combined wear behavior and long-term implant-bone fixation of total knee replacement: a novel in...

aor_972 177..183

Combined Wear Behavior and Long-Term Implant-BoneFixation of Total Knee Replacement:

A Novel In Vitro Set-up

*†Michele Spinelli, *Saverio Affatato, *‡Luca Cristofolini, *Paolo Erani,§Domenico Tigani, and *Marco Viceconti

*Laboratorio di Tecnologia Medica, Istituto Ortopedico Rizzoli, Bologna; †DMTI, Engineering Faculty, Universityof Florence, Florence; ‡Engineering Faculty, University of Bologna; and §7th Division, Istituto Ortopedico Rizzoli,

Bologna, Italy

Abstract: The success of a total knee replacement (TKR)strongly depends on the prosthetic design; this includes onone hand the best choice of the bearing materials to mini-mize wear, on the other hand a good orientation of theprosthetic components with respect to the loadingdirections. The aim of this study was to investigate thefeasibility of a new experimental setup combining two fun-damental aspects for the long-term success of kneeimplants: wear and micromotions. A novel procedure wasused to simulate working conditions as close as possible toin vivo ones and to measure implant-bone micromotion, bymeans of fixing the femoral component of the prosthesis tothe distal part of a synthetic femur to be tested through aknee simulator. Gravimetric wear of the tibial specimens

was assessed at regular intervals. Implant-bone induciblemicromotions and permanent migrations were measured atthree locations throughout the test. Wear patterns on tibialspecimens were characterized through a standardized pro-tocol based on digital image analysis; fatigue damage in thecement was quantified. Some initial conditioning wasnoticed both in the wear process and microcracking distri-bution within the cement mantle. Similarity in wear tracksobserved on tibial inserts and other retrieval studies,coupled with clinically consistent migration patternsfor TKR, supports the efficacy of the new in vitromethod presented. Key Words: Knee simulator—Micromotion—In vitro stability—Femoral fixation—Meniscus wear—Total knee replacement.

A large and increasing number of total kneereplacements (TKRs) are performed yearly (1–3) tosatisfy younger and more active patients under-going TKR (1–3) with respect to better long-termexpectations. In spite of the good results obtainedwith in vitro wear tests, massive polyethylene wearremains a serious complication in many TKR cases(4–6). This suggests that there are one or more cofac-tors which are not accounted in standard ISO 14243-1/3 wear tests that might induce severe wear in vivo.Cadaver studies suggest that changes in the orienta-tion of the prosthetic components with respect to the

loading directions as small as 3° can dramaticallychange the contact pressure, one of the prime deter-minants of wear (7–11).Thus, it might be possible thatone of those cofactors driving the clinical results farfrom the ones observed in vitro might be related tothe fact that, in in vitro wear tests, we usually assumethe prosthetic components are rigidly fixed withrespect to the loading directions.

Current laboratory methods make possible toevaluate with very good accuracy the elastic and per-manent changes of position and orientation of theprosthetic components with respect to the host boneand, thus, with respect to the loading directions (12).This aspect, however, has been poorly investigated byliterature reports (12). In TKR, in fact, load transferbetween the bone and the prosthesis plays a relevantrole in the observed dynamic behavior, especially inthe femoral part, and this might be attributable tothe implant-bone fixation and alignment. It is also

doi:10.1111/j.1525-1594.2009.00972.x

Received June 2009; revised October 2009.Address correspondence and reprint requests to Dr. Saverio

Affatato, Laboratorio di Tecnologia Medica—Istituto OrtopedicoRizzoli, Via di Barbiano, 1/10, 40136 Bologna, Italy. E-mail:[email protected]

Artificial Organs34(5):E177–E183, Wiley Periodicals, Inc.© 2010, Copyright the AuthorsJournal compilation © 2010, International Center for Artificial Organs and Transplantation and Wiley Periodicals, Inc.

E177

possible to measure the wear induced by cyclicloading and motion, but assuming that the relativeposition of the prosthetic components does notchange with respect to the loading directions (13–17).

What so far has not been possible is to evaluatethe wear of a TKR under realistic biomechanical con-ditions, which include not only a detailed reproduc-tion of the kinematics and dynamic conditions, butalso the influence that the load-induced micromotionmight have on the instantaneous contact pressure.

