80

Technical Notes in BPR usually originate when a progress report goes beyondthe usual form to present a wealth of detail about some single aspect of thesubject under investigation . In other cases, a progress report may simply requirea more generous physical format to do justice to its content, in the opinion of theEditor . BPR Technical Notes are carefully studied by the publication's staff, butare not formally refereed by reviewers from BPR's Editorial Board or ad hocreviewers, as are the papers on the preceding pages.

TECHNICAL NOTES

In Vitro Evaluation of the Effect of Acetabular Pros-thesis Implantation on Human Cadaver Pelves

WILLIAM PETTY, M .DAssociate Professor, Department of OrthopaedicsUniversity of Florida College of MedicineCo-Chief, Orthopaedic SurgeryVeterans Administration HospitalGainesville, Florida 32602

GARY J . MILLER, Ph . D.Assistant Professor, Department of OrthopaedicsUniversity of Florida College of MedicineGainesville, Florida 32610

GEORGE PIOTROWSKI, Ph . D.Associate ProfessorDepartment of Mechanical EngineeringUniversity of Florida College of EngineeringGainesville, Florida 32610

INTRODUCTION

During the past 15 years, total hip arthroplasty, asdeveloped by Charnley, McKee, Watson-Farrar, and others,has revolutionized treatment of hip disease . The procedureis the most successful one ever developed for the treatmentof the arthritic hip, and most patients achieve excellentresults . However, a small percentage of patients are not sofortunate and develop a complication severe enough tocause failure of the total hip arthroplasty . Failure may occurdue to several causes, and these can basically be brokendown into three general groups, including (i) biologicalfailures, the most prominent of which is infection, (ii)technical failure resulting from such problems as poorpreparation of the bony beds, poor cementing technique,and poor placement of the prosthetic component, and (iii)mechanical failure of the prostheses, including loosening,migration, and breakage.

Early in the evolution of total hip arthroplasty, greatattention was paid to the complications of infection andwear . Infection remains a major problem because of itssevere consequences, but its incidence has been reducedby various measures, including improving the operativeenvironment and the use of antibiotics . In some circum-stances, wear of the components has been a problem ; but,in general that problem has been a minor one, especially inthe majority of total hip arthroplasties in which metal andpolyethylene components are used.

The most common cause of failure in total hip arthro-plasty is loosening . Loosening has received the greatestattention in the femur because of the relatively earlyrecognition of this problem . There has been less concernabout loosening of the acetabular component, but as

duration of followup becomes longer, it is becomingapparent that failure of fixation on the socket side of totalhip arthroplasty may become at least as common as, if notmore common than, failure on the femoral side . Forexample,Muller and co-workers have reported that in 81patients followed for 10 years or longer, 16 hips (20percent) required re-operation, most commonly for a looseacetabular component . In Charnley's series, there has beena steady increase in the incidence of acetabular looseningas followup times have become longer . Review of 141 hipswith an average followup of 10 .1 years revealed that 70percent had roentgenographic demarcation between thebone cement and bone on the socket side . Thirteen percentof the 70 percent of sockets had actually migrated, and 9percent of the total groups of patients demonstratedmigration of the acetabular component after 10 years . Mostof these patients had symptoms.

The only factor that could be correlated with increasedlikelihood of acetabular failure was that it appeared to bemore common in those patients with rheumatoid arthritis.Beckenbaugh and others reported in their series, after 4 to7 years of followup, that 99 percent of their patients hadroentgenographic demarcation between the cement andbone of the acetabulum.

Implantation of the prosthetic acetabulum involves ream-ing of the bony acetabulum, and socket fixation with bonecement . Several details of socket implantation technique(such as depth of reaming, use of a pilot hole, use ofanchoring holes, and methods for reinforcing a weakenedbony acetabulum) remain controversial.

The experiments reported here were performed toevaluate quantitatively the effects of these techniques onthe stress and strain pattern of human cadaver pelves.

MATERIALS AND METHODS

Twelve pairs of cadaveric hemipelves were acquired bytranssection of the sacroiliac joint and the pubic symphysisand cleaned of soft tissue . The mid-portion of the arcuateline of the ilium was degreased with Chlorothene-NU a andcarefully sanded with 400 grit sandpaper . Three strain gagerosettes (EA-13-062RB-120) a were installed, using stand-ard installation techniques, in positions as illustrated inFigure 1 . Each of the three gages of a rosette measuredstrain in a different direction, and these three independentmeasurements were used to compute the largest tensile,compressive, and shear strains in the bony material at thesite of the rosette, using the Mohr's circle transformationtechnique.

