Acromion-ﬁxation of glenoidcomponents in total shoulder arthroplasty

7/30/2019 Acromion-fixation of glenoidcomponents in total shoulder arthroplasty

http://slidepdf.com/reader/full/acromion-xation-of-glenoidcomponents-in-total-shoulder-arthroplasty 1/10

Journal of Biomechanics 38 (2005) 1702–1711

Acromion-fixation of glenoid components in total shoulderarthroplasty

Linda A. Murphy, Patrick J. PrendergastÃ

aCentre for Bioengineering, Department of Mechanical Engineering, Trinity College, Dublin, Ireland

Accepted 1 June 2004

Abstract

Successful design of components for total shoulder arthroplasty has proven to be challenging. This is because of the difficulties in

maintaining fixation of the component that inserts into the scapula; i.e., the glenoid component. Glenoid components that are

fixated to both the glenoid and acromion (a long process extending medially on the dorsal aspect of the scapula) have the possible

advantage of greater stability over those that are fixated to the glenoid alone. In this study, a finite element analysis is used to

investigate whether or not acromion fixation is advantageous for glenoid components. Full muscle loading and joint reaction forces

are included in the finite element model. Reflective photoelasticity of five scapulae is used to obtain experimental data to compare

with results from the finite element analysis, and it confirms the structural behaviour of the finite element model. When implanted

with an acromion-fixated prosthesis, it is found that high unphysiological stresses occur in the scapula bone, and that stresses in the

fixation are not reduced. Very high stresses are predicted in that part of the prosthesis which connects the acromion to the glenoid. It

is found that the very high stresses are partly in response to the muscle and joint reaction forces acting at the acromion. It is

concluded that, because of the relatively high forces acting at the acromion, fixation to it may not be the way forward in glenoid

component design.

r 2004 Elsevier Ltd. All rights reserved.

Keywords: Glenoid; Acromion-fixation; Prosthesis; Shoulder arthroplasty

1. Introduction

Whilst arthroplasty of the hip and knee has improved

significantly over the past 30 years, developments in

total shoulder arthroplasty have not yet led to an

implant that has adequate long-term results (Limb,

2002). This slower development rate for shoulder

prostheses may be partly attributed to the lower number

of shoulder replacements performed; in the USAapproximately 20,000 shoulder replacements are carried

out annually compared to 258,000 hip and 299,000 knee

operations (American Academy of Orthopaedic Sur-

geons, 2003). Furthermore, the complex biomechanics

of the shoulder, with motion and force transmission at

the glenohumeral and acromioclavicular joints, creates

many design challenges. The high forces at the glenoid

surface and the small volume of the glenoid cavity create

problems for fixation. The small amount of glenoid bone

is often further diminished by the effects of rheumatoid

arthritis, and this disease also affects the stabilising

function of the rotator cuff. Frich (1994) reported that

complications in total shoulder arthroplasty due toloosening of the glenoid component have been as high as

15% in rheumatoid arthritic patients. Skirving (1999)

also reported that patients with osteoarthritis generally

have a better clinical outcome compared to rheumatoid

arthritic patients. It is clear, therefore, that pre-clinical

testing of glenoid prosthesis performance should be

extended to the rheumatoid arthritic case.

It has been established that constrained (or fixed

fulcrum) total shoulder arthroplasty designs were

ARTICLE IN PRESS

www.elsevier.com/locate/jbiomech

www.JBiomech.com

0021-9290/$ - see front matterr 2004 Elsevier Ltd. All rights reserved.

doi:10.1016/j.jbiomech.2004.06.030

ÃCorresponding author. Tel: +353-1-608-1383; fax: +353-1-679-

5554

E-mail address: [email protected] (P.J. Prendergast).

http://www.elsevier.com/locate/jbiomech

http://www.elsevier.com/locate/jbiomech



problematic (Neer, 1990). These prostheses consisted of

a cemented glenoid component hinged to a humeral

component. The design rationale was based on the

hypothesis that the glenohumeral instability resulting

from a weakened or torn rotator cuff could be

eliminated by a direct constraint. However, since the

healthy joint has such a large range of motion, theconstrained nature of the device resulted in a number of

complications; these included loss of bony fixation,

bending or breakage of the implant, and dysfunction

related to joint stiffness and non-physiological tension-

ing of the surrounding tissues. Designs in current use are

mainly unconstrained (i.e. the glenoid and humeral

components are not hinged). One class of unconstrained

designs have sought to use the acromion to aid in

fixation of the glenoid prosthesis and to prevent

impingement of the humeral head in patients with

rotator cuff deficiency. Many acromion-fixated designs,

which may be termed semi-constrained , have been

introduced, some showing good short-term follow-up

results, as described in detail in Table 1. Acromion-

fixation appeared promising; however, to date, clinical

studies are few and long-term follow-up studies are

generally not available. A definite conclusion as to

whether these acromion-fixated designs are the way

forward, particularly for patients with rheumatoid

arthritis, is still awaited.