The scope of this study is to verify the feasibilityof an experimental setup where these two aspects,wear of polyethylene insert and micromotion ofthe femoral component, are combined effectively,without compromising the general accuracy andrepeatability of the methods.

MATERIALS AND METHODS

A novel, consistent procedure was used to simulateworking conditions as close as possible to in vivoones and to measure implant-bone micromotion;weight loss of meniscal inserts and cement damagewere assessed after test completion (12).

SpecimensIn order to test an anatomic cement mantle sup-

ported in a way that replicates real bone mechanicalproperties, the femoral components were implantedon composite femur models (Mod. 3106, Pacific-Research, Vashon, WA, USA) (12). Compositemodels allow a significant reduction of variabilitycompared to human bones, thus obtaining a bettersensitivity for comparative studies (18–22). A com-puter numerical control milling machine was used toresect the distal part of the femurs to accomplish thehighest repeatability in the bone-cutting geometry.Cobalt–chromium TKR femoral components (size 3,Multigen-Plus, Lima, Udine, Italy) (23) werecemented by an experienced knee surgeon (pre-chilled -4°C bowl-mixed bone cement, Palacos-R40,Schering-Plough, Brussels, Belgium). As failure mostlikely initiates at the prosthesis–cement interface(24), it was essential to ensure strong cement–bonebonding in the in vitro model; this was guaranteed asexplained in (12). Nine composite femur specimenswere prepared following the same procedure to allowrobust mechanical testing (12); they were dividedinto groups of three in order to

• perform mechanical test (N = 3);• guarantee a soak control (N = 3);• ensure absence of artifacts due to inspection pro-

cedure (N = 3, blank specimens).

During the mechanical test, the three test speci-mens were coupled with three ultra high molecularweight polyethylene (UHMWPE) meniscal bear-ings supported by cobalt–chromium tibial trays(Multigen-Plus).

Wear test detailsMechanical tests were performed using a three-

plus-one-stations knee simulator (Shore Western,Los Angeles, CA, USA) in accordance with theISO14243-1,3 requirements. Femoral and tibial com-ponents were mounted on the simulator following acustom-designed setup (12). Five million cycles wererun at a frequency of 1.1 Hz. Twenty-five percentsterile bovine calf serum (Sigma, St. Louis, MO,USA) balanced with deionized water and 0.2%sodium azide (Merck, Darmstadt, Germany) todelay bacterial degradation was used as lubricant(at 37 � 2°C). Ethylenediaminetetraacetic acid(20 mmol/dm3) was added to minimize precipitationof calcium phosphates. Gravimetric wear of the tibialspecimens was assessed at 500 000 cycle intervals.Weight loss was measured through a microbalance(Sartorius AG, Göttingen, Germany; resolution0.01 mg and accuracy �0.1 mg). A standardized pro-tocol based on black and white image analysis wasused to examine the surface of the tibial inserts after5 million cycles. In the first phase, digital images ofthe specimens (tested and retrieved) were calibrated(geometrically and optically) to equalize informationcontent; in the second phase, a worn area wasdetected for each compartment and expressed as apercentage of the total working area (12).

Coefficient of variation (CV) was calculated forthe mean wear rate in order to account for test reli-ability through the comparison with other in vitrowear tests on similar implant designs. A value of CVclose to 0 indicates low dispersion of the measuredphenomena.

Implant-bone micromotion measurementEach implanted femur was instrumented with

three waterproof linear variable differential trans-former (LVDT) displacement transducers providedwith a flat tip (12). The transducers were aligned soas to measure displacement of the femoral compo-nent along the femoral long axis. This is the direc-tion where the largest micromotions (either due totranslation or to component tilting) are observedclinically with radiostereometric analysis (RSA)(25,26). The frame of the LVDTs were fixed at givenlocations to the femur (thus, they rocked togetherwith the implanted femur during the test), while the

M. SPINELLI ET AL.E178

Artif Organs, Vol. 34, No. 5, 2010

LVDT flat tip was in contact with the superior edgeof the femoral component at three locations (27):

1 Anterior prosthetic surface (LVDT-A);2 Posterior lateral condyle (LVDT-PL);3 Posterior medial condyle (LVDT-PM).

This allowed tracking (Fig. 1):

• inducible (elastic) micromotion;• permanent migration (unrecoverable).

