- Email: [email protected]
Patterns of stress distribution at the proximal femur after implantation of a modular neck prosthesis. A biomechanical study
Clinical Biomechanics 28 (2013) 415–422
Contents lists available at SciVerse ScienceDirect
Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech
Patterns of stress distribution at the proximal femur after implantation of a modular neck prosthesis. A biomechanical study Angelos N. Politis a, b,⁎, George K. Siogkas b, c, Ioannis D. Gelalis b, Theodore A. Xenakis b a b c
Department of Orthopaedic Surgery, Jewish General Hospital, McGill University, Montreal, Canada Laboratory of Biomechanics, University of Ioannina, School of Medicine, Ioannina, Greece Department of Electrical and Computer Engineering, University of Patras, Greece
a r t i c l e
i n f o
Article history: Received 30 July 2012 Accepted 5 February 2013 Keywords: Strain Modular Neck Femur Arthroplasty
a b s t r a c t Background: Modular total hip arthroplasty incorporating a double taper design is an evolution offering potential advantages compared to single head–neck taper or monolithic designs. Changes in femoral offset, neck length or femoral anteversion are expected to alter the strain distribution. Methods: We therefore analyzed the strain patterns after usage of all types of necks of a modular neck prosthesis, implanted in composite femurs. Findings: The load distribution presented a repeatable pattern. Anteverted neck combinations resulted in higher stress at the anterior surface, whereas the retroverted ones at the posterior (e.g. at the middle frontal site, stress is 13.63% higher when we shifted from the long neutral neck to the long 15° anteverted neck and at the middle back site 19.73% higher when we shifted from the long neutral to the long 15° retroverted neck). Compressive stress was larger at the calcar region and exacerbated by the use of the varus neck (e.g. at the frontal 1 site stress increased by 44.01% when we used the long 8° varus neck in comparison to the long neutral neck). Anteverted neck combinations resulted in higher strain at the anterior cortex around the tip of the prosthesis. Short necks exhibited lower stress at the femoral shaft and higher at the trans-trochanteric area. Interpretation: Anteverted neck combinations could be more prone to anterior thigh pain. Because of the possible risk of adaptive hypertrophy and early mechanical failure due to increased stress, the surgeon should be cautious when using necks with combined characteristics or short necks. Crown Copyright © 2013 Published by Elsevier Ltd. All rights reserved.
1. Introduction Modular total hip arthroplasty (THA) incorporating a double taper design is an innovation offering potential advantages compared to single head–neck taper or monolithic designs (Dunbar, 2010). Those include the adjustment of leg length and offset via the head–neck taper, femoral anteversion via the neck–stem taper, easier revision when there is no need to revise a well-fixed femoral stem and optimal restoration of soft tissue tension and patient biomechanics (Dunbar, 2010). The use of modular necks has thus increased in the recent years and authors reported good mid- and long-term clinical outcomes (Benazzo et al., 2010; Sakai et al., 2010). Adjusting femoral offset, leg length and orientation of the components are of crucial importance. Offset correlates to abductor muscle function, wear and impingement (Dastane et al., 2011). On the other hand, over-lengthening of the limb can be a problem (Konyves and Bannister, 2005), whereas failure to comply with the recommendations for acetabular inclination/anteversion and femoral anteversion may ⁎ Corresponding author at: Department of Orthopaedic Surgery, Jewish General Hospital, McGill University, 447, Ave Prince Albert, Apt. 4, Westmount, Montreal, Canada H3YP6. E-mail address: [email protected] (A.N. Politis).
lead to edge loading and prosthetic impingement, which can cause dislocation, mechanical loosening, wear or breakage of the polyethylene liner, metallosis or metal ion release in metal-on-metal bearings, and squeaking or breakage of ceramic-on-ceramic bearings (De Haan et al., 2008; Miki and Sugano, 2011). Are changes in femoral offset, neck length or femoral anteversion expected to alter the strain distribution at the femur? The aim of this biomechanical study is to analyze and compare the strain patterns at the proximal femur after usage of all types of necks of a commercially available modular neck stem prosthesis and possibly connect the results with the clinical praxis. 2. Methods In this study a set of cementless modular PROFEMUR-E® Total Hip Replacement System (Wright Medical Technology Inc., Arlington, TN, USA) was used. The aforementioned stem is a modular prosthesis manufactured from Ti6Al4V and has a 500 μm thick coating of pure titanium plasma spray. The surgeon may choose between six interchangeable necks available in two lengths and a total of 22 combinations can be used. We used custom-made, commercially available, medium left, fourth generation, medium composite femoral models
0268-0033/$ – see front matter. Crown Copyright © 2013 Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.clinbiomech.2013.02.004
416
A.N. Politis et al. / Clinical Biomechanics 28 (2013) 415–422
(model # 3403-99, Sawbones, Pacific Research Laboratories, Vashon Island, WA, USA). The bones were already osteotomized and specifically machined for use with a size 3 Profemur-E stem. A custom-made fixture designed to reproduce loading conditions during the single-leg stance phase of walking, as described by McLeish and Charnley (1970), was attached to the load cell of a computer-controlled hydraulic testing machine (MTS 858 Mini Bionix, MTS Systems Corp, Eden Prairie, MN, USA). The femur was tilted into 12° of valgus and was positioned neutral on the sagittal plane. Hip abductors were simulated by a small chain attached to a custom-made base that was fixed to the lateral aspect of the greater trochanter. The abductor force simulation applied the load at an angle of abduction 15° to the sagittal plane (Finlay et al., 1989; McLeish and Charnley, 1970). The distal end of the femur was embedded in a steel pot with radiopaque bone cement. A modified universal ball joint was mounted between the distal construct and the base of the machine (Fig. 1). The circumference of the femoral model was divided into 3 parts and strain gages were fixed along the lateral, medial-anterior and medial-posterior surface of the femur at positions 60° apart. Three 350-Ohm tri-axial rosette strain gages (KYOWA, KFG-2-350-D17-11, Kyowa Electronic Instrument, Tokyo, Japan) were bonded on the transtrochanteric surface, where a more complex strain pattern was expected. One rosette was made up of three strain gages mounted at 60° angles. The median of the peak-to-peak value of the sinusoidal strain over time was computed for each of the three gages. Uni-axial 350-Ohm strain gages (KYOWA, KFG-2-350-C1-11, Kyowa Electronic Instrument, Tokyo, Japan) were used along the shaft of the femur, where the strain pattern was expected to be simpler. The uni-axial strain gages were distributed at three horizontal levels at 48, 96, and 144 mm below the level of the lesser trochanter, so that the middle gages were around the tip of the stem. The leads of the gages were connected to a Wheatstone bridge configuration (Kyowa SS-24R Switching and Balancing Box, Kyowa Electronic Instrument, Tokyo, Japan). The gage outputs were transferred to a signal amplifier module and consequently to an (MTS TestStar II® data acquisition system, MTS Systems Corp, Eden Prairie, MN, USA). Load cycles were programmed to simulate single-leg stance of a normal-weight subject. Applying a vertical force five sixths of the body weight, with the weight of the lower extremity subtracted, would yield a physiological resultant hip joint force in the hip simulator (Wik et al., 2011). Thus, an axial load of 600 N was applied to simulate the single-leg stance of a subject weighing 70 kg. A 3-step
testing sequence was used as follows: 1. Ramp-up to − 300 N (rate 100 N/s). 2. Sinusoidal axial loading between − 100 and − 600 N applied at a frequency of 1 Hz for 200 cycles. 3. Ramp-down to − 50 N (rate 500 N/s). The material testing system operated under force control. Load cell limit was set at − 750 N. Strain values were recorded for 200 full cycles. Tests were repeated three times for every composite bone to obtain the average strain for each gage. The set of the three tests was repeated on three different composite femurs. The position of the strain gages was checked constant with the use of a phantom femoral model. All conditions were tested on every composite bone. Prior to testing, the abductor chain was pre-tensioned until the level arm was balanced at the horizontal plane in the beginning of every load cycle. Three specimens were tested and statistics were performed on the three sets of data obtained from the three specimens. Every set of data included all neck variations and was retrieved from the same specimen. The following neck variations were tested: • • • • •
•
• • • • •
•
Long neutral Long 8° anteverted (long 8 DG A)/long 8° retroverted (long 8 DG R) Long 8° varus (long 8 DG VAR)/long 8° valgus (long 8 DG VAL) Long 15° anteverted (long 15 DG A)/long 15° retroverted (long 15 DG R) Long varus valgus 1 anteverted (long VAR VAL 1 A = anteverted and valgus)/long varus valgus 1 retroverted (long VAR VAL 1 R = retroverted and varus) Long varus valgus 2 anteverted (long VAR VAL 2 A = anteverted and varus)/long varus valgus 2 retroverted (long VAR VAL 2 R = retroverted and valgus) Short neutral Short 8° anteverted (short 8 DG A)/short 8° retroverted (short 8 DG R) Short 8° varus (short 8 DG VAR)/short 8° valgus (short 8 DG VAL) Short 15° anteverted (short 15 DG A)/short 15° retroverted (short 15 DG R) Short varus valgus 1 anteverted (short VAR VAL 1 A = anteverted and valgus)/short varus valgus 1 retroverted (short VAR VAL 1 R = retroverted and varus) Short varus valgus 2 anteverted (short VAR VAL 2 A = anteverted and varus)/short varus valgus 2 retroverted (short VAR VAL 2 R = retroverted and valgus)
One-way analysis of variance (ANOVA) of the data was performed using MatLab (The MathWorks Inc, Natick, MA, USA). Post-hoc analysis was performed using the Tukey–Kramer test. A P value b 0.05 was considered statistically significant. 3. Results 3.1. Long neutral neck vs long 8° anteverted neck (long 8 DG A) vs long 8° retroverted neck (long 8 DG R) (Fig. 2, Table 1)
Fig. 1. Experimental set-up.
