- Email: [email protected]
Effects of anterior offsetting of humeral head component in posteriorly unstable total shoulder arthroplasty: Finite element modeling of cadaver specimens
Author’s Accepted Manuscript Effects of Anterior Offsetting of Humeral Head Component in Posteriorly Unstable Total Shoulder Arthroplasty: Finite Element Modeling of Cadaver Specimens Gregory S. Lewis, William K. Conaway, Hwabok Wee, H. Mike Kim www.elsevier.com/locate/jbiomech
PII: DOI: Reference:
S0021-9290(17)30011-8 http://dx.doi.org/10.1016/j.jbiomech.2017.01.010 BM8082
To appear in: Journal of Biomechanics Received date: 22 July 2016 Revised date: 14 November 2016 Accepted date: 2 January 2017 Cite this article as: Gregory S. Lewis, William K. Conaway, Hwabok Wee and H. Mike Kim, Effects of Anterior Offsetting of Humeral Head Component in Posteriorly Unstable Total Shoulder Arthroplasty: Finite Element Modeling of Cadaver Specimens, Journal of Biomechanics, http://dx.doi.org/10.1016/j.jbiomech.2017.01.010 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Humeral head anterior offsetting shoulder arthroplasty Ms. No. BM-D-16-00728
Effects of Anterior Offsetting of Humeral Head Component in Posteriorly Unstable Total Shoulder Arthroplasty: Finite Element Modeling of Cadaver Specimens
Gregory S. Lewis, PhD1 William K. Conaway, BS1 Hwabok Wee, PhD1 H. Mike Kim, MD1
1
Department of Orthopaedics and Rehabilitation, Penn State College of Medicine Milton S. Hershey Medical Center, Hershey, PA 17033
Corresponding Author: H. Mike Kim, MD Assistant Professor Department of Orthopaedics and Rehabilitation Penn State College of Medicine Milton S. Hershey Medical Center Mail Code EC089 Hershey, PA 17033, U.S.A. Email: [email protected] Fax: (717) 531-0349 Office Phone: (717) 531-4826 Key words: posterior instability, total shoulder arthroplasty, glenoid retroversion, anterior offsetting technique, finite element model Word Count: 3890 1
Humeral head anterior offsetting shoulder arthroplasty ABSTRACT A novel technique of “anterior offsetting” of the humeral head component to address posterior instability in total shoulder arthroplasty has been proposed, and its biomechanical benefits have been previously demonstrated experimentally. The present study sought to characterize the changes in joint mechanics associated with anterior offsetting with various amounts of glenoid retroversion using cadaver specimen-specific 3-dimensional finite element models. Specimenspecific computational finite element models were developed through importing digitized locations of six musculotendinous units of the rotator cuff and deltoid muscles based off three cadaveric shoulder specimens implanted with total shoulder arthroplasty in either anatomic or anterior humeral head offset. Additional glenoid retroversion angles (0°, 10°, 20°, and 30°) other than each specimen’s actual retroversion were modeled. Contact area, contact force, peak pressure, center of pressure, and humeral head displacement were calculated at each offset and retroversion for statistical analysis. Anterior offsetting was associated with significant anterior shift of center of pressure and humeral head displacement upon muscle loading (p<0.05). Although statistically insignificant, anterior offsetting was associated with increased contact area and decreased peak pressure (p > 0.05). All study variables showed significant differences when compared between the 4 different glenoid retroversion angles (p > 0.05) except for total force (p > 0.05). The study finding suggests that the anterior offsetting technique may contribute to joint stability in posteriorly unstable shoulder arthroplasty and may reduce eccentric loading on glenoid components although the long term clinical results are yet to be investigated in future.
INTRODUCTION
2
Humeral head anterior offsetting shoulder arthroplasty Total shoulder arthroplasty is commonly performed for treatment of glenohumeral osteoarthritis, which is often complicated by posterior bone loss and excessive retroversion of the glenoid (Denard et al., 2013; Iannotti et al., 2003; Steinmann et al., 2000). This pathologic change is known to cause posterior humeral subluxation and potentially lead to poor clinical outcomes following shoulder arthroplasty (Denard et al., 2013; Hill et al., 2001; Iannotti et al., 2003; Neer et al., 1988; Nowak et al., 2009; Steinmann et al., 2000; Walch et al., 2012). Various surgical techniques have been used to address this issue: eccentric glenoid reaming, (Denard et al., 2013; Hill et al., 2001; Neer et al., 1988; Steinmann et al., 2000) bone grafting of a defective posterior glenoid, (Denard et al., 2013; Hill et al., 2001; Neer et al., 1988; Steinmann et al., 2000) posteriorly augmented glenoid components, (Rice et al., 2008) posterior capsule plication, (Walch et al., 2012) and implanting the humeral stem in less retroversion (Spencer et al., 2005). However, each of these has its own limitations. A novel surgical technique of implanting the humeral head component in an anterior offset position as opposed to an anatomic posterior offset position was proposed to address this issue (Kim et al., 2016) and has been used clinically. The normal humeral head center of rotation is, on average, 2 mm posterior and 7 mm medial to the mid-axis of the humeral medullary canal (Boileau et al., 1997; Robertson et al., 2000). Modular eccentric-offset head components were developed to better restore each patient’s unique offset and are typically implanted in a posterior offset position. Conversely, the novel anterior offsetting technique involves implanting the humeral head component in an anterior offset position rather than in the “anatomic” posterior offset position (Kim et al., 2016). Although there are no published studies reporting the clinical outcomes of the anterior offsetting technique, the early anecdotal results are promising and show concentric containment of the humeral head even in the presence of severe glenoid retroversion. 3
Humeral head anterior offsetting shoulder arthroplasty Previously, we have demonstrated experimentally that anterior offsetting of the humeral head increased the resistance to posterior humeral head translation, shifted joint contact pressures anteriorly, and increased joint contact area in cadaver specimens with simulated glenoid retroversion of 10° or 20° (Kim et al., 2016). These promising findings indicate that anterior offsetting is potentially a simple surgical technique for increasing the joint stability in total shoulder arthroplasty complicated by posterior instability. However, glenoid retroversion could not be varied within a specimen in our previous experimental cadaver study because of the permanent nature of cement fixation of glenoid components in intraosseous holes, and thus, we were unable to assess the within-specimen effects of changes in glenoid retroversion magnitude on changes associated with anterior offsetting of the humeral head component. Computational finite element modeling can provide a tool for systematically analyzing such effects. The purpose of this study was to develop cadaver specimen-specific 3-dimensional finite element models of glenohumeral musculoskeletal mechanics following total shoulder arthroplasty and to characterize the changes in joint mechanics with various degrees of glenoid retroversion. We hypothesized that 1) anterior offsetting of the humeral head component in the presence of glenoid retroversion would shift contact pressures anteriorly and would increase the contact area and 2) the effects of anterior offsetting would increase proportionally to the degree of glenoid retroversion.
MATERIALS AND METHODS Specimen-specific computational finite element models were developed from six cadaveric shoulders from three donors with no gross abnormality or arthritic changes verified with radiographic screening. The mean age of the donors was 53 years (range 35 - 63). There
4
Humeral head anterior offsetting shoulder arthroplasty were 2 males. Total shoulder arthroplasty was performed on each specimen by a single experienced shoulder and elbow surgeon using the standard surgical technique as previously described.6 The humeral head was resected at 30° retroversion and 138° neck-shaft angle using a single cutting guide provided by the manufacturer. Humeral stems with a single size of 9-mm diameter and 130-mm length (Bigliani/Flatow system, Zimmer, Warsaw, IN) were implanted in a press-fit fashion. Then, the humerus was amputated 10 cm proximal to the elbow joint. The glenoid was reamed in eccentric posterior inclination to create 10° or 20° of retroversion. Within
each pair of shoulders, one shoulder was randomly assigned 10° retroversion, and the other 20° retroversion. Three-pegged all-polyethylene glenoid components with 46 mm (circumference diameter) x 46 mm (surface curvature diameter) size were implanted. Modular offset humeral head components with a 46-mm circumference diameter and 18-mm thickness were used. The offset heads had a female cavity for a Morse taper, which was at an eccentric position (3 mm from the geometric center) to reproduce the normal anatomical offset of the humeral head. The head component was implanted first in the anatomic (posterior) offset position and then in the anterior offset position. The anatomic offset position was determined individually for each specimen by finding the rotational position of the head component where the posterosuperior native head was best covered by the head component. The anterior offset position was found by dialing the head component anteriorly to a position that mirrored the magnitude of anteroposterior offset of the anatomic offset position.