Possible bone and cement swelling effects, derivingfrom lubricant adsorption, were monitored on threeadditional specimens (12). A robust statistical analy-sis was carried out to give an appropriate meaning todefined quality indicators; with the aim of identifyinga good indicator of the risk of failure, different micro-motion indicators were obtained for each specimenand each sensor:

1 Ninety-five percent percentile of the induciblemicromotion all over the test (to discard occasionalhigher peaks, which are not representative of theaverage trend);

2 Ninety-five percent percentile of the induciblemicromotion over the first and the last 100 000cycles;

3 Total permanent migration from the beginning tothe end of the test;

4 Total permanent migration over the first and last100 000 cycles.

To avoid inducible micromotion overestimation(caused by the distance on the bone–prosthesisinterface and bone elastic deformation [19,21]), theinducible trend was used as an indicator that is lessaffected by such elastic deformation; this indicatorquantifies the difference between the induciblemicromotions at the end (last 100 000 cycles) and atthe beginning (first 100 000 cycles) of the test. Apositive value of the inducible trend would indicatean increment of inducible micromotion, associatedwith loosening. At the same time, to prevent devia-tion due to the swelling of the composite bone andcement layer, the ratio between the migration accu-mulated during the first 100 000 cycles and at theend of the whole test was computed (stabilizationratio [19,21]). A value of the stabilization ratio closeto 1 indicates that most of the migration takes placeinitially, while the component tends to stabilize sub-sequently. Confidence intervals were computed foreach measurement location. Sample size was esti-mated for future tests. After test completion, the

FIG. 1. (a) Mean volumetric wear trend (�SD) for all tibial inserts tested under synthetic bone fixation. Represented values are obtainedby dividing the weight loss for the UHMWPE density declared by the manufacturer. (b and c) Inducible and permanent implant-bonemigration in the short (1Mc) and long-term (5Mc).

COMBINED IN VITRO WEAR AND STABILITY ASSESSMENT FOR TKR E179


cement layer was inspected by means of dye pen-etrants (AVIO-B, Rotvel, American Gas & Chemi-cal Company, Northwale, NJ, USA) to detect fatiguecracks. First, the exposed surface was inspected.Then, the internal cement was inspected. Additionaltests were performed to exclude artifacts due toswelling or creep of the composite femur models,and due to the inspection procedure for the cementlayers (12).

RESULTS

Wear testThe wear behavior of the tibial inserts is summa-

rized through the mean trend in Fig. 1a.Visual exami-nation of the three tibial inserts, after 5 million cycles,revealed intra-specimens similarity. On tibial inserts,two different worn areas were observed for themedial and lateral compartment (12); Fig. 2a,b showstwo of the three tested specimens. Predominance ofadhesive wear and absence of micropits confirm find-ings from other authors (4). A comparison betweenin vitro experience and clinical retrievals (Fig. 2c,d)proves acceptable qualitative matching of wear pat-terns (12,28). CV, calculated for the mean wear rate,was 0.29. The CV of the wear rates, reported in pub-lished studies describing wear tests done according toISO 14243-1,3 and performed on similar designs ofmeniscal insert, ranged between 0.07 and 0.40, withan average value of 0.24 (Table 1).

MicromotionsThe observed migration of the implant caused by

swelling of the cement and bone in the soak speci-

mens was in the order of 0.1 mm over 15 days; thisswelling can only affect estimated implant permanentmigrations, recorded over the entire test, and not theestimates of the implant inducible micromotion. Themicromotion patterns were different between mea-surement locations but always highly correlated withthe applied load over the simulated gait cycle, indi-cating a reliable measurement from the LVDTs (12).Inducible micromotions were highly repeatable; forthe entire test, their variations were not statisticallysignificant (t-test, P > 0.1). The trend of migrationover time (Fig. 2) did not indicate any particular signof loosening (migration rate increasing over time).The average migration sign (<0) indicates distal sepa-ration of components (Fig. 1c). A general trend isevident both in the initial phase (subjected to someconditioning processes) that during the steady staterunning:

• medio-lateral differences indicated a tilt of theprosthetic component in the valgus direction in thefrontal plane with respect to the femur;

• antero-posterior differences indicated a tilt inextension direction in a sagittal plane.

The stabilization ratio (12) indicates that most ofthe migration takes place initially with a tendency tostabilization with the increasing of cycle number.Cement mantle was inspected for damage assess-ment; the exposed surface did not show particularevidences, while a consistent number of cracks werevisible, mainly on the anterior part and in the femoralpeg region. The application of the same inspectiveprocedure to the blank specimens guaranteed againstany artificially induced conditioning (12).