The purpose of this paper is to investigate the

biomechanical rationale for acromion-fixation over

non-acromion-fixation devices. One acromion-fixated

prosthesis that showed good results is that of Gagey and

Mazas (1990). We use the method of three-dimensionalfinite element analysis to test the hypothesis that

attachment of the glenoid component to the acromion

increases the durability of the component’s fixation

(Gagey and Mazas, 1990). In particular, the aim is to

determine if load sharing occurs between the acromion

and glenoid so as to reduce the high stresses experienced

in the cement mantle relative to a prosthesis without

acromion-fixation.

2. Materials and methods

2.1. Development of the FE model

A three-dimensional representation of the scapula was

generated from CT scans. This procedure involved

scanning an embalmed left scapula, immersed in water,

with the following parameters (Siemens Somoton Plus

4): 140 kV, 111mA, and 1.0 s scan time along the sagittal

plane and in the coronal direction. Axial images were

reconstructed every 1 mm. Images were used to create a

finite element mesh in MARC (MARC, Palo Alto, CA,

USA); see Lacroix et al. (1997) for the precise details.

Three-dimensional isoparametric elements were used in

the model. Element sizes varied with mesh refinement

around the glenoid neck of the scapula and at the

borders of cortical and cancellous bone (in this region,

the average element size isE1 mm), with good aspect

ratios maintained throughout.

The density of cancellous or cortical bone can vary

greatly in the glenoid (Frich et al., 1997, Mansat et al.,1998). Therefore, to incorporate this potentially im-

portant influence on the stress transfer within the bone,

a CT or Hounsfield number was retrieved for each pixel

of each CT scan and correlated to the density of the

material by a linear interpolation (Hvid et al., 1989).

This CT number is then correlated to the Young’s

modulus (Rice et al., 1988, Schlaffler and Burr, 1988)

allowing the heterogeneous elasticity of the bone to be

defined. In this study, the Young’s modulus, E, was

calculated using the equation E ¼ 0:06þ 0:9r2 derived

by Rice et al. (1988) when ro1.54 g/cm3 and E ¼

0:09r7:4 derived by Schaffler and Burr (1988) when

r41.54 g/cm3; r being the bone density. The calculated

Young’s moduli were rounded to obtain 35 distinct

values for use in the finite element model. The Poisson’s

ratio of the bone was taken to be 0.3. It may be possible

that some artefacts were present due to partial volume

effect at the surfaces. This could lead to an over-

estimated cortical volume with underestimated density.

Young’s moduli of the bone cement and polyethylene

(PE) were taken to be 2.2 and 0.5 GPa, respectively. The

Poisson’s ratio of bone cement and PE were taken to be

0.3 and 0.4, respectively. Model accuracy was analysed

by a mesh convergence of three different mesh densities:

33,990 degrees-of-freedom (dof), 51,309 dof, and 71,703dof.

Three-dimensional muscle and joint loads were

applied based on data reported in van der Helm (1991,

1994) and van der Helm and Pronk (1995). Muscle

insertion areas in this study were based on their

descriptions and also according to the description in

Gray’s anatomy (Williams, 1995). The muscles were

divided into six elements in order to take into account

the fact that some muscles have a large contact area.

This therefore allows the orientation of some muscles to

differ from just one direction, see Fig. 1(a). Magnitudes

of the muscle and joint forces ( f x, f y, f z) were input in alocal co-ordinate system, defined by the most dorsal

point on the acromioclavicular joint, the trigonum

spinae, and the angulus inferior, see van der Helm,

(1991).

Static muscle and joint load data for 901 to 1801 of

abduction were applied which included the maximum

joint reaction force of 406 N at 901 arm-abduction angle

(personal communication from Prof. Frans van der

Helm, Delft University of Technology, The Nether-

lands). Fig. 1(b) shows the full set of muscle loads which

were applied. Load sharing from the glenoid to the

acromion can therefore be accurately analysed. In this

ARTICLE IN PRESS

L.A. Murphy, P.J. Prendergast / Journal of Biomechanics 38 (2005) 1702–1711 1703





study we were interested in how the load is shared

between the acromion and the glenoid and in particularhow the glenohumeral joint load and the acromion

muscle loads are distributed. Fig. 2(a) isolates these

forces to schematically show the location and distribu-

tion of these acromion muscle forces acting in the finite

element model (note: for all load cases the full set of

muscles were used). The glenohumeral joint load was

modelled by applying a parabolic load based on a

diametrical mismatch of 6 mm, as recommended by

Iannotti et al. (1992); further details are given in Lacroix

et al. (2000).

