reverse shoulder arthroplasty indications, technique, and results

Techniques in Shoulder and Elbow Surgery 6(3):135–149, 2005 � 2005 Lippincott Williams & Wilkins, Philadelphia

Reverse Shoulder ArthroplastyIndications, Technique, and ResultsArmodios M. Hatzidakis, MDWestern Orthopaedics P.C.Denver, CO

Tom R. Norris, MDCalifornia Pacific Medical CenterDepartment of Orthopaedic SurgerySan Francisco, CA

Pascal Boileau, MDHopital de L’Archet—University of NiceDepartment of Orthopaedic Surgery & Sports TraumatologyNice, France

n ABSTRACT

The surgical treatment of glenohumeral arthritis with ro-tator cuff deficiency is a difficult challenge. Hemiarthro-plasty, the standard treatment at this time, is associatedwith satisfactory results in a ‘‘limited goals’’ perspective,but often the clinical results are unpredictable. Elevationafter hemiarthroplasty approximately 90�. Pain reliefcan be inconsistent and can deteriorate over time.Constrained prostheses, including ball and socketand reverse ball and socket designs, were introducedin the 1970s to improve upon the results of arthroplastyin this challenging population. Unfortunately, clinicalresults were inconsistent using these designs, and ratesof mechanical loosening and revision were high. The onlydesign that survived has been the prosthesis of PaulGrammont (Dijon, France). His Delta III Prosthesis(DePuy, Warsaw IN) has been in use in its current formsince 1992, with good clinical results and relatively lowmechanical loosening rates compared with the ball andsocket and reversed ball and socket designs of the past.This design is also utilized by the Tornier AequalisReversed Prosthesis (Tornier SA, Montbonnot, FR). InEurope, and more recently in the United States, this pros-thetic design has proven useful in treating patients withglenohumeral arthritis with extensive cuff deficiency,

proximal humeral fracture nonunions and malunions,and failed arthroplasty with a deficient rotator cuff. Pre-dictably good results can be obtained in these difficultcircumstances, with good pain relief and elevation oftenexceeding the horizontal. Active rotation, however, isusually not improved. Complication rates are low inpatients with cuff tear arthritis and relatively high in re-vision arthroplasty cases. As with all revision operations,be they failed cuff repairs or failed prosthetic replace-ments, there is a higher risk of lingering low grade infec-tions. Results can be optimized and complicationsminimized with proper patient selection, fastidious surgi-cal technique, and proper postoperative rehabilitation.

n HISTORICAL PERSPECTIVE

Treatment of the rotator cuff deficient shoulder with ar-thritis has proven a difficult surgical challenge. Becausepain relief was the goal and results of function wereusually poor, Neer stated that arthroplasty for rotator cuffarthropathy should be judged by ‘‘limited goals’’ criteria.Multiple studies have confirmed empirical evidence thathemiarthroplasty for cuff deficient arthritis has somewhatunpredictable results.1–8 Total shoulder arthroplasty hasbeen abandoned for this group due to the loss of forcecouple of the supraspinatus-deltoid and subsequent earlyglenoid loosening. This is attributable to the ‘‘rocking horsephenomenon’’ with upward subluxation onto the glenoidrim, leading to loosening.9 Although hemiarthroplasty

Address correspondence and reprint requests to: Pascal Boileau, MD,Professor and Chairman. Dept of Orthopaedic Surgery & SportsTraumatology, Hopital de L’Archet—University of Nice, 151 Route deSt Antoine de Ginestiere, 06202 Nice, France (e-mail: [email protected]).

R E V I E W

135Volume 6, Issue 3

provides a smooth surface for articulation with the nativeglenoid and acromion, the biomechanical stabilization ofthe fulcrum for elevation is still deficient. Sometimesthe intact remaining cuff, deltoid, and ‘‘acetabularized’’acromial arch can compensate for this lack of fulcrum,but good early results can deteriorate secondary to pro-gressive erosion of the prosthetic head into the glenoidand acromion (Fig. 1). Pain relief is often good initially,but can deteriorate over time with glenoid arthritis andacromial wear or fracture. The results of hemiarthro-plasty are even more inconsistent in patients who havea ‘‘pseudoparalytic’’ shoulder, in whom attempts at ele-vation are rewarded only with an ineffective shrug. As wehave gained more experience in teasing out subtle causesof this ineffectual shrug associated with massive rotatorcuff tears, we wish to differentiate between those withirreparable cuff tears with true pseudo paralysis of theshoulder (PPS) and anterior superior subluxation dueto muscle imbalance, and those patients with painful lossof elevation (PLE) due to an hourglass or otherwise pain-ful biceps. With PPS humeral head arthroplasty andtendon reconstructions have been ineffectual, whereaswith PLE arthroscopic biceps tenotomy has allowedmany to regain overhead elevation.

