- Email: [email protected]
The effect of radial head implant shape on radiocapitellar kinematics during in vitro forearm rotation
J Shoulder Elbow Surg (2015) 24, 258-264
www.elsevier.com/locate/ymse
The effect of radial head implant shape on radiocapitellar kinematics during in vitro forearm rotation Hannah L. Shannon, MESc, Simon R. Deluce, MESc, Joshua W. Giles, PhD, James A. Johnson, PhD, Graham J.W. King, MD, MSc, FRCSC* Hand and Upper Limb Centre Bioengineering Laboratory, St. Joseph’s Health Care, Western University, London, ON, Canada Background: A number of radial head implants are in clinical use for the management of radial head fractures and their sequelae. However, the optimal shape of a radial head implant to ensure proper tracking relative to the capitellum has not been established. This in vitro biomechanical study compared radiocapitellar joint kinematics for 3 radial head implant designs as well as the native head. Methods: Eight cadaveric upper extremities were tested using a forearm rotation simulator with the elbow at 90 of flexion. Motion of the radius relative to the capitellum was optically tracked. A stem was navigated into a predetermined location and cemented in place. Three unipolar implant shapes were tested: axisymmetric, reverse-engineered patient-specific, and population-based quasi-anatomic. The patientspecific and quasi-anatomic implants were derived from measurements performed on computed tomography models. Results: Medial-lateral and anterior-posterior translation of the radial head with respect to the capitellum varied with forearm rotation and radial head condition. A significant difference in medial-lateral (P ¼ .03) and anterior-posterior (P ¼ .03) translation was found between the native radial head and the 3 implants. No differences were observed among the radial head conditions except for a difference in medial-lateral translation between the axisymmetric and patient-specific implants (P ¼ .04). Conclusions: Radiocapitellar kinematics of the tested radial head implants were similar in all but one comparison, and all had different kinematics from the native radial head. Patient-specific radial head implants did not prove advantageous relative to conventional implant designs. The shape of the fixed stem unipolar radial head implants had little influence on radiocapitellar kinematics when optimally positioned in this testing model. Level of evidence: Basic Science, Biomechanics. Ó 2015 Journal of Shoulder and Elbow Surgery Board of Trustees. Keywords: Arthroplasty; radial head; kinematics; implant design; patient specific; in vitro
No specific Institutional Review Board approval was required for this study. *Reprint requests: Graham J.W. King, MD, FRCSC, Hand and Upper Limb Centre, St. Joseph’s Health Care, 268 Grosvenor Street, Rm D0-202, London, ON N6A 4L6, Canada. E-mail address: [email protected] (G.J.W. King).
Approximately 33% of elbow fractures involve the radial head.13,15 This structure is an important stabilizer of the elbow, especially in the case of associated ligamentous injuries.15 In the setting of comminuted, displaced, unreconstructable radial head fractures associated with ligament
1058-2746/$ - see front matter Ó 2015 Journal of Shoulder and Elbow Surgery Board of Trustees. http://dx.doi.org/10.1016/j.jse.2014.09.019
Radial implant shape and radiocapitellar kinematics injuries, replacing the radial head with an implant to maintain joint stability is essential.1 A number of commercially available radial head implants have been developed and are in common clinical use for the management of acute radial head fractures and late reconstruction for nonunions and malunions. Currently, most of these implants have a head that is rotationally symmetric around the stem; also termed axisymmetric.2 However, anatomic studies have shown that the radial head is elliptical and that the articulating dish is located eccentrically with respect to the neck.11,12,18 The complex shape of the native radial head poses a question of whether an axisymmetric implant with a fixed stem can adequately replicate normal radiocapitellar joint kinematics during forearm rotation. During this motion, the dish of the radial head rotates about the capitellum, and the margin of the radial head rotates within the radial notch of the ulna. If the native radial head is elliptical and the dish is eccentric, it is possible that a fixed-stem axisymmetric unipolar implant may exhibit abnormal kinematics that may cause pain, stem loosening, and lead to clinical failure. More anatomic implant designs have been developed, but their effectiveness relative to more traditional axisymmetric implants during forearm motion is unknown.3 Furthermore, the concept of a reverse-engineered patient-specific radial head implant based on imaging of the contralateral uninjured elbow is of interest because, theoretically, this should most closely replicate the native radial head and, hence, should reproduce normal kinematics if positioned correctly. In view of the foregoing, the objective of this study was to compare the radiocapitellar kinematics of the native articulation with axisymmetric, population-based quasianatomic, and reverse-engineered patient-specific unipolar implant designs. We hypothesized that more anatomically shaped radial head implants would have similar radiocapitellar kinematics to the native radial heads and that the axisymmetric radial head implants would differ.
Materials and methods Specimen preparation Eight fresh frozen cadaveric upper extremities from male donors (age, 75 8 years) were thawed overnight at room temperature before testing. The specimens were prepared by suturing the biceps and triceps tendons using nylon braided string and the pronator teres with No. 5 Ethibond polyester braided suture (Ethicon Inc/Johnson and Johnson, Somerville, NJ, USA). These sutures were attached to servomotors to achieve active forearm rotation, as described below. The wrist was fixed in neutral flexion by passing a pin through the long finger metacarpal into the distal radius. The surgical incisions were closed using 2-0 Vicryl suture (Ethicon Inc/Johnson and Johnson). The specimens were kept at room temperature and were hydrated with normal saline throughout testing.