For insertion of the acromion-fixated design, com-

bined cement and screw fixation was recommended

(Stryker Howmedica Osteonics, 1999; Gagey and

Mazas, 1990). This involved inclusion of a 1.5 mmcement mantle under the base-plate and a 0.5 mm

cement mantle around the glenoid peg. For screw

fixation of the baseplate, two cortical bone screws were

modelled with two further screws modelled to create

acromion-fixation. Regarding the acromion-fixation

screws, it is recommended that they be allowed to break

through the acromion bone and later trimmed to

prevent the sharp tips inserting into the surrounding

soft tissue (personal communication, Mr. R. Christie,

Stryker Howmedica Osteonics). Fig. 2(b) shows the FE

model of the scapula, cement mantle and acromion

prosthesis.

ARTICLE IN PRESS

Fig. 1. (a) Muscle force distribution of the trapezius muscle (after van der Helm, 1991) and (b) muscle loads applied to the finite element model of the

scapula at 901 arm abduction, schematic (data from F.C.T. van der Helm, personal communication).

Fig. 2. Various views of the FE model of the scapula showing (a) the muscles of interest in this study, the trapezius and deltoid, both acting on the

acromion and spine of the scapula (these and all the muscles shown in Fig. 1(b) were included in our model) and (b) the acromion-fixation component

inserted in the scapula.




Since the prosthesis is often implanted into rheuma-

toid arthritic bone, a prosthesis implanted into this

type of diseased joint was also analysed. For this case,

a Larsen Grade IV-type destruction (Kelly, 1994)

was reproduced and proximal subluxed loads

were applied associated with a deficient rotator

cuff for 0–1801

in flexion and abduction. Thesubluxed loads were simulated by moving the joint load

superiorly by 5 mm, as described in Murphy et al.

(2001).

2.2. Confirmation of the finite element model

The scapula used for our finite element model was

damaged in an attempt at validation using the strain

gauge technique. As a result samples of five left scapulae

were obtained and together with our original scapula

were deemed to be of average geometric and material

properties based on the parameters measured by Mallonet al. (1992). All five scapulae were coated with a

photoelastic coating (PL-8 liquid coating thick-

ness=1 mm, PL-1 adhesive thickness 0.1 mm), these

were embedded in bone cement along their medial

borders (Fig. 3). When loaded the fringe patterns were

viewed using a 030 series reflective polariscope (Vishay

Measurements Group UK Ltd., UK). Loading was

applied using the modular head of a humeral prosthesis,

see Fig. 3. Incremental loads were applied in steps of

0.2 kN to a maximum of 1.6 kN. Resulting strains were

then observed using the reflection polariscope and

measurements were taken using the null balance

compensator.

To enable comparison of like with like, a layer of

elements to represent the photoelastic material with

E =2.9 GPa and n=0.36 (Vishay, 2003) was added to

the surface of the scapula. The coating was assumed to

be bonded to the underlying bone. The finite element

model of the scapula was loaded and constrained in the

same way as the experiment, i.e. along the medial border

and included the joint surface. The glenoid was loaded

parabolically with the same load magnitude as in the

experiment.

3. Results

The plot of element material volume at each stresslevel for the three mesh densities shows that similar

results are achieved with each model (Fig. 4). Therefore,

a mesh density of 33,990dof was used for subsequent

analyses.

The average shear stress, average principal strain

difference and average principal strain direction values

were calculated over six areas (see Fig. 5(a)) which

correspond to regions of high stress in the FE model,

Fig 5(b); measurements were taken in the regions which

correspond to stress levels of approximately 10 MPa in

the FE model at a load of 1.6 kN, these results are listed

in Table 2.Cement stresses plotted for the entire mantle show

that they are generally low for both normal bone

(Fig. 6(a)) and rheumatoid arthritic bone (Fig. 6(b)).

