A FINITE ELEMENT MODELING APPROACH TO UNDERSTANDING CRITICAL MECHANICAL TRADE-OFFS IN REVERSE SHOULDER ARTHROPLASTY

Vijay Permeswaran, Jessica E. Goetz, Carolyn M. Hettrich, and Donald D. Anderson

The University of Iowa, Iowa City, IA, USA

email: vijay-permeswaran@uiowa.edu – web: http://poppy.obrl.uiowa.edu/

INTRODUCTION

Reverse shoulder arthroplasty (RSA) holds in-creasing attraction as a means to reliably restore pain-free function to patients with glenohumeral arthritis who are rotator cuff-deficient [1]. RSA uti-lizes a reverse ball-in-socket design, with the hu-meral component as the socket and the glenoid component the ball (glenosphere), to medialize and distalize the humeral center of rotation (Fig. 1).

In the wake of favorable post-op results, RSA has exhibited high early to mid-term complication rates. Scapular notching has been reported to occur after RSA in 31-97% of patients. Notching involves lateral bone loss in the scapula, attributed to contact between the humeral component and the scapula in terminal adduction. This recognition has led to modifications in implant design and surgical tech-nique to reduce impingement risk. A common change has been to lateralize the center of rotation, either by placing bone graft behind the glenoid baseplate or by extending the glenosphere. Both of these modifications decrease the deltoid moment arm, and extending the glenosphere increases the risk of glenoid baseplate loosening or failure.

There is clear need for a better understanding of the mechanical trade-offs involved in RSA implant designs and implantation. A systematic finite ele-ment modeling approach to assess mechanical stresses in the implant and the neighboring bony tissues is here introduced.

METHODS A finite element model of an Aequalis RSA de-

sign (Tornier, Montbonnot, France) was created. Two common implantation techniques were mod-eled: a 10° inferior tilt reamed into the glenoid sur-face and a Bony Increased Offset (BIO) achieved using 10 mm of bone harvested from the humeral head [2]. The scapular geometry was segmented from CT scans of the Visible Female. The implant geometry was obtained from 3D laser scans of the actual implants, and surfaces were fit to the implant geometries using Geomagic Studio software (Ge-omagic USA, Morrisville, NC).

A finite element mesh was created (Fig. 2) from the generated surfaces using TrueGrid (v 2.3 XYZ Scientific Applications, Inc., Livermore, CA). The

scapula, BIO bone wafer, humeral polyethylene cup, and cup liner were modeled as hexahedral three-dimensional elements while the glenosphere and humeral stem were modeled as rigid surfaces.

The humeral component was initially placed in 30° of abduction. A force/moment combination rep-resenting the weight of the arm was applied to the end of the humeral stem. The humeral component was then allowed to freely adduct until contact with the scapula occurred. The adduction angle was then fixed for both models at the angle for which one of

SIMULATED JOINT AND MUSCLE FORCES IN REVERSED AND ANATOMIC SHOULDER PROSTHESES 755

VOL. 90-B, No. 6, JUNE 2008

reversed prosthesis with a rotator cuff, or an anatomicalprosthesis without a rotator cuff. When the arm is horizon-tal, the arm weight moment of force is maximal, and is two-thirds balanced by the deltoid and one-third balanced bythe rotator cuff muscles, according to the muscle forceratios and moment arms (Table I and Fig. 2). A completeremoval of the rotator cuff muscles would increase the del-toid force by 50%. On the other hand, the rotation centremedialisation of the reversed prosthesis very approximatelydoubles the deltoid moment arms, and would thus reduceby half the force required to balance the weight of the arm.

For the anatomical prosthesis, the amplitude and direc-tion of the joint force were consistent with most bio-mechanical models of a healthy shoulder. The joint forcewas maximal when the arm was approximately horizontaland this has been reported previously.22,33,34 The maximumjoint force was nearly equal to the body weight and this hasbeen shown in vivo with an instrumented shoulderimplant.35 The displacement of the contact point on the gle-noid surface, related to the well-known rocking-horseeffect, was also consistent with other models.22,33 On thehumeral side, the location of the contact point also con-firmed the results of a cadaver study.36

In spite of their increasing clinical use, quantitative bio-mechanical analyses of reversed shoulder prostheses arestill rare. Another numerical model of the shoulder alsoreported that abduction is possible without rotator cuffmuscles.37 An increase of the maximum moment arm of thedeltoid was also predicted (from 35 mm to 52 mm),inducing a reduction of the total muscle by a factor of 5.The increase of the moment arm is consistent with ourresults, but the comparison of the total muscle force is notpossible as that model also included the scapulothoracicmuscles. In another comparative study, a shoulder modelwas used to calculate the moment arm of the deltoid duringabduction in the scapular plane.38 The moment arms werethen used to estimate the maximum muscle performance.