FIG. 2. (a and b) Two of the three testedspecimens. (c and d) Two clinical retriev-als that have, respectively, 58- and63-month follow-up. The damaged areaseems independent from size and designof the knee components. This qualitativecomparison proves acceptable matchingof wear patterns. Arrows put in evidencelongitudinal (AP direction) microwear pat-terns accounting for the rolling movementtypical of the knee kinematics.



DISCUSSION

The aim of this study was to verify the feasibility ofan experimental setup able to effectively combinetwo fundamental aspects in the preclinical assess-ment of TKR: the wear of meniscal polyethylene andmicromotion of the femoral component.

The observed wear mechanism was in accordancewith in vivo findings of other authors [13,15,28] andwith two clinical retrievals analyzed with the sameprotocol (12), thus supporting the efficacy of the pre-sented in vitro wear assessment method; in particular,longitudinal wear patterns, in the AP direction, arerelated to the rolling movement typical of the kneekinematics (Fig. 2). Moreover, reliability of thepresent wear test is confirmed by the computed CV(0.29) well within the range of CV reported in othercomparable studies (Table 1). In fact, the low valuefor this indicator accounts for sound and stable wearmeasurement, whose repeatability seems comparableto that achievable with more conventional testingsetups. Medio-lateral tilting of the femoral compo-nents was consistent with larger worn areas locatedposterior on the medial plateau of tested polyethyl-ene specimens. The same worn area location isevident

on retrieved inserts, confirming that the method iscapable of providing good insight on in vivo kneebiomechanics. The amount of motion recorded at thebone-femoral component interface was comparableto that found in vivo (RSA) for well-fixed implants(26,29). The trend of micromotion over the tests didnot show any tendency to loosening (12). Inter-specimen standard deviation for the inducible micro-motions and for the permanent migrations wascalculated (12); in vivo variability, as assessed byRSA, is one order of magnitude larger (25,26,30).Theinducible trend and the stabilization ratio are indica-tors particularly representative of the situation, asthey exhibited a low inter-specimen variability; theyindicate a settling tendency over time (12).The crack-ing pattern observed during cement mantle inspec-tion, was consistent in terms of localization andmorphology of cracks (12). An overall comparisonwith clinical outcomes on well-fixed cemented TKR,reveals good compatibility both in the case of induc-ible micromotions and permanent migrations (26,31);the last parameters accounted for values one orderof magnitude lower than those in loose implants(25,29,31). The largest micromotions were recordednear the posterior condyles, mimicking patternsrecorded in vivo (25). With regard to the cement

TABLE 1. Summary of other wear tests on meniscal polyethylene inserts of similar design. Coefficient of variation wascalculated for each of the following test setups: CVmean (0.24 � 0.1) was then evaluated in order to account for reliability of

the present wear test with the others

Design Test setup (ISO 14243 part 1) Wear rate (mg/Mc) CV

14Stryker, PCS, SXPE Three specimens 0.4 � 0.1 CV = 0.25Thickness = 9 mm 10 million cycles14Stryker, PCS, XLPE Three specimens 3.0 � 0.2 CV = 0.07Thickness = 9 mm 10 million cycles18DePuy, PFC Six specimens 38.3 � 13.1 CV = 0.34Thickness = 10 mm 5 million cycles18DePuy, PFC Sigma Six specimens 21.5 � 5.5 CV = 0.26Thickness = 10 mm 5 million cycles18DePuy, PFC Sigma Rotating Platform Six specimens 12.1 � 3.6 CV = 0.30Thickness = 10 mm 5 million cycles18DePuy, LCS Rotating Platform Six specimens 15.2 � 3.7 CV = 0.24Thickness = 10 mm 5 million cycles15Biomet, Maxim (UHMWPE) Four specimens 14.9 � 3.7 CV = 0.25Thickness = 12 mm 4 million cycles15Biomet, Maxim (GUR 1050) Four specimens 17.6 � 1.7 CV = 0.10Thickness = 12 mm 4 million cycles16Smith and Nephew Profix, CR Three specimens 20.0 � 2.1 CV = 0.10Thickness = 9 mm 5 million cycles16Smith and Nephew, Profix CR Three specimens 11.6 � 1.3 CV = 0.11Thickness = 9 mm 5 million cycles17Zimmer, Durasul UHMWPE (unaged) Eight specimens 8.4 � 1.9 CV = 0.23Thickness = 10 mm 2 million cycles17Zimmer, Durasul UHMWPE (aged) Eight specimens 9.3 � 3.7 CV = 0.41Thickness = 10 mm 2 million cyclesLima, Multigen-Plus Three specimens 2.7 � 0.8 CV = 0.29Thickness = 10 mm 5 million cycles

