A Biomechanical Analysis

COPYRIGHT 2003 BY THE JOURNAL OF BONE AND JOINT SURGERY, INCORPORATED

A Biomechanical Analysis of Polyethylene Liner

Cementation into a Fixed Metal Acetabular Shell

BY GEOFFREY F. HAFT, MD, ANNELIESE D. HEINER, PHD, LAWRENCE D. DORR, MD, THOMAS D. BROWN, PHD, AND JOHN J. CALLAGHAN, MD

Investigation performed at the Departments of Orthopaedic Surgery and Biomedical Engineering, University of Iowa, Iowa City, Iowa

Background: A common clinical scenario encountered by an orthopaedic surgeon is a patient with a secure cement-less acetabular shell and a failed polyethylene liner. One treatment option is to cement a new liner into the fixedshell. The purpose of this study was to evaluate technical variables to improve the mechanical strength of such ce-mented liner constructs.

Methods: The contributions of shell texturing, liner texturing, and cement mantle thickness (between the liner andthe shell) were evaluated by comparing torsional strength (among nine groups of constructs) and lever-out strength(among eight groups of constructs).

Results: Failure almost always occurred at the cement-liner interface. The two exceptions (failure at the shell-cementinterface) occurred with a polished, untextured shell with no screw-holes. This finding indicates that if a shell has ex-isting texturing (such as holes), further intraoperative scoring of the shell is unnecessary, but some sort of texturingis necessary to avoid construct failure at the shell-cement interface. Textured liners had significantly ( = 0.05)greater torsional and lever-out strength than untextured liners. The greatest construct strength occurred when linergrooves were oriented so as to oppose the applied loading. A 4-mm-thick cement mantle resulted in slightly greatertorsional strength than a 2-mm-thick cement mantle, and a 2-mm-thick cement mantle resulted in considerablygreater lever-out strength than a 4-mm-thick cement mantle, but these differences were not significant.

Conclusions: When cementing a liner into a well-fixed shell, a surgeon should ensure that both the shell and theliner are textured, as interdigitation of the cement with the shell and the liner is crucial to the mechanical strength ofthis construct.

urgeons who perform total joint arthroplasty are fre-quently faced with a patient who has a cementless acetab-ular shell that is securely fixed to the pelvis but the

polyethylene liner has failed. Polyethylene wear and osteolysisassociated with a secure cementless acetabular shell are becom-ing frequent clinical problems, as better designs of the compo-nent last into the second decade1 and less optimal designs failearlier2-10. In addition to wear, another complication is dislodg-ment of the modular liner. Numerous case reports over the lastdecade have established the ubiquity of this complication in as-sociation with a variety of acetabular components11-22.

Several options are available to the surgeon in this situa-tion. One option is to perform a complete acetabular revision,which may be the preferred treatment in patients who havepoor fixation of the acetabular shell, malpositioning of theshell, or a very small shell in which a cemented liner would

unduly compromise the polyethylene thickness. However, com-plete revision of the acetabular component is accompanied byserious potential complications, including pelvic discontinu-ity and severe bone loss. Typically, a 6 to 8-mm increase in thediameter of the replacement acetabular shell is required withthis option, and cutting the original shell into pie-slice pieces(with the potential for particulate debris generation) is some-times necessary. Also, although revision acetabular compo-nents demonstrate stable interfaces at five to ten years, theyare more commonly associated with radiolucent lines than areprimary components23,24. Another option is simply to replacethe damaged or worn liner with a new version of the original.However, there are several common situations in which this isnot an option: the locking mechanism of the liner may bedamaged, a replacement polyethylene liner may be unavailableor of questionable quality (as a result of gamma irradiation in

S

TH E JO U R NA L OF BONE & JOINT SURGER Y JBJS .ORGVO LU M E 85-A NU M B E R 6 JU N E 2003

A BIOMECH ANICAL ANALYSIS OF POLYE T HY LENE LINER CEMENTAT ION INTO A FI XE D ME TA L ACE TA BU LA R SH E L L

air or a long or unknown shelf life), or a constrained liner maybe needed to combat instability after liner removal and capsu-lar dbridement. In these latter situations, the surgeons bestoption may be to leave the fixed shell in place and cement a re-placement liner into the shell.

Heck and Murray25, in 1986, were the first to describe, asfar as we know, the cemented liner technique in a case reporton the revision of a metal-on-metal prosthesis. While therehave been only a few additional case reports on this techniquein the orthopaedic literature12,14,20,26-30, discussions at nationalmeetings have suggested that many surgeons throughout theUnited States have been cementing liners into fixed shells inrecent years. In a retrospective review, seventeen patients witha cemented acetabular liner were followed for an average of 2.5years (range, one to 4.7 years)27. There were no radiographicchanges in the bone-shell interface during the follow-up inter-val, but the first patient in whom this method was used had arevision because of failure of the cement-liner interface.

