Migration and cyclic motion of a new short-stemmed hip prosthesis – a biomechanical in vitro study

Migration and cyclic motion of a new short-stemmed hip prosthesis – a biomechanical in vitro study

Clinical Biomechanics 21 (2006) 834–840 www.elsevier.com/locate/clinbiomech Migration and cyclic motion of a new short-stemmed hip prosthesis – a bio...

332KB Sizes 1 Downloads 44 Views

Clinical Biomechanics 21 (2006) 834–840 www.elsevier.com/locate/clinbiomech

Migration and cyclic motion of a new short-stemmed hip prosthesis – a biomechanical in vitro study F.M. Westphal a, N. Bishop a, M. Honl b, E. Hille c, K. Pu¨schel d, M.M. Morlock a

a,*

Hamburg University of Technology, Biomechanics Section, Denickestrasse 15, D-21073 Hamburg, Germany b Rush University Medical Center, 1653W Congress Pkwy, Chicago, IL 60612, USA c General Hospital Eilbek, Friedrichsberger Str. 60, 22081 Hamburg, Germany d UKE University Hospital Hamburg, Martinistr. 52, 20246 Hamburg, Germany Received 31 October 2005; accepted 4 April 2006

Abstract Background. Uncemented, short-stemmed hip prostheses have been developed to reduce the risk of stress shielding and to preserve femural bone stock. The long-term success of these implants is yet uncertain. Prerequisite for osseointegration is sufficient primary stability. In this study the cyclic motion and migration patterns of a new short-stemmed hip implant were compared with those for two clinically successful shaft prostheses. Methods. The prostheses were implanted in paired fresh human femura and loaded dynamically (gait cycle) with increasing load (max 2100 N) up to 15,000 cycles. Relative displacements between prosthesis and bone were recorded using a 3D-video analysis system. Findings. The short stem displayed a biphasic migration pattern with stabilisation at maximum load. Initial migration was predominantly into varus and was greater than that for the shaft prostheses. Failure occurred in cases of poor bone quality and malpositioning. Cyclic motion of the short prosthesis was less than that for the shaft prostheses. Surface finish showed no effect. System stiffness for the new stem was lower than for the shaft prostheses. Interpretation. The new stem tended to migrate initially more than the shaft prostheses, but stabilised when cortical contact was achieved or the cancellous bone was compacted sufficiently. Bone quality and correct positioning were important factors for the short stem. The lower cyclic motion of the new stem should be favourable for bony ingrowth. The lower system bending stiffness with the new implant indicated a more physiological loading of the bone and should thereby reduce the effects of stress shielding. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: Total hip replacement; Short-stemmed implant; Primary stability

1. Introduction Uncemented total hip prostheses are used predominantly in younger patients. In recent years implant design changes have been made to achieve more proximal load transfer into the femur to reduce proximal stress shielding and thus preserve bone stock for potential revision surgery. The short and midterm results for these prostheses when implanted by the design surgeons are promising, with sur*

Corresponding author. E-mail addresses: [email protected], [email protected] (M.M. Morlock). 0268-0033/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.clinbiomech.2006.04.004

vival rates above 97% (Buergi et al., 2005; Morrey et al., 2000; Thomas et al., 2004). However, other authors have reported less satisfactory midterm results with survival rates of 90.5% after 8 years. (Ishaque et al., 2004) and 85.2% after 26.1 months for patients with polyarthritis (Fink et al., 2000). Therefore, the long-term performance of short-stemmed prostheses remains uncertain. The success of cementless total hip arthroplasty relies on osseointegration of the implants. Prerequisite is primary stability, which can be achieved by the fixation principle of ‘‘press-fitting’’ (Morscher et al., 2002; Henry et al., 1993). Clinical studies investigating the migrational behaviour of femoral components, using ‘radio stereometric analysis’

