anatomy. Thesurgeon is presented with a struc-tural disruption that can be repairedin one of several ways. The difficultyis in choosing the repair pattern(form) that will restore function, re-alizing that function is derived onlyfrom the proper application of thephysical principles of biomechanics.

Recognizing the direct relation-ship between normal anatomy andnormal biomechanics is the key tounderstanding how to properly re-pair disrupted tissue, whether by ar-throscopic or open techniques. Inthe shoulder, as elsewhere in thebody, biomechanical function is aconsequence of the basic physiolog-ic anatomic structure as well as ofthe interaction of tendons, muscles,and ligaments.

Balance of Force CouplesThe intuitive solution to the

problem of a torn rotator cuff is to“cover the hole.” This can be doneby techniques that disregard biome-chanics (eg, subscapularis tendontransfer, freeze-dried allograft) to re-pair large rotator cuff defects. Unfor-tunately, the history of rotator cuffrepair is filled with biomechanicallyunsound procedures.

The muscles of the rotator cuffand the extrinsic shoulder musclesare positioned to create momentsabout the shoulder that will producespecific rotational motion. Further-more, the shoulder can maintain astable fulcrum of motion only whenit maintains balanced force couples(ie, balanced moments) in both thecoronal and transverse planes3-5 (Fig-ure 1). When these force couples aredisrupted by a massive rotator cuff

Figure 1

A, Coronal plane force couple. The inferior portion of the rotator cuff (below thecenter of rotation [O]) creates a moment that must balance the deltoid moment.B, Axillary view of transverse plane force couple. The subscapularis anteriorly isbalanced against the infraspinatus and teres minor posteriorly. a = moment arm ofthe inferior portion of the rotator cuff, A = moment arm of the deltoid, C = resultantof rotator cuff forces, D = deltoid force, ΣM0 = the sum of the moments about thecenter of rotation (O), I = infraspinatus, r = moment arm of the subscapularis,R = moment arm of the infraspinatus and teres minor, S = subscapularis

Figure 2

A, The transverse plane force couple (left) and the coronal plane force couple (right)are disrupted by a massive rotator cuff tear involving the posterior rotator cuff,infraspinatus, and teres minor. B, An alternative pattern of disruption of thetransverse plane force couple. The transverse plane force couple is disrupted by amassive tear involving the anterior rotator cuff (ie, subscapularis). D = deltoid, I =infraspinatus, O = center of rotation, S = subscapularis, TM = teres minor

Arthroscopic Rotator Cuff Repair

334 Journal of the American Academy of Orthopaedic Surgeons

tear, the shoulder can no longermaintain a stable fulcrum, and activemotion is lost (Figure 2). Thus, anoverriding principle of cuff repairmust be the restoration of balancedforce couples. When rotator cuff re-pair restores the anterior and poste-rior forces so that the moments be-

tween the two are balanced, functionwill be restored, even in the presenceof a persistent defect in the superiorrotator cuff (supraspinatus).

The Rotator CableAnother strong anatomic clue is

related to the biomechanical func-

tion of the rotator cuff. The intactrotator cuff demonstrates an arch-ing, cable-like thickening surround-ing a thinner crescent of tissue thatinserts into the greater tuberosity ofthe humerus; this is known as thecable-crescent complex.6 This cable-like structure represents a thicken-ing of the coracohumeral ligamentand consistently is located at themargin of the avascular zone.7 Therotator cable extends from its anteri-or attachment just posterior to thebiceps tendon to its posterior attach-ment near the inferior border of theinfraspinatus tendon (Figure 3). Thisrotator cable may function in a waythat is analogous to a load-bearingsuspension bridge. By this model,stress is transferred from the cuffmuscles to the rotator cable as a dis-tributed load, thereby stress-shielding the thinner, avascular cres-cent tissue, particularly in olderindividuals.

A rotator cuff tear similarly canbe modeled after a suspensionbridge, with the free margin of thetear corresponding to the cable andthe anterior and posterior attach-ments of the tear corresponding tothe supports at each end of the ca-ble’s span8 (Figure 4). By this model,the supraspinatus muscle, even witha supraspinatus tendon tear, can stillexert its compressive effect on theshoulder joint by means of its dis-tributed load along the span of thesuspension bridge configuration.Halder et al9 confirmed the validityof this suspension bridge model inan in vitro biomechanical study.

Seeing and Reachingthe Pathology

In general, if we can see the tear andreach it with arthroscopic instru-ments, we can repair it arthroscopi-cally. Seeing the tear involves con-trolling bleeding, which requires athorough understanding of fluid me-chanics and turbulence control. Toreach and repair the tear, the surgeonmust possess an in-depth knowledge

Figure 3

Superior (A) and posterior (B) projections of the rotator cable and crescent. Therotator cable extends from the biceps tendon to the inferior margin of theinfraspinatus, spanning the supraspinatus and infraspinatus insertions. B =mediolateral diameter of the rotator crescent, BT = biceps tendon, C = width ofrotator cable, I = infraspinatus, S = supraspinatus, TM = teres minor

Figure 4

A, A rotator cuff tear can be modeled after a suspension bridge. B, The free margincorresponds to the cable, and the anterior and posterior attachments of the tearcorrespond to the supports at each end of the cable’s span.

Stephen S. Burkhart, MD, and Ian K. Y. Lo, MD

Volume 14, Number 6, June 2006 335

of the anatomy of the shoulder to es-tablish safe portals at an angle of ap-proach that allows repair.