259
Implant design The unipolar radial head implant system created for this study consisted of a cemented generic stem that could be used with all implants and the implant head.4 Three radial head implants were developed for assessment: (1) an axisymmetric, (2) a populationbased quasi-anatomic, and (3) a reverse-engineered patient-specific implant. Computed tomography scans of 34 male upper extremities were used to design a series of 3 sizes of population-based quasi-anatomic implants. Reverse-engineered patient-specific implants were also generated for the 8 specimens used in the current study. Mimics image processing software (Materialize NV, Leuven, Belgium) was used to generate a preoperative 3-dimensional (3D) model of the radius. Implant parameters were measured by ellipse fitting the surface model cross sections using custom software based on the Visualization Toolkit (VTK). The quasi-anatomic design used averages derived from these measured parameters, whereas the patient-specific model used each specimen’s unique parameters.4 The shape of the axisymmetric radial head was modeled after the EVOLVE Proline Radial Head System (Wright Medical Technology Inc, Arlington, TN, USA). Unlike the commercially available implant, which incorporates a cylindrical polished stem, the axisymmetric implant was modified to fit on the same custommade stem common to all implant shapes. All radial head implants were formed out of ABS M30 plastic (Fig. 1) using a rapid prototyping machine with an accuracy of 0.127 mm (Stratasys Fortus 400 MC, Eden Prairie, MN, USA).
Testing apparatus The arm was mounted into a custom upper extremity motion simulator that was previously developed in our laboratory and has been validated through the peer-review process6 (Fig. 2). The biceps, triceps, and pronator teres sutures were attached to servomotors to allow for computer-controlled movement. The triceps was activated to maintain 90 of flexion by resting the forearm on a support bar. The biceps and pronator teres were activated to achieve active forearm supination from the pronated position during each stage of testing while the kinematics were measured with Optotrak Certus optical tracking system (NDI, Waterloo, ON, Canada).
Testing protocol After the intact specimen was tested, the lateral collateral ligament (LCL) was sectioned from the lateral epicondyle to provide access to the radial head and subsequently repaired using a braided No. 2 HiFi ultra-high-weight polyethylene suture (CONMED Linvatec, Largo, FL, USA). The suture ends were passed into a hole placed at the isometric point on the lateral epicondyle and out through 2 tunnels located proximal to the site of entry. This allowed the ligament to be repaired such that the original line of action was restored.8 The annular ligament was kept intact. A pneumatic actuator was used to apply a load of 20 N to the LCL before the ligament cable was clamped during testing. Because repeated access to the radiocapitellar joint was required for placement of the implants, the kinematics of the LCL repair model with the native radial head was compared with the intact state to confirm that it could be used as a control.
260
H.L. Shannon et al.
Figure 1 Examples of the radial head implants and stem: (A) reverse-engineered patient-specific, (B) population-based quasi-anatomic, (C) axisymmetric, and (D) custom-made generic stem.
The 3D optical trackers were fixed to the radius and the navigation tool with the implant stem. Feedback to the surgeon was given in real-time to allow corrections of the implant location using a custom LabVIEW program (National Instruments, Austin, TX, USA).5 Once the target location was achieved, the stem was cemented into place using Surgical Simplex bone cement (Stryker, Mahwah, NJ, USA).
Kinematics measurements
Figure 2 ligament.
The forearm motion simulator. LCL, lateral collateral
The stem was cemented into the canal of the radius such that the radial heads would be in an optimal location (Fig. 3). The radial head implants were tested in a random order.
Registration, navigation, and implantation Computed tomography scans were performed on all of the specimens tested, and 3D surface models of the radii were generated as described above. A custom program was used to determine an ideal location for the implant stem based on measurements of the native radius.5 The radial head was then removed using an oscillating saw, and the canal was reamed. The minor diameter of the native radial head was used to select the diameter of axisymmetric implant. The implant height was defined by the commercially available implant system on which the custom implant was modelled.