F.M. Westphal et al. / Clinical Biomechanics 21 (2006) 834–840

(RSA) (Freeman and Plante-Bordeneuve, 1994; Donnelly et al., 1997) or ‘single image X-ray analysis – femoral component analysis’ (EBRA-FCA) (Krismer et al., 1999) have shown that the failure rate of uncemented stems correlates with migration of more than 1.2 mm/year (Freeman and Plante-Bordeneuve, 1994) and 1.5 mm in the first two years after operation (Krismer et al., 1999). Furthermore, in vitro tests have shown that bony ingrowth of an implant is only possible with sufficiently low relative interface motions (<150 lm) (Shirazi-Adl et al., 1993; Jasty et al., 1997). A new short-stemmed prosthesis, PROXIMATM (DePuy; Leeds, UK), has been designed with a close anatomical fit to the proximal cortex, with the aim of maximising primary stability, particularly in torsion. It is also proposed that the shorter shaft leads to more physiological loading of the femur, thereby limiting potential bone resorption due to stress shielding. The purpose of the present study was to compare the cyclic relative interface motion and migration of the PROXIMATM stem with that of clinically successful hip prostheses in a pre-clinical in vitro setting.

Table 1 The input variables; the factor of interest in each test series is indicated by italic writing Test series

Prosthesis type

Head offset

Surface finish

Specimen age [yrs]

Specimen sex

1

PROXIMA (HO) IPS (SO)

+5 mm +17 mm

Standard Standard

49 (SD 8)

Male

2

PROXIMA (HO) PROXIMA (HO)

+5 mm +5 mm

Standard ZTT

49 (SD 4)

Male

3

PROXIMA (HO) SUMMIT (HO)

+5 mm +5 mm

ZTT ZTT

58 (SD 4)

Male

HO: high offset version, SO: standard offset version.

835

2. Methods Three test series were performed, labelled 1–3 in sequential order. In each test the prostheses were implanted into 5 human femur pairs. The bone specimen age (39–62 years) was chosen to represent the predominantly lower age of patients who receive short-stemmed implants (Table 1). After retrieval, the bones were stored at 32 °C and thawed to room temperature prior to testing. Tests 1 and 3 compared the short-stemmed PROXIMATM implant with the IPSTM and SUMMITTM (DePuy, Leeds, UK) cementless shaft prostheses, respectively. In test series 1 different head offsets were used to achieve the same total offset for both implants to be compared. Test 2 compared two surface finishes of the PROXIMATM prosthesis (Fig. 1): the ZTTTM surface (which is used in the SUMMITTM prosthesis) and the standard PorocoatTM surface (used in the IPSTM prosthesis). 2.1. Test set-up The test procedure was based on the ISO norm ‘‘Application of a cyclic axial and torsional load to the head of a stemmed femoral component’’ (ISO 7206-4). Loads were applied to simulate physiological conditions during a normal gait cycle, according to those measured in vivo by Bergmann et al. (1993). Implantation was carried out by a medical doctor according to manufacturer’s specifications with repeated X-ray control for correct positioning. The femura were then cut 18 cm distal to the Lesser Trochanter. The prosthesis–bone system was positioned in a jig so that the femural shaft could be reproducibly aligned in 10° adduction and 10° flexion. In this orientation 10 cm of the femural shaft was embedded in a metal tube using RenCastÒ FC 53 (Huntsman, Cambridge, UK). To

Fig. 1. The prostheses investigated: (a) from left to right: PROXIMATM, SUMMITTM, IPSTM; (b) two surface finishes. Top: standard PorocoatTM surface, bottom: ZTTTM surface.