Although this model is only geometrical and does not solvethe equations associated with equilibrium, it also reportsthat reversed implants increase the moment arm and theperformance of the deltoid.

Although a direct comparison with clinical results is dif-ficult, the model’s predictions were consistent with clinicalexperience. The mechanism of the reversed prosthesis wasclearly reproduced and quantified here, confirming the eff-icancy of this design to balance the missing stabilising andmotor function of the rotator cuff muscles. The model alsoconfirms the crucial role of the deltoid for correct functionof this implant, and even predicts that activity of this mus-cle is reduced by 20%.

The current model also provides several valuable predic-tions. First, the deltoid moment arms indicate that theincrease in deltoid efficiency is effective mainly at the startof abduction, which means that the reversed prosthesis inparticular improves the initiation of the movement. Themoment arm also stresses the importance of the anteriordeltoid, which can be damaged during surgery, particularlyrevision surgery. Concerning the joint force, the predictedreduction associated with more congruent articular sur-faces should also significantly reduce the pressure on, andhence the wear of, the polyethylene component. This poten-tial advantage could, however, be overwhelmed by theimpingement between the humeral polyethylene compo-nent and the glenoid neck, which is still a major source ofwear.6 It has been assumed that the rocking-horse effect,which is neutralised at the glenoid site, might be transferredto the humeral component, weakening its fixation.16 As norocking-horse effect was observed on the humeral compo-nent, its higher rate of failure may be associated with non-biomechanical factors, such as a poor bone support or oste-olysis.

In conclusion, this force analysis elucidates the ability ofthe reversed implant to allow abduction without rotatorcuff muscles. The quantified gain in the moment arms was

150°120°90°60°30°

150°120°90°

60°

30°

30° 60°90°120°

150°

30–150°

30°60°90°

120°150°

30–150°

Fig. 4b

Diagrams showing the direction of the glenohumeral joint force, on the glenoid and humeral side, every 30° of abduction in the scapular plane, for a)the anatomical prosthesis, b) the reversed prosthesis without any rotator cuff muscle, and c) the reversed prosthesis without supraspinatus only.

Fig. 4a Fig. 4c

SIMULATED JOINT AND MUSCLE FORCES IN REVERSED AND ANATOMIC SHOULDER PROSTHESES 755

VOL. 90-B, No. 6, JUNE 2008

reversed prosthesis with a rotator cuff, or an anatomicalprosthesis without a rotator cuff. When the arm is horizon-tal, the arm weight moment of force is maximal, and is two-thirds balanced by the deltoid and one-third balanced bythe rotator cuff muscles, according to the muscle forceratios and moment arms (Table I and Fig. 2). A completeremoval of the rotator cuff muscles would increase the del-toid force by 50%. On the other hand, the rotation centremedialisation of the reversed prosthesis very approximatelydoubles the deltoid moment arms, and would thus reduceby half the force required to balance the weight of the arm.

For the anatomical prosthesis, the amplitude and direc-tion of the joint force were consistent with most bio-mechanical models of a healthy shoulder. The joint forcewas maximal when the arm was approximately horizontaland this has been reported previously.22,33,34 The maximumjoint force was nearly equal to the body weight and this hasbeen shown in vivo with an instrumented shoulderimplant.35 The displacement of the contact point on the gle-noid surface, related to the well-known rocking-horseeffect, was also consistent with other models.22,33 On thehumeral side, the location of the contact point also con-firmed the results of a cadaver study.36

In spite of their increasing clinical use, quantitative bio-mechanical analyses of reversed shoulder prostheses arestill rare. Another numerical model of the shoulder alsoreported that abduction is possible without rotator cuffmuscles.37 An increase of the maximum moment arm of thedeltoid was also predicted (from 35 mm to 52 mm),inducing a reduction of the total muscle by a factor of 5.The increase of the moment arm is consistent with ourresults, but the comparison of the total muscle force is notpossible as that model also included the scapulothoracicmuscles. In another comparative study, a shoulder modelwas used to calculate the moment arm of the deltoid duringabduction in the scapular plane.38 The moment arms werethen used to estimate the maximum muscle performance.