Multiple constrained and semiconstrained prostheseshave been designed to compensate for the loss of rota-tional fulcrum caused by severe degeneration of the ro-tator cuff. Previous reverse prostheses were designedby Reeves, Liverpool, Bayley, Neer and Averill, Kolbel,Kessel, Fenlin, and Gerard in the 1970s and 1980s. Dis-appointing results and high complication rates led toabandonment of these prostheses.10,11 The common de-nominator for failure was loosening of the glenoid

component. The center of rotation was lateral to the gle-noid and created high torque on the relatively frailglenoid. The reverse prosthesis (RP) of Professor PaulGrammont (Dijon, France) was also designed specificallyfor the rotator cuff deficient shoulder.12–14 It differedfrom previous prostheses in that the glenoid componentwas larger, and its center of rotation was medializedon to the glenoid face, thus decreasing torque on thecomponent-bone fixation. The humeral neck angle isnonanatomic and made more horizontal. This increasesstability of the new glenohumeral joint. In short tomedium term studies up to 10 years, this prosthesishas succeeded where others have failed, predominatelybecause of its favorable biomechanical characteristics.Short and medium-term follow-up studies in Europe havedetermined that this specific design leads to functionalrestoration of elevation and favorable survivorship,with relatively low rates of early glenoid loosening.15–22

Two current versions of the prosthesis have been usedin Europe; since 1992 for the Delta and since 1998 forthe Tornier. Respectively, these prostheses have beenavailable in the United States since FDA approval inNovember 2003 (DePuy, Warsaw IN) and May 2004(Tornier SA, Montbonnot, FR).

BiomechanicsThe RP design of Grammont combines a large hemi-spherical glenoid component (36 or 42 mm diameter)that has a medialized center of rotation and secure screwfixation to the glenoid, with a cemented humeral compo-nent that has a more horizontally aligned (155�) meta-physeal neck with a reverse hemispherical polyethylenecup that perfectly conforms to the glenosphere (Fig. 2).

FIGURE 1. A, Humeral head replacement placed for shoulder arthritis with a massive irreconstructible deficiency of thesupraspinatus and subscapularis. B, The humeral head ceases to ‘‘roll’’ on the glenoid and turns into a ‘‘wedge,’’ leadingto (C) superior erosion of the acromion and glenoid. D, The functional result is poor, with forward flexion of approximately30� and persistent pain.

Hatzidakis et al

136 Techniques in Shoulder and Elbow Surgery

This design allows for more range of motion than reverseball and socket designs of the past (Fig. 3). With properplacement of this prosthesis, the center of humeral rota-tion is made more distal, placing the deltoid muscle atgreater tension (Fig. 4), and medialized, which recruitsmore deltoid muscle fibers for elevation and increasesthe deltoid lever arm.3 These factors increase the del-toid’s biomechanical ability to generate rotational torque,elevating the arm on a new stable fulcrum (the gleno-sphere). The medialized center of rotation decreasestorsional forces on glenosphere fixation, decreasing thetendency of the glenoid component to loosen (Fig. 5).

n INDICATIONS/CONTRAINDICATIONS

There are 3 specific circumstances where we have foundthe RP to be useful. The common characteristic of theseindications is that the biomechanical fulcrum for eleva-tion is lost, either via irreparable loss of the rotator cuffor severe malposition of the tuberosities.

The first indication is primary osteoarthritis of theshoulder with massive irreparable cuff tear, in patientswith secondary osteoarthritis or osteonecrosis who havea ‘‘decompensated’’ cuff deficiency (‘‘Cuff Tear Arthri-tis,’’ CTA). These patients typically suffer from signifi-cant pain and cannot lift their arms to the horizontaldespite attempts to strengthen the deltoid and remainingintact cuff. Many of these patients have had one or manyattempts to repair the rotator cuff, and some may exhibit

FIGURE 2. The Reversed Shoulder Prosthesis� (TornierInc., Tornier SA, Montbonnot, FR).

FIGURE 3. A reversed design of thepast utilizes a small prosthetic head withhighly constrained humeral cup, leadingto decreased potential range of motioncompared with the Grammont design.

Reverse Shoulder Arthroplasty


marked anterosuperior escape of the humeral head withattempted elevation. Patients with anterosuperior escapeare usually not helped by humeral hemiarthroplasty, evenwith attempted coracoacromial arch reconstruction,making them ideal candidates for reversed shoulderarthroplasty.

The second indication is severe fracture sequelae(FS). These patients have had a displaced, usually com-minuted fracture of the proximal humerus with tuberositymalposition or nonunion. Although an unconstrainedprosthesis can reliably provide pain relief and restore

function in those cases where the prosthesis can be‘‘adapted’’ to slight tuberosity malalignment, patientswho have unhealed tuberosities or require osteotomy ofthe greater tuberosity do not typically fare as well.23 AnRP can consistently provide both pain relief and restora-tion of active elevation in this specific patient group.

The third indication for reverse shoulder arthroplastyis revision of previously failed arthroplasty. Typicallythese patients have had a hemiarthroplasty placed forCTA or comminuted proximal humerus fracture. InCTA cases, common causes for failure are loss of the

FIGURE 4. A, Deltoid elevation torquein the cuff deficient shoulder is de-creased secondary to proximal humeralmigration, whereas elevation torque inthe Grammont reverse shoulder (B) isincreased because of a medializedand lowered center of rotation (L2 3

F2 . L1 3 F1).

FIGURE 5. A, The lateralized center ofrotation and constrained design of previ-ous reverse designs leads to moreshear forces on glenoid fixation com-pared with the less constrained, medial-ized center of rotation of the Grammontdesign (B).