All soft tissues were removed with
exception of the rotator cuff musculotendinous units and joint capsule. Each rotator cuff musculotendinous unit was isolated from the scapula and joint capsule. They were then fixed with nylon anchors medially to facilitate loading (Fig. 1). The deltoid was approximated with 5
Humeral head anterior offsetting shoulder arthroplasty two inflexible nylon ropes running from eyelet screws inserted into the deltoid tuberosity through two eyelet screws fixed on the superolateral acromion and coracoid process. Each specimen was fixed in a custom testing frame. The scapular body was fixed with a clamp in a neutral rotation in the coronal and horizontal planes, and 20˚ downward-tilted position in the sagittal plane with the glenoid fossa perpendicular to the floor to mimic the anatomic position of the scapula (Fig. 1). The distal end of the humerus was held by a ring stand in either 15° or 30˚ of abduction. The humerus was placed in a 30˚ external rotation position so that the humeral head component faced the glenoid parallel to the scapular body axis. Hanging weights, cables, and pulleys were used to apply 15N forces independently to each musculotendinous unit (with two 15N forces applied in total to the two deltoid ropes). For each of the six specimens, the 3-dimensional positions of the subscapularis, supraspinatus, infraspinatus, teres minor, anterior deltoid, and posterior deltoid muscles/tendons were digitized while specimens were under stable loading with use of a 3D digitizer (MicroScribe, Revware, Raleigh, NC) at their insertion points on the humeral head and the points where they wrapped around the glenoid neck. These latter points behaved as musculotendinous wrap points and moved approximately 1 mm or less during glenohumeral movements. The humerus was held with the joint reduced in during digitization, and points were referenced to a coordinate system based on glenoid and scapular landmarks. Points were digitized in both the anatomic and the experimental anterior offset positions of the humeral head component. Specimen-specific computational finite element models of the above experimental setup were developed using Abaqus (Implicit, v. 6.13, Dessault Systems) (Fig. 2). Manufacturersupplied .IGS files of the humeral head and dual-radius glenoid of the Bigliani/Flatow system® (Zimmer, Warsaw, IN) were imported. Components were meshed with quadratic tetrahedral
6
Humeral head anterior offsetting shoulder arthroplasty elements (C3D10). Mesh convergence testing with one full model and five different meshes resulted in choosing a model size of 59,000 elements. This model produced differences in anterior-posterior center of contact less than 1% different than a model with 603,000 elements. Linear elastic isotropic material properties were assumed: glenoid component with E= 1.0 GPa & ν=0.4 (ultra-high molecular weight polyethylene), (Yongpravat et al., 2013) and humeral head component with E= 200 GPa and ν=0.29 (cobalt chromium).(Lacroix et al., 1997) An arm weight of 5% of donor bodyweight (Winter, 2005) was applied, fixed with respect to the humeral head at the approximated center of the arm (based on known abduction and flexion angles of digitized specimens). The medial surface of the glenoid was fixed in space. To test joint stability under fixed shoulder abduction and flexion, the humeral head was allowed to translate in any direction and rotate about the arm axis, but other rotations were fixed to zero. A sliding contact model between the components with Coulomb friction ( = 0.07) (Hopkins et al., 2006; Zhang et al., 2013), pressure-overclosure relationship without physical softening (hard contact), and contact controls, was modeled. Self-actuated axial connectors (CONN3D2 elements) were used to simulate each of the six musculotendinous units: subscapularis, supraspinatus, infraspinatus, teres minor, anterior deltoid, and posterior deltoid. Initial endpoints were the 3D points as digitized on the cadaver specimens (Fig. 2). Anterior-posterior coordinates for left-sided specimens were multiplied by -1 to create all right-sided models. Humeral insertion points were constrained to move with the humeral head, whereas musculotendinous wrap points around the scapula were fixed along with the glenoid neck. An initial simulation step was used to translate the humeral head (without rotations) toward the glenoid to bring the two into concentric contact.
Then the six
musculotendinous units were activated with 15N each, the same magnitudes for muscle forces
7
Humeral head anterior offsetting shoulder arthroplasty used in the previous physical experiment (Kim et al., 2016). Sensitivity of results to magnitude of muscle forces was tested by using magnitudes of 50 N in additional models of one specimen. Static stress analysis with large displacement formulation was used in Abaqus. Humeral head displacement relative to initial concentric position, glenohumeral contact area (Abaqus CAREA), total contact force (due to contact pressure, CFNM), center of pressure location (XN), and peak pressure (CPRESS) were computed. The center of pressure location was relative to the center of the glenoid component as determined by coordinate transformation to a glenoid local coordinate system.
Displacements and positions used the convention of anterior and superior being
positive, and posterior and inferior being negative. In Part I, models were developed corresponding to the six separate digitized shoulders, each with the two humeral head offset conditions. Statistical comparisons between the anatomic and anterior-offset humeral head models were performed using Wilcoxon signed ranks tests (n = 6).
In Part II, for the three shoulders digitized with 30° arm abduction (S1, S2, and S3),
additional glenoid retroversion conditions were modeled to investigate the effect of glenoid retroversion (n = 3). S1 and S2 had been implanted and experimentally digitized in 20° glenoid retroversion, and S3 was implanted and digitized in 10° retroversion. Four glenoid retroversion angles were modeled: 0°, 10°, 20°, and 30°. Additional retroversion angles other than the angle physically created on the specimens were modeled by rotating the glenoid about the humeral head center in the transverse plane to a new retroversion angle. The musculotendinous scapular wrap points were translated the same as the glenoid center translation associated with a new retroversion angle. Wilcoxon signed ranks tests were used to compare the variables between the two offset positions in each sample at each retroversion angle. In order to evaluate the effect of glenoid retroversion angles on the variables, the data were compared between the 4 glenoid
8
Humeral head anterior offsetting shoulder arthroplasty retroversion angles using Friedman test each specimen at each offset position. Statistical significance was set at p = 0.05. The data are shown as mean ± standard deviation.
RESULTS Part I Significant differences between anatomic and anterior offset positions were detected in the center of pressure locations in both anteroposterior and superoinferior directions, and in the humeral head displacement in the anteroposterior direction (* in Fig. 3).
The mean
anteroposterior coordinate of the center of pressure was -6.1 ± 0.8 mm and -4.7 ± 0.7 mm for the anatomic and anterior offset positions, respectively (p = 0.028), indicating that the anterior offset position shifted the center of pressure anteriorly upon muscle loading. The mean superoinferior coordinate of the center of pressure was -2.5 ± 2.2 mm and -3.5 ± 2.4 mm for the anatomic and anterior offset positions, respectively (p = 0.028). The mean anteroposterior displacement of the humeral head was -0.05 ± 0.15 mm and 0.09 ± 0.18 mm for the anatomic and anterior offset positions, respectively (p = 0.028), indicating that the anterior offset position resulted in a more anterior translation of the humeral head upon muscle loading. No significant differences in contact area, total force, and peak pressure were detected comparing the anatomic and anterior offset humeral head component positions (p > 0.05) (Fig. 3). The mean contact area was 153 ± 51 mm2 and 190 ± 63 mm2 for the anatomic and anterior offset positions, respectively (p = 0.12). The mean total force was 71.0 ± 4.6 N and 72.2 ± 5.2 N for the anatomic and anterior offset positions, respectively (p = 0.75). The mean peak pressure was 2.9 ± 1.0 MPa and 2.5 ± 1.1 MPa for the anatomic and anterior offset positions, respectively (p = 0.17).
9
Humeral head anterior offsetting shoulder arthroplasty For the anatomical offset the resultant force vector acting on the humerus from all muscles (at the start of the simulation) was angulated anteromedially an average of 7.7º ± 2.6º away from the scapula body plane in the transverse plane, and for the anterior offset this angle averaged 16.3º ± 2.3º indicating a more anteriorly directed component of muscle force. Tests of result sensitivity to muscle force magnitude in models of one specimen showed that, as expected, contact area, total force, and peak pressure were sensitive to muscle force magnitude (Fig. S-1). Superoinferior coordinate of the center of pressure was sensitive to muscle force; however anteroposterior coordinate of the center of pressure was not very sensitive, changing from -6.0 mm to -6.2 mm for the anatomic offset position with 15 and 50 N muscle forces, respectively, and changing from -3.7 mm to -3.8 mm for the anterior offset position with 15 and 50 N muscle forces, respectively. Part II The results from the additional glenoid retroversion angles modeled for the three 30° abduction specimens revealed fairly consistent trends, but the models for specimen #3 (S3) showed aberrant results at 30° glenoid retroversion (Fig. 4 and S-2) . Paired comparisons between the anatomic and anterior offset positions at each glenoid retroversion angle (Wilcoxon signed ranks test, n=3) showed no significant difference in any of the variables (p > 0.05) although there was a trend that anterior offsetting was associated with increased contact area, decreased peak pressure, anterior shifting of center of pressure, and anterior shifting of humeral head displacement (Fig. 4 and S-2).
Repeated comparisons between 4 different glenoid
retroversion angles (Friedman test) showed that all variables were significantly different between the retroversion angles (p < 0.05) († and ‡ in Fig. 4 and S-2) except total force (p > 0.05). The distribution of contact pressures on the glenoid component is shown in Figure 5. There was a 10
Humeral head anterior offsetting shoulder arthroplasty trend that the magnitude of changes associated with anterior offsetting compared to anatomic offsetting in contact area, center of pressure, and peak pressure increased with increasing glenoid retroversion (Fig. S-3), but statistical tests were not performed on these data due to the small sample size and unknown normality of the data distribution.