Given that surgeons are actively employing this tech-nique, laboratory work is clearly necessary to help to deter-mine the best construct that can be created. Previous studiesof the strength of cemented liner constructs have investigateda few of the variables involved29-32. The purpose of the presentstudy was to expand upon these earlier studies and determinewhich surgeon-controlled variables would lead to the stron-gest mechanical construct when a polyethylene liner is ce-mented into a fixed acetabular shell. The specific objectives ofthe study were to determine the contributions of shell textur-ing, liner texturing, and cement mantle thickness to the over-all mechanical strength of the construct.

Materials and Methodshe contributions of the shell-cement and cement-liner in-terfaces to the cemented liner construct were evaluated

with use of torsional and lever-out tests. Torsional and lever-outtests were chosen on the basis of previously published studies of

mechanical testing of modular acetabular components33,34 andcemented all-polyethylene liners32 as well as proposed mecha-nisms for liner dislodgment in failed modular acetabular com-ponents11,14,16,18-20. Push-out testing was rejected as a clinicallyunrealistic failure mode, despite its previous use in the test-ing of liner-locking mechanisms of newly introduced modu-lar components33-35. For both the torsional and lever-out tests,yield strength and maximum strength were determined.

Experimental groups were designated (Table I) to deter-mine the effects of (1) unscored compared with scored ace-tabular shells, with and without holes; (2) smooth comparedwith textured liners; and (3) cement mantle thickness. Fourdifferent shells and five different liners were used to evaluatethese variables (Fig. 1), with five specimens for each combina-tion studied. The groups were chosen on the basis of com-ponent availability. All implants were provided by DePuyOrthopaedics (Warsaw, Indiana).

The shells included Summit cluster hole size-54 (unmod-ified) and highly polished Duraloc Enduron size-54 (central-hole-eliminated) components. The shells were scored (byelectron discharge machining) to simulate the intraoperativescratching that is often done ostensibly to improve cementinterdigitation25,29,31. The scoring consisted of channels cen-tered about the cup apex. The channels had a width of 2 mmand a depth of 1 mm. The shells were designated as cluster hole-unscored, cluster hole-scored, polished-no hole-unscored,and polished-no hole-scored.

The liners used in the present study included Summitsize-50, Summit size-46, Duraloc Enduron size-48, and Ul-tima size-44 implants. Equatorial nubs and ridges were re-moved from all liners except the Ultima liners (which wereunmodified). The Summit liners and the Duraloc liners wereaxisymmetric. These liners were designated as smooth andwere further designated by their cement mantle thicknesses(approximately 2, 3, and 4 mm for the Summit size-50, Dur-aloc Enduron size-48, and Summit size-46 liners, respectively).

T

TABLE I Combinations of Shells and Liners Used in Torsion and Lever-out Testing*

Shells

Liners

Smooth 2-mm (Summit 50)

Smooth 3-mm (Duraloc 48)

Smooth 4-mm (Summit 46)

Vertically Scored

(Summit 46)

Circumferentially Grooved

and Nubbed (Ultima 44)

Cluster hole-unscored (Summit 54)

T, L (2.05 mm) T, L (4.05 mm) T, L (4.05 mm) T, L (3.24 mm)

Cluster hole-scored (Summit 54)

T (4.05 mm) T (3.24 mm)

Polished-no hole-unscored (Duraloc 54)

T, L (3.15 mm) T, L (3.20 mm)

Polished-no hole-scored (Duraloc 54)

T, L (3.15 mm) L (3.20 mm)

*Groups were chosen on the basis of component availability. Five replicate shells and liners were tested in each combination. The compo-nents are designated by name and size. Cement mantle thicknesses for shell-liner combinations are also shown (in parentheses). All im-plants were manufactured by DePuy Orthopaedics (Warsaw, Indiana). T = torsion testing, and L = lever-out testing.



The Ultima liners, which had three circumferential groovesand a circle of six spacer nubs with a radius of 2 mm, weredesignated as circumferentially grooved and nubbed. Sum-mit size-46 liners were vertically scored with a specially groundcutter on a vertical computer numeric controlled (CNC) mill-ing machine. The scoring consisted of a cross pattern centeredabout the cup apex, with 2-mm-wide and 1-mm-deep scores.The scored Summit liners were designated as vertically scored.

The acetabular shells were potted with dental acrylicinto cylindrical containers (torsional specimens) or rectangu-lar containers (lever-out specimens) to simulate well-fixedshell backing. Care was taken to keep the potting medium outof the screw-holes and central impaction hole in the shellswith cluster holes. For the scored shells that were to be evalu-ated with lever-out testing, the scores were aligned with thesides of the acrylic pot. For the shells with cluster holes thatwere to be evaluated with lever-out testing, the screw-holeswere aligned so that they would be on the tensile side of thespecimen (and aligned with the top of the acrylic pot).