836

F.M. Westphal et al. / Clinical Biomechanics 21 (2006) 834–840

eliminate lateral forces, this tube was mounted rigidly to a horizontal low friction XY-table on the load cell of a servohydraulic materials testing machine (MTS Bionix 858.2, Eden Prairie, USA). Load was applied to the prosthesis head under force control using a steel tube with an inner diameter of 8 mm to locate the B32 mm spherical head. Load was applied dynamically at a frequency of 2 Hz, increasing stepwise up to 3000 cycles and then continuing to 15,000 cycles or failure. Initially, 20 conditioning cycles were applied with an axial load cycling between 50 and 200 N. From cycle 20 to cycle 1000 a load of 50–800 N was applied, from cycle 1001 to cycle 2000 a load of 50– 1200 N, from cycle 2001 to cycle 3000 a load of 50– 1600 N and from cycle 3001 until the end of the test a load of 50–2100 N was applied. Motion data were recorded for the first 100 cycles and then for 15 cycles each at 500, 1000, 1500, 2000 and every subsequent 1000 cycles. 2.2. Assessment of motion data Motion data were recorded using a 4 camera video system with retroflective markers, at a sampling frequency of 120 Hz (VICON 470, Vicon-Peak, Oxford, UK). The static accuracy of this system for the calibrated volume used in this study was 100 lm. To measure displacement magnitudes a marker set (cluster of 4 markers) was attached to bone and implant, respectively. Clusters were stiff, light weight and rigidly mounted, to minimise dynamic errors. One marker set was fixed to the prosthetic head by a clamp (‘A’ in Fig. 2) and the other was screwed onto the Greater Trochanter (‘B’ in Fig. 2). Positioning of the marker sets on each femur was standardised based on anatomical landmarks. Relative displacements between prosthesis and bone marker set coordinate systems were referred to a new local head coordinate system. This was defined with its origin at the centre of the prosthetic head, and registered by recording the positions of a line of markers at a known distance from the head centre before the start of each test (‘D’ in Fig. 2). Temporary markers were also placed on the femur and recorded to define the orientation of the local orthogonal coordinate system, which was aligned with its z-axis parallel to the femural axis and with the y-axis in the plane of the femoral neck (‘C’ in Fig. 2). As output variables three relative bone–prosthesis translations (x = anterior–posterior, y = medial–lateral, z = superior–inferior) and three relative rotations (Cardan angles around x (varus–valgus), y (anterior–posterior), z (retrotorsion–antetorsion) axes), respectively, were computed. All movements were reported in the local head coordinate system relative to the unloaded orientation. The resultant ‘total’ translation and rotation magnitudes were also calculated. Two parameters, ‘migration’ and ‘cyclic motion’, were evaluated. ‘Migration’ was defined as the permanent displacement of the implant with respect to the bone relative to the initial unloaded situation. ‘Cyclic motion’ was defined as the amplitude of the cyclic relative stem–bone displacement for a load cycle.

Fig. 2. Experimental test set-up. (A) Marker set representing the prosthesis system. (B) marker set representing the bone system. (C) triplet defining the output coordinate system direction components. (D) marker pair defining the centre of the output coordinate system. (1) Loading cylinder on the axial testing machine.

2.3. System stiffness The system stiffness was evaluated for cycles 14,990– 15,015 at the maximal load of 2100 N. Stiffness was defined as the quotient of the applied axial load and the measured axial displacement of the loading cylinder of the material testing machine. This parameter was used to compare differences in overall deformation of the femur-prosthesis system. 2.4. Statistics Statistical analysis was performed using one way ANOVA (SPSS Inc., Chicago, USA). 3. Results 3.1. Test series 1 (PROXIMA vs. IPS) Both implants showed a biphasic pattern of cyclic motion with increasing load, consisting of a relatively steep initial increase, followed by a relatively low displacement phase,

F.M. Westphal et al. / Clinical Biomechanics 21 (2006) 834–840

1.0 Proxima Std

(a)

IPS

Translation [mm]

0.8 0.6 0.4 0.2 0.0 0

3000

6000

9000 cycle [#]

12000

15000

1.0 Proxima Std

(b)

Proxima ZTT

Translation [mm]

0.8 0.6 0.4 0.2 0.0 0

3000

6000 9000 cycle [#]

12000

15000

1.0 Proxima ZTT

(c)

SUMMIT

Translation [mm]

0.8 0.6 0.4 0.2 0.0 0

3000

6000 9000 cycle [#]