DISCUSSION The biomechanical benefits of anterior offsetting of the humeral component have been previously demonstrated in posteriorly unstable shoulder arthroplasty simulated on cadaveric shoulders with artificially created glenoid retroversion (Kim et al., 2016). It was found that the force and energy required to displace the humeral head posteriorly increased significantly with the anterior offset position compared to the anatomic offset position. The joint contact pressures were significantly shifted anteriorly, and the joint contact area significantly increased with the anterior offset position. The study concluded that anterior offsetting of the humeral head component increased the resistance to posterior humeral head translation, shifted joint contact pressures anteriorly, and increased joint contact area, thus, potentially increasing the joint stability. The present study was performed to investigate this finding in various degrees of glenoid retroversion with use of 3-dimensional finite element models produced based off cadaveric shoulder specimens. The present study found that anterior offsetting of the humeral component resulted in a significant anterior shift of the joint contact pressure and humeral head position upon muscle loading. This finding is in agreement with the results of the previous ex vivo biomechanical study (Kim et al., 2016) and supports the first study hypothesis. Although its exact clinical implications are unknown at this point, this finding together with the observation of successful
11
Humeral head anterior offsetting shoulder arthroplasty early clinical results suggests that the anterior offsetting technique may contribute to the joint stability in posteriorly unstable total shoulder arthroplasty due to increased glenoid retroversion and may reduce the eccentric loading on glenoid components implanted in retroversion (Kim et al., 2016). This study also demonstrated that the anterior offsetting led to a more anterior orientation of the resultant muscle force vector acting on the humeral head, providing a potential explanation for the mechanism underlying the resulting changes in shoulder biomechanics. The mean shift of the anteroposterior joint contact pressure was similar in the models (1.37 mm) and corresponding specimen experiments (1.35 mm). However superoinferior mean shifts were not similar, and no significant correlations were detected for pairwise comparison between model and experimental data on a specimen-specific basis (p>.05). It has been previously shown (Kim et al., 2016) that the biomechanical benefits of anterior offsetting were more prominent in glenoids with more severe retroversion (10° vs. 20° retroversion). This apparent incremental effect of anterior offsetting with increasing glenoid retroversion was tested in the present study as the second hypothesis. It was found that all three simulated samples showed a trend for an incremental effect in contact area, peak pressure, and anteroposterior location of center of pressure with increasing glenoid retroversion except for one sample (S3) at 30° retroversion. This trend failed to reach the statistical significance possibly due to the small sample size of the study and the aberrant behavior of the one sample at 30° retroversion. Our modeling focused on the shoulder stabilization in the midrange of motion, in which the passive capsule and ligamentous tissues are more lax, and stability is obtained by compression of the humeral head into the concave glenoid by active muscle contraction (Lippitt et al., 1993). Our focus on smaller loads applied to the rotator cuff and deltoid is based on
12
Humeral head anterior offsetting shoulder arthroplasty Nyffeler et al.’s (Nyffeler et al., 2006) biomechanical study of glenohumeral stability following shoulder replacement, and our own recent cadaver study (Kim et al., 2016). According to our sensitivity testing, the important anteroposterior coordinate of the center of pressure was not very sensitive to changes in muscle force magnitude, although other results were sensitive to these changes. Buchler and Farron (Buchler et al., 2004) used a finite element model to assess the effects of humeral head geometry on replaced shoulder biomechanics; rotator cuff muscles were included but the deltoid was not. Models of active glenohumeral stability with substantially greater complexity have been reported, such as Viehofer et al. (Viehofer et al., 2016) who modeled 27 individual muscle segments and muscle forces estimated by optimization. However simplified loading schemes have been used in experimental and computational studies of effects of implant design on stability. For instance, Hopkins et al. (Hopkins et al., 2007) used finite element modeling to investigate the effects of implant curvatures on glenohumeral stability. In that study the authors applied an assumed resultant compressive and shear load to the humeral head, instead of applying individual muscle loads. This modeling study had several limitations. Models were constructed from digitization of non-arthritic cadaver specimens implanted with arthroplasty components.
Physical
digitization in cadaver specimens (for which we also had experimental biomechanical measurements) was used because of the ability to directly visualize the musculotendon soft tissues. Image-based modeling from CT or MRI would be advantageous for future work in modeling arthritic patients.
Muscle forces were simplified as linear forces of a constant
magnitude acting at single points on the bones. This simplification was consistent between the humeral head offset conditions, and assumed that more realistic muscles and muscle attachments would not substantially affect the relative effects of humeral head offset condition on joint
13
Humeral head anterior offsetting shoulder arthroplasty kinematics and contact pressures.
It has been demonstrated for other anatomical sites that
muscle modeling and boundary condition assumptions can impact results (Phillips et al., 2007; Speirs et al., 2007) Our omission of muscles such as latissimus dorsi is supported by a recent complex model of active glenohumeral stability (Favre et al., 2012). That study also showed force contributions from the long head biceps and pectoralis major at specific arm elevation angles, although contributions from the deltoid and rotator cuff were more substantial.
In some
cases, there was passage of the muscle force line through a small portion of the humeral head, indicating in reality that the musculotendinous tissue wrapped around the component and wrap points should be implemented in further modeling efforts.
Having muscles crossing the
components is not physiological and will change their moment arm and, therefore, their contribution to the biomechanics of the shoulder joint. Dynamic shoulder movements were not modeled. Scapulothoracic motion (Braman et al., 2009) and its effects on muscle forces crossing the glenohumeral joint were ignored. Other soft tissues surrounding the glenohumeral joint, especially the joint capsule, were not taken into account. Multiple glenoid retroversion angles were modeled with hypothetical shifts in muscle wrap point locations instead of digitization of additional specimens implanted with those retroversion angles. The glenoid component was fixed at its medial surface instead of modeling the bone and cement. Cement and bone have been neglected previously in modeling (Oosterom et al., 2003) and experimental (Nyffeler et al., 2006) studies of glenohumeral stability, and the approach is supported by our modeling (Wee et al., 2015) and experimental (Lewis et al., 2016) observations of motions at the component medial border typically less than 0.05 mm relative to the fixed glenoid bone (Fig. S-4). However, large retroversion angles can potentially have impacts on implant micromotions relative to the cement and bone (Farron et al., 2006). Due to the sample size and unknown distribution of the data,
14
Humeral head anterior offsetting shoulder arthroplasty nonparametric tests were used throughout the study. As nonparametric tests are generally regarded less powerful tests than parametric tests in detecting true statistical significance, the overall impact of exclusive usage of nonparametric tests on the study results is unknown. Another potential limitation is that the present study used glenoid components that had the same radius of curvature as the humeral heads, which might have resulted in a higher degree of constraint and a lower degree of contact than the constructs with a radial mismatch. However, the overall study results would not have been affected significantly because the same construct was used across the groups and specimens in the present study. It should be noted that the anatomic variations of the individual cadaveric specimens were not considered in this study. Although the exact impact of these anatomic variations is unknown, it is unlikely that the study results were affected significantly as all the specimens were prepared in the same way for the implantation, and statistical comparisons were conducted either within a single specimen or within a pair that had almost identical anatomy. Lastly, the effects of varying degrees of anterior offset were not investigated in this study due to the practical difficulty in estimating the locations of the rotator cuff insertion sites upon offset changes. Only one anterior offset was used for each sample in this study, which was the exact opposite of the anatomic offset of each sample.
ACKNOWLEDGEMENT Funding for related experiments provided by Orthopaedic Research and Education Foundation, who were not involved further in the study and manuscript.
CONFLICT OF INTEREST STATEMENT
15
Humeral head anterior offsetting shoulder arthroplasty H.M.K. received materials from Zimmer to support this research. G.S.L. has received materials from Depuy-Synthes for unrelated research.
