Evaluation of abduction range of motion and avoidance ofinferior scapular impingement in a reverse shoulder model

Sergio Gutierrez, MS,a,b Jonathan C. Levy, MD,c Mark A. Frankle, MD,a Derek Cuff, MD,a Tony S. Keller, PhD,a

Derek R. Pupello, MBA,a and William E. Lee III, PhD,b Temple Terrace, Tampa, and Fort Lauderdale, FL

The purpose of this study was to determine the effects ofprosthetic design and surgical technique of reverseshoulder implants on total abduction range of motion andimpingement on the inferior scapular neck. Customimplants in three glenosphere diameters (30, 36, and 42mm), with 3 different centers of rotation offsets (0, +5,and +10 mm), were placed into a Sawbones scapula(Pacific Research Laboratories, Vashon, WA) in 3different positions: superior, center, and inferior glenoid.Humeral sockets were manufactured with a 130�, 150�,and 170� neck-shaft angle. Four independent factors(glenosphere diameter, center of rotation offset,glenosphere position on the glenoid, and humeral neck-shaft angle) were compared with the 2 dependent factorsof range of motion and inferior scapular impingement.Center of rotation offset had the largest effect on range ofmotion, followed by glenosphere position. Neck-shaftangle had the largest effect on inferior scapularimpingement, followed by glenosphere position. Thisinformation may be useful to the surgeon when decidingon the appropriate reverse implant. (J ShoulderElbow Surg 2008;17:608-615.)

Reverse shoulder arthroplasty is a successful surgicalprocedure to treat pain and provide functional im-provement in patients with glenohumeral arthritis androtator cuff deficiency.1,3,11,14 However, careful ex-amination of the functional outcomes seen with thereverse shoulder replacement reveals variable im-provements in range of motion. Valenti et al12 andBoulahia et al2 showed active elevation ranging

From the aFlorida Orthopaedic Institute Research Foundation; bBio-mechanics Laboratory, Biomedical Engineering, University ofSouth Florida; and cOrthopaedic Institute at Holy Cross.

This research was supported in part by the Florida OrthopaedicInstitute Research Foundation.

Reprint requests: Dr Mark Frankle, Florida Orthopaedic Institute,13020 N Telecom Pkwy, Temple Terrace, FL 33637. (E-mail:frankle@pol.net).

Copyright ª 2008 by Journal of Shoulder and Elbow SurgeryBoard of Trustees.

1058-2746/2008/$34.00doi:10.1016/j.jse.2007.11.010

608

from 30� to 100� and external rotation from 20� to50�. Frankle et al3 showed active elevation rangingfrom 30� to 180� and external rotation from 10� to65�. This variability is likely caused by multiple factors,including severity of disease, variable degrees of mus-cle loss, surgical technique, and prosthetic design.

Inferior impingement of the reverse shoulder im-plant on the inferior scapular neck has been noted tobe the mechanism for the development of scapularnotching.1,11 Typically, this impingement occurswhen the arm is in a resting position and, biomechan-ically, has been referred to as an adduction deficit.7

Reduction of the adduction deficit is of particular inter-est, because progressive scapular notching has beenobserved to a variable degree radiographically, in-cluding 56% by Valenti et al,12 63% by Boulahia et al,1

65% by Sirveaux et al,11 74% by Boileau et al,1 and96% by Werner et al14 and has even been implicatedas the cause of failure in several patients.13 A previousstudy by Nyffeler et al7 demonstrated adduction defi-cit was decreased by placing the base plate flushwith the inferior edge of the glenoid, with the gleno-sphere extending below the inferior border of thescapula. This result suggested that surgical techniquecould help to reduce adduction deficit.