To track the motion of the forearm, optical tracker mounts were fixed to the radius and ulna directly on the bone surface using custom-made mounts and cortical bone screws. The trackers were attached to the ulnar mount and the humerus. A ring containing 6 optical marker sets that were calibrated to perform as one rigid body was attached to the radius. This was done to maintain visualization of the radius by the tracking system throughout forearm rotation. Kinematic data were analyzed using custom LabVIEW software to quantify the motion of the radius with respect to the humerus during active forearm supination. The humeral and radial coordinate systems were defined using a series of anatomic digitizations.7 For the humerus, the capitellum surface and the trochlear sulcus were traced, and a point was digitized in the center of the humeral shaft at the mid-diaphysis. The capitellum was sphere fit, the trochlear sulcus was circle fit, and their centers were calculated. A vector was made between these 2 center points in the medial direction (the Zhum þ axis). The bisector of these 2 center points was found, and a vector was made from this bisector to the shaft point (long axis). The Yhum þ axis was the cross product of the Zhum þ axis and the long axis. The Xhum þ axis was the cross product of the Yhum þ axis and the Zhum þ axis. The origin of the humeral coordinate system was the center of the capitellum, such that radial head translation about the capitellum during rotation could be measured (Fig. 4). To produce the radial coordinate system, 10 points around the rim of the radial head were digitized in addition to the radial styloid and the dorsal and volar aspects of the lesser sigmoid notch of the distal radius. The 10 rim points were circle fit,and the center was found. The bisector of the dorsal and volar aspects of the
Radial implant shape and radiocapitellar kinematics
261
Figure 3 This flow chart depicts the phases of the testing protocol. Radiocapitellar kinematics were measured during simulated active forearm supination. LCL, lateral collateral ligament; RH, radial head. lesser sigmoid notch was found, and a vector was made from this bisector to the radial styloid point (medial vector). The bisector of this vector was found, and a vector from this bisector to the center of the radial head was made (the Xrad þ axis). The Yrad þ axis was the cross product of the Xrad þ axis and the medial vector. The Zrad þ axis was the cross product of the Xrad þ axis and the Yrad þ axis. The origin of the radial coordinate system was located at the center of the radial head (Fig. 4, B). The outcome variables examined were medial-lateral (ML) translation and anteriorposterior (AP) translation of the center of the radial head with respect to the center of the capitellum (Fig. 4, C and D).
Statistical analyses A 2-way repeated-measures analysis of variance was performed with radial head condition and forearm rotation angle as the independent variables to determine if there were differences between the intact state and the LCL repair (the 2 levels of the condition variable) for both ML and AP translations of the radial head. This was done to determine whether the LCL repair condition could be used as a control. A 2-way repeated measures analysis of variance was also performed with radial head condition and rotation angle as the independent variables to determine if there were differences in kinematics between the native radial head and the 3 implant shapes during forearm rotation. A post hoc pairwise comparison (a ¼ 0.05) was used to determine if there were differences between each implant and the native radial head.
This same technique was used to compare the 3 implant morphologies with each other. SPSS software (IBM, Armonk, NY, USA) was used for all analyses.
Results Owing to the use of cadaveric specimens from elderly donors and the need to compare motion using a repeatedmeasures design, kinematic data were truncated to include an arc from 40 of pronation to 50 of supination, which could be achieved in all specimens. There was no significant difference (P > .05) when ML or AP translations for the intact elbows were compared with the native radial head and after LCL sectioning and repair. For all cases, the 2 conditions were within 1-mm translation of each other. Thus, the LCL sectioned and repaired condition with the native radial head was used as the control for comparison with the radial head implants. Significant differences in ML translation were found among the radial head morphologies (P ¼ .03); however, pairwise comparisons only showed a significant difference between the axisymmetric and patient-specific implants (P ¼ .04; Fig. 5). As the forearm moved from pronation to supination, the radial head translated medially (P ¼ .04). There was a significant effect of radial head morphology on
262
H.L. Shannon et al.
Figure 4 Anterior-posterior (AP) and medial-lateral (ML) translations of the radial head with respect to the center of the capitellum were examined. (A) Anterior view of the humerus shows the humeral coordinate system with the center of the capitellum as the origin. (B) Proximal view of the radial head shows the radial coordinate system with the center of the radial head as the origin. (C) Direction of ML translation of the radial head is shown relative to the capitellum, and (D) direction of AP translation of the radial head is shown relative to the capitellum.
Figure 5 This graph displays the mean data for active radial head (RH) translations in the medial-lateral (ML) direction (n ¼ 8). A significant difference in ML translation was found between the native RH and the 3 different implant shapes (P ¼ .03). Pairwise comparisons did not demonstrate any differences between the 4 RH conditions, except for a significant difference in ML translation between the axisymmetric and patient-specific implant (P ¼ .04). The standard deviations, which ranged from 0.86 to 2.42 mm, were removed for clarity.
Figure 6 This graph displays the mean data for active radial head (RH) translations in the anterior-posterior (AP) direction (n ¼ 8). A significant difference in AP translation was found between the native RH and the 3 different implant shapes (P ¼ .03). The standard deviations, which ranged from 1.38 to 2.68 mm, were removed for clarity.
AP translation (P ¼ .03); however, pairwise comparisons failed to detect a difference between the native radial head and the 3 different implant shapes (P > .05; Fig. 6). AP translation did not change significantly during forearm rotation (P ¼ .41).