Strain measurement of the uni-axial strain gages revealed that longitudinal deformation was compressive on the medial side of the femur and tensile on the lateral. Strain analysis from the rosette gages on the anterior surface of the trans-trochanteric area revealed that gages 1 and 2 (the gage parallel to the longitudinal axis of the femur and the gage 60° deviated towards the medial side) showed compressive signal, while gage 3 (the gage 60° deviated towards the lateral side) showed tensile signal. Strain analysis from the rosette gages on the posterior surface of the trans-trochanteric area revealed that strain gage 1 (the gage 60° deviated towards the lateral side) showed tensile signal, while strain gages 2 and 3 (the gage parallel to the longitudinal axis of the femur and the gage 60° deviated towards the medial side) showed compressive signal. A consistent finding was that the long 8 DG A neck “conducted” stresses towards the anterior surface of the femur, while the long 8 DG R neck towards the posterior.
A.N. Politis et al. / Clinical Biomechanics 28 (2013) 415–422
417
Fig. 2. Long neutral vs long 8° anteverted vs long 8° retroverted. The aim of the comparison is to examine the effect of the change of the neck anteversion on the femoral strain.
This was valid for all anterior and posterior uni-axial strain gages along the shaft. The long neutral neck presented intermediate signal between the A and the R neck. For example, at the middle frontal strain gage site, stress was 13.63% higher when we shifted from the long neutral neck to the long 15 DG A neck and at the middle back strain gage site 19.73% higher when we shifted from the long neutral to the long 15 DG R neck. This principle was not valid for the trans-trochanteric region, where no correlation between the version of the prosthetic neck and transmission of stress was found. Stress was less on the anterior than the posterior surface of the trans-trochanteric region (e.g. comparison at the frontal 2 and back 2 strain gage sites revealed that using the long neutral neck resulted in 37.62% higher stress on the posterior surface in comparison to the anterior). Statistical analysis revealed that there were statistically significant differences between the three necks at all strain gage sites. 3.2. Long neutral neck vs long 15° anteverted neck (long 15 DG A) vs long 15° retroverted neck (long 15 DG R) The exact same conclusions were drawn in the aforementioned comparison. The pattern of strain distribution was the same and statistical analysis revealed that there were statistically significant differences between the three necks at all strain gage sites, but back 1 strain gage at the posterior trans-trochanteric area. 3.3. Long neutral neck vs long 8° anteverted neck (long 8 DG A) vs long 15° anteverted neck (long 15 DG A) (Fig. 3, Table 2) Our data revealed that stress on the anterior surface of the femur was greater using the long 15 DG A neck than using the long 8 DG A neck than using the long neutral neck. This was valid at the femoral
shaft. For example, at the bottom frontal strain gage site, stress was 16.97% and 38.5% higher when we shifted from the long neutral to the long 8 DG A and to the long 15 DG A neck respectively. Stress at the posterior surface had the opposite pattern. This was nevertheless again not valid for the trans-trochanteric region, where stress pattern was more irregular. There were statistically significant differences between the three necks at all strain gage sites. 3.4. Long neutral neck vs long 8° retroverted neck (long 8 DG R) vs long 15° retroverted neck (long 15 DG R) Similarly to the previous comparison, stress at the posterior surface of the model was greater using the long 15 DG R neck than using the long 8 DG R than using the long neutral. This was again valid at the shaft of the specimen. Stress at the anterior surface had the opposite pattern. The aforementioned pattern was once more not valid for the trans-trochanteric region. There were statistically significant differences between the three necks at all strain gage sites. 3.5. Long neutral neck vs long 8° varus neck (long 8 DG VAR) vs long 8° valgus neck (long 8 DG VAL) (Fig. 4, Table 3) Comparison of these three necks leads to the conclusion that the differences in the stress at the femoral shaft were smaller. For example, at the upper frontal strain gage site shifting from the long neutral neck to the long 8 DG VAR and to the long 8 DG VAL neck resulted in a 1.23% and 3.96% strain difference respectively. We observed that stress was larger at the posterior than the anterior surface of the model and this was also true for the trans-trochanteric area (e.g. using the long 8 DG VAR neck the stress at the upper back strain gage site was 107.5% higher compared to the stress at the upper frontal strain
Table 1 Indicative strains for the long neutral vs long 8° anteverted vs long 8° retroverted comparison (+ = compressive strain/− = tensile strain). The median of the peak-to-peak value of the sinusoidal strain over time εmax (με) is indicated. Neck/strain gage
Middle frontal
Middle back
Frontal 2
Back 2
1: Long neutral (SD) 2: Long 8 DG A (SD) 3: Long 8 DG R (SD) P Couples P b 0.05
+616 (±3.55) +700 (±8.12) +486 (±2.71) 0.000000 1–2, 1–3, 2–3
+608 (±3.17) +465 (±5.11) +728 (±2.90) 0.000000 1–2, 1–3, 2–3
+404 (±3.91) +329 (±3.04) +186 (±12.36) 0.000000 1–2, 1–3, 2–3
+556 (±3.02) +379 (±249) +572 (±21.08) 0.000000 1–2, 1–3, 2–3
418
A.N. Politis et al. / Clinical Biomechanics 28 (2013) 415–422
Fig. 3. Long neutral vs long 8° anteverted vs long 15° anteverted. The aim of the comparison is to examine the effect of the change of the amount of the neck anteversion on the femoral strain.
gage site). A safe conclusion was that compressive stress was larger at the calcar region and this was exacerbated by the use of the long 8 DG VAR neck (e.g. at the frontal 1 strain gage site stress increased by 44.01% when we used the long 8 DG VAR neck in comparison to the long neutral neck). This was also affirmed at the posterior surface of the trans-trochanteric region. 3.6. Long neutral neck vs long varus valgus 1 anteverted neck (long VAR VAL 1 A = anteverted and valgus) vs long varus valgus 1 retroverted neck (long VAR VAL 1 R = retroverted and varus) (Fig. 5, Table 4) Usage of necks with combined characteristics led to a synergistic stress pattern. The anteverted component of the neck led to larger strain at the anterior surface of the model, whereas the retroverted at the posterior. Similarly, the varus component led to higher strain at the calcar region. The use of necks with combined characteristics led to higher stress at the trans-trochanteric region (e.g. at the frontal 2 strain gage site stress increased by 48.51% and 50.74% when we shifted from the long neutral neck to the long VARVAL1A and to the long VARVAL1R neck respectively). This observation applied to all comparisons, concerning necks with combined characteristics: - Long neutral vs long VAR VAL 2 A (A and VAR) vs long VAR VAL 2 R (R and VAL) - Long VAR VAL 1 A (A and VAL) vs long 8 DG A vs long 8 DG VAL - Long VAR VAL 1 R (R and VAR) vs long 8 DG R vs long 8 DG VAR - Long VAR VAL 2 A (A and VAR) vs long 8 DG A vs long 8 DG VAR - Long VAR VAL 2 R (R and VAL) vs long 8 DG R vs long 8 DG VAL
3.7. Long vs short necks (Fig. 6) A general observation in all comparisons between long and short necks was that use of long necks resulted in higher stress at the shaft of the femur, whereas use of short necks led to higher stresses at the trans-trochanteric area. For example, at the middle frontal strain gage site stress increased by 6.2% when we shifted from the short neutral to the long neutral neck and at the Lateral 2 strain gage site the stress increased by 14.78% when we shifted from the long neutral to the short neutral neck. This finding applied to all comparisons between short and long necks. 4. Discussion Modularity of the femoral neck is increasing in popularity with implants available from several manufacturers. Modular stems allow increased flexibility by the surgeon, where a standard design with a fixed neck geometry will not restore the optimal hip biomechanics. Clinical advantages of these stems include the adjustment of leg length and offset via the head–neck taper, femoral anteversion via the neck–stem taper, easier revision when there is no need to revise a well-fixed femoral stem and optimal restoration of soft tissue tension and patient biomechanics. On the other hand, concerns over modularity have been reported, including the risk of mechanical failure of the trunion, disassociation of the components and femoral subsidence. Other possible complications include the generation of particle debris, metal ions, fretting and crevice corrosion. It has been reported that even with modern taper designs and corrosion-resistant materials,
Table 2 Indicative strains for the long neutral vs long 8° anteverted vs long 15° Αnteverted comparison (+ = compressive strain/− = tensile strain). The median of the peak-to-peak value of the sinusoidal strain over time εmax (με) is indicated. Neck/strain gage
Bottom frontal
Bottom back
Lateral 2
Frontal 2
1: Long neutral (SD) 2: Long 8 DG A (SD) 3: Long 15 DG Α (SD) P Couples P b 0.05
+483 (±3.21) +565 (±7.00) +669 (±3.58) 0.000001 1–2, 1–3, 2–3
+795 (±3.74) +625 (±7.07) +594 (±3.26) 0.000001 1–2, 1–3, 2–3
−1150 (±6.84) −949 (±6.33) −1170 (±4.44) 0.000000 1–2, 1–3, 2–3
+404 (±3.91) +329 (±3.04) +641 (±3.16) 0.000000 1–2, 1–3, 2–3
A.N. Politis et al. / Clinical Biomechanics 28 (2013) 415–422
419
Fig. 4. Long neutral vs long 8° varus vs long 8° valgus. The aim of the comparison is to examine the effect of the change of the caput-collum-diaphyseal angle on the femoral strain.