PCS, posterior cruciate stabilized; SXPE, sequentially enhanced crosslinked polyethylene; XLPE, crosslinked polyethylene; PFC, press fitcondylar; LCS, low contact stress; GUR, granular UHMWPE Ruhrchemie; CR, cruciate retaining.



layer, only a few cracks were observed and they wereconfined in the region around the pegs. The goodsuccess rate reported clinically for the Multigen-Plusdesign (23) is reflected by the low overall looseningtendency exhibited in this in vitro test.

CONCLUSIONS

While these results are very positive, they shouldbe taken with some caution, due to some limitation ofthe present study. First of all, a small sample size wasanalyzed, which suggests prudence in the consider-ations both on wear performance of these tibialinserts and on implant-bone stability. However, itshould be noted that the use of composite femurs forin vitro testing, although still controversial, signifi-cantly reduced inter-specimen variability we wouldhave with cadaveric specimens, thus increasing statis-tical power (18,20,32). This study presented a novelapproach for quantifying wear behavior and femoralcomponent micromotions in TKR, integrating assess-ment of cement damage. Its application to a commer-cial TKR design has demonstrated its feasibility,repeatability, and clinical relevance of the resultsprovided. Similarly to clinical experience (25),femoral components were found to tilt toward valgus(in the frontal plane) and in extension (in a sagittalplane).

Differences in wear patterns as a consequence ofdifferent fixation frames indicate significance of suchtest parameter. Moreover, the contemporary focus ontwo relevant clinical problems with the same experi-mental setup constitutes an innovative and cost-efficient approach for the preclinical validation ofsuch medical devices.

Acknowledgments: The authors would like tothank Luigi Lena for the illustrations and Mara Zaval-loni for the support during the experiments. Lima spa(Italy) promoted and partially funded this work.

REFERENCES

1. Stea S, Bordini B, De Clerico M, Toni A. Report of ortho-paedic prosthetic implantology. Overall data: hip and kneearthroplasty in Emilia Romagna region: Istituti OrtopediciRizzoli. 2007. Available at: https://ripo.cineca.it/pdf/global_report_en_2007_rev1.pdf

2. NIH. NIH Consensus Statement on total knee replacement.NIH Consens State Sci Statements 2003;20:1–34.

3. DoOLU H. Annual report 2006: the Swedish knee arthro-plasty register, 2006. Available at: http://www.knee.nko.se/english/online/thePages/publication.php

4. Schmalzried TP, Callaghan JJ. Wear in total hip and kneereplacements. J Bone Joint Surg Am 1999;81:115–36.

5. Miura H, Higaki H, Nakanishi Y, et al. Prediction of total kneearthroplasty polyethylene wear using the wear index. J Arthro-plasty 2002;17:760–6.

6. Naudie DD, Ammeen DJ, Engh GA, Rorabeck CH. Wear andosteolysis around total knee arthroplasty. J Am Acad OrthopSurg 2007;15:53–64.

7. Martin KJ, Neu CP, Hull ML. An MRI-based method to alignthe compressive loading axis for human cadaveric knees. JBiomech Eng 2007;855–62.

8. Benjamin J. Component alignment in total knee arthroplasty.Instr Course Lect 2006;55:405–12.

9. Romero J, Duronio JF, Sohrabi A, et al. Varus and valgusflexion laxity measurement methods in loaded cadavericknees. Clin Orthop Relat Res 2002;394:243–53.

10. Tetsworth K, Paley D. Malalignment and degenerativearthropathy. Orthop Clin North Am 1994;25:367–77.

11. Werner FW, Ayers DC, Maletsky LP, Rullkoetter PJ. Theeffect of valgus/varus malalignment on load distribution intotal knee replacements. J Biomech 2005;38:349–55.

12. Affatato S, Cristofolini L, Leardini W, et al. A new method ofin-vitro wear assessment of the UHMWPE tibial insert in totalknee replacement. Artif Organs 2008;32:942–8.