The effects of 3 fixed-stem unipolar radial head implant shapes (ie, the axisymmetric, population-based quasianatomic, and reverse-engineered patient-specific) on radiocapitellar kinematics during rotation were examined. Small but significant differences in kinematics were evident
Discussion
Radial implant shape and radiocapitellar kinematics among radial head morphologies during forearm rotation. The change in kinematics after radial head arthroplasty may be partly related to the small errors (<2 mm translation and <11 rotation) in placement of the stem despite image-based computer navigation.5 The errors measured in rotation would be expected to have a greater effect on the 2 anatomic implants than on the axisymmetric implant because the anatomic implants have off-centered dishes that rely on accurate placement to replicate native radial head kinematics. It is likely that up to 11 of malrotation altered the position of the dish of both the quasi-anatomic and patient-specific implants. This error in rotational orientation would not have had a significant effect on the axisymmetric implant because the dish is located centrally on this implant. The observation that the kinematics of the more anatomic implants and the axisymmetric implants were similar to the native articulation, despite rotational malpositioning, suggests other factors such as translational malpositioning or annular ligament and interosseous membrane tension may have a greater influence on radiocapitellar kinematics than implant shape. Further studies are needed to quantify the articular contact area, location, and pressure at the radiocapitellar joint and to further elucidate the influence of errors in implant positioning. The reverse-engineered patientspecific radial head implants used in the current study had kinematics that were similar to the more conventional offthe-shelf implant designs, suggesting that the increased costs associated with these custom devices may not be justified. The radial head has a complex elliptical shape. The articular dish is typically offset from the center of the head, and the head is offset from the center of the neck.11,12,18 Hence, changes in kinematics after radial head arthroplasty are expected unless the implant is perfectly positioned during surgery and the implant shape and size replicates the native anatomy. Proper tracking within the joint is essential to maintain the health of the articular cartilage. Abnormal contact between the radius and capitellum during elbow movement could lead to increased stress on parts of the articulation causing pain, premature cartilage, wear of the capitellum,17 and an early onset of osteoarthritis. This may lead to the need for removal of the radial head implant or revision to a unicompartmental radiocapitellar arthroplasty.10,16 The implant stem was cemented into the canal of the radius to ensure the quasi-anatomic and patient-specific radial head implants were both located in an optimal position. Because we used a custom stem to fit all implants, the axisymmetric implant was fixed in place as well. Although a fixed position is necessary for anatomically shaped implants and is used by some commercially available axisymmetric implants, other unipolar implants are designed to have a loose-fitting stem or have a bipolar articulation, such that the implant has some ability to selfalign the dish against the capitellum.2 During forearm motion, the radial heads of these implants may move
263 slightly with respect to the proximal radius such that they stay in optimal contact with the capitellum. Future studies should compare the radiocapitellar kinematics of fixedstem unipolar radial head implants with the kinematics of a loose-fitting stem or a bipolar articulation. The native radial head with an LCL repair was used as the control in this experiment so that repeated access to the radiocapitellar joint could be achieved. The LCL resection and repair technique described by Fraser et al.8 was successfully repeated in this study. By removing the LCL from the lateral epicondyle, the radial head could be accessed easily. The annular ligament, an important radial head stabilizer, was kept intact. By using strong braided sutures and tensioning the LCL repair to 20 N, the control closely replicated the radiocapitellar kinematics of the intact elbow during forearm rotation, as evidenced by the fact that no difference was observed in ML or AP measurements between intact and repaired testing groups. The kinematic pathways of the radial head with respect to the capitellum observed in the current study are similar to those reported by Galik et al.9 They found that the radial head translates an average of 2.1 mm in the AP direction and 1.6 mm in the ML direction during forearm rotation. Our data for mean AP and ML translations are within these ranges. In our study we kept the annular ligament intact, unlike in the study by Galik et al,9 where the annular ligament was sectioned. This important elbow stabilizer, as well as the concavity compression of the curved articular dish of the radial head with respect to the spherical capitellum, likely helped to control the motions of the native radial head and the various implants tested. The current study has several limitations. We used cadaveric specimens from elderly donors, whereas most patients undergoing radial head arthroplasty tend to be younger. The motion simulator may not precisely replicate muscle activation and forearm kinematics in vivo; however, the repeated-measures experimental design and precise kinematic measurements using an optical tracking system made it possible to quantify significant differences between radial head morphologies. A post hoc power analysis showed that our sample size of 8 specimens was sufficient to determine these differences. We only studied forearm rotation kinematics at 90 of flexion and did not quantify kinematics during elbow motion. Owing to the limitation of testing with fresh cadaveric specimens, we chose to focus on this position because it reflects the position of function for the upper extremity and thus is most often used in studies of radiocapitellar kinematics. Our computer and image-assisted implant positioning did not precisely replicate the target position of the implant stems as we had hoped. However, the error in positioning a radial head implant without navigation has not been quantified but is likely considerably higher than those achieved with the precise techniques used in the current study.14 Despite this degree of inaccuracy in implant positioning (<2 mm in translation and <11 in rotation), it
264 did not appear to affect the kinematics of the implants, which were similar to the native radial head. Further studies are needed to determine whether the favorable kinematics of the fixed-stem unipolar radial head implants used in this in vitro study can be achieved clinically with standard nonnavigated techniques of implant positioning.
Conclusions The kinematics of the radial head implants and the native articulation were similar except for a difference in ML translation between the axisymmetric and patientspecific implants. This study suggests that the kinematics of the radiocapitellar joint are not sensitive to radial head implant shape during simulated forearm rotation at 90 of flexion when using an optimally positioned fixed-stem unipolar device. The reverseengineered patient-specific radial head implants used in the current study had similar kinematics as the more conventional off-the-shelf implant designs.