corrosion, fretting and particulate debris were observed to a greater extent in the second neck–stem junction and that the degradation of the junction can contribute to metallosis and generation of the aseptic lymphocyte-laminated vascular-associated lesions that can lead to revision (Kop et al., 2012). Even if a change in the stress pattern at the femur is expected, it is not known whether the different neck orientation results in a significant change in bone loading. Measurement of bone surface strains is an indirect way to elucidate the implications that exist in the literature about extreme biomechanical behavior after usage of certain types of modular necks, such as the long varus or retroverted necks (Ellman and Levine, 2013; Wilson et al., 2010). It is reported that the long VAR neck component increases the bending moment by 32.7%, when compared to the short VAR neck and that failures of modular necks are attributed to neck offset (neck length+ head length) (Traina et al., 2009). A recent study from the Swedish Arthroplasty Registry has reported an increased revision rate for small-sized stems with increased neck length (Thien and Kärrholm, 2010). According to our measurements, the aforesaid necks (long 15 DG R and long 8 DG VAR) did not exhibit extreme loading behavior, with the maximum microstrain value recorded, being εmax = − 1440με. Another clinical issue is the possible link of the use of the different neck configurations of this specific prosthesis with the development of stress shielding. Preservation of the proximal bone stock is a fundamental goal after THA (Kim et al., 2001). The largest loss in bone mineral density is reported to occur in the proximomedial aspect of the femur, while evidence exists that a significant amount of bone loss may also occur in the proximolateral (Gibbons et al., 2001). The area below the calcar is also particularly vulnerable in strain decrease
(Aamodt et al., 2001; Kim et al., 2001). Clinically, the consequences of stress shielding are not entirely clear, although it is often cited as a reason for concern causing osteolysis, and thus aceptic loosening in the long term. The exact level of strain required to alter or preserve bone remodeling is not known. We observed that stress was larger at the posterior than the anterior surface of the model and this was also true for the trans-trochanteric area. This makes this area particularly vulnerable to stress shielding. According to our data, the lateral or posterior transtrochanteric area are in lower risk for stress shielding, as they did not exhibit very low strains in any neck combination. Studies conducted to understand the response of human femurs under loading conditions, reported that loading led to a bending load in the proximal two-thirds of the femur, whereas almost homogeneous strain conditions prevail in the lower third. It is cited that longitudinal strain decreases from proximal to distal on the surface of the native femur in vitro (Helwig et al., 2011). On the contrary, Kim et al. (2001) found that strain increased from proximal to distal in the intact femur, while Jasty et al. (1994) reported that strain was relatively uniform along the diaphysis of the femur. In our study, strains were compressive at the medial side and tensile at the lateral side and this finding was consistent for all neck comparisons. This was observed for both the transtrochanteric area and the shaft of the femur. Our data suggest that the load distribution pattern does not increase nor decrease in a uniform manner along the shaft, but presents a certain pattern that was validated throughout our study: at the anterior surface of the femur the load distribution is the following: strain at the middle horizontal level (around the tip of the prosthesis) greater than strain at the lower horizontal level (the mid-shaft of the femur) greater than strain at the upper horizontal level (around the middle of the prosthesis). The posterior surface presents the opposite strain
Table 3 Indicative strains for the long neutral vs long 8° varus vs long 8° valgus comparison (+ = compressive strain/− = tensile strain). The median of the peak-to-peak value of the sinusoidal strain over time εmax (με) is indicated. Neck/strain gage
Frontal 1
Back 2
Upper back
Upper frontal
1: Long neutral (SD) 2: Long 8 DG VAR (SD) 3: Long 8 DG VΑL (SD) P Couples P b 0.05
+443 (±3.68) +638 (±3.12) +453 (±2.62) 0.000000 1–2, 1–3, 2–3
+557 (±3.02) +756 (±3.24) +623 (±3.18) 0.000000 1–2, 1–3, 2–3
+940 (±4.10) +828 (±3.77) +968 (±4.98) 0.000003 1–2, 1–3, 2–3
+404 (±2.49) +399 (±1.77) +388 (±3.02) 0.060422 –
420
A.N. Politis et al. / Clinical Biomechanics 28 (2013) 415–422
Fig. 5. Long neutral vs long varus valgus anteverted 1 (anteverted and valgus) vs long varus valgus retroverted 1 (retroverted and varus). The aim of the comparison is to examine the effect of the use of the necks with combined characteristics on the femoral strain.
pattern: strain at the upper horizontal level greater than strain at the lower horizontal level greater than strain at the middle level. We deduce that the area around the tip of the prosthesis at the anterior cortex that presents the higher strain values is the area of load transfer from the prosthesis to the bone, and is subject to a higher risk of provoking anterior thigh pain. This is in agreement with the fact that thigh pain is perceived at the anterior side of the femur in most patients. On the contrary the pattern of stress distribution at the transtrochanteric area was more complex throughout the experiment, and no safe conclusions can be drawn. We attribute this to the fact that three rosette strain gages provide tri-axial strain measurement in only three points of this, anatomically complex, part of the femoral bone and cannot depict the true strain distribution in the area. Furthermore, the transtrochanteric area is subject to strain exerted from the hip abductors to the greater trochanter. To the best of our knowledge, the biomechanical behavior of femoral stems with modular necks was examined in two previous studies. Simpson et al. (2009) performed a finite element analysis on the strain pattern after use of modular necks in anteversion, retroversion and of two different offsets. They reported that strain on the medial bone surface increased distally along the femur. Overall, lower strains were observed on the medial bone when an anteverted neck was used, as compared to a retroverted neck, and the authors concluded that the different neck geometries did not result in a significant change in bone mechanics. Umeda et al. (2003) performed strain gage measurements on composite femurs implanted with a cementless stem and combined with retroverted and anteverted femoral necks. The authors reported that on the medial aspect compressive strains were much larger than tensile strains, while on the lateral aspect tensile strains were much larger than compressive strains regardless of neck version. On the
aspect toward which the prosthetic neck was oriented compressive strains tended to be larger than tensile, and on the opposite side tensile strains tended to be larger than compressive. This was most marked around the level of the stem tip. Our data agree with Umeda et al. The anteverted necks in our study conducted stress towards the anterior surface, while the retroverted ones towards the posterior. This was valid for all uni-axial strain gages along the shaft of the femur. Neutral necks presented an intermediate signal between the anteverted and the retroverted ones. This principle was once more not valid for the trans-trochanteric region and this is, to the authors' judgment, due to the complex anatomy of the area. Our data also verified the assumption that a 15 DG ante/retroverted neck swifts more stress towards the direction of neck orientation, than an 8 DG neck. According to our data, the VAR/VAL neck component had a smaller effect on the stress at the femoral shaft than the anteverted/retroverted neck variation and a constant finding was that compressive stress was larger at the calcar region and calcar strain was exacerbated by the use of the 8 DG VAR neck. This is in accordance with the universal agreement that a large percentage of strain after THA is transmitted through the crucial calcar region. Significant thigh pain after implantation of a cementless device occurs in as many as 22% of patients and is an important concern. Focal mismatch in flexural rigidity around the stem tip has been implicated as an important factor contributing to mid-thigh pain, but sources of pain may also include implant micromotion and instability (Campbell et al., 1992). There are no published data about the strain threshold that can possibly elicit thigh pain. According to the authors' opinion, high strains may lead to thigh pain via “channeling” high values of stress through the anterior femoral cortex, whereas low strains may be indicative of a possible loosening and instability of
Table 4 Indicative strains for the long neutral vs long varus valgus anteverted 1 (anteverted and valgus) vs long varus valgus retroverted 1 (retroverted and varus) comparison (+ = compressive strain/− = tensile strain). The median of the peak-to-peak value of the sinusoidal strain over time εmax (με) is indicated. Neck/strain gage
Middle frontal
Middle back
Frontal 1
Frontal 2
1: Long neutral (SD) 2: Long VARVAL 1 A (SD) 3: Long VARVAL 1 R (SD) P Couples P b 0.05
+616 (±3.55) +633 (±3.22) +565 (±48.70) 0.000023 1–2, 1–3, 2–3
+605 (±3.17) +527 (±3.29) +686 (±50.28) 0.000000 1–2, 1–3, 2–3
+442 (±3.68) +550 (±9.88) +576 (±5.41) 0.000000 1–2, 1–3, 2–3
+404 (±3.91) +597 (±9.99) +609 (±6.08) 0.000000 1–2, 1–3
A.N. Politis et al. / Clinical Biomechanics 28 (2013) 415–422
421
Fig. 6. Long neutral vs short neutral. The aim of the comparison is to examine the effect of the change of the neck length on the femoral strain.