Seeing the TearThere are four factors to consider

in the control of bleeding: (1) patientblood pressure, (2) arthroscopicpump pressure, (3) rate of fluid flow,and (4) turbulence within the sys-tem. Generally, in the absence ofmedical contraindications, the pa-tient’s systolic blood pressure ismaintained between 90 and 100 mmHg. The arthroscopic pump pressureis run at 60 mm Hg; if bleeding im-pairs visualization, the pressure maybe temporarily increased to 75 mmHg for short periods (ie, 10 to 15min). When possible, a dedicatedseparate 8-mm inflow cannula isused to maximize flow. It is best notto use arthroscopic instrumentsthrough the inflow cannula becausethey may create turbulence thathampers visualization.

The single most important (andleast recognized) factor in control-ling bleeding is turbulence con-trol.10 Turbulence results from rapidfluid flow out of the shoulder, whichcan occur through a skin portal that

does not have a cannula. In such asituation, the Bernoulli effect cre-ates a force at right angles to thestreaming fluid that sucks the bloodfrom the many capillaries in the sub-acromial space. Bleeding caused byturbulence can be controlled by ob-structing fluid egress using simpledigital pressure over the skin portals.This mechanism is separate frompressure distention of the subacro-mial space. In fact, increasing thepressure in the presence of turbulentflow makes the problem worse be-cause it increases the speed of egressof the fluid from the skin portal,thereby increasing turbulence. Sim-ilarly, “chasing the bleeders” withelectrocautery is counterproductivebecause the surgeon can never cau-terize all of the bleeders as long asthere is turbulence.

Reaching the TearTo reach all parts of the cuff safe-

ly, the surgeon must be familiar witha variety of portals, including themodified Neviaser portal and thesubclavicular portal.11 Some sur-geons prefer retrograde suture pas-sage, with or without suture shut-tles. When tissue quality is good,

antegrade suture passage through alateral portal is satisfactory. Howev-er, when tissue quality is suspect,the surgeon should perform retro-grade suture passage to capture alarger bridge of tendon tissue formore secure fixation.

Tear Patterns

Open rotator cuff repair is performedthrough an anterolateral window (ie,incision). The surgeon must bringthe cuff to that window to work onit; this spatial constraint can make itvery difficult for the surgeon per-forming open cuff repair to properlyevaluate and repair the tear. Further-more, the anterolateral window hasfostered a mindset of medial-to-lateral repair in surgeons performingopen cuff repair.

Arthroscopy, however, frees thesurgeon from spatial constraints.The surgeon may assess and ap-proach the rotator cuff from severaldifferent angles to fully delineate thetear pattern and then repair it ana-tomically. Tear patterns are classi-fied based on the shape and mobili-ty of the tear margins. In ourexperience, most rotator cuff tearscan be classified into four major cat-egories: (1) crescent-shaped, (2) U-shaped, (3) L-shaped, and (4) massive,contracted, immobile tears.

Crescent-shaped tears are theclassic standard tears, with excellentmedial-to-lateral mobility, regard-less of size (Figure 5, A). These tearsmay be repaired directly to bonewith minimal tension (Figure 5, B).

U-shaped tears extend much far-ther medially than do crescent-shaped tears, with the apex of thetear adjacent to or medial to the gle-noid rim (Figure 6, A). Recognizingthis tear pattern is critical becauseattempting to medially mobilize andrepair the apex of the tear to a later-al bone bed will result in extremetensile stresses in the middle of therepaired cuff margin, causing tensileoverload and subsequent failure. Se-quential side-to-side suturing, from

Figure 5

A, Superior view of a crescent-shaped rotator cuff tear involving the supraspinatus(SS) and infraspinatus (IS) tendons. B, Crescent-shaped tears demonstrateexcellent mobility from a medial-to-lateral direction and may be repaired directly tobone.

Arthroscopic Rotator Cuff Repair

336 Journal of the American Academy of Orthopaedic Surgeons

medial to lateral, of the anterior andposterior leaves of the tear causesthe free margin of the rotator cuff toconverge toward the bone bed on thehumerus (Figure 6, B). The free mar-gin of the rotator cuff can then beeasily repaired to the bone bed in atension-free manner (Figure 6, C).The technique of margin conver-gence not only allows repair of seem-ingly irreparable tears but also min-

imizes strain at the repair site,thereby providing an added degree ofprotection for the tendon-to-bonecomponent of the repair.

L-shaped and reverse L-shapedtears are similar to U-shaped tears.However, in the L-shaped tear, oneof the leaves is more mobile than theother leaf and can be more easilybrought to the bone bed and to theother leaf (Figure 7, A). The apex of

the L is identified, and the longitudi-nal split is sutured in a side-to-sidemanner (Figure 7, B), after which theconverged margin is repaired to bone(Figure 7, C).

In the chronic L-shaped tear, thephysiologic pull of the rotator cuffmuscles posteriorly causes the tearto assume a more U-shaped configu-ration (Figure 8, A). It is imperativethat the surgeon determine which

Figure 6

A, Superior view of a U-shaped rotator cuff tear involving the supraspinatus (SS) and infraspinatus (IS) tendons. B, The firststep in repair is done with side-to-side sutures using the principle of margin convergence. C, The free margin is then repaired tobone in a tension-free manner.

Figure 7

A, Superior view of an acute L-shaped rotator cuff tear involving the supraspinatus (SS) and rotator interval (RI). B, The acuteL-shaped tear should be initially repaired along the longitudinal split. C, The converged margin is then repaired to bone. CHL =coracohumeral ligament, IS = infraspinatus, Sub = subscapularis

Stephen S. Burkhart, MD, and Ian K. Y. Lo, MD

Volume 14, Number 6, June 2006 337

leaf is more mobile and where the“corner” of the L-shaped tear needsto be restored. First, a traction sutureis placed at the corner to visually es-tablish its location. Subsequently,side-to-side suture along the longitu-dinal split will accomplish marginconvergence (Figure 8, B), followedby repair of the converged margin tobone (Figure 8, C).

