- Email: [email protected]
In vitro initial stability of a stemless humeral implant
JCLB-04095; No of Pages 5 Clinical Biomechanics xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech
In vitro initial stability of a stemless humeral implant Philippe Favre ⁎, Jörn Seebeck, Paul A.E. Thistlethwaite, Marc Obrist, Jason G. Steffens, Andrew R. Hopkins, Paul A. Hulme Zimmer Biomet, Sulzerallee 8, CH-8404 Winterthur, Switzerland
a r t i c l e
i n f o
Article history: Received 9 July 2015 Accepted 1 December 2015 Keywords: Stemless Humerus Primary stability Micromotion Image analysis
a b s t r a c t Background: Stemless humeral prostheses have been recently introduced. We measured for the first time their in vitro primary stability and analyzed the influence of three clinically important parameters (bone quality, implant size and post-operative loading) on micromotion. We also assessed if displacement sensors are appropriate to measure implant micromotion. Methods: A stemless humeral implant (Sidus® Stem-Free Shoulder, Zimmer GmbH, Winterthur, Switzerland) was implanted in 18 cadaveric humeri. Three-dimensional motion of the implant was measured under dynamic loading at three load magnitudes with displacement sensors. Additionally, the relative motion at the bone–implant interface was measured with an optical system in four specimens. Results: Micromotion values derived from the displacement sensors were significantly higher than those measured by the optical system (P b 0.005). Analysis of variance (ANOVA) indicated that bone density (P b 0.0005) and load (P b 0.0001) had a significant effect on implant micromotion, however the effect of implant size was not statistically significant (P = 0.123). Interpretation: Micromotion of this stemless design was shown to be significantly dependent on cancellous bone density. Patients must therefore have adequate bone quality for this procedure. The influence of load magnitude on micromotion emphasizes the need for controlled post-operative rehabilitation. Measurements with displacement sensors overestimate true interface micromotion by up to 50% and correction by an optical system is strongly recommended. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Recently introduced stemless humeral prostheses facilitate an anatomic reconstruction of the humeral head, minimize intraoperative humeral fractures and preserve bone stock (Berth and Pap, 2013). While most complications in total shoulder arthroplasty involve the glenoid component, loosening of humeral stems has also been reported (Cil et al., 2009; Matsen et al., 2003; Roper et al., 1990; Sanchez-Sotelo et al., 2001; Torchia et al., 1997). Stemless humeral implants rely on a surface coating and press-fit with the cancellous bone for their stability and may have a different response to loading than stems that have cancellous and cortical bone contact. Initial clinical studies suggest that stemless shoulder implants achieve good fixation (Berth and Pap, 2013; Huguet et al., 2010; Kadum et al., 2011), but knowledge on how some of the key clinical parameters affect initial stability is missing.
The first aim of this study was to determine the influence of bone quality, implant size and post-operative loading on micromotion. In vitro measurement of reverse glenoid baseplates displacement can include elastic system deformation resulting in a large overestimation of interface micromotion (Favre et al., 2011; Hopkins et al., 2008). The second objective of the study was to assess if the use of the established displacement sensor method (Harman et al., 2005; Harris et al., 2000; Kwon et al., 2010; Peppers et al., 1998; Poon et al., 2010; Virani et al., 2008) is nevertheless appropriate for micromotion testing of a stemless humeral implant where comparatively more interface micromotion than elastic system deformation might occur, or if recent image-based solutions (Codsi and Iannotti, 2008; Favre et al., 2011) would yield more accurate results. 2. Methods 2.1. Specimens
⁎ Corresponding author. E-mail addresses: [email protected] (P. Favre), [email protected] (J. Seebeck), [email protected] (P.A.E. Thistlethwaite), [email protected] (M. Obrist), [email protected] (J.G. Steffens), [email protected] (A.R. Hopkins), [email protected] (P.A. Hulme).
Eighteen cadaveric stripped humeri (five bilateral female, three bilateral male shoulders and two unilateral male shoulders, average age 60 SD 10 years) were used. The specimens were cut to keep approximately 10 cm of the proximal humerus. Bone density was assessed from CT scans of the proximal humerus. Three cylindrical areas were
http://dx.doi.org/10.1016/j.clinbiomech.2015.12.004 0268-0033/© 2015 Elsevier Ltd. All rights reserved.
Please cite this article as: Favre, P., et al., In vitro initial stability of a stemless humeral implant, Clin. Biomech. (2015), http://dx.doi.org/10.1016/ j.clinbiomech.2015.12.004
2
P. Favre et al. / Clinical Biomechanics xxx (2015) xxx–xxx
analyzed corresponding to the cancellous bone volumes surrounding the three different implant sizes (small, medium and large). Hounsfield values calibrated with respect to a density phantom (Gammex 467, Gammex Inc, Middleton, WI, USA) were converted into apparent density using a linear relationship (Esses et al., 1989). 2.2. Implantation Sidus® Stem-Free Shoulder implants (Zimmer GmbH, Winterthur, Switzerland) were tested in this study. These implants are indicated for cementless use in hemi or total shoulder arthroplasty. They are approved and were launched in Europe in 2012. They are not approved for sale and distribution in the US market but are the subject of an ongoing clinical study to support a premarket approval. The system comprises a rough blasted titanium alloy anchor and polished cobalt chromium humeral head (Fig. 1). The anchors were implanted into the humeri according to the surgical technique. A small size anchor was implanted in bones for which a medium size anchor may have been more suitable. This was done to ensure we had an appropriate number of small size implants for statistical analysis. The humerus was cemented (Osteobond® Copolymer Bone Cement, Zimmer, Warsaw, IN, USA) in the specimen holder while ensuring that the anchor baseplate was aligned with the specimen holder. The humeral head with a fixed measurement stage was impacted onto the taper of the implant. 2.3. Primary stability measurements Four differential variable reluctance transducer (DVRT) displacement sensors (SG-DVRT-8, 2 μm resolution, Lord Microstrain, Cary, NC, USA) were used to measure the motion of the implant (data acquisition system: Spider 8-30 TF 600 Hz HBM; software: Catman, HBM, Darmstadt, Germany). The sensors were positioned to measure the inferior–
superior tilt around the anterior–posterior axis and the inferior–superior, medial–lateral and anterior–posterior displacements during loading (Fig. 2). Attaching the sensors to the specimen holder ensured that the orientation of the sensors was kept the same for all specimens. From the DVRT output, the implant motion was calculated relative to a central point on the implant using Matlab (MathWorks Inc., Natick, MA, USA) and took into account the influence of implant tilting. To determine maximum implant motion, the resultant motion of six points on the implant periphery (Fig. 1) was calculated from translations and rotations of the central point using rigid body transformations. Similarly, for the direct comparison of the DVRT and camera based systems, implant micromotion was calculated for the same region that was imaged using the camera system. Fourteen specimens were analysed using only the DVRT system and four specimens with varying bone quality were analysed using both the DVRT and camera systems. A factor that corrected for the elastic motion of the cancellous bone included in the DVRT measurements was derived by comparing interface micromotion using a more accurate, previously published optical method (Fig. 3) (Favre et al., 2011). An approximately 1 × 1 cm cutout was made on the side of the bone to gain visible access to the cancellous bone–implant interface. Images of the interface were taken with a highresolution camera system (Prosilica GX1920, AVT, Stadtroda, Germany) equipped with a telecentric lens (S5LPJ4425, Sill Optics GmbH & Co, Wendelstein, Germany). The imaging axis of the lens was kept perpendicular to the analyzed plane. Pre-load (50 N) and full load images were compared using a Matlab script that called the image analysis software Fiji (Schindelin et al., 2012) as a subroutine (Favre et al., 2011). This script aligned the two images to remove any rotation and translation of the system or the camera, identified landmarks common to both images and evaluated the relative implant-bone motion during loading (see Fig. 3 bottom image). Micromotion for all landmarks within the region of interest was averaged.
