IRB: As these co

approval was re

Funding: This w

SUNY Upstate M

*Reprint req

pedic Surgery, S

Syracuse, NY 13

E-mail addre

J Shoulder Elbow Surg (2012) 21, 917-924

1058-2746/$ - s

doi:10.1016/j.jse

www.elsevier.com/locate/ymse

Does glenoid baseplate geometry affect its fixationin reverse shoulder arthroplasty?

Jaison James, MDa, Kayla R. Huffman, BSa, Frederick W. Werner, MMEa,*,Levi G. Sutton, MSa, Vipul N. Nanavati, MDb

aDepartment of Orthopedic Surgery, SUNY Upstate Medical University, Syracuse, NY, USAbEastern Orthopedics and Sports Medicine, S. Windsor, CT, USA

Background: The effect of glenoid baseplate geometry has not been studied as it pertains to reverseshoulder arthroplasty. The purpose of this study was to compare 2 baseplate designs whose major differ-ence is being either a flat backed design or a convex baseplate, with regard to their bone interface area,screw engagement, and bone volume removed using 3-dimensional modeling.Methods: Three-dimensional models of 6 scapulae were used to virtually implant models of a flat backedand a convex backed glenoid baseplate. Additional reaming was performed in 1 mm increments, up to 5mm, and the amount of baseplate screw engagement was calculated at each increment. Statistical differ-ences between flat and convex implants were calculated.Results: Insertion of the convex baseplate required statistically greater removal of bone as compared to theflat baseplate (P ¼ .003). No statistical changes in total area were observed with reaming of the glenoid forthe convex baseplate (P > .095). However, for the flat baseplate, 1 mm of reaming caused a statisticaldecrease in area available for fixation. The amount of total bone area in contact with a convex baseplatewas statistically greater than with a flat baseplate (P ¼ .004). The amount of screw engagement was statis-tically less with the convex baseplate, compared to the flat (P ¼ .026).Discussion: A convex backed glenoid baseplate can improve the contact surface area at the bone implantinterface as compared to a flat backed design. However, better screw engagement and less bone volumeremoved during reaming favors a flat backed design, particularly when adequate bone-implant contactcannot be achieved.Level of evidence: Basic Science Study, Biomechanical Study.� 2012 Journal of Shoulder and Elbow Surgery Board of Trustees.

Keywords: Reverse shoulder arthroplasty; glenoid fixation

mputational models were based on cadaver bones, no IRB

quired.

as funded by the Department of Orthopedic Surgery,

edical University.

uests: Frederick W. Werner, MME, Department of Ortho-

UNY Upstate Medical University, 750 E. Adams Street,

210, USA.

ss: wernerf@upstate.edu (F.W. Werner).

ee front matter � 2012 Journal of Shoulder and Elbow Surgery

.2011.04.017

Reverse shoulder arthroplasty has been increasing inpopularity, despite early reports of failures,6,21 due to newergenerations of designs tending to be more successful.6

Reverse shoulder arthroplasty is claiming its niche asa procedure offered to low-demand elderly patientssuffering from painful glenohumeral arthritis with super-imposed rotator cuff insufficiency.6 Since FDA approvalfor use in the U.S. in 2004, reverse shoulder design

Board of Trustees.

Figure 1 Geometry of 1 of the 6 scapulae before reaming. Allsubsequent figures of a scapula are after surgical modifications tothe 3-D representation of this scapula.

918 J. James et al.

improvements have focused on reducing the classiccomplications associated with initial fixation and implantlongevity. Biomechanical testing has revealed a number offactors that may effect an implant’s clinical success.1-3,20

Glenoid component fixation failure has been heralded asone of the most common complications of reverse shoulderarthroplasty leading to revision surgery.1-3,12 The impor-tance of initial glenoid fixation as a precursor to long-termfixation of reverse shoulder implants via bone ingrowth hasbecome increasingly recognized in the literature.10,11 Intheory, a convex baseplate design would improve glenoidfixation15; however, most of the literature looking atdifferences between flat and convex glenoid baseplates isfocused on standard total shoulder designs.

Many glenoid baseplate (metaglene) design modifica-tions have attempted to reduce the incidence of glenoidcomponent loosening. Previous biomechanical and finiteelement analysis studies have demonstrated the advantages/disadvantages of different keel designs and materials17;cement versus cementless fixation2 and smooth versusrough fixation surface1 in standard glenoid implant designs;and screw orientation3 and glenoid baseplate position interms of tilt9 and position14 in reversed implant designs.Although the effect of glenoid baseplate geometry on initialreversed-anatomy glenoid fixation has been studied usingfinite element analysis,11 additional studies on the amountof bone available for bone fixation are needed to comparethe 2 current reverse shoulder glenoid designs: curved(convex) back and flat back.