13. Tsukamoto R, Chen S, Asano T, et al. Improved wear perfor-mance with crosslinked UHMWPE and zirconia implants inknee simulation. Acta Orthop 2006;77:505–11.

14. Benson LC, DesJardins JD, LaBerge M. Effects of in vitrowear of machined and molded UHMWPE tibial inserts onTKR kinematics. J Biomed Mater Res 2001;58:496–504.

15. Ezzet KA, Hermida JC, Colwell CW Jr., D’Lima DD. Oxidizedzirconium femoral components reduce polyethylene wear in aknee wear simulator. Clin Orthop Relat Res 2004;120–4.

16. Muratoglu OK, Burroughs BR, Bragdon CR, Christensen S,Lozynsky A, Harris WH. Knee simulator wear of polyethylenetibias articulating against explanted rough femoral com-ponents. Clin Orthop Relat Res 2004;108–13.

17. McEwen HM, Barnett PI, Bell CJ, et al. The influence ofdesign, materials and kinematics on the in vitro wear of totalknee replacements. J Biomech 2005;38:357–65.

18. Cristofolini L, Viceconti M, Cappello A, Toni A. Mechanicalvalidation of whole bone composite femur models. J Biomech2005;38:525–35.

19. Cristofolini L, Teutonico AS, Monti L, Cappello A, Toni A.Comparative in vitro study on the long term performance ofcemented hip stems: validation of a protocol to discriminatebetween “good” and “bad” designs. J Biomech 2003;36:1603–15.

20. Cristofolini L, Erani P, Savigni P, Bordini B, Viceconti M.Preclinical assessment of the long-term endurance ofcemented hip stems. Part 2: in-vitro and ex-vivo fatiguedamage of the cement mantle. Proc Inst Mech Eng [H]2007;221:585–99.

21. Britton JR, Prendergast PJ. Preclinical testing of femoralhip components: an experimental investigation with fourprostheses. J Biomech Eng 2005;127:872–80.

22. Maher SA, Prendergast PJ. Discriminating the looseningbehaviour of cemented hip prostheses using measurements ofmigration and inducible displacement. J Biomech 2002;35:257–65.

23. Lizaur A, Miralles-Muñoz FA, Martin-Ballesteros O. Resulta-dos patellares con una pròtesis total de rodilla con com-ponente femoral simétrico: estudio prospectivo con unseguimiento de 3 a 7 años. Revista de Ortopedia y Trauma-tologìa 2006;50:441–6.

24. Barink M, Verdonschot N, de Waal Malefijt M. A differentfixation of the femoral component in total knee arthroplastymay lead to preservation of femoral bone stock. Proc InstMech Eng [H] 2003;217:325–32.

25. Nilsson KG, Karrholm J, Linder L. Femoral component migra-tion in total knee arthroplasty: randomized study comparingcemented and uncemented fixation of the Miller-Galante Idesign. J Orthop Res 1995;13:347–56.

26. Ryd L. Roentgen stereophotogrammetric analysis of pros-thetic fixation in the hip and knee joint. Clin Orthop Relat Res1992;56–65.



27. Cristofolini L, Affatato S, Erani P, Leardini W, Tigani D, Vice-conti M, Viceconti M. Long-term implant-bone fixation of thefemoral component in total knee replacement . Proc Inst MechEng [H] 2008;222:319–31.

28. Harman MK, Markovich GD, Banks SA, Hodge WA. Wearpatterns on tibial plateaus from varus and valgus osteoarthriticknees. Clin Orthop Relat Res 1998;149–58.

29. Brostrom LA, Goldie I, Selvik G. Micromotion of the totalknee. Acta Orthop Scand 1989;60:443–5.

30. Karrholm J, Gill RH, Valstar ER. The history and future ofradiostereometric analysis. Clin Orthop Relat Res 2006;448:10–21.

31. Green DL, Bahniuk E, Liebelt RA, Fender E, Mirkov P.Biplane radiographic measurements of reversible displace-ment (including clinical loosening) and migration of total jointreplacements. J Bone Joint Surg Am 1983;65:1134–43.

32. Papini M, Zdero R, Schemitsch EH, Zalzal P. The biomechan-ics of human femurs in axial and torsional loading: comparisonof finite element analysis, human cadaveric femurs, and syn-thetic femurs. J Biomech Eng 2007;129:12–9.



combined wear behavior and long-term implant-bone fixation of total knee replacement: a novel in...

Documents