Fig. 1. Sidus® Stem-Free Shoulder (left) with locations analyzed for micromotion (the point marked with an arrow indicates the location of peak micromotion), implanted Sidus® Anchor (upper right), cemented construct (lower right).
Please cite this article as: Favre, P., et al., In vitro initial stability of a stemless humeral implant, Clin. Biomech. (2015), http://dx.doi.org/10.1016/ j.clinbiomech.2015.12.004
P. Favre et al. / Clinical Biomechanics xxx (2015) xxx–xxx
3
Fig. 2. DVRT setup to measure primary stability of the stemless shoulder implant. (1) Load application device connected to the testing machine, (2) custom humeral head (anchor and humeral bone are below), (3) measurement stage, (4a) 2× DVRT oriented in medial–lateral axis, (4b) DVRT oriented in anterior–posterior axis, and (4c) DVRT oriented in superior–inferior axis.
2.4. Loading Three load steps were applied reaching 220 N, 520 N and 820 N. The lower load step (220 N) was applied to precondition the implant. During the first 2 months of physiotherapy following shoulder arthroplasty, forces measured by a telemetric shoulder implant reached 40% body weight (BW) (392 N) during training against resistance but only 20% BW (196 N) for other movements (Bergmann et al., 2007). These loads were measured for one patient only so the middle load step (520 N) could represent a conservative estimate of post-operative shoulder loading for a broader range of patients. The upper load (820 N) was set to represent the peak loading during “normal” use with no weight in the hand (Bergmann et al., 2011; Westerhoff et al., 2009). The loads were applied in the coronal plane at a 30° angle from the implant central axis. This angle corresponds to the ranges measured in vivo (Bergmann et al., 2011; Westerhoff et al., 2009). Loads were applied to the humeral head cyclically for 100 cycles per load step in force control at 300 N/s using a single axis material testing machine (Zwick 1456, Zwick Roell, Ulm, Germany). Micromotion was compared at 1, 25, 50 and 100 cycles and was expected to stabilize quickly with little variation between the first few cycles and after 100 cycles (Favre et al., 2011; Kwon et al., 2010; Virani et al., 2008). 2.5. Statistical analyses All statistical analyses were performed in Matlab using a three-way analysis of variance (ANOVA) to test the influence of implant size (n = 7 for large, n = 5 for medium and n = 6 for small), joint loading (n = 18 for each load step) and bone quality (n = 18). A paired t test was used to test the difference between the DVRT and the camera measurements (n = 12). 3. Results The DVRT micromotion measurements were significantly higher than the camera measurements (P b 0.005), but the trends with respect to the loads were similar (Table 1). The ratio between the camera based interface micromotion and DVRT measured implant motion for 520 and
Fig. 3. Top: high-resolution camera with telecentric lens and light source to image the bone–implant interface under load. Bottom: bone–implant interface. The rectangular blue window shows the resulting image analysis on a portion of the image only. The arrows represent the 2D relative displacement vectors of the identified landmarks. The white horizontal bar represents 1 mm, the arrows length is magnified by a factor five for visualization purpose.
820 N was 0.5 to 0.75. The 0.75 factor was multiplied to all DVRT results to yield the maximum DVRT micromotion values. The micromotion showed clear evidence of stabilization after 25 cycles. The maximum micromotion was consistently found at the anchor point that is most lateral and distal (see Fig. 1, point marked with an arrow). Bone density (P b 0.0005) and load (P b 0.0001) had a significant effect on implant micromotion (Table 2 and Fig. 4); however, the effect of implant size was not statistically significant (P = 0.123).