Looking specifically at prosthetic design, we findseveral different reverse shoulder implants are cur-rently available and many others are likely in develop-ment. Each implant differs in several basic designparameters, including center of rotation (COR) offset,glenosphere diameter, and humeral neck-shaft anglerelative to the horizontal plane. The COR offsets canvary from 0 to 10 mm lateral to the glenoid fossa.The diameter of available glenospheres also variesfrom 32 to 42 mm, and humeral neck-shaft anglesrange from 135� to 155�. The implication of these dif-ferent design factors on shoulder kinematics is poorlyunderstood and may have a dramatic influence on out-comes after surgical reconstruction. To date, no biome-chanical study has systematically evaluated the effectof reverse shoulder prosthesis design and implant posi-tioning on glenohumeral motion.

The purpose of this study was not to create a surgicaltechnique but to determine how different parameterscontribute to the total glenohumeral abduction rangeof motion (ROM) and adduction deficit in a reverse

J Shoulder Elbow Surg Gutierrez et al 609Volume 17, Number 4

shoulder model. Our hypothesis was that glenosphereposition, COR offset, glenosphere diameter, andhumeral neck-shaft angle had different effects on theabduction ROM and adduction deficit.

MATERIALS AND METHODS

Reverse shoulder implant components consisted of a ballthat was attached to the glenoid (glenosphere) and a humeralsocket that was attached to a wooden dowel. These compo-nents were manufactured using Delrin (DuPont, Wilmington,DE), which is a wear-resistant, low-friction plastic. The gleno-spheres were manufactured with 3 diameters (30, 36, and42 mm) and 3 COR offsets (0 mm or hemispherical, +5 mm,and +10 mm offset from the glenoid), as summarized in Table I.The glenoid components were rigidly attached to the glenoidsurface of a Sawbones shoulder model of a large left scapula(model1050-10,PacificResearchLaboratories,VashonWA).

To implant the glenospheres in a consistent manner, weused the block on the medial side of the Sawbones scapulaas a reference for measurement. The glenoid on each Saw-bones scapula was reamed flat so that the plane of the gle-noid was parallel to the plane of the medial border of theblock of the scapula.

Three different positions on the glenoid were studied: su-perior, neutral, and inferior (Figure 1). The neutral positionwas centered in the glenoid, and the superior and inferiorpositions were halfway between the center and the superiorand inferior edges of the glenoid, respectively. Variations inglenosphere component geometry and placement on theglenoid were consistent with clinical practice, with the ex-ception of the superiorly placed glenospheres.1,3-5,7,10,11

Although rarely used in practice, the superior position wasincluded in this analysis to understand its effect on ROMand inferior scapular impingement.

Humeral components were manufactured for each gleno-sphere with the 3 humeral neck-shaft angles of 130�, 150�,and 170�. The inside diameter of the humeral socketmatched the glenosphere diameter, and the socket was de-signed with a constant depth/radius (d/r) ratio of 0.56.This d/r ratio was chosen as the mean of the commercial re-verse implants (range, 0.46-0.67). A hole was machined inthe humeral socket component to orient the sockets at eachof the 3 neck-shaft angles (Figure 2, A). Machining toler-ances were approximately 0.05 mm, and machined compo-nent geometries were measured using a digital caliper(0.025-mm precision). The humeral socket outer diametersfor this study were held constant at 50 mm throughout alldevices, which is a typical diameter for the normal humeralhead. Table I summarizes the humeral component depth foreach of the 3 socket diameters and the 3 humeral angles. Awooden dowel was inserted into the hole to simulate the hu-meral shaft. The dowel was 33 cm long, which is the approx-imate length of the average humerus.4

The Sawbones scapula model was used in conjunctionwith a 3-dimensional coordinate measurement system tomeasure total glenohumeral abduction ROM of the humeralsocket component in the scapular plane (Figure 2, B). Thescapula was rigidly fixed and oriented to simulate the 30�

angle of the scapular plane and tilted 23� anteriorly to thesagittal plane. The scapula was held in neutral abduc-tion with the glenoid face perpendicular to the floor. A

6-degrees-of-freedom, electromagnetic goniometer (Flockof Birds, Ascension Technology Corporation, Burlington,VT), with an accuracy of 0.05 mm and 0.15�, was rigidlyattached to the distal end of the wooden dowel.