The loading of each hemipelvis was produced by a fixture(Figs . 2 and 3) designed to simulate conditions of single-legstance phase in gait . The sacroiliac joint of each specimenwas encased in a block of aluminum-filled epoxy b to

81

Bulletin of Prosthetics Research BPR 10-33 (Vol . 17 No. 1) Spring 1980

FIGURE 1.Paired Hemipelves each instrumented withthree strain gage rosettes, here labelled A,B, and C . The sacroiliac joints are encased

in blocks of aluminum-filled epoxy .

FIGURE 2Lateral view of hemipelvis (HP) bolted,through the epoxy block (EB), to thepositioning fixture (PF) . A Charnley femoralhead prosthesis (FP) is mounted on theloading bar (LB), along with cables (C) andhooks (H) which simulate the abductormuscle pull . The load cell (LC) monitors thesimulated ground reaction, applied to theother end of the loading bar.

FIGURE 3.Another view of the hemipelvis mounted inthe loading fixture . (See caption of Figure 2for identification of parts .)

82

TECHNICAL NOTE : PETTY et at.

provide a convenient means for attaching the bone to thepositioning fixture . This fixture permitted the bone to beoriented in the position it assumes in erect standing, andrigidly held in that position . The loading bar connects to thehemipelvis through a femoral head prosthesis mounted onthe loading bar, and by cables attached to the iliac crest byfour small hooks . These hooks and cables simulate the pullof the abductor muscles of the hip . The opposite end of theloading bar is connected to a load cell which measures theupward force applied to the loading bar . This forcesimulates the ground reaction force on the standing leg.

The hemipelvis on the positioning fixture, the load cell,and the loading bar are mounted in a universal testingmachine in such a way that the load cell pulls up on theloading bar, thereby producing the joint reaction force andthe pull of the musculature simultaneously.

Each hemipelvis was mounted in the positioning fixtureand carefully oriented . The positioning fixture was locked inplace by tightening its clamps, and the loading barinstalled . Two photographs of the test set-up, representingA-P and lateral views, were made and used to determinegraphically the ratio between the applied load, as measuredby the load cell, and the joint force.

The pubic symphysis was left unloaded by this loadingfixture . This was deemed to be a reasonable approximation,since individuals with disrupted symphyses have beenknown to heal with a gap between the pubes which hasnot compromised their ability to walk normally.

The instrumentation (Fig .4) was designed to minimize thenumber of connections to be made in connecting eachbone's nine strain gages . Only 10 conductors (one per gageplus power common), rather than 18, are required.Corresponding gages from paired bones are connected ashalf of a Wheatstone bridge; this insures automatictemperature compensation . The connections between thegage pairs are made within the calibration box which alsocontains the bridge completion resistors, trimming resistorswhich allow the output from each gage to be nulled at thestart of each test run, and a calibration resistor . Thecalibration resistor electrically simulates a fixed value of

LEFT BONE

CAL . BOX

>

strain to the amplifying and recording equipment ; switchS1 can be used to select the sign of the calibration signal,and switch S2 allows the calibration signal to be generatedfor each gage in turn . The strains produced by each gage,as well as the force measured by the load cell, wererecorded simultaneously on an oscillograph.

After each test, the photograph of the test set-up wasused to determine graphically the ratio of joint force toapplied load . At each of 15 positions along the oscillographtraces, values for the load cell and strain gage readingswere read and input to a computer program whichperformed a linear regression of strain reading versus jointforce for each gage reading . A typical graph of gage readingas a function of joint force for three gages from one bone isshown in Figure 5 and illustrates the linear relationshipobserved . Thus, all data are presented as strain per unitjoint force.

The computer program further calculated the principal(maximum tensile and compressive) strains and theirorientations for each rosette, as well as the maximumshear strain . These three strain values represent thelargest deformations to which the material under therosette was being subjected.