Table 1 Comparison of micromotion data obtained using a camera system and DVRT (before correction with the 0.75 factor). All values are given for the location that was imaged using the camera system. Specimen
S110935
S110851
S110823
S110870
Implant size
S
M
L
L
Bone density (g/cm3)
0.12
0.15
0.24
0.41
Micromotion (μm) Camera DVRT Camera DVRT Camera DVRT Camera DVRT 220 N 37 Load 520 N 72 820 N 112
54 115 189
45 106 194
52 142 271
18 49 69
24 70 113
6 12 21
9 24 42
Please cite this article as: Favre, P., et al., In vitro initial stability of a stemless humeral implant, Clin. Biomech. (2015), http://dx.doi.org/10.1016/ j.clinbiomech.2015.12.004
4
P. Favre et al. / Clinical Biomechanics xxx (2015) xxx–xxx Table 2 Influence of load on micromotion. Data displayed as mean micromotion (1 standard deviation). Load [N]
Micromotion [μm]
220 520 820
44 (SD 21) 79 (SD 32) 125 (SD 64)
4. Discussion This is the first study investigating the in vitro micromotion behavior of a stemless humeral implant. The influence of clinically relevant parameters (bone quality, implant size and post-operative loading) has been assessed, and the feasibility of such measurements with displacement sensors has been gauged. The results of the current investigation are within the 26–162 μm range of reported in vitro reverse glenoid motion (Bicknell et al., 2003; Codsi and Iannotti, 2008; Favre et al., 2011; Harman et al., 2005; Kwon et al., 2010; Poon, 2010; Virani et al., 2008), although direct comparisons are affected by differences in loading, measurement and fixation conditions. There exist very little literature on micromotion testing of stemmed humeral implants (Harris et al., 2000; Peppers et al., 1998), and only one study evaluating cementless fixation (Peppers et al., 1998). Bone density did not have any significant effect on humeral stem axial micromotion. Humeral stems are designed to have a distal interference fit with the cortical bone and are therefore less sensitive to the proximal cancellous bone quality. In contrast, a stemless shoulder anchor is completely and solely reliant on the cancellous bone for support so that the surrounding cancellous bone density strongly influences anchor micromotion (Fig. 4). Further studies are required for a side-by-side comparison of the primary stability behavior of stemless, short stems and stemmed humeral prostheses. Maximum micromotion was found in low bone density specimens when the highest loads were applied. The trabecular structure experiences more elastic deformation and destruction with higher loads. Similarly, in lower bone quality specimens, fewer or thinner trabeculae more easily deform and fracture, provide less contact area at the interface with the anchor and are therefore less able to resist implant motion. The combination of low bone quality and high joint loads should therefore be avoided. As recommended in the Sidus® surgical technique, patients eligible for a stemless implant should have good bone quality and loads should be only gradually increased following surgery to allow for sufficient osseointegration. A distinct bone density threshold for which stemless implants would be appropriate cannot
Fig. 4. Influence of bone density on peak implant micromotion (values corrected with factor). Data are displayed according to implant size (different marker shapes) and applied load (520 N for empty marker shapes and 820 N for filled marker shapes).
be determined at this point. Micromotion limits for bone ingrowth are imprecise (Pilliar et al., 1986). Also a much larger number of specimens should be tested in order to define an exact level with a high enough confidence for the full range of bone material, implant sizes and loads. The effect of anchor size was not statistically significant for implant motion although a trend towards larger micromotions for smaller anchors was observed. The second objective of this study was to assess if the use of DVRT type sensors was appropriate for the assessment of micromotion of a cementless implant relying on cancellous press-fit alone for fixation. While both measurement systems were able to similarly capture the influence of load, anchor size and bone quality on micromotion (Table 1), the DVRT system overestimated micromotion. Instead of measuring the motion of the implant relative to the bone, DVRT measurements assess implant motion under the assumption that no bone deformation occurs. The effect of elastic deformation could not be entirely eliminated even though special attention was given to enhance DVRT measurements by using four sensors to track 3D implant motion and remove the effect of sensor location by calculating motion on implant surface points. In the current study, the system deformation still accounted for up to 50% of the DVRT measurement. On the other hand, the camera method measures both the implant and the bone motion in the same image and allows the direct calculation of the true relative motion by subtraction of the bone motion from the implant motion. The overestimation of micromotion through DVRT measurements reached two orders of magnitude difference using a similar interface image-based analysis for reverse glenoid baseplates (Favre et al., 2011). Stemless humeral implants may provide a less robust fixation than a baseplate fixed with four screws, allowing more relative movement to the bone in comparison to elastic system deformation. Displacement sensors are therefore appropriate for a relative comparison of the influence on micromotion of different parameters (design, loading, surgical technique, etc.) within one investigation. Nevertheless, a comparison between different studies would inevitably be biased given that every test setup will deform in its own manner and include a different proportion of elastic deformation, preventing a proper one-to-one comparison. To allow comparison of absolute values between studies, more accurate systems such as the optical method presented in this paper should be used either as the primary micromotion measurement technique or to correct DVRT values. Limitations of this study include that a small anchor was sometimes implanted in bones where a medium size may have been more suitable. This was necessary to test the influence of anchor size within the range of specimens available. Anchor undersizing carried out in this study can be considered acceptable because anchor size was shown to not significantly influence micromotion. Only four specimens were analyzed with both the sensor and optical methods because the optical system was not yet available at the start of the study. Using both measurement systems on the full set of bones would have increased the statistical validity of the comparison of the two measurement methods. This would have also avoided the use of a single factor to derive the micromotion from the DVRT measurements for all specimens. However, the worst-case factor was used to generate a conservative estimate of interface micromotion. This factor was defined for the setup, loading and implant specific to this study and should not be generalized to other tests. Experimental micromotion assessment does not provide the full micromotion distribution on the whole bone–implant interface. Techniques such as finite element analysis would be beneficial to determine this. However, the majority of the implant’s surface should have micromotion values less than the reported values because care was taken to include calculation points furthest away from the expected center of rotation. Postoperative in vivo interface micromotion is expected to be lower than the values measured during this study because loads in the ranges that were measured by Bergmann et al. (2007) during post-operative rehabilitation were lower than the loads we applied. Loading for in vitro primary stability assessment should represent the loading during the months following surgery when the process of osseointegration
Please cite this article as: Favre, P., et al., In vitro initial stability of a stemless humeral implant, Clin. Biomech. (2015), http://dx.doi.org/10.1016/ j.clinbiomech.2015.12.004
P. Favre et al. / Clinical Biomechanics xxx (2015) xxx–xxx
occurs. In the current study, we also applied higher loads (820 N) representative of normal daily use in order to challenge the fixation. The effect of anchor size on implant motion showed a trend towards larger micromotions for smaller anchors but was not statistically significant. The small sample size and the large variability may have had an influence on the conclusions drawn regarding implant size. Finally, one single implant design was tested and conclusions are only valid for this specific anchor design and amount of press-fit. 5. Conclusions This study characterized the micromotion of the Sidus® Stem-Free Shoulder implant. Caution must be used when treating patients with lower bone density because implant micromotion was shown to be significantly influenced by cancellous bone quality. In the case of any doubt about bone quality affecting the stable fixation of the anchor, a stemmed shoulder prosthesis must be preferred. Loading significantly influenced implant micromotion of the implant, supporting the need for controlled reintroduction of function to the joint through appropriate postoperative management. The effect of anchor size was not statistically significant for implant micromotion. DVRT sensors overestimated interface micromotion and are appropriate for a relative comparison of different parameters within an investigation. Optical techniques, such as the one presented in this paper, are not influenced by elastic system deformation and thus measurements may be more representative of the true interface micromotion. Acknowledgments The study was financed entirely by Zimmer GmbH. References Bergmann, G., Graichen, F., Bender, A., Kaab, M., Rohlmann, A., Westerhoff, P., 2007. In vivo glenohumeral contact forces—measurements in the first patient 7 months postoperatively. J. Biomech. 40, 2139–2149. Bergmann, G., Graichen, F., Bender, A., Rohlmann, A., Halder, A., Beier, A., Westerhoff, P., 2011. In vivo gleno-humeral joint loads during forward flexion and abduction. J. Biomech. 44, 1543–1552. Berth, A., Pap, G., 2013. Stemless shoulder prosthesis versus conventional anatomic shoulder prosthesis in patients with osteoarthritis: a comparison of the functional outcome after a minimum of two years follow-up. J. Orthop. Traumatol. 14, 31–37. Bicknell, R.T., Liew, A.S., Danter, M.R., Patterson, S.D., King, G.J., Chess, D.G., Johnson, J.A., 2003. Does keel size, the use of screws, and the use of bone cement affect fixation of a metal glenoid implant? J. Shoulder Elbow Surg. 12, 268–275. Cil, A., Veillette, C.J., Sanchez-Sotelo, J., Sperling, J.W., Schleck, C., Cofield, R.H., 2009. Revision of the humeral component for aseptic loosening in arthroplasty of the shoulder. J. Bone Joint Surg. (Br.) 91, 75–81.