Stresses produced in the metal were greatest in the

acromion arm, and were predicted to reach magnitudes

of 100 MPa, see tensile stress in Fig. 7(a) and

compressive stress in Fig. 7(b). From 601, 1201 and

1801 of abduction a twisting of the prosthesis arm

occurs. Deformation plots of the acromion base-plate

show from an inferior view that a shifting of the

prosthesis upwards at the acromion arm and down-

wards on the anterior face occur, see Fig. 7(c). From a

view above it is clear that the prosthesis is twistingslightly about the superior–inferior plane and also

shifting in an anterior–inferior direction.

Stress distributions in the bone were significantly

affected by the insertion of an acromion-fixation, see

Fig. 8. This may indicate stress transfer away from the

glenoid causing cement stresses to be lower there.

ARTICLE IN PRESS

Prosthetic humeralhead

Coated scapula

Medial borderrestrained by cement

PMMA filled aluminiumbox clamped to Instron

Fig. 3. Experimental set-up showing the scapula embedded in bone

cement and loaded for the purpose of confirming the FE model.

0

2

4

6

8

10

12

14

< 0 2-3 5-6 8-9 11-12 14-15 17-18 20-21 23-24 26-27 29-30 >32

Stress (MPa)

% o

f t o t a l c e m e n t v o l u m e

Mesh 1 = 33990 d.o.f

Mesh 2 = 51309 d.o.f.

Mesh 3 = 71703 d.o.f.

Fig. 4. The total volume at each stress level throughout the entire

mesh was generated for the three models with varying mesh densities.

Mesh 1 was used in this analysis.

L.A. Murphy, P.J. Prendergast / Journal of Biomechanics 38 (2005) 1702–17111706



4. Discussion

Glenoid component loosening has become a see-

mingly inevitable part of total shoulder arthroplasty. In

the attempt to find a well-performing glenoid compo-

nent, a plethora of designs have been introduced, some

of which, although ingenious (Table 1), have no

apparent biomechanical basis. A concept that has

frequently arisen in total shoulder arthroplasty is that

glenoid prostheses would derive greater durability if

fixated not only to the glenoid but to the acromion as

well (Gagey and Mazas, 1990), refer to Table 1. This

paper has focused on testing the hypothesis that

additional attachment of the glenoid component to theacromion increases the component’s potential durabil-

ity.

A finite element model, such as the one presented

here, is limited in its predictive value insofar as we

analyse one bone only. However, the dimensions of the

scapula analysed here fall well within those of typical

scapulae measured by Mallon et al. (1992); hence the

results are expected to be representative of scapulae

generally. Since there is such a large variability in the

geometric pathology of the rheumatoid glenohumeral

joint, and in the rate of progression of the disease (Kelly,

1994), simulation of a shape changed rheumatoid

arthritis glenohumeral joint was beyond the scope of

this study. A general approach to modelling thedestructive effects of rheumatoid arthritis was carried

out based on the experimental work of Frich (1994a) (to

date, Frich’s study is the only study which has data for

rheumatoid arthritis glenoid bone material properties).

This consisted of simulating a Larsen Grade IV-type

destruction with a subluxed joint load.

Shoulder prostheses fixation depends to a great extent

on very small details (localised contacts between a

narrow rim and a thin layer of cortical bone, the

contacts of the screw threads etc.). These details were

included in the model for a general case but cannot

predict the huge variability that occurs from patient topatient. The present work is only concerned with an

analysis of glenoid replacement in the direct post-

operative situation; no account was taken of soft tissue

formation on interfaces or postoperative remodelling of

the bone.

Similarity of strain data obtained from experimental

and FE models is necessary to confirm that valid results

are being realised. The finite element model is adequate

for the question posed; the photoelastic technique has a

full-field capability, which allows observation and

measurement of strain directions and magnitudes for

complex geometries, under varying complex loading

ARTICLE IN PRESS

Fig. 5. (a) FE model of the scapula showing the areas (1–6, which correspond to regions of high stress in the FE model (b)) where average shear

stress, average principal strain difference and principal strain direction values were calculated for both the experimental and FE scapulae.




modes, regardless of material homogeneity. Point

measurements were also calculated over an area and

the average value was then compared to the same

regions in the FE mesh; this was carried out for six

regions. Agreement of averaged strains was within 20%.

A large error occurred on the costal region for area two,

see Table 2, showing a difference of 301

, this may be dueto some bubble formation or overstressing of the

coating during the contouring procedure. On inspection

some slight bubble formation occurred on scapulae 4

and 5 at area 2. Relative similarity of the photoelastic

and finite element results confirms that the finite element

model is a sufficiently good description of the standard

behaviour of a real scapula for the purpose of analysing

prosthesis performance. Errors associated with the

photoelastic technique include Poisson’s coefficient

mismatch, incorrect light incidence, uneven coating

thickness and, as reported by Cristofolini et al., (1994),

the reinforcing effect of the photoelastic coating.