Although this model is only geometrical and does not solvethe equations associated with equilibrium, it also reportsthat reversed implants increase the moment arm and theperformance of the deltoid.

Although a direct comparison with clinical results is dif-ficult, the model’s predictions were consistent with clinicalexperience. The mechanism of the reversed prosthesis wasclearly reproduced and quantified here, confirming the eff-icancy of this design to balance the missing stabilising andmotor function of the rotator cuff muscles. The model alsoconfirms the crucial role of the deltoid for correct functionof this implant, and even predicts that activity of this mus-cle is reduced by 20%.

The current model also provides several valuable predic-tions. First, the deltoid moment arms indicate that theincrease in deltoid efficiency is effective mainly at the startof abduction, which means that the reversed prosthesis inparticular improves the initiation of the movement. Themoment arm also stresses the importance of the anteriordeltoid, which can be damaged during surgery, particularlyrevision surgery. Concerning the joint force, the predictedreduction associated with more congruent articular sur-faces should also significantly reduce the pressure on, andhence the wear of, the polyethylene component. This poten-tial advantage could, however, be overwhelmed by theimpingement between the humeral polyethylene compo-nent and the glenoid neck, which is still a major source ofwear.6 It has been assumed that the rocking-horse effect,which is neutralised at the glenoid site, might be transferredto the humeral component, weakening its fixation.16 As norocking-horse effect was observed on the humeral compo-nent, its higher rate of failure may be associated with non-biomechanical factors, such as a poor bone support or oste-olysis.

In conclusion, this force analysis elucidates the ability ofthe reversed implant to allow abduction without rotatorcuff muscles. The quantified gain in the moment arms was

150°120°90°60°30°

150°120°90°

60°

30°

30° 60°90°120°

150°

30–150°

30°60°90°

120°150°

30–150°

Fig. 4b

Diagrams showing the direction of the glenohumeral joint force, on the glenoid and humeral side, every 30° of abduction in the scapular plane, for a)the anatomical prosthesis, b) the reversed prosthesis without any rotator cuff muscle, and c) the reversed prosthesis without supraspinatus only.

Fig. 4a Fig. 4c

SIMULATED JOINT AND MUSCLE FORCES IN REVERSED AND ANATOMIC SHOULDER PROSTHESES 755

VOL. 90-B, No. 6, JUNE 2008

reversed prosthesis with a rotator cuff, or an anatomicalprosthesis without a rotator cuff. When the arm is horizon-tal, the arm weight moment of force is maximal, and is two-thirds balanced by the deltoid and one-third balanced bythe rotator cuff muscles, according to the muscle forceratios and moment arms (Table I and Fig. 2). A completeremoval of the rotator cuff muscles would increase the del-toid force by 50%. On the other hand, the rotation centremedialisation of the reversed prosthesis very approximatelydoubles the deltoid moment arms, and would thus reduceby half the force required to balance the weight of the arm.

For the anatomical prosthesis, the amplitude and direc-tion of the joint force were consistent with most bio-mechanical models of a healthy shoulder. The joint forcewas maximal when the arm was approximately horizontaland this has been reported previously.22,33,34 The maximumjoint force was nearly equal to the body weight and this hasbeen shown in vivo with an instrumented shoulderimplant.35 The displacement of the contact point on the gle-noid surface, related to the well-known rocking-horseeffect, was also consistent with other models.22,33 On thehumeral side, the location of the contact point also con-firmed the results of a cadaver study.36

In spite of their increasing clinical use, quantitative bio-mechanical analyses of reversed shoulder prostheses arestill rare. Another numerical model of the shoulder alsoreported that abduction is possible without rotator cuffmuscles.37 An increase of the maximum moment arm of thedeltoid was also predicted (from 35 mm to 52 mm),inducing a reduction of the total muscle by a factor of 5.The increase of the moment arm is consistent with ourresults, but the comparison of the total muscle force is notpossible as that model also included the scapulothoracicmuscles. In another comparative study, a shoulder modelwas used to calculate the moment arm of the deltoid duringabduction in the scapular plane.38 The moment arms werethen used to estimate the maximum muscle performance.