Two sets of data were obtained for each hemipelvis . Thefirst was obtained with the acetabulum in its natural state(i .e ., without an acetabular prosthesis), using a properlyfitted Austin Moore prosthesis to apply the joint force . Thesecond data set was obtained using an acetabularprosthesis and one of eight variations in the basicinstallation technique (Table 1).

a Micromeasurements Inc ., Romulus, Michigan

b Devcon Corp ., Danvers, Massachusetts

RESULTS

The "Normal" Pattern

A total of 24 hemipelves were tested with an Austin Mooreprosthesis against an acetabulum without implant . The straindata are summarized in Figure 6, which includes the nine

strain gage readings, the computed principal (maximum tensileand compressive) strains, and the maximum shear strain.Thus, Figure 6 represents the "normal" pattern of strain foundalong the arcuate line of the ilium in single leg stance . Eachrosette yields a similar pattern of strain : the first and second

FIGURE 4.Schematic diagram of the instrumentationused to balance and calibrate the straingage outputs . The strain gages arerepresented by L, - Lo ; D, and T, are bridgecompletion and trimming resistors, whichare replicated for each of the nine straingage pairs . The polarity of the calibrationsignal generated by R, can be selectedwith switch S,, while Sz determines whichstrain gage is being calibrated.

RIGHT BONE

R9'

) }> o+V

D :>

> D i

OF F, 0 o?-a . . . .o

2OFF

To-RECORDER

AMP.

83

Bulletin of Prosthetics Research BPR 10-33 (Vol . 17 No . 1) Spring 1980

TABLE 1.

Types of Implantations Performed

Set aSide

Installation Technique

V

L

No center hole, removal of cartilage only

V

R

10 mm center hole, removal of cartilage only

VI

L

20 mm center hole, reamed to pelvic cortex : less than 1 mm of bone at hole edge

VI

R

Same as VI L, but with addition of protrusio ring

X

L

10 mm hole, reamed to 4 mm

X

R

Same as X L, but with three anchoring holes

XI

L

30-mm hole, reamed to pelvic cortex : less than 1 mm of bone at hole edge

XI

R

Same as X( L, but with protrusio ring

a Each set contains 3 pelves, with the left and right pelves treated differently.

-1500

-1000

-500

FIGURE 5.A representative graph of gage readingsas a function of joint force . Each linerepresents the best least-squares fit tothe data points shown, with the bandsrepresenting ± one standard deviationindicated.

JOINT FORCE(N)

7525

50

Gage 2

-500

gages of each rosette exhibit moderate compressive strains,while the third gage (situated posteriorly but oriented almostvertically) shows a tensile strain . The two principal strains foreach rosette are nearly equal in magnitude but of oppositesign, with the highest strain found at rosette C . The drawing(Fig . 7) illustrates the orientation of the principal strains, aswell as their relative magnitudes . The difference between theprincipal strains equals the maximum shear strain experiencdby the material under a rosette ; the "normal" values for maxi-mum shear strain are also shown in Figure 6 . Calculationsindicate that the average normal strains at these locations arealmost zero. This indicates that the bony material in the regionis undergoing almost pure shear.

Side-to-side differences for all measured and computedstrain variables were calculated, and are summarized onFigure 8 . In all cases, the average value of the right-leftdifferences was within two standard errors of zero, indicatingno left-right bias in pelves without acetabular cups . It is alsointeresting to note that the scatter in the left-right differences

is relatively uniform for all nine strain gages comprisingrosettes A, B and C, as are their principal strains ; but thescatter in the maximum shear strains is larger.

Effects of Various Installation Techniques c

1 . NO CENTRAL HOLEThree acetabular implants were installed with no central

guide and with no bone removal . (i .e ., curetting the cartilageonly) . Only the principal strains and maximum shear strainsare presented here, since they are insensitive to gageorientation, while the individual gage readings are not . Figure10a summarizes the principal and maximum shear strain data

obtained for these three specimens along with the strainsobtained prior to implantation . Also shown is the "normal"strain pattern for unimplanted pelves, from Figure 6 . Figure10b shows the difference in strain readings obtained beforeand after installation of the prosthesis (difference =value after— value before) . All changes are very small compared to theindividual gage readings, indicating that if the bone of the

84

'TECHNICAL NOTE : PETTY et al.

pelvis is not disturbed while applying an acetabular cup, anessentially unchanged pattern of strain results after implanta-tion of an acetabular prosthesis.