Disclaimer Grants from the Natural Sciences and Engineering Research Council of Canada (grant no. 153063-2013) and the Canadian Institutes of Health Research (grant no. MOP-106500) supported the acquisition of cadaveric specimens, surgical supplies, imaging, and personnel required for this study. Graham King is a consultant and receives royalties from Tornier Inc. and Wright Medical Technology, which are bith related to the subject of this article. The other authors, their immediate families, and any research foundations with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.
References 1. Beingessner DM, Dunning CE, Gordon KD, Johnson JA, King GJ. The effect of radial head excision and arthroplasty on elbow kinematics and stability. J Bone Joint Surg Am 2004;86:1730-9. 2. Calfee R, Madom I, Weiss AP. Radial head arthroplasty. J Hand Surg Am 2006;31:314-21. http://dx.doi.org/10.1016/j.jhsa.2005.12.005.
H.L. Shannon et al. 3. Chanlalit C, Shukla DR, Fitzsimmons JS, An KN, O’Driscoll SW. Influence of prosthetic design on radiocapitellar concavitycompression stability. J Shoulder Elbow Surg 2011;20:885-90. http:// dx.doi.org/10.1016/j.jse.2011.03.009 4. Deluce SR, Shannon H, Lalone E, Ferreira LM, King GJ, Johnson JA. Comparison of population-based and patient-specific radial head implants using surface matching. Presented at the Orthopaedic Research Society 2013 Annual Meeting, San Antonio, TX; Jan 26-29, 2013. 5. Deluce SR, Shannon H, Lalone E, Ferreira LM, King GJ, Johnson JA. Development of a computer- and image-assisted guidance system for radial head arthroplasty. Presented at the Orthopaedic Research Society 2013 Annual Meeting, San Antonio, TX; Jan 26-29, 2013. 6. Ferreira LM, Johnson JA, King GJ. Development of an active elbow flexion simulator to evaluate joint kinematics with the humerus in the horizontal position. J Biomech 2010;43:2114-9. http://dx.doi.org/10. 1016/j.jbiomech.2010.04.007 7. Ferreira LM, King GJ, Johnson JA. Motion-derived coordinate systems reduce inter-subject variability of elbow flexion kinematics. J Orthop Res 2011;29:596-601. http://dx.doi.org/10.1002/jor.21278 8. Fraser GS, Pichora JE, Ferreira LM, Brownhill JR, Johnson JA, King GJ. Lateral collateral ligament repair restores the initial varus stability of the elbow: an in vitro biomechanical study. J Orthop Trauma 2008;22:615-23. http://dx.doi.org/10.1097/BOT.0b013e31818 86f37 9. Galik K, Baratz ME, Butler AL, Dougherty J, Cohen MS, Miller MC. The effect of the annular ligament on kinematics of the radial head. J Hand Surg Am 2007;32:1218-24. http://dx.doi.org/10.1016/j.jhsa. 2007.06.008. 10. Heijink A, Morrey BF, Cooney WP III. Radiocapitellar hemiarthroplasty for radiocapitellar arthritis: a report of three cases. J Shoulder Elbow Surg 2008;17:e12-5. http://dx.doi.org/10.1016/j.jse. 2007.04.009. 11. King GJ, Zarzour ZD, Patterson SD, Johnson JA. An anthropometric study of the radial head: implications in the design of a prosthesis. J Arthroplasty 2001;16:112-6. 12. Koslowsky TC, Germund I, Beyer F, Mader K, Krieglstein CF, Koebke J. Morphometric parameters of the radial head: an anatomical study. Surg Radiol Anat 2007;29:225-30. http://dx.doi.org/10.1007/ s00276-007-0197-1 13. Mason ML. Some observations on fractures of the head of the radius with a review of one hundred cases. Br J Surg 1954;42:123-32. 14. McDonald CP, Johnson JA, Peters TM, King GJ. Image-based navigation improves the positioning of the humeral component in total elbow arthroplasty. J Shoulder Elbow Surg 2010;19:533-43. http://dx. doi.org/10.1016/j.jse.2009.10.010 15. Morrey BF. The elbow and its disorders. Philadelphia, PA: Saunders Elsevier, ISBN 1416029028; 2008. p. 419. 16. Sabo MT, Shannon H, De Luce S, Lalone E, Ferreira LM, Johnson JA, et al. Elbow kinematics after radiocapitellar arthroplasty. J Hand Surg Am 2012;37:1024-32. http://dx.doi.org/10.1016/j.jhsa.2012.02.021 17. van Riet RP, Van Glabbeek F, Neale PG, Bimmel R, Bortier H, Morrey BF, et al. Anatomical considerations of the radius. Clin Anat 2004;17:564-9. http://dx.doi.org/10.1002/ca.10256 18. van Riet RP, Van Glabbeek F, Neale PG, Bortier H, An KN, O’Driscoll SW. The noncircular shape of the radial head. J Hand Surg Am 2003;28:972-8. http://dx.doi.org/10.1016/S0363-5023(03) 00426-X.