12000

15000

Fig. 3. Total translational cyclic motion (mean and SD) over 15,000 cycles. (a) Test series 1 (PROXIMATM vs. IPSTM). (b) Test series 2 (ZTTTM vs. Standard surface). (c) Test series 3 (PROXIMATM vs. SUMMITTM).

termed ‘stabilised’ (Fig. 3a). The cyclic motion magnitude was lower for the PROXIMATM than for the IPSTM throughout the test, which was significant at 15,000 cycles (0.29 (SD 0.06) mm vs. 0.66 (SD 0.11) mm, respectively, P < 0.001; Fig. 3a). ‘Stabilisation’ was observed for both stem types at the maximum load of 2100 N at around 4000 cycles. A difference in migration behaviour was also observed between the two implants. The total migration magnitudes were not-significantly different between PROXIMATM and IPSTM (5.22 (SD 3.5) mm vs. 4.71 (SD 0.6) mm, respectively, at 15,000 cycles, P = 0.808, Fig. 4a). Similarly to the cyclic relative motion, the total migration pattern of the PROXIMATM was biphasic. ‘Stabilisation’ of migration was

837

observed for all PROXIMATM implantations at around 4000 cycles. The IPSTM showed a more steady increase, without a biphasic migration pattern. The major difference observed was the displacement direction, which was predominantly inferior–medial translation with varus rotation for the PROXIMATM, and predominantly inferior–posterior translation with rotation into retrotorsion for the IPSTM (Fig. 4a). Two PROXIMATM implantations failed. This occurred early in the test period and no ‘stabilisation’ was observed in the migration behaviour. One medial cutout (cycle 1090) apparently occurred due to poor bone quality, with very soft cancellous bone stock observed. In the second case (failure at cycle 8310) broaching had been performed suboptimally with a dorsal gap remaining visible after impaction of the implant. The IPSTM also resulted in two early failures: one due to poor bone quality (cycle 1050) and the other (cycle 7933) due to a crack in the Calcar, which occurred during impaction. The system stiffness for the PROXIMATM was slightly lower than for the IPSTM (1486 (SD 484) N/mm, 1604 (SD 197) N/mm, respectively, P = 0.376). 3.2. Test series 2 (PROXIMATM: Standard vs. ZTTTM surfaces) No significant difference or trend was observed for either migration or cyclic motion between the two surface coatings. The cyclic motion ‘stabilised’ for the Standard and for the ZTTTM surfaces after about 4000 cycles (Fig. 3b). At 15,000 cycles the magnitudes were 0.39 (SD 0.10) mm and 0.43 (SD 0.08) mm, respectively, (P = 0.596). The migrational biphasic patterns and magnitudes, as well as the direction of movement predominantly into varus, were similar (Fig. 4b). One early failure occurred for each surface finish. In the group with the standard surface finish suboptimal implantation led to medial cut-out in one case (cycle 4033). The test which failed in the ZTTTM group had to be restarted four times during the first 500 cycles due to technical problems with the testing apparatus (cycle 3045). This may have caused additional stresses which led to failure. The system stiffness for the PROXIMATM with the ZTTTM surface was lower than that for the Standard surface, but the difference was not significant (1656 (SD 140) N/mm, 2440 (SD 1009) N/mm, respectively, P = 0.420). 3.3. Test series 3 (PROXIMATM vs. SUMMITTM) The cyclic motion for both prostheses displayed similar patterns but showed a trend for a slightly lower magnitude at 15,000 cycles for the PROXIMATM than for the SUMMITTM (0.33 (SD 0.06) mm vs. 0.40 (SD 0.04) mm, respectively, P = 0.079, Fig. 3c). The migration magnitude was significantly higher for the PROXIMATM than for the SUMMITTM (at 15,000 cycles: 7.88 (SD 2.5) mm vs. 4.33 (SD 1.9) mm, respectively, P = 0.035, Fig. 4c). The migration patterns were comparable to those observed for Test

838

F.M. Westphal et al. / Clinical Biomechanics 21 (2006) 834–840

Rotation [°]

Translation [mm] anterior medial superior total

(a) 12 10 8 6 4 2 0 -2 -4 -6 -8

valgus retrotorsion anterior total PROXIMA Std vs.