Previously, we reported on the optimization of glenoidfixation for a reverse flat backed glenoid component.18 Thegoal of this study was to compare a convex baseplate toa flat backed design using our existing methodology18 todetermine whether this design factor may have an impacton glenoid fixation. However, no current designs differ onlyby this single design factor. They also differ by their sizeand peg or post design. Therefore, this study wasa comparison of 2 reverse shoulder baseplate designs whosemajor difference, in our opinion, is whether the surface isflat or convex. The purpose of this study was to compare 2baseplate designs whose major difference is whether theywere either a flat backed design or a convex baseplate, withregard to their bone interface area, screw engagement, andbone volume removed using 3-dimensional modeling.

Methods and materials

Three-dimensional models of 6 fresh scapulas were created usingcomputer tomography (CT) scans of each scapula.18 Cortical andcancellous bone were differentiated by thresholding the CT scanimages. Using medical imaging software, Slice-O-Matic (Tomo-vision, Quebec, Canada), the outer boundaries of the cortical boneand the cancellous bone were identified. Two models were createdof each scapula, 1 that included both cancellous and corticalbone and the other that included only cancellous bone.

Three-dimensional models were also created of a flat backed2-screw design (Zimmer Trabecular Metal Reverse Shoulder;Zimmer, Warsaw, IN, USA) and of a convex backed 2-screwdesign (Zimmer Anatomical Shoulder Inverse/Reverse; Zimmer)using an optical measuring device and reverse engineering soft-ware (Geomagic Studio; Geomagic, Research Triangle Park, NC,USA). The flat implant consisted of a 28-mm diameter baseplatewith a 16.7-mm long and 8-mm diameter peg. The convex implanthad a height and width of 34 and 23 mm, respectively. Bothimplants use 4.5 mm diameter screws, located superiorly andinferiorly, that can be variably oriented within a 30o arc beforebeing locked in position.

Each baseplate design was virtually implanted into each of the6 scapula models (Fig. 1).18 Based on the published surgicaltechniques,22,23 each glenoid was virtually reamed to the appro-priate depth and orientation (Figs. 2 and 3), and the baseplateinserted into the glenoid. As indicated by the surgical guides, eachglenoid was reamed just enough to create either a uniformly flat orspherical surface beneath the reamer. As described below, thisamount of reaming was defined to be 0 mm of reaming for eachdesign. A 36-mm diameter reamer was used for the flat implant. Aspherical reamer with a radius of curvature of 31 mm was used forthe convex implant, thus achieving a spherical recess. The glenoidbaseplate of each device was then positioned to have the inferioredge of the baseplate coincide with the inferior edge of the glenoidand centered in the anterior-posterior direction. The baseplate wasrotated so that the superior and inferior screws were aligned withthe direction of the longest dimension of the glenoid before it wasreamed. The screws for each design were then repeatedly insertedat 5o increments (within the manufacturer’s guidelines and limits)to determine the orientation at which there was maximal screwengagement (Figs. 4 and 5).

Three geometrical measurements were computed in thecomputer automated design (CAD) assemblies of the bone andimplant (Solid Works, Concord, MA, USA). First, the volume ofbone removed from the glenoid due to surface reaming, insertionof the central peg and the screws was determined. Second, themaximum amount of surface area available for a baseplate after

Figure 2 Surface of the same glenoid shown in Figure 1, afterinitial reaming for the flat baseplate. This corresponds to the 0 mmreaming condition.

Figure 3 Surface of the same glenoid shown in Figure 1, afterinitial reaming for the convex baseplate. This corresponds to the0 mm reaming condition.

Figure 4 View of the flat baseplate and screw insertion at the0 mm reaming condition. Medial half of scapula has beenremoved for visualization.

Reverse shoulder arthroplasty 919

reaming was computed. Third, the actual bone surface area incontact with the undersurface of the baseplate as well as the postor keel was determined (Figs. 6 and 7). These 3 measurementswere calculated first at the minimal (initial) amount of reaming,defined as 0 mm reaming, and then following additional amountsof glenoid reaming. These measurements were computed for boththe entire bone model of each scapula as well as the model of justthe cancellous bone, allowing the computation of these measure-ments for just the cortical and cancellous bone.