www.elsevier.com/locate/ymse
The effect of radial head implant shape on radiocapitellar kinematics during in vitro forearm rotation Hannah L. Shannon, MESc, Simon R. Deluce, MESc, Joshua W. Giles, PhD, James A. Johnson, PhD, Graham J.W. King, MD, MSc, FRCSC* Hand and Upper Limb Centre Bioengineering Laboratory, St. Joseph’s Health Care, Western University, London, ON, Canada Background: A number of radial head implants are in clinical use for the management of radial head fractures and their sequelae. However, the optimal shape of a radial head implant to ensure proper tracking relative to the capitellum has not been established. This in vitro biomechanical study compared radiocapitellar joint kinematics for 3 radial head implant designs as well as the native head. Methods: Eight cadaveric upper extremities were tested using a forearm rotation simulator with the elbow at 90 of flexion. Motion of the radius relative to the capitellum was optically tracked. A stem was navigated into a predetermined location and cemented in place. Three unipolar implant shapes were tested: axisymmetric, reverse-engineered patient-specific, and population-based quasi-anatomic. The patientspecific and quasi-anatomic implants were derived from measurements performed on computed tomography models. Results: Medial-lateral and anterior-posterior translation of the radial head with respect to the capitellum varied with forearm rotation and radial head condition. A significant difference in medial-lateral (P ¼ .03) and anterior-posterior (P ¼ .03) translation was found between the native radial head and the 3 implants. No differences were observed among the radial head conditions except for a difference in medial-lateral translation between the axisymmetric and patient-specific implants (P ¼ .04). Conclusions: Radiocapitellar kinematics of the tested radial head implants were similar in all but one comparison, and all had different kinematics from the native radial head. Patient-specific radial head implants did not prove advantageous relative to conventional implant designs. The shape of the fixed stem unipolar radial head implants had little influence on radiocapitellar kinematics when optimally positioned in this testing model. Level of evidence: Basic Science, Biomechanics. Ó 2015 Journal of Shoulder and Elbow Surgery Board of Trustees. Keywords: Arthroplasty; radial head; kinematics; implant design; patient specific; in vitro
No specific Institutional Review Board approval was required for this study. *Reprint requests: Graham J.W. King, MD, FRCSC, Hand and Upper Limb Centre, St. Joseph’s Health Care, 268 Grosvenor Street, Rm D0-202, London, ON N6A 4L6, Canada. E-mail address: [email protected] (G.J.W. King).
Approximately 33% of elbow fractures involve the radial head.13,15 This structure is an important stabilizer of the elbow, especially in the case of associated ligamentous injuries.15 In the setting of comminuted, displaced, unreconstructable radial head fractures associated with ligament
1058-2746/$ - see front matter Ó 2015 Journal of Shoulder and Elbow Surgery Board of Trustees. http://dx.doi.org/10.1016/j.jse.2014.09.019
Radial implant shape and radiocapitellar kinematics injuries, replacing the radial head with an implant to maintain joint stability is essential.1 A number of commercially available radial head implants have been developed and are in common clinical use for the management of acute radial head fractures and late reconstruction for nonunions and malunions. Currently, most of these implants have a head that is rotationally symmetric around the stem; also termed axisymmetric.2 However, anatomic studies have shown that the radial head is elliptical and that the articulating dish is located eccentrically with respect to the neck.11,12,18 The complex shape of the native radial head poses a question of whether an axisymmetric implant with a fixed stem can adequately replicate normal radiocapitellar joint kinematics during forearm rotation. During this motion, the dish of the radial head rotates about the capitellum, and the margin of the radial head rotates within the radial notch of the ulna. If the native radial head is elliptical and the dish is eccentric, it is possible that a fixed-stem axisymmetric unipolar implant may exhibit abnormal kinematics that may cause pain, stem loosening, and lead to clinical failure. More anatomic implant designs have been developed, but their effectiveness relative to more traditional axisymmetric implants during forearm motion is unknown.3 Furthermore, the concept of a reverse-engineered patient-specific radial head implant based on imaging of the contralateral uninjured elbow is of interest because, theoretically, this should most closely replicate the native radial head and, hence, should reproduce normal kinematics if positioned correctly. In view of the foregoing, the objective of this study was to compare the radiocapitellar kinematics of the native articulation with axisymmetric, population-based quasianatomic, and reverse-engineered patient-specific unipolar implant designs. We hypothesized that more anatomically shaped radial head implants would have similar radiocapitellar kinematics to the native radial heads and that the axisymmetric radial head implants would differ.
Materials and methods Specimen preparation Eight fresh frozen cadaveric upper extremities from male donors (age, 75 8 years) were thawed overnight at room temperature before testing. The specimens were prepared by suturing the biceps and triceps tendons using nylon braided string and the pronator teres with No. 5 Ethibond polyester braided suture (Ethicon Inc/Johnson and Johnson, Somerville, NJ, USA). These sutures were attached to servomotors to achieve active forearm rotation, as described below. The wrist was fixed in neutral flexion by passing a pin through the long finger metacarpal into the distal radius. The surgical incisions were closed using 2-0 Vicryl suture (Ethicon Inc/Johnson and Johnson). The specimens were kept at room temperature and were hydrated with normal saline throughout testing.