16
Humeral head anterior offsetting shoulder arthroplasty REFERENCES
Boileau, P., Walch, G. 1997. The three-dimensional geometry of the proximal humerus. Implications for surgical technique and prosthetic design. The Journal of bone and joint surgery British volume. 79, 857-65. Braman, J.P., Engel, S.C., Laprade, R.F., Ludewig, P.M. 2009. In vivo assessment of scapulohumeral rhythm during unconstrained overhead reaching in asymptomatic subjects. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 18, 960-7. Buchler, P., Farron, A. 2004. Benefits of an anatomical reconstruction of the humeral head during shoulder arthroplasty: a finite element analysis. Clinical biomechanics. 19, 16-23. Denard, P.J., Walch, G. 2013. Current concepts in the surgical management of primary glenohumeral arthritis with a biconcave glenoid. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 22, 1589-98. Farron, A., Terrier, A., Buchler, P. 2006. Risks of loosening of a prosthetic glenoid implanted in retroversion. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 15, 521-6. Favre, P., Sentelera, M., Hippa, J., Scherrera, S., Gerbera, C., Snedeker, J. 2012. An integrated model of active glenohumeral stability. J Biomechanics. 45, 2248-55. Hill, J.M., Norris, T.R. 2001. Long-term results of total shoulder arthroplasty following bonegrafting of the glenoid. The Journal of bone and joint surgery American volume. 83-A, 877-83. Hopkins, A.R., Hansen, U.N., Amis, A.A., Taylor, M., Emery, R.J. 2007. Glenohumeral kinematics following total shoulder arthroplasty: a finite element investigation. Journal of 17
Humeral head anterior offsetting shoulder arthroplasty orthopaedic research : official publication of the Orthopaedic Research Society. 25, 10815. Hopkins, A.R., Hansen, U.N., Amis, A.A., Taylor, M., Gronau, N., Anglin, C. 2006. Finite element modelling of glenohumeral kinematics following total shoulder arthroplasty. Journal of biomechanics. 39, 2476-83. Iannotti, J.P., Norris, T.R. 2003. Influence of preoperative factors on outcome of shoulder arthroplasty for glenohumeral osteoarthritis. The Journal of bone and joint surgery American volume. 85-A, 251-8. Kim, H.M., Chacon, A.C., Andrews, S.H., Roush, E.P., Cho, E., Conaway, W.K., Kunselman, A.R., Lewis, G.S. 2016. Biomechanical benefits of anterior offsetting of humeral head component in posteriorly unstable total shoulder arthroplasty: A cadaveric study. Journal of orthopaedic research : official publication of the Orthopaedic Research Society. 34, 666-74. Lacroix, D., Prendergast, P.J. 1997. Stress analysis of glenoid component designs for shoulder arthroplasty. Proceedings of the Institution of Mechanical Engineers Part H, Journal of engineering in medicine. 211, 467-74. Lewis, G.S., Brenza, J.B., Paul, E.M., Armstrong, A.D. 2016. Construct damage and loosening around glenoid implants: A longitudinal micro-CT study of five cadaver specimens. Journal of orthopaedic research : official publication of the Orthopaedic Research Society. 34, 1053-60. Lippitt, S., Matsen, F. 1993. Mechanisms of glenohumeral joint stability. Clinical orthopaedics and related research. 20-8.
18
Humeral head anterior offsetting shoulder arthroplasty Neer, C.S., 2nd, Morrison, D.S. 1988. Glenoid bone-grafting in total shoulder arthroplasty. The Journal of bone and joint surgery American volume. 70, 1154-62. Nowak, D.D., Bahu, M.J., Gardner, T.R., Dyrszka, M.D., Levine, W.N., Bigliani, L.U., Ahmad, C.S. 2009. Simulation of surgical glenoid resurfacing using three-dimensional computed tomography of the arthritic glenohumeral joint: the amount of glenoid retroversion that can be corrected. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 18, 680-8. Nyffeler, R.W., Sheikh, R., Atkinson, T.S., Jacob, H.A., Favre, P., Gerber, C. 2006. Effects of glenoid component version on humeral head displacement and joint reaction forces: an experimental study. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 15, 625-9. Oosterom, R., Herder, J.L., van der Helm, F.C., Swieszkowski, W., Bersee, H.E. 2003. Translational stiffness of the replaced shoulder joint. Journal of biomechanics. 36, 1897907. Phillips, A.T., Pankaj, P., Howie, C.R., Usmani, A.S., Simpson, A.H. 2007. Finite element modelling of the pelvis: inclusion of muscular and ligamentous boundary conditions. Medical engineering & physics. 29, 739-48. Rice, R.S., Sperling, J.W., Miletti, J., Schleck, C., Cofield, R.H. 2008. Augmented glenoid component for bone deficiency in shoulder arthroplasty. Clinical orthopaedics and related research. 466, 579-83. Robertson, D.D., Yuan, J., Bigliani, L.U., Flatow, E.L., Yamaguchi, K. 2000. Three-dimensional analysis of the proximal part of the humerus: relevance to arthroplasty. The Journal of bone and joint surgery American volume. 82-A, 1594-602.
19
Humeral head anterior offsetting shoulder arthroplasty Speirs, A.D., Heller, M.O., Duda, G.N., Taylor, W.R. 2007. Physiologically based boundary conditions in finite element modelling. Journal of biomechanics. 40, 2318-23. Spencer, E.E., Jr., Valdevit, A., Kambic, H., Brems, J.J., Iannotti, J.P. 2005. The effect of humeral component anteversion on shoulder stability with glenoid component retroversion. The Journal of bone and joint surgery American volume. 87, 808-14. Steinmann, S.P., Cofield, R.H. 2000. Bone grafting for glenoid deficiency in total shoulder replacement. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 9, 361-7. Viehofer, A.F., Snedeker, J.G., Baumgartner, D., Gerber, C. 2016. Glenohumeral joint reaction forces increase with critical shoulder angles representative of osteoarthritis-A biomechanical analysis. Journal of orthopaedic research : official publication of the Orthopaedic Research Society. 34, 1047-52. Walch, G., Moraga, C., Young, A., Castellanos-Rosas, J. 2012. Results of anatomic nonconstrained prosthesis in primary osteoarthritis with biconcave glenoid. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 21, 152633. Wee, H., Armstrong, A.D., Flint, W.W., Kunselman, A.R., Lewis, G.S. 2015. Peri-implant stress correlates with bone and cement morphology: Micro-FE modeling of implanted cadaveric glenoids. Journal of orthopaedic research : official publication of the Orthopaedic Research Society. 33, 1671-9. Winter, D.A. Biomechanics and Motor Control of Human Movement. Hoboken, NJ: John Wiley & Sons; 2005.
20
Humeral head anterior offsetting shoulder arthroplasty Yongpravat, C., Kim, H.M., Gardner, T.R., Bigliani, L.U., Levine, W.N., Ahmad, C.S. 2013. Glenoid implant orientation and cement failure in total shoulder arthroplasty: a finite element analysis. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 22, 940-7. Zhang, J., Yongpravat, C., Kim, H.M., Levine, W.N., Bigliani, L.U., Gardner, T.R., Ahmad, C.S. 2013. Glenoid articular conformity affects stress distributions in total shoulder arthroplasty. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 22, 350-6.
21
Humeral head anterior offsetting shoulder arthroplasty FIGURE LEGENDS Figure 1. Cadaveric shoulder specimen setup on a testing frame. Each rotator cuff musculotendinous unit was isolated from the scapula and joint capsule. They were then fixed with nylon anchors medially to facilitate loading. The deltoid was approximated with two inflexible nylon ropes running from eyelet screws inserted into the deltoid tuberosity through two eyelet screws fixed on the acromion and coracoid process. The scapular body was fixed with a clamp in a neutral rotation in the coronal and horizontal planes, and 20˚ downward-tilted position in the sagittal plane with the glenoid fossa perpendicular to the floor. The distal end of the humerus was held by a ring stand. Hanging weights, cables, and pulleys were used to apply 15N forces independently to each musculotendinous unit (with two 15N forces applied in total to the two deltoid ropes).
Figure 2. Axial (left) and anteroposterior (right) views of the finite element model of a specimen depicting the humeral head and glenoid components, initial lines of action of the six musculotendinous units (red lines), fixed scapular wrap points (blue triangles), moving humeral attachment points (green circles), long axis of the humerus (black line), and fixed glenoid constraint (light blue box). Meshed humeral head and glenoid components are also shown at the bottom. Ant = anterior, Post = posterior, Sup = superior, Inf = inferior.
Figure 3. Comparisons between the anatomic and anterior offset positions of the humeral head components constructed from experimentally digitized cadaveric shoulder specimens (n = 6), each with glenoid retroversion of either 10° or 20°. * p < 0.05 for Wilcoxon signed ranks tests
22
Humeral head anterior offsetting shoulder arthroplasty comparing between anatomic and anterior offset positions.
Error bars are +/- 1 standard
deviation.
Figure 4. Results from the additional glenoid retroversion angles modeled for the three 30° abduction specimens. † = p < 0.05 for Friedman test comparing between retroversion angles at anatomic offset position. ‡ = p < 0.05 for Friedman test comparing between retroversion angles at anterior offset position. A-P = anteroposterior.
Figure 5.
Predicted glenoid contact pressure distributions for model of specimen S2, for
anatomic and anterior humeral head offset and four glenoid retroversion angles. Center of contact pressure is shown with the red plus sign. Note the increased anterior shift associated with the anterior humeral head offset for larger glenoid retroversion angles. Anterior is to the right, posterior to the left.
Supplementary Data Figure S-1. Results of sensitivity analysis of change in muscle force magnitudes from 15 N to 50 N in one randomly selected specimen model.
Figure S-2. Results from the additional glenoid retroversion angles modeled for the three 30° abduction specimens. † = p < 0.05 for Friedman test comparing between retroversion angles at
23
Humeral head anterior offsetting shoulder arthroplasty anatomic offset position. ‡ = p < 0.05 for Friedman test comparing between retroversion angles at anterior offset position. S-I = superoinferior.
Figure S-3. Changes in study variables associated with humeral head anterior offsetting vs. glenoid retroversion angle. Each of the three 30° abduction specimens (S1, S2, and S3) is plotted, and each plotted data point represents the result for anterior offset position value minus the result for anatomic offset position.