Although this model is only geometrical and does not solvethe equations associated with equilibrium, it also reportsthat reversed implants increase the moment arm and theperformance of the deltoid.

Although a direct comparison with clinical results is dif-ficult, the model’s predictions were consistent with clinicalexperience. The mechanism of the reversed prosthesis wasclearly reproduced and quantified here, confirming the eff-icancy of this design to balance the missing stabilising andmotor function of the rotator cuff muscles. The model alsoconfirms the crucial role of the deltoid for correct functionof this implant, and even predicts that activity of this mus-cle is reduced by 20%.

The current model also provides several valuable predic-tions. First, the deltoid moment arms indicate that theincrease in deltoid efficiency is effective mainly at the startof abduction, which means that the reversed prosthesis inparticular improves the initiation of the movement. Themoment arm also stresses the importance of the anteriordeltoid, which can be damaged during surgery, particularlyrevision surgery. Concerning the joint force, the predictedreduction associated with more congruent articular sur-faces should also significantly reduce the pressure on, andhence the wear of, the polyethylene component. This poten-tial advantage could, however, be overwhelmed by theimpingement between the humeral polyethylene compo-nent and the glenoid neck, which is still a major source ofwear.6 It has been assumed that the rocking-horse effect,which is neutralised at the glenoid site, might be transferredto the humeral component, weakening its fixation.16 As norocking-horse effect was observed on the humeral compo-nent, its higher rate of failure may be associated with non-biomechanical factors, such as a poor bone support or oste-olysis.

In conclusion, this force analysis elucidates the ability ofthe reversed implant to allow abduction without rotatorcuff muscles. The quantified gain in the moment arms was

150°120°90°60°30°

150°120°90°

60°

30°

30° 60°90°120°

150°

30–150°

30°60°90°

120°150°

30–150°

Fig. 4b

Diagrams showing the direction of the glenohumeral joint force, on the glenoid and humeral side, every 30° of abduction in the scapular plane, for a)the anatomical prosthesis, b) the reversed prosthesis without any rotator cuff muscle, and c) the reversed prosthesis without supraspinatus only.

Fig. 4a Fig. 4c

SIMULATED JOINT AND MUSCLE FORCES IN REVERSED AND ANATOMIC SHOULDER PROSTHESES

755

VOL. 90-B, No. 6, JUNE 2008

reversed prosthesis with a rotator cuff, or an anatomical

prosthesis without a rotator cuff. When the arm is horizon-

tal, the arm weight moment of force is maximal, and is two-

thirds balanced by the deltoid and one-third balanced by

the rotator cuff muscles, according to the muscle force

ratios and moment arms (Table I and Fig. 2). A complete

removal of the rotator cuff muscles would increase the del-

toid force by 50%. On the other hand, the rotation centre

medialisation of the reversed prosthesis very approximately

doubles the deltoid moment arms, and would thus reduce

by half the force required to balance the weight of the arm.

For the anatomical prosthesis, the amplitude and direc-

tion of the joint force were consistent with most bio-

mechanical models of a healthy shoulder. The joint force

was maximal when the arm was approximately horizontal

and this has been reported previously.22,33,34 The maximum

joint force was nearly equal to the body weight and this has

been shown in vivo with an instrumented shoulder

implant.35 The displacement of the contact point on the gle-

noid surface, related to the well-known rocking-horse

effect, was also consistent with other models.22,33 On the

humeral side, the location of the contact point also con-

firmed the results of a cadaver study.36

In spite of their increasing clinical use, quantitative bio-

mechanical analyses of reversed shoulder prostheses are

still rare. Another numerical model of the shoulder also

reported that abduction is possible without rotator cuff

muscles.37 An increase of the maximum moment arm of the

deltoid was also predicted (from 35 mm to 52 mm),

inducing a reduction of the total muscle by a factor of 5.

The increase of the moment arm is consistent with our

results, but the comparison of the total muscle force is not

possible as that model also included the scapulothoracic

muscles. In another comparative study, a shoulder model

was used to calculate the moment arm of the deltoid during

abduction in the scapular plane.38 The moment arms were

then used to estimate the maximum muscle performance.

Although this model is only geometrical and does not solve

the equations associated with equilibrium, it also reports

that reversed implants increase the moment arm and the

performance of the deltoid.