the prosthesis, leading again to thigh pain. Analysis of our data showed that anteverted neck combinations result in higher strain at the anterior cortex around the tip of the prosthesis and could thus be more prone to anterior thigh pain. This was more intense with the use of the 15 DG anteverted necks. For example, at the middle frontal strain gage site which is around the tip of the prosthesis, strain increased by 13.63% and 34.25% when we shifted from the long neutral neck to the long 8 DG A and to the long 15 DG A respectively. Another assumption confirmed was that usage of combined necks will lead to a more complex stress pattern. The authors believe that necks with combined characteristics should be used prudently, since they lead to significantly higher strain at the transtrochanteric area with irregular strain pattern. The greater degree of freedom that these necks provide cannot compensate for a bad surgical technique. The surgeon should also be cautious in the use of the short necks. Short necks exhibited lower stress at the femoral shaft and seemed to “relieve” the strain at the trans-trochanteric area. These changes in strain patterns raise concerns regarding adaptive hypertrophy and mechanical failure due to increased stress. This study has several limitations. First, the THA model we used did not emulate all surrounding soft tissues, due to the inherent difficulty of such a simulation in a laboratory setting. Nevertheless, the hip joint contact and abductor muscle forces have been demonstrated to have the greatest effect on strain patterns in the proximal aspect of the femur. Our model did not include the iliotibial tract which could reduce mediolateral bending of the femur (Stolk et al., 2001). Stolk et al. stated that additional inclusion of the iliotibial tract, the adductors and the vastii produced relatively small effects during all gait phases, with their most prominent effect being a slight reduction of bone strains at the level of the stem tip during heel-strike. Another limitation was that only axial compression tests were performed. A third limitation was that the implants were tested in the immediate post-operative condition, without any degree of bone ingrowth on the component that would allow for a greater stress transfer from the stem to the femoral cortex. In the real world, this ongoing bone remodeling would alter implant-femur biomechanics. A fourth limitation was that the location of the strain gages may not reflect the exact position of the maximum strain differentiations. Finally, the composite bones used may resemble but do not match the exact
anatomical geometry and local material property variations of the native femur. On the other hand, this study has some distinct strengths. A study showed that inter-specimen variability of physiological strain could be up to 62% (Davy et al., 1988). This is the reason that the data of every data set were obtained from the same composite bone each time and all neck variations were tested on the same specimen every time. Moreover, conditions of the experiment were kept constant at the beginning of each load cycle and tests were repeated on three different composite bones. Strain values measured were highly repeatable, indicating that there was little or no degradation of bone properties caused by repeated testing. To the best of the authors' knowledge, this is the only biomechanical study examining all the modular neck variations of the aforementioned commercially available prosthesis. 5. Conclusion This study has the advantage of measuring the strain in ideal conditions, changing one factor every time. This is often not the case in the clinical setting though. In the operating theater the choice of the femoral stem geometry is made regarding the preoperative templating, restoration of the leg length, joint stability and the quality of the bone. The surgeon should thus always keep in mind that every patient's case is unique and be very careful translating these experimental results. Nevertheless, one should anticipate that usage of the anteverted combinations could be more prone to anterior thigh pain or that the varus combinations could exacerbate calcar stress. The surgeon could also fine-tune parts of the surgical technique such as cup anteversion or height of the neck cut if that could lead to avoidance of using the short or combined variations, keeping in mind that there is a trade-off between capacity for correcting deformities and the resulting biomechanical stability of the bone. The effect of modular necks on different femoral implant designs, femoral models with torsional abnormalities and different implant sizes should be evaluated in further research and compared to data acquired from cadaveric femoral models. The authors believe that assessment of the biomechanical behavior could be a part of the preclinical evaluation of a new endoprosthesis.
422
A.N. Politis et al. / Clinical Biomechanics 28 (2013) 415–422
Conflict of interest The authors declare that they have no any financial or personal relationships with other people or organizations that could inappropriately bias the work published in this article. References Aamodt, A., Lund-Larsen, J., Eine, J., Andersen, E., Benum, P., Husby, O.S., 2001. Changes in proximal femoral strain after insertion of uncemented standard and customised femoral stems. An experimental study in human femora. J. Bone Joint Surg. Br. 83, 921–929. Benazzo, F., Rossi, S.M., Cecconi, D., Piovani, L., Ravasi, F., 2010. Mid-term results of an uncemented femoral stem with modular neck options. Hip. Int. 20, 427–433. Campbell, A.C., Rorabeck, C.H., Bourne, R.B., Chess, D., Nott, L., 1992. Thigh pain after cementless hip arthroplasty. Annoyance or ill omen. J. Bone Joint Surg. Br. 74, 63–66. Dastane, M., Dorr, L.D., Tarwala, R., Wan, Z., 2011. Hip offset in total hip arthroplasty: quantitative measurement with navigation. Clin. Orthop. Relat. Res. 469, 429–436. Davy, D.T., Kotzar, G.M., Brown, R.H., Heiple, K.G., Goldberg, V.M., Heiple, K.G., et al., 1988. Telemetric force measurements across the hip after total arthroplasty. J. Bone Joint. Surg. Am. 70, 45–50. De Haan, R., Pattyn, C., Gill, H.S., Murray, D.W., Campbell, P.A., De Smet, K., 2008. Correlation between inclination of the acetabular component and metal ion levels in metal-on-metal hip resurfacing replacement. J. Bone Joint Surg. Br. 90, 1291–1297. Dunbar, M.J., 2010. The proximal modular neck in THA: a bridge too far: affirms. Orthopedics 33, 640. Ellman, M.B., Levine, B.R., 2013. Fracture of the modular femoral neck component in total hip arthroplasty. J. Arthroplasty 196, e1–e5. Finlay, J.B., Rorabeck, C.H., Bourne, R.B., Tew, W.M., 1989. In vitro analysis of proximal femoral strains using PCA femoral implants and a hip-abductor muscle simulator. J. Arthroplasty 4, 335–345. Gibbons, C.E., Davies, A.J., Amis, A.A., Olearnik, H., Parker, B.C., Scott, J.E., 2001. Periprosthetic bone mineral density changes with femoral components of differing design philosophy. Int. Orthop. 25, 89–92. Helwig, P., Hindenlang, U., Hirschmüller, A., Konstantinidis, L., Südkamp, N., Schneider, R., 2011. A femoral model with all relevant muscles and hip capsule ligaments. Comput. Methods Biomech. Biomed. Engin. http://dx.doi.org/10.1080/ 10255842.2011.631918 (Dec 8 [Electronic publication ahead of print].
Jasty, M., O'Connor, D.O., Henshaw, R.M., Harrigan, T.P., Harris, W.H., 1994. Fit of the uncemented femoral component and the use of cement influence the strain transfer the femoral cortex. J. Orthop. Res. 12, 648–656. Kim, Y.H., Kim, J.S., Cho, S.H., 2001. Strain distribution in the proximal human femur. An in vitro comparison in the intact femur and after insertion of reference and experimental femoral stems. J. Bone Joint Surg. Br. 83, 295–301. Konyves, A., Bannister, G.C., 2005. The importance of leg length discrepancy after total hip arthroplasty. J. Bone Joint Surg. Br. 87, 155–157. Kop, A.M., Keogh, C., Swarts, E., 2012. Proximal component modularity in THA — at what cost? An implant retrieval study. Clin. Orthop. Relat. Res. 470, 1885–1894. McLeish, R.D., Charnley, J., 1970. Abduction forces in the one-legged stance. J. Biomech. 3, 191–209. Miki, H., Sugano, N., 2011. Modular neck for prevention of prosthetic impingement in cases with excessively anteverted femur. Clin. Biomech. 26, 944–949. Sakai, T., Ohzono, K., Nishii, T., Miki, H., Takao, M., Sugano, N., 2010. A modular femoral neck and head system works well in cementless total hip replacement for patients with developmental dysplasia of the hip. J. Bone Joint Surg. Br. 92, 770–776. Simpson, D.J., Little, J.P., Gray, H., Murray, D.W., Gill, H.S., 2009. Effect of modular neck variation on bone and cement mantle mechanics around a total hip arthroplasty stem. Clin. Biomech. 24, 274–285. Stolk, J., Verdonschot, N., Huiskes, R., 2001. Hip-joint and abductor-muscle forces adequately represent in vivo loading of a cemented total hip reconstruction. J. Biomech. 34, 917–936. Thien, T.M., Kärrholm, J., 2010. Design-related risk factors for revision of primary cemented stems. Acta Orthop. 81, 407–412. Traina, F., De Clerico, M., Biondi, F., Pilla, F., Tassinari, E., Toni, A., 2009. Sex differences in hip morphology: is stem modularity effective for total hip replacement? J. Bone Joint. Surg. Am. 91, 121–128. Umeda, N., Saito, M., Sugano, N., Ohzono, K., Nishii, T., Sakai, T., et al., 2003. Correlation between femoral neck version and strain on the femur after insertion of femoral prosthesis. J. Orthop. Sci. 8, 381–386. Wik, T.S., Enoksen, C., Klaksvik, J., Østbyhaug, P.O., Foss, O.A., Ludvigsen, J., et al., 2011. In vitro testing of the deformation pattern and initial stability of a cementless stem coupled to an experimental femoral head, with increased offset and altered femoral neck angles. Proc. Inst. Mech. Eng. H 225, 797–808. Wilson, D.A., Dunbar, M.J., Amirault, J.D., Farhat, Z., 2010. Early failure of a modular femoral neck total hip arthroplasty component: a case report. J. Bone Joint. Surg. Am. 92, 1514–1517.