However, each of these errors, except for uneven coating

thickness and incorrect light incidence, is addressed in

this paper by also applying the coating in the FE model.

The difficulty in creating a homogenous coating thick-

ness experimentally was experienced. However, by

measuring each coating with a micrometer prior to

adhesion, associated errors could be minimised.

Furthermore, the scapulae used were different in size

and material property distribution from the one

modelled in the FE analysis. However, we tried to

eliminate this problem by testing five scapulae and

comparing to our FE model. Point-to-point agreement

cannot be expected as the experimental specimen used tocreate the FE model was not available for testing. To

compare the cancellous bone material properties of our

scapula with others, the glenoid was divided into four

volumes: superior–anterior, superior–posterior, inter-

ior–anterior, and inferior–posterior. In each volume, the

volume fraction for each distinct Young’s modulus

value was calculated and then the weighted–average

Young’s modulus was calculated. The calculated aver-

age Young’s modulus for the superior–anterior, super-

ior–posterior, inferior–anterior, and inferior–posterior

volumes were 0.264, 0.321, 0.231, and 0.215 GPa,

respectively which is within the normal range. Thedensity was highest in the supero-posterior part of the

glenoid in accordance with other studies, see Lacroix et

al. (2000).

Complete validation of a model can never be achieved

perfectly since numerical models are only a representa-

tion of reality, not reality itself. However, the closer the

correspondence between the model and experiment the

more confidence we have in the model. In this study the

FE model of the intact scapula, fully constrained along

its medial edge and loaded by a single force is used to

provide an experimental confirmation for the implanted

bone, which is loaded in a more complex way via the

ARTICLE IN PRESS

T a b l e 2

P h o t o e l a s t i c i t y a n d F E m o d e l c o m p a r i s o n

s o f m a x i m u m

s h e a r s t r e s s , p r i n c i p a l s t r a i n d i f f e r e n c e ,

d i r e c t i o n o f p r i n c i p a l s t r a i n v a l u e s c a l c u l a t e d a t a r e a s o f h i g h s t r e s s ( t h e s e v a

l u e s a r e a v e r a g e d o v e r

a r e a s 1 – 6 , s e e F i g .

5 ( a ) )

S u r f a c e A r e a

E x p .

1

E

x p .

2

E x p .

3

E x p .

4

E x p .

5

A v e r a g e o f 5 s c a p u l a e

F E s c a p u l a

t m a x ( M P a ) e 1 - e 2

e X

( 1 ) t

m a x ( M P a ) e 1 - e 2

e X

( 1 ) t m a x ( M P a ) e 1 - e 2

e X

( 1 ) t m a x ( M P a ) e 1 - e 2

e X

( 1 ) t m a x ( M P a )

e 1 - e 2

e X

( 1 ) t m a x ( M P a )

e 1 - e 2

e X

( 1 ) t m

a x ( M P a ) e 1 - e 2

e X

( 1 )