IPS

10 8 6 4 2 0 -2 -4 -6 -8

0

3000

6000

9000

(b)

12000

0

15000

PROXIMA Std

vs.

3000

6000

9000

12000

15000

12000

15000

PROXIMA ZTT

10

12 10 8 6 4 2 0 -2 -4 -6 -8

8 6 4 2 0 -2 -4 -6 -8

0

3000

6000

9000

(c)

12000

0

15000

PROXIMA ZTT

vs.

3000

6000

9000

SUMMIT

10

12 10 8 6 4 2 0 -2 -4 -6 -8

8 6 4 2 0 -2 -4 -6 -8

0

3000

6000

9000

12000

15000

Cycles [#]

0

3000

6000

9000

12000

15000

Cycles [#]

Fig. 4. Migration components for the three experimental tests over 15,000 cycles. Mean values plotted. Total migration is shown bold with standard deviation. Left hand graphs: translation components; right hand graphs: rotation components. (a) Test series 1 (PROXIMATM vs. IPSTM). (b) Test series 2 (Standard vs. ZTTTM). (c) Test series 3 (ZTTTM vs. SUMMITTM).

series 1. The shaft-prosthesis (SUMMITTM) migrated into retrotorsion, whereas the PROXIMATM migrated predominantly into varus (Fig. 4-3). Neither the SUMMITTM nor the PROXIMATM failed during the test. The system stiffness for the PROXIMATM implant was significantly lower than for the SUMMITTM shaft prosthesis (2117 (SD 418) N/mm vs. 3530 (SD 669) N/mm, respectively, P = 0.015). 4. Discussion In this study the migration and cyclic motion were compared between a new short-stemmed hip prosthesis and two

clinically successful shaft prostheses in order to estimate the primary stability of the new stem. Primary stability is known to be essential to the long-term success of uncemented implants by allowing bony ingrowth and secondary stability. It is therefore assumed that the clinically successful stems used in this study can be taken as a reference with regard to the relative motion of the stem and bone measured in this study. It should be noted that the test set-up described simulates a worst-case loading scenario. Since no muscle forces or biological effects, such as trabecular regeneration or bony ingrowth, were simulated it could be expected that clinical migration of the implants would be less pronounced than

F.M. Westphal et al. / Clinical Biomechanics 21 (2006) 834–840

that observed in this study. This is supported by the fact that magnitudes measured in this study for clinically successful stems are much higher than migration magnitudes that have been found to correlate with failure clinically. A subsidence of 1.5 mm has been found to correlate with implant-loosening after 5 years with an accuracy of 91% (Krismer et al., 1999). Published data show though, that the IPSTM can achieve a survival rate of 100% after 6.3 years (Kim, 2002). Despite this discrepancy between experimental and clinical results for established uncemented shaft stems, the tendencies and differences in the parameters measured allow a relative comparison of the input variables with the new short-stemmed prosthesis. Quantitative differences in migration of the same implant between the test series are mostly related to differences in bone quality. Therefore inter-test comparisons should be interpretated with caution. Nevertheless, the similar qualitative results of the PROXIMATM prostheses between all test series and similar quantitative results within each test series indicate that reproducibility can be achieved with this test set-up.

839

significantly higher migration magnitudes compared to the SUMMITTM emphasise that short-stemmed implants such as the PROXIMATM require good cancellous bone stock in the Calcar region, as no additional distal stem fixation, as for the SUMMITTM, is available. In contrast, the IPSTM has a similar proximal geometry to the PROXIMATM and, in comparison with the SUMMITTM, a smaller stem diameter and therefore less stability distally. Consequently the migration magnitudes are only slightly smaller for the IPSTM than for the PROXIMATM. Failure of the PROXIMATM occurred due to poor quality bone stock or suboptimal implantation. Both factors led to medial cutout, as the stem rotated into varus. Malpositioning as well as undersizing were important implantation criteria, which apparently led to failure. Such failures are to be expected, since an in vitro test is dominated by mechanical factors: the higher lever arm due to varus and/or medialised positioning causes higher stresses in the Calcar bone, leading to a higher risk of failure. Therefore, both surgical technique and careful patient selection would appear to be crucial factors for the PROXIMATM prosthesis.