Hatzidakis et al


fulcrum for elevation (dysfunctional cuff or postopera-tive subscapularis rupture with resulting instability)and progressive erosion into the glenoid and/or acro-mion. This leads to increased pain and decreased func-tion. Common causes for failure in fracture cases aremalposition of the component and tuberosity malunionor nonunion. In CTA and post-fracture cases, revisionhemiarthroplasty or conversion to total shoulder arthro-plasty with soft tissue and/or bony procedures is eitherunpredictable or contraindicated. At times, even with ap-parent cuff restoration, the scarring and multiple opera-tions result in the cuff not functioning effectively. Motionand function are lost despite heroic efforts with surgeryand rehabilitation. In our experience, RP placement inthese cases is less beneficial than in primary CTApatients, but the results seem to be more satisfactory thanwith other procedures that have been described. With therevision prosthesis cases, particularly those with multipleprevious operations, low grade infections are more com-mon. All precautions are observed to anticipate, diag-nose, and treat any potential sepsis.

Relative IndicationsAcute fractures and old nonunions in the very elderlywhere rotator cuff rehabilitation is suboptimal is an exam-ple where early results are reported to be good, but lon-ger follow-up is lacking. Early European (D. Mole) andU.S. experience with the RP have been encouraging forthe very elderly patient with acute proximal humeral4-part fractures. Using the prosthesis in these patientscan be beneficial, as the need for tuberosity union isbypassed and post operative rehabilitation is expedited.

ContraindicationsThis prosthesis should not be used in primary osteoarthri-tis or osteonecrosis of the shoulder when the articularsurface–tuberosity relationships are normal and the rota-tor cuff is intact and functional. These patients are bestserved with a third generation anatomic unconstrainedtotal shoulder arthroplasty. The RP also should not beused if the patient exhibits marked deltoid deficiency,as the shoulder will not function well and will be proneto dislocate. A history of previous infection is a relativecontraindication, as recurrence in these patients is high.Finally, the RP should be used only sparingly in patientsless than 65 years old, as long-term survivorship andcomplication rates are unknown.

n COMPONENTS ANDMANUFACTURERS

Currently the RP design of Grammont is offered by twoprosthetic manufacturers. The Delta III� prosthesis isavailable through DePuy (Warsaw IN), and the Aequalis

Reversed Shoulder Prosthesis� is available throughTornier SA, (Montbonnot, FR). For practical purposesof this technique description, the Aequalis ReversedShoulder Prosthesis will be referenced, although thesteps are somewhat similar for the Delta III. The glenoidcomponent consists of a circular baseplate (metaglene),which is fixed to the native glenoid with 2 compressionscrews and 2 locked screws (Fig. 6). In the United States,the surface of the plate that contacts the glenoid bone isgrit blasted and roughened to aid with ongrowth. The su-perior and inferior holes in the baseplate are threadedand angled 20� superiorly and inferiorly, respectively.The anterior and posterior screw holes are not threaded,leaving flexibility to place the anterior and posteriorcompression screws into the best bone that is available.The articulating glenoid component, or glenosphere, isa smooth cobalt-chrome hemisphere that is fixed to thebaseplate via a Morse taper and countersunk set screw(Fig. 7), placing the center of rotation at the level ofthe glenoid/baseplate interface.

The humeral component is designed for cementeduse only, with a cup shaped metaphyseal component thatis fixed to stems of variable lengths via a screw-on mech-anism. The metaphyseal cut–shaft angle is 155�. Thishorizontal inclination leads to stability of the metaphy-seal polyethylene cup when articulating with the gleno-sphere. In the Aequalis Reversed Shoulder� system,polyethylene inserts of increasing diameter (6 mm,9 mm, and 12 mm) (Fig. 8) are available, in additionto a metal metaphyseal offset extension (9 mm). Withthe metaphyseal extension, 15 mm, 18 mm, and 21 mmoffsets become available. Availability of these differentpolyethylene thicknesses and offset lengths are particu-larly useful in the revision situation, when bone lossmakes prosthetic tension more difficult to set preciselyvia cementation of the humeral component. At the timeof this writing, the Delta III prosthesis (DePuy, WarsawIN) has a 6 mm and 9 mm insert, the 9 mm metal exten-sion, and a deeper ‘‘retentive’’ cup option (6 mm) witha deeper thinner polyethylene insert.

FIGURE 6. The baseplate is attached to the native glenoidby a press-fit central peg, 2 compression screws (anteriorand posterior), and 2 locking screws (superior and inferior).



n PREOPERATIVE PLANNING

Preoperative radiographs include a true AP of the shoul-der with the arm in neutral rotation (Grashey view), anaxillary lateral view, and a scapular lateral view. We typ-ically obtain a long AP of the humeral shaft with the APshoulder view to ensure that the humeral canal is normaland a cemented humeral component can be inserted with-out complication. In addition, a Computerized Tomogra-phy scan is obtained of the shoulder, including the entirescapula with 2 mm cuts, so that fatty atrophy of the cuff25