With the scapula-glenoid component fixed, each of the 9glenospheres was evaluated using the 3 different humeralneck-shaft angled components. Glenohumeral abductionROM was limited superiorly by impingement of the socketon either the superior edge of the glenoid or the acromion,whereas glenohumeral adduction was limited by impinge-ment on the inferior glenoid or scapula (adduction deficit)or 0� (neutral position of the humeral shaft), whicheveroccurred first. The humeral component (dowel) was manu-ally manipulated from minimum adduction to maximumabduction. The X, Y, and Z coordinates were recorded atminimum adduction and at maximum abduction, wherethe X and Z coordinates corresponded to the abductionplane (Figure 2, B). The adduction deficit was determinedby the resting position in maximal adduction. If adductionwas 0�, no adduction deficit was present. Total glenohum-eral abduction ROM was determined from the differencebetween maximal adduction and maximum abduction.

Statistical analyses were conducted using JMP software(SAS Institute, Cary, NC). Four independent factors (diame-ter, COR offset, glenoid placement, and humeral neck-shaftangle) were compared with the dependent factors (abduc-tion ROM and adduction deficit angle). Descriptive statisticswere performed using a standard least-squares regression.A multivariate analysis of variance was used to analyzethe effect of each factor on the dependent variables. A bal-anced factorial design, with the same number of observa-tions for each factor, was used. The significance level wasset at P < .05 for all statistics.

RESULTS

Total abduction range of motion

The greatest total abduction ROM was 117.5� (42mm, +10-mm COR, and inferior, 170�), whereas theleast maximum total abduction ROM was 40.2� (30mm, 0-mm COR, and neutral, 170�; and 30 mm,0-mm COR, and neutral, 150�; Table II). Maximal

Table I Glenosphere and humerosocket component geometry

Glenosphere Humerosocket

Diameter(mm)

COR offset(mm)

Humeral angle(�)

Depth(mm)

30 0 130 8.430 5 150 8.430 10 170 8.436 0 130 10.136 5 150 10.136 10 170 10.142 0 130 11.842 5 150 11.842 10 170 11.8

COR, center of rotation.

610 Gutierrez et al J Shoulder Elbow SurgJuly/August 2008

Figure 1 Photograph sequence illustrates the 9 glenoid component arrangements, consisting of the 3 center of ro-tation offsets of 0, +5 and +10 mm and the 3 glenosphere positions of superior (S), neutral (N), and inferior (I), foreach of the 3 different diameter glenospheres (30, 36 and 42 mm).

abduction was limited by impingement on either theacromion or the superior edge of the glenoid. Signifi-cant effects on total glenohumeral abduction ROMwere found for all the factors studied (P < .0001).The factor with the greatest effect on total abductionROM was the glenosphere COR offset (P < .0001,F ¼ 2118), followed by glenoid position (P < .0001,F ¼ 1740), glenosphere diameter (P < .0001, F ¼79), and humeral angle (P < .0001, F ¼ 77). Gleno-spheres with a positive COR offset improved the totalabduction ROM for all glenoid positions examined.Glenospheres with a COR offset of +10 mm were as-sociated with up to a 91% increase (neutral glenoidposition) in total abduction ROM, compared with gle-nospheres with no COR offset (0 mm; Figure 3).

Adduction deficit

The largest adduction deficit was 64.4� (30 mm,0-mm COR, and superior, 170�), whereas the minimum

adduction deficit was 0� or no adduction deficit (TableIII). Significant effects on adduction deficit were foundfor all the factors studied (P < .0001). The factor withthe greatest effect on decreasing the adduction deficitwas the humeral neck-shaft angle (P < .0001, F ¼3264), followed by glenosphere position (P < .0001,F ¼ 2054), glenosphere COR offset (P < .0001, F ¼1212), and glenosphere diameter (P < .0001, F ¼116). The 3 specific factors that had the greatest effecton adduction deficit were the 130� humeral neck-shaftangle, inferior position, and +10 mm COR offset (P <.0001; Figure 4).