In our experience, these first threetear patterns represent >90% of pos-terosuperior rotator cuff tears.12 Us-ing the principles outlined above,most rotator cuff tears, even massiveones, can be arthroscopically re-paired without extensive mobiliza-tion. Repairing a tear according to itspattern can produce excellent re-sults. However, the fourth tear pat-

tern (ie, the massive, contracted, im-mobile rotator cuff tear) requiresmore sophisticated techniques.12

These tears demonstrate no mobilityfrom medial to lateral or from ante-rior to posterior; thus, they cannot berepaired directly to bone or side-to-side with margin convergence. Suchmassive, contracted, immobile cufftears represent 9.6% of the massive

Figure 8

A, Superior view of a chronic L-shaped tear that has assumed a U-shaped configuration. B, L-shaped tear demonstratingexcellent mobility from an anterior to posterior direction; however, one of the tear margins (usually the posterior leaf) is moremobile. The tear is initially repaired using side-to-side sutures (A' → A) via the principle of margin convergence. C, Theconverged margin is then repaired to bone in a tension-free manner. CHL = coracohumeral ligament, IS = infraspinatus, RI =rotator interval, SS = supraspinatus, Sub = subscapularis

Figure 9

A, Superior view of a massive, contracted, immobile longitudinal rotator cuff tear. B, An anterior interval slide is performed,incising the interval between the supraspinatus tendon (SS) and the rotator interval (RI). C, The improved mobility from therelease allows repair of the supraspinatus tendon to a lateral bone bed. D, The posterior leaf of the tear, consisting of theinfraspinatus and teres minor (IS/TM), is then advanced superiorly and laterally, and the residual longitudinal defect is closed byside-to-side sutures. CHL = coracohumeral ligament, Sub = subscapularis

Arthroscopic Rotator Cuff Repair

338 Journal of the American Academy of Orthopaedic Surgeons

tears in the senior author’s referral-based practice and probably a muchlesser percentage in most other prac-tices.12 They are difficult to managebecause simple arthroscopic débride-ment typically does not result in anyincrease in strength, motion, or func-tion, although pain relief may beachieved.13

These massive tears present inone of two patterns: massive con-tracted longitudinal (Figure 9) andmassive contracted crescent (Fig-ure 10). In massive contracted longi-tudinal tears, the “tongue” of su-praspinatus tendon at the anteriormargin of the tear is useful during re-pair after appropriate mobilization. In

general, massive contracted crescenttears are wider in an anterior-to-posterior direction than the massivecontracted longitudinal tears, makingthem more difficult to repair.

Although these two contractedimmobile tear patterns were previ-ously considered irreparable (a termthat continues to be redefined),many large tears may now be re-paired via advanced arthroscopicmobilization techniques. These ad-vanced arthroscopic mobilizationtechniques are modifications of in-terval slide techniques performedduring open surgery.14-19

The arthroscopic anterior intervalslide, originally described by Tau-

ro,18 improves the mobility of the ro-tator cuff by releasing the intervalbetween the supraspinatus tendonand the rotator interval, effectivelyincising the coracohumeral liga-ment, which becomes contracted inthese tears. A cutting instrument isintroduced through a lateral or ac-cessory lateral portal, and the releaseis oriented toward the base of thecoracoid (Figure 9, B). Using an an-terior interval slide typically gains1 to 2 cm of additional lateral excur-sion of the supraspinatus tendon. Formassive contracted longitudinaltears, this is often adequate to allowtendon-to-bone repair of the su-praspinatus (Figure 9, C), followed

Figure 10

A, Superior view of a massive, contracted, immobile crescent rotator cuff tear. B, A double interval slide is performed, consistingfirst of an anterior interval slide followed by a posterior interval slide, releasing the interval between the supraspinatus (SS) andinfraspinatus (IS) tendons. C, After release there is improved mobility (arrows) of the supraspinatus tendon and the infraspinatus/teres minor (IS/TM) posteriorly. D, The supraspinatus is then repaired to a lateral bone bed in a tension-free manner, and theinfraspinatus/teres minor tendons are advanced laterally and superiorly. E, The residual defect is closed with side-to-sidesutures. CHL = coracohumeral, Sub = subscapularis

Stephen S. Burkhart, MD, and Ian K. Y. Lo, MD

Volume 14, Number 6, June 2006 339

by side-to-side suture plus advance-ment of the infraspinatus on thebone bed, if possible (Figure 9, D).

For massive contracted crescenttears, 1 to 2 cm of additional lateralexcursion of the tendon usually isnot enough to allow tendon-to-bonerepair. In such cases, a double inter-val slide (ie, anterior interval slideplus posterior interval slide) oftencan achieve dramatic improvementin lateral excursion (up to 5 cm of ad-ditional lateral mobility of both thesupraspinatus and the infraspinatus)(Figure 10). The posterior intervalslide generally improves the lateralexcursion of the infraspinatus tendonsufficiently to repair most or all ofthe infraspinatus to bone. Repair of atleast the inferior half of the in-fraspinatus is critical because thatamount of functional infraspinatus isnecessary to restore the posterior mo-ment, thereby balancing the trans-verse plane force couple to reestab-lish a normal glenohumeral fulcrum.

