- Email: [email protected]
Effect of undersizing on the long-term stability of the Exeter hip stem: A comparative in vitro study
Clinical Biomechanics 25 (2010) 899–908
Contents lists available at ScienceDirect
Clinical Biomechanics j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / c l i n b i o m e c h
Effect of undersizing on the long-term stability of the Exeter hip stem: A comparative in vitro study Luca Cristofolini a,b,⁎, Paolo Erani a, Ewa Bialoblocka-Juszczyk a,b, Hirotsugu Ohashi c, Satoshi Iida d, Izumi Minato e, Marco Viceconti a a
Laboratorio di Tecnologia Medica, Istituto Ortopedico Rizzoli, Via di Barbiano 1/10, 40136 Bologna, Italy DIEM, Università degli Studi di Bologna, Viale Risorgimento 2, 40136 Bologna, Italy Department of Orthopaedic Surgery, Saiseikai Nakatsu Hospital, 2-10-39, Shibata, Kitaku, Osaka 530-0012, Japan d Department of Orthopaedic Surgery, Matsudo City Hospital, Kamihongou 4005, Matsudo, Chiba 271-0064, Japan e Department of Orthopaedic Surgery, Niigata Rinko Hospital, Momoyamacho 1-114-3, Niigata City, Niigata 950-0051, Japan b c
a r t i c l e
i n f o
Article history: Received 8 March 2010 Accepted 5 July 2010 Keywords: Polished cemented hip prosthesis Stem undersizing Long-term implant-bone stability Fatigue failure Inducible micromotion Permanent migration In vitro testing
a b s t r a c t Background: Even for clinically successful hip stems such as the Exeter-V40 occasional failures are reported. It has been reported that sub-optimal pre-operative planning, leading to implant undersizing and/or thin cement mantle, can explain such failures. The scope of this study was to investigate whether stem undersizing and a thin cement mantle are sufficient to cause implant loosening. Methods: A comparative in vitro study was designed to compare hip implants prepared with optimal and smaller than optimal stem size. Exeter-V40, a highly polished cemented hip stem, was used in both cases. Tests were carried out simulating 24 years of activity of active hip patients. A multifaceted approach was taken: inducible and permanent micromotions were recorded throughout the test; cement micro-cracks were quantified using dye penetrants and statistically analyzed. Findings: The implants with an optimal stem size withstood the entire mechanical test, with low and stable inducible micromotions and permanent migrations during the test, and with moderate fatigue damage in the cement mantle after test completion. Conversely, the undersized specimens showed large and increasing micromotions, and failed after few loading cycles, because of macroscopic cracks in the proximal part of the cement mantle. While results for the optimal stem size are typical for stable hip stems, those for the undersize stem indicate a critical scenario. Interpretation: These results confirm that even a clinically successful hip prosthesis such as the Exeter-V40 is prone to early loosening if a stem smaller than the optimal size is implanted. © 2010 Elsevier Ltd. All rights reserved.
1. Introduction The Exeter is one of the most frequently used hip stems, accounting for 21.8% of the cemented stems implanted in Europe (Scheerlinck et al., 2004). It is a clinically successful stem, with a survival rate of 98.5% at 7 years and over 95% at 10 years (Kärrholm et al., 2008; Stea et al., 2008). While the average survival rate in itself is excellent, there are hospitals/divisions where higher failure rates are reported (Kärrholm et al., 2008; Stea et al., 2008). Implant failure may be due to a number of factors, related to the patient, the device, and the surgical technique. There are two causes of failure of cemented hip stems by aseptic loosening, which are related to the surgical technique: (i) sub-optimal pre-operative planning (Ebramzadeh et al., 2004), and (ii) excessively thin cement mantle (Ebramzadeh et al., 1994). ⁎ Corresponding author. Laboratorio di Tecnologia Medica, Istituto Ortopedico Rizzoli, Via di Barbiano, 1/10, 40136 Bologna, Italy. E-mail address: [email protected] (L. Cristofolini). 0268-0033/$ – see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.clinbiomech.2010.07.003
Incorrect choice of the stem size is a relatively common problem related to pre-operative planning, both in cemented and cementless hip stems (Barrack, 2000; Giannikas et al., 2002; Viceconti et al., 2003). It has been reported that stem undersizing by two stem sizes occurs in 10% of the cases with anatomical uncemented hip stems, when pre-operative templating is based on standard radiographs (Viceconti et al., 2003). Stem undersizing in cemented arthroplasty has been associated to excessive stress in the cement mantle (Harrington et al., 2002; Janssen et al., 2009), which can lead to early aseptic loosening (Verdonschot, 1996). It has been suggested that the use of pre-operative planning software enabling a full threedimensional virtual visualisation would reduce the incidence of such problems (Lattanzi et al., 2003; Viceconti et al., 2003). Insufficient cement mantle thickness is primarily caused by insufficient reaming of the femoral canal (Scheerlinck et al., 2006). In fact, it has often been shown that a mantle thicker than 2–3 mm reduces cement stress and micromotions (Ramaniraka et al., 2000). Conversely, more cracks are found both ex vivo (Kawate et al., 1998) and in vitro (Mann et al., 2004) where the cement mantle is thin. In
900
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
fact, even in the case when cracks grow at the same rate across the cement mantle, a reduced thickness is expected to decrease the time needed for a crack to break through the entire mantle thickness (Hertzler et al., 2002). In vitro testing is able to discriminate between successful and critical hip stems (Maher and Prendergast, 2002; Britton and Prendergast, 2005; Cristofolini et al., 2007a,c; Zhang et al., 2008). In vitro testing has also been able to explain early failure of a specific cemented stem in relation with excessively thin cement mantles (Cristofolini et al., 2007b). The scope of this work was to assess if, given a clinically successful hip stem, sub-optimal planning (stem undersizing) and sub-optimal implantation (thin cement mantle) is sufficient to cause implant failure. In order to test this hypothesis, composite femurs implanted with Exeter-V40 stems of different sizes were tested in vitro to investigate implant micromotions and cement fatigue damage.
2. Methods A mechanical testing procedure that was extensively validated (also in comparison against ex vivo retrievals) was recently proposed for quantifying long-term hip implant damage both in terms of implant-bone micromotions and of cement fatigue damage (Cristotofolini et al., 2007a,c; 2003). As revision rates are extremely low, failures are most likely to occur in the most demanding patients for implant survival, i.e. the young, active and heavy patients that are most likely to apply higher and most frequent loads (Dorey and Amstutz, 2002; Kilgus et al., 1991). A loading profile has recently been proposed that enables simulating 24 years of activity of a very demanding patient (Cristofolini et al., 2007c; 2002).
2.1. Specimens The present study was carried out on a total of ten Exeter hip stems (Exeter-V40, Howmedica, Mahwah, NJ, USA), which is a Cr–Co–Mo highly polished cemented stem. They were implanted by a pool of skilled surgeons in composite femurs (Mod. 3103, Pacific Research Labs, Vashon Island, WA), using the instrumentation recommended by the manufacturer. An anatomic reference system was established (Cristofolini, 1997; Ruff and Hayes, 1983) to assist in maintaining consistent implantation, and increase repeatability during specimen preparation and testing. To enable reproducible preparation, bone cement (Simplex-P, Stryker-Howmedica, Mahwah, NJ, USA) was mixed using a sealed mixing device (Summit Medical Syringe Type, Summit Medical Group, Bourton-on-the-Water, UK) at 23–25 °C, 40– 55% relative humidity. To replicate the worst-case clinical scenario in which the hip stems must still function well, mixing was performed without application of vacuum (which is one of the options indicated by the Manufacturer for Simplex-P). In fact, hand-mixing is still used in the clinical practice between 6% (Gheduzzi et al., 2004) and 50% of cases (Breusch et al., 2000; Stea et al., 2008). Cement was injected in a
retrograde fashion prior to stem insertion. The stems were provided with the standard Exeter distal centralizer. Ten femurs were implanted as follows (specimens were X-rayed to document alignment and mantle thickness): – Optimal specimens: four femurs were prepared with the optimal stem size (Exeter-V40 femoral stem: 37.5 mm offset, No. 2) based on pre-operative templating following the manufacturer's instructions. – Undersized specimens: four femurs were rasped to a smaller size than the Optimal one (Exeter-V40 femoral stem: 35.5 mm offset). A stem two sizes smaller than the Optimal one was chosen, as errors of between one and three sizes are commonly reported for hip stems (Viceconti et al., 2003). Such stem was slightly shorter, and narrower both in the frontal and sagittal planes than the previous one. – Dummy specimens: two additional specimens were prepared for each type. They were used for optimizing the sectioning procedure and for assessing the absence of artefacts due to stem extraction and cement mantle sectioning (see below). To provide a comparison against a largely used stem, this study was compared with a previous publication for the Lubinus-SPII cemented hip stem (Bialoblocka-Juszczyk et al., 2010). In that study, six Lubinus-SPII stems implanted in composite bones with the same bone cement (Simplex-P) underwent the same in vitro testing as in the present study, using the same type of composite femurs, the same loading and testing equipment, and the same procedure for inspection of the cement mantles. The Lubinus-SPII was chosen for comparison because it has a high survival rate, which is similar to the Exeter-V40: nearly 98% after 10 years (Kärrholm et al., 2008). The use of the Lubinus-SPII as a benchmark for pre-clinical testing has been established in a large European Project (Maher and Prendergast, 2002; Stolk et al., 2002a; Waide et al., 2001). 2.2. Load history As the focus was on most demanding patients, a severe load history was simulated that replicated 24 years of activity of a very active patient (Cristofolini et al., 2007c). As walking is less detrimental than other tasks for hip joint loosening (Kassi et al., 2005; Stolk et al., 2002b), only stair-climbing and more severe motor tasks were included (Bergmann et al., 2001). All the most relevant activities from a fatigue point of view were included in the accelerated test (Table 1) (Cristofolini et al., 2007c) for a total of 1,000,348 cycles at 0.75 Hz. This procedure has been proven to yield results that are comparable to those found in retrieved cement mantles after in vivo cycling (Cristofolini et al., 2007a). To enable comparisons between Optimal and Undersized specimens, both series were tested so that load was applied with the same offset, resulting in identical bending moment and torsional moments being applied to all specimens (Cristofolini et al., 2007c).