The liners were then cemented into the shells. All speci-mens were cleaned with soap and water before cementing. Thecement (Surgical Simplex P provided by Howmedica, Ruther-ford, New Jersey) was mixed by hand, and the liner was ce-mented five minutes after the cement-mixing was started. Forthe purpose of testing reproducibility, the thickness of the ce-ment mantle and the centering of the liner within each shellwere precisely controlled, with use of specially designed hemi-spherical spacers and an axial-torsional materials testing ma-

chine (model 858; MTS Systems, Eden Prairie, Minnesota).The hemispherical spacer was first placed in the shell, and theliner was lowered into the spacer with use of the materials test-ing machine until a compressive load was detected. The axialposition of the materials testing machine with the liner restingin the spacer was noted, the liner was raised, and the spacerwas removed. The cement was then poured into the shell, andthe liner was lowered into the cement to the previously notedlevel. Excess cement was removed with use of a straight, nar-row osteotome, taking care to avoid pulling cement out of theinterface. The liner was kept at this level for fifteen minutes,and then the specimen (the potting medium, shell, cement,and liner) was carefully removed from the fixturing of the ma-terials testing machine. The specimens were placed in an incu-bator at 37C, and the cement was cured for four to five daysbefore testing.

Torsional TestingNine experimental groups of five specimens each, for a total offorty-five specimens, were tested in torsion (Table I). The lin-ers were rigidly fixed to the actuator of the materials testingmachine with use of twelve small, circumferentially spacedscrews. The potted acetabular shells were fixed to the load-torque cell of the materials testing machine (Fig. 2-A). An x-ystage allowed free horizontal motion. An axial load of 70 kg(690 N) was applied to the liner. The liner was then rotatedabout its symmetry axis (Fig. 2-A) at a rate of 1/sec until fail-ure occurred. Torque and rotation angle data were collected

Fig. 1

Acetabular shells (top) and liners (bottom) used for torsional and lever-out testing of cemented liner constructs. The scoring simulated

intraoperative scratching that is done ostensibly to improve cement interdigitation; the channels are 2 mm wide and 1 mm deep. All

implants were manufactured by DePuy Orthopaedics.



at 0.002-second intervals by the materials testing machinesoftware (MTS) and analyzed with use of Excel software(Microsoft, Seattle, Washington). The failure interface wasrecorded and photographed.

Yield torque and maximum (ultimate) torque were de-termined from each torque-angle recording (Fig. 2-B). Yieldtorque was defined as either the first abrupt change of slopefor the torque-angle recording or (to account for nonabruptslope changes) the intersection of the torque-angle curve witha 0.01 offset initial slope, whichever was the smaller value.Each torque measurement was averaged for each shell-linercombination. We used two methods to examine our data.First, after testing the assumption of homogeneity of varianceamong the nine experimental groups with use of the Brownand Forsythe method36, one-way analysis of variance was usedto determine whether there was a significant mean difference( = 0.05) between any of the shell-liner combinations. TheTukey-Welsch multiple comparison procedure was then usedto determine which shell-liner combinations were signifi-cantly different from one another. Second, when the differ-ence between the mean values did not quite reach significancewith the Tukey-Welsch test, we used the common Student ttest. The first method was chosen because of variance hetero-

geneity and because it provides a conservative estimate of sig-nificance. The second method ignores variance heterogeneityand thus provides a lower limit to the estimate of p. Statisticalpower was calculated for = 0.05 for nine experimentalgroups of five specimens each and for the realized effect sizesand a benchmark effect size of one standard deviation37.

Lever-out TestingEight experimental groups, with five specimens in each for a to-tal of forty specimens, were evaluated with lever-out testing.The liners were attached to a 6.35-mm-thick (0.25-in) grippingring, with use of twelve small circumferentially spaced screws(Fig. 3-A). A lever arm (a 3/8"-24 bolt) was screwed into thecenter of the bearing surface, and a retaining nut was screwed tothe level of the ring. A low-melting-point (70C) bismuth alloy(Cerrobend; Cerro Metal Products, Bellefonte, Pennsylvania)was then poured into the liner. For the smooth 3-mm liners, thevertically scored liners, and the circumferentially grooved andnubbed liners, eight holes were drilled into the bearing surfacefor interdigitation of the bismuth alloy, to improve purchase.The specimen was then attached to the load cell of the materialstesting machine. Lever-out torque was applied by means of acylindrical platen eccentrically contacting the lever arm. Theplaten was lowered at a rate of 1.33 mm/sec (corresponding to a

Fig. 2-A

Figs. 2-A and 2-B Torsional testing of the cemented liner

specimens. Fig. 2-A Cutaway schematic for a smooth 4-mm

liner with a cluster hole-unscored shell. Z indicates the sym-

metry axis of the liner.