5
Codsi, M.J., Iannotti, J.P., 2008. The effect of screw position on the initial fixation of a reverse total shoulder prosthesis in a glenoid with a cavitary bone defect. J. Shoulder Elbow Surg. 17, 479–486. Esses, S.I., Lotz, J.C., Hayes, W.C., 1989. Biomechanical properties of the proximal femur determined in vitro by single-energy quantitative computed tomography. J. Bone Miner. Res. 4, 715–722. Favre, P., Perala, S., Vogel, P., Fucentese, S.F., Goff, J.R., Gerber, C., Snedeker, J.G., 2011. In vitro assessments of reverse glenoid stability using displacement gages are misleading—recommendations for accurate measurements of interface micromotion. Clin. Biomech. 26, 917–922. Harman, M., Frankle, M., Vasey, M., Banks, S., 2005. Initial glenoid component fixation in “reverse” total shoulder arthroplasty: a biomechanical evaluation. J. Shoulder Elbow Surg. 14, 162s–167s. Harris, T., Jobe, C.M., Dai, Q.G., 2000. Fixation of proximal humeral prostheses and rotational micromotion. J. Shoulder Elbow Surg. 9, 205–210. Hopkins, A.R., Hansen, U.N., Bull, A.M., Emery, R., Amis, A.A., 2008. Fixation of the reversed shoulder prosthesis. J. Shoulder Elbow Surg. 17, 974–980. Huguet, D., DeClercq, G., Rio, B., Teissier, J., Zipoli, B., 2010. Results of a new stemless shoulder prosthesis: radiologic proof of maintained fixation and stability after a minimum of three years' follow-up. J. Shoulder Elbow Surg. 19, 847–852. Kadum, B., Mafi, N., Norberg, S., Sayed-Noor, A.S., 2011. Results of the total evolutive shoulder system (TESS): a single-centre study of 56 consecutive patients. Arch. Orthop. Trauma Surg. 131, 1623–1629. Kwon, Y.W., Forman, R.E., Walker, P.S., Zuckerman, J.D., 2010. Analysis of reverse total shoulder joint forces and glenoid fixation. Bull. NYU Hosp. Jt. Dis. 68, 273–280. Matsen III, F.A., Iannotti, J.P., Rockwood Jr., C.A., 2003. Humeral fixation by press-fitting of a tapered metaphyseal stem: a prospective radiographic study. J. Bone Joint Surg. Am. 85, 304–308. Peppers, T.A., Jobe, C.M., Dai, Q.G., Williams, P.A., Libanati, C., 1998. Fixation of humeral prostheses and axial micromotion. J. Shoulder Elbow Surg. 7, 414–418. Pilliar, R.M., Lee, J.M., Maniatopoulos, C., 1986. Observations on the effect of movement on bone ingrowth into porous-surfaced implants. Clin. Orthop. Relat. Res. 208, 108–113. Poon, P., 2010. Biomechanical evaluation of different designs of glenospheres in the SMR reverse shoulder prosthesis: micromotion of the baseplate and risk of loosening. 2 pp. 94–99. Poon, P., Chou, J., Young, D., Malak, S., Anderson, I.A., 2010. Biomechanical evaluation of different designs of glenospheres in the SMR reverse shoulder prosthesis: micromotion of the baseplate and risk of loosening. Should. Elb. 2, 94–99. Roper, B.A., Paterson, J.M., Day, W.H., 1990. The Roper–Day total shoulder replacement. J. Bone Joint Surg. (Br.) 72, 694–697. Sanchez-Sotelo, J., O'Driscoll, S.W., Torchia, M.E., Cofield, R.H., Rowland, C.M., 2001. Radiographic assessment of cemented humeral components in shoulder arthroplasty. J. Shoulder Elbow Surg. 10, 526–531. Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., Preibisch, S., Rueden, C., Saalfeld, S., Schmid, B., Tinevez, J.Y., White, D.J., Hartenstein, V., Eliceiri, K., Tomancak, P., Cardona, A., 2012. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–682. Torchia, M.E., Cofield, R.H., Settergren, C.R., 1997. Total shoulder arthroplasty with the Neer prosthesis: long-term results. J. Shoulder Elbow Surg. 6, 495–505. Virani, N.A., Harman, M., Li, K., Levy, J., Pupello, D.R., Frankle, M.A., 2008. In vitro and finite element analysis of glenoid bone/baseplate interaction in the reverse shoulder design. J. Shoulder Elbow Surg. 17, 509–521. Westerhoff, P., Graichen, F., Bender, A., Halder, A., Beier, A., Rohlmann, A., Bergmann, G., 2009. In vivo measurement of shoulder joint loads during activities of daily living. J. Biomech. 42, 1840–1849.