Table 1 Activities in the simulated physiological loading (load values and frequency of occurrence). The load components were based on the literature so as to replicate the most critical scenario for the axial and torsional stability (Cristofolini et al., 2007c). The load history consisted of 1252 simulated weeks (corresponding to a total of 1,000,348 loading cycles). Axial force (compression)
Bending moment (frontal plane)
ACTIVITY
N
% BW
Nm
Stairs up Stairs down Bath tub entry Bath tub exit Car entry Car exit
2037 2223 2741 2741 3229 2939
370 404 498 498 587 534
16.50 27.94 22.17 22.17 38.01 28.93
Axial torque (intra-rotation) % BWm 3.00 5.08 4.03 4.03 6.91 5.26
Cycles/simulated day
Nm
% BWm
N
25.30 24.20 34.05 34.05 34.87 28.38
4.60 4.40 6.19 6.19 6.34 5.16
54 54 1 1 2 2
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
All transducers, were recorded throughout the test to obtain:
2.3. Implant-bone micromotion measurement The specimens were mounted on a biaxial testing machine (MiniBionix, MTS, Minneapolis, MN, USA), and were equipped with five displacement transducers (overall accuracy: better than 2.3 μm (Cristofolini et al., 2003; Monti et al., 1999)): – Four LVDTs (Mod. D5/40, RDP, Wolverhampton, UK) were mounted in 3.5 mm transcortical holes following a validated protocol, to measure the stem–cement interface shear micromotion: LVDT1, LVDT2 and LVDT3 measured micromotion in the direction of rotation on the medial aspect, respectively close to the calcar, at mid stem, and close to the stem tip; LVDT4 measured axial micromotions at the anterior side, close to the stem tip (Fig. 1). – A single-arm extensometer (Mod. 632.06-H20, MTS) was mounted proximally (on the greater trochanter, Fig. 1), to measure the axial micromotion of the stem with respect to the bone. This included stem–cement but also cement–bone micromotion.
901
– The inducible micromotion for each loading cycle (i.e., the amount of relative micromotion that is recovered after load release); – The permanent migration (i.e., the non-recoverable relative slippage at the interface which is not recovered, and which accumulates cycle after cycle). For each specimen and each sensor, data were processed to obtain (Cristofolini et al., 2003, 2007c): – The 95th percentile of the inducible micromotion throughout the duration of the test (95th percentile values were chosen so as to discard occasional higher peaks, which are not representative of the average trend); – The total permanent migration of the stem from the beginning to the end of the test. 2.4. Cement mantle inspection for fatigue damage At the end of the mechanical testing the specimens were opened axially, so as to extract the stems. To avoid loss of information in the regions where most cracks are normally found, the cutting plane was a sagittal one (close to the neutral plan when the femur is bent in its frontal plane). The stem–cement interface was inspected under an optical microscope, both in its original condition, and with the aid of dye penetrants (AVIO-B spray, Rotvel, American Gas&Chemical company, Northwale, NJ, USA) (Cristofolini et al., 2007a). Specimens were subsequently reassembled, and bone-cement constructs were sectioned transversely every 6.5 mm with a diamond blade (MDP200, Remet, Bologna, Italy) on a water-cooled disk-saw (TR60, Remet, Bologna, Italy). The cut surface was polished with silicon-carbidepaper (Hermes-P600, Stone-Boss, New York, USA) on a rotating polishing machine (D-2000, Exakt-Aparatebau, Norderstedt, Germany). Both faces of each slice were prepared with dye penetrants (AVIO-B spray). Images were acquired digitally with high spatial resolution (1 pixel = 4 μm) under an optical microscope. They were subsequently processed with a semi-automated custom-written software that allowed (Cristofolini et al., 2007a): – – – – –
Measurement of the cement mantle area Measurement of the cement mantle thickness Measurement of cracks longer than 0.05 mm Counting and recording the position of cracks Recording the presence of voids and defects.
The four Dummy implants that did not undergo fatigue tests were inspected with the same procedure to verify the absence of cracks due to the preparation and sectioning procedure, confirming the absence of artifactual cracks. 2.5. Statistical analysis The criterion of Peirce was applied to screen for doubtful data (Peirce, 1852; Ross, 2003): no specimen needed to be excluded from this study. To compare the Optimal against the Undersized specimens, and the Exeter specimens of this study against the Lubinus-SPII specimens, a nonparametric Kolmogorov–Smirnov test was applied to:
Fig. 1. In vitro setup for the cyclic loading in an anterior view, with the constraints used to avoid any undesirable load component to be transmitted to the specimen. The position of the transducers is visible: LVDT1 (proximal, medial), LVDT2 (mid stem, medial) and LVDT3 (stem tip, medial) are recording stem cement interface micromotions in the direction of rotation. The single-arm extensometer (mounted on the greater trochanter) and LVDT4 (same level as LVDT3, anterior) are recording micromotions in the longitudinal direction.
– The cement mantle thickness in the different regions of the cement mantle; – The inducible micromotions and permanent migrations at the different measurement positions; – The indicators of cement damage at the different positions along and around the stem. All Statistical analyses were performed using StatView (v.5.0.1, SAS Institute, Cary, NC, USA).
902
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
Fig. 2. Detail of the macroscopic crack that opened in the medial region in the Undersized specimens after few loading cycles.
3. Results All Optimal specimens survived the entire mechanical test, and none was visibly loose at the end of the test. Conversely, none of the Undersized specimens survived the mechanical test because of macroscopic implant failure: – The cement mantle in the antero-medial region fractured as early as the first simulated week of patient activity in all specimens (Fig. 2). Secondary cracks, in postero-medial aspect, often appeared within the 10th week of simulated patient activity; – Inducible micromotions and permanent migrations in the direction of rotation increased very quickly (Fig. 3), and after a few cycles exceeded the transducer range (1 mm); – The implants became so loose that the testing machine could apply the assigned loads with an error of more than 10% both on the torque and on the axial load, even after fine-tuning of the load-feedback loop.
motions in the direction of rotation about the stem axis never exceeded 1 μm (which is close to the sensitivity of the LVDTs, Fig. 5). Axial inducible micromotions were larger, but never exceeded 20 μm. Current results were compared against the Lubinus-SPII: although the axial inducible micromotions of the Optimal Exeter specimens were slightly larger than those of the Lubinus-SPII, this difference was not significant (Kolmogorov–Smirnov, P N 0.6); inducible micromotions in the direction of rotation did not differ significantly (Kolmogorov– Smirnov, P N 0.999). 3.2. Cement mantle thickness The cement mantle thickness (measured in the specimens sectioned after test completion, (Table 2) was on average (standard deviation) 2.79 (1.43) mm for the Optimal implants, and 1.92 (0.60)
Therefore, all tests on the Undersized specimens had to be terminated long before the end of the planned load history of 1252 simulated weeks: between 3 simulated weeks and 40 simulated weeks. 3.1. Permanent migrations and inducible micromotions A difference of one order of magnitude appeared between the permanent migrations of the Optimal and Undersized specimens even in the first few cycles (Fig. 3). In fact, while permanent migrations for the Undersized specimens steadily grew until failure occurred, for the Optimal they exhibited a first increase, while they grew much less during the rest of the test, according to a stabilization pattern. In the Optimal specimens, permanent migrations (Fig. 4) were much larger in the axial direction (up to 108 μm) than in the direction of rotation about the stem axis (approximately 20–30 μm). Current results were compared against previous results for the Lubinus-SPII (BialoblockaJuszczyk et al., 2010): axial permanent migrations were significantly larger for the Optimal Exeter specimens than for the Lubinus-SPII (Kolmogorov–Smirnov, P = 0.016), while such difference was less pronounced for the rotations about the stem axis (Kolmogorov– Smirnov, P N 0.6 for LVDT1 and LVDT2, but P = 0.02 for LVDT3). Inducible micromotions recorded in the Optimal specimens were stable over time, and were one order of magnitude lower than in the Undersized specimens. For the Optimal specimens, inducible micro-
Fig. 3. Permanent migrations over time for two typical specimens. The Optimal implants exhibited a smooth trend, and a tendency to settle over time. The Undersized specimens exhibited a sequence of sudden steps associated with opening of cement cracks, while the transducers went out-of-range within a few cycles. While LVDT1, LVDT2, and LVDT3 recorded micromotions associated with a rotation of the stem about its long axis (with a positive value corresponding to a rotation towards posterior), the extensometer and LVDT4 measured the axial component (with a negative value corresponding to the stem sinking towards distal).
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
903
Fig. 4. Permanent migrations (average and standard deviation): for the Optimal specimens they were recorded until the end of the simulated load history; for the Undersized specimens they were recorded when the tests was stopped due to macroscopic loosening (before 3% of the planned cyclic loading, as they exceeded the range of the displacement transducers). For comparison, previously published results (Bialoblocka-Juszczyk et al., 2010) for the Lubinus-SPII stem are plotted. While LVDT1, LVDT2, and LVDT3 recorded micromotions associated with a rotation of the stem about its long axis (with a positive value corresponding to a rotation towards posterior), the extensometer and LVDT4 measured the axial component (with a negative value corresponding to the stem sinking towards distal).
mm for the Undersized ones. This difference was statistically significant in the most proximal region (Kolmogorov–Smirnov, P = 0.04), while it was not significant at mid stem and distally (Kolmogorov–Smirnov, P N 0.2). The mantle in the anterior and lateral parts of the implant was thicker in the Optimal implants (Kolmogorov–Smirnov, P = 0.01 and P b 0.0001 respectively), while the thickness was comparable in the medial and posterior sides (Kolmogorov–Smirnov, P = 0.17).
After sectioning the Undersized specimens, some long cracks were visible on the inner cement surface. The location and the pattern of the cracks were consistent for all specimens (Fig. 6):
3.3. Fatigue damage in the cement mantle
In addition to the qualitative observations above, quantitative information about the crack distribution was obtained. It was not possible to compare the amount of cracks for the Optimal and the Undersized specimens, because of the inability of the Undersized specimens to complete the test. The first fatigue damage indicator was the number of cracks per unit volume of cement (Fig. 7): the same number of crack means more or less damage depending on the volume of cement where it is distributed (Cristofolini et al., 2007a). The distribution and the number of cracks per unit volume varied along the stem length in the Optimal specimens, in a way that was consistent between specimens: most cracks were localized near the distal centralizer; a few additional cracks were observed in the proximal region, medially. Current results were compared against the Lubinus-SPII (Bialoblocka-Juszczyk et al., 2010):
After sectioning the specimens longitudinally and extracting the stems, very little damage was visible on the inner cement surface of the Optimal specimens. The location and the pattern of the cracks were consistent for all specimens, with differences of less than 20% in terms of longitudinal extension of the cracks (Fig. 6): – A single crack was visible in the medial part, originating from the proximal region, and propagating typically for 10 mm towards distal. – Some porosity was found at the stem–cement interface proximally, mainly on the medial side (this was visible as stained areas). – Some short cracks were visible near the stem tip, corresponding to the distal centralizer.
– One or two cracks were visible in the medial part, originating from the proximal region (similar to the Optimal specimens), and propagating for more than half the stem length. – Because of the severe cracking in the proximal region, and of the extremely low number of cycles applied to these specimens, no cracks developed in the distal region.