12 can be examined and accurate 2-dimensional coronalreconstructions can be obtained. The axial cuts give thesurgeon an estimate of glenoid size and the feasibility ofglenoid implantation, as the glenoid can occasionally beasymmetrically worn. The axial cuts also are good indi-cators of the degree of fatty infiltration that has occurredin the posterior cuff, an important prognostic indicatorfor regaining some functional external rotation16. Thecoronal reconstructions show the amount of superior

wear of the glenoid, giving the surgeon an idea of howmuch reaming is required inferiorly to create ideal gleno-sphere inferior tilt (0–15�). In cases of very severe supe-rior glenoid erosion, an iliac crest bone graft can beplanned to fill the bony deficiency. In special circum-stances, the humeral head may also be harvested afterremoving any remaining articular cartilage and then usedto bone graft the glenoid defect with or without supple-mental iliac crest graft (Fig. 9).

n TECHNIQUE

For the purposes of this paper, our technique for implan-tation will be described for the ‘‘virgin’’ CTA case.Although Grammont initially described a trans-acromial approach for placement of this prosthesis, mostsurgeons now use a deltopectoral or anterosuperior ap-proach. We prefer the deltopectoral approach, as it sparesthe deltoid insertion and axillary nerve and is highly use-ful in prosthetic revision surgery.

Exposure and Humeral PreparationThe deltopectoral approach is standard, with an incisionstarted just superior and medial to the coracoid processand extended obliquely distally to the deltoid insertionat the mid-arm. The subcutaneous tissues are dissectedwith knife and electrocautery. The cephalic vein is re-tracted laterally with the deltoid and the deltopectoralinterval fully dissected from the clavicle to the insertionof the pectoralis major on the proximal humerus. Theproximal humerus has typically migrated superiorly.The coracoacromial ligament is visualized, followedby removal of the subscapularis bursa and identificationof the conjoint tendon. The deltoid is retracted laterallyusing a broad Richardson retractor, and the conjoint ten-don gently retracted medially using a narrow bluntRichardson retractor. The remaining subscapularis istagged with medially placed #5 braided suture, and thesubscapularis is tenotomized 1 cm from its insertion. Re-lease of the subscapularis continues inferiorly, releasingthe inferior muscular fibers and inferior capsule from thehumerus while an assistant provides gentle progressiveexternal rotation to the arm. The superior 1 cm of thepectoralis insertion is also occasionally released to assistin ease of exposure.

Superiorly, the cuff is usually absent. In cases of pre-vious failed cuff repair, suture and scar are removed, andthe interval between the cuff and acromion dissectedsharply or with electrocautery. The posterior cuff and tu-berosity is freed from the overlying deltoid sharply orwith electrocautery, taking care not to injure the axillarynerve inferiorly.

At this point the humerus is readily dislocated ante-riorly and superiorly for humeral preparation. The

FIGURE 7. The glenosphere is fixed securely to the base-plate via a Morse taper and countersunk set screw.

FIGURE 8. The Reversed Shoulder Prosthesis� (TornierInc., Tornier SA, Montbonnot, FR) has 3 humeral polyeth-ylene inserts of different thicknesses, in addition to a 9 mmmetal extension (not pictured), to optimize deltoid tensionand prosthetic stability.

Hatzidakis et al


highest, most lateral point on the humeral head is iden-tified as a reference point. A hole is created with an awl;the head cutting guide is inserted into the humeral shaftvia this hole, and retroversion is determined using a pinset at the desired degree. Grammont advocated a versionof 0�. We typically set the cut between 0 and 20� of ret-roversion, as this version seems to allow ‘‘filling’’ of themetaphysis without anterior cortical protrusion, and mayallow for slightly improved functional external rotation.Nonetheless, the optimal retroversion for this componenthas yet to be defined. A minimal head cut is made par-

allel to the undersurface of the neck cutting guide, takingcare not to hit the shaft of the cutting guide with the saw(Fig. 10). The guide is removed and the cut completedwith either the saw or an osteotome. Cancellous boneis harvested from the remaining humeral head and savedto graft the glenoid central hole prior to inserting thebaseplate. Then the metaphyseal ‘‘cheese grater’’ reameris used to remove remaining cancellous bone from theproximal humerus to allow metaphyseal componentplacement (Fig. 11). There are 2 sizes of reamers, 36and 42 mm. The size that fits best is usually 36 mm,

FIGURE 9. One potential solution for glenoid bone loss (TRN): Rheumatoid arthritis and Arthrokatadysis of the upper gle-noid. The defect in the glenoid was bone grafted with the osteotomized humeral head (HH) as it fit the erosion perfectly (A).The HH had to be osteotomized more vertically than usual for the standard approach for the Reverse, then temporarily fixedwith Steinmann pins and then fixed with the baseplate (B). Ideally the post of the baseplate reaches the native scapula.

FIGURE 10. The humeral cutting guideand version rod (A). The version rod isaligned with the forearm to determinethe version of the humeral head cut (B).



but for larger individuals or those with previous prostheticinstability, preparation for a 42 mm component may bepreferred. The decision on reamer size is based on pre-operative planning and intraoperative estimation of hu-meral and glenoid size.