Please cite this article as: Favre, P., et al., In vitro initial stability of a stemless humeral implant, Clin. Biomech. (2015), http://dx.doi.org/10.1016/ j.clinbiomech.2015.12.004
Contents lists available at ScienceDirect
Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech
In vitro initial stability of a stemless humeral implant Philippe Favre ⁎, Jörn Seebeck, Paul A.E. Thistlethwaite, Marc Obrist, Jason G. Steffens, Andrew R. Hopkins, Paul A. Hulme Zimmer Biomet, Sulzerallee 8, CH-8404 Winterthur, Switzerland
a r t i c l e
i n f o
Article history: Received 9 July 2015 Accepted 1 December 2015 Keywords: Stemless Humerus Primary stability Micromotion Image analysis
a b s t r a c t Background: Stemless humeral prostheses have been recently introduced. We measured for the first time their in vitro primary stability and analyzed the influence of three clinically important parameters (bone quality, implant size and post-operative loading) on micromotion. We also assessed if displacement sensors are appropriate to measure implant micromotion. Methods: A stemless humeral implant (Sidus® Stem-Free Shoulder, Zimmer GmbH, Winterthur, Switzerland) was implanted in 18 cadaveric humeri. Three-dimensional motion of the implant was measured under dynamic loading at three load magnitudes with displacement sensors. Additionally, the relative motion at the bone–implant interface was measured with an optical system in four specimens. Results: Micromotion values derived from the displacement sensors were significantly higher than those measured by the optical system (P b 0.005). Analysis of variance (ANOVA) indicated that bone density (P b 0.0005) and load (P b 0.0001) had a significant effect on implant micromotion, however the effect of implant size was not statistically significant (P = 0.123). Interpretation: Micromotion of this stemless design was shown to be significantly dependent on cancellous bone density. Patients must therefore have adequate bone quality for this procedure. The influence of load magnitude on micromotion emphasizes the need for controlled post-operative rehabilitation. Measurements with displacement sensors overestimate true interface micromotion by up to 50% and correction by an optical system is strongly recommended. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Recently introduced stemless humeral prostheses facilitate an anatomic reconstruction of the humeral head, minimize intraoperative humeral fractures and preserve bone stock (Berth and Pap, 2013). While most complications in total shoulder arthroplasty involve the glenoid component, loosening of humeral stems has also been reported (Cil et al., 2009; Matsen et al., 2003; Roper et al., 1990; Sanchez-Sotelo et al., 2001; Torchia et al., 1997). Stemless humeral implants rely on a surface coating and press-fit with the cancellous bone for their stability and may have a different response to loading than stems that have cancellous and cortical bone contact. Initial clinical studies suggest that stemless shoulder implants achieve good fixation (Berth and Pap, 2013; Huguet et al., 2010; Kadum et al., 2011), but knowledge on how some of the key clinical parameters affect initial stability is missing.
The first aim of this study was to determine the influence of bone quality, implant size and post-operative loading on micromotion. In vitro measurement of reverse glenoid baseplates displacement can include elastic system deformation resulting in a large overestimation of interface micromotion (Favre et al., 2011; Hopkins et al., 2008). The second objective of the study was to assess if the use of the established displacement sensor method (Harman et al., 2005; Harris et al., 2000; Kwon et al., 2010; Peppers et al., 1998; Poon et al., 2010; Virani et al., 2008) is nevertheless appropriate for micromotion testing of a stemless humeral implant where comparatively more interface micromotion than elastic system deformation might occur, or if recent image-based solutions (Codsi and Iannotti, 2008; Favre et al., 2011) would yield more accurate results. 2. Methods 2.1. Specimens
⁎ Corresponding author. E-mail addresses: [email protected] (P. Favre), [email protected] (J. Seebeck), [email protected] (P.A.E. Thistlethwaite), [email protected] (M. Obrist), [email protected] (J.G. Steffens), [email protected] (A.R. Hopkins), [email protected] (P.A. Hulme).
Eighteen cadaveric stripped humeri (five bilateral female, three bilateral male shoulders and two unilateral male shoulders, average age 60 SD 10 years) were used. The specimens were cut to keep approximately 10 cm of the proximal humerus. Bone density was assessed from CT scans of the proximal humerus. Three cylindrical areas were
http://dx.doi.org/10.1016/j.clinbiomech.2015.12.004 0268-0033/© 2015 Elsevier Ltd. All rights reserved.
Please cite this article as: Favre, P., et al., In vitro initial stability of a stemless humeral implant, Clin. Biomech. (2015), http://dx.doi.org/10.1016/ j.clinbiomech.2015.12.004
2
P. Favre et al. / Clinical Biomechanics xxx (2015) xxx–xxx
analyzed corresponding to the cancellous bone volumes surrounding the three different implant sizes (small, medium and large). Hounsfield values calibrated with respect to a density phantom (Gammex 467, Gammex Inc, Middleton, WI, USA) were converted into apparent density using a linear relationship (Esses et al., 1989). 2.2. Implantation Sidus® Stem-Free Shoulder implants (Zimmer GmbH, Winterthur, Switzerland) were tested in this study. These implants are indicated for cementless use in hemi or total shoulder arthroplasty. They are approved and were launched in Europe in 2012. They are not approved for sale and distribution in the US market but are the subject of an ongoing clinical study to support a premarket approval. The system comprises a rough blasted titanium alloy anchor and polished cobalt chromium humeral head (Fig. 1). The anchors were implanted into the humeri according to the surgical technique. A small size anchor was implanted in bones for which a medium size anchor may have been more suitable. This was done to ensure we had an appropriate number of small size implants for statistical analysis. The humerus was cemented (Osteobond® Copolymer Bone Cement, Zimmer, Warsaw, IN, USA) in the specimen holder while ensuring that the anchor baseplate was aligned with the specimen holder. The humeral head with a fixed measurement stage was impacted onto the taper of the implant. 2.3. Primary stability measurements Four differential variable reluctance transducer (DVRT) displacement sensors (SG-DVRT-8, 2 μm resolution, Lord Microstrain, Cary, NC, USA) were used to measure the motion of the implant (data acquisition system: Spider 8-30 TF 600 Hz HBM; software: Catman, HBM, Darmstadt, Germany). The sensors were positioned to measure the inferior–
superior tilt around the anterior–posterior axis and the inferior–superior, medial–lateral and anterior–posterior displacements during loading (Fig. 2). Attaching the sensors to the specimen holder ensured that the orientation of the sensors was kept the same for all specimens. From the DVRT output, the implant motion was calculated relative to a central point on the implant using Matlab (MathWorks Inc., Natick, MA, USA) and took into account the influence of implant tilting. To determine maximum implant motion, the resultant motion of six points on the implant periphery (Fig. 1) was calculated from translations and rotations of the central point using rigid body transformations. Similarly, for the direct comparison of the DVRT and camera based systems, implant micromotion was calculated for the same region that was imaged using the camera system. Fourteen specimens were analysed using only the DVRT system and four specimens with varying bone quality were analysed using both the DVRT and camera systems. A factor that corrected for the elastic motion of the cancellous bone included in the DVRT measurements was derived by comparing interface micromotion using a more accurate, previously published optical method (Fig. 3) (Favre et al., 2011). An approximately 1 × 1 cm cutout was made on the side of the bone to gain visible access to the cancellous bone–implant interface. Images of the interface were taken with a highresolution camera system (Prosilica GX1920, AVT, Stadtroda, Germany) equipped with a telecentric lens (S5LPJ4425, Sill Optics GmbH & Co, Wendelstein, Germany). The imaging axis of the lens was kept perpendicular to the analyzed plane. Pre-load (50 N) and full load images were compared using a Matlab script that called the image analysis software Fiji (Schindelin et al., 2012) as a subroutine (Favre et al., 2011). This script aligned the two images to remove any rotation and translation of the system or the camera, identified landmarks common to both images and evaluated the relative implant-bone motion during loading (see Fig. 3 bottom image). Micromotion for all landmarks within the region of interest was averaged.