904
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
Fig. 5. Inducible micromotions (average and standard deviation, absolute value): for the Optimal specimens they were recorded throughout the simulated load history; for the Undersized specimens they were recorded when the tests was stopped due to macroscopic loosening (before 3% of the planned cyclic loading, as they exceeded the range of the displacement transducers). For comparison, previously published results (Bialoblocka-Juszczyk et al., 2010) for the Lubinus-SPII stem are plotted. While LVDT1, LVDT2, and LVDT3 recorded micromotions associated with a rotation of the stem about its long axis, the extensometer and LVDT4 measured the axial component.
on average, the number of cracks per unit volume of the Exeter with an Optimal mantle did not differ significantly from those of the Lubinus-SPII (Fig. 7, Kolmogorov–Smirnov, P = 0.98). However, if regions were analyzed separately, the Optimal Exeter implants had fewer cracks than the Lubinus-SPII in all regions except in the most distal part. In fact, if the cracks involving the distal centralizer were excluded, the Optimal Exeter
Table 2 Thickness of the cement mantle in the two types of implant, in the most proximal, mid and distal third of the stem length (average and standard deviation between slices and between specimens, in millimetres). Cement thickness was measured after test completion, on the sectioned specimens. Position along the stem
Position around the stem
Optimal specimens: average (standard deviation)
Undersized specimens: average (standard deviation)
Proximal third
Anterior Posterior Medial Lateral Anterior Posterior Medial Lateral Anterior Posterior Medial Lateral
2.16 1.88 4.39 4.74 2.22 1.8 1.62 3.87 2.60 2.23 2.19 2.86
1.3 (0.14) 1.16 (0.19) 2.13 (0.43) 1.95 (0.71) 1.36 (0.25) 1.64 (0.13) 2.38 (0.35) 2.08 (0.38) 2.15 (0.46) 2.13 (0.30) 2.71 (0.46) 1.86 (0.54)
Mid stem
Distal third
(0.86) (0.90) (2.09) (1.00) (1.06) (0.73) (0.72) (0.86) (0.45) (0.52) (1.20) (1.06)
implants had nearly ten times fewer cracks per unit volume than the Lubinus-SPII (Kolmogorov–Smirnov, P = 0.024). Another indicator of the extent of the fatigue damage is the extension of the cracks (Cristofolini et al., 2007a). The area of the crack surface was estimated based on the length of the crack visible on each face of the implant slices (Fig. 7). Indication of the amount of damage was obtained when the crack extension was related to the total cement volume (Cristofolini et al., 2007a). The crack extension per unit volume was uneven in the Optimal specimens: most damage was localized near the distal centralizer. Current results were compared against the Lubinus-SPII: on average, the crack extension per unit volume of the Exeter with an Optimal mantle did not differ significantly from that of the Lubinus-SPII (Fig. 7, Kolmogorov–Smirnov, P N 0.999). If regions were analyzed separately, the Optimal Exeter implants had less crack extension than the Lubinus-SPII in all regions, except the most distal part. In fact, if the distal centralizer was excluded, the Optimal Exeter implants had nearly eight times less damage than the LubinusSPII (Kolmogorov–Smirnov, P = 0.024).
4. Discussion The scope of this work was to assess if stem undersizing and a thin cement mantle would significantly increase the risk of implant loosening for a clinically successful hip stem (the highly polished cemented Exeter-V40).
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
905
Fig. 6. Inner cement surface after sectioning the femurs, with dye penetrants highlighting the cracks: typical Optimal (left) and Undersized (right) specimens. It must be noted that while the cracks of the Optimal specimens were observed after completion of the entire planned cyclic loading, the cracks in the Undersized specimens were observed after the test was stopped due to macroscopic loosening (in all cases before 3% of the planned cyclic loading was achieved).
In our study, Exeter-V40 stems of the Optimal size were compared to Undersized implants, under identical in vitro loading conditions. The Undersized implants underwent early loosening, with all the indicators of implant failure (macroscopic fractures of the cement mantle, and large and increasing implant-bone micromotions). All the indicators of cement fatigue indicated more extensive damage in the case of Undersized implants. None of the Undersized implants was able to withstand the in vitro test (all tests were stopped at most after 3% of the planned load history). It must be noted that even clinically unsuccessful designs such as the Müller-Curved stem completed the proposed mechanical test, although reporting extensive cement damage and large and increasing implant micromotions (Cristofolini et al., 2007a,c). Therefore, this can be assumed as an indicator of extremely negative outcome. Conversely, when the same stem was implanted with an Optimal size, cement damage was moderate. Some damage was also present in the Optimal specimens, but this did not seem to significantly compromise mantle integrity. The number of cracks per unit volume and their distribution was comparable to that found in vitro for other successful implants where the same methods were applied (Bialoblocka-Juszczyk et al., 2010; Cristofolini et al., 2007a). Indeed, cracks were found also around ex vivo well-fixed hip implants autopsy-retrieved (Jasty et al., 1991; Kawate et al., 1998). Also, the inducible micromotions and permanent migrations observed
for the Optimal implants were relatively small, and compatible with successful implants (Cristofolini et al., 2007c). The trend of migrations over time (Fig. 3) for the Undersized implants clearly showed a steep and loosening trend, while the Optimal implants tended to stabilize over time. Therefore, in our study it was confirmed that an undersized Exeter-V40 implant has higher chances of loosening than a correctly implanted one. As in this study only one stem type was tested under different implantation conditions, it is not possible to state whether such a dramatic change is to be expected for any cemented hip stem, or this is related to some specific feature of the Exeter-V40. Ongoing investigation on retrieved failed Exeter implants can help elucidating how many failures are affected by poor planning or implantation. The scenario observed for the Undersized implants in our in vitro study is more critical than one would expect based on clinical experience with the Exeter-V40 stem. In fact, our protocol did not aim at replicating a typical hip patient, but a very active one (Cristofolini et al., 2007c). The rationale is that failure is a rare event, which occurs in a small percentage of patients (2–10% fail between few weeks and 10 years post-op (Kärrholm et al., 2008; Stea et al., 2008)), typically the most demanding, heavy and active patients (Kilgus et al., 1991). Therefore, in order for an in vitro test to be sensitive to small design/implantation variations, such group of patients at risk should be simulated (Dorey and Amstutz, 2002)
906
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
Fig. 7. Indicators of cement damage for the Optimal Exeter specimens: the graphs on top detail the number of cracks per unit volume of bone cement (count/mm3); the lower graphs report the sum of crack extensions over total cement volume (mm2/mm3). The crack statistics are analyzed as follows: (i) distribution along the stem; (ii) distribution around the stem; (iii) summary for the entire specimen (both including and excluding the cracks involving the distal centralizer). The bars correspond to the average and standard deviation. Results of the Undersized specimens are not reported as the test was stopped due to macroscopic loosening (before 3% of the planned cyclic loading). For comparison, published results (Bialoblocka-Juszczyk et al., 2010) for the Lubinus-SPII stem are plotted.
by applying a very severe load history. This way, our in vitro protocol is capable of capturing the conditions under which such implants fail (Cristofolini et al., 2007a,c). In other words, our tests do not show that all undersized implants would fail, but that undersizing, in combination with active patients, is likely to induce early loosening. Smaller micromotions and reduced cement damage should be expected when lower loads are applied (both in vitro and in vivo) (Cristofolini et al., 2007a,c). In the present study, bone cement was mixed without application of vacuum. There is no consensus about the relevance of vacuum mixing, both in vitro and clinically (Ling and Lee, 1998; Messick et al., 2007). In all cases, if an effect exists, hand-mixing represents the worst-case clinical scenario in which the hip stems must still function well. As the same mixing conditions were applied to all the specimens in the present study, quantitative comparisons are possible between the two Exeter groups. Our findings concerning the detrimental effect of Undersized implants are in agreement with other numerical (Janssen et al., 2009), in vitro (Harrington et al., 2002), and clinical studies. A comparison of the results obtained in this study with previous ones for a different stem (Bialoblocka-Juszczyk et al., 2010) corroborates the present findings. In fact: – Comparable micromotions were measured in vitro by others: (Britton and Prendergast, 2005) and (Zhang et al., 2008) measured an axial migration of the order of 20–30 μm after one million cycles
–
–
–
–
(in both such studies a constant load amplitude was applied, which corresponded to the smallest peak of our study) with Exeter stems implanted in composite femurs. The Optimal Exeter implants had larger permanent migrations and inducible micromotions than the Lubinus-SPII (which was tested in an identical fashion in a previous study). This is somewhat expected, as the Exeter (and in general polished tapered stem) shows larger micromotions in vivo (Malchau et al., 1994; Williams et al., 2002). Numerical models predicted similar patterns (Stolk et al., 2007). Similar to our in vitro results for the Optimal specimens, also (Ling, 2006) reported that most cracks are localized near the Exeter distal centralizer. At the same time, the overall indicators of cement damage did not differ between the Optimal Exeter implants and the Lubinus-SPII: this is in agreement with the fact that comparable survival rates are reported for the two designs (Kärrholm et al., 2008; Stea et al., 2008). A similar result was found in vitro by (Gravius et al., 2008): the Optimal Exeter implants induced 20% less cracks than the Lubinus-SPII, but this difference was not statistically significant. An explanation to these observations was provided by (Huiskes et al., 1998), who introduced the concept of “force-closed design” for the Exeter, as opposed to “shape-closed design”. Force-closed designs tend to move more, but this is not necessarily associated to cement damage, as cement creep can accommodate relatively large migrations (Ling, 2006).