Distal metaphyseal reaming is then performed witha long conical reamer, inserting the reamer so that theheight landmark is level with the highest part of the hu-meral cut. Diaphyseal reaming is then performed sequen-tially with reamers of 6.5, 9, 12, and 15 mm until thereamer comes into contact with diaphyseal cortical bone.If the reamer of a specific size reaches the appropriatedepth as noted by the height landmark with chatter,reaming is stopped. Preparing the canal more aggressivelyrisks fracturing the humerus as well as not having an ad-equate cement mantle.

The humeral trial is then assembled and retroversionrod set to 0 to 20�, according to the preference of thesurgeon. The trial is then inserted with manual pressure,followed by light impaction with a mallet. A plasticcut protector is then placed into the metaphysis of thecomponent to protect the trial component during glenoidpreparation. At times the medial head still protrudesand needs trimming to facilitate retraction and glenoidexposure.

Glenoid PreparationHoman and Kolbel retractors are used to retract the softtissues and proximal humerus while a circumferentialcapsulotomy and labral excision are performed. The

origin of the long head of the triceps is released fromthe inferior glenoid, and the lateral pillar of the scapulais palpated. A 2-prong Tiemannn Capsular retractor isplaced against the lateral pillar/scapular neck. This servesas a reference guide later for the superior-inferior orien-tation of the baseplate, so that the inferior screw beplaced down the center of the axial border of the scapula.Glenoid osteophytes are removed to reveal the true gle-noid anatomic shape and more correctly identify the baseof glenoid bone that is most solid for baseplate (meta-glene) placement. The central hole guide is assembledand placed with the handle inferior for the deltopectoralapproach. The inferior most edge of the guide is placedagainst the inferior most edge of the glenoid. This en-sures that the metaglene is placed against the glenoidas inferior as possible. Alternatively, one may prefer toinvert the baseplate and use it as a guide to ensure moreinferior placement. The goal is to have the glenosphereinferior enough and tilted inferiorly a few degrees to de-crease any contact of the polyethylene insert with thescapula. This nuance is the recent attempt to preventscapular notching. Based on preoperative CT scanningthe guide is angled inferiorly to create an inferior tiltto the reamed surface of 0–15�. The central hole is thendrilled with a 6 mm bit, which has a self-stop (Fig. 12).The guide is then removed, followed by placing the flatglenoid reamer (Fig. 12). The reamer is started awayfrom the bone, then pushed slowly into the glenoid togently flatten the glenoid face, conservatively removingbone to preserve as much of the subchondral plate as pos-sible. Once a uniformly flat surface has been createdand a 1–2 mm groove is created circumferentially, thereaming is complete. The groove must be created toallow placement of the baseplate flush against the bone.The reamer is removed and the glenoid central hole drillguide placed, followed by over-drilling of the central holewith a 7.5 mm drill with a self-stop (Fig. 13). The can-cellous bone graft is impacted in the post hole. The base-plate inserter is assembled and the 8.0 mm baseplatecentral post is impacted into the 7.5 mm drilled hole fora press-fit into the scapula.

There is an ‘‘up’’ and ‘‘down’’ marker on the inserterwith a vertical line that allows orientation of the base-plate so that the inferior drill hole is aligned with the in-ferior pillar of the scapular neck (Fig. 14). This ensuresplacement of an inferior screw that is as long as possible.The baseplate is impacted against the glenoid bone withseveral sharp blows with a mallet. The inserter is gentlyremoved, leaving the baseplate in place. The positioningof the superior and inferior locking screws is most crit-ical. We prefer to place the inferior screw first to set therotation of the baseplate. The threaded drill guide isscrewed into the metaglene screw hole. A long 3 mm drillbit is inserted into the guide, drilling through the face of

FIGURE 11. A cup is made in the native proximal humerususing the power ‘‘cheese grater’’ reamer to accept the hu-meral metaphyseal component.

Hatzidakis et al


the glenoid and along the inferior pillar of the scapula(Fig. 15). After the far cortex is breached, the length ismeasured and the inferior screw placed. It is fully ad-vanced temporarily, ‘‘locked’’ into the baseplate, keepingit from rotating during subsequent screw placement. Thesuperior screw is then drilled, measured, and inserted.The anterior and posterior screws are then drilled andproper lengths measured (Fig. 16). These 2 screws arethen inserted but not tightened until the superior andinferior screws are temporarily loosened, unlocking themfrom the baseplate. The anterior and posterior screws arethen tightened sequentially to achieve compression of thebaseplate against the prepared glenoid surface. Once fullcompression is obtained with the anterior and posteriorscrews, the superior and inferior screws are locked toachieve final baseplate fixation. The baseplate should

be very well fixed to the bone. This requires at least3 of the 4 screws having good purchase.

The glenosphere of choice (36 or 42 mm) is thenguided to the metaglene using a screwdriver (Fig. 17).The glenosphere is gently pressed onto the Morse taperof the metaglene and the screwdriver is removed. Theglenosphere is firmly seated onto the Morse taper of themetaglene using a polyethylene impactor and severalsharp blows of the mallet. Final glenosphere fixation issecured by tightening the center set screw, which threadsinto a mating thread on the inside of the center peg ofthe metaglene (Fig. 17). Care is taken throughout thesesteps to ensure that soft tissue is completely clearedfrom the periphery of the glenoid so that incompleteglenosphere-metaglene fixation and postoperative lucentlines do not occur.