Fig. 1. Sidus® Stem-Free Shoulder (left) with locations analyzed for micromotion (the point marked with an arrow indicates the location of peak micromotion), implanted Sidus® Anchor (upper right), cemented construct (lower right).
Please cite this article as: Favre, P., et al., In vitro initial stability of a stemless humeral implant, Clin. Biomech. (2015), http://dx.doi.org/10.1016/ j.clinbiomech.2015.12.004
P. Favre et al. / Clinical Biomechanics xxx (2015) xxx–xxx
3
Fig. 2. DVRT setup to measure primary stability of the stemless shoulder implant. (1) Load application device connected to the testing machine, (2) custom humeral head (anchor and humeral bone are below), (3) measurement stage, (4a) 2× DVRT oriented in medial–lateral axis, (4b) DVRT oriented in anterior–posterior axis, and (4c) DVRT oriented in superior–inferior axis.
2.4. Loading Three load steps were applied reaching 220 N, 520 N and 820 N. The lower load step (220 N) was applied to precondition the implant. During the first 2 months of physiotherapy following shoulder arthroplasty, forces measured by a telemetric shoulder implant reached 40% body weight (BW) (392 N) during training against resistance but only 20% BW (196 N) for other movements (Bergmann et al., 2007). These loads were measured for one patient only so the middle load step (520 N) could represent a conservative estimate of post-operative shoulder loading for a broader range of patients. The upper load (820 N) was set to represent the peak loading during “normal” use with no weight in the hand (Bergmann et al., 2011; Westerhoff et al., 2009). The loads were applied in the coronal plane at a 30° angle from the implant central axis. This angle corresponds to the ranges measured in vivo (Bergmann et al., 2011; Westerhoff et al., 2009). Loads were applied to the humeral head cyclically for 100 cycles per load step in force control at 300 N/s using a single axis material testing machine (Zwick 1456, Zwick Roell, Ulm, Germany). Micromotion was compared at 1, 25, 50 and 100 cycles and was expected to stabilize quickly with little variation between the first few cycles and after 100 cycles (Favre et al., 2011; Kwon et al., 2010; Virani et al., 2008). 2.5. Statistical analyses All statistical analyses were performed in Matlab using a three-way analysis of variance (ANOVA) to test the influence of implant size (n = 7 for large, n = 5 for medium and n = 6 for small), joint loading (n = 18 for each load step) and bone quality (n = 18). A paired t test was used to test the difference between the DVRT and the camera measurements (n = 12). 3. Results The DVRT micromotion measurements were significantly higher than the camera measurements (P b 0.005), but the trends with respect to the loads were similar (Table 1). The ratio between the camera based interface micromotion and DVRT measured implant motion for 520 and
Fig. 3. Top: high-resolution camera with telecentric lens and light source to image the bone–implant interface under load. Bottom: bone–implant interface. The rectangular blue window shows the resulting image analysis on a portion of the image only. The arrows represent the 2D relative displacement vectors of the identified landmarks. The white horizontal bar represents 1 mm, the arrows length is magnified by a factor five for visualization purpose.
820 N was 0.5 to 0.75. The 0.75 factor was multiplied to all DVRT results to yield the maximum DVRT micromotion values. The micromotion showed clear evidence of stabilization after 25 cycles. The maximum micromotion was consistently found at the anchor point that is most lateral and distal (see Fig. 1, point marked with an arrow). Bone density (P b 0.0005) and load (P b 0.0001) had a significant effect on implant micromotion (Table 2 and Fig. 4); however, the effect of implant size was not statistically significant (P = 0.123).