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
There are some limitations of this study that need to be discussed: – Accelerated test: our method simulated a selected range of activities, ignoring many others that are known for generating lower loads (e.g. walking, sitting down etc. (Bergmann et al., 2001)), as well as resting intervals. Load components were simplified to allow testing on a biaxial machine. It has been shown that the cement damage obtained using our in vitro testing is similar to that found in cement mantles retrieved upon revision hip surgery for the same stem type (Cristofolini et al., 2007a, 2003). However, for the Exeter stem, where cement creep is important, it is possible that our approach leads to underestimate implant migrations. In fact, migrations observed in this study are slightly larger but comparable to those found in other in vitro studies with similar limitations (Britton and Prendergast, 2005). – Composite femurs: The use of these bone models as a benchmark for comparative pre-clinical testing has been established in a large European Project (Maher and Prendergast, 2002; Waide et al., 2001), as they keep the same stiffness properties as human bones (Cristofolini et al., 1996) while allowing a significant reduction of the specimen variability compared to human bones. Previous research (Cristofolini, 1997; Harman et al., 1995) has demonstrated that the stability of the prosthesis and the sensitivity to the stem design are of the same order of magnitude in synthetic and human femurs. In conclusion, the hypothesis that undersizing the Exeter hip stem can cause implant loosening is corroborated. This has serious clinical implications: in fact, our results show that even an excellent stem can turn into a complete failure if implantation is not adequately planned and performed. Surgeons should always pay extra care in selecting the correct stem size, and in avoiding under-reaming of the femoral cavity. Conflict of interest statement This research was financially supported by Aesculap AG, Tuttlingen, Germany. However, this study does not aim at demonstrating that a prosthetic design is better or worse than any other. This study simply investigates whether the same stem, implanted improperly, can generate a higher failure risk. Acknowledgements The authors thank Luigi Lena for the illustrations. This research was financially supported by Aesculap AG. References Barrack, R.L., 2000. Early failure of modern cemented stems. J. Arthroplasty 15, 1036–1050. Bergmann, G., Deuretzbacher, G., Heller, M., Graichen, F., Rohlmann, A., Strauss, J., et al., 2001. Hip contact forces and gait patterns from routine activities. J. Biomech. 34, 859–871. Bialoblocka-Juszczyk, E., Cristofolini, L., Erani, P., Viceconti, M., 2010. Effect of long-term physiological activity on the long-term stem stability of cemented hip arthroplasty: in vitro comparison of three commercial bone cements. Proc. Inst. Mech. Eng. H 224, 53–65. Breusch, S.J., Lukoschek, M., Schneider, U., Ewerbeck, V., 2000. “State of the art” der zementierten Hüftendoprothetik: Qualität der Zementiertechnik ist entscheidend. Dtsch. Arzteblatt. 97, 2030–2033. Britton, J.R., Prendergast, P.J., 2005. Preclinical testing of femoral hip components: an experimental investigation with four prostheses. J. Biomech. Eng. 127, 872–880. Cristofolini, L., 1997. A critical analysis of stress shielding evaluation of hip prostheses. Crit. Rev. Biomed. Eng. 25, 409–483. Cristofolini, L., Viceconti, M., Cappello, A., Toni, A., 1996. Mechanical validation of whole bone composite femur models. J. Biomech. 29, 525–535. Cristofolini, L., Savigni, P., Saponara Teutonico, A., Toni, A., 2002. Ex-vivo and in-vitro cement mantle fatigue damage around femoral stems: validation of a protocol to simulate real-life loading in hip replacement patients. In: Bedzinski, R.P.C., Scigala, K. (Eds.), 13th Conference of the European Society of Biomechanics (Clinical
907
Biomechanics Award 2002). Oficyna Wydawnicza Politechniki Wroclawskiej Publ., Wroclaw, Poland, Wroclaw, Poland, pp. 48–49. Cristofolini, L., Saponara Teutonico, A., Monti, L., Cappello, A., Toni, A., 2003. Comparative in vitro study on the long term performance of cemented hip stems: validation of a protocol to discriminate between “good” and “bad” designs. J. Biomech. 36, 1603–1615. Cristofolini, L., Erani, P., Savigni, P., Bordini, B., Viceconti, M., 2007a. Pre-clinical assessment of the long-term endurance of cemented hip stems. Part 2: in-vitro and ex-vivo fatigue damage of the cement mantle. Proc. Inst. Mech. Eng. H 221, 585–599. Cristofolini, L., Erani, P., Savigni, P., Grupp, T., Thies, O., Viceconti, M., 2007b. Increased long-term failure risk associated with excessively thin cement mantle in cemented hip arthroplasty: a comparative in vitro study. Clin. Biomech. 22, 410–421. Cristofolini, L., Saponara Teutonico, A., Savigni, P., Erani, P., Viceconti, M., 2007c. Pre-clinical assessment of the long-term endurance of cemented hip stems. Part 1: effect of daily activities - a comparison of two load histories. Proc. Inst. Mech. Eng. H 221, 569–584. Dorey, F.D., Amstutz, H.C., 2002. Editorial: the need to account for patient activity when evaluating results of total hip arthroplasty with survivorship analysis. J. Bone Joint Surg. Am. 84-A, 709–710. Ebramzadeh, E., Sarmiento, A., McKellop, H.A., Llinas, A., Gogan, W., 1994. The cement mantle in total hip arthroplasty. Analysis of long-term radiographic results. J. Bone Joint Surg. Am. 76, 77–87. Ebramzadeh, E., Sangiorgio, S.N., Longjohn, D.B., Buhari, C.F., Dorr, L.D., 2004. Initial stability of cemented femoral stems as a function of surface finish, collar, and stem size. J. Bone Joint Surg. Am. 86-A, 106–115. Gheduzzi, S., Nedungayil, S., Learmonth, I.D., 2004. Know your cement: a review of current cementing practice in the UK. Hip Int. 14, 128–129. Giannikas, K.A., Din, R., Sadiq, S., Dunningham, T.H., 2002. Medium-term results of the ABG total hip arthroplasty in young patients. J. Arthroplasty 17, 184–188. Gravius, S., Wirtz, D.C., Siebert, C.H., Andereya, S., Mueller-Rath, R., Maus, U., et al., 2008. In vitro interface and cement mantle analysis of different femur stem designs. J. Biomech. 41, 2021–2028. Harman, M.K., Toni, A., Cristofolini, L., Viceconti, M., 1995. Initial stability of uncemented hip stems: an in-vitro protocol to measure torsional interface motion. Med. Eng. Phys. 17, 163–171. Harrington, M.A.J., O'Connor, D.O., Lozynsky, A.J., Kovach, I., Harris, W.H., 2002. Effects of femoral neck length, stem size, and body weight on strains in the proximal cement mantle. J. Bone Joint Surg. Am. 84, 573–579. Hertzler, J., Miller, M.A., Mann, K.A., 2002. Fatigue crack growth rate does not depend on mantle thickness: an idealized cemented stem construct under torsional loading. J. Orthop. Res. 20, 676–682. Huiskes, R., Verdonschot, N., Nivbrant, B., 1998. Migration, stem shape, and surface finish in cemented total hip arthroplasty. Clin. Orthop. Relat. Res. 103–112. Janssen, D., van Aken, J., Scheerlinck, T., Verdonschot, N., 2009. Finite element analysis of the effect of cementing concepts on implant stability and cement fatigue failure. Acta Orthop. 80, 319–324. Jasty, M., Maloney, W.J., Bragdon, C.R., O'Connor, D.O., Haire, T., Harris, W.H., 1991. The initiation of failure in cemented femoral components of hip arthroplasties. J. Bone Joint Surg. Br. 73, 551–558. Kärrholm, J., G., G., C., R., P., H. Swedish hip arthroplasty register. Annual report 2007 (available from: http://www.jru.orthop.gu.se/): Swedish HIp Arthoplasty Register; 2008. Kassi, J.-P., Heller, M.O., Stoeckle, U., Perka, C., Guda, G.N., 2005. Stair climbing is more critical than walking in pre-clinical assessment of primary stability in cementless THA in vitro. J. Biomech. 38, 1143–1154. Kawate, K., Maloney, W.J., Bragdon, C.R., Biggs, S., Jasty, M., Harris, W.H., 1998. Importance of a thin cement mantle. Autopsy studies of eight hips. Clin. Orthop. Relat. Res. 355, 70–76. Kilgus, D.J., Dorey, F.D., Finerman, G.A.M., Amstutz, H.C., 1991. Patient activity, sports participation, and impact loading on the durability of cemented total hip replacement. Clin. Orthop. Relat. Res. 269, 25–31. Lattanzi, R., Grazi, E., Testi, D., Viceconti, M., Cappello, A., Toni, A., 2003. Accuracy and repeatability of cementless total hip replacement surgery in patients with deformed anatomies. Med. Inform. Internet Med. 28, 59–71. Ling, R.S.M., 2006. The History and Development of the Exeter Hip2nd Edition. Stryker, UK. Ling, R.S., Lee, A.J., 1998. Porosity reduction in acrylic cement is clinically irrelevant. Clin. Orthop. Relat. Res. 249–253. Maher, S.A., Prendergast, P.J., 2002. Discriminating the loosening behaviour of cemented hip prosthesis using measurements of migration and inducible displacement. J. Biomech. 35, 257–265. Malchau, H., Snorrason, F., Herberts, P., Karrholm, J., 1994. Micromotion of femoral stems in total hip arthroplasty. A randomized study of cemented, hydroxyapatitecoated, and porous-coated stems with roentgen stereophotogrammetric analysis. J. Bone Joint Surg. Am. 76, 1692–1705. Mann, K.A., Gupta, S., Race, A., Miller, M.A., Cleary, R.J., Ayers, D.C., 2004. Cement microcracks in thin-mantle regions after in vitro fatigue loading. J. Arthroplasty 19, 605–612. Messick, K.J., Miller, M.A., Damron, L.A., Race, A., Clarke, M.T., Mann, K.A., 2007. Vacuum-mixing cement does not decrease overall porosity in cemented femoral stems: an in vitro laboratory investigation. J. Bone Joint Surg. Br. 89, 1115–1121. Monti, L., Cristofolini, L., Viceconti, M., 1999. An in-vitro quantitative analysis of the primary stability of uncemented hip stems. Artif. Organs 23, 851–859. Peirce, B., 1852. Criterion for the rejection of doubtful observations. Astron. J. 2, 161–163. Ramaniraka, N.A., Rakatomanana, L.R., Leyvraz, P.-F., 2000. The fixation of cemented femoral component. Effect of stem stiffness, cement thickness and roughness of the cement–bone surface. J. Bone Joint Surg. Br. 82-B, 297–303. Ross, S.M., 2003. Peirce's criterion for the elimination of suspect experimental data. J. Eng. Tech. 2003, 1–12.
908
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
Ruff, C.B., Hayes, W.C., 1983. Cross-sectional Geometry of Pecos Pueblo Femora and Tibiae — a biomechanical investigation: I. Method and general patterns of variation. Am. J. Phys. Anthropol. 60, 359–381. Scheerlinck, T., Druyts, P., Casteleyn, P.P., 2004. The use of primary total hip arthroplasty in university hospitals of the European Union. Acta Orthop. Belg. 70, 231–239. Scheerlinck, T., de Mey, J., Deklerck, R., Noble, P.C., 2006. CT analysis of defects of the cement mantle and alignment of the stem: in vitro comparison of Charnley– Kerboul femoral hip implants inserted line-to-line and undersized in paired femora. J. Bone Joint Surg. Br. 88, 19–25. Stea, S., Bordini, B., De Clerico, M., Toni, A., 2008. Report of orthopaedic prosthetic implantology. Overall data: hip, knee and shoulder arthroplasty in Emilia Romagna region: Istituto Ortopedico Rizzoli. Stolk, J., Verdonschot, N., Cristofolini, L., Toni, A., Huiskes, R., 2002a. Finite element and experimental models of cemented hip joint reconstructions can produce similar bone and cement strain in pre-clinical tests. J. Biomech. 35, 499–510. Stolk, J., Verdonschot, N., Huiskes, R., 2002b. Stair climbing is more detrimental to the cement in hip replacement than walking. Clin. Orthop. Relat. Res. 405, 294–305.
Stolk, J., Janssen, D., Huiskes, R., Verdonschot, N., 2007. Finite element-based preclinical testing of cemented total hip implants. Clin. Orthop. Relat. Res. 456, 138–147. Verdonschot, N.J., 1996. Biomechanical failure Scenarios for Cemented Total Hip Replacement. Ponsen & Looijen bv publ, Wageningen. Viceconti, M., Lattanzi, R., Antonietti, B., Paderni, S., Olmi, R., Sudanese, A., et al., 2003. CT-based surgical planning software improves the accuracy of total hip replacement preoperative planning. Med. Eng. Phys. 25, 371–377. Waide, V., Cristofolini, L., Toni, A., 2001. A CAD-CAM methodology to produce bone remodeled composite femurs for preclinical investigations. Proc. Inst. Mech. Eng. H 215, 459–469. Williams, H.D., Browne, G., Gie, G.A., Ling, R.S., Timperley, A.J., Wendover, N.A., 2002. The Exeter universal cemented femoral component at 8 to 12 years. A study of the first 325 hips. J. Bone Joint Surg. Br. 84, 324–334. Zhang, H.Y., Blunt, L., Jiang, X.Q., Brown, L., Barrans, S., Zhao, Y., 2008. Femoral stem wear in cemented total hip replacement. Proc. Inst. Mech. Eng. H 222, 583–592.