Although a direct comparison with clinical results is dif-

ficult, the model’s predictions were consistent with clinical

experience. The mechanism of the reversed prosthesis was

clearly reproduced and quantified here, confirming the eff-

icancy of this design to balance the missing stabilising and

motor function of the rotator cuff muscles. The model also

confirms the crucial role of the deltoid for correct function

of this implant, and even predicts that activity of this mus-

cle is reduced by 20%.

The current model also provides several valuable predic-

tions. First, the deltoid moment arms indicate that the

increase in deltoid efficiency is effective mainly at the start

of abduction, which means that the reversed prosthesis in

particular improves the initiation of the movement. The

moment arm also stresses the importance of the anterior

deltoid, which can be damaged during surgery, particularly

revision surgery. Concerning the joint force, the predicted

reduction associated with more congruent articular sur-

faces should also significantly reduce the pressure on, and

hence the wear of, the polyethylene component. This poten-

tial advantage could, however, be overwhelmed by the

impingement between the humeral polyethylene compo-

nent and the glenoid neck, which is still a major source of

wear.6 It has been assumed that the rocking-horse effect,

which is neutralised at the glenoid site, might be transferred

to the humeral component, weakening its fixation.16 As no

rocking-horse effect was observed on the humeral compo-

nent, its higher rate of failure may be associated with non-

biomechanical factors, such as a poor bone support or oste-

olysis.

In conclusion, this force analysis elucidates the ability of

the reversed implant to allow abduction without rotator

cuff muscles. The quantified gain in the moment arms was

150°120°90°60°30°

150°120°90°60°

30°

30° 60°90°

120°150°

30–150°

30°60°90°120°

150°

30–150°

Fig. 4b

Diagrams showing the direction of the glenohumeral joint force, on the glenoid and humeral side, every 30° of abduction in the scapular plane, for a)

the anatomical prosthesis, b) the reversed prosthesis without any rotator cuff muscle, and c) the reversed prosthesis without supraspinatus only.

Fig. 4a

Fig. 4c

SIMULATED JOINT AND MUSCLE FORCES IN REVERSED AND ANATOMIC SHOULDER PROSTHESES

755

VOL. 90-B, No. 6, JUNE 2008

reversed prosthesis with a rotator cuff, or an anatomical

prosthesis without a rotator cuff. When the arm is horizon-

tal, the arm weight moment of force is maximal, and is two-

thirds balanced by the deltoid and one-third balanced by

the rotator cuff muscles, according to the muscle force

ratios and moment arms (Table I and Fig. 2). A complete

removal of the rotator cuff muscles would increase the del-

toid force by 50%. On the other hand, the rotation centre

medialisation of the reversed prosthesis very approximately

doubles the deltoid moment arms, and would thus reduce

by half the force required to balance the weight of the arm.

For the anatomical prosthesis, the amplitude and direc-

tion of the joint force were consistent with most bio-

mechanical models of a healthy shoulder. The joint force

was maximal when the arm was approximately horizontal

and this has been reported previously.22,33,34 The maximum

joint force was nearly equal to the body weight and this has

been shown in vivo with an instrumented shoulder

implant.35 The displacement of the contact point on the gle-

noid surface, related to the well-known rocking-horse

effect, was also consistent with other models.22,33 On the

humeral side, the location of the contact point also con-

firmed the results of a cadaver study.36

In spite of their increasing clinical use, quantitative bio-

mechanical analyses of reversed shoulder prostheses are

still rare. Another numerical model of the shoulder also

reported that abduction is possible without rotator cuff

muscles.37 An increase of the maximum moment arm of the

deltoid was also predicted (from 35 mm to 52 mm),

inducing a reduction of the total muscle by a factor of 5.

The increase of the moment arm is consistent with our

results, but the comparison of the total muscle force is not

possible as that model also included the scapulothoracic

muscles. In another comparative study, a shoulder model

was used to calculate the moment arm of the deltoid during

abduction in the scapular plane.38 The moment arms were

then used to estimate the maximum muscle performance.

Although this model is only geometrical and does not solve

the equations associated with equilibrium, it also reports

that reversed implants increase the moment arm and the

performance of the deltoid.