C o s t a l

1

5 . 2

5

0 . 0

0 3 1

4 0

9

. 4 8

0 . 0

0 5 6

4 5

8 . 6

3

0 . 0

0 5 1

4 0

7 . 9

2

0 . 0

0 4 7

5 0

9 . 1

4

0 . 0

0 5 4

5 0

8 . 0

8

0 . 0

0 4 8

4 5

1 0

. 1 5

0 . 0

0 6 0

4 5

2

9 . 4

2

0 . 0

0 5 6

7 0

7

. 1 1

0 . 0

0 4 2

6 5

9 . 1

2

0 . 0

0 5 4

7 0

8 . 4

6

0 . 0

0 5 0

2 5

9 . 3

9

0 . 0

0 5 6

2 0

8 . 7

0

0 . 0

0 5 2

5 0

9

. 8 3

0 . 0

0 5 8

8 0

3

4 . 2

3

0 . 0

0 2 5

1 0

5

. 8 3

0 . 0

0 3 5

2 0

3 . 8

9

0 . 0

0 2 3

2 0

3 . 8

9

0 . 0

0 2 3

1 5

6 . 2

6

0 . 0

0 3 7

1 0

4 . 8

2

0 . 0

0 2 9

1 5

5

. 7 4

0 . 0

0 2 8

1 5

D o r s a l

4

7 . 3

2

0 . 0

0 4 3

5 5

6

. 0 9

0 . 0

0 3 6

5 5

7 . 2

8

0 . 0

0 4 3

5 0

7 . 7

8

0 . 0

0 4 6

4 5

6 . 5

9

0 . 0

0 3 9

3 0

7 . 0

1

0 . 0

0 4 1

4 7

6

. 6 9

0 . 0

0 4 0

4 5

5

8 . 5

1

0 . 0

0 5 0

5 0

7

. 9 5

0 . 0

0 4 7

3 0

9 . 1

1

0 . 0

0 5 4

3 5

7 . 4

5

0 . 0

0 4 4

3 0

8 . 6

3

0 . 0

0 5 1

3 0

8 . 3

3

0 . 0

0 4 9

3 5

8

. 5 6

0 . 0

0 5 0

3 5

6

8 . 1

2

0 . 0

0 4 8

8 5

8

. 8 0

0 . 0

0 5 2

7 5

8 . 8

4

0 . 0

0 5 2

7 0

7 . 4

5

0 . 0

0 4 4

7 5

*

*

*

8 . 3

0

0 . 0

0 4 9

7 6 . 2

5 7

. 5 9

0 . 0

0 4 5

8 5

t m a x

i s t h e a v e r a g e m a x i m u m

s h e a r s t r e s s

o v e r a r e a s p e c i fi e d .

e 1 - e 2 i s t h e a v e r a g e p r i n

c i p a l s t r a i n d i f f e r e n c e o v e r a r e a s p e c i fi e d .

e

X

i s t h e a v e r a g e p r i n c i p a l s t r a i n d i r e c t i o n o v

e r a r e a s p e c i fi e d .

* N o

m e a s u r e m e n t s a t t h i s p o i n t d u e t o c o a t i n g

d a m a g e .




muscle and joint forces. We confirmed our model using

the joint load only because the application of muscle

loads by straps (as done in Britton et al., 2003) would

have covered a large area of the scapular surface limiting

the area over which photoelastic coating could be

viewed.

The relative motion between the acromion and the

glenoid of the unimplanted scapula is affected by

insertion of the acromion-fixated glenoid prosthesis. If

the natural relative motion is large then it would be

resisted by an implanted prosthesis and this may be the

reason why the acromion prosthesis arm is highly

stressed, see Fig. 7(a). The finite element analysis shows

that the relative displacement between the glenoid and

the acromion under the action of the muscle loads is

significant at approximately 1 mm in the absence of

acromion-fixation, see Fig. 9. At the higher magnitude

loads (301, 601 and 901 of abduction) the highest relative

ARTICLE IN PRESS

Fig. 6. Cement maximum principal stresses for (a) normal bone and (b) rheumatoid arthritis bone — a comparison of the acromion design at

30–1801 of abduction.

Fig. 7. (a) Maximum (0–60 MPa) and (b) minimum (0–60 MPa) principal stress plots for the acromion base-plate for 601 of abduction in normal

bone. (c) plots the deformation of the base-plate for 601, 1201 and 1801 of abduction glenohumeral joint load (i–vi), magnification 15.




motion occurs (0.8–1.2 mm). This motion is resisted by

the acromion device, which causes high stresses in the

acromion arm and a corresponding unphysiological

stressing of the bone. At 1201 and 1501 of abduction,

with no acromion-fixation, the relative motion is smaller

for the natural scapula than for the case with acromion- fixation (Fig. 9). This is an unexpected result and may be

explained as follows: Relative motion between the

glenoid and acromion is calculated as a sum of the

motion in the x, y and z directions. With acromion-

fixation, load sharing occurs between the glenoid and

acromion. It is the deformations of the glenoid base-

plate, and in particular the acromion, which results in

the high stresses seen in Fig. 7(a). Clearly, twisting of the

acromion arm is occurring as a result of the deforma-

tions of the prosthesis (see Fig. 7 (c)), producing a

resulting motion of the glenoid which is not in the same

direction as the acromion, thus giving the relative

motion results shown in Fig. 9. Furthermore, because

of the rigid attachment of the prosthesis acromion arm,

a similar magnitude of relative motion occurs at each

angle of abduction, which is clearly not evident for the

natural case. This proves that the acromion-fixation

forces a non-physiological stress state and loading

pattern at the glenohumeral joint.Another consequence of acromion-fixation is that

stresses produced in the entire scapula are altered

compared to the case of reconstruction without acro-

mion-fixation. Significant global bone stress changes

were not found with non-acromion-fixated designs

(Lacroix et al., 2000; Murphy et al., 2001). It is a usual

design requirement for prostheses to reproduce a natural

distribution of stresses (Prendergast, 2001; Murphy,

2002); however, for the acromion design the acromion

bone stresses show different distributions and are higher

than the case with no prosthesis fixation (16MPa

higher).