DISCUSSION

A careful analysis of the outcomes after reverseshoulder replacement reveals variable improvementin shoulder elevation.2,3,12 To judge these improve-ments accurately, isolated glenohumeral motion must

J Shoulder Elbow Surg Gutierrez et al 611Volume 17, Number 4

Figure 2 A, Photographs show the 3 different humeral neck-shaft angles. The 170� humeral neck shaft angle isnot currently available in clinical practice. B, Schematic illustration shows the experimental setup used for adduction-abduction range of motion measurements.

be evaluated; however, this information has beenlargely lacking up to now. Seebauer et al9,10 con-ducted the only clinical study to isolate the improve-ment in glenohumeral elevation after a reverseshoulder implant. On the basis of dynamic fluoro-scopic radiographs, they reported that the maximumactive glenohumeral abduction ROM in the scapularplane, using the Delta III prosthesis (DePuy, Inc, War-saw, IN), was 53�. A similar amount of glenohumeralmotion was seen in a cadaver model using the sameprosthesis.7 Nyffler et al7 evaluated the abductionROM of the Delta III with a 36-mm glenosphere.When implanted using the manufacturer’s recommen-ded surgical technique, the mean abduction ROM inthe scapular plane ranged from 25� to 67� (totalabduction ROM, 42�). When implanted in an inferiorposition on the glenoid, the mean abduction ROMranged from 1� to 81� (total abduction ROM, 80�).7

Thus, modification of surgical technique not only im-proved the overall motion but also helped to limit theadduction deficit from 25� for the manufacturer’s rec-ommended placement to 1� for an inferior placementon the glenoid.

In the current study, evaluation of abduction ROMnoted statistically significant differences for differentimplant designs and changes in implant position onthe glenoid. The variable that resulted in the greatestimprovement in ROM was the COR offset (P <.0001, F ¼ 2118). The larger the COR offset, thegreater the abduction motion. In addition, placementof the glenosphere inferiorly on the glenoid resultedin improved motion (P < .0001, F ¼ 1740). Movingthe center of rotation farther away from the scapula,or placing the glenosphere more inferiorly, gives thehumeral socket more clearance before impinging onthe acromion or superior glenoid, thereby maximizingglenohumeral abduction ROM.

Although glenosphere diameter and humeral angleresulted in improvement in motion, these were smallcompared with the COR offset and glenosphere posi-tion. This can be exemplified by comparing differencesin ROMbetweendifferent diameters anddifferentCORoffsets as well as comparisons between different gleno-sphere positions and different neck-shaft angles (TableII). For example, changes in diameter netted a ROMimprovement of only 5.5� (30 to 42 mm, 0-mm offset,

612 Gutierrez et al J Shoulder Elbow SurgJuly/August 2008

Table II Glenohumeral abduction range of motion measurements (mean 6 standard deviation) for the 4 different design factors studied

Abduction range of motion (�)

Glenosphereposition COR offset

Humeralcomponent 30 mm 36 mm 42 mm

130� 48.2 6 5.8 45.2 6 4.8 43.2 6 3.30 mm 150� 47.9 6 5.8 43.9 6 3.8 43.4 6 4.5

170� 48.6 6 5.3 45.2 6 3.6 43.7 6 3.7130� 58.8 6 5.4 55.0 6 5.2 54.4 6 5.3

Superior +5 mm 150� 58.1 6 4.9 55.4 6 5.2 53.8 6 4.8170� 58.9 6 4.5 55.8 6 4.6 53.9 6 4.2130� 70.5 6 6.5 67.6 6 5.0 64.8 6 3.5

+10 mm 150� 69.4 6 6.3 66.8 6 5.3 66.1 6 5.1170� 69.7 6 6.0 67.9 6 5.7 66.3 6 5.2130� 41.8 6 1.1 50.0 6 1.8 59.2 6 2.3

0 mm 150� 40.2 6 0.8 49.9 6 0.9 57.6 6 1.2170� 40.2 6 2.0 50.6 6 0.5 58.2 6 1.3130� 68.1 6 2.3 75.5 6 1.2 83.1 6 2.0