The additional reaming was performed in 1 mm increments, upto a total of 5 mm (Figs. 8 and 9), to simulate surgical adjustmentsfor eroded bone or surgical error. The orientation of each screwwas adjusted, within the amount allowed by the implant design,for each amount of reaming to maximize the amount of boneengagement. Screw engagement was defined as the length fromwhere the screw enters the bone until it begins to protrude fromthe bone.

A 2-factor repeated measures analysis of variance (ANOVA)was used to examine if therewere statistical differences between theflat and convex implants and for different amounts of reaming forthe (a) total volume of bone removed, (b) maximum glenoid surfacearea available for a baseplate, and (c) amount of bone in contactwith the baseplate. If differences with reaming were found at P <.05, then a 1-way repeated measures ANOVAwas used to determineat what levels of reaming there were statistical differences. ABonferroni adjustment for multiple comparisons was used. A 3factor repeated measures ANOVA was used to examine screwengagement. The 3 factors were screw location, baseplate design,and the amount of reaming. To differentiate between the 2 baseplatedesigns, additional 2-way repeated measures ANOVAs were per-formed comparing screw location and the amount of reaming.

Results

Insertion of the convex baseplate required a statisticallygreater removal of bone as compared to the flat baseplate(P ¼ .003; Figs. 10 and 11). As would be expected for bothbaseplate designs, additional reaming decreased the bonevolume. In the case of the flat baseplate, just 1 mm ofreaming caused a statistically greater volume of bone to be

Figure 5 View of the convex baseplate and screw insertion atthe 0 mm reaming condition. Medial half of scapula has beenremoved for visualization.

Figure 6 Surface of the glenoid after 0 mm reaming for the flatbaseplate. The light brown region is the portion of the scapula thatis in contact with the flat baseplate.

Figure 7 Surface of the glenoid after 0 mm reaming for theconvex baseplate. The dark brown region is the portion of thescapula that is in contact with the convex baseplate.

Figure 8 Surface of the glenoid after 5 mm reaming for the flatbaseplate.

920 J. James et al.

removed (P ¼ .001). Each additional millimeter of reamingalso caused a statistically significant amount of boneremoval. For the convex baseplate, after 2 mm of reaming,there was a statistical increase in the amount of boneremoved (P ¼ .027) compared to the initial reaming statefor the convex baseplate.

The maximum glenoid surface area available for fixationof a baseplate was statistically different between the convexand flat baseplates (Fig. 12). The available area for theconvex baseplate was greater than the flat for the combinedcortical and cancellous bone surfaces (P ¼ .006), for thecortical (P ¼ .005) and cancellous portions of the area (P ¼.021). No statistical changes in total area (sum of corticaland cancellous) were observed with reaming of the glenoid

for the convex baseplate (P > .095). However, for the flatbaseplate, with just one 1 mm of reaming, there wasa statistical decrease in the area available for fixation(Fig. 12). Additional statistical decreases in total areaavailable occurred with 2 and 3 mm of reaming.

The amount of total bone area (sum of cortical andcancellous bone areas) in contact with a convex baseplatewas statistically greater than with a flat baseplate (P ¼ .004;Fig. 13). The cortical portion of the total area was also

Figure 9 Surface of the glenoid after 5 mm reaming for theconvex baseplate.

Figure 10 AP cross-sectional cut view of 1 scapula, for theconvex baseplate, showing how much bone was initially removed(0 mm of reaming level) in blue and the additional amountremoved with 5 mm of reaming level in red. The sectional viewwas taken at the center of the glenoid.

Figure 11 The volume of bone (sum of the cortical andcancellous) removed as a function of the amount of bone reamingfor both the flat and convex baseplates.

Reverse shoulder arthroplasty 921

greater with the convex baseplate (P ¼ .001), but there wasno difference in the amount of cancellous bone in contactwith the 2 baseplates (P > .713). With incrementalreaming of the glenoid, there was no change in the corticalbone area beneath the baseplates (P > .666). However,there was a decrease in the combined cortical and cancel-lous bone area beneath the convex baseplate by 3 mm of

reaming (P ¼ .033) and beneath the flat baseplate by 5 mmof reaming (P ¼ .038).

The amount of screw engagement was statistically less(Fig. 14) with the convex baseplate, compared to the flat(P ¼ .026). While there was no difference in the amount ofengagement between the superior and inferior screws of theflat design (P ¼ .769), the inferior screw had moreengagement than the superior screw of the convex design(P ¼ .047). Reaming of as little of 1 mm statisticallydecreased the amount of screw engagement (P < .025).Additional statistical decreases in screw engagementoccurred with 2 mm of reaming.