Contents lists available at ScienceDirect
Clinical Biomechanics j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / c l i n b i o m e c h
Effect of undersizing on the long-term stability of the Exeter hip stem: A comparative in vitro study Luca Cristofolini a,b,⁎, Paolo Erani a, Ewa Bialoblocka-Juszczyk a,b, Hirotsugu Ohashi c, Satoshi Iida d, Izumi Minato e, Marco Viceconti a a
Laboratorio di Tecnologia Medica, Istituto Ortopedico Rizzoli, Via di Barbiano 1/10, 40136 Bologna, Italy DIEM, Università degli Studi di Bologna, Viale Risorgimento 2, 40136 Bologna, Italy Department of Orthopaedic Surgery, Saiseikai Nakatsu Hospital, 2-10-39, Shibata, Kitaku, Osaka 530-0012, Japan d Department of Orthopaedic Surgery, Matsudo City Hospital, Kamihongou 4005, Matsudo, Chiba 271-0064, Japan e Department of Orthopaedic Surgery, Niigata Rinko Hospital, Momoyamacho 1-114-3, Niigata City, Niigata 950-0051, Japan b c
a r t i c l e
i n f o
Article history: Received 8 March 2010 Accepted 5 July 2010 Keywords: Polished cemented hip prosthesis Stem undersizing Long-term implant-bone stability Fatigue failure Inducible micromotion Permanent migration In vitro testing
a b s t r a c t Background: Even for clinically successful hip stems such as the Exeter-V40 occasional failures are reported. It has been reported that sub-optimal pre-operative planning, leading to implant undersizing and/or thin cement mantle, can explain such failures. The scope of this study was to investigate whether stem undersizing and a thin cement mantle are sufficient to cause implant loosening. Methods: A comparative in vitro study was designed to compare hip implants prepared with optimal and smaller than optimal stem size. Exeter-V40, a highly polished cemented hip stem, was used in both cases. Tests were carried out simulating 24 years of activity of active hip patients. A multifaceted approach was taken: inducible and permanent micromotions were recorded throughout the test; cement micro-cracks were quantified using dye penetrants and statistically analyzed. Findings: The implants with an optimal stem size withstood the entire mechanical test, with low and stable inducible micromotions and permanent migrations during the test, and with moderate fatigue damage in the cement mantle after test completion. Conversely, the undersized specimens showed large and increasing micromotions, and failed after few loading cycles, because of macroscopic cracks in the proximal part of the cement mantle. While results for the optimal stem size are typical for stable hip stems, those for the undersize stem indicate a critical scenario. Interpretation: These results confirm that even a clinically successful hip prosthesis such as the Exeter-V40 is prone to early loosening if a stem smaller than the optimal size is implanted. © 2010 Elsevier Ltd. All rights reserved.
1. Introduction The Exeter is one of the most frequently used hip stems, accounting for 21.8% of the cemented stems implanted in Europe (Scheerlinck et al., 2004). It is a clinically successful stem, with a survival rate of 98.5% at 7 years and over 95% at 10 years (Kärrholm et al., 2008; Stea et al., 2008). While the average survival rate in itself is excellent, there are hospitals/divisions where higher failure rates are reported (Kärrholm et al., 2008; Stea et al., 2008). Implant failure may be due to a number of factors, related to the patient, the device, and the surgical technique. There are two causes of failure of cemented hip stems by aseptic loosening, which are related to the surgical technique: (i) sub-optimal pre-operative planning (Ebramzadeh et al., 2004), and (ii) excessively thin cement mantle (Ebramzadeh et al., 1994). ⁎ Corresponding author. Laboratorio di Tecnologia Medica, Istituto Ortopedico Rizzoli, Via di Barbiano, 1/10, 40136 Bologna, Italy. E-mail address: [email protected] (L. Cristofolini). 0268-0033/$ – see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.clinbiomech.2010.07.003
Incorrect choice of the stem size is a relatively common problem related to pre-operative planning, both in cemented and cementless hip stems (Barrack, 2000; Giannikas et al., 2002; Viceconti et al., 2003). It has been reported that stem undersizing by two stem sizes occurs in 10% of the cases with anatomical uncemented hip stems, when pre-operative templating is based on standard radiographs (Viceconti et al., 2003). Stem undersizing in cemented arthroplasty has been associated to excessive stress in the cement mantle (Harrington et al., 2002; Janssen et al., 2009), which can lead to early aseptic loosening (Verdonschot, 1996). It has been suggested that the use of pre-operative planning software enabling a full threedimensional virtual visualisation would reduce the incidence of such problems (Lattanzi et al., 2003; Viceconti et al., 2003). Insufficient cement mantle thickness is primarily caused by insufficient reaming of the femoral canal (Scheerlinck et al., 2006). In fact, it has often been shown that a mantle thicker than 2–3 mm reduces cement stress and micromotions (Ramaniraka et al., 2000). Conversely, more cracks are found both ex vivo (Kawate et al., 1998) and in vitro (Mann et al., 2004) where the cement mantle is thin. In
900
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
fact, even in the case when cracks grow at the same rate across the cement mantle, a reduced thickness is expected to decrease the time needed for a crack to break through the entire mantle thickness (Hertzler et al., 2002). In vitro testing is able to discriminate between successful and critical hip stems (Maher and Prendergast, 2002; Britton and Prendergast, 2005; Cristofolini et al., 2007a,c; Zhang et al., 2008). In vitro testing has also been able to explain early failure of a specific cemented stem in relation with excessively thin cement mantles (Cristofolini et al., 2007b). The scope of this work was to assess if, given a clinically successful hip stem, sub-optimal planning (stem undersizing) and sub-optimal implantation (thin cement mantle) is sufficient to cause implant failure. In order to test this hypothesis, composite femurs implanted with Exeter-V40 stems of different sizes were tested in vitro to investigate implant micromotions and cement fatigue damage.
2. Methods A mechanical testing procedure that was extensively validated (also in comparison against ex vivo retrievals) was recently proposed for quantifying long-term hip implant damage both in terms of implant-bone micromotions and of cement fatigue damage (Cristotofolini et al., 2007a,c; 2003). As revision rates are extremely low, failures are most likely to occur in the most demanding patients for implant survival, i.e. the young, active and heavy patients that are most likely to apply higher and most frequent loads (Dorey and Amstutz, 2002; Kilgus et al., 1991). A loading profile has recently been proposed that enables simulating 24 years of activity of a very demanding patient (Cristofolini et al., 2007c; 2002).
2.1. Specimens The present study was carried out on a total of ten Exeter hip stems (Exeter-V40, Howmedica, Mahwah, NJ, USA), which is a Cr–Co–Mo highly polished cemented stem. They were implanted by a pool of skilled surgeons in composite femurs (Mod. 3103, Pacific Research Labs, Vashon Island, WA), using the instrumentation recommended by the manufacturer. An anatomic reference system was established (Cristofolini, 1997; Ruff and Hayes, 1983) to assist in maintaining consistent implantation, and increase repeatability during specimen preparation and testing. To enable reproducible preparation, bone cement (Simplex-P, Stryker-Howmedica, Mahwah, NJ, USA) was mixed using a sealed mixing device (Summit Medical Syringe Type, Summit Medical Group, Bourton-on-the-Water, UK) at 23–25 °C, 40– 55% relative humidity. To replicate the worst-case clinical scenario in which the hip stems must still function well, mixing was performed without application of vacuum (which is one of the options indicated by the Manufacturer for Simplex-P). In fact, hand-mixing is still used in the clinical practice between 6% (Gheduzzi et al., 2004) and 50% of cases (Breusch et al., 2000; Stea et al., 2008). Cement was injected in a
retrograde fashion prior to stem insertion. The stems were provided with the standard Exeter distal centralizer. Ten femurs were implanted as follows (specimens were X-rayed to document alignment and mantle thickness): – Optimal specimens: four femurs were prepared with the optimal stem size (Exeter-V40 femoral stem: 37.5 mm offset, No. 2) based on pre-operative templating following the manufacturer's instructions. – Undersized specimens: four femurs were rasped to a smaller size than the Optimal one (Exeter-V40 femoral stem: 35.5 mm offset). A stem two sizes smaller than the Optimal one was chosen, as errors of between one and three sizes are commonly reported for hip stems (Viceconti et al., 2003). Such stem was slightly shorter, and narrower both in the frontal and sagittal planes than the previous one. – Dummy specimens: two additional specimens were prepared for each type. They were used for optimizing the sectioning procedure and for assessing the absence of artefacts due to stem extraction and cement mantle sectioning (see below). To provide a comparison against a largely used stem, this study was compared with a previous publication for the Lubinus-SPII cemented hip stem (Bialoblocka-Juszczyk et al., 2010). In that study, six Lubinus-SPII stems implanted in composite bones with the same bone cement (Simplex-P) underwent the same in vitro testing as in the present study, using the same type of composite femurs, the same loading and testing equipment, and the same procedure for inspection of the cement mantles. The Lubinus-SPII was chosen for comparison because it has a high survival rate, which is similar to the Exeter-V40: nearly 98% after 10 years (Kärrholm et al., 2008). The use of the Lubinus-SPII as a benchmark for pre-clinical testing has been established in a large European Project (Maher and Prendergast, 2002; Stolk et al., 2002a; Waide et al., 2001). 2.2. Load history As the focus was on most demanding patients, a severe load history was simulated that replicated 24 years of activity of a very active patient (Cristofolini et al., 2007c). As walking is less detrimental than other tasks for hip joint loosening (Kassi et al., 2005; Stolk et al., 2002b), only stair-climbing and more severe motor tasks were included (Bergmann et al., 2001). All the most relevant activities from a fatigue point of view were included in the accelerated test (Table 1) (Cristofolini et al., 2007c) for a total of 1,000,348 cycles at 0.75 Hz. This procedure has been proven to yield results that are comparable to those found in retrieved cement mantles after in vivo cycling (Cristofolini et al., 2007a). To enable comparisons between Optimal and Undersized specimens, both series were tested so that load was applied with the same offset, resulting in identical bending moment and torsional moments being applied to all specimens (Cristofolini et al., 2007c).
Table 1 Activities in the simulated physiological loading (load values and frequency of occurrence). The load components were based on the literature so as to replicate the most critical scenario for the axial and torsional stability (Cristofolini et al., 2007c). The load history consisted of 1252 simulated weeks (corresponding to a total of 1,000,348 loading cycles). Axial force (compression)
Bending moment (frontal plane)
ACTIVITY
N
% BW
Nm
Stairs up Stairs down Bath tub entry Bath tub exit Car entry Car exit
2037 2223 2741 2741 3229 2939
370 404 498 498 587 534
16.50 27.94 22.17 22.17 38.01 28.93
Axial torque (intra-rotation) % BWm 3.00 5.08 4.03 4.03 6.91 5.26
Cycles/simulated day
Nm
% BWm
N
25.30 24.20 34.05 34.05 34.87 28.38
4.60 4.40 6.19 6.19 6.34 5.16
54 54 1 1 2 2
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
All transducers, were recorded throughout the test to obtain:
2.3. Implant-bone micromotion measurement The specimens were mounted on a biaxial testing machine (MiniBionix, MTS, Minneapolis, MN, USA), and were equipped with five displacement transducers (overall accuracy: better than 2.3 μm (Cristofolini et al., 2003; Monti et al., 1999)): – Four LVDTs (Mod. D5/40, RDP, Wolverhampton, UK) were mounted in 3.5 mm transcortical holes following a validated protocol, to measure the stem–cement interface shear micromotion: LVDT1, LVDT2 and LVDT3 measured micromotion in the direction of rotation on the medial aspect, respectively close to the calcar, at mid stem, and close to the stem tip; LVDT4 measured axial micromotions at the anterior side, close to the stem tip (Fig. 1). – A single-arm extensometer (Mod. 632.06-H20, MTS) was mounted proximally (on the greater trochanter, Fig. 1), to measure the axial micromotion of the stem with respect to the bone. This included stem–cement but also cement–bone micromotion.