FIGURE 12. A center hole is drilled forreaming using the appropriate guide(A), followed by reaming the glenoid flat(B), preferably with a 10� downward tilt.

FIGURE 13. The center hole is widenedslightly by the 7.5 mm drill bit and guideso it will accept the 8 mm diametercentral peg of the baseplate with a goodpress-fit.



Glenohumeral Trialing, Humeral ComponentImplantation, and Final ReductionAfter the glenosphere is firmly implanted, the humeruscan be redislocated out of the incision. Care is takennot to scratch the glenosphere during the dislocationmaneuver. A 6 mm trial humeral insert is then placedand impacted into the humeral trial metaphyseal cup(Fig. 18). A trial reduction is performed. The humeralcomponent should rotate nicely without signs of instabil-ity. Gentle traction is applied to the arm to perform a‘‘shuck’’ maneuver. There should be less than 1 mm ofdiastasis between the humeral cup and glenosphere dur-ing this maneuver to ensure proper stability. If the shoul-

der cannot be reduced with a 6 mm insert, then the trial

is removed, adhesions lysed, and additional proximal

humeral bone resected if necessary.Once it is determined that the shoulder will reduce,

the proximal humeral component is cemented. Acement restrictor is inserted and the humeral shaftmeticulously dried. Cementing technique is very impor-tant for this prosthesis, as there may be more problemswith humeral loosening with the Reverse Prosthesis thanwith unconstrained total shoulder replacement. Press fitplacement of the humeral component is not recom-mended at this time following some earlier experiencewith subsidence.

FIGURE 14. The glenoid baseplate (inthis case coated with hydroxyapatite)is impacted onto the glenoid, with thelower mark aligned with the lateral pillarof the scapula.

FIGURE 15. Correct position of thebaseplate is confirmed by first drillingthe inferior locking screwholewith a longguide that threads into the plate.

Hatzidakis et al


After the humeral component is well fixed with ce-ment, the 6 mm trial insert is placed into the metaphysisand the shoulder reduced. If the tension is adequate, theshoulder is redislocated and the final polyethylene linerimpacted into the metaphyseal cup (Fig. 19). If there isinsufficient tension, gradually increasing sizes of linertrials are used to obtain the optimum fit. In the primaryCTA case it is rare that a liner over 9 mm is required, butthe larger 12 mm liner and 9 mm metaphyseal extensionare often essential for the revision situation.

After the shoulder is reduced with final implants inposition, the subscapularis is repaired via 3 to 4 tendonto tendon sutures, occasionally augmenting the repairwith transosseous sutures if the remaining tendon tag

on the humerus is insufficient. The deltopectoral intervalis closed over a drain, with a running absorbable suture.The skin is closed in routine fashion. Postoperative radio-graphs are obtained in the recovery room (Fig. 20).

n POSTOPERATIVE MANAGEMENT

The shoulder is immobilized in slight abduction and innear-neutral rotation. Gentle passive motion in all planesis started immediately, taking care not to push the limit ofintraoperative motion measurements. Active assistedmotion is started at 3 to 4 weeks postoperatively, fol-lowed by active motion at approximately 6 weekspostoperative. The immobilizer is discontinued at this

FIGURE 16. With the superior and infe-rior locking screws partially advanced(not locked into the plate), the anteriorand posterior screws are drilled andfilled with 4.5 mm screws, compressingthe baseplate to the glenoid bone. Thesuperior and inferior screws are then ad-vanced, ‘‘locking’’ the fixation construct.

FIGURE 17. The glenosphere is lightlyadvanced onto the reverse Morse taperof the baseplate using the center screw-driver. The screwdriver is then removedand a polyethylene impactor applied tothe glenosphere, followed by severalsharp blows with the mallet to seat theprosthesis. The center screwdriver isthen used again, this time to advancethe recessed central screw firmly intothe peg of the baseplate.



time. Activities as tolerated are allowed at 3 to 4 monthspostoperatively.

n RESULTS

We recently performed a retrospective review of 45 re-verse shoulder arthroplasty patients, with an average fol-low up of 42 months.16 The indication for the procedurewas end-stage glenohumeral arthritis with massive irrep-

arable rotator cuff tear in 21 patients, revision of failedhemiarthroplasty in 19 patients, and fracture sequelaewith severe malunion or nonunion of the tuberosities in5 patients. All 3 groups showed a significant increasein active elevation (from 55� pre-operatively to 121� post-operatively) and Constant Score24 (from 17 to 59), butno significant change in active external rotation (from7� to 11�) or internal rotation (S1 pre and postoperatively).Seventy-eight percent of patients were satisfied or verysatisfied with the result, and 67% of patients had no orslight pain. However, the postoperative Constant Score,adjusted Constant Score, and ASES Shoulder Score wereall significantly higher in the CTA group as comparedwith the revision group (P = 0.01, 0.004, and 0.002,respectively). Severe fatty infiltration of the teres minor25

was associated with lower external rotation (0� vs. 15�,P = 0.02) and lower functional results (Constant score:46 vs. 66, P = 0.007).