4.1. Cyclic motion 4.3. Surface finish The PROXIMATM prosthesis resulted in a cyclic motion of 290–430 lm at 15,000 cycles in all test series. The surface finish had no significant effect. Lower cyclic motions were observed for the PROXIMATM compared to the shaft prostheses. The significantly lower cyclic motions for the shortstemmed implant in comparison with the IPSTM prosthesis indicate that it seats much better in the bone during loading. The shaft of the IPSTM prosthesis is smaller than the diameter of the canal but may impinge during loading as the bone bends. This could cause toggling of the IPSTM proximally. In contrast to the IPSTM, the SUMMITTM showed only slightly higher cyclic motion than the PROXIMATM, which indicates good seating for both implants. The SUMMITTM prosthesis has a larger diameter distal stem than the IPSTM, which forms a tight press fit in the canal, leading to a good fit. However, stress shielding is potentially increased due to the stiffer system. Lower cyclic motions potentially lead to better bony ingrowth (Jasty et al., 1997) and, since the stemmed implants tested in this study are successful clinically, the PROXIMATM would indicate good possibilities for longterm consolidation with the bone. 4.2. Migration The PROXIMATM prosthesis tended to migrate into varus. A relatively large initial migration phase was characteristically observed, followed by a lower magnitude ‘stable’ phase. The initially high migration can be attributed to gradual compaction of the cancellous bone in the proximal femur. Based on X-ray observation before and after loading, stabilisation apparently occurred once the implant had made contact with the lateral cortex, or the cancellous bone had become sufficiently compacted in this region. The

The surface finish did not influence migration or cyclic motion in this test set-up. As this test only simulates the biomechanical performance, the biological effect of these two surfaces with regard to bony ingrowth cannot be assessed. However, the same surfaces have been used on the clinically successful IPSTM and SUMMITTM implants and can therefore be associated with osseointegration in press-fit systems. 4.4. Loading of the bone The system stiffness was significantly lower for the PROXIMATM than for the SUMMITTM prosthesis, which indicates that loading of the bone may be more physiological using a proximal stem. Only slightly lower stiffness was observed for PROXIMATM than for the IPSTM, which does not have such rigid distal fixation as the SUMMITTM prosthesis. Since the system stiffness was related mainly to bending in the frontal plane the PROXIMATM should allow the bone to bend more than for the more rigid stemmed prostheses. In turn, this indicates that the PROXIMATM subjects the bone to less stress shielding than distally rigid stemmed implants. Stress shielding remains a clinical problem, whereby the under-stressed proximal bone resorbs, leaving compromised bone-stock for revision procedures (Engh et al., 2003; Turner et al., 2005). Thus, shortstemmed prostheses, such as the PROXIMATM, indicate potential to minimise this effect. 4.5. Limitations The relative motion magnitudes reported cannot be directly compared to in vivo magnitudes or relative

840

F.M. Westphal et al. / Clinical Biomechanics 21 (2006) 834–840

interface motion values reported in the literature. This is due to the measurement of displacements between two discrete points in this study, both of them remote from the interface. The relative motion between the externally mounted marker sets includes elastic and plastic deformation of the bone. This would tend to result in overestimation of displacement magnitudes, such that smaller and safer magnitudes would be expected in vivo. 5. Conclusion The cyclic motion of the PROXIMATM prosthesis is lower than that for clinically successful stemmed implants; bony ingrowth and long-term stability should therefore be expected. The PROXIMATM implant tended to migrate more than the longer shaft prostheses initially, in particular in the varus direction, but then stabilised after about 4000 loading cycles. Lack of stabilisation due to insufficient medial and lateral support, either by suboptimal alignment or poor bone quality, led to medial cutout. Careful surgical technique and patient selection are therefore necessary for the PROXIMATM implant. The lower bending stiffness of the PROXIMATM indicates more physiological load transfer, which may reduce the risk of stress shielding. Acknowledgements DePuy, Leeds, UK. Ministry of Science and Health, Hamburg, Germany for financial support. References Bergmann, G., Graichen, F., Rohlmann, A., 1993. Hip joint loading during walking and running, measured in two patients. J. Biomech. 26, 969–990. Buergi, M.L., Stoffel, K.K., Jacob, H.A., Bereiter, H.H., 2005. Radiological findings and clinical results of 102 thrust-plate femoral hip prostheses: a follow-up of 2 to 8 years. J. Arthroplasty 20, 108–117.