Table 1 Comparison of micromotion data obtained using a camera system and DVRT (before correction with the 0.75 factor). All values are given for the location that was imaged using the camera system. Specimen
S110935
S110851
S110823
S110870
Implant size
S
M
L
L
Bone density (g/cm3)
0.12
0.15
0.24
0.41
Micromotion (μm) Camera DVRT Camera DVRT Camera DVRT Camera DVRT 220 N 37 Load 520 N 72 820 N 112
54 115 189
45 106 194
52 142 271
18 49 69
24 70 113
6 12 21
9 24 42
Please cite this article as: Favre, P., et al., In vitro initial stability of a stemless humeral implant, Clin. Biomech. (2015), http://dx.doi.org/10.1016/ j.clinbiomech.2015.12.004
4
P. Favre et al. / Clinical Biomechanics xxx (2015) xxx–xxx Table 2 Influence of load on micromotion. Data displayed as mean micromotion (1 standard deviation). Load [N]
Micromotion [μm]
220 520 820
44 (SD 21) 79 (SD 32) 125 (SD 64)
4. Discussion This is the first study investigating the in vitro micromotion behavior of a stemless humeral implant. The influence of clinically relevant parameters (bone quality, implant size and post-operative loading) has been assessed, and the feasibility of such measurements with displacement sensors has been gauged. The results of the current investigation are within the 26–162 μm range of reported in vitro reverse glenoid motion (Bicknell et al., 2003; Codsi and Iannotti, 2008; Favre et al., 2011; Harman et al., 2005; Kwon et al., 2010; Poon, 2010; Virani et al., 2008), although direct comparisons are affected by differences in loading, measurement and fixation conditions. There exist very little literature on micromotion testing of stemmed humeral implants (Harris et al., 2000; Peppers et al., 1998), and only one study evaluating cementless fixation (Peppers et al., 1998). Bone density did not have any significant effect on humeral stem axial micromotion. Humeral stems are designed to have a distal interference fit with the cortical bone and are therefore less sensitive to the proximal cancellous bone quality. In contrast, a stemless shoulder anchor is completely and solely reliant on the cancellous bone for support so that the surrounding cancellous bone density strongly influences anchor micromotion (Fig. 4). Further studies are required for a side-by-side comparison of the primary stability behavior of stemless, short stems and stemmed humeral prostheses. Maximum micromotion was found in low bone density specimens when the highest loads were applied. The trabecular structure experiences more elastic deformation and destruction with higher loads. Similarly, in lower bone quality specimens, fewer or thinner trabeculae more easily deform and fracture, provide less contact area at the interface with the anchor and are therefore less able to resist implant motion. The combination of low bone quality and high joint loads should therefore be avoided. As recommended in the Sidus® surgical technique, patients eligible for a stemless implant should have good bone quality and loads should be only gradually increased following surgery to allow for sufficient osseointegration. A distinct bone density threshold for which stemless implants would be appropriate cannot
Fig. 4. Influence of bone density on peak implant micromotion (values corrected with factor). Data are displayed according to implant size (different marker shapes) and applied load (520 N for empty marker shapes and 820 N for filled marker shapes).
be determined at this point. Micromotion limits for bone ingrowth are imprecise (Pilliar et al., 1986). Also a much larger number of specimens should be tested in order to define an exact level with a high enough confidence for the full range of bone material, implant sizes and loads. The effect of anchor size was not statistically significant for implant motion although a trend towards larger micromotions for smaller anchors was observed. The second objective of this study was to assess if the use of DVRT type sensors was appropriate for the assessment of micromotion of a cementless implant relying on cancellous press-fit alone for fixation. While both measurement systems were able to similarly capture the influence of load, anchor size and bone quality on micromotion (Table 1), the DVRT system overestimated micromotion. Instead of measuring the motion of the implant relative to the bone, DVRT measurements assess implant motion under the assumption that no bone deformation occurs. The effect of elastic deformation could not be entirely eliminated even though special attention was given to enhance DVRT measurements by using four sensors to track 3D implant motion and remove the effect of sensor location by calculating motion on implant surface points. In the current study, the system deformation still accounted for up to 50% of the DVRT measurement. On the other hand, the camera method measures both the implant and the bone motion in the same image and allows the direct calculation of the true relative motion by subtraction of the bone motion from the implant motion. The overestimation of micromotion through DVRT measurements reached two orders of magnitude difference using a similar interface image-based analysis for reverse glenoid baseplates (Favre et al., 2011). Stemless humeral implants may provide a less robust fixation than a baseplate fixed with four screws, allowing more relative movement to the bone in comparison to elastic system deformation. Displacement sensors are therefore appropriate for a relative comparison of the influence on micromotion of different parameters (design, loading, surgical technique, etc.) within one investigation. Nevertheless, a comparison between different studies would inevitably be biased given that every test setup will deform in its own manner and include a different proportion of elastic deformation, preventing a proper one-to-one comparison. To allow comparison of absolute values between studies, more accurate systems such as the optical method presented in this paper should be used either as the primary micromotion measurement technique or to correct DVRT values. Limitations of this study include that a small anchor was sometimes implanted in bones where a medium size may have been more suitable. This was necessary to test the influence of anchor size within the range of specimens available. Anchor undersizing carried out in this study can be considered acceptable because anchor size was shown to not significantly influence micromotion. Only four specimens were analyzed with both the sensor and optical methods because the optical system was not yet available at the start of the study. Using both measurement systems on the full set of bones would have increased the statistical validity of the comparison of the two measurement methods. This would have also avoided the use of a single factor to derive the micromotion from the DVRT measurements for all specimens. However, the worst-case factor was used to generate a conservative estimate of interface micromotion. This factor was defined for the setup, loading and implant specific to this study and should not be generalized to other tests. Experimental micromotion assessment does not provide the full micromotion distribution on the whole bone–implant interface. Techniques such as finite element analysis would be beneficial to determine this. However, the majority of the implant’s surface should have micromotion values less than the reported values because care was taken to include calculation points furthest away from the expected center of rotation. Postoperative in vivo interface micromotion is expected to be lower than the values measured during this study because loads in the ranges that were measured by Bergmann et al. (2007) during post-operative rehabilitation were lower than the loads we applied. Loading for in vitro primary stability assessment should represent the loading during the months following surgery when the process of osseointegration
Please cite this article as: Favre, P., et al., In vitro initial stability of a stemless humeral implant, Clin. Biomech. (2015), http://dx.doi.org/10.1016/ j.clinbiomech.2015.12.004
P. Favre et al. / Clinical Biomechanics xxx (2015) xxx–xxx
occurs. In the current study, we also applied higher loads (820 N) representative of normal daily use in order to challenge the fixation. The effect of anchor size on implant motion showed a trend towards larger micromotions for smaller anchors but was not statistically significant. The small sample size and the large variability may have had an influence on the conclusions drawn regarding implant size. Finally, one single implant design was tested and conclusions are only valid for this specific anchor design and amount of press-fit. 5. Conclusions This study characterized the micromotion of the Sidus® Stem-Free Shoulder implant. Caution must be used when treating patients with lower bone density because implant micromotion was shown to be significantly influenced by cancellous bone quality. In the case of any doubt about bone quality affecting the stable fixation of the anchor, a stemmed shoulder prosthesis must be preferred. Loading significantly influenced implant micromotion of the implant, supporting the need for controlled reintroduction of function to the joint through appropriate postoperative management. The effect of anchor size was not statistically significant for implant micromotion. DVRT sensors overestimated interface micromotion and are appropriate for a relative comparison of different parameters within an investigation. Optical techniques, such as the one presented in this paper, are not influenced by elastic system deformation and thus measurements may be more representative of the true interface micromotion. Acknowledgments The study was financed entirely by Zimmer GmbH. References Bergmann, G., Graichen, F., Bender, A., Kaab, M., Rohlmann, A., Westerhoff, P., 2007. In vivo glenohumeral contact forces—measurements in the first patient 7 months postoperatively. J. Biomech. 40, 2139–2149. Bergmann, G., Graichen, F., Bender, A., Rohlmann, A., Halder, A., Beier, A., Westerhoff, P., 2011. In vivo gleno-humeral joint loads during forward flexion and abduction. J. Biomech. 44, 1543–1552. Berth, A., Pap, G., 2013. Stemless shoulder prosthesis versus conventional anatomic shoulder prosthesis in patients with osteoarthritis: a comparison of the functional outcome after a minimum of two years follow-up. J. Orthop. Traumatol. 14, 31–37. Bicknell, R.T., Liew, A.S., Danter, M.R., Patterson, S.D., King, G.J., Chess, D.G., Johnson, J.A., 2003. Does keel size, the use of screws, and the use of bone cement affect fixation of a metal glenoid implant? J. Shoulder Elbow Surg. 12, 268–275. Cil, A., Veillette, C.J., Sanchez-Sotelo, J., Sperling, J.W., Schleck, C., Cofield, R.H., 2009. Revision of the humeral component for aseptic loosening in arthroplasty of the shoulder. J. Bone Joint Surg. (Br.) 91, 75–81.