Gagey and Mazas (1990) also reported on some

preliminary clinical trials with this type of prosthesis.

Problems were reported to occur due to overstressing of

the acromion arm resulting in metal fatigue and failure

of the device. In subsequent designs the acromion arm

was reinforced by thickening; however, this study

suggests that even the thickened acromion arm analysed

in this study may not reduce the stresses sufficiently. In

conclusion, attachment of the glenoid prosthetic com-

ponent to the acromion, with the idea of sharing the

load away from the overstressed glenoid, may appear to

be simple and beneficial. However, any of the joint

reaction force shared out to the acromion is alsocoupled with a muscle load (152 N: deltoid muscle,

116 N: Trapezius muscle) shared out from the acromion

to the glenoid. This, it would seem, creates no advantage

for the fixation and a significant disadvantage of highly

stressing the component itself.

Acknowledgements

Stryker Howmedica Osteonics, Raheen Business

Park, Raheen, Limerick and Enterprise Ireland pro-

vided financial support for this work. Ir Peter Nuijtenand Mr Robert Christie are thanked for their advice.

References

American Academy of Orthopaedic Surgeons, 2003. Arthroplasty and

total joint replacement procedures 1990 to 1999. www.Aaos.org/

wordhtml/research/arthrop.htm.

Apoil, A., Koechlin, P.H., Augereau, B., Hongier, J., 1983. Prosthese

totale d’e ´ paule a ` appui acromio-coracoı ¨dien prosthe ` se de re ´ section

hume ´ rale. Acta Orthopaedica Belgica 49 (5), 571–578.

Britton, J.R., Walsh, L.A., Prendergast, P.J., 2003. Mechanical

simulation of muscle loading on the proximal femur, analysis of

ARTICLE IN PRESS

Fig. 8. Maximum principal stress plots in (a) the natural scapula (no

prosthesis inserted) and (b) with the acromion prosthesis inserted.

Relative motion between the acromion andglenoid

0

0.2

0.4

0.6

0.8

1

1.2

1.4

30 60 90 120 150 180Angle of abduction

R e l a t i v e m o t i o n ( m m )

Normal bone- Acromion design

RA bone- Acromion design

Normal bone- No acromion fixation

Fig. 9. Plot of the relative motion between the glenoid and acromion

of the scapula for no acromion-fixation and acromion-fixation, 30–180

degrees of abduction.


http://www.aaos.org/wordhtml/research/arthrop.htm







cemented femoral component migration with and without muscle

loading. Clinical Biomechanics 18, 637–646.

Copeland, S., 1990. Cementless total shoulder replacement. In: Post,

M., Morrey, B.F., Hawkins, R.J. (Eds.), Surgery of the Shoulder.

Mosby Yearbook, St Louis, pp. 289–293.

Cristofolini, L., Cappello, A., Toni, A., 1994. Experimental errors in

the application of photoelastic coatings on human femurs with

uncemented hip stems. Strain 30, 95–103.Frich, L.H., 1994a. Strength and structure of glenoidal bone. Ph.D.

thesis, A ˚ rhus University, A ˚ rhus, Denmark.

Frich, L.H., 1994. Architectural and mechanical properties of the

glenoid in the normal state and the rheumatoid arthritic shoulder.

Danish Medical Bulletin 41 (4), 446–447.

Frich, L.H., Jensen, N.C., Odgaard, A., Pedersen, C.M., Soıjbjerg,

J.O., Dalstra, M., 1997. Bone Strength and Material Properties of

the Glenoid. Journal of Shoulder and Elbow Surgery 6, 97–104.

Gagey, O., Mazas, F., 1990. A new total shoulder prosthesis with

acromion-fixation. In: Post, M., Morrey, B.F., Hawkins, R.J.

(Eds.), Surgery of the Shoulder. Mosby Yearbook, St Louis,

pp. 282–284.

van der Helm, F.C.T., 1991. The shoulder mechanism, a dynamic

approach. Ph.D. Thesis, Delft University Press, Delft.

van der Helm, F.C.T., 1994. A finite element musculoskeletal model of the shoulder mechanism. Journal of Biomechanics 27 (5), 551–569.

van der Helm, F.C.T., Pronk, G.M., 1995. Three-dimensional

recording and description of motions of the shoulder mechanism.

Journal of Biomechanical Engineering 117, 27–40.