901
– The inducible micromotion for each loading cycle (i.e., the amount of relative micromotion that is recovered after load release); – The permanent migration (i.e., the non-recoverable relative slippage at the interface which is not recovered, and which accumulates cycle after cycle). For each specimen and each sensor, data were processed to obtain (Cristofolini et al., 2003, 2007c): – The 95th percentile of the inducible micromotion throughout the duration of the test (95th percentile values were chosen so as to discard occasional higher peaks, which are not representative of the average trend); – The total permanent migration of the stem from the beginning to the end of the test. 2.4. Cement mantle inspection for fatigue damage At the end of the mechanical testing the specimens were opened axially, so as to extract the stems. To avoid loss of information in the regions where most cracks are normally found, the cutting plane was a sagittal one (close to the neutral plan when the femur is bent in its frontal plane). The stem–cement interface was inspected under an optical microscope, both in its original condition, and with the aid of dye penetrants (AVIO-B spray, Rotvel, American Gas&Chemical company, Northwale, NJ, USA) (Cristofolini et al., 2007a). Specimens were subsequently reassembled, and bone-cement constructs were sectioned transversely every 6.5 mm with a diamond blade (MDP200, Remet, Bologna, Italy) on a water-cooled disk-saw (TR60, Remet, Bologna, Italy). The cut surface was polished with silicon-carbidepaper (Hermes-P600, Stone-Boss, New York, USA) on a rotating polishing machine (D-2000, Exakt-Aparatebau, Norderstedt, Germany). Both faces of each slice were prepared with dye penetrants (AVIO-B spray). Images were acquired digitally with high spatial resolution (1 pixel = 4 μm) under an optical microscope. They were subsequently processed with a semi-automated custom-written software that allowed (Cristofolini et al., 2007a): – – – – –
Measurement of the cement mantle area Measurement of the cement mantle thickness Measurement of cracks longer than 0.05 mm Counting and recording the position of cracks Recording the presence of voids and defects.
The four Dummy implants that did not undergo fatigue tests were inspected with the same procedure to verify the absence of cracks due to the preparation and sectioning procedure, confirming the absence of artifactual cracks. 2.5. Statistical analysis The criterion of Peirce was applied to screen for doubtful data (Peirce, 1852; Ross, 2003): no specimen needed to be excluded from this study. To compare the Optimal against the Undersized specimens, and the Exeter specimens of this study against the Lubinus-SPII specimens, a nonparametric Kolmogorov–Smirnov test was applied to:
Fig. 1. In vitro setup for the cyclic loading in an anterior view, with the constraints used to avoid any undesirable load component to be transmitted to the specimen. The position of the transducers is visible: LVDT1 (proximal, medial), LVDT2 (mid stem, medial) and LVDT3 (stem tip, medial) are recording stem cement interface micromotions in the direction of rotation. The single-arm extensometer (mounted on the greater trochanter) and LVDT4 (same level as LVDT3, anterior) are recording micromotions in the longitudinal direction.
– The cement mantle thickness in the different regions of the cement mantle; – The inducible micromotions and permanent migrations at the different measurement positions; – The indicators of cement damage at the different positions along and around the stem. All Statistical analyses were performed using StatView (v.5.0.1, SAS Institute, Cary, NC, USA).
902
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
Fig. 2. Detail of the macroscopic crack that opened in the medial region in the Undersized specimens after few loading cycles.
3. Results All Optimal specimens survived the entire mechanical test, and none was visibly loose at the end of the test. Conversely, none of the Undersized specimens survived the mechanical test because of macroscopic implant failure: – The cement mantle in the antero-medial region fractured as early as the first simulated week of patient activity in all specimens (Fig. 2). Secondary cracks, in postero-medial aspect, often appeared within the 10th week of simulated patient activity; – Inducible micromotions and permanent migrations in the direction of rotation increased very quickly (Fig. 3), and after a few cycles exceeded the transducer range (1 mm); – The implants became so loose that the testing machine could apply the assigned loads with an error of more than 10% both on the torque and on the axial load, even after fine-tuning of the load-feedback loop.
motions in the direction of rotation about the stem axis never exceeded 1 μm (which is close to the sensitivity of the LVDTs, Fig. 5). Axial inducible micromotions were larger, but never exceeded 20 μm. Current results were compared against the Lubinus-SPII: although the axial inducible micromotions of the Optimal Exeter specimens were slightly larger than those of the Lubinus-SPII, this difference was not significant (Kolmogorov–Smirnov, P N 0.6); inducible micromotions in the direction of rotation did not differ significantly (Kolmogorov– Smirnov, P N 0.999). 3.2. Cement mantle thickness The cement mantle thickness (measured in the specimens sectioned after test completion, (Table 2) was on average (standard deviation) 2.79 (1.43) mm for the Optimal implants, and 1.92 (0.60)
Therefore, all tests on the Undersized specimens had to be terminated long before the end of the planned load history of 1252 simulated weeks: between 3 simulated weeks and 40 simulated weeks. 3.1. Permanent migrations and inducible micromotions A difference of one order of magnitude appeared between the permanent migrations of the Optimal and Undersized specimens even in the first few cycles (Fig. 3). In fact, while permanent migrations for the Undersized specimens steadily grew until failure occurred, for the Optimal they exhibited a first increase, while they grew much less during the rest of the test, according to a stabilization pattern. In the Optimal specimens, permanent migrations (Fig. 4) were much larger in the axial direction (up to 108 μm) than in the direction of rotation about the stem axis (approximately 20–30 μm). Current results were compared against previous results for the Lubinus-SPII (BialoblockaJuszczyk et al., 2010): axial permanent migrations were significantly larger for the Optimal Exeter specimens than for the Lubinus-SPII (Kolmogorov–Smirnov, P = 0.016), while such difference was less pronounced for the rotations about the stem axis (Kolmogorov– Smirnov, P N 0.6 for LVDT1 and LVDT2, but P = 0.02 for LVDT3). Inducible micromotions recorded in the Optimal specimens were stable over time, and were one order of magnitude lower than in the Undersized specimens. For the Optimal specimens, inducible micro-
Fig. 3. Permanent migrations over time for two typical specimens. The Optimal implants exhibited a smooth trend, and a tendency to settle over time. The Undersized specimens exhibited a sequence of sudden steps associated with opening of cement cracks, while the transducers went out-of-range within a few cycles. While LVDT1, LVDT2, and LVDT3 recorded micromotions associated with a rotation of the stem about its long axis (with a positive value corresponding to a rotation towards posterior), the extensometer and LVDT4 measured the axial component (with a negative value corresponding to the stem sinking towards distal).
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
903
Fig. 4. Permanent migrations (average and standard deviation): for the Optimal specimens they were recorded until the end of the simulated load history; for the Undersized specimens they were recorded when the tests was stopped due to macroscopic loosening (before 3% of the planned cyclic loading, as they exceeded the range of the displacement transducers). For comparison, previously published results (Bialoblocka-Juszczyk et al., 2010) for the Lubinus-SPII stem are plotted. While LVDT1, LVDT2, and LVDT3 recorded micromotions associated with a rotation of the stem about its long axis (with a positive value corresponding to a rotation towards posterior), the extensometer and LVDT4 measured the axial component (with a negative value corresponding to the stem sinking towards distal).
mm for the Undersized ones. This difference was statistically significant in the most proximal region (Kolmogorov–Smirnov, P = 0.04), while it was not significant at mid stem and distally (Kolmogorov–Smirnov, P N 0.2). The mantle in the anterior and lateral parts of the implant was thicker in the Optimal implants (Kolmogorov–Smirnov, P = 0.01 and P b 0.0001 respectively), while the thickness was comparable in the medial and posterior sides (Kolmogorov–Smirnov, P = 0.17).
After sectioning the Undersized specimens, some long cracks were visible on the inner cement surface. The location and the pattern of the cracks were consistent for all specimens (Fig. 6):
3.3. Fatigue damage in the cement mantle
In addition to the qualitative observations above, quantitative information about the crack distribution was obtained. It was not possible to compare the amount of cracks for the Optimal and the Undersized specimens, because of the inability of the Undersized specimens to complete the test. The first fatigue damage indicator was the number of cracks per unit volume of cement (Fig. 7): the same number of crack means more or less damage depending on the volume of cement where it is distributed (Cristofolini et al., 2007a). The distribution and the number of cracks per unit volume varied along the stem length in the Optimal specimens, in a way that was consistent between specimens: most cracks were localized near the distal centralizer; a few additional cracks were observed in the proximal region, medially. Current results were compared against the Lubinus-SPII (Bialoblocka-Juszczyk et al., 2010):
After sectioning the specimens longitudinally and extracting the stems, very little damage was visible on the inner cement surface of the Optimal specimens. The location and the pattern of the cracks were consistent for all specimens, with differences of less than 20% in terms of longitudinal extension of the cracks (Fig. 6): – A single crack was visible in the medial part, originating from the proximal region, and propagating typically for 10 mm towards distal. – Some porosity was found at the stem–cement interface proximally, mainly on the medial side (this was visible as stained areas). – Some short cracks were visible near the stem tip, corresponding to the distal centralizer.
– One or two cracks were visible in the medial part, originating from the proximal region (similar to the Optimal specimens), and propagating for more than half the stem length. – Because of the severe cracking in the proximal region, and of the extremely low number of cycles applied to these specimens, no cracks developed in the distal region.