ComplicationsIn our series of 45 patients, 14 complications occurredin 11 patients (24%). The most common complicationswere dislocation (3 cases), deep infection (3 cases), peri-prosthetic humeral fracture (2 cases), and late acromialfracture (2 cases). One case of postoperative hematoma,axillary nerve palsy, and symptomatic humeral looseningwere also observed. Complication and revision rateswere highest in the revision group (47% and 26%,respectively).

Ten patients (22%) needed further surgery for treat-ment of a complication. All 3 dislocation patients (1 inthe CTA group, 2 in the revision group) were broughtback to the operating room for liner exchange or additionof the 9 mm metal spacer with retentive 6 mm liner. Two

FIGURE 18. The trial insert is impacted into the trial hu-meral prosthesis, followed by a reduction to assess tensionand stability.

FIGURE 19. The final polyethyleneinsert is impacted into the humeralprosthesis.

Hatzidakis et al


patients in the revision group required prosthetic removalfor infection. One infection in the revision group wastreated successfully with 1-stage irrigation, debridement,and polyethylene liner exchange. One patient in the re-vision group developed aseptic humeral loosening. Thehumeral component was revised 1 year postoperatively.Another patient required revision for treatment of a peri-prosthetic humeral fracture sustained in a motor vehicleaccident. Although no cases of glenoid loosening wereobserved, 1 patient required removal of the glenosphereon postoperative day number 1. The glenoid was frac-tured during reaming, but the metaglene and glenosphereappeared to be stable after implantation. The glenospherewas noted to lose fixation on postoperative radiographs,and the patient was brought back to the operating roomfor conversion to a humeral head replacement.

Scapular notching, or loss of inferior glenoid bonesecondary to impingement of the prosthesis against theinferior glenoid neck (Fig. 21), was observed on finalradiographs in 24 cases (74%). Twenty-eight percentof patients exhibited scapular notching that extendedbeyond the inferior screw. Despite the high percentageof cases that exhibited scapular notching, none weresymptomatic and no cases of glenoid loosening were noted.Heterotopic ossification was noted on postoperativeradiographs in 17 cases (45%), always occurring at thelower margin of the glenoid. Grade I ossification 18was found in 11 cases, grade 2 in 2 cases, and gradeIII in 4 cases26. Grade III ossification cases tended tohave greater restriction in range of motion (passive andactive) than patients with less severe heterotopic ossifica-tion. A bony spur at the inferior margin of the scapularneck was noted in 24 cases (63%) and was always as-sociated with a scapular notch (Fig. 21).

n FUTURE OF THE TECHNIQUE

Our most pressing concern regarding this prosthesis isthat it is used for the correct indications, as outlined above.The Tornier Aequalis Reversed Shoulder Prosthesis�was designed based upon the principles of Paul Grammont,including a horizontal humeral neck cut, hemispher-ical glenoid component with secure locked screw fixa-tion, and a medialized and lowered center of rotationthat increases the effectiveness of the deltoid while min-imizing glenoid loosening. The prosthesis is designedfor individuals with end-stage glenohumeral arthritiswith an extensive irreparable cuff deficiency. Althoughshort and intermediate term data regarding the longevityof this prosthesis are quite favorable, long term data are

FIGURE 20. Final radiographs. Notethe slight inferior tilt of the glenosphere.

FIGURE 21. ‘‘Notching’’ of the inferior glenoid by themedial humeral prosthetic neck. Note the hypertrophicosteophyte just medial to the notch.



not available. Therefore, at this time we favor restrictingthe use of this prosthesis to older individuals. The majorbenefit of this prosthesis is that shoulder level functionand above can reliably be achieved in patients with un-complicated rotator cuff deficient arthritis (Fig. 22). Thisconsistency of results has not been achieved in our orothers’ hands with the use of conventional hemiarthro-plasty. Although the prosthesis offers some hope of painrelief and restoration of function in patients who requirerevision of failed shoulder prosthesis, complication ratesare high (approximately 50%). Dislocation rates may bedecreased by the availability of multiple length optionsfor the humeral polyethylene insert and with use of the

larger component. We anticipate that this prosthetic de-sign will continue to be a valuable tool in treating pa-tients for whom satisfactory options have often beenlacking in the past.

n REFERENCES

1. Cofield RH. Total shoulder arthroplasty with the Neerprosthesis. J Bone Joint Surg Am. 1984;66-A:899–906.

2. Field LD, Dines DM, Zabinski SJ, et al. Hemiarthroplastyof the shoulder for rotator cuff arthropathy. J ShoulderElbow Surg. 1997;6:18–23.

FIGURE 22. A, Anteroposterior radiograph of glenohumeral arthritis with extensive cuff deficiency in a patient with a ‘‘pseu-doparalytic’’ painful shoulder. Axial computerized tomography slice reveals rupture of the subscapularis and significant at-rophy of the external rotators. B, Anteroposterior radiograph of the shoulder at maximum elevation. Postoperative radiographof the patient’s reverse prosthesis, and her forward flexion 9 months postoperatively.

Hatzidakis et al


3. Neer CS, Watson KC, Stanton FJ. Recent experience in

total shoulder replacement. J Bone Joint Surg Am. 1982;

64-A:319–337.

4. Neer CS, Craig EV, Fukuda H. Cuff-tear arthropathy.

J Bone Joint Surg Am. 1983;65:1232–1244.