2. EFFECTS OF A CENTRAL HOLE AND REAMINGThe effects of the presence of a central hole drilled during

the installation of the prosthesis is depicted in Figures 11a and11b through 14a and 14b . A 10 mm central hole with noaccompanying reaming of the acetabulum (Group V R) causesonly a slight increase in the tensile strains for rosettes A and B,with a modest increase in the shear strain at rosette A (Figs.1 1 a and 11 b) . Rosette C shows little change.

Drilling a 10-mm central hole and reaming the bone to a4-mm thickness (Group X L) increases the strainexperienced by the bone (Fig . 12a and 12b). While thestrain data from this group conformed to the "normal"pattern before implantation, all principal strains undergo anapproximately equal increase of about Y2 1st/N (or Y2microstrain per Newton) in the magnitude of the strainpost-implantation ; consequently the maximum shear strainincreases by roughly twice as much, on the average.

Enlarging the hole to 20 mm and reaming to the pelviccortex (less than 1 mm of bone at hole edges, Group VI L)yields a similar increase in the tensile principal strains andmaximum shear strain . Figures 13a and 13b illustrate thatthe post-implantation strain pattern clearly shows largerstrains, with rosette A located more anteriorly, showinggreater increases in maximum tensile strain than theposterior rosette C . The large increases for one of thespecimens were not echoed by the other two specimens,especially posteriorly, and thus their significance isquestionable. The data suggest a shift in the loaddistribution within the bone, with the antero-medialmaterial (rosette A) experiencing increased strain.

Curiously, when there was more extensive removal (orloss) of acetabular bone, the above pattern of uniformlyincreased strain was reversed . This condition wassimulated by drilling a 30-mm central hole and reaming theacetabulum to 1 mm of bone remaining at the hole edge(Group XI L) . Figures 14a and 14b illustrate that rosette Ais relatively unaffected, the effect on rosette B isinconsistent, while rosette C exhibits a decrease in themagnitude of the principal strains as well as the maximumshear strain . As with Group X L 10-mm central hole andreaming to 4-mm thickness (Fig . 12a and 12b), the strainunder rosette A increased relative to the strain in rosette C,suggesting an antero-medial shift in the load distributionwithin the bone . This antero-medial shift was similar tothat found in Group VI L (Figs . 13a and 13b).

3. REINFORCEMENT WITH A PROTRUSIO RING

The bones of one set (Group VI R) were first perforatedwith a 20-mm center hole, similar to the bones representedby Group VI L (Figs. 13a and 13b)—but with the acetabularcup installation reinforced with a protrusio ring . d AsFigures 15a and 15b show, the maximum tensile strain forall rosettes increased equally from pre-installation values.As in the case without the reinforcement (Group VI L), the

strains increase, but the use of the protrusio ring inhibitsthe anterior shift of strains seen in the previous preparationwithout reinforcement . The maximum compressive strainremained essentially unchanged for all three rosettes.

Bones of another group of specimens (Group XI R, Figs.16a and 16b) were drilled with a 30 mm hole and reamedto simulate extensive bone loss (similar to the over-reamedbut unreinforced bones represented by Group XI L, Figs.14a and 14b) but with the additional installation of aprotrusio ring for reinforcement . Tensile strains for rosettesA and B increased, concurrent with an increase in themaximum shear strain . Rosette C, on the otherhand,experienced decreases in the maximum tensile strainand the maximum shear strain observed . This pattern isquite similar to that shown in the unreinforced case . (Thedata in both of these cases exhibit much scatter, makingthem difficult to interpret, however .)

4. EFFECT OF ANCHORING HOLES

In one group of three hemipelves (Group X R) eachreceived three anchoring holes, which are intended toprovide greater stability to the fixation of the acetabularprosthesis . As Figures 17a and 17b illustrate, the principalstrains and the shear strains all increased in magnitude,with the most consistent effect being a substantial increasein the shear strain at rosette B . The results for this groupare quite similar to the results for Group X L (Figs . 12a and12b) which had similar preparation but without anchoringholes . (In the case of one bone, the shear strain at rosetteA nearly doubled; this bone will be sectioned in order tostudy the spatial relationship between the anchoring holesand the strain gages .)

DISCUSSION

The strain pattern found in hemipelves without anacetabular prosthesis was quite consistent from specimento specimen . All three rosettes are in almost pure shear,and there is a monotonic increase in strain values fromrosette A to rosette C.