904
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
Fig. 5. Inducible micromotions (average and standard deviation, absolute value): for the Optimal specimens they were recorded throughout the simulated load history; for the Undersized specimens they were recorded when the tests was stopped due to macroscopic loosening (before 3% of the planned cyclic loading, as they exceeded the range of the displacement transducers). For comparison, previously published results (Bialoblocka-Juszczyk et al., 2010) for the Lubinus-SPII stem are plotted. While LVDT1, LVDT2, and LVDT3 recorded micromotions associated with a rotation of the stem about its long axis, the extensometer and LVDT4 measured the axial component.
on average, the number of cracks per unit volume of the Exeter with an Optimal mantle did not differ significantly from those of the Lubinus-SPII (Fig. 7, Kolmogorov–Smirnov, P = 0.98). However, if regions were analyzed separately, the Optimal Exeter implants had fewer cracks than the Lubinus-SPII in all regions except in the most distal part. In fact, if the cracks involving the distal centralizer were excluded, the Optimal Exeter
Table 2 Thickness of the cement mantle in the two types of implant, in the most proximal, mid and distal third of the stem length (average and standard deviation between slices and between specimens, in millimetres). Cement thickness was measured after test completion, on the sectioned specimens. Position along the stem
Position around the stem
Optimal specimens: average (standard deviation)
Undersized specimens: average (standard deviation)
Proximal third
Anterior Posterior Medial Lateral Anterior Posterior Medial Lateral Anterior Posterior Medial Lateral
2.16 1.88 4.39 4.74 2.22 1.8 1.62 3.87 2.60 2.23 2.19 2.86
1.3 (0.14) 1.16 (0.19) 2.13 (0.43) 1.95 (0.71) 1.36 (0.25) 1.64 (0.13) 2.38 (0.35) 2.08 (0.38) 2.15 (0.46) 2.13 (0.30) 2.71 (0.46) 1.86 (0.54)
Mid stem
Distal third
(0.86) (0.90) (2.09) (1.00) (1.06) (0.73) (0.72) (0.86) (0.45) (0.52) (1.20) (1.06)
implants had nearly ten times fewer cracks per unit volume than the Lubinus-SPII (Kolmogorov–Smirnov, P = 0.024). Another indicator of the extent of the fatigue damage is the extension of the cracks (Cristofolini et al., 2007a). The area of the crack surface was estimated based on the length of the crack visible on each face of the implant slices (Fig. 7). Indication of the amount of damage was obtained when the crack extension was related to the total cement volume (Cristofolini et al., 2007a). The crack extension per unit volume was uneven in the Optimal specimens: most damage was localized near the distal centralizer. Current results were compared against the Lubinus-SPII: on average, the crack extension per unit volume of the Exeter with an Optimal mantle did not differ significantly from that of the Lubinus-SPII (Fig. 7, Kolmogorov–Smirnov, P N 0.999). If regions were analyzed separately, the Optimal Exeter implants had less crack extension than the Lubinus-SPII in all regions, except the most distal part. In fact, if the distal centralizer was excluded, the Optimal Exeter implants had nearly eight times less damage than the LubinusSPII (Kolmogorov–Smirnov, P = 0.024).
4. Discussion The scope of this work was to assess if stem undersizing and a thin cement mantle would significantly increase the risk of implant loosening for a clinically successful hip stem (the highly polished cemented Exeter-V40).
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
905
Fig. 6. Inner cement surface after sectioning the femurs, with dye penetrants highlighting the cracks: typical Optimal (left) and Undersized (right) specimens. It must be noted that while the cracks of the Optimal specimens were observed after completion of the entire planned cyclic loading, the cracks in the Undersized specimens were observed after the test was stopped due to macroscopic loosening (in all cases before 3% of the planned cyclic loading was achieved).
In our study, Exeter-V40 stems of the Optimal size were compared to Undersized implants, under identical in vitro loading conditions. The Undersized implants underwent early loosening, with all the indicators of implant failure (macroscopic fractures of the cement mantle, and large and increasing implant-bone micromotions). All the indicators of cement fatigue indicated more extensive damage in the case of Undersized implants. None of the Undersized implants was able to withstand the in vitro test (all tests were stopped at most after 3% of the planned load history). It must be noted that even clinically unsuccessful designs such as the Müller-Curved stem completed the proposed mechanical test, although reporting extensive cement damage and large and increasing implant micromotions (Cristofolini et al., 2007a,c). Therefore, this can be assumed as an indicator of extremely negative outcome. Conversely, when the same stem was implanted with an Optimal size, cement damage was moderate. Some damage was also present in the Optimal specimens, but this did not seem to significantly compromise mantle integrity. The number of cracks per unit volume and their distribution was comparable to that found in vitro for other successful implants where the same methods were applied (Bialoblocka-Juszczyk et al., 2010; Cristofolini et al., 2007a). Indeed, cracks were found also around ex vivo well-fixed hip implants autopsy-retrieved (Jasty et al., 1991; Kawate et al., 1998). Also, the inducible micromotions and permanent migrations observed
for the Optimal implants were relatively small, and compatible with successful implants (Cristofolini et al., 2007c). The trend of migrations over time (Fig. 3) for the Undersized implants clearly showed a steep and loosening trend, while the Optimal implants tended to stabilize over time. Therefore, in our study it was confirmed that an undersized Exeter-V40 implant has higher chances of loosening than a correctly implanted one. As in this study only one stem type was tested under different implantation conditions, it is not possible to state whether such a dramatic change is to be expected for any cemented hip stem, or this is related to some specific feature of the Exeter-V40. Ongoing investigation on retrieved failed Exeter implants can help elucidating how many failures are affected by poor planning or implantation. The scenario observed for the Undersized implants in our in vitro study is more critical than one would expect based on clinical experience with the Exeter-V40 stem. In fact, our protocol did not aim at replicating a typical hip patient, but a very active one (Cristofolini et al., 2007c). The rationale is that failure is a rare event, which occurs in a small percentage of patients (2–10% fail between few weeks and 10 years post-op (Kärrholm et al., 2008; Stea et al., 2008)), typically the most demanding, heavy and active patients (Kilgus et al., 1991). Therefore, in order for an in vitro test to be sensitive to small design/implantation variations, such group of patients at risk should be simulated (Dorey and Amstutz, 2002)
906
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
Fig. 7. Indicators of cement damage for the Optimal Exeter specimens: the graphs on top detail the number of cracks per unit volume of bone cement (count/mm3); the lower graphs report the sum of crack extensions over total cement volume (mm2/mm3). The crack statistics are analyzed as follows: (i) distribution along the stem; (ii) distribution around the stem; (iii) summary for the entire specimen (both including and excluding the cracks involving the distal centralizer). The bars correspond to the average and standard deviation. Results of the Undersized specimens are not reported as the test was stopped due to macroscopic loosening (before 3% of the planned cyclic loading). For comparison, published results (Bialoblocka-Juszczyk et al., 2010) for the Lubinus-SPII stem are plotted.
by applying a very severe load history. This way, our in vitro protocol is capable of capturing the conditions under which such implants fail (Cristofolini et al., 2007a,c). In other words, our tests do not show that all undersized implants would fail, but that undersizing, in combination with active patients, is likely to induce early loosening. Smaller micromotions and reduced cement damage should be expected when lower loads are applied (both in vitro and in vivo) (Cristofolini et al., 2007a,c). In the present study, bone cement was mixed without application of vacuum. There is no consensus about the relevance of vacuum mixing, both in vitro and clinically (Ling and Lee, 1998; Messick et al., 2007). In all cases, if an effect exists, hand-mixing represents the worst-case clinical scenario in which the hip stems must still function well. As the same mixing conditions were applied to all the specimens in the present study, quantitative comparisons are possible between the two Exeter groups. Our findings concerning the detrimental effect of Undersized implants are in agreement with other numerical (Janssen et al., 2009), in vitro (Harrington et al., 2002), and clinical studies. A comparison of the results obtained in this study with previous ones for a different stem (Bialoblocka-Juszczyk et al., 2010) corroborates the present findings. In fact: – Comparable micromotions were measured in vitro by others: (Britton and Prendergast, 2005) and (Zhang et al., 2008) measured an axial migration of the order of 20–30 μm after one million cycles
–
–
–
–
(in both such studies a constant load amplitude was applied, which corresponded to the smallest peak of our study) with Exeter stems implanted in composite femurs. The Optimal Exeter implants had larger permanent migrations and inducible micromotions than the Lubinus-SPII (which was tested in an identical fashion in a previous study). This is somewhat expected, as the Exeter (and in general polished tapered stem) shows larger micromotions in vivo (Malchau et al., 1994; Williams et al., 2002). Numerical models predicted similar patterns (Stolk et al., 2007). Similar to our in vitro results for the Optimal specimens, also (Ling, 2006) reported that most cracks are localized near the Exeter distal centralizer. At the same time, the overall indicators of cement damage did not differ between the Optimal Exeter implants and the Lubinus-SPII: this is in agreement with the fact that comparable survival rates are reported for the two designs (Kärrholm et al., 2008; Stea et al., 2008). A similar result was found in vitro by (Gravius et al., 2008): the Optimal Exeter implants induced 20% less cracks than the Lubinus-SPII, but this difference was not statistically significant. An explanation to these observations was provided by (Huiskes et al., 1998), who introduced the concept of “force-closed design” for the Exeter, as opposed to “shape-closed design”. Force-closed designs tend to move more, but this is not necessarily associated to cement damage, as cement creep can accommodate relatively large migrations (Ling, 2006).