Discussion

The key factors contributing to long-term glenoid fixationin reverse shoulder arthroplasty include bone-implantcontact, screw fixation/engagement, and, ultimately, boneingrowth.11 In standard glenoid implants, glenoid compo-nent fixation is viewed as the weak link in total shoulderarthroplasty, the failure of which commonly leads to revi-sion surgery.16 Glenoid loosening of standard implantsultimately leads to revision surgery, which can beextremely difficult as a result of inadequate glenoid bonestock available for the fixation of a new glenoid baseplate.8

These issues should also be evident in reverse shoulderdesigns. Therefore, it is crucial to reduce glenoid looseningin an effort to promote longevity of the reverse shoulderprosthesis.

In theory, a convex baseplate would serve to improvereverse shoulder glenoid fixation.15 This phenomenon hasbeen studied extensively in other areas of joint arthroplasty.In fact, clinical and experimental studies with regards tototal hip arthroplasty have indicated that a close geometricfit between the supporting bone and the femoral component

Figure 13 The amount of area beneath the convex and flatbaseplates is shown for the sum of the cortical and cancellousbone and for just the cortical bone, as a function of the amount ofbone reaming.

Figure 12 The maximum glenoid surface area (sum of corticaland cancellous) available for baseplate fixation as a function of theamount of bone reaming. The convex baseplate has a greateravailable area than the flat baseplate, as the spherical reamer isprogressing deeper into the glenoid and creating a greater spher-ical surface.

922 J. James et al.

is critical to implant fixation.5 Bone-implant contact is alsorelated to implant geometry19; therefore, intuitively, thenatural concave contour of the native glenoid would favora convex baseplate. An overwhelming majority of theliterature examining the difference between convex and flatglenoid components was specific to standard total shoulderarthroplasty. Although biomechanical studies havedemonstrated improved performance of convex poly-ethylene glenoid component design,1,13 the effect of gle-noid baseplate geometry has not been thoroughlyinvestigated with respect to reverse shoulder arthroplasty.

A biomechanical study dealing with glenoid designdemonstrated that a convex backed glenoid would ulti-mately lead to stress transmission in compression morethan shear as compared to a flat backed component.1 Thisstudy1 also suggested that a curved shape will preservemore bone during implantation. In contrast, the presentstudy showed that convex baseplate implantation requiredstatistically greater removal of bone as compared to a flatbaseplate. This may be due to (a) the convex baseplate keelbeing larger than the flat baseplate post, and (b) in theseglenoids, the reamer for the convex baseplate has a smallerradius of curvature than the actual glenoid curvature.

Judicious reaming is supported by our results showingthat the volume of bone removed statistically increasedwith incremental reaming. This is especially important inlight of the study by Frich et al,7 where they showed that aslittle as 1 mm of reaming would result in a 25% reductionin compressive strength, and 2 mm would decrease thelevel of compressive strength by 70%.

Biomechanical studies of standard glenoid designs haveshown that reaming that improves conformity between theglenoid component and glenoid bone can minimize normal

eccentric loads,4 and that, in the presence of eccentricloading, a curved back glenoid component demonstratedless tilt and displacement when compared to a flat backdesign.13 Although our study focused on reverse glenoiddesigns, our results demonstrate that the glenoid surfacearea available for fixation was greater with the convexbaseplate compared to the flat and thus may help explainthese biomechanical studies. This is further emphasized byour results showing that the amount of total bone area incontact with a convex base plate was greater than witha flat plate.

Screw fixation/engagement, as noted in the literature, isvitally important to the reduction of glenoid loosening ofboth standard glenoid designs2 and in reverse shoulderdesigns.3 As stated by Virani et al, a stable interface willallow bone ingrowth which provides long-term attachmentbetween the baseplate and the glenoid.20 Screw fixation hasbeen shown to counteract liftoff tensile displacementsproduced by eccentric loading of the glenohumeral joint ina standard glenoid design.2 Initial rigid fixation of reverseglenoid baseplates is dependent on the surgical placementof the screws and the quality of the glenoid bone stock.10

Stable screw fixation of reverse glenoid designs has beencorrelated with an increase of screw surface area within theglenoid bone.12 In a study of reverse glenoid components,Hopkins et al11 demonstrated that using a convex backedglenoid implant would allow screws to be placed furtherapart than with a flat backed design, thereby resulting ingreater resistance to interface motion. Our results demon-strated that there was no difference between superior andinferior screw engagement for the flat baseplate; however,inferior screw purchase was greater compared to superiorscrew engagement for the convex baseplate. We alsoobserved there was less overall screw engagement witha convex baseplate compared to a flat baseplate. The

Figure 14 The amount of screw engagement for the flat andconvex baseplates, for both the inferior and superior screws, asa function of the amount of bone removed.