259
Implant design The unipolar radial head implant system created for this study consisted of a cemented generic stem that could be used with all implants and the implant head.4 Three radial head implants were developed for assessment: (1) an axisymmetric, (2) a populationbased quasi-anatomic, and (3) a reverse-engineered patient-specific implant. Computed tomography scans of 34 male upper extremities were used to design a series of 3 sizes of population-based quasi-anatomic implants. Reverse-engineered patient-specific implants were also generated for the 8 specimens used in the current study. Mimics image processing software (Materialize NV, Leuven, Belgium) was used to generate a preoperative 3-dimensional (3D) model of the radius. Implant parameters were measured by ellipse fitting the surface model cross sections using custom software based on the Visualization Toolkit (VTK). The quasi-anatomic design used averages derived from these measured parameters, whereas the patient-specific model used each specimen’s unique parameters.4 The shape of the axisymmetric radial head was modeled after the EVOLVE Proline Radial Head System (Wright Medical Technology Inc, Arlington, TN, USA). Unlike the commercially available implant, which incorporates a cylindrical polished stem, the axisymmetric implant was modified to fit on the same custommade stem common to all implant shapes. All radial head implants were formed out of ABS M30 plastic (Fig. 1) using a rapid prototyping machine with an accuracy of 0.127 mm (Stratasys Fortus 400 MC, Eden Prairie, MN, USA).
Testing apparatus The arm was mounted into a custom upper extremity motion simulator that was previously developed in our laboratory and has been validated through the peer-review process6 (Fig. 2). The biceps, triceps, and pronator teres sutures were attached to servomotors to allow for computer-controlled movement. The triceps was activated to maintain 90 of flexion by resting the forearm on a support bar. The biceps and pronator teres were activated to achieve active forearm supination from the pronated position during each stage of testing while the kinematics were measured with Optotrak Certus optical tracking system (NDI, Waterloo, ON, Canada).
Testing protocol After the intact specimen was tested, the lateral collateral ligament (LCL) was sectioned from the lateral epicondyle to provide access to the radial head and subsequently repaired using a braided No. 2 HiFi ultra-high-weight polyethylene suture (CONMED Linvatec, Largo, FL, USA). The suture ends were passed into a hole placed at the isometric point on the lateral epicondyle and out through 2 tunnels located proximal to the site of entry. This allowed the ligament to be repaired such that the original line of action was restored.8 The annular ligament was kept intact. A pneumatic actuator was used to apply a load of 20 N to the LCL before the ligament cable was clamped during testing. Because repeated access to the radiocapitellar joint was required for placement of the implants, the kinematics of the LCL repair model with the native radial head was compared with the intact state to confirm that it could be used as a control.
260
H.L. Shannon et al.
Figure 1 Examples of the radial head implants and stem: (A) reverse-engineered patient-specific, (B) population-based quasi-anatomic, (C) axisymmetric, and (D) custom-made generic stem.
The 3D optical trackers were fixed to the radius and the navigation tool with the implant stem. Feedback to the surgeon was given in real-time to allow corrections of the implant location using a custom LabVIEW program (National Instruments, Austin, TX, USA).5 Once the target location was achieved, the stem was cemented into place using Surgical Simplex bone cement (Stryker, Mahwah, NJ, USA).
Kinematics measurements
Figure 2 ligament.
The forearm motion simulator. LCL, lateral collateral
The stem was cemented into the canal of the radius such that the radial heads would be in an optimal location (Fig. 3). The radial head implants were tested in a random order.
Registration, navigation, and implantation Computed tomography scans were performed on all of the specimens tested, and 3D surface models of the radii were generated as described above. A custom program was used to determine an ideal location for the implant stem based on measurements of the native radius.5 The radial head was then removed using an oscillating saw, and the canal was reamed. The minor diameter of the native radial head was used to select the diameter of axisymmetric implant. The implant height was defined by the commercially available implant system on which the custom implant was modelled.