24
Humeral head anterior offsetting shoulder arthroplasty
25
Humeral head anterior offsetting shoulder arthroplasty
26
Humeral head anterior offsetting shoulder arthroplasty
27
PII: DOI: Reference:
S0021-9290(17)30011-8 http://dx.doi.org/10.1016/j.jbiomech.2017.01.010 BM8082
To appear in: Journal of Biomechanics Received date: 22 July 2016 Revised date: 14 November 2016 Accepted date: 2 January 2017 Cite this article as: Gregory S. Lewis, William K. Conaway, Hwabok Wee and H. Mike Kim, Effects of Anterior Offsetting of Humeral Head Component in Posteriorly Unstable Total Shoulder Arthroplasty: Finite Element Modeling of Cadaver Specimens, Journal of Biomechanics, http://dx.doi.org/10.1016/j.jbiomech.2017.01.010 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Humeral head anterior offsetting shoulder arthroplasty Ms. No. BM-D-16-00728
Effects of Anterior Offsetting of Humeral Head Component in Posteriorly Unstable Total Shoulder Arthroplasty: Finite Element Modeling of Cadaver Specimens
Gregory S. Lewis, PhD1 William K. Conaway, BS1 Hwabok Wee, PhD1 H. Mike Kim, MD1
1
Department of Orthopaedics and Rehabilitation, Penn State College of Medicine Milton S. Hershey Medical Center, Hershey, PA 17033
Corresponding Author: H. Mike Kim, MD Assistant Professor Department of Orthopaedics and Rehabilitation Penn State College of Medicine Milton S. Hershey Medical Center Mail Code EC089 Hershey, PA 17033, U.S.A. Email: [email protected] Fax: (717) 531-0349 Office Phone: (717) 531-4826 Key words: posterior instability, total shoulder arthroplasty, glenoid retroversion, anterior offsetting technique, finite element model Word Count: 3890 1
Humeral head anterior offsetting shoulder arthroplasty ABSTRACT A novel technique of “anterior offsetting” of the humeral head component to address posterior instability in total shoulder arthroplasty has been proposed, and its biomechanical benefits have been previously demonstrated experimentally. The present study sought to characterize the changes in joint mechanics associated with anterior offsetting with various amounts of glenoid retroversion using cadaver specimen-specific 3-dimensional finite element models. Specimenspecific computational finite element models were developed through importing digitized locations of six musculotendinous units of the rotator cuff and deltoid muscles based off three cadaveric shoulder specimens implanted with total shoulder arthroplasty in either anatomic or anterior humeral head offset. Additional glenoid retroversion angles (0°, 10°, 20°, and 30°) other than each specimen’s actual retroversion were modeled. Contact area, contact force, peak pressure, center of pressure, and humeral head displacement were calculated at each offset and retroversion for statistical analysis. Anterior offsetting was associated with significant anterior shift of center of pressure and humeral head displacement upon muscle loading (p<0.05). Although statistically insignificant, anterior offsetting was associated with increased contact area and decreased peak pressure (p > 0.05). All study variables showed significant differences when compared between the 4 different glenoid retroversion angles (p > 0.05) except for total force (p > 0.05). The study finding suggests that the anterior offsetting technique may contribute to joint stability in posteriorly unstable shoulder arthroplasty and may reduce eccentric loading on glenoid components although the long term clinical results are yet to be investigated in future.
INTRODUCTION
2
Humeral head anterior offsetting shoulder arthroplasty Total shoulder arthroplasty is commonly performed for treatment of glenohumeral osteoarthritis, which is often complicated by posterior bone loss and excessive retroversion of the glenoid (Denard et al., 2013; Iannotti et al., 2003; Steinmann et al., 2000). This pathologic change is known to cause posterior humeral subluxation and potentially lead to poor clinical outcomes following shoulder arthroplasty (Denard et al., 2013; Hill et al., 2001; Iannotti et al., 2003; Neer et al., 1988; Nowak et al., 2009; Steinmann et al., 2000; Walch et al., 2012). Various surgical techniques have been used to address this issue: eccentric glenoid reaming, (Denard et al., 2013; Hill et al., 2001; Neer et al., 1988; Steinmann et al., 2000) bone grafting of a defective posterior glenoid, (Denard et al., 2013; Hill et al., 2001; Neer et al., 1988; Steinmann et al., 2000) posteriorly augmented glenoid components, (Rice et al., 2008) posterior capsule plication, (Walch et al., 2012) and implanting the humeral stem in less retroversion (Spencer et al., 2005). However, each of these has its own limitations. A novel surgical technique of implanting the humeral head component in an anterior offset position as opposed to an anatomic posterior offset position was proposed to address this issue (Kim et al., 2016) and has been used clinically. The normal humeral head center of rotation is, on average, 2 mm posterior and 7 mm medial to the mid-axis of the humeral medullary canal (Boileau et al., 1997; Robertson et al., 2000). Modular eccentric-offset head components were developed to better restore each patient’s unique offset and are typically implanted in a posterior offset position. Conversely, the novel anterior offsetting technique involves implanting the humeral head component in an anterior offset position rather than in the “anatomic” posterior offset position (Kim et al., 2016). Although there are no published studies reporting the clinical outcomes of the anterior offsetting technique, the early anecdotal results are promising and show concentric containment of the humeral head even in the presence of severe glenoid retroversion. 3
Humeral head anterior offsetting shoulder arthroplasty Previously, we have demonstrated experimentally that anterior offsetting of the humeral head increased the resistance to posterior humeral head translation, shifted joint contact pressures anteriorly, and increased joint contact area in cadaver specimens with simulated glenoid retroversion of 10° or 20° (Kim et al., 2016). These promising findings indicate that anterior offsetting is potentially a simple surgical technique for increasing the joint stability in total shoulder arthroplasty complicated by posterior instability. However, glenoid retroversion could not be varied within a specimen in our previous experimental cadaver study because of the permanent nature of cement fixation of glenoid components in intraosseous holes, and thus, we were unable to assess the within-specimen effects of changes in glenoid retroversion magnitude on changes associated with anterior offsetting of the humeral head component. Computational finite element modeling can provide a tool for systematically analyzing such effects. The purpose of this study was to develop cadaver specimen-specific 3-dimensional finite element models of glenohumeral musculoskeletal mechanics following total shoulder arthroplasty and to characterize the changes in joint mechanics with various degrees of glenoid retroversion. We hypothesized that 1) anterior offsetting of the humeral head component in the presence of glenoid retroversion would shift contact pressures anteriorly and would increase the contact area and 2) the effects of anterior offsetting would increase proportionally to the degree of glenoid retroversion.
MATERIALS AND METHODS Specimen-specific computational finite element models were developed from six cadaveric shoulders from three donors with no gross abnormality or arthritic changes verified with radiographic screening. The mean age of the donors was 53 years (range 35 - 63). There
4
Humeral head anterior offsetting shoulder arthroplasty were 2 males. Total shoulder arthroplasty was performed on each specimen by a single experienced shoulder and elbow surgeon using the standard surgical technique as previously described.6 The humeral head was resected at 30° retroversion and 138° neck-shaft angle using a single cutting guide provided by the manufacturer. Humeral stems with a single size of 9-mm diameter and 130-mm length (Bigliani/Flatow system, Zimmer, Warsaw, IN) were implanted in a press-fit fashion. Then, the humerus was amputated 10 cm proximal to the elbow joint. The glenoid was reamed in eccentric posterior inclination to create 10° or 20° of retroversion. Within
each pair of shoulders, one shoulder was randomly assigned 10° retroversion, and the other 20° retroversion. Three-pegged all-polyethylene glenoid components with 46 mm (circumference diameter) x 46 mm (surface curvature diameter) size were implanted. Modular offset humeral head components with a 46-mm circumference diameter and 18-mm thickness were used. The offset heads had a female cavity for a Morse taper, which was at an eccentric position (3 mm from the geometric center) to reproduce the normal anatomical offset of the humeral head. The head component was implanted first in the anatomic (posterior) offset position and then in the anterior offset position. The anatomic offset position was determined individually for each specimen by finding the rotational position of the head component where the posterosuperior native head was best covered by the head component. The anterior offset position was found by dialing the head component anteriorly to a position that mirrored the magnitude of anteroposterior offset of the anatomic offset position.