Contents lists available at SciVerse ScienceDirect
Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech
Patterns of stress distribution at the proximal femur after implantation of a modular neck prosthesis. A biomechanical study Angelos N. Politis a, b,⁎, George K. Siogkas b, c, Ioannis D. Gelalis b, Theodore A. Xenakis b a b c
Department of Orthopaedic Surgery, Jewish General Hospital, McGill University, Montreal, Canada Laboratory of Biomechanics, University of Ioannina, School of Medicine, Ioannina, Greece Department of Electrical and Computer Engineering, University of Patras, Greece
a r t i c l e
i n f o
Article history: Received 30 July 2012 Accepted 5 February 2013 Keywords: Strain Modular Neck Femur Arthroplasty
a b s t r a c t Background: Modular total hip arthroplasty incorporating a double taper design is an evolution offering potential advantages compared to single head–neck taper or monolithic designs. Changes in femoral offset, neck length or femoral anteversion are expected to alter the strain distribution. Methods: We therefore analyzed the strain patterns after usage of all types of necks of a modular neck prosthesis, implanted in composite femurs. Findings: The load distribution presented a repeatable pattern. Anteverted neck combinations resulted in higher stress at the anterior surface, whereas the retroverted ones at the posterior (e.g. at the middle frontal site, stress is 13.63% higher when we shifted from the long neutral neck to the long 15° anteverted neck and at the middle back site 19.73% higher when we shifted from the long neutral to the long 15° retroverted neck). Compressive stress was larger at the calcar region and exacerbated by the use of the varus neck (e.g. at the frontal 1 site stress increased by 44.01% when we used the long 8° varus neck in comparison to the long neutral neck). Anteverted neck combinations resulted in higher strain at the anterior cortex around the tip of the prosthesis. Short necks exhibited lower stress at the femoral shaft and higher at the trans-trochanteric area. Interpretation: Anteverted neck combinations could be more prone to anterior thigh pain. Because of the possible risk of adaptive hypertrophy and early mechanical failure due to increased stress, the surgeon should be cautious when using necks with combined characteristics or short necks. Crown Copyright © 2013 Published by Elsevier Ltd. All rights reserved.
1. Introduction Modular total hip arthroplasty (THA) incorporating a double taper design is an innovation offering potential advantages compared to single head–neck taper or monolithic designs (Dunbar, 2010). Those include the adjustment of leg length and offset via the head–neck taper, femoral anteversion via the neck–stem taper, easier revision when there is no need to revise a well-fixed femoral stem and optimal restoration of soft tissue tension and patient biomechanics (Dunbar, 2010). The use of modular necks has thus increased in the recent years and authors reported good mid- and long-term clinical outcomes (Benazzo et al., 2010; Sakai et al., 2010). Adjusting femoral offset, leg length and orientation of the components are of crucial importance. Offset correlates to abductor muscle function, wear and impingement (Dastane et al., 2011). On the other hand, over-lengthening of the limb can be a problem (Konyves and Bannister, 2005), whereas failure to comply with the recommendations for acetabular inclination/anteversion and femoral anteversion may ⁎ Corresponding author at: Department of Orthopaedic Surgery, Jewish General Hospital, McGill University, 447, Ave Prince Albert, Apt. 4, Westmount, Montreal, Canada H3YP6. E-mail address: [email protected] (A.N. Politis).
lead to edge loading and prosthetic impingement, which can cause dislocation, mechanical loosening, wear or breakage of the polyethylene liner, metallosis or metal ion release in metal-on-metal bearings, and squeaking or breakage of ceramic-on-ceramic bearings (De Haan et al., 2008; Miki and Sugano, 2011). Are changes in femoral offset, neck length or femoral anteversion expected to alter the strain distribution at the femur? The aim of this biomechanical study is to analyze and compare the strain patterns at the proximal femur after usage of all types of necks of a commercially available modular neck stem prosthesis and possibly connect the results with the clinical praxis. 2. Methods In this study a set of cementless modular PROFEMUR-E® Total Hip Replacement System (Wright Medical Technology Inc., Arlington, TN, USA) was used. The aforementioned stem is a modular prosthesis manufactured from Ti6Al4V and has a 500 μm thick coating of pure titanium plasma spray. The surgeon may choose between six interchangeable necks available in two lengths and a total of 22 combinations can be used. We used custom-made, commercially available, medium left, fourth generation, medium composite femoral models
0268-0033/$ – see front matter. Crown Copyright © 2013 Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.clinbiomech.2013.02.004
416
A.N. Politis et al. / Clinical Biomechanics 28 (2013) 415–422
(model # 3403-99, Sawbones, Pacific Research Laboratories, Vashon Island, WA, USA). The bones were already osteotomized and specifically machined for use with a size 3 Profemur-E stem. A custom-made fixture designed to reproduce loading conditions during the single-leg stance phase of walking, as described by McLeish and Charnley (1970), was attached to the load cell of a computer-controlled hydraulic testing machine (MTS 858 Mini Bionix, MTS Systems Corp, Eden Prairie, MN, USA). The femur was tilted into 12° of valgus and was positioned neutral on the sagittal plane. Hip abductors were simulated by a small chain attached to a custom-made base that was fixed to the lateral aspect of the greater trochanter. The abductor force simulation applied the load at an angle of abduction 15° to the sagittal plane (Finlay et al., 1989; McLeish and Charnley, 1970). The distal end of the femur was embedded in a steel pot with radiopaque bone cement. A modified universal ball joint was mounted between the distal construct and the base of the machine (Fig. 1). The circumference of the femoral model was divided into 3 parts and strain gages were fixed along the lateral, medial-anterior and medial-posterior surface of the femur at positions 60° apart. Three 350-Ohm tri-axial rosette strain gages (KYOWA, KFG-2-350-D17-11, Kyowa Electronic Instrument, Tokyo, Japan) were bonded on the transtrochanteric surface, where a more complex strain pattern was expected. One rosette was made up of three strain gages mounted at 60° angles. The median of the peak-to-peak value of the sinusoidal strain over time was computed for each of the three gages. Uni-axial 350-Ohm strain gages (KYOWA, KFG-2-350-C1-11, Kyowa Electronic Instrument, Tokyo, Japan) were used along the shaft of the femur, where the strain pattern was expected to be simpler. The uni-axial strain gages were distributed at three horizontal levels at 48, 96, and 144 mm below the level of the lesser trochanter, so that the middle gages were around the tip of the stem. The leads of the gages were connected to a Wheatstone bridge configuration (Kyowa SS-24R Switching and Balancing Box, Kyowa Electronic Instrument, Tokyo, Japan). The gage outputs were transferred to a signal amplifier module and consequently to an (MTS TestStar II® data acquisition system, MTS Systems Corp, Eden Prairie, MN, USA). Load cycles were programmed to simulate single-leg stance of a normal-weight subject. Applying a vertical force five sixths of the body weight, with the weight of the lower extremity subtracted, would yield a physiological resultant hip joint force in the hip simulator (Wik et al., 2011). Thus, an axial load of 600 N was applied to simulate the single-leg stance of a subject weighing 70 kg. A 3-step
testing sequence was used as follows: 1. Ramp-up to − 300 N (rate 100 N/s). 2. Sinusoidal axial loading between − 100 and − 600 N applied at a frequency of 1 Hz for 200 cycles. 3. Ramp-down to − 50 N (rate 500 N/s). The material testing system operated under force control. Load cell limit was set at − 750 N. Strain values were recorded for 200 full cycles. Tests were repeated three times for every composite bone to obtain the average strain for each gage. The set of the three tests was repeated on three different composite femurs. The position of the strain gages was checked constant with the use of a phantom femoral model. All conditions were tested on every composite bone. Prior to testing, the abductor chain was pre-tensioned until the level arm was balanced at the horizontal plane in the beginning of every load cycle. Three specimens were tested and statistics were performed on the three sets of data obtained from the three specimens. Every set of data included all neck variations and was retrieved from the same specimen. The following neck variations were tested: • • • • •
•
• • • • •
•
Long neutral Long 8° anteverted (long 8 DG A)/long 8° retroverted (long 8 DG R) Long 8° varus (long 8 DG VAR)/long 8° valgus (long 8 DG VAL) Long 15° anteverted (long 15 DG A)/long 15° retroverted (long 15 DG R) Long varus valgus 1 anteverted (long VAR VAL 1 A = anteverted and valgus)/long varus valgus 1 retroverted (long VAR VAL 1 R = retroverted and varus) Long varus valgus 2 anteverted (long VAR VAL 2 A = anteverted and varus)/long varus valgus 2 retroverted (long VAR VAL 2 R = retroverted and valgus) Short neutral Short 8° anteverted (short 8 DG A)/short 8° retroverted (short 8 DG R) Short 8° varus (short 8 DG VAR)/short 8° valgus (short 8 DG VAL) Short 15° anteverted (short 15 DG A)/short 15° retroverted (short 15 DG R) Short varus valgus 1 anteverted (short VAR VAL 1 A = anteverted and valgus)/short varus valgus 1 retroverted (short VAR VAL 1 R = retroverted and varus) Short varus valgus 2 anteverted (short VAR VAL 2 A = anteverted and varus)/short varus valgus 2 retroverted (short VAR VAL 2 R = retroverted and valgus)
One-way analysis of variance (ANOVA) of the data was performed using MatLab (The MathWorks Inc, Natick, MA, USA). Post-hoc analysis was performed using the Tukey–Kramer test. A P value b 0.05 was considered statistically significant. 3. Results 3.1. Long neutral neck vs long 8° anteverted neck (long 8 DG A) vs long 8° retroverted neck (long 8 DG R) (Fig. 2, Table 1)
Fig. 1. Experimental set-up.