Although a direct comparison with clinical results is dif-

ficult, the model’s predictions were consistent with clinical

experience. The mechanism of the reversed prosthesis was

clearly reproduced and quantified here, confirming the eff-

icancy of this design to balance the missing stabilising and

motor function of the rotator cuff muscles. The model also

confirms the crucial role of the deltoid for correct function

of this implant, and even predicts that activity of this mus-

cle is reduced by 20%.

The current model also provides several valuable predic-

tions. First, the deltoid moment arms indicate that the

increase in deltoid efficiency is effective mainly at the start

of abduction, which means that the reversed prosthesis in

particular improves the initiation of the movement. The

moment arm also stresses the importance of the anterior

deltoid, which can be damaged during surgery, particularly

revision surgery. Concerning the joint force, the predicted

reduction associated with more congruent articular sur-

faces should also significantly reduce the pressure on, and

hence the wear of, the polyethylene component. This poten-

tial advantage could, however, be overwhelmed by the

impingement between the humeral polyethylene compo-

nent and the glenoid neck, which is still a major source of

wear.6 It has been assumed that the rocking-horse effect,

which is neutralised at the glenoid site, might be transferred

to the humeral component, weakening its fixation.16 As no

rocking-horse effect was observed on the humeral compo-

nent, its higher rate of failure may be associated with non-

biomechanical factors, such as a poor bone support or oste-

olysis.

In conclusion, this force analysis elucidates the ability of

the reversed implant to allow abduction without rotator

cuff muscles. The quantified gain in the moment arms was

150°120°90°60°30°

150°120°90°60°

30°

30° 60°90°

120°150°

30–150°

30°60°90°120°

150°

30–150°

Fig. 4b

Diagrams showing the direction of the glenohumeral joint force, on the glenoid and humeral side, every 30° of abduction in the scapular plane, for a)

the anatomical prosthesis, b) the reversed prosthesis without any rotator cuff muscle, and c) the reversed prosthesis without supraspinatus only.

Fig. 4a

Fig. 4c

distalizeand

medialize

native shoulder

reverse shoulder

arthroplasty

total shoulder

arthroplasty

CORCOR

center of rotation (COR)

scapula

humerus

glenoid

deltoid

Figure 1. Comparison of the centers of rotation (COR) for the humerus (from left to right) in the native shoulder, in total shoulder arthroplasty, and in reverse shoulder arthroplasty.

10° inferior tilt BIO implantation

Figure 2. The finite elment model to the left shows a gleno-sphere implanted in a 10° inferior tilt, and to the right a bony increased offset (BIO) implantation.

the two models first contacted the scapula. An in-ternal-external rotation moment profile was then applied to the humeral implants while they were constrained to rotate about the center of the gleno-sphere, and contact stresses at the scapula-humeral cup interface were recorded. All jobs were run using Abaqus Explicit 6.11.1 (Dassault Systèmes, Vélizy-Villacoublay, France).

RESULTS AND DISCUSSION

Impingement between the humeral cup compo-nent and the inferior lateral scapular border was de-tected in both models (Fig. 3), but the contact stress was elevated and persisted across more angles in the 10° inferiorly tilted glenosphere compared to the BIO implantation. The point of contact stress was located in close proximity with the previously re-ported highest incidence of scapular notching.

These results suggest that a finite element mod-eling approach can provide insights into the nature

of contact between the scapula and the humeral component in RSA. Prior developments with com-putational wear formulations [3] provide a means to study the progression from frank impingement mo-tions to scapular notching. Additional work is need-ed to provide suitable kinematic/force-moment pro-files reflective of activities of daily living in these cuff-deficient patients, specifically in poses where impingement is expected. REFERENCES 1. Frankle M et al. J Bone Joint Surg Am. 87(8):1697-1705, 2005. 2. Boileau P et al. Clin Orthop Relat Res. 469(9):2558-2567, 2011. 3. Maxian TA et al. J Orthop Res. 14(4):668-675, 1996

ACKNOWLEDGEMENTS

This work was funded by faculty startup grant from the Department of Orthopaedics and Rehabilitation at the University of Iowa.

External)rotation

Internal(rotation

External)rotation

Internal(rotation

10° inferior tilt BIO implantation

Figure 3. Contact stress at the in-terface between the glenosphere implanted in a 10° inferior tilt and the humeral cup was more elevated and persisted over a wider range of int/ext rotation compared to that of a bony increased offset (BIO) im-plantation.