When performing a posterior in-terval slide, the scapular spine mustbe cleared of its surrounding sub-acromial fibroadipose tissue so thatit can be clearly seen through a later-

al viewing portal (Figure 11, A).When cleared of soft tissue, the scap-ular spine resembles the keel of aboat, and it serves as a marker be-tween the supraspinatus and in-fraspinatus (Figure 11, B).

The suprascapular nerve is poten-tially at risk during the posterior in-terval slide. It curves tightly aroundthe base of the scapular spine at itsjunction with the posterior glenoidneck, enveloped within a fat pad.20

Care must be taken to lift the softtissue away from the bone when cut-ting, avoiding the bone surfaces withthe cutting instrument, and stoppingthe release when the fat pad at thebase of the scapular spine is visible(Figure 11, C). After the release, if acomplete repair is not possible, asmuch of a partial repair as is possibleshould be done.21

Patients with single- and double-interval slides have demonstratedmarked improvement not only inpain, but also in active range of mo-tion, strength, and overall shoulderfunction.12 Because these results arefar superior to débridement alone,we no longer perform arthroscopicdébridement alone.

Biomechanics ofRotator Cuff Fixation

During the past 12 years, several bio-mechanical studies have been con-ducted with the collective intent ofoptimizing the strength of rotatorcuff fixation.22-28 As a result, currentcuff fixation constructs are greatlysuperior to those of just a few yearsago. Even so, the importance of tearpattern recognition must not be un-derestimated. An incorrectly re-paired tear may have such largestrains at the tear margin that a se-cure repair cannot be accom-plished—for example, a U-shapedtear repaired strictly by medial-to-lateral fixation to bone, withoutmargin convergence sutures.

For U- and L-shaped tears, side-to-side sutures that achieve marginconvergence have the unique biome-chanical property of significantly re-ducing strain at the tear margin.29

Side-to-side closure of two thirds ofa U-shaped tear reduces strain at thetear margin by a factor of six, thusdramatically enhancing the securityof the fixation and safeguarding

Figure 11

Arthroscopic images of the steps in a posterior interval slide. A, Arthroscopic view of a right shoulder from a lateral portaldemonstrating the arching column of the scapular spine (Sp), which serves as a marker between the supraspinatus (SS) andinfraspinatus (IS) tendons. A Viper suture passer (Arthrex, Naples, FL) is used to pass traction sutures into each tendon.B, Release of the posterior margin of the supraspinatus tendon from the infraspinatus using a basket punch. C, Completedposterior interval slide demonstrating complete release of the infraspinatus from the supraspinatus, revealing the scapularspine. Care is taken to avoid injury to the underlying suprascapular nerve. G = glenoid

Arthroscopic Rotator Cuff Repair

340 Journal of the American Academy of Orthopaedic Surgeons

against failure of the repair con-struct.

Under cyclic loads, transosseousrotator cuff repair constructs failat low numbers of cycles becauseof cutting of the suture throughbone.22,23 Transosseous rotator cuffrepair constructs were particularlysusceptible to this mode of failurewhen the distal bone holes weremade through metaphyseal bonerather than through thick corticalbone.

In contrast, under cyclic loads, ro-tator cuff repair constructs securedto bone by suture anchors did notfail at the bone-suture anchor inter-face but rather because the suturecut through the tendon. On average,these failures occurred at higher cy-cles than similar transosseous re-pairs. Thus, by using suture anchors,the weak link was effectively trans-ferred from the bone to the rotatorcuff tendon.22

After transferring the weak linkto the suture-tendon interface, thenext task was to optimize that weaklink by finding the best way to min-imize suture cutout through tendon.Doubling the number of fixationpoints of suture to tendon would re-duce the load in each suture by 50%.The easiest way to do this was todouble-load each suture anchor.Burkhart et al24 confirmed the loadreduction of this construct experi-mentally.

Optimization of the anchor pull-out strength is desirable. It has beenshown mathematically that the pull-out angle (ie, deadman angle), whichis the angle of insertion of the sutureanchor, and the tension-reductionangle both should be ≤45° (Figure12). A deadman angle ≤45° will sig-nificantly resist anchor pullout.30

A second factor in choosing ananchor relates to suture abrasion.Suture abrasion that primarily arisesduring arthroscopic knot tying canoccur at the knot itself or as the su-ture slides through the anchor eye-let, abrades against bone, or slidesthrough the knot pusher. Lo et al25

tested a variety of metal and biode-gradable polymer anchors for sutureabrasion and confirmed the findingsof Bardana et al,26 who showed thatmetal anchors demonstrate marked-ly more suture abrasion than dopolymer biodegradable anchors.However, the results among variousdesigns of biodegradable anchorswere markedly different, as well.One common biodegradable anchordesign incorporates a hole throughthe polymer body to form the sutureeyelet (Panalok RC anchor and3.5-mm Panalok; Mitek, Westwood,MA). This design failed by cutting ofthe Ethibond suture (Ethicon, Som-erville, NJ) through the suture eyeletat low cycles (11.2 cycles to failurefor the Panalok RC; 12.5 cycles tofailure for the 3.5-mm Panalok). Inanother biodegradable suture anchordesign, the eyelet consists of a flex-ible loop of no. 5 polyester suturethat is insert-molded into the body

of the anchor (Biocorkscrew; Ar-threx, Naples, FL). Cyclic abrasiontesting demonstrated failure throughthe Ethibond suture rather thanthrough the eyelet of the anchor, atsignificantly higher cycles than theother biodegradable anchor designs.