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
There are some limitations of this study that need to be discussed: – Accelerated test: our method simulated a selected range of activities, ignoring many others that are known for generating lower loads (e.g. walking, sitting down etc. (Bergmann et al., 2001)), as well as resting intervals. Load components were simplified to allow testing on a biaxial machine. It has been shown that the cement damage obtained using our in vitro testing is similar to that found in cement mantles retrieved upon revision hip surgery for the same stem type (Cristofolini et al., 2007a, 2003). However, for the Exeter stem, where cement creep is important, it is possible that our approach leads to underestimate implant migrations. In fact, migrations observed in this study are slightly larger but comparable to those found in other in vitro studies with similar limitations (Britton and Prendergast, 2005). – Composite femurs: The use of these bone models as a benchmark for comparative pre-clinical testing has been established in a large European Project (Maher and Prendergast, 2002; Waide et al., 2001), as they keep the same stiffness properties as human bones (Cristofolini et al., 1996) while allowing a significant reduction of the specimen variability compared to human bones. Previous research (Cristofolini, 1997; Harman et al., 1995) has demonstrated that the stability of the prosthesis and the sensitivity to the stem design are of the same order of magnitude in synthetic and human femurs. In conclusion, the hypothesis that undersizing the Exeter hip stem can cause implant loosening is corroborated. This has serious clinical implications: in fact, our results show that even an excellent stem can turn into a complete failure if implantation is not adequately planned and performed. Surgeons should always pay extra care in selecting the correct stem size, and in avoiding under-reaming of the femoral cavity. Conflict of interest statement This research was financially supported by Aesculap AG, Tuttlingen, Germany. However, this study does not aim at demonstrating that a prosthetic design is better or worse than any other. This study simply investigates whether the same stem, implanted improperly, can generate a higher failure risk. Acknowledgements The authors thank Luigi Lena for the illustrations. This research was financially supported by Aesculap AG. References Barrack, R.L., 2000. Early failure of modern cemented stems. J. Arthroplasty 15, 1036–1050. Bergmann, G., Deuretzbacher, G., Heller, M., Graichen, F., Rohlmann, A., Strauss, J., et al., 2001. Hip contact forces and gait patterns from routine activities. J. Biomech. 34, 859–871. Bialoblocka-Juszczyk, E., Cristofolini, L., Erani, P., Viceconti, M., 2010. Effect of long-term physiological activity on the long-term stem stability of cemented hip arthroplasty: in vitro comparison of three commercial bone cements. Proc. Inst. Mech. Eng. H 224, 53–65. Breusch, S.J., Lukoschek, M., Schneider, U., Ewerbeck, V., 2000. “State of the art” der zementierten Hüftendoprothetik: Qualität der Zementiertechnik ist entscheidend. Dtsch. Arzteblatt. 97, 2030–2033. Britton, J.R., Prendergast, P.J., 2005. Preclinical testing of femoral hip components: an experimental investigation with four prostheses. J. Biomech. Eng. 127, 872–880. Cristofolini, L., 1997. A critical analysis of stress shielding evaluation of hip prostheses. Crit. Rev. Biomed. Eng. 25, 409–483. Cristofolini, L., Viceconti, M., Cappello, A., Toni, A., 1996. Mechanical validation of whole bone composite femur models. J. Biomech. 29, 525–535. Cristofolini, L., Savigni, P., Saponara Teutonico, A., Toni, A., 2002. Ex-vivo and in-vitro cement mantle fatigue damage around femoral stems: validation of a protocol to simulate real-life loading in hip replacement patients. In: Bedzinski, R.P.C., Scigala, K. (Eds.), 13th Conference of the European Society of Biomechanics (Clinical
907
Biomechanics Award 2002). Oficyna Wydawnicza Politechniki Wroclawskiej Publ., Wroclaw, Poland, Wroclaw, Poland, pp. 48–49. Cristofolini, L., Saponara Teutonico, A., Monti, L., Cappello, A., Toni, A., 2003. Comparative in vitro study on the long term performance of cemented hip stems: validation of a protocol to discriminate between “good” and “bad” designs. J. Biomech. 36, 1603–1615. Cristofolini, L., Erani, P., Savigni, P., Bordini, B., Viceconti, M., 2007a. Pre-clinical assessment of the long-term endurance of cemented hip stems. Part 2: in-vitro and ex-vivo fatigue damage of the cement mantle. Proc. Inst. Mech. Eng. H 221, 585–599. Cristofolini, L., Erani, P., Savigni, P., Grupp, T., Thies, O., Viceconti, M., 2007b. Increased long-term failure risk associated with excessively thin cement mantle in cemented hip arthroplasty: a comparative in vitro study. Clin. Biomech. 22, 410–421. Cristofolini, L., Saponara Teutonico, A., Savigni, P., Erani, P., Viceconti, M., 2007c. Pre-clinical assessment of the long-term endurance of cemented hip stems. Part 1: effect of daily activities - a comparison of two load histories. Proc. Inst. Mech. Eng. H 221, 569–584. Dorey, F.D., Amstutz, H.C., 2002. Editorial: the need to account for patient activity when evaluating results of total hip arthroplasty with survivorship analysis. J. Bone Joint Surg. Am. 84-A, 709–710. Ebramzadeh, E., Sarmiento, A., McKellop, H.A., Llinas, A., Gogan, W., 1994. The cement mantle in total hip arthroplasty. Analysis of long-term radiographic results. J. Bone Joint Surg. Am. 76, 77–87. Ebramzadeh, E., Sangiorgio, S.N., Longjohn, D.B., Buhari, C.F., Dorr, L.D., 2004. Initial stability of cemented femoral stems as a function of surface finish, collar, and stem size. J. Bone Joint Surg. Am. 86-A, 106–115. Gheduzzi, S., Nedungayil, S., Learmonth, I.D., 2004. Know your cement: a review of current cementing practice in the UK. Hip Int. 14, 128–129. Giannikas, K.A., Din, R., Sadiq, S., Dunningham, T.H., 2002. Medium-term results of the ABG total hip arthroplasty in young patients. J. Arthroplasty 17, 184–188. Gravius, S., Wirtz, D.C., Siebert, C.H., Andereya, S., Mueller-Rath, R., Maus, U., et al., 2008. In vitro interface and cement mantle analysis of different femur stem designs. J. Biomech. 41, 2021–2028. Harman, M.K., Toni, A., Cristofolini, L., Viceconti, M., 1995. Initial stability of uncemented hip stems: an in-vitro protocol to measure torsional interface motion. Med. Eng. Phys. 17, 163–171. Harrington, M.A.J., O'Connor, D.O., Lozynsky, A.J., Kovach, I., Harris, W.H., 2002. Effects of femoral neck length, stem size, and body weight on strains in the proximal cement mantle. J. Bone Joint Surg. Am. 84, 573–579. Hertzler, J., Miller, M.A., Mann, K.A., 2002. Fatigue crack growth rate does not depend on mantle thickness: an idealized cemented stem construct under torsional loading. J. Orthop. Res. 20, 676–682. Huiskes, R., Verdonschot, N., Nivbrant, B., 1998. Migration, stem shape, and surface finish in cemented total hip arthroplasty. Clin. Orthop. Relat. Res. 103–112. Janssen, D., van Aken, J., Scheerlinck, T., Verdonschot, N., 2009. Finite element analysis of the effect of cementing concepts on implant stability and cement fatigue failure. Acta Orthop. 80, 319–324. Jasty, M., Maloney, W.J., Bragdon, C.R., O'Connor, D.O., Haire, T., Harris, W.H., 1991. The initiation of failure in cemented femoral components of hip arthroplasties. J. Bone Joint Surg. Br. 73, 551–558. Kärrholm, J., G., G., C., R., P., H. Swedish hip arthroplasty register. Annual report 2007 (available from: http://www.jru.orthop.gu.se/): Swedish HIp Arthoplasty Register; 2008. Kassi, J.-P., Heller, M.O., Stoeckle, U., Perka, C., Guda, G.N., 2005. Stair climbing is more critical than walking in pre-clinical assessment of primary stability in cementless THA in vitro. J. Biomech. 38, 1143–1154. Kawate, K., Maloney, W.J., Bragdon, C.R., Biggs, S., Jasty, M., Harris, W.H., 1998. Importance of a thin cement mantle. Autopsy studies of eight hips. Clin. Orthop. Relat. Res. 355, 70–76. Kilgus, D.J., Dorey, F.D., Finerman, G.A.M., Amstutz, H.C., 1991. Patient activity, sports participation, and impact loading on the durability of cemented total hip replacement. Clin. Orthop. Relat. Res. 269, 25–31. Lattanzi, R., Grazi, E., Testi, D., Viceconti, M., Cappello, A., Toni, A., 2003. Accuracy and repeatability of cementless total hip replacement surgery in patients with deformed anatomies. Med. Inform. Internet Med. 28, 59–71. Ling, R.S.M., 2006. The History and Development of the Exeter Hip2nd Edition. Stryker, UK. Ling, R.S., Lee, A.J., 1998. Porosity reduction in acrylic cement is clinically irrelevant. Clin. Orthop. Relat. Res. 249–253. Maher, S.A., Prendergast, P.J., 2002. Discriminating the loosening behaviour of cemented hip prosthesis using measurements of migration and inducible displacement. J. Biomech. 35, 257–265. Malchau, H., Snorrason, F., Herberts, P., Karrholm, J., 1994. Micromotion of femoral stems in total hip arthroplasty. A randomized study of cemented, hydroxyapatitecoated, and porous-coated stems with roentgen stereophotogrammetric analysis. J. Bone Joint Surg. Am. 76, 1692–1705. Mann, K.A., Gupta, S., Race, A., Miller, M.A., Cleary, R.J., Ayers, D.C., 2004. Cement microcracks in thin-mantle regions after in vitro fatigue loading. J. Arthroplasty 19, 605–612. Messick, K.J., Miller, M.A., Damron, L.A., Race, A., Clarke, M.T., Mann, K.A., 2007. Vacuum-mixing cement does not decrease overall porosity in cemented femoral stems: an in vitro laboratory investigation. J. Bone Joint Surg. Br. 89, 1115–1121. Monti, L., Cristofolini, L., Viceconti, M., 1999. An in-vitro quantitative analysis of the primary stability of uncemented hip stems. Artif. Organs 23, 851–859. Peirce, B., 1852. Criterion for the rejection of doubtful observations. Astron. J. 2, 161–163. Ramaniraka, N.A., Rakatomanana, L.R., Leyvraz, P.-F., 2000. The fixation of cemented femoral component. Effect of stem stiffness, cement thickness and roughness of the cement–bone surface. J. Bone Joint Surg. Br. 82-B, 297–303. Ross, S.M., 2003. Peirce's criterion for the elimination of suspect experimental data. J. Eng. Tech. 2003, 1–12.
908
L. Cristofolini et al. / Clinical Biomechanics 25 (2010) 899–908
Ruff, C.B., Hayes, W.C., 1983. Cross-sectional Geometry of Pecos Pueblo Femora and Tibiae — a biomechanical investigation: I. Method and general patterns of variation. Am. J. Phys. Anthropol. 60, 359–381. Scheerlinck, T., Druyts, P., Casteleyn, P.P., 2004. The use of primary total hip arthroplasty in university hospitals of the European Union. Acta Orthop. Belg. 70, 231–239. Scheerlinck, T., de Mey, J., Deklerck, R., Noble, P.C., 2006. CT analysis of defects of the cement mantle and alignment of the stem: in vitro comparison of Charnley– Kerboul femoral hip implants inserted line-to-line and undersized in paired femora. J. Bone Joint Surg. Br. 88, 19–25. Stea, S., Bordini, B., De Clerico, M., Toni, A., 2008. Report of orthopaedic prosthetic implantology. Overall data: hip, knee and shoulder arthroplasty in Emilia Romagna region: Istituto Ortopedico Rizzoli. Stolk, J., Verdonschot, N., Cristofolini, L., Toni, A., Huiskes, R., 2002a. Finite element and experimental models of cemented hip joint reconstructions can produce similar bone and cement strain in pre-clinical tests. J. Biomech. 35, 499–510. Stolk, J., Verdonschot, N., Huiskes, R., 2002b. Stair climbing is more detrimental to the cement in hip replacement than walking. Clin. Orthop. Relat. Res. 405, 294–305.
Stolk, J., Janssen, D., Huiskes, R., Verdonschot, N., 2007. Finite element-based preclinical testing of cemented total hip implants. Clin. Orthop. Relat. Res. 456, 138–147. Verdonschot, N.J., 1996. Biomechanical failure Scenarios for Cemented Total Hip Replacement. Ponsen & Looijen bv publ, Wageningen. Viceconti, M., Lattanzi, R., Antonietti, B., Paderni, S., Olmi, R., Sudanese, A., et al., 2003. CT-based surgical planning software improves the accuracy of total hip replacement preoperative planning. Med. Eng. Phys. 25, 371–377. Waide, V., Cristofolini, L., Toni, A., 2001. A CAD-CAM methodology to produce bone remodeled composite femurs for preclinical investigations. Proc. Inst. Mech. Eng. H 215, 459–469. Williams, H.D., Browne, G., Gie, G.A., Ling, R.S., Timperley, A.J., Wendover, N.A., 2002. The Exeter universal cemented femoral component at 8 to 12 years. A study of the first 325 hips. J. Bone Joint Surg. Br. 84, 324–334. Zhang, H.Y., Blunt, L., Jiang, X.Q., Brown, L., Barrans, S., Zhao, Y., 2008. Femoral stem wear in cemented total hip replacement. Proc. Inst. Mech. Eng. H 222, 583–592.