Fig. 2-B

Representative recording, indicating yield and maximum

torque values.



liner rotation of approximately 1/sec) until failure occurred.Force (Fig. 3-A) and displacement data were collected in0.01-sec intervals by the materials testing machine software(MTS), were converted to moment and angular data, andwere analyzed with Excel software (Microsoft). The failureinterface was recorded and photographed.

Yield moment and maximum (ultimate) moment weredetermined from each moment-angle recording (Fig. 3-B).Yield moment was defined as the intersection of the moment-angle curve with a 0.05 offset initial slope. Each moment mea-sure was averaged for each shell-liner combination. Again, aftertesting the assumption of homogeneity of variance among theeight experimental groups with use of the Brown and Forsythemethod36, one-way analysis of variance was used to deter-mine whether there was a significant mean difference ( =0.05) between any of the shell-liner combinations. The Tukey-Welsch multiple comparison procedure was used first, and thenthe Student t test was used to determine which shell-liner com-binations were significantly different from one another. Statisti-cal power was calculated for = 0.05 for eight experimentalgroups of five specimens each and for the realized effect sizesand a benchmark effect size of one standard deviation37.

Resultshell texturing, liner texturing, and the cement mantlethickness each affected both the torsional and the lever-out

strength of the cemented liner constructs (see Appendix). Asan a priori measure, for an effect size of one standard devia-

tion, power was 0.997 for the torsional measures and 0.995 forthe lever-out measures. For the actual one-way analyses ofvariance, power was 1.000 for all four measures. Visible failureoccurred only at the cement-liner interface, with the exceptionof only two specimens. The two exceptions were both polished-no hole-unscored shells; one involved failure of the shell-cement interface with a circumferentially grooved and nubbedliner, and the other was a combined failure of the cement-linerand shell-cement interfaces with a smooth 3-mm liner. Be-cause of differences in the design features among the implants,not all cemented liner constructs could be directly compared;therefore, only certain subsets are discussed below.

Effect of Shell TexturingThe groups were not compared with respect to the effect ofshell texturing because failure did not occur at the shell-cement interface except in the two specimens discussed above.As long as the shell had some type of texturing, whether exist-ing features (screw-holes) or intraoperative scoring, no fail-ure occurred at the shell-cement interface.

Effect of Liner TexturingThe effect of liner texturing was studied with the cluster hole-unscored shells as a constant factor; the smooth 4-mm liners,vertically scored liners, and circumferentially grooved andnubbed liners were compared. In torsion, yield torque wassignificantly higher for the vertically scored liners ( = 0.05)(Fig. 4-A). Maximum torque was significantly higher ( =

S

Fig. 3-A

Figs. 3-A and 3-B Lever-out testing of cemented liner specimens.

Fig. 3-A Cutaway schematic for a circumferentially grooved and

nubbed liner with a cluster hole-unscored shell. P = force.



0.05) for the two textured liners (the vertically scored linersand the circumferentially grooved and nubbed liners), andthese two liners were not significantly different from one an-other (Fig. 4-A). In the lever-out test, the yield and maximummoments were significantly higher for the circumferentiallygrooved and nubbed liners ( = 0.05) (Fig. 4-B).Effect of Cement Mantle ThicknessThe effect of cement mantle thickness was studied with thecluster-hole-unscored shell as a constant factor; the smooth4-mm and smooth 2-mm liners were compared. In torsion,yield and maximum torque were slightly higher for smooth4-mm liners, but the difference was not significant (Fig. 5-A).In the lever-out test, the yield moment for the 4-mm cementmantle was a mean (and standard deviation) of 6.63 2.36N-m, whereas the 2-mm cement mantle resisted with a meanof 22.85 6.41 N-m (Fig. 5-B). The difference between themeans (16.2 N-m) did not meet the requirement of a meandifference of 17.5 N-m for significance with use of the Tukey-Welsch test. However, when the conventional Student two-tailed t test was used, the p value was 0.0031. For the maximummoment in the lever-out test, the 4-mm cement mantle re-sisted with a mean (and standard deviation) of 23.14 6.32N-m, whereas the 2-mm cement mantle resisted with a meanof 42.37 4.40 N-m (Fig. 5-B). The difference between thesemeans (19.2 N-m) also did not meet the requirement for a

mean difference of 21.6 N-m for significance with use of theTukey-Welsch test; however, when examined with the Studenttwo-tailed t test, the p value was 0.0008 (Fig. 5-B).

Discussionhe increasing use of the cemented liner technique has gen-erated a compelling need to compare the strength of this

construct with that of modular acetabular components. Mel-drum and Hollis31 established that the strength of a cementedliner construct was similar to that of modular components inlever-out and push-out testing. The present study evaluatedthe mechanical strength of the shell-cement-liner construct intorsional and lever-out tests, as a function of individual factors(shell texturing, liner texturing, and cement mantle thickness)that characterize the construct. The results of the current studyoffer guidance to a surgeon who chooses to use a cementedliner technique.