Neutral +5 mm 150� 67.4 6 0.7 74.7 6 1.3 83.1 6 1.4170� 67.6 6 1.0 75.4 6 1.2 83.6 6 0.7130� 90.1 6 1.3 92.3 6 0.9 94.5 6 1.0

+10 mm 150� 89.1 6 0.5 95.9 6 1.0 102.7 6 1.1170� 89.3 6 0.7 96.2 6 0.4 102.9 6 0.6130� 62.3 6 0.7 65.7 6 0.9 67.8 62.4

0 mm 150� 65.3 6 0.9 73.7 6 1.4 79.5 6 0.8170� 65.5 6 1.2 73.0 6 1.1 79.1 6 0.7130� 73.3 6 1.1 76.5 6 1.3 77.0 6 1.5

Inferior +5 mm 150� 86.4 6 1.1 92.2 6 1.3 96.2 6 1.0170� 86.2 6 0.7 92.5 6 1.4 98.4 6 0.6130� 84.3 6 1.1 87.1 6 2.1 89.2 6 2.7

+10 mm 150� 102.2 6 0.7 105.1 6 1.2 106.6 6 0.7170� 106.0 6 0.9 112.1 6 0.9 117.5 6 0.8

COR, center of rotation.

inferior placement, 130� neck-shaft angle), whereaschanges in the COR offset netted a larger change of22� (0- to +10-mm offset, 30 mm, inferior placement,130� neck-shaft angle). Changes in neck-shaft angleshowed a small change of 3.2� (130� to 170� neck-shaft angle, 0-mm offset, 30 mm, inferior placement)compared with 20.5� for a change in the glenosphereposition (neutral to inferior placement, 0-mm offset, 30mm, 130� neck-shaft angle). Thus, maximizing abduc-tion ROM is best achieved with a larger COR offset andinferior translation of the glenosphere placement.

Examination of the adduction deficit revealed sig-nificant differences, depending on the design exam-ined and the position of implantation. In general, theadduction deficit was primarily dependent on the hu-meral component angle (P < .0001, F ¼ 3264), fol-lowed by glenosphere position (P < .0001, F ¼2054) and glenosphere COR offset (P < .0001, F ¼1212). Larger glenosphere diameters were able tolimit the adduction deficit only minimally (P < .0001,F ¼ 116). Several of the constructs displayed noadduction deficit and were, therefore, able to be ad-ducted to at least 0�. Thus, modifications in both surgi-

cal technique (inferior translation) and prostheticdesign (more varus neck-shaft angle and larger CORoffset) resulted in a reduction of the adduction deficit.

A Sawbones scapula model was used to evaluatebiomechanically the effects of changing the COR off-set, glenosphere position, glenosphere diameter,and humeral neck-shaft angle on glenohumeral abduc-tion ROM and adduction deficit in reverse shoulderimplants. The major advantage of using a Sawbonesscapula model was the ability to test inherent differ-ences in ROM related to the geometry of the devices,independent of anatomic differences present whencadaver models are used.6,8 Using a cadaver model,Nyffler et al7 noted that motion was always limited byimpingement on areas of the scapula. The Sawbonesmodel was able to replicate a consistent model of thescapular anatomy best in an effort to study how motionis limited by scapular impingement.

Limitations of this study include omission of the prox-imal humeral anatomy, lack of variation in gleno-sphere tilt, changes relating to human scapularmorphology (including inclination of the inferior gle-noid neck and its intersection with the lateral body of

J Shoulder Elbow Surg Gutierrez et al 613Volume 17, Number 4

the scapula), no scapulothoracic motion, notching inlocations other than inferior to the glenoid component,and truncation of the glenoid vault, which can occurduring reaming.7 In the anatomic shoulder, ROM is

limited by mechanical impingement as well as bysoft tissue tension. Presumably, similar impingementpoints are present in reverse shoulder arthroplasty,but actual impingement can vary greatly depending

Figure3 Graph shows the percentage difference in abduction range of motion (ROM) between components with +5and +10 mm center of rotation (COR) offsets (arranged according to glenosphere position). The mean combinedROM and COR offset data (n ¼ 45) is presented with the standard deviation (error bars).