Strain measurement of the uni-axial strain gages revealed that longitudinal deformation was compressive on the medial side of the femur and tensile on the lateral. Strain analysis from the rosette gages on the anterior surface of the trans-trochanteric area revealed that gages 1 and 2 (the gage parallel to the longitudinal axis of the femur and the gage 60° deviated towards the medial side) showed compressive signal, while gage 3 (the gage 60° deviated towards the lateral side) showed tensile signal. Strain analysis from the rosette gages on the posterior surface of the trans-trochanteric area revealed that strain gage 1 (the gage 60° deviated towards the lateral side) showed tensile signal, while strain gages 2 and 3 (the gage parallel to the longitudinal axis of the femur and the gage 60° deviated towards the medial side) showed compressive signal. A consistent finding was that the long 8 DG A neck “conducted” stresses towards the anterior surface of the femur, while the long 8 DG R neck towards the posterior.
A.N. Politis et al. / Clinical Biomechanics 28 (2013) 415–422
417
Fig. 2. Long neutral vs long 8° anteverted vs long 8° retroverted. The aim of the comparison is to examine the effect of the change of the neck anteversion on the femoral strain.
This was valid for all anterior and posterior uni-axial strain gages along the shaft. The long neutral neck presented intermediate signal between the A and the R neck. For example, at the middle frontal strain gage site, stress was 13.63% higher when we shifted from the long neutral neck to the long 15 DG A neck and at the middle back strain gage site 19.73% higher when we shifted from the long neutral to the long 15 DG R neck. This principle was not valid for the trans-trochanteric region, where no correlation between the version of the prosthetic neck and transmission of stress was found. Stress was less on the anterior than the posterior surface of the trans-trochanteric region (e.g. comparison at the frontal 2 and back 2 strain gage sites revealed that using the long neutral neck resulted in 37.62% higher stress on the posterior surface in comparison to the anterior). Statistical analysis revealed that there were statistically significant differences between the three necks at all strain gage sites. 3.2. Long neutral neck vs long 15° anteverted neck (long 15 DG A) vs long 15° retroverted neck (long 15 DG R) The exact same conclusions were drawn in the aforementioned comparison. The pattern of strain distribution was the same and statistical analysis revealed that there were statistically significant differences between the three necks at all strain gage sites, but back 1 strain gage at the posterior trans-trochanteric area. 3.3. Long neutral neck vs long 8° anteverted neck (long 8 DG A) vs long 15° anteverted neck (long 15 DG A) (Fig. 3, Table 2) Our data revealed that stress on the anterior surface of the femur was greater using the long 15 DG A neck than using the long 8 DG A neck than using the long neutral neck. This was valid at the femoral
shaft. For example, at the bottom frontal strain gage site, stress was 16.97% and 38.5% higher when we shifted from the long neutral to the long 8 DG A and to the long 15 DG A neck respectively. Stress at the posterior surface had the opposite pattern. This was nevertheless again not valid for the trans-trochanteric region, where stress pattern was more irregular. There were statistically significant differences between the three necks at all strain gage sites. 3.4. Long neutral neck vs long 8° retroverted neck (long 8 DG R) vs long 15° retroverted neck (long 15 DG R) Similarly to the previous comparison, stress at the posterior surface of the model was greater using the long 15 DG R neck than using the long 8 DG R than using the long neutral. This was again valid at the shaft of the specimen. Stress at the anterior surface had the opposite pattern. The aforementioned pattern was once more not valid for the trans-trochanteric region. There were statistically significant differences between the three necks at all strain gage sites. 3.5. Long neutral neck vs long 8° varus neck (long 8 DG VAR) vs long 8° valgus neck (long 8 DG VAL) (Fig. 4, Table 3) Comparison of these three necks leads to the conclusion that the differences in the stress at the femoral shaft were smaller. For example, at the upper frontal strain gage site shifting from the long neutral neck to the long 8 DG VAR and to the long 8 DG VAL neck resulted in a 1.23% and 3.96% strain difference respectively. We observed that stress was larger at the posterior than the anterior surface of the model and this was also true for the trans-trochanteric area (e.g. using the long 8 DG VAR neck the stress at the upper back strain gage site was 107.5% higher compared to the stress at the upper frontal strain
Table 1 Indicative strains for the long neutral vs long 8° anteverted vs long 8° retroverted comparison (+ = compressive strain/− = tensile strain). The median of the peak-to-peak value of the sinusoidal strain over time εmax (με) is indicated. Neck/strain gage
Middle frontal
Middle back
Frontal 2
Back 2
1: Long neutral (SD) 2: Long 8 DG A (SD) 3: Long 8 DG R (SD) P Couples P b 0.05
+616 (±3.55) +700 (±8.12) +486 (±2.71) 0.000000 1–2, 1–3, 2–3
+608 (±3.17) +465 (±5.11) +728 (±2.90) 0.000000 1–2, 1–3, 2–3
+404 (±3.91) +329 (±3.04) +186 (±12.36) 0.000000 1–2, 1–3, 2–3
+556 (±3.02) +379 (±249) +572 (±21.08) 0.000000 1–2, 1–3, 2–3
418
A.N. Politis et al. / Clinical Biomechanics 28 (2013) 415–422
Fig. 3. Long neutral vs long 8° anteverted vs long 15° anteverted. The aim of the comparison is to examine the effect of the change of the amount of the neck anteversion on the femoral strain.
gage site). A safe conclusion was that compressive stress was larger at the calcar region and this was exacerbated by the use of the long 8 DG VAR neck (e.g. at the frontal 1 strain gage site stress increased by 44.01% when we used the long 8 DG VAR neck in comparison to the long neutral neck). This was also affirmed at the posterior surface of the trans-trochanteric region. 3.6. Long neutral neck vs long varus valgus 1 anteverted neck (long VAR VAL 1 A = anteverted and valgus) vs long varus valgus 1 retroverted neck (long VAR VAL 1 R = retroverted and varus) (Fig. 5, Table 4) Usage of necks with combined characteristics led to a synergistic stress pattern. The anteverted component of the neck led to larger strain at the anterior surface of the model, whereas the retroverted at the posterior. Similarly, the varus component led to higher strain at the calcar region. The use of necks with combined characteristics led to higher stress at the trans-trochanteric region (e.g. at the frontal 2 strain gage site stress increased by 48.51% and 50.74% when we shifted from the long neutral neck to the long VARVAL1A and to the long VARVAL1R neck respectively). This observation applied to all comparisons, concerning necks with combined characteristics: - Long neutral vs long VAR VAL 2 A (A and VAR) vs long VAR VAL 2 R (R and VAL) - Long VAR VAL 1 A (A and VAL) vs long 8 DG A vs long 8 DG VAL - Long VAR VAL 1 R (R and VAR) vs long 8 DG R vs long 8 DG VAR - Long VAR VAL 2 A (A and VAR) vs long 8 DG A vs long 8 DG VAR - Long VAR VAL 2 R (R and VAL) vs long 8 DG R vs long 8 DG VAL
3.7. Long vs short necks (Fig. 6) A general observation in all comparisons between long and short necks was that use of long necks resulted in higher stress at the shaft of the femur, whereas use of short necks led to higher stresses at the trans-trochanteric area. For example, at the middle frontal strain gage site stress increased by 6.2% when we shifted from the short neutral to the long neutral neck and at the Lateral 2 strain gage site the stress increased by 14.78% when we shifted from the long neutral to the short neutral neck. This finding applied to all comparisons between short and long necks. 4. Discussion Modularity of the femoral neck is increasing in popularity with implants available from several manufacturers. Modular stems allow increased flexibility by the surgeon, where a standard design with a fixed neck geometry will not restore the optimal hip biomechanics. Clinical advantages of these stems include the adjustment of leg length and offset via the head–neck taper, femoral anteversion via the neck–stem taper, easier revision when there is no need to revise a well-fixed femoral stem and optimal restoration of soft tissue tension and patient biomechanics. On the other hand, concerns over modularity have been reported, including the risk of mechanical failure of the trunion, disassociation of the components and femoral subsidence. Other possible complications include the generation of particle debris, metal ions, fretting and crevice corrosion. It has been reported that even with modern taper designs and corrosion-resistant materials,
Table 2 Indicative strains for the long neutral vs long 8° anteverted vs long 15° Αnteverted comparison (+ = compressive strain/− = tensile strain). The median of the peak-to-peak value of the sinusoidal strain over time εmax (με) is indicated. Neck/strain gage
Bottom frontal
Bottom back
Lateral 2
Frontal 2
1: Long neutral (SD) 2: Long 8 DG A (SD) 3: Long 15 DG Α (SD) P Couples P b 0.05
+483 (±3.21) +565 (±7.00) +669 (±3.58) 0.000001 1–2, 1–3, 2–3
+795 (±3.74) +625 (±7.07) +594 (±3.26) 0.000001 1–2, 1–3, 2–3
−1150 (±6.84) −949 (±6.33) −1170 (±4.44) 0.000000 1–2, 1–3, 2–3
+404 (±3.91) +329 (±3.04) +641 (±3.16) 0.000000 1–2, 1–3, 2–3
A.N. Politis et al. / Clinical Biomechanics 28 (2013) 415–422
419
Fig. 4. Long neutral vs long 8° varus vs long 8° valgus. The aim of the comparison is to examine the effect of the change of the caput-collum-diaphyseal angle on the femoral strain.