Shell texturing, in any form, prevented failure at the shell-cement interface. As long as the shell had holes or was scored,visible failure occurred only at the cement-liner interface. Thisresult suggests that the practice of intraoperative scoring of theacetabular shell, to improve the strength of the cement-shell in-terface, is unnecessary provided that the shell has screw-holes orother existing texturing. The data reported by Dunlop et al.38

and the surgical technique described by LaPorte et al.29 also sup-port the concept that, as long as a shell has screw-holes, addi-tional scoring is not needed. Avoiding scoring of the shellprevents the creation of metal particulate debris (a likely sourceof third-body wear) during the scoring process. However, if ashell lacks texturing features, a surgeon should score the shell toavoid failure at the shell-cement interface.

Regarding the effect of cement thickness, the present re-sults suggest that untextured (smooth) liners with a 2-mm-thick cement mantle had greater construct strength in thelever-out test than did the constructs with the 4-mm-thickcement mantle. We say this because there may be differingopinions as to the type of statistical test that is appropriate forthese data. The most conservative test, the Tukey-Welsch test,just missed significance (the test provided no p values), whereasthe less restrictive Student two-tailed t test provided p values(yield moment, 0.00031; maximum moment, 0.0008) thatwere highly significant. Under torsional loading, the differ-ences between the liners with a 4-mm-thick cement mantleand those with a 2-mm-thick mantle did not appear to besignificant. Bensen et al.30 observed that the mean lever-outstrength for cemented liners with a 4-mm cement mantle was37 N-m, whereas the liner with a 2-mm cement mantle wouldnot lever out before the polyethylene yielded about the holeinto which the lever-out rod was inserted (moments as high as68 N-m were recorded). In torsion, the 2-mm-thick and the4-mm-thick cement mantles demonstrated no significant dif-ferences, although a trend toward better performance wasnoted for the thicker (4-mm) cement mantle. A possible ex-planation may be that interface stresses change minimally asthe cement thickness changes. In a finite element study ofcemented liners under axial loading, Kurtz et al.39 found a

T

Fig. 3-B

Representative recording, indicating yield and maximum

moment values.



Effect of liner texturing in torsion testing (Fig. 4-A) and lever-out testing (Fig. 4-B). The cluster hole-unscored shell was com-

mon to all specimens in this evaluation. Asterisks indicate groups with significantly different average values ( = 0.05).

Fig. 4-B

Fig. 4-A



change of only 9% in cement compressive principal stress asthe cement mantle thickness was changed between 0.5 and 2.0mm. With no compelling data pointing toward a particularcement thickness, a surgeon is faced with a choice. One optionis to choose a thicker cement mantle. For example, to ensure acomplete cement mantle of at least 3 mm throughout (and toprevent the liner from bottoming out), a 4-mm-thick differ-ential (in the radius) between the liner and the shell is proba-bly a good starting point. By erring on the side of caution andcementing a slightly undersized liner, the surgeon can ensureadequate liner positioning, containment of the componentwithin the shell, and at least minimal cement mantle thicknessthroughout the construct. (In the present laboratory trials, ce-ment mantle thickness and centering were more meticulouslycontrolled than is practical intraoperatively.) The other optionis to choose a thinner cement mantle, which means a thickerliner will be used, resulting in the availability of more polyeth-ylene for wear and for any necessary liner scoring32. However,a thin cement mantle could lead to bottoming out of the lineror (in an effort to avoid bottoming out) placement of a proudliner. It should be noted, though, that if a liner with spacernubs was used, bottoming out and malpositioning of the linerwould not be problems.

The most important variable for the surgeon to controlis liner texturing. In almost every cemented liner combinationtested, failure occurred at the cement-liner interface. In tor-sion, the cemented liner constructs with a vertically scoredliner had the highest average yield torque and maximumtorque; this difference was significant as compared with theother constructs, with one exception for each measure (bothexceptions having a circumferentially grooved and nubbedliner) (see Appendix). In lever-out testing, constructs with acircumferentially grooved and nubbed liner had a significantlygreater yield moment and maximum moment than any otherconstructs (see Appendix). Thus, when the groove orienta-tion was directed so as to oppose the applied loading, therewas a much greater resistance to failure. These results suggestthat if a smooth liner is to be used, the surgeon should textureit with a series of orthogonal grooves prior to cementing it inplace. This conclusion is consistent with the data reported byOh32, who tested the effect of grooves on the torsional strengthof a cemented all-polyethylene liner and found that the addi-tion of 1-mm-deep grooves to a liner increased torsionalstrength from approximately 28 to 154 N-m. This conclusionis also supported by Dunlop et al.38, who measured the tor-sional strength of cemented liner constructs as 2.4 to 14.6N-m with a smooth liner, 15.6 to 44.4 N-m with a scored liner,and 40.4 to 78.4 N-m with an all-polyethylene liner.