Reverse shoulder arthroplasty 923

importance of careful reaming was underscored by ourresults, which showed that reaming as little as 1 mm ofbone statistically decreased the amount of screw engage-ment in the glenoid.

A possible limitation to this study is that the baseplatedesigns considered here have different baseplate surfaceareas; however, these designs were considered to be thebest comparison of a flat and convex baseplate. Thesedesigns also differ by the size of the post or keel and thesize and shape of the baseplate, each of which may affectbaseplate fixation. Bone ingrowth, although assumed to becorrelated with glenoid baseplate fixation, was notmeasured in this study. In this study, we also did notquantify the bone density of either the cortical or cancel-lous bone, because here we inserted both implants into eachscapula, they served as their own control. Finally, maximalscrew engagement was obtained for each screw that wasvirtually implanted. This screw engagement did notdifferentiate between cortical and cancellous bone due tothe obscurity of this measurement when the screw was incontact with both types of bone. Also, the ability to obtainthese screw trajectories, in vivo, was not explored. Apossible area of future research would be to compare theamount of micromotion associated with physiologicloading in a convex back versus flat back glenoid baseplatein a cadaveric model.

Conclusion

Glenoid implant fixation in reverse shoulder arthroplastyrelies on a variety of factors. Preservation of bone stockthrough judicious reaming is critically important.Reduction of micromotion through secure initial screwfixation, maximization of bone-implant contact, and,ultimately, bone ingrowth can help prevent the

disastrous complication of glenoid loosening in reverseshoulder replacement. In this study’s comparison of 2baseplate designs, whose major difference is whetherthey were either a flat backed design or a convex base-plate, demonstrated that the use of a convex back gle-noid baseplate can improve the contact surface areabetween bone and implant, but at the cost of decreasedscrew engagement and increased bone volume removed.In situations where adequate bone-implant contactcannot be achieved, better screw engagement wouldmake a flat back glenoid baseplate a more favorableoption.

Disclaimer

The authors, their immediate family, and any researchfoundation with which they are affiliated did not receiveany financial payments or other benefits from anycommercial entity related to the subject of this article.

References

1. Anglin C, Wyss UP, Pichora DR. Mechanical testing of shoulder

prostheses and recommendations for glenoid design. J Shoulder Elbow

Surg 2000;9:323-31.

2. Bicknell RT, Liew ASL, Danter MR, Patterson SD, King GJW,

Chess DG, et al. Does keel size, the use of screws, and the use of bone

cement affect fixation of a metal glenoid implant? J Shoulder Elbow

Surg 2003;12:268-75. doi:10.1016/S1058-2746(02)00028-9

3. Chebli C, Huber P, Watling J, Bertelsen A, Bicknell RT, Matsen F III.

Factors affecting fixation of the glenoid component of a reverse total

shoulder prosthesis. J Shoulder Elbow Surg 2008;17:323-7. doi:10.

1016/j.jse.2007.07.015

4. Collins D, Tencer A, Sidles J, Matsen F III. Edge displacement and

deformation of glenoid components in response to eccentric loading.

The effect of preparation of the glenoid bone. J Bone Joint Surg -

Series A 1992;74:501-7.

5. Dujardin FH, Mollard R, Toupin JM, Coblentz A, Thomine JM.

Micromotion, fit, and fill of custom made femoral stems designed with

an automated process. Clin Ortho Rel Res 1996:276-89.

6. Frankle M, Levy JC, Pupello D, Siegal S, Saleem A, Mighell M, et al.

The reverse shoulder prosthesis for glenohumeral arthritis associated

with severe rotator cuff deficiency. A minimum two-year follow-up

(study of sixty patients surgical technique. J Bone Joint Surg Am

2006;88(Suppl 1 Pt 2):178-90. doi:10.1302/0301-620X.89B2.18161

7. Frich LH, Jensen NC, Odgaard A, Pedersen CM, Sojbjerg JO,

Dalstra M. Bone strength and material properties of the glenoid. J

Shoulder Elbow Surg 1997;6:97-104.