This strain pattern appears to remain unaltered by theinstallation of an acetabular prosthesis—if the acetabulumis not reamed or perforated by a central hole . Thus, thepresence of the prosthesis appears to be less significantthan the disruption of the bony structures during itsinstallation . A 10-mm central hole has a relatively smalleffect on the strain pattern, but the increase in tensile andshear strain for rosette A strongly suggests a shift in theload distribution ; the bone situated anteriorly (and medially)to the acetabulum is strained more . This shift is much morepronounced for bones with 20-mm central holes andfurther reaming (in 2 out of 3 cases).

C Figure 9 describes the key used in Figures 10-17.

d DePuy Inc ., Warsaw, Indiana

85

Bulletin of Prosthetics Research BPR 10-33 (Vol . 17 No. 1) Spring 1980

pst/N

S .D. pst/N

pst/N

4

5

3

2

GageReadings

5KEY :

+2S .E.AVE.2 S.E .

0

4

3

2

4PrincipalStrains

3

2

— 0

5

-

-2A

-2B

-2A B C

Rosette

Rosette

FIGURE S.Summary of strain data (strain per unitjoint force) obtained from 24 hemipelvesloaded through an Austin Moore prosthesiswithout acetabular implant.

FIGURE 7.A pictorial representation of the averageorientation and magnitude of calculatedprincipal strains for the hemipelves loadedthrough an Austin Moore prosthesiswithout acetabular implant . Solid arrowsrepresent tensile strains, while dashedarrows represent compressive strains .

KEY

2 S.D.pst/N

.2 S .E.AVE.

GageReadings

- 2 S.E.

-2S.D .

pst/N 3 3

MaximumShearStrain

22

0 0

ust/ N 3

-2-2 -2-

86

CA

BRosette

A

BRosette

A B CRosette

FIGURE 8.Summary of side-to-side differences in strain data from the 12 pairs of hemipelves ofFigure 6.

Further removal of bone by reaming elevates all strain

SOURCES AND RELATED READINGlevels more or less uniformly. If almost all cancellous boneis removed, the picture changes . The changes in strain,from pre-implantation to post-implantation, appear to bequite varied and arbitrary . Surprisingly, however, themaximum shear strains appear to have decreased, mostdramatically at rosette C . (It should be noted that this set ofbones appeared to deviate from the normal patterns ofstrain considerably, pre-implantation, and any conclusionsdrawn are very tentative .)

The use of a protrusio ring to reinforce the acetabularimplantation after moderate bone removal eliminated theantero-medial shift in the strain pattern . Compressivestrains are unaffected, but tensile and shear strains arealmost uniformly increased by about V2 pst/N . Whenextensive bone removal (and/or destruction) has occurred,the protrusio ring does not have the same effect, as shownby the increases in tensile and shear strain antero-medially(rosettes A and B) . (The large scatter in the data is probablydue to variations in the amount of bone remaining andindicates the need to replicate this group of specimens togain additional confidence in the data .)

Reinforcing holes do not appear to affect the strainpatterns overall, but local effects anteriorly in the vicinity ofrosette A were noted . This is undoubtedly related to theproximity of one of the anchoring holes to that rosette.

ACKNOWLEDGEMENT

The authors acknowledge the excellent technicalassistance of Mina Robinson and the editorial assistance ofLinda Hensdell .

1. Anderson, C .B .J . ; M .A .R . Freeman ; and S .A .V . Swanson:Loosening of the Cemented Acetabular Cup in Total HipReplacement . J . Bone Joint Surg . 548 :590, 1972.

2. Beckenbaugh, R .D . and D .M . llstrup : Looking Back at TotalHip Arthroplasty, A Review of 333 Hips Four to Seven Yearsafter Surgery, with Special Emphasis on Loosening andWear . Presented at the American Academy of OrthopaedicSurgeons Meeting, Dallas, February 26, 1978.

3. Beckenbaugh, R .D . and D .M . llstrup : Total Hip Arthroplasty.A Review of 333 Cases with Long Follow-Up . J . Bone JointSurg . 60A :306-313, 1978

4. Charnley, J. : The Long-Term Results of Low-FrictionArthroplasty of the Hip Performed as a Primary Intervention.J . Bone Joint Surg . 5413 ;61-76, 1972.