corrosion, fretting and particulate debris were observed to a greater extent in the second neck–stem junction and that the degradation of the junction can contribute to metallosis and generation of the aseptic lymphocyte-laminated vascular-associated lesions that can lead to revision (Kop et al., 2012). Even if a change in the stress pattern at the femur is expected, it is not known whether the different neck orientation results in a significant change in bone loading. Measurement of bone surface strains is an indirect way to elucidate the implications that exist in the literature about extreme biomechanical behavior after usage of certain types of modular necks, such as the long varus or retroverted necks (Ellman and Levine, 2013; Wilson et al., 2010). It is reported that the long VAR neck component increases the bending moment by 32.7%, when compared to the short VAR neck and that failures of modular necks are attributed to neck offset (neck length+ head length) (Traina et al., 2009). A recent study from the Swedish Arthroplasty Registry has reported an increased revision rate for small-sized stems with increased neck length (Thien and Kärrholm, 2010). According to our measurements, the aforesaid necks (long 15 DG R and long 8 DG VAR) did not exhibit extreme loading behavior, with the maximum microstrain value recorded, being εmax = − 1440με. Another clinical issue is the possible link of the use of the different neck configurations of this specific prosthesis with the development of stress shielding. Preservation of the proximal bone stock is a fundamental goal after THA (Kim et al., 2001). The largest loss in bone mineral density is reported to occur in the proximomedial aspect of the femur, while evidence exists that a significant amount of bone loss may also occur in the proximolateral (Gibbons et al., 2001). The area below the calcar is also particularly vulnerable in strain decrease
(Aamodt et al., 2001; Kim et al., 2001). Clinically, the consequences of stress shielding are not entirely clear, although it is often cited as a reason for concern causing osteolysis, and thus aceptic loosening in the long term. The exact level of strain required to alter or preserve bone remodeling is not known. We observed that stress was larger at the posterior than the anterior surface of the model and this was also true for the trans-trochanteric area. This makes this area particularly vulnerable to stress shielding. According to our data, the lateral or posterior transtrochanteric area are in lower risk for stress shielding, as they did not exhibit very low strains in any neck combination. Studies conducted to understand the response of human femurs under loading conditions, reported that loading led to a bending load in the proximal two-thirds of the femur, whereas almost homogeneous strain conditions prevail in the lower third. It is cited that longitudinal strain decreases from proximal to distal on the surface of the native femur in vitro (Helwig et al., 2011). On the contrary, Kim et al. (2001) found that strain increased from proximal to distal in the intact femur, while Jasty et al. (1994) reported that strain was relatively uniform along the diaphysis of the femur. In our study, strains were compressive at the medial side and tensile at the lateral side and this finding was consistent for all neck comparisons. This was observed for both the transtrochanteric area and the shaft of the femur. Our data suggest that the load distribution pattern does not increase nor decrease in a uniform manner along the shaft, but presents a certain pattern that was validated throughout our study: at the anterior surface of the femur the load distribution is the following: strain at the middle horizontal level (around the tip of the prosthesis) greater than strain at the lower horizontal level (the mid-shaft of the femur) greater than strain at the upper horizontal level (around the middle of the prosthesis). The posterior surface presents the opposite strain
Table 3 Indicative strains for the long neutral vs long 8° varus vs long 8° valgus comparison (+ = compressive strain/− = tensile strain). The median of the peak-to-peak value of the sinusoidal strain over time εmax (με) is indicated. Neck/strain gage
Frontal 1
Back 2
Upper back
Upper frontal
1: Long neutral (SD) 2: Long 8 DG VAR (SD) 3: Long 8 DG VΑL (SD) P Couples P b 0.05
+443 (±3.68) +638 (±3.12) +453 (±2.62) 0.000000 1–2, 1–3, 2–3
+557 (±3.02) +756 (±3.24) +623 (±3.18) 0.000000 1–2, 1–3, 2–3
+940 (±4.10) +828 (±3.77) +968 (±4.98) 0.000003 1–2, 1–3, 2–3
+404 (±2.49) +399 (±1.77) +388 (±3.02) 0.060422 –
420
A.N. Politis et al. / Clinical Biomechanics 28 (2013) 415–422
Fig. 5. Long neutral vs long varus valgus anteverted 1 (anteverted and valgus) vs long varus valgus retroverted 1 (retroverted and varus). The aim of the comparison is to examine the effect of the use of the necks with combined characteristics on the femoral strain.
pattern: strain at the upper horizontal level greater than strain at the lower horizontal level greater than strain at the middle level. We deduce that the area around the tip of the prosthesis at the anterior cortex that presents the higher strain values is the area of load transfer from the prosthesis to the bone, and is subject to a higher risk of provoking anterior thigh pain. This is in agreement with the fact that thigh pain is perceived at the anterior side of the femur in most patients. On the contrary the pattern of stress distribution at the transtrochanteric area was more complex throughout the experiment, and no safe conclusions can be drawn. We attribute this to the fact that three rosette strain gages provide tri-axial strain measurement in only three points of this, anatomically complex, part of the femoral bone and cannot depict the true strain distribution in the area. Furthermore, the transtrochanteric area is subject to strain exerted from the hip abductors to the greater trochanter. To the best of our knowledge, the biomechanical behavior of femoral stems with modular necks was examined in two previous studies. Simpson et al. (2009) performed a finite element analysis on the strain pattern after use of modular necks in anteversion, retroversion and of two different offsets. They reported that strain on the medial bone surface increased distally along the femur. Overall, lower strains were observed on the medial bone when an anteverted neck was used, as compared to a retroverted neck, and the authors concluded that the different neck geometries did not result in a significant change in bone mechanics. Umeda et al. (2003) performed strain gage measurements on composite femurs implanted with a cementless stem and combined with retroverted and anteverted femoral necks. The authors reported that on the medial aspect compressive strains were much larger than tensile strains, while on the lateral aspect tensile strains were much larger than compressive strains regardless of neck version. On the
aspect toward which the prosthetic neck was oriented compressive strains tended to be larger than tensile, and on the opposite side tensile strains tended to be larger than compressive. This was most marked around the level of the stem tip. Our data agree with Umeda et al. The anteverted necks in our study conducted stress towards the anterior surface, while the retroverted ones towards the posterior. This was valid for all uni-axial strain gages along the shaft of the femur. Neutral necks presented an intermediate signal between the anteverted and the retroverted ones. This principle was once more not valid for the trans-trochanteric region and this is, to the authors' judgment, due to the complex anatomy of the area. Our data also verified the assumption that a 15 DG ante/retroverted neck swifts more stress towards the direction of neck orientation, than an 8 DG neck. According to our data, the VAR/VAL neck component had a smaller effect on the stress at the femoral shaft than the anteverted/retroverted neck variation and a constant finding was that compressive stress was larger at the calcar region and calcar strain was exacerbated by the use of the 8 DG VAR neck. This is in accordance with the universal agreement that a large percentage of strain after THA is transmitted through the crucial calcar region. Significant thigh pain after implantation of a cementless device occurs in as many as 22% of patients and is an important concern. Focal mismatch in flexural rigidity around the stem tip has been implicated as an important factor contributing to mid-thigh pain, but sources of pain may also include implant micromotion and instability (Campbell et al., 1992). There are no published data about the strain threshold that can possibly elicit thigh pain. According to the authors' opinion, high strains may lead to thigh pain via “channeling” high values of stress through the anterior femoral cortex, whereas low strains may be indicative of a possible loosening and instability of
Table 4 Indicative strains for the long neutral vs long varus valgus anteverted 1 (anteverted and valgus) vs long varus valgus retroverted 1 (retroverted and varus) comparison (+ = compressive strain/− = tensile strain). The median of the peak-to-peak value of the sinusoidal strain over time εmax (με) is indicated. Neck/strain gage
Middle frontal
Middle back
Frontal 1
Frontal 2
1: Long neutral (SD) 2: Long VARVAL 1 A (SD) 3: Long VARVAL 1 R (SD) P Couples P b 0.05
+616 (±3.55) +633 (±3.22) +565 (±48.70) 0.000023 1–2, 1–3, 2–3
+605 (±3.17) +527 (±3.29) +686 (±50.28) 0.000000 1–2, 1–3, 2–3
+442 (±3.68) +550 (±9.88) +576 (±5.41) 0.000000 1–2, 1–3, 2–3
+404 (±3.91) +597 (±9.99) +609 (±6.08) 0.000000 1–2, 1–3
A.N. Politis et al. / Clinical Biomechanics 28 (2013) 415–422
421
Fig. 6. Long neutral vs short neutral. The aim of the comparison is to examine the effect of the change of the neck length on the femoral strain.