Table III Adduction deficit measurements (mean 6 standard deviation) for the 4 different design factors studied

Abduction deficit angle (�)

Glenosphere position COR offset Humeral component 30 mm 36 mm 42 mm

130� 27.9 6 1.2 24.6 6 1.0 21.5 + 1.60 mm 150� 46.4 6 1.7 44.2 6 0.7 39.9 6 1.7

170� 64.4 6 1.5 61.4 6 0.7 58.4 6 1.6130� 15.6 6 1.1 13.4 6 0.9 9.9 6 2.0

Superior +5 mm 150� 34.2 6 1.7 32.1 6 0.5 29.5 6 1.5170� 52.6 6 1.2 50.2 6 0.4 47.2 6 1.9130� 3.7 6 0.4 1.9 6 0.7 2.0 6 0.6

+10 mm 150� 21.9 6 1.3 19.9 6 0.8 17.4 6 1.6170� 40.3 6 1.4 37.4 6 2.1 35.4 6 1.2130� 24.0 6 1.4 21.2 6 1.6 16.0 6 1.9

0 mm 150� 43.7 6 1.3 39.8 6 1.0 34.8 6 1.6170� 62.6 6 2.1 57.1 6 0.9 53.1 6 1.3130� 11.5 6 3.0 7.7 6 1.5 2.8 6 2.1

Neutral +5 mm 150� 29.6 6 0.8 27.0 6 1.5 21.5 6 2.1170� 47.8 6 0.9 44.2 6 1.3 39.9 6 1.4130� 1.1 6 0.5 0�* 0�*

+10 mm 150� 18.4 6 0.6 14.7 6 1.7 10.5 6 1.5170� 36.5 6 0.8 33.4 6 0.9 29.2 6 1.3130� 0�* 0�* 0�*

0 mm 150� 15.1 6 1.4 10.4 + 1.8 6.0 6 0.9170� 32.9 6 1.3 28.7 6 1.5 25.0 6 1.4

Inferior 130� 0�* 0�* 0�*+5 mm 150� 5.4 6 1.2 3.4 6 1.9 0�*

170� 23.4 6 1.0 20.1 6 2.3 16.1 6 1.4130� 0�* 0�* 0�*

+10 mm 150� 0�* 0�* 0�*170� 15.0 6 0.8 11.5 6 1.1 7.9 6 1.0

COR, center of rotation.*No adduction deficit.

614 Gutierrez et al J Shoulder Elbow SurgJuly/August 2008

on the placement of the humeral component in the hu-meral shaft and glenosphere orientation. Given therelatively large number of design factors consideredin this study, we elected to omit considerations of gle-nosphere tilt and proximal humeral geometry andfocused on the effects of humeral and glenoid compo-nent geometry on abduction ROM and inferior scapu-lar impingement.

One other limitation was the lack of soft tissue ten-sion (muscle and tendon forces) in the mechanicalmodel. Readers should be cautioned that our findingsmay have involved prosthetic combinations and posi-tions that are clinically unfeasible owing to the exces-sive soft tissue tension they would generate that couldlead to limited motion and stiffness (ie, overstuffing thejoint) or to the lack of soft tissue tension that could lead

Figure 4 Photographs show the differences in adduction deficit. A and B, Center of rotation (COR) offset of 0 mm vsa COR offset of +10 mm. C and D, Superior placement on the glenoid vs inferior placement on the glenoid. E and F,A 170� neck-shaft (N-S) angle vs a 130� N-S angle.

J Shoulder Elbow Surg Gutierrez et al 615Volume 17, Number 4

to instability. It should be stated that this study did notdetermine the safe limits of any of the parameterstested, and component size and position must be indi-vidualized for each clinical situation.