Many factors influence the choiceof suture anchor. To be effective, theanchors must possess certain mini-mal strength characteristics. In gen-eral, nearly all commercially avail-able suture anchors satisfy theseminimal strength requirements.

Optimizing suture type forstrength, abrasion resistance, andknot security was the next step inimproving the strength of rotatorcuff fixation. Lo et al25 comparedno. 2 Ethibond, the most commonlyused suture in rotator cuff repair,with no. 2 Fiberwire (Arthrex). No. 2Fiberwire is a recently introducedbraided, nonabsorbable, polyblendsuture with an ultimate strength tofailure that is roughly equivalent tothat of no. 5 Ethibond. Interestingly,no. 2 Fiberwire failed at notablyhigher cycles compared with no. 2Ethibond under all tested abrasionconditions, even when using metal-lic anchors.25 In fact, no. 2 Fiberwirefailed at cycles 5 to 51 times higherthan did no. 2 Ethibond and at cyclicloading levels at which abrasionfrom the suture eyelet is probablyclinically irrelevant. Other high-strength sutures have recently be-come available and may be preferredby some surgeons.

The final component needed tooptimize rotator cuff repair is the ar-throscopic knot. The ideal knotmust achieve the optimal balance ofknot security and loop security.Knot security, which is the effective-ness of the knot at resisting slippagewhen load is applied, depends onthree factors: friction, internal inter-ference, and slack between throws.Loop security is the ability to main-tain a tight suture loop around tissueas the knot is tied. Thus, it is possi-ble for any knot to have good knotapproximation and security but poor

Figure 12

Proper anchor insertion at the deadmanangle. θ1 is the pullout angle for theanchor (angle the suture makesperpendicular to the anchor). θ2 is thetension-reduction angle (angle thesuture makes with the direction of pullof the rotator cuff). Ideally, θ1 and θ2

should both be ≤45°.

Stephen S. Burkhart, MD, and Ian K. Y. Lo, MD

Volume 14, Number 6, June 2006 341

loop security (eg, a loose suture loop)and, therefore, to be ineffective atapproximating the tissue edges to berepaired (Figure 13).

The best combination of knot se-curity and loop security is the ar-throscopic surgeon’s knot, which isa static (nonsliding) knot composed

of a base knot of three half hitches inthe same direction, followed bythree reversing half-hitches on alter-nating posts (RHAPs)27 (Figure 14).When a complex sliding knot isused, it is optimized and fully“locked” (ie, it will fail by breakagerather than slippage) only if the slid-ing knot is followed by three stackedRHAPs on top of it. Of 12 differentsliding knots that were tested, theRoeder knot with three RHAPs pro-vided the optimal balance of knot se-curity and loop security.27 Even so,the static arthroscopic surgeon’sknot displayed even better charac-teristics of knot security and loop se-curity than any of the sliding knots.

Burkhart et al28 performed a bio-mechanical study comparing theloop security of various half-hitchcombinations when tied with a stan-dard single-hole knot pusher withthe loop security of the same knotstied with the double-diameter knotpusher (Surgeon’s Sixth Finger; Ar-threx). The cannulated double-diameter knot pusher maintainsloop security by means of pressurefrom the tip of the cannulated innerdiameter tube against the first throwof the knot; subsequent throws areadvanced with a sliding plasticouter-diameter tube (Figure 15).Thus, sequential half hitches may besecured without the suture loop be-coming lax. The loop security ofknots tied with the double-diameterknot pusher was superior to that ofother methods.

Finally, restoration of the entirerotator cuff footprint, or tendon-to-bone contact area, will more effec-tively transmit rotator cuff forces tobone and may improve the potentialfor complete healing. A recent studyby Galatz et al31 showed a 94% inci-dence of recurrent rotator cuff de-fects after arthroscopic repair by asingle-row medial-to-lateral tech-nique.

The information on arthroscopicrepair must be interpreted in thecontext of established results foropen rotator cuff repair. Larger tear

Figure 13

A, A tight suture loop holds the soft tissue tightly apposed to the prepared bonebed. B, A loose loop allows the soft tissue to pull away from the prepared bone bed,regardless of how securely the knot is tied.

Figure 14

Static surgeon’s knot. RHAPs = reversing half-hitches on alternating posts

Arthroscopic Rotator Cuff Repair

342 Journal of the American Academy of Orthopaedic Surgeons

size is a predisposing factor forrecurrence.32-36 In fact, Harryman etal33 showed by postoperative ultra-sound that 68% of three-tendon cufftears repaired by open surgery hadrecurrent defects. Many technicaland biologic factors influence the re-sults of rotator cuff repair.

The rotator cuff footprint may becompletely reestablished arthroscop-ically by means of a double-row su-ture anchor repair, assuming thatthere is enough lateral excursion ofthe tendons to allow this. A medialrow of anchors (one or two rows) isplaced adjacent to the articular mar-gin, and sutures are passed to securethe cuff with mattress sutures. Thena lateral row of anchors is placed,with simple sutures to secure thetendon edge.37

The Optimum RepairConstruct

A series of studies was performed todetermine the optimum rotator cuffrepair construct.22-25,27-30,37 The bestconstruct would optimize suture-to-bone fixation, suture-to-tendon fixa-tion, abrasion resistance of the su-ture, suture strength, knot security,and loop security. Research indicatesthat the optimized construct con-sists of doubly loaded biodegradablepolymer suture anchors with insert-molded suture eyelets; no. 2 Fiber-wire suture; and six-throw arthro-scopic surgeon’s knots with threeRHAPs, tied with a double-diameterknot pusher.22-25,27,28,30 When possi-ble, a double-row repair may be used

in an attempt to optimize the foot-print of the repaired rotator cuff. Amedial row of suture anchors isplaced at the articular margin of thebone bed, and the rotator cuff is se-cured with mattress sutures. In addi-tion, a lateral row of suture anchorsis placed on the greater tuberosity,securing the tendon edge with sim-ple sutures.