the prosthesis, leading again to thigh pain. Analysis of our data showed that anteverted neck combinations result in higher strain at the anterior cortex around the tip of the prosthesis and could thus be more prone to anterior thigh pain. This was more intense with the use of the 15 DG anteverted necks. For example, at the middle frontal strain gage site which is around the tip of the prosthesis, strain increased by 13.63% and 34.25% when we shifted from the long neutral neck to the long 8 DG A and to the long 15 DG A respectively. Another assumption confirmed was that usage of combined necks will lead to a more complex stress pattern. The authors believe that necks with combined characteristics should be used prudently, since they lead to significantly higher strain at the transtrochanteric area with irregular strain pattern. The greater degree of freedom that these necks provide cannot compensate for a bad surgical technique. The surgeon should also be cautious in the use of the short necks. Short necks exhibited lower stress at the femoral shaft and seemed to “relieve” the strain at the trans-trochanteric area. These changes in strain patterns raise concerns regarding adaptive hypertrophy and mechanical failure due to increased stress. This study has several limitations. First, the THA model we used did not emulate all surrounding soft tissues, due to the inherent difficulty of such a simulation in a laboratory setting. Nevertheless, the hip joint contact and abductor muscle forces have been demonstrated to have the greatest effect on strain patterns in the proximal aspect of the femur. Our model did not include the iliotibial tract which could reduce mediolateral bending of the femur (Stolk et al., 2001). Stolk et al. stated that additional inclusion of the iliotibial tract, the adductors and the vastii produced relatively small effects during all gait phases, with their most prominent effect being a slight reduction of bone strains at the level of the stem tip during heel-strike. Another limitation was that only axial compression tests were performed. A third limitation was that the implants were tested in the immediate post-operative condition, without any degree of bone ingrowth on the component that would allow for a greater stress transfer from the stem to the femoral cortex. In the real world, this ongoing bone remodeling would alter implant-femur biomechanics. A fourth limitation was that the location of the strain gages may not reflect the exact position of the maximum strain differentiations. Finally, the composite bones used may resemble but do not match the exact
anatomical geometry and local material property variations of the native femur. On the other hand, this study has some distinct strengths. A study showed that inter-specimen variability of physiological strain could be up to 62% (Davy et al., 1988). This is the reason that the data of every data set were obtained from the same composite bone each time and all neck variations were tested on the same specimen every time. Moreover, conditions of the experiment were kept constant at the beginning of each load cycle and tests were repeated on three different composite bones. Strain values measured were highly repeatable, indicating that there was little or no degradation of bone properties caused by repeated testing. To the best of the authors' knowledge, this is the only biomechanical study examining all the modular neck variations of the aforementioned commercially available prosthesis. 5. Conclusion This study has the advantage of measuring the strain in ideal conditions, changing one factor every time. This is often not the case in the clinical setting though. In the operating theater the choice of the femoral stem geometry is made regarding the preoperative templating, restoration of the leg length, joint stability and the quality of the bone. The surgeon should thus always keep in mind that every patient's case is unique and be very careful translating these experimental results. Nevertheless, one should anticipate that usage of the anteverted combinations could be more prone to anterior thigh pain or that the varus combinations could exacerbate calcar stress. The surgeon could also fine-tune parts of the surgical technique such as cup anteversion or height of the neck cut if that could lead to avoidance of using the short or combined variations, keeping in mind that there is a trade-off between capacity for correcting deformities and the resulting biomechanical stability of the bone. The effect of modular necks on different femoral implant designs, femoral models with torsional abnormalities and different implant sizes should be evaluated in further research and compared to data acquired from cadaveric femoral models. The authors believe that assessment of the biomechanical behavior could be a part of the preclinical evaluation of a new endoprosthesis.
422
A.N. Politis et al. / Clinical Biomechanics 28 (2013) 415–422
Conflict of interest The authors declare that they have no any financial or personal relationships with other people or organizations that could inappropriately bias the work published in this article. References Aamodt, A., Lund-Larsen, J., Eine, J., Andersen, E., Benum, P., Husby, O.S., 2001. Changes in proximal femoral strain after insertion of uncemented standard and customised femoral stems. An experimental study in human femora. J. Bone Joint Surg. Br. 83, 921–929. Benazzo, F., Rossi, S.M., Cecconi, D., Piovani, L., Ravasi, F., 2010. Mid-term results of an uncemented femoral stem with modular neck options. Hip. Int. 20, 427–433. Campbell, A.C., Rorabeck, C.H., Bourne, R.B., Chess, D., Nott, L., 1992. Thigh pain after cementless hip arthroplasty. Annoyance or ill omen. J. Bone Joint Surg. Br. 74, 63–66. Dastane, M., Dorr, L.D., Tarwala, R., Wan, Z., 2011. Hip offset in total hip arthroplasty: quantitative measurement with navigation. Clin. Orthop. Relat. Res. 469, 429–436. Davy, D.T., Kotzar, G.M., Brown, R.H., Heiple, K.G., Goldberg, V.M., Heiple, K.G., et al., 1988. Telemetric force measurements across the hip after total arthroplasty. J. Bone Joint. Surg. Am. 70, 45–50. De Haan, R., Pattyn, C., Gill, H.S., Murray, D.W., Campbell, P.A., De Smet, K., 2008. Correlation between inclination of the acetabular component and metal ion levels in metal-on-metal hip resurfacing replacement. J. Bone Joint Surg. Br. 90, 1291–1297. Dunbar, M.J., 2010. The proximal modular neck in THA: a bridge too far: affirms. Orthopedics 33, 640. Ellman, M.B., Levine, B.R., 2013. Fracture of the modular femoral neck component in total hip arthroplasty. J. Arthroplasty 196, e1–e5. Finlay, J.B., Rorabeck, C.H., Bourne, R.B., Tew, W.M., 1989. In vitro analysis of proximal femoral strains using PCA femoral implants and a hip-abductor muscle simulator. J. Arthroplasty 4, 335–345. Gibbons, C.E., Davies, A.J., Amis, A.A., Olearnik, H., Parker, B.C., Scott, J.E., 2001. Periprosthetic bone mineral density changes with femoral components of differing design philosophy. Int. Orthop. 25, 89–92. Helwig, P., Hindenlang, U., Hirschmüller, A., Konstantinidis, L., Südkamp, N., Schneider, R., 2011. A femoral model with all relevant muscles and hip capsule ligaments. Comput. Methods Biomech. Biomed. Engin. http://dx.doi.org/10.1080/ 10255842.2011.631918 (Dec 8 [Electronic publication ahead of print].
Jasty, M., O'Connor, D.O., Henshaw, R.M., Harrigan, T.P., Harris, W.H., 1994. Fit of the uncemented femoral component and the use of cement influence the strain transfer the femoral cortex. J. Orthop. Res. 12, 648–656. Kim, Y.H., Kim, J.S., Cho, S.H., 2001. Strain distribution in the proximal human femur. An in vitro comparison in the intact femur and after insertion of reference and experimental femoral stems. J. Bone Joint Surg. Br. 83, 295–301. Konyves, A., Bannister, G.C., 2005. The importance of leg length discrepancy after total hip arthroplasty. J. Bone Joint Surg. Br. 87, 155–157. Kop, A.M., Keogh, C., Swarts, E., 2012. Proximal component modularity in THA — at what cost? An implant retrieval study. Clin. Orthop. Relat. Res. 470, 1885–1894. McLeish, R.D., Charnley, J., 1970. Abduction forces in the one-legged stance. J. Biomech. 3, 191–209. Miki, H., Sugano, N., 2011. Modular neck for prevention of prosthetic impingement in cases with excessively anteverted femur. Clin. Biomech. 26, 944–949. Sakai, T., Ohzono, K., Nishii, T., Miki, H., Takao, M., Sugano, N., 2010. A modular femoral neck and head system works well in cementless total hip replacement for patients with developmental dysplasia of the hip. J. Bone Joint Surg. Br. 92, 770–776. Simpson, D.J., Little, J.P., Gray, H., Murray, D.W., Gill, H.S., 2009. Effect of modular neck variation on bone and cement mantle mechanics around a total hip arthroplasty stem. Clin. Biomech. 24, 274–285. Stolk, J., Verdonschot, N., Huiskes, R., 2001. Hip-joint and abductor-muscle forces adequately represent in vivo loading of a cemented total hip reconstruction. J. Biomech. 34, 917–936. Thien, T.M., Kärrholm, J., 2010. Design-related risk factors for revision of primary cemented stems. Acta Orthop. 81, 407–412. Traina, F., De Clerico, M., Biondi, F., Pilla, F., Tassinari, E., Toni, A., 2009. Sex differences in hip morphology: is stem modularity effective for total hip replacement? J. Bone Joint. Surg. Am. 91, 121–128. Umeda, N., Saito, M., Sugano, N., Ohzono, K., Nishii, T., Sakai, T., et al., 2003. Correlation between femoral neck version and strain on the femur after insertion of femoral prosthesis. J. Orthop. Sci. 8, 381–386. Wik, T.S., Enoksen, C., Klaksvik, J., Østbyhaug, P.O., Foss, O.A., Ludvigsen, J., et al., 2011. In vitro testing of the deformation pattern and initial stability of a cementless stem coupled to an experimental femoral head, with increased offset and altered femoral neck angles. Proc. Inst. Mech. Eng. H 225, 797–808. Wilson, D.A., Dunbar, M.J., Amirault, J.D., Farhat, Z., 2010. Early failure of a modular femoral neck total hip arthroplasty component: a case report. J. Bone Joint. Surg. Am. 92, 1514–1517.