Ultimately, several important design and surgicalfactors must be considered when a reverse shoulderimplant is selected. These include, but are not limitedto, baseplate-host bone fixation, stability (resistanceto subluxation and dislocation), muscular weaknessor deficiency, the degree of bone loss, and soft tissuetension. In cases where optimal baseplate fixation canbe achieved and the risk of instability is minimal, max-imization of function may be considered. In thesecases, surgeons may wish to select an implant that al-lows for the largest ROM and the least amount of ad-duction deficit. The results of this study indicate thatglenospheres with a greater distance from the glenoidto the COR and an inferior placement on the glenoidprovide for greater potential ROM. Adduction deficitcan best be improved by selecting prosthesis witha varus neck-shaft angle and inferior placement ofthe glenosphere on the glenoid.

We thank Karmen A. Anderson and Milton Bertrand fortheir technical assistance. We dedicate this work to the mem-ory of Dr Tony S. Keller and Karmen A. Anderson. You willbe missed.

REFERENCES

1. Boileau P, Watkinson DJ, Hatzidakis AM, Balg F. Grammont re-verse prosthesis: design, rationale, and biomechanics. J ShoulderElbow Surg 2005;14:147S-61S.

2. Boulahia A, Edwards TB, Walch G, Baratta RV. Early results of a re-verse design prosthesis in the treatment of arthritis of the shoulder inelderly patients with a large rotator cuff tear. Orthopedics 2002;25:129-33.

3. Frankle M, Siegal S, Pupello D, Saleem A, Mighell M, Vasey M.The Reverse Shoulder Prosthesis for glenohumeral arthritis associ-ated with severe rotator cuff deficiency. A minimum two-year fol-low-up study of sixty patients. J Bone Joint Surg Am 2005;87:1697-705.

4. Grammont PM, Baulot E. Delta shoulder prosthesis for rotator cuffrupture. Orthopedics 1993;16:65-8.

5. Harman M, Frankle M, Vasey M, Banks S. Initial glenoid compo-nent fixation in ‘‘reverse’’ total shoulder arthroplasty: a biomechan-ical evaluation. J Shoulder Elbow Surg 2005;14:162S-7S.

6. Lucas B, Asher M, McIff T, Lark R, Burton D. Estimation of transverseplane pelvic rotation using a posterior-anterior radiograph. Spine2005;30:E20-7.

7. Nyffeler RW, Werner CM, Gerber C. Biomechanical rele-vance of glenoid component positioning in the reverse DeltaIII total shoulder prosthesis. J Shoulder Elbow Surg 2005;14:524-8.

8. Schep NW, van Walsum T, De Graaf JS, Broeders IA, van derWerken C. Validation of fluoroscopy-based navigation in the hipregion: what you see is what you get? Comput Aided Surg2002;7:279-83.

9. Seebauer L. Reverse prosthesis through a superior approach forcuff tear arthropathy. Tech Shoulder Elbow Surg 2006;7:13-26.

10. Seebauer L, Walter W, Keyl W. Reverse total shoulder arthro-plasty for the treatment of defect arthropathy. Oper Orthop Trau-matol 2005;17:1-24.

11. Sirveaux F, Favard L, Oudet D, Huquet D, Walch G, Mole D.Grammont inverted total shoulder arthroplasty in the treatment ofglenohumeral osteoarthritis with massive rupture of the cuff. Resultsof a multicentre study of 80 shoulders. J Bone Joint Surg Br 2004;86:388-95.

12. Valenti PH, Boutens D, Nerot C. Delta 3 reversed prosthesis for os-teoarthritis with massive rotator cuff tear: long term results (>5years). Shoulder Prosthesis 2000:253-9.

13. Vanhove B, Beugnies A. Grammont’s reverse shoulder prosthesisfor rotator cuff arthropathy. A retrospective study of 32 cases.Acta Orthop Belg 2004;70:219-25.

14. Werner CM, Steinmann PA, Gilbart M, Gerber C. Treatment ofpainful pseudoparesis due to irreparable rotator cuff dysfunctionwith the Delta III reverse-ball-and-socket total shoulder prosthesis.J Bone Joint Surg Am 2005;87:1476-86.