Donnelly, W.J., Kobayashi, A., Freeman, M.A., Chin, T.W., Yeo, H., West, M., Scott, G., 1997. Radiological and survival comparison of four methods of fixation of a proximal femoral stem. J. Bone Joint Surg. Br. 79, 351–360. Engh Jr., C.A., Young, A.M., Engh Sr., C.A., Hopper Jr., R.H., 2003. Clinical consequences of stress shielding after porous-coated total hip arthroplasty. Clin. Orthop. Relat. Res., 157–163. Fink, B., Siegmuller, C., Schneider, T., Conrad, S., Schmielau, G., Ruther, W., 2000. Short- and medium-term results of the thrust plate prosthesis in patients with polyarthritis. Arch. Orthop. Trauma Surg. 120, 294– 298. Freeman, M.A., Plante-Bordeneuve, P., 1994. Early migration and late aseptic failure of proximal femoral prostheses. J. Bone Joint Surg. Br. 76, 432–438. Henry, J.D., Reilly, D., Poss, R., 1993. Two- to four-year experience with cemented, press-fit, and porous coated applications of the profile total hip system. Acta Orthop. Belg. 59 (Suppl. 1), 190–194. Ishaque, B.A., Wienbeck, S., Sturz, H., 2004. Midterm results and revisions of the thrust plate prosthesis (TPP). Z. Orthop. Ihre Grenzgeb. 142, 25–32. Jasty, M., Bragdon, C., Burke, D., O’Connor, D., Lowenstein, J., Harris, W.H., 1997. In vivo skeletal responses to porous-surfaced implants subjected to small induced motions. J. Bone Joint Surg. Am. 79, 707– 714. Kim, Y.H., 2002. Cementless total hip arthroplasty with a close proximal fit and short tapered distal stem (third-generation) prosthesis. J. Arthroplasty 17, 841–850. Krismer, M., Biedermann, R., Stockl, B., Fischer, M., Bauer, R., Haid, C., 1999. The prediction of failure of the stem in THR by measurement of early migration using EBRA-FCA. Einzel-Bild-Roentgen-Analysefemoral component analysis. J. Bone Joint Surg. Br. 81, 273–280. Morrey, B.F., Adams, R.A., Kessler, M., 2000. A conservative femoral replacement for total hip arthroplasty. A prospective study. J. Bone Joint Surg. Br. 82, 952–958. Morscher, E.W., Widmer, K.H., Bereiter, H., Elke, R., Schenk, R., 2002. Cementless socket fixation based on the ‘‘press-fit’’ concept in total hip joint arthroplasty. Acta Chir Orthop. Traumatol. Cech. 69, 8–15. Shirazi-Adl, A., Dammak, M., Paiement, G., 1993. Experimental determination of friction characteristics at the trabecular bone/porouscoated metal interface in cementless implants. J. Biomed. Mater. Res. 27, 167–175. Thomas, W., Lucente, L., Mantegna, N., Grundei, H., 2004. ESKA (CUT) endoprosthesis. Orthopade 33, 1243–1248. Turner, A.W., Gillies, R.M., Sekel, R., Morris, P., Bruce, W., Walsh, W.R., 2005. Computational bone remodelling simulations and comparisons with DEXA results. J. Orthop. Res. 23, 705–712.