8. Fukuda K, Chen CM, Cofield RH, Chao EYS. Biomechanical analysis

of stability and fixation strength of total shoulder prostheses. Ortho-

pedics 1988;11:141-9.

9. Gutierrez S, Greiwe RM, Frankle MA, Siegal S, Lee WE III.

Biomechanical comparison of component position and hardware

failure in the reverse shoulder prosthesis. J Shoulder Elbow Surg 2007;

16(3 Suppl):S9-12. doi:10.1016/j.jse.2005.11.008

10. Harman M, Frankle M, Vasey M, Banks S. Initial glenoid component

fixation In "reverse" total shoulder arthroplasty: A biomechanical

924 J. James et al.

evaluation. J Shoulder Elbow Surg 2005;14:162S-7S. doi:10.1016/j.

jse.2004.09.030

11. Hopkins AR, Hansen UN. Primary stability in reversed-anatomy gle-

noid components. Proceedings of the institution of mechanical engi-

neers, Part H. J Eng Med 2009;223:805-12. doi:10.1243/

09544119JEIM557

12. Hopkins AR, Hansen UN, Bull AMJ, Emery R, Amis AA. Fixation of

the reversed shoulder prosthesis. J Shoulder Elbow Surg 2008;17:974-

80. doi:10.1016/j.jse.2008.04.012

13. Lacaze F, Kempf JF, Bonnomet F, Boutemy P, Colin F. Primary

fixation of glenoid implants: An in vitro study. In: Gilles Walch PB,

editor. Shoulder arthroplasty. Berlin: Springer-Verlag; 1999. p. 141-6.

14. Levigne C, Boileau P, Favard L, Garaud P, Mole D, Sirveaux F, et al.

Scapular notching in reverse shoulder arthroplasty. J Shoulder Elbow

Surg 2008;17:925-35. doi:10.1016/j.jse.2008.02.010

15. Middernacht B, De Roo PJ, Van Maele G, De Wilde LF. Consequences

of scapular anatomy for reversed total shoulder arthroplasty. Clin

Ortho Rel Res 2008;466:1410-8. doi:10.1007/s11999-008-0187-6

16. Nguyen D, Ferreira LM, Brownhill JR, Faber KJ, Johnson JA. Design

and development of a computer assisted glenoid implantation tech-

nique for shoulder replacement surgery. Comput Aid Surg 2007;12:

152-9. doi:10.1080/10929080701374315

17. Stone KD, Grabowski JJ, Cofield RH, Morrey BF, An KN. Stress

analyses of glenoid components in total shoulder arthroplasty.

J Shoulder Elbow Surg 1999;8:151-8.

18. Sutton LG, Werner FW, Jones AK, Close CA, Nanavati VN. Opti-

mization of glenoid fixation in reverse shoulder arthroplasty using

3-dimensional modeling. J Shoulder Elbow Surg 2010;19:664-9. doi:

10.1016/j.jse.2009.12.003

19. Szabo I, Buscayret F, Edwards TB, Nemoz C, Boileau P, Walch G.

Radiographic comparison of flat-back and convex-back glenoid

components in total shoulder arthroplasty. J Shoulder Elbow Surg

2005;14:636-42. doi:10.1016/j.jse.2005.05.004

20. Virani NA, Harman M, Li K, Levy J, Pupello DR, Frankle MA. In

vitro and finite element analysis of glenoid bone/baseplate interaction

in the reverse shoulder design. J Shoulder Elbow Surg 2008;17:509-

21. doi:10.1016/j.jse.2007.11.003

21. Werner CML, Steinmann PA, Gilbart M, Gerber C. Treatment of

painful pseudoparesis due to irreparable rotator cuff dysfunction

with the delta Iii reverse-ball-and-socket total shoulder prosthesis.

J Bone Joint Surg - Series A 2005;87:1476-86. doi:10.2106/JBJS.D.

02342

22. Zimmer. Anatomical shoulder inverse/reverse system: Surgical tech-

nique. 2010. http://www.zimmer.com/content/pdf/en-US/Anatomical_

Shoulder_Inverse_Reverse_System_Surgical_Technique_97-4223-102-

00_Rev2_10_2010_.pdf.

23. Zimmer. Zimmer Trabecular metal reverse shoulder system: Surgical

technique. 2010. http://www.zimmer.com/content/pdf/en-US/Trab

ecular_Metal_Reverse_Shoulder_System_Surgical_Technique_97-43

09-003-00_Rev_4_07_2010.pdf.