The U.S. Food and Drug Administration requires test-ing of the strength of liner capture mechanisms for all com-mercially available modular acetabular components. The resultsof the present study of cemented liner constructs with tex-tured liners compare favorably with those in previously pub-lished studies of modular acetabular components. For thecircumferentially grooved and nubbed liner cemented into thecluster hole-unscored shell in the present study, the average

maximum lever-out moment was 146 N-m. Tradonsky et al.34

determined the lever-out strength of eight contemporarymodular components; the results ranged from a high of 77.3N-m (Duraloc; DePuy, Warsaw, Indiana) to a low of 4.9 N-m(Triloc; DePuy), with a median of 37.5 N-m (Omnifit; Os-teonics, Allendale, New Jersey). Bailey et al.33 determined thatthe lever-out strength of a Durasul Inter-Op liner (Sulzer Or-thopedics, Austin, Texas) was 65.5 N-m.

Davidson et al.40 determined the frictional torque of theinterface of a cobalt-chromium-alloy head articulating againstan ultra-high molecular weight polyethylene shell with useof a nonrocking biaxial hip-joint simulator. With a 32-mm-diameter head, 5000-N maximum load, and water as a lubri-cant, the frictional torque measured an average of 0.94 N-m.This value is lower than even the lowest average maximumtorque measured for the cemented liners (6.1 N-m for a smooth3-mm liner cemented into a polished-no hole-unscored shell)and considerably lower than the average maximum torquesmeasured for the cemented liners that had texturing (51.9 to65.7 N-m) (see Appendix).

The current study has several limitations. First, the pro-cess used to cement the liners into the shells was an idealizedrepresentation of the actual intraoperative procedure. Cementthicknesses, liner centering, and creation of reproducible scoresin shells and liners were carefully controlled to avoid addingconfounding variables to the model, but they clearly repre-sented a best-case scenario. Second, a larger number of speci-mens in each shell-liner combination could have provided agreater statistical power; for instance, the differences in out-come measures between the 2-mm and 4-mm cement mantlethicknesses might have reached significance with a larger sam-ple size. Third, fatigue failure, which would have been apotentially informative mode of testing, was not considered inthe current study; liner fatigue damage and subsequent fail-ure could result from numerous instances in which the yieldstrength of the liner is exceeded. Fourth, it was not possible toobtain completely axisymmetric shells, to directly test the hy-pothesis that at least some shell texturing is necessary; the de-signs of the shells in this study included an outer liner-lockingmechanism, which provided some stability in torque andlever-out testing. However, most shells in patients who needan acetabular revision include these liner-locking mecha-nisms, making the results in the present study clinically appli-cable. Finally, the shells and liners tested came from only onecompany; the quantitative values might be somewhat differ-ent with shells and liners from a different company. How-ever, if similar variables of shell and liner texturing werestudied, the qualitative results and subsequent conclusionsobtained should be similar, regardless of which companysimplants were considered.

Given the importance of liner texturing, the intraopera-tive time limitations for adequately performing the texturing,and the potential for weakening a liner by over-texturing itwith a high-speed burr, the authors recommend that manu-facturers consider producing polyethylene liners specificallydesigned for cementing into a shell. The features of these lin-



Fig. 5-A

Effect of cement mantle thickness in torsion testing (Fig. 5-A) and lever-out testing (Fig. 5-B). The cluster hole-unscored

shell was common to all specimens in this evaluation. No groups had significantly different average values ( = 0.05).

Fig. 5-B



ers would resemble the backside designs of cemented all-polyethylene acetabular components. The liners should havevertical grooves to increase torsional strength at the cement-liner interface and circumferential grooves to increase thelever-out strength. The liners should also incorporate spac-ers to prevent bottoming out or malpositioning of the linerin the shell; these spacers can also contribute to the strengthof the cement-liner interface. Trial liners, which the surgeoncan use to ensure a proper fit in a variety of shell types, arealso crucial. Along these lines, manufacturers should alsoconsider making a constrained liner with these characteris-tics, as the cemented liner technique is gaining popularity inpatients with recurrent dislocations26,27. Given the increasingprevalence of patients who will require cemented acetabu-lar liners in the coming years, the availability of a line ofprefabricated textured liners would be useful in revision hipsurgery.

AppendixResults of cemented liner constructs tested in torsionand lever-out are available with the electronic versions of

this article, on our web site at www.jbjs.org (go to the articlecitation and click on Supplementary Material) and on ourquarterly CD-ROM (call our subscription department, at 781-449-9780, to order the CD-ROM).

References

1. Schulte KR, Callaghan JJ, Kelley SS, Johnston RC. The outcome of Charnley total hip arthroplasty with cement after a minimum twenty-year follow-up. The results of one surgeon. J Bone Joint Surg Am. 1993;75:961-75.