5. Charnley, J . : Total Hip Replacement by Low-FrictionArthroplasty . Clin . Orthop . 72 :7-21, 1970.

6. Charnley, J . and Z . Cupic : The Nine and Ten Year Results ofthe Low-Friction Arthroplasty of the Hip . Clin . Orthop. 95 :9-25, 1973.

7. DeLee, J .G . and J . Charnley : Radiological Demarcation ofCemented Sockets in Total Hip Replacement . Clin . Orthop.121 :20-32, 1976.

8. Eichler, J . : Ein Vorschlag Zur Operativen Behandlung derProtrusio Acetabuli . Arch . orthop . Unfall-Chir . 75 :76, 1973.

9. Harris, W .H . : Loosening in The Hip . Proceedings of the SixthOpen Scientific Meeting of The Hip Society, pp . 162-175,1978.

10. Harris, W .H . and W .N. Jones : The Use of Wire Mesh in TotalHip Replacement . Clin . Orthop . 106 :117, 1975

11. McCollum, D .E . and J .A . Nunley : Bone Grafting inAcetabular Protrusio : A Biologic Buttress . In The Hip,Proceedings of the Sixth Open Scientific Meeting of the HipSociety, pp . 124-146, 1978.

12. Muller, M . : New Aspects in Total Hip Replacement.Presented at the Eastern Orthopaedic Association Meeting,Southhamptom, Bermuda, October 14, 1977.

13. Reckling, F .W . ; M .A . Asher ; and W .L . Dillon : A LongitudinalStudy of the Radiolucent Line at the Bone-Cement InterfaceFollowing Total Joint Replacement Procedures . J .Bone JointSurg . 59A :355-358, 1977.

14. Salvati, E .A . ; P . Bullough; and P .D . Wilson : IntrapelvicProtrusio of the Acetabular Component Following Total HipReplacement . Clin . Orthop . 111 :212, 1975.

15. Wilson, P .D ., Jr . : Frequency of Mechanical Failure of TotalJoint Replacement . In Proccedings of the Workshop onMechanical Failure of Total Joint Replacement, Atlanta,Georgia, May 31, 1978

87

Bulletin of Prosthetics Research BPR 10-33 (Vol . 17 No . 1) Spring 1980

KEY

Overall average ofPre-implantationvalues (from Fig. 6 )~

Prior to „--0. —\AfterImplantation

"°

Implantation

FIGURE 9.Key for data presented in Figures 10 through 17 . The threecrossbars on the vertical line represent the three values of strainrecorded or computed for each of the three hemipelves in groupstested prior to installation of the acetabular component with thecircle indicating the average value for "normal" hemipelves, asshown in Figures 6 and 7 . The three crossbars of the rectangleindicate values of strain obtained after component installation.

A

B

C

Maximum

pst/N ShearStrainspst/N

2

}1st/N

2

0

0

MaximumShearStrainsPrincipal

Strains

A B C

ROSETTES

PrincipalStrains

B CA

B

CROSETTE ROSETTE

0

0

B

0

ROSETTES

FIGURE 10a (top)

FIGURE 11a (top)Summary of strain data for hemipelves in set V L (no center hole,

Summary of strain data for hemipelves in set V R (10-mm centerremoval of cartilage only) .

hole, and removal of cartilage only).

FIGURE 10b (bottom)

FIGURE 11 b (bottom)Summary of difference in strain readings obtained before and after

Summary of difference in strain readings obtained before and aftercomponent installation for hemipelves in set V L .

component installation for hemipelves in set V R .

pst/

N

pst

/NM

axim

umS

hear

Stra

in 3 -

4.6

4.5

39

37

4.2

4.6

pst/

NM

axim

umps

t/ N

She

ar

0P

rinc

ipal

Str

ains

Str

ain

2

0

A B

C

A

B

C

Max

imum

Pri

ncip

al

She

arS

trai

n

Str

ains

2

A

B C

0

L.

00 _IL

Max

imum

—S

hear

Str

ain

pst/N 2 0

A

B

C

-1R

OS

ET

TE8

A B

C

RO

SE

TT

E

A

B

C

RO

SE

TT

EA

B C

RO

SE

TT

E

pst

/N M

axim

um

Shea

rS

trai

ns3

Pri

ncip

alS

trai

ns

o

2 0A

B C

A Pri

ncip

alSt

rain

s2

r

Max

imum

She

arS

trai

ns

0

i

A

B

C

RO

SETT

ES

L A

B C

RO

SE

TT

ES

FIG

UR

E 1

2a (

top)

Sum

rnary

of

str

ain

data

for

hem

ipelv

es in s

et

X L

(10 m

mcente

r hole

, re

am

ed to 4

mm

).