5
Codsi, M.J., Iannotti, J.P., 2008. The effect of screw position on the initial fixation of a reverse total shoulder prosthesis in a glenoid with a cavitary bone defect. J. Shoulder Elbow Surg. 17, 479–486. Esses, S.I., Lotz, J.C., Hayes, W.C., 1989. Biomechanical properties of the proximal femur determined in vitro by single-energy quantitative computed tomography. J. Bone Miner. Res. 4, 715–722. Favre, P., Perala, S., Vogel, P., Fucentese, S.F., Goff, J.R., Gerber, C., Snedeker, J.G., 2011. In vitro assessments of reverse glenoid stability using displacement gages are misleading—recommendations for accurate measurements of interface micromotion. Clin. Biomech. 26, 917–922. Harman, M., Frankle, M., Vasey, M., Banks, S., 2005. Initial glenoid component fixation in “reverse” total shoulder arthroplasty: a biomechanical evaluation. J. Shoulder Elbow Surg. 14, 162s–167s. Harris, T., Jobe, C.M., Dai, Q.G., 2000. Fixation of proximal humeral prostheses and rotational micromotion. J. Shoulder Elbow Surg. 9, 205–210. Hopkins, A.R., Hansen, U.N., Bull, A.M., Emery, R., Amis, A.A., 2008. Fixation of the reversed shoulder prosthesis. J. Shoulder Elbow Surg. 17, 974–980. Huguet, D., DeClercq, G., Rio, B., Teissier, J., Zipoli, B., 2010. Results of a new stemless shoulder prosthesis: radiologic proof of maintained fixation and stability after a minimum of three years' follow-up. J. Shoulder Elbow Surg. 19, 847–852. Kadum, B., Mafi, N., Norberg, S., Sayed-Noor, A.S., 2011. Results of the total evolutive shoulder system (TESS): a single-centre study of 56 consecutive patients. Arch. Orthop. Trauma Surg. 131, 1623–1629. Kwon, Y.W., Forman, R.E., Walker, P.S., Zuckerman, J.D., 2010. Analysis of reverse total shoulder joint forces and glenoid fixation. Bull. NYU Hosp. Jt. Dis. 68, 273–280. Matsen III, F.A., Iannotti, J.P., Rockwood Jr., C.A., 2003. Humeral fixation by press-fitting of a tapered metaphyseal stem: a prospective radiographic study. J. Bone Joint Surg. Am. 85, 304–308. Peppers, T.A., Jobe, C.M., Dai, Q.G., Williams, P.A., Libanati, C., 1998. Fixation of humeral prostheses and axial micromotion. J. Shoulder Elbow Surg. 7, 414–418. Pilliar, R.M., Lee, J.M., Maniatopoulos, C., 1986. Observations on the effect of movement on bone ingrowth into porous-surfaced implants. Clin. Orthop. Relat. Res. 208, 108–113. Poon, P., 2010. Biomechanical evaluation of different designs of glenospheres in the SMR reverse shoulder prosthesis: micromotion of the baseplate and risk of loosening. 2 pp. 94–99. Poon, P., Chou, J., Young, D., Malak, S., Anderson, I.A., 2010. Biomechanical evaluation of different designs of glenospheres in the SMR reverse shoulder prosthesis: micromotion of the baseplate and risk of loosening. Should. Elb. 2, 94–99. Roper, B.A., Paterson, J.M., Day, W.H., 1990. The Roper–Day total shoulder replacement. J. Bone Joint Surg. (Br.) 72, 694–697. Sanchez-Sotelo, J., O'Driscoll, S.W., Torchia, M.E., Cofield, R.H., Rowland, C.M., 2001. Radiographic assessment of cemented humeral components in shoulder arthroplasty. J. Shoulder Elbow Surg. 10, 526–531. Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., Preibisch, S., Rueden, C., Saalfeld, S., Schmid, B., Tinevez, J.Y., White, D.J., Hartenstein, V., Eliceiri, K., Tomancak, P., Cardona, A., 2012. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–682. Torchia, M.E., Cofield, R.H., Settergren, C.R., 1997. Total shoulder arthroplasty with the Neer prosthesis: long-term results. J. Shoulder Elbow Surg. 6, 495–505. Virani, N.A., Harman, M., Li, K., Levy, J., Pupello, D.R., Frankle, M.A., 2008. In vitro and finite element analysis of glenoid bone/baseplate interaction in the reverse shoulder design. J. Shoulder Elbow Surg. 17, 509–521. Westerhoff, P., Graichen, F., Bender, A., Halder, A., Beier, A., Rohlmann, A., Bergmann, G., 2009. In vivo measurement of shoulder joint loads during activities of daily living. J. Biomech. 42, 1840–1849.
Please cite this article as: Favre, P., et al., In vitro initial stability of a stemless humeral implant, Clin. Biomech. (2015), http://dx.doi.org/10.1016/ j.clinbiomech.2015.12.004