2. Bankston AB, Cates H, Ritter MA, Keating EM, Faris PM. Polyethylene wear in total hip arthroplasty. Clin Orthop. 1995;317:7-13.

3. Callaghan JJ, Kim YS, Brown TD, Pedersen DR, Johnston RC. Concerns and improvements with cementless metal-backed acetabular components. Clin Orthop. 1995;311:76-84.

4. Clohisy JC, Harris WH. The Harris-Galante porous-coated acetabular compo-nent with screw fixation. An average ten-year follow-up study. J Bone Joint Surg Am. 1999;81:66-73.

5. Engh CA Jr, Culpepper WJ 2nd, Engh CA. Long-term results of use of the ana-tomic medullary locking prosthesis in total hip arthroplasty. J Bone Joint Surg Am. 1997;79:177-84.

6. Hernandez JR, Keating EM, Faris PM, Meding JB, Ritter MA. Polyethylene wear in uncemented acetabular components. J Bone Joint Surg Br. 1994;76:263-6.

7. Latimer HA, Lachiewicz PF. Porous-coated acetabular components with screw fixation. Five to ten-year results. J Bone Joint Surg Am. 1996;78:975-81.

8. Maloney WJ, Peters P, Engh CA, Chandler H. Severe osteolysis of the pelvis in association with acetabular replacement without cement. J Bone Joint Surg Am. 1993;75:1627-35.

9. Nashed RS, Becker DA, Gustilo RB. Are cementless acetabular components the cause of excess wear and osteolysis in total hip arthroplasty? Clin Orthop. 1995;317:19-28.

10. Sychterz CJ, Moon KH, Hashimoto Y, Terefenko KM, Engh CA Jr, Bauer TW. Wear of polyethylene cups in total hip arthroplasty. A study of specimens retrieved post mortem. J Bone Joint Surg Am. 1996;78:1193-200.

11. Barrack RL, Burke DW, Cook SD, Skinner HB, Harris WH. Complications related to modularity of total hip components. J Bone Joint Surg Br. 1993;75:688-92.

12. Beaver RJ, Schemitsch EH, Gross AE. Disassembly of a one-piece metal-backed acetabular component. A case report. J Bone Joint Surg Br. 1991;73:908-10.

13. Brien WW, Salvati EA, Wright TM, Nelson CL, Hungerford DS, Gilliam DL. Dissociation of acetabular components after total hip arthroplasty. Report of four cases. J Bone Joint Surg Am. 1990;72:1548-50.

14. Gonzalez della Valle A, Ruzo PS, Li S, Pellicci P, Sculco TP, Salvati EA. Dis-lodgment of polyethylene liners in first and second-generation Harris-Galante acetabular components. A report of eighteen cases. J Bone Joint Surg Am. 2001;83:553-9.

15. Kauschke T, Zilch H. [An unusual early complication in cementless replace-ment of the hip joint. Case report]. Unfallchirurgie. 1994;20:329-32. German.

16. Kitziger KJ, DeLee JC, Evans JA. Disassembly of a modular acetabular com-ponent of a total hip-replacement arthroplasty. A case report. J Bone Joint Surg Am. 1990;72:621-3.

17. Louwerse RT, Heyligers IC. Late failure of the polyethylene liner fixation in an uncemented total hip arthroplasty. J Arthroplasty. 1999;14:391-6.

18. OBrien RF, Chess D. Late disassembly of a modular acetabular component. A case report. J Arthroplasty. 1992;7 Suppl:453-5.

19. Retpen JB, Solgaard S. Late disassembly of modular acetabular components. A report of two cases. Acta Orthop Scand. 1993;64:193-5.

20. Ries MD, Collis DK, Lynch F. Separation of the polyethylene liner from acetabular cup metal backing. A report of three cases. Clin Orthop. 1992;282:164-9.

21. Star MJ, Colwell CW Jr, Donaldson WF 3rd, Walker RH. Dissociation of mod-ular hip arthroplasty components after dislocation. A report of three cases at differing dissociation levels. Clin Orthop. 1992;278:111-5.

22. Castro FP Jr, Chimento G, Munn BG, Levy RS, Timon S, Barrack RL. An anal-ysis of Food and Drug Administration medical device reports relating to total joint components. J Arthroplasty. 1997;12:765-71.

23. Silverton CD, Rosenberg AG, Sheinkop MB, Kull LR, Galante JO. Revision of the acetabular component without cement after total hip arthroplasty. A follow-up note regarding results at seven to eleven years. J Bone Joint Surg Am. 1996;78:1366-70.

24. Weber KL, Callaghan JJ, Goetz DD, Johnston RC. Revision of a failed ce-mented total hip prosthesis with insertion of an acetabular component with-out cement and a femoral component with cement. A five to eight-year follow-up study. J Bone Joint Surg Am. 1996;78:982-94.