5. Pollock RG, Deliz ED, McIlveen SJ, et al. Prosthetic

replacement in rotator cuff deficient shoulders. J Shoulder

Elbow Surg. 1992;1:173–186.

6. Sanchez-Sotelo J, Cofield RH, Rowland CM. Shoulder

hemiarthroplasty for glenohumeral arthritis associated with

severe rotator cuff deficiency. J Bone Joint Surg Am. 2001;

83-A:1814–1822.

7. Williams GR Jr, Rockwood CA Jr. Hemiarthroplasty in

rotator cuff-deficient shoulders. J Shoulder Elbow Surg.

1996;5:362–367.

8. Zuckerman JD, Scott AJ, Gallagher MA. Hemiarthroplasty

for cuff tear arthropathy. J Shoulder Elbow Surg. 2000;9:

169–172.

9. Franklin JL, Barrett WP, Jackins SE, et al. Glenoid loosen-

ing in total shoulder arthroplasty. Association with rotator

cuff deficiency. J Arthroplasty. 1988;3:39–46.

10. Brostrom LA, Wallenstein R, Olsson E, et al. The Kessel

prosthesis in total shoulder arthroplasty. A five-year expe-

rience. Clin Orthop. 1992;277:155–160.

11. Gross RM.History of total shoulder arthroplasty. In: Crosby

L., ed. Total Shoulder Arthroplasty. Rosemont Il: American

Academy of Orthopaedic Surgeons; 1999.

12. Grammont P, Trouilloud P, Laffay JP, et al. Etude et real-

isation d’une nouvelle prothese d’epaule. Rhumatologie.

1987;39:17–22.

13. Grammont PM, Baulot E. Delta shoulder prosthesis for

rotator cuff rupture. Orthopedics. 1993;16:65–68.

14. Grammont PM, Baulot E, Chabernaud D. R esultats des 16

premiers cas d’arthroplastie totale d’epaule inversee sans

ciment pour des omarthroses avec grande rupture de coiffe.

Rev Chir Orthop Reparatrice Appar Mot. 1996;82(suppl I):

169.

15. Baulot E, Chabernaud D, Grammont P. R esultats de la

prothese inversee de Grammont pour les omarthroses

associees a de grandes destructions de la coiffe. A propos

de 16 cas. Acta Orthop Belg. 1995;61(Suppl I):112–

119.

16. Boileau P,Watkinson D, Hatzidakis A, et al. The GrammontReverse Prosthesis: Design, rationale and biomechanics.J Shoulder Elbow Surg. 2005;14(Suppl S):147S–161S.

17. Boulahia A, Edwards TB, Walch G, et al. Early results ofa reverse design prosthesis in the treatment of arthritis ofthe shoulder in elderly patients with a large rotator cuff tear.Orthopedics. 2002;25:129–133.

18. DeWilde L, Mombert M, Van Petegem P, et al. Revision ofshoulder replacement with a reversed shoulder prosthesis(Delta III). Report of five cases. Acta Orthop Belg.2001;67:348–353.

19. Favard L, Lautmann S, Sirveaux F, et al. Hemi arthroplastyversus reverse arthroplasty in the treatment of osteoarthritiswith massive rotator cuff tear. In: Walch G, Boileau P,Mole D, eds. 2000 Shoulder Prosthesis. Two to Ten Year

Follow-Up. Montpellier France: Sauramps Medical; 2001:261–268.

20. Jacobs R, DeBeer P, De Smet L. Treatment of rotator cuffarthropathy with a reversed delta shoulder prosthesis. Acta

Orthop Belg. 2001;67:344–347.

21. Sirveaux F, Favard L, Oudet D, et al. Grammont invertedtotal shoulder arthroplasty in the treatment of glenohumeral

osteoarthritis with massive and non repairable cuff rup-ture. In: Walch G, Boileau P, Mole D, eds. 2000 ShoulderProsthesis . Two to Ten Year Follow-Up. MontpellierFrance: Sauramps Medical; 2001:247–252.

22. Valenti Ph, Boutens D, Nerot C. Delta 3 reversed prosthesisfor arthritis with massive rotator cuff tear: long term results(.5 years). In: Walch G, Boileau P, Mole D, eds. 2000Shoulder Prosthesis . Two to Ten Year Follow-Up. Mont-pellier France: Sauramps Medical; 2001:253–259.

23. Boileau P, Trojani C, Walch G, et al. Shoulder arthroplastyfor the treatment of the sequelae of fractures of the proximalhumerus. J Shoulder Elbow Surg. 2001;10:299–308.

24. Constant CR, Murley AHG. A. Clinical method of func-tional assessment of the shoulder. Clin Orthop. 1987;214:160–164.

25. Goutallier D, Postel JM, Bernageau J, et al. Fatty muscledegeneration in cuff ruptures. Pre- and postoperative eval-uation by CT scan. Clin Orthop. 1994;304:78–83.

26. Kjaersgaard-Andersen P, Frich LH, Sojbjerg JO, et al.Heterotopic bone formation following total shoulderarthroplasty. J Arthroplasty. 1989;4:99–104.



reverse shoulder arthroplasty indications, technique, and results

Documents