12b (

bottom

)y o

f diffe

rence in s

train

readin

gs o

bta

ined b

efo

re a

nd

con a

nent in

sta

llation for

hem

ipelv

es in s

et X

L.

FIG

UR

E 1

3a (

top

)S

um

mary

of

str

ain

data

for

hem

ipelv

es in s

et

VI

L (

20 m

mcente

r hole

, re

am

ed to p

elv

ic c

ort

ex —

less than 1

mm

of

bone a

t hole

edge).

FIG

UR

E13b (

bottom

)S

um

mary

of diffe

rence in s

train

readin

gs o

bta

ined b

efo

re a

nd

after

com

ponent in

sta

llation for

hem

ipelv

es in s

et V

I L

.

FIG

UR

E 1

4a (

top

)S

um

mary

of str

ain

data

for

hem

ipelv

es in s

et X

I L (

30 m

mcente

r hole

, re

am

ed to p

elv

ic c

ort

ex —

less than 1

mm

of

bone a

t hole

edge).

FIG

UR

E 1

4b

(b

ott

om

)S

um

mary

of diffe

rence in s

train

readin

gs o

bta

ined b

efo

re a

nd

after

com

ponent in

sta

llation for

hem

ipelv

es in s

et X

I L

.

Max

imum

She

arS

train

pst

/NP

rinci

pal

Str

ain

4r

3

pst

/N

45

4

6.3

56 i4 8

pst

/Npst/

N

43

3.8

4.2 E

.]E

JP

rinci

pal

Str

ains

Max

imum

_31

She

arS

train

s

pst/

N

3

Max

imum

She

arS

trai

n2

22

2

2

0

0

0

0

A B

C

0

A

B C

-2

2

A

8

C

Pri

nci

pal

Str

ains

-2 2

Max

imum

She

arS

trai

n

A 8

c

Prin

cipa

lS

trai

ns

A

B

C

-1 —

2

Max

imum

She

arS

trai

n

Max

imum

Pri

ncip

al

She

ar S

trai

n

Str

ains

0

0

A

B

C

A

B C

RO

SE

TT

E

RO

SE

TT

E

FIG

UR

E 1

5a (

top

)S

um

mary

of str

ain

data

for

hem

ipelv

es in s

et

VI

R (

20 m

m c

ente

rhole

, re

am

ed to p

elv

ic c

ort

ex —

less t

han 1

mm

of

bone a

t hole

edge p

lus p

rotr

usio

rin

g).

FIG

UR

E 1

5b (

bottom

)S

um

mary

of diffe

rence in s

train

readin

gs o

bta

ined b

efo

re a

nd a

fter

com

ponent in

sta

llation for

hem

ipelv

es in s

et

VI

R.

A

8

RO

SE

TT

E

-2A

S C

RO

SE

TT

E

FIG

UR

E 1

6a (

top

)S

um

mary

of str

ain

data

for

hem

ipelv

es in s

et X

I R

(30

mm

cente

r hole

, re

am

ed to p

elv

ic c

ort

ex —

less than 1

mm

of bone a

t hole

edge p

lus p

rotr

usio

rin

g).

FIG

UR

E 1

6b

(b

ott

om

)S

um

mary

of diffe

rence in s

train

readin

gs o

bta

ined b

efo

reand a

fter

com

ponent

insta

llation f

or

hem

ipelv

es in s

et

XI

R.

A

B

C

A

B C

2L

RO

SE

TT

E

-2L

RO

SE

TT

E

FIG

UR

E 1

7a (

top

)S

um

mary

of

str

ain

data

for

hem

ipelv

es in s

et

X R

(10 m

mcente

r hole

, re

am

ed to 4

mm

but w

ith a

dditio

n o

f th

ree

anchori

ng h

ole

s).

FIG

UR

E 1

7b

(b

ott

om

)S

um

mary

of diffe

rence in s

train

readin

gs o

bta

ined b

efo

re a

nd

after

com

ponent in

sta

llation for

hem

ipelv

es in s

et X

R.

CO W

00

-2

00

El