Subscapularis TendonTears

Arthroscopic repair of the torn sub-scapularis tendon poses unique tech-nical challenges. First, the availableworking space for arthroscopic in-strumentation in this part of theshoulder is quite confined, so we typ-ically repair the subscapularis first,before soft-tissue swelling compro-mises the space even further. Second,coracoid impingement with subcora-coid stenosis is frequently part of theproblem, often necessitating arthro-scopic coracoplasty.38-40 Finally, sub-scapularis tears may be accompaniedby medial subluxation of the biceps,requiring the surgeon to perform ar-throscopic tenotomy or tenodesis ofthe long head of the biceps.41

With chronic retracted subscapu-laris tears, identification of the sub-scapularis tendon can be problemat-ic. We have identified a consistentstructure that appears arthroscopi-cally as a comma-shaped arc of liga-mentous tissue that is always locat-ed at the superolateral border of thesubscapularis and can be used to lo-cate that border. The comma sign is

formed by a portion of the superiorglenohumeral/coracohumeral liga-ment complex that has torn awayfrom the humerus42 (Figure 16). Thisligament complex, which functionsas the medial sling of the biceps ten-don before disruption, is consistent-ly present as a guide to the subscap-ularis.

Figure 16

The comma sign is visible in this rightshoulder with a retracted subscapularistear, as viewed through a posteriorportal. The comma is an arc-shapedstructure adjoining the superolateralborder of the subscapularis tendon(SSc). This structure is formed whenthe medial sling of the biceps tearsaway from its footprint on the lessertuberosity of the humerus, directlyadjacent to the footprint of the uppersubscapularis. The lateral border of thecomma defines the lateral border ofthe subscapularis (dotted line). G =glenoid, H = humerus, M = medial slingof biceps (comma), * = junction ofmedial sling and upper border ofsubscapularis muscle

Figure 15

The double-diameter knot pusher (Surgeon’s Sixth Finger; Arthrex), consists of an inner (I) metallic tube with a sliding plasticouter (O) tube. A suture threader (T) is used to pass the suture limb through the inner metallic tube.

Stephen S. Burkhart, MD, and Ian K. Y. Lo, MD

Volume 14, Number 6, June 2006 343

Massive, Contracted,ImmobileAnterosuperior RotatorCuff Tears

The arthroscopic approach to mobi-lizing contracted immobile tearsmust be different when the subscap-

ularis tendon is chronically tornalong with the supraspinatus and in-fraspinatus (anterosuperior tears) be-cause the coracohumeral ligamenthas a propensity to rigidly shortenand tether the cuff medially. The in-terval slide in continuity is a usefultechnique in these complex cases.43

In this procedure, the coraco-humeral ligament is released by dis-secting it “in continuity” from thebase of the coracoid, preserving anintact lateral bridge of rotator inter-val tissue containing the commasign. This release generally pro-duces enough additional lateral ex-

Figure 17

Interval slide in continuity. A, Arthroscopic lateral portal view of a left shoulder. A shaver is introduced through an accessorylateral portal and is used to dissect and expose the coracoid neck, in essence releasing and excising the origin of coracohumeralligament from the lateral aspect of the coracoid. Here, a coracoplasty has been performed. B, Lateral portal view. A shaver isintroduced through an accessory lateral portal and is used to excise the medial rotator interval while preserving an intact lateralbridge of rotator interval tissue containing the comma sign. The coracohumeral ligament is completely released and excisedfrom the lateral aspect of the base of the coracoid. C, Posterior portal view demonstrating a completed interval slide in continuity.A lateral bridge of rotator interval tissue has been maintained, which contains the comma sign and leads to the superolateralaspect of the torn subscapularis tendon. The coracoid tip and coracoid base are visible on either side of the comma sign. D, Acompleted subscapularis repair. Posterior portal view using a 30° arthroscope demonstrating the intra-articular perspective. Thecomma sign is still intact. The hooked probe demonstrates the interval slide in continuity. E, Lateral portal view demonstratinga residual U-shaped posterosuperior rotator cuff defect, which can be closed with margin convergence. The intact lateral portionof the rotator interval serves as an anterior leaf that can be repaired to the posterior leaf of the rotator cuff. A hooked probe hasbeen placed within the defect created by the interval slide in continuity. F, Lateral portal view following repair of a U-shapedrotator cuff tear. The intact lateral portion of the rotator interval, including the tissue of the comma sign, is incorporated into theside-to-side sutures (arrows). CB = coracoid base, CN = coracoid neck, CT = coracoid tip, G = glenoid, H = humerus, SSc =subscapularis tendon, * = comma sign

Arthroscopic Rotator Cuff Repair

344 Journal of the American Academy of Orthopaedic Surgeons

cursion of the subscapularis to per-mit its direct repair to its bone bedon the lesser tuberosity. The resid-ual defect in the cuff is then closedaccording to standard principles.When margin convergence is indi-cated, the intact lateral portion ofthe rotator interval serves as an an-terior leaf that can be repaired to theposterior leaf of the rotator cuff (Fig-ure 17).