Geoffrey F. Haft, MDThomas D. Brown, PhDDepartment of Biomedical Engineering, University of Iowa, 2181 Westlawn Building, Iowa City, IA 52242. E-mail address for G.F. Haft: [email protected]. E-mail address for T.D. Brown: [email protected]

Anneliese D. Heiner, PhDJohn J. Callaghan, MDDepartment of Orthopaedic Surgery, University of Iowa, 200 Hawkins Drive, Iowa City, IA 52242. E-mail address for J.J. Callaghan: [email protected]. E-mail address for A.D. Heiner: [email protected]

Lawrence D. Dorr, MDThe Dorr Arthritis Institute, Centinela Hospital Medical Center, 501 East Hardy Street, Suite 300, Inglewood, CA 90301. E-mail address: [email protected].

In support of their research or preparation of this manuscript, one or more of the authors received grants or outside funding from DePuy, War-saw, Indiana. In addition, one or more of the authors received payments or other benefits or a commitment or agreement to provide such benefits from commercial entities (DePuy and Howmedica, Rutherford, New Jer-sey). Also, a commercial entity (DePuy) paid or directed, or agreed to pay or direct, benefits to a research fund, foundation, educational institution, or other charitable or nonprofit organization with which the authors are affiliated or associated.



25. Heck DA, Murray DG. In vivo construction of a metal-backed, high-molecular-weight polyethylene cup during McKee-Farrar revision total joint arthroplasty. A case report. J Arthroplasty. 1986;1:203-6.

26. Goetz DD, Capello WN, Callaghan JJ, Brown TD, Johnston RC. Salvage of a recurrently dislocating total hip prosthesis with use of a constrained acetabu-lar component. A retrospective analysis of fifty-six cases. J Bone Joint Surg Am. 1998;80:502-9.

27. Haft GF, Heiner AD, Callaghan JJ, Brown TD. Stability of acetabular liners ce-mented into cementless shells. Presented as a poster exhibit at the Annual Meeting of the American Academy of Orthopaedic Surgeons, San Francisco, California, February 28-March 4, 2001.

28. Implex. Continuum hip system: Hedrocel acetabular cup: replacement insert technique. Product literature. Allendale, NJ: Implex; 1999.

29. LaPorte DM, Mont MA, Pierre-Jacques H, Peyton RS, Hungerford DS. Tech-nique for acetabular liner revision in a nonmodular metal-backed component. J Arthroplasty. 1998;13:348-50.

30. Bensen CV, Del Schutte H Jr, Weaver KD. Mechanical stability of polyethyl-ene liners cemented into acetabular shells. Crit Rev Biomed Eng. 2000;28 (1-2):7-10.

31. Meldrum RD, Hollis JM. The strength of a cement acetabular locking mecha-nism. J Arthroplasty. 2001;16:748-52.

32. Oh I. A comprehensive analysis of the factors affecting acetabular cup fixa-tion and design in total hip replacement arthroplasty: a series of experimen-tal and clinical studies. In: The hip proceedings of the eleventh open scientific meeting of The Hip Society. St. Louis: CV Mosby; 1983. p 129-77.

33. Bailey A, Radefeld R, Krevolin J. Liner stability in acetabular component fa-

tigue performance. Trans Am Soc Biomech. 2000;24:165-6.

34. Tradonsky S, Postak PD, Froimson AI, Greenwald AS. A comparison of the disassociation strength of modular acetabular components. Clin Orthop. 1993;296:154-60.

35. American Society for Testing and Materials, editors. Standard test method for determining the axial disassembly force of a modular acetabular device. In: Annual book of ASTM standards. Volume 13, Medical devices; emergency medical services. Philadelphia: American Society for Testing and Materials; 2003. E1820-97.

36. Conover WJ, Johnson ME, Johnson MM. Comparative study of tests for ho-mogeneity of variances, with applications to the outer continental shelf bid-ding data. Technometrics. 1981;23:351-61.

37. Borenstein MT, Rothstein H, Cohen J, Shoenfeld D, Berlin J, Lakatos E. Power and precision, version 2. A computer program for statistical power analysis and confidence intervals. Mahwah, NJ: Lawrence Erlbaum Associ-ates; 2001.

38. Dunlop D, Oxland T, Gordon J, Garbuz D, Masri B, Duncan C. The biomechan-ics of cementing a new liner into a well-fixed metal acetabular shell. Trans Orthop Res Soc. 2002;27:973.

39. Kurtz S, Hozak W, Ray R, Edidin A. Biomechanical analysis of cementing an acetabular liner in a metal shell for revision total hip arthroplasty. Trans Orthop Res Soc. 2001;26:1078.

40. Davidson JA, Schwartz G, Lynch G, Gir S. Wear, creep, and frictional heating of femoral implant articulating surfaces and the effect on long-term perfor-mancePart II, Friction, heating, and torque. J Biomed Mater Res. 1988;22 (A1 Suppl):69-91.

Documents

A Biomechanical Analysis