All soft tissues were removed with
exception of the rotator cuff musculotendinous units and joint capsule. Each rotator cuff musculotendinous unit was isolated from the scapula and joint capsule. They were then fixed with nylon anchors medially to facilitate loading (Fig. 1). The deltoid was approximated with 5
Humeral head anterior offsetting shoulder arthroplasty two inflexible nylon ropes running from eyelet screws inserted into the deltoid tuberosity through two eyelet screws fixed on the superolateral acromion and coracoid process. Each specimen was fixed in a custom testing frame. The scapular body was fixed with a clamp in a neutral rotation in the coronal and horizontal planes, and 20˚ downward-tilted position in the sagittal plane with the glenoid fossa perpendicular to the floor to mimic the anatomic position of the scapula (Fig. 1). The distal end of the humerus was held by a ring stand in either 15° or 30˚ of abduction. The humerus was placed in a 30˚ external rotation position so that the humeral head component faced the glenoid parallel to the scapular body axis. Hanging weights, cables, and pulleys were used to apply 15N forces independently to each musculotendinous unit (with two 15N forces applied in total to the two deltoid ropes). For each of the six specimens, the 3-dimensional positions of the subscapularis, supraspinatus, infraspinatus, teres minor, anterior deltoid, and posterior deltoid muscles/tendons were digitized while specimens were under stable loading with use of a 3D digitizer (MicroScribe, Revware, Raleigh, NC) at their insertion points on the humeral head and the points where they wrapped around the glenoid neck. These latter points behaved as musculotendinous wrap points and moved approximately 1 mm or less during glenohumeral movements. The humerus was held with the joint reduced in during digitization, and points were referenced to a coordinate system based on glenoid and scapular landmarks. Points were digitized in both the anatomic and the experimental anterior offset positions of the humeral head component. Specimen-specific computational finite element models of the above experimental setup were developed using Abaqus (Implicit, v. 6.13, Dessault Systems) (Fig. 2). Manufacturersupplied .IGS files of the humeral head and dual-radius glenoid of the Bigliani/Flatow system® (Zimmer, Warsaw, IN) were imported. Components were meshed with quadratic tetrahedral
6
Humeral head anterior offsetting shoulder arthroplasty elements (C3D10). Mesh convergence testing with one full model and five different meshes resulted in choosing a model size of 59,000 elements. This model produced differences in anterior-posterior center of contact less than 1% different than a model with 603,000 elements. Linear elastic isotropic material properties were assumed: glenoid component with E= 1.0 GPa & ν=0.4 (ultra-high molecular weight polyethylene), (Yongpravat et al., 2013) and humeral head component with E= 200 GPa and ν=0.29 (cobalt chromium).(Lacroix et al., 1997) An arm weight of 5% of donor bodyweight (Winter, 2005) was applied, fixed with respect to the humeral head at the approximated center of the arm (based on known abduction and flexion angles of digitized specimens). The medial surface of the glenoid was fixed in space. To test joint stability under fixed shoulder abduction and flexion, the humeral head was allowed to translate in any direction and rotate about the arm axis, but other rotations were fixed to zero. A sliding contact model between the components with Coulomb friction ( = 0.07) (Hopkins et al., 2006; Zhang et al., 2013), pressure-overclosure relationship without physical softening (hard contact), and contact controls, was modeled. Self-actuated axial connectors (CONN3D2 elements) were used to simulate each of the six musculotendinous units: subscapularis, supraspinatus, infraspinatus, teres minor, anterior deltoid, and posterior deltoid. Initial endpoints were the 3D points as digitized on the cadaver specimens (Fig. 2). Anterior-posterior coordinates for left-sided specimens were multiplied by -1 to create all right-sided models. Humeral insertion points were constrained to move with the humeral head, whereas musculotendinous wrap points around the scapula were fixed along with the glenoid neck. An initial simulation step was used to translate the humeral head (without rotations) toward the glenoid to bring the two into concentric contact.
Then the six
musculotendinous units were activated with 15N each, the same magnitudes for muscle forces
7
Humeral head anterior offsetting shoulder arthroplasty used in the previous physical experiment (Kim et al., 2016). Sensitivity of results to magnitude of muscle forces was tested by using magnitudes of 50 N in additional models of one specimen. Static stress analysis with large displacement formulation was used in Abaqus. Humeral head displacement relative to initial concentric position, glenohumeral contact area (Abaqus CAREA), total contact force (due to contact pressure, CFNM), center of pressure location (XN), and peak pressure (CPRESS) were computed. The center of pressure location was relative to the center of the glenoid component as determined by coordinate transformation to a glenoid local coordinate system.
Displacements and positions used the convention of anterior and superior being
positive, and posterior and inferior being negative. In Part I, models were developed corresponding to the six separate digitized shoulders, each with the two humeral head offset conditions. Statistical comparisons between the anatomic and anterior-offset humeral head models were performed using Wilcoxon signed ranks tests (n = 6).
In Part II, for the three shoulders digitized with 30° arm abduction (S1, S2, and S3),
additional glenoid retroversion conditions were modeled to investigate the effect of glenoid retroversion (n = 3). S1 and S2 had been implanted and experimentally digitized in 20° glenoid retroversion, and S3 was implanted and digitized in 10° retroversion. Four glenoid retroversion angles were modeled: 0°, 10°, 20°, and 30°. Additional retroversion angles other than the angle physically created on the specimens were modeled by rotating the glenoid about the humeral head center in the transverse plane to a new retroversion angle. The musculotendinous scapular wrap points were translated the same as the glenoid center translation associated with a new retroversion angle. Wilcoxon signed ranks tests were used to compare the variables between the two offset positions in each sample at each retroversion angle. In order to evaluate the effect of glenoid retroversion angles on the variables, the data were compared between the 4 glenoid
8
Humeral head anterior offsetting shoulder arthroplasty retroversion angles using Friedman test each specimen at each offset position. Statistical significance was set at p = 0.05. The data are shown as mean ± standard deviation.
RESULTS Part I Significant differences between anatomic and anterior offset positions were detected in the center of pressure locations in both anteroposterior and superoinferior directions, and in the humeral head displacement in the anteroposterior direction (* in Fig. 3).
The mean
anteroposterior coordinate of the center of pressure was -6.1 ± 0.8 mm and -4.7 ± 0.7 mm for the anatomic and anterior offset positions, respectively (p = 0.028), indicating that the anterior offset position shifted the center of pressure anteriorly upon muscle loading. The mean superoinferior coordinate of the center of pressure was -2.5 ± 2.2 mm and -3.5 ± 2.4 mm for the anatomic and anterior offset positions, respectively (p = 0.028). The mean anteroposterior displacement of the humeral head was -0.05 ± 0.15 mm and 0.09 ± 0.18 mm for the anatomic and anterior offset positions, respectively (p = 0.028), indicating that the anterior offset position resulted in a more anterior translation of the humeral head upon muscle loading. No significant differences in contact area, total force, and peak pressure were detected comparing the anatomic and anterior offset humeral head component positions (p > 0.05) (Fig. 3). The mean contact area was 153 ± 51 mm2 and 190 ± 63 mm2 for the anatomic and anterior offset positions, respectively (p = 0.12). The mean total force was 71.0 ± 4.6 N and 72.2 ± 5.2 N for the anatomic and anterior offset positions, respectively (p = 0.75). The mean peak pressure was 2.9 ± 1.0 MPa and 2.5 ± 1.1 MPa for the anatomic and anterior offset positions, respectively (p = 0.17).
9
Humeral head anterior offsetting shoulder arthroplasty For the anatomical offset the resultant force vector acting on the humerus from all muscles (at the start of the simulation) was angulated anteromedially an average of 7.7º ± 2.6º away from the scapula body plane in the transverse plane, and for the anterior offset this angle averaged 16.3º ± 2.3º indicating a more anteriorly directed component of muscle force. Tests of result sensitivity to muscle force magnitude in models of one specimen showed that, as expected, contact area, total force, and peak pressure were sensitive to muscle force magnitude (Fig. S-1). Superoinferior coordinate of the center of pressure was sensitive to muscle force; however anteroposterior coordinate of the center of pressure was not very sensitive, changing from -6.0 mm to -6.2 mm for the anatomic offset position with 15 and 50 N muscle forces, respectively, and changing from -3.7 mm to -3.8 mm for the anterior offset position with 15 and 50 N muscle forces, respectively. Part II The results from the additional glenoid retroversion angles modeled for the three 30° abduction specimens revealed fairly consistent trends, but the models for specimen #3 (S3) showed aberrant results at 30° glenoid retroversion (Fig. 4 and S-2) . Paired comparisons between the anatomic and anterior offset positions at each glenoid retroversion angle (Wilcoxon signed ranks test, n=3) showed no significant difference in any of the variables (p > 0.05) although there was a trend that anterior offsetting was associated with increased contact area, decreased peak pressure, anterior shifting of center of pressure, and anterior shifting of humeral head displacement (Fig. 4 and S-2).
Repeated comparisons between 4 different glenoid
retroversion angles (Friedman test) showed that all variables were significantly different between the retroversion angles (p < 0.05) († and ‡ in Fig. 4 and S-2) except total force (p > 0.05). The distribution of contact pressures on the glenoid component is shown in Figure 5. There was a 10
Humeral head anterior offsetting shoulder arthroplasty trend that the magnitude of changes associated with anterior offsetting compared to anatomic offsetting in contact area, center of pressure, and peak pressure increased with increasing glenoid retroversion (Fig. S-3), but statistical tests were not performed on these data due to the small sample size and unknown normality of the data distribution.