The Future

The future of arthroscopic rotatorcuff repair can be discussed from twostandpoints: educational initiativesin arthroscopic rotator cuff repairand refinements in technique. TheAANA and the AAOS have bothbeen active in providing hands-oncadaveric skills courses at the Or-thopaedic Learning Center in Rose-mont, Illinois. A twice-yearly courseoffered by the AANA to orthopaedicresidents at a reduced price exposessurgeons-in-training to advancedskills at an early stage. Remotetransmission of live arthroscopic ro-tator cuff surgery by the AANA tolarge meetings of surgeons allows in-teractive discussion between theshoulder surgeon and the audienceas the case is being performed.

Refinement of technique is pro-gressing rapidly. Already, the expertarthroscopic surgeon can achievemechanically stable rotator cuff con-structs that are as good as, or betterthan, those that are achievable byopen means. Furthermore, theseconstructs produce excellent clinicaland anatomic results. These resultsshould not be surprising because me-chanical stability always enhancesbiologic healing. Knotless fixation,although appealing, requires furtherdevelopment. Even if knotless an-chors are developed that are as se-cure as those requiring knots, therewill still be a frequent need for side-to-side sutures. Currently, side-to-side suture generally requires knots.

Summary

Arthroscopic rotator cuff repair hasundergone dramatic refinementsduring the past few years. The sur-geon who is accomplished in thesetechniques can now produce resultsthat are as good as or better thanthose produced by open techniques.Achieving a biomechanically stableconstruct is critically important, andthis can now be reproducibly accom-plished via arthroscopic means. Inthe next 10 years, the trend towardall-arthroscopic rotator cuff repairwill increase, and the results willdemonstrate clinical effectiveness.

References

Evidence-based Medicine: There areno level I or level II randomized con-trolled comparative studies refer-enced. Level III, IV, case reported, orcase-controlled studies are presentedalong with several level V expertopinion referenced articles describ-ing surgical techniques.

Citation numbers printed in boldtype indicate references publishedwithin the past 5 years.

1. Fineberg M, Wind W, Stoeckl A, Smo-linski R: Current practices and opin-ions in shoulder surgery: Results of asurvey of members of the AmericanOrthopaedic Society for Sports Medi-cine and the Arthroscopy Associationof North America. 22nd AnnualMeeting Abstracts of Podium Presen-tations. Phoenix, AZ: ArthroscopyAssociation of North America, 2003,p 19.

2. Abrams JS, Savoie FH III: Arthroscop-ic rotator cuff repair: Is it the new goldstandard? 72nd Annual Meeting Pro-ceedings. Rosemont, IL: AmericanAcademy of Orthopaedic Surgeons,2005, p 71.

3. Inman VT, Saunders JB, Abbott LC:Observations on the function of theshoulder joint. J Bone Joint Surg Am1944;26:1-30.

4. Burkhart SS: Arthroscopic debride-ment and decompression for selectedrotator cuff tears: Clinical results,pathomechanics, and patient selec-tion based on biomechanical parame-

ters. Orthop Clin North Am 1993;24:111-123.

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

6. Burkhart SS, Esch JC, Jolson RC: Therotator crescent and rotator cable:An anatomic description of theshoulder’s “suspension bridge.”Arthroscopy 1993;9:611-616.

7. Clark JM, Harryman DT II: Tendons,ligaments, and capsule of the rotatorcuff: Gross and microscopic anatomy.J Bone Joint Surg Am 1992;74:713-725.

8. Burkhart SS: Fluoroscopic compari-son of kinematic patterns in massiverotator cuff tears: A suspension bridgemodel. Clin Orthop Relat Res 1992;284:144-152.

9. Halder AM, O’Driscoll SW, Heers G,et al: Biomechanical comparison of ef-fects of supraspinatus tendon detach-ments, tendon defects, and muscle re-tractions. J Bone Joint Surg Am 2002;84:780-785.

10. Burkhart SS, Danaceau SM, Athana-siou KA: Turbulence control as a fac-tor in improving visualization duringsubacromial shoulder arthroscopy.Arthroscopy 2001;17:209-212.

11. Nord KD, Mauck BM: The new sub-clavian portal and modified Neviaserportal for arthroscopic rotator cuffrepair. Arthroscopy 2003;19:1030-1034.

12. Lo IK, Burkhart SS: Arthroscopic re-pair of massive, contracted, immobilerotator cuff tears using single and dou-ble interval slides: Technique and pre-liminary results. Arthroscopy 2004;20:22-33.

13. Ellman H, Kay SP, Wirth M: Arthro-scopic treatment of full-thickness ro-tator cuff tears: 2- to 7-year follow-upstudy. Arthroscopy 1993;9:195-200.

14. Bigliani LU, Cordasco FA, McIlveenSJ, Musso ES: Operative repair of mas-sive rotator cuff tears: Long-term re-sults. J Shoulder Elbow Surg 1992;1:120-130.

15. Warren RF: Surgical considerationsfor rotator cuff tears in athletes, inJackson DW (ed): Shoulder Surgery inthe Athlete. Rockville, MD: AspenPublications, 1985, pp 73-82.

16. Codd TP, Flatow EL: Anterior acro-mioplasty, tendon mobilization, anddirect repair of massive rotator cufftears, in Burkhead WZ Jr (ed): RotatorCuff Disorders. Baltimore, MD: Wil-liams & Wilkins, 1996, pp 323-334.

17. Arroyo JS, Flatow EL: Management of

Stephen S. Burkhart, MD, and Ian K. Y. Lo, MD

Volume 14, Number 6, June 2006 345

rotator cuff disease: Intact and repair-able cuff, in Iannotti JP, Williams GRJr (eds): Disorders of the Shoulder: Di-agnosis and Management. Philadel-phia, PA: Lippincott Williams &Wilkins, 1999, pp 31-56.