Hvid, I., Bentzen, S.M., Linde, F., Mosekilde, L., Pongsoitpetch, B.,

1989. X-ray Quantitative Computed Tomography: The Relations

to Physical Properties of Proximal Tibial Trabecular Bone Speci-

mens. Journal of Biomechanics 22, 837–844.

Iannotti, J.P., Gabriel, J.P., Schneck, S.L., Evans, B.G., Misra, S.,

1992. The normal glenohumeral relationships- and anatomical

study of 140 shoulders. Journal of Bone and Joint Surgery 74A,

491–500.

Kelly, I.G., 1994. Unconstrained shoulder arthroplasty in rheumatoid

arthritis. Clinical Orthopaedics and Related Research 307, 94–102.

Lacroix, D., Prendergast, P.J., Murray, R., McAlinden, S., D’Arcy, E.,1997. The use of quantitative computed tomography to generate a

finite element model of the scapula bone. In: Monaghan, J., Lyons,

C.G. (Eds.), Proceedings of IMC 14. Sustainable Technologies in

Manufacturing Industries, pp 257–262.

Lacroix, D., Murphy, L.A., Prendergast, P.J., 2000. Three-dimensional

finite element analysis of glenoid replacement prostheses: a

comparison of keeled and pegged anchorage systems. Journal of

Biomechanical Engineering 122 (4), 430–436.

Laurence, M., 1991. Replacement arthroplasty of the rotator cuff

deficient shoulder. Journal of Bone and Joint Surgery 73B,

916–919.

Limb, D., 2002. Shoulder replacement—current problems. Current

Orthopaedics 16, 15–20.

Mallon, W.J., Brown, H.R., Volger III, J.B., Martinez, S., 1992.

Radiographic and geometric anatomy of the scapula. Clinical

Orthopaedics and Related Research 277, 142–154.Mansat, P., Barea, C., Hobatho, M.C., Darmana, R., Mansat, M.,

1998. Anatomic Variation of the Mechanical Properties of the

Glenoid. Journal of Shoulder and Elbow Surgery 7, 109–115.

Mazas, F., de la Caffiniere, J.Y., 1977. Une nouvelle prosthe ` se totale

d’e ´ paule. Revue de chirurgie Orthopedique et Reparatrice de l’

appareil moteur 63, 113–115.

Mazas, F., de la Caffiniere, J.Y., 1982. Total shoulder replacement by

an unconstrained prosthesis. Report of 38 cases. Revue de chirurgie

Orthopedique et Reparatrice de l’ appareil moteur 68, 161–170.

McElwain, J.P., English, E., 1987. The early results of porous-coated

total shoulder arthroplasty. Clinical Orthopaedics and Related

Research 218, 217–224.

Murphy, L.A., 2002. Hypothesis testing for glenoid replacement in

total shoulder arthroplasty. PhD Thesis, University of Dublin.

Murphy, L.A., Prendergast, P.J., Resch, H., 2001. Structural analysisof an offset-keel design glenoid component compared to a

centre keel design. Journal of Shoulder and Elbow Surgery 10,

568–579.

Neer, C.S.II., 1990. Shoulder reconstruction. Saunders, Philadelphia,

pp. 143–271.

Prendergast, P.J., 2001, Bone Prostheses and Implants. In: Cowin, S.C.

(Ed.), Bone Mechanics Handbook, second ed. (Chapter 35).

Redfern, T.R., Wallace, W.A., 1998. History of shoulder replacement

surgery. In: Wallace, W.A. (Ed.), Joint Replacement in the

Shoulder and Elbow. Butterworth and Heinmann, Oxford,

pp. 6–16.

Rice, J.C., Cowin, S.C., Bowman, J.A., 1988. On the Dependence of

the Elasticity and Strength of Cancellous Bone on Apparent

Density. Journal Biomechanics 21, 155–168.

Schaffler, M.B., Burr, D.B., 1988. Stiffness of Compact Bone: Effectsof Porosity and Density. Journal of Biomechanics 21, 13–16.

Skirving, A.P., 1999. Total shoulder arthroplasty – current problems

and possible solutions. Journal of Orthopaedic Science 4, 42–53.

Stryker Howmedica Osteonics, 1999. The Howmedica Anatomic

Shoulder, Anatomic Adaptation booklet.

Vishay, 2003. Coating materials and adhesives. www.vishay.com/

Williams, P.L., 1995. Gray’s Anatomy. 38th ed. Churchill Livingstone,

Inc., pp. 615-634.

ARTICLE IN PRESS


http://www.vishay.com/

http://www.vishay.com/

Documents

Acromion-ﬁxation of glenoidcomponents in total shoulder arthroplasty