DISCUSSION The biomechanical benefits of anterior offsetting of the humeral component have been previously demonstrated in posteriorly unstable shoulder arthroplasty simulated on cadaveric shoulders with artificially created glenoid retroversion (Kim et al., 2016). It was found that the force and energy required to displace the humeral head posteriorly increased significantly with the anterior offset position compared to the anatomic offset position. The joint contact pressures were significantly shifted anteriorly, and the joint contact area significantly increased with the anterior offset position. The study concluded that anterior offsetting of the humeral head component increased the resistance to posterior humeral head translation, shifted joint contact pressures anteriorly, and increased joint contact area, thus, potentially increasing the joint stability. The present study was performed to investigate this finding in various degrees of glenoid retroversion with use of 3-dimensional finite element models produced based off cadaveric shoulder specimens. The present study found that anterior offsetting of the humeral component resulted in a significant anterior shift of the joint contact pressure and humeral head position upon muscle loading. This finding is in agreement with the results of the previous ex vivo biomechanical study (Kim et al., 2016) and supports the first study hypothesis. Although its exact clinical implications are unknown at this point, this finding together with the observation of successful
11
Humeral head anterior offsetting shoulder arthroplasty early clinical results suggests that the anterior offsetting technique may contribute to the joint stability in posteriorly unstable total shoulder arthroplasty due to increased glenoid retroversion and may reduce the eccentric loading on glenoid components implanted in retroversion (Kim et al., 2016). This study also demonstrated that the anterior offsetting led to a more anterior orientation of the resultant muscle force vector acting on the humeral head, providing a potential explanation for the mechanism underlying the resulting changes in shoulder biomechanics. The mean shift of the anteroposterior joint contact pressure was similar in the models (1.37 mm) and corresponding specimen experiments (1.35 mm). However superoinferior mean shifts were not similar, and no significant correlations were detected for pairwise comparison between model and experimental data on a specimen-specific basis (p>.05). It has been previously shown (Kim et al., 2016) that the biomechanical benefits of anterior offsetting were more prominent in glenoids with more severe retroversion (10° vs. 20° retroversion). This apparent incremental effect of anterior offsetting with increasing glenoid retroversion was tested in the present study as the second hypothesis. It was found that all three simulated samples showed a trend for an incremental effect in contact area, peak pressure, and anteroposterior location of center of pressure with increasing glenoid retroversion except for one sample (S3) at 30° retroversion. This trend failed to reach the statistical significance possibly due to the small sample size of the study and the aberrant behavior of the one sample at 30° retroversion. Our modeling focused on the shoulder stabilization in the midrange of motion, in which the passive capsule and ligamentous tissues are more lax, and stability is obtained by compression of the humeral head into the concave glenoid by active muscle contraction (Lippitt et al., 1993). Our focus on smaller loads applied to the rotator cuff and deltoid is based on
12
Humeral head anterior offsetting shoulder arthroplasty Nyffeler et al.’s (Nyffeler et al., 2006) biomechanical study of glenohumeral stability following shoulder replacement, and our own recent cadaver study (Kim et al., 2016). According to our sensitivity testing, the important anteroposterior coordinate of the center of pressure was not very sensitive to changes in muscle force magnitude, although other results were sensitive to these changes. Buchler and Farron (Buchler et al., 2004) used a finite element model to assess the effects of humeral head geometry on replaced shoulder biomechanics; rotator cuff muscles were included but the deltoid was not. Models of active glenohumeral stability with substantially greater complexity have been reported, such as Viehofer et al. (Viehofer et al., 2016) who modeled 27 individual muscle segments and muscle forces estimated by optimization. However simplified loading schemes have been used in experimental and computational studies of effects of implant design on stability. For instance, Hopkins et al. (Hopkins et al., 2007) used finite element modeling to investigate the effects of implant curvatures on glenohumeral stability. In that study the authors applied an assumed resultant compressive and shear load to the humeral head, instead of applying individual muscle loads. This modeling study had several limitations. Models were constructed from digitization of non-arthritic cadaver specimens implanted with arthroplasty components.
Physical
digitization in cadaver specimens (for which we also had experimental biomechanical measurements) was used because of the ability to directly visualize the musculotendon soft tissues. Image-based modeling from CT or MRI would be advantageous for future work in modeling arthritic patients.
Muscle forces were simplified as linear forces of a constant
magnitude acting at single points on the bones. This simplification was consistent between the humeral head offset conditions, and assumed that more realistic muscles and muscle attachments would not substantially affect the relative effects of humeral head offset condition on joint
13
Humeral head anterior offsetting shoulder arthroplasty kinematics and contact pressures.
It has been demonstrated for other anatomical sites that
muscle modeling and boundary condition assumptions can impact results (Phillips et al., 2007; Speirs et al., 2007) Our omission of muscles such as latissimus dorsi is supported by a recent complex model of active glenohumeral stability (Favre et al., 2012). That study also showed force contributions from the long head biceps and pectoralis major at specific arm elevation angles, although contributions from the deltoid and rotator cuff were more substantial.
In some
cases, there was passage of the muscle force line through a small portion of the humeral head, indicating in reality that the musculotendinous tissue wrapped around the component and wrap points should be implemented in further modeling efforts.
Having muscles crossing the
components is not physiological and will change their moment arm and, therefore, their contribution to the biomechanics of the shoulder joint. Dynamic shoulder movements were not modeled. Scapulothoracic motion (Braman et al., 2009) and its effects on muscle forces crossing the glenohumeral joint were ignored. Other soft tissues surrounding the glenohumeral joint, especially the joint capsule, were not taken into account. Multiple glenoid retroversion angles were modeled with hypothetical shifts in muscle wrap point locations instead of digitization of additional specimens implanted with those retroversion angles. The glenoid component was fixed at its medial surface instead of modeling the bone and cement. Cement and bone have been neglected previously in modeling (Oosterom et al., 2003) and experimental (Nyffeler et al., 2006) studies of glenohumeral stability, and the approach is supported by our modeling (Wee et al., 2015) and experimental (Lewis et al., 2016) observations of motions at the component medial border typically less than 0.05 mm relative to the fixed glenoid bone (Fig. S-4). However, large retroversion angles can potentially have impacts on implant micromotions relative to the cement and bone (Farron et al., 2006). Due to the sample size and unknown distribution of the data,
14
Humeral head anterior offsetting shoulder arthroplasty nonparametric tests were used throughout the study. As nonparametric tests are generally regarded less powerful tests than parametric tests in detecting true statistical significance, the overall impact of exclusive usage of nonparametric tests on the study results is unknown. Another potential limitation is that the present study used glenoid components that had the same radius of curvature as the humeral heads, which might have resulted in a higher degree of constraint and a lower degree of contact than the constructs with a radial mismatch. However, the overall study results would not have been affected significantly because the same construct was used across the groups and specimens in the present study. It should be noted that the anatomic variations of the individual cadaveric specimens were not considered in this study. Although the exact impact of these anatomic variations is unknown, it is unlikely that the study results were affected significantly as all the specimens were prepared in the same way for the implantation, and statistical comparisons were conducted either within a single specimen or within a pair that had almost identical anatomy. Lastly, the effects of varying degrees of anterior offset were not investigated in this study due to the practical difficulty in estimating the locations of the rotator cuff insertion sites upon offset changes. Only one anterior offset was used for each sample in this study, which was the exact opposite of the anatomic offset of each sample.
ACKNOWLEDGEMENT Funding for related experiments provided by Orthopaedic Research and Education Foundation, who were not involved further in the study and manuscript.
CONFLICT OF INTEREST STATEMENT
15
Humeral head anterior offsetting shoulder arthroplasty H.M.K. received materials from Zimmer to support this research. G.S.L. has received materials from Depuy-Synthes for unrelated research.