18. Tauro JC: Arthroscopic “intervalslide” in the repair of large rotator cufftears. Arthroscopy 1999;15:527-530.

19. Flatow EL, Klepps S: Arthroscopicmobilization and rotator cuff repair.Video. Rosemont, IL: American Acad-emy of Orthopaedic Surgeons, 2004.

20. Warner JJP, Krushell RJ, Masquelet A,Gerber C: Anatomy and relationshipsof the suprascapular nerve: Anatomi-cal constraints to mobilization of thesupraspinatus and infraspinatus mus-cles in the management of massiverotator-cuff tears. J Bone Joint SurgAm 1992;74:36-45.

21. Burkhart SS: Partial repair of massiverotator cuff tears: The evolution of aconcept. Orthop Clin North Am1997;28:125-132.

22. Burkhart SS, Diaz-Pagàn JL, WirthMA, Athanasiou KA: Cyclic loadingof anchor-based rotator cuff repairs:Confirmation of the tension overloadphenomenon and comparison of su-ture anchor fixation with transos-seous fixation. Arthroscopy 1997;13:720-724.

23. Burkhart SS, Johnson TC, Wirth MA,Athanasiou KA: Cyclic loading oftransosseous rotator cuff repairs:Tension overload as a possible causeof failure. Arthroscopy 1997;13:172-176.

24. Burkhart SS, Wirth MA, Simonich M,Salem D, Lanctot D, Athanasiou K:Knot security in simple sliding knotsand its relationship to rotator cuff re-pair: How secure must the knot be?Arthroscopy 2000;16:202-207.

25. Lo IK, Burkhart SS, Athanasiou K:

Abrasion resistance of two typesof nonabsorbable braided suture.Arthroscopy 2004;20:407-413.

26. Bardana DD, Burks RT, West JR, GreisPE: The effect of suture anchor designand orientation on suture abrasion:An in vitro study. Arthroscopy 2003;19:274-281.

27. Lo IK, Burkhart SS, Chan KC, Athana-siou K: Arthroscopic knots: Deter-mining the optimal balance of loopsecurity and knot security.Arthroscopy 2004;20:489-502.

28. Burkhart SS, Wirth MA, Simonich M,Salem D, Lanctot D, Athanasiou K:Loop security as a determinant of tis-sue fixation security. Arthroscopy1998;14:773-776.

29. Burkhart SS, Athanasiou KA, WirthMA: Margin convergence: A methodof reducing strain in massive rotatorcuff tears. Arthroscopy 1996;12:335-338.

30. Burkhart SS: The deadman theory ofsuture anchors: Observations along aSouth Texas fence line. Arthroscopy1995;11:119-123.

31. Galatz LM, Ball CM, Teefey SA, Mid-dleton WD, Yamaguchi K: The out-come and repair integrity of complete-ly arthroscopically repaired large andmassive rotator cuff tears. J BoneJoint Surg Am 2004;86:219-224.

32. Gazielly DF, Gleyze P, Montagnon C:Functional and anatomic results afterrotator cuff repair. Clin Orthop RelatRes 1994;304:43-53.

33. Harryman DT II, Mack LA, Wang KY,Jackins SE, Richardson ML, MatsenFA III: Repairs of the rotator cuff: Cor-relation of functional results with in-tegrity of the cuff. J Bone Joint SurgAm 1991;73:982-989.

34. Iannotti JP, Bernot MP, Kuhlman JR,Kelly MJ, Williams GR: Postoperativeassessment of shoulder function: A

prospective study of full-thickness ro-tator cuff tears. J Shoulder ElbowSurg 1996;5:449-457.

35. Liu SH, Baker CL: Arthroscopicallyassisted rotator cuff repair: Correla-tion of functional results with integri-ty of the cuff. Arthroscopy 1994;10:54-60.

36. Thomazeau H, Boukobza E, MorcetN, Chaperon J, Langlais F: Predictionof rotator cuff repair results by mag-netic resonance imaging. ClinOrthop Relat Res 1997;344:275-283.

37. Lo IK, Burkhart SS: Double-rowarthroscopic rotator cuff repair: Re-establishing the footprint of therotator cuff. Arthroscopy 2003;19:1035-1042.

38. Lo IK, Burkhart SS: The etiology andassessment of subscapularis tendontears: A case for subcoracoid impinge-ment, the roller-wringer effect, andTUFF lesions of the subscapularis.Arthroscopy 2003;19:1142-1150.

39. Lo IK, Parten PM, Burkhart SS: Com-bined subcoracoid and subacromialimpingement in association with an-terosuperior rotator cuff tears: An ar-throscopic approach. Arthroscopy2003;19:1068-1078.

40. Lo IK, Burkhart SS: Arthroscopic cora-coplasty through the rotator interval.Arthroscopy 2003;19:667-671.

41. Lo IK, Burkhart SS: Arthroscopicbiceps tenodesis using a bioabsorb-able interference screw. Arthroscopy2004;20:85-95.

42. Lo IK, Burkhart SS: The comma sign:An arthroscopic guide to the torn sub-scapularis tendon. Arthroscopy 2003;19:334-337.

43. Lo IK, Burkhart SS: The interval slidein continuity: A method of mobilizingthe anterosuperior rotator cuff with-out disrupting the tear margins.Arthroscopy 2004;20:435-441.

Arthroscopic Rotator Cuff Repair

346 Journal of the American Academy of Orthopaedic Surgeons