16
Humeral head anterior offsetting shoulder arthroplasty REFERENCES
Boileau, P., Walch, G. 1997. The three-dimensional geometry of the proximal humerus. Implications for surgical technique and prosthetic design. The Journal of bone and joint surgery British volume. 79, 857-65. Braman, J.P., Engel, S.C., Laprade, R.F., Ludewig, P.M. 2009. In vivo assessment of scapulohumeral rhythm during unconstrained overhead reaching in asymptomatic subjects. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 18, 960-7. Buchler, P., Farron, A. 2004. Benefits of an anatomical reconstruction of the humeral head during shoulder arthroplasty: a finite element analysis. Clinical biomechanics. 19, 16-23. Denard, P.J., Walch, G. 2013. Current concepts in the surgical management of primary glenohumeral arthritis with a biconcave glenoid. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 22, 1589-98. Farron, A., Terrier, A., Buchler, P. 2006. Risks of loosening of a prosthetic glenoid implanted in retroversion. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 15, 521-6. Favre, P., Sentelera, M., Hippa, J., Scherrera, S., Gerbera, C., Snedeker, J. 2012. An integrated model of active glenohumeral stability. J Biomechanics. 45, 2248-55. Hill, J.M., Norris, T.R. 2001. Long-term results of total shoulder arthroplasty following bonegrafting of the glenoid. The Journal of bone and joint surgery American volume. 83-A, 877-83. Hopkins, A.R., Hansen, U.N., Amis, A.A., Taylor, M., Emery, R.J. 2007. Glenohumeral kinematics following total shoulder arthroplasty: a finite element investigation. Journal of 17
Humeral head anterior offsetting shoulder arthroplasty orthopaedic research : official publication of the Orthopaedic Research Society. 25, 10815. Hopkins, A.R., Hansen, U.N., Amis, A.A., Taylor, M., Gronau, N., Anglin, C. 2006. Finite element modelling of glenohumeral kinematics following total shoulder arthroplasty. Journal of biomechanics. 39, 2476-83. Iannotti, J.P., Norris, T.R. 2003. Influence of preoperative factors on outcome of shoulder arthroplasty for glenohumeral osteoarthritis. The Journal of bone and joint surgery American volume. 85-A, 251-8. Kim, H.M., Chacon, A.C., Andrews, S.H., Roush, E.P., Cho, E., Conaway, W.K., Kunselman, A.R., Lewis, G.S. 2016. Biomechanical benefits of anterior offsetting of humeral head component in posteriorly unstable total shoulder arthroplasty: A cadaveric study. Journal of orthopaedic research : official publication of the Orthopaedic Research Society. 34, 666-74. Lacroix, D., Prendergast, P.J. 1997. Stress analysis of glenoid component designs for shoulder arthroplasty. Proceedings of the Institution of Mechanical Engineers Part H, Journal of engineering in medicine. 211, 467-74. Lewis, G.S., Brenza, J.B., Paul, E.M., Armstrong, A.D. 2016. Construct damage and loosening around glenoid implants: A longitudinal micro-CT study of five cadaver specimens. Journal of orthopaedic research : official publication of the Orthopaedic Research Society. 34, 1053-60. Lippitt, S., Matsen, F. 1993. Mechanisms of glenohumeral joint stability. Clinical orthopaedics and related research. 20-8.
18
Humeral head anterior offsetting shoulder arthroplasty Neer, C.S., 2nd, Morrison, D.S. 1988. Glenoid bone-grafting in total shoulder arthroplasty. The Journal of bone and joint surgery American volume. 70, 1154-62. Nowak, D.D., Bahu, M.J., Gardner, T.R., Dyrszka, M.D., Levine, W.N., Bigliani, L.U., Ahmad, C.S. 2009. Simulation of surgical glenoid resurfacing using three-dimensional computed tomography of the arthritic glenohumeral joint: the amount of glenoid retroversion that can be corrected. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 18, 680-8. Nyffeler, R.W., Sheikh, R., Atkinson, T.S., Jacob, H.A., Favre, P., Gerber, C. 2006. Effects of glenoid component version on humeral head displacement and joint reaction forces: an experimental study. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 15, 625-9. Oosterom, R., Herder, J.L., van der Helm, F.C., Swieszkowski, W., Bersee, H.E. 2003. Translational stiffness of the replaced shoulder joint. Journal of biomechanics. 36, 1897907. Phillips, A.T., Pankaj, P., Howie, C.R., Usmani, A.S., Simpson, A.H. 2007. Finite element modelling of the pelvis: inclusion of muscular and ligamentous boundary conditions. Medical engineering & physics. 29, 739-48. Rice, R.S., Sperling, J.W., Miletti, J., Schleck, C., Cofield, R.H. 2008. Augmented glenoid component for bone deficiency in shoulder arthroplasty. Clinical orthopaedics and related research. 466, 579-83. Robertson, D.D., Yuan, J., Bigliani, L.U., Flatow, E.L., Yamaguchi, K. 2000. Three-dimensional analysis of the proximal part of the humerus: relevance to arthroplasty. The Journal of bone and joint surgery American volume. 82-A, 1594-602.
19
Humeral head anterior offsetting shoulder arthroplasty Speirs, A.D., Heller, M.O., Duda, G.N., Taylor, W.R. 2007. Physiologically based boundary conditions in finite element modelling. Journal of biomechanics. 40, 2318-23. Spencer, E.E., Jr., Valdevit, A., Kambic, H., Brems, J.J., Iannotti, J.P. 2005. The effect of humeral component anteversion on shoulder stability with glenoid component retroversion. The Journal of bone and joint surgery American volume. 87, 808-14. Steinmann, S.P., Cofield, R.H. 2000. Bone grafting for glenoid deficiency in total shoulder replacement. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 9, 361-7. Viehofer, A.F., Snedeker, J.G., Baumgartner, D., Gerber, C. 2016. Glenohumeral joint reaction forces increase with critical shoulder angles representative of osteoarthritis-A biomechanical analysis. Journal of orthopaedic research : official publication of the Orthopaedic Research Society. 34, 1047-52. Walch, G., Moraga, C., Young, A., Castellanos-Rosas, J. 2012. Results of anatomic nonconstrained prosthesis in primary osteoarthritis with biconcave glenoid. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 21, 152633. Wee, H., Armstrong, A.D., Flint, W.W., Kunselman, A.R., Lewis, G.S. 2015. Peri-implant stress correlates with bone and cement morphology: Micro-FE modeling of implanted cadaveric glenoids. Journal of orthopaedic research : official publication of the Orthopaedic Research Society. 33, 1671-9. Winter, D.A. Biomechanics and Motor Control of Human Movement. Hoboken, NJ: John Wiley & Sons; 2005.
20
Humeral head anterior offsetting shoulder arthroplasty Yongpravat, C., Kim, H.M., Gardner, T.R., Bigliani, L.U., Levine, W.N., Ahmad, C.S. 2013. Glenoid implant orientation and cement failure in total shoulder arthroplasty: a finite element analysis. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 22, 940-7. Zhang, J., Yongpravat, C., Kim, H.M., Levine, W.N., Bigliani, L.U., Gardner, T.R., Ahmad, C.S. 2013. Glenoid articular conformity affects stress distributions in total shoulder arthroplasty. Journal of shoulder and elbow surgery / American Shoulder and Elbow Surgeons [et al]. 22, 350-6.
21
Humeral head anterior offsetting shoulder arthroplasty FIGURE LEGENDS Figure 1. Cadaveric shoulder specimen setup on a testing frame. Each rotator cuff musculotendinous unit was isolated from the scapula and joint capsule. They were then fixed with nylon anchors medially to facilitate loading. The deltoid was approximated with two inflexible nylon ropes running from eyelet screws inserted into the deltoid tuberosity through two eyelet screws fixed on the acromion and coracoid process. The scapular body was fixed with a clamp in a neutral rotation in the coronal and horizontal planes, and 20˚ downward-tilted position in the sagittal plane with the glenoid fossa perpendicular to the floor. The distal end of the humerus was held by a ring stand. Hanging weights, cables, and pulleys were used to apply 15N forces independently to each musculotendinous unit (with two 15N forces applied in total to the two deltoid ropes).
Figure 2. Axial (left) and anteroposterior (right) views of the finite element model of a specimen depicting the humeral head and glenoid components, initial lines of action of the six musculotendinous units (red lines), fixed scapular wrap points (blue triangles), moving humeral attachment points (green circles), long axis of the humerus (black line), and fixed glenoid constraint (light blue box). Meshed humeral head and glenoid components are also shown at the bottom. Ant = anterior, Post = posterior, Sup = superior, Inf = inferior.
Figure 3. Comparisons between the anatomic and anterior offset positions of the humeral head components constructed from experimentally digitized cadaveric shoulder specimens (n = 6), each with glenoid retroversion of either 10° or 20°. * p < 0.05 for Wilcoxon signed ranks tests
22
Humeral head anterior offsetting shoulder arthroplasty comparing between anatomic and anterior offset positions.
Error bars are +/- 1 standard
deviation.
Figure 4. Results from the additional glenoid retroversion angles modeled for the three 30° abduction specimens. † = p < 0.05 for Friedman test comparing between retroversion angles at anatomic offset position. ‡ = p < 0.05 for Friedman test comparing between retroversion angles at anterior offset position. A-P = anteroposterior.
Figure 5.
Predicted glenoid contact pressure distributions for model of specimen S2, for
anatomic and anterior humeral head offset and four glenoid retroversion angles. Center of contact pressure is shown with the red plus sign. Note the increased anterior shift associated with the anterior humeral head offset for larger glenoid retroversion angles. Anterior is to the right, posterior to the left.
Supplementary Data Figure S-1. Results of sensitivity analysis of change in muscle force magnitudes from 15 N to 50 N in one randomly selected specimen model.
Figure S-2. Results from the additional glenoid retroversion angles modeled for the three 30° abduction specimens. † = p < 0.05 for Friedman test comparing between retroversion angles at
23
Humeral head anterior offsetting shoulder arthroplasty anatomic offset position. ‡ = p < 0.05 for Friedman test comparing between retroversion angles at anterior offset position. S-I = superoinferior.
Figure S-3. Changes in study variables associated with humeral head anterior offsetting vs. glenoid retroversion angle. Each of the three 30° abduction specimens (S1, S2, and S3) is plotted, and each plotted data point represents the result for anterior offset position value minus the result for anatomic offset position.
24
Humeral head anterior offsetting shoulder arthroplasty
25
Humeral head anterior offsetting shoulder arthroplasty
26
Humeral head anterior offsetting shoulder arthroplasty
27