To track the motion of the forearm, optical tracker mounts were fixed to the radius and ulna directly on the bone surface using custom-made mounts and cortical bone screws. The trackers were attached to the ulnar mount and the humerus. A ring containing 6 optical marker sets that were calibrated to perform as one rigid body was attached to the radius. This was done to maintain visualization of the radius by the tracking system throughout forearm rotation. Kinematic data were analyzed using custom LabVIEW software to quantify the motion of the radius with respect to the humerus during active forearm supination. The humeral and radial coordinate systems were defined using a series of anatomic digitizations.7 For the humerus, the capitellum surface and the trochlear sulcus were traced, and a point was digitized in the center of the humeral shaft at the mid-diaphysis. The capitellum was sphere fit, the trochlear sulcus was circle fit, and their centers were calculated. A vector was made between these 2 center points in the medial direction (the Zhum þ axis). The bisector of these 2 center points was found, and a vector was made from this bisector to the shaft point (long axis). The Yhum þ axis was the cross product of the Zhum þ axis and the long axis. The Xhum þ axis was the cross product of the Yhum þ axis and the Zhum þ axis. The origin of the humeral coordinate system was the center of the capitellum, such that radial head translation about the capitellum during rotation could be measured (Fig. 4). To produce the radial coordinate system, 10 points around the rim of the radial head were digitized in addition to the radial styloid and the dorsal and volar aspects of the lesser sigmoid notch of the distal radius. The 10 rim points were circle fit,and the center was found. The bisector of the dorsal and volar aspects of the
Radial implant shape and radiocapitellar kinematics
261
Figure 3 This flow chart depicts the phases of the testing protocol. Radiocapitellar kinematics were measured during simulated active forearm supination. LCL, lateral collateral ligament; RH, radial head. lesser sigmoid notch was found, and a vector was made from this bisector to the radial styloid point (medial vector). The bisector of this vector was found, and a vector from this bisector to the center of the radial head was made (the Xrad þ axis). The Yrad þ axis was the cross product of the Xrad þ axis and the medial vector. The Zrad þ axis was the cross product of the Xrad þ axis and the Yrad þ axis. The origin of the radial coordinate system was located at the center of the radial head (Fig. 4, B). The outcome variables examined were medial-lateral (ML) translation and anteriorposterior (AP) translation of the center of the radial head with respect to the center of the capitellum (Fig. 4, C and D).
Statistical analyses A 2-way repeated-measures analysis of variance was performed with radial head condition and forearm rotation angle as the independent variables to determine if there were differences between the intact state and the LCL repair (the 2 levels of the condition variable) for both ML and AP translations of the radial head. This was done to determine whether the LCL repair condition could be used as a control. A 2-way repeated measures analysis of variance was also performed with radial head condition and rotation angle as the independent variables to determine if there were differences in kinematics between the native radial head and the 3 implant shapes during forearm rotation. A post hoc pairwise comparison (a ¼ 0.05) was used to determine if there were differences between each implant and the native radial head.
This same technique was used to compare the 3 implant morphologies with each other. SPSS software (IBM, Armonk, NY, USA) was used for all analyses.
Results Owing to the use of cadaveric specimens from elderly donors and the need to compare motion using a repeatedmeasures design, kinematic data were truncated to include an arc from 40 of pronation to 50 of supination, which could be achieved in all specimens. There was no significant difference (P > .05) when ML or AP translations for the intact elbows were compared with the native radial head and after LCL sectioning and repair. For all cases, the 2 conditions were within 1-mm translation of each other. Thus, the LCL sectioned and repaired condition with the native radial head was used as the control for comparison with the radial head implants. Significant differences in ML translation were found among the radial head morphologies (P ¼ .03); however, pairwise comparisons only showed a significant difference between the axisymmetric and patient-specific implants (P ¼ .04; Fig. 5). As the forearm moved from pronation to supination, the radial head translated medially (P ¼ .04). There was a significant effect of radial head morphology on
262
H.L. Shannon et al.
Figure 4 Anterior-posterior (AP) and medial-lateral (ML) translations of the radial head with respect to the center of the capitellum were examined. (A) Anterior view of the humerus shows the humeral coordinate system with the center of the capitellum as the origin. (B) Proximal view of the radial head shows the radial coordinate system with the center of the radial head as the origin. (C) Direction of ML translation of the radial head is shown relative to the capitellum, and (D) direction of AP translation of the radial head is shown relative to the capitellum.
Figure 5 This graph displays the mean data for active radial head (RH) translations in the medial-lateral (ML) direction (n ¼ 8). A significant difference in ML translation was found between the native RH and the 3 different implant shapes (P ¼ .03). Pairwise comparisons did not demonstrate any differences between the 4 RH conditions, except for a significant difference in ML translation between the axisymmetric and patient-specific implant (P ¼ .04). The standard deviations, which ranged from 0.86 to 2.42 mm, were removed for clarity.
Figure 6 This graph displays the mean data for active radial head (RH) translations in the anterior-posterior (AP) direction (n ¼ 8). A significant difference in AP translation was found between the native RH and the 3 different implant shapes (P ¼ .03). The standard deviations, which ranged from 1.38 to 2.68 mm, were removed for clarity.
AP translation (P ¼ .03); however, pairwise comparisons failed to detect a difference between the native radial head and the 3 different implant shapes (P > .05; Fig. 6). AP translation did not change significantly during forearm rotation (P ¼ .41).
The effects of 3 fixed-stem unipolar radial head implant shapes (ie, the axisymmetric, population-based quasianatomic, and reverse-engineered patient-specific) on radiocapitellar kinematics during rotation were examined. Small but significant differences in kinematics were evident
Discussion
Radial implant shape and radiocapitellar kinematics among radial head morphologies during forearm rotation. The change in kinematics after radial head arthroplasty may be partly related to the small errors (<2 mm translation and <11 rotation) in placement of the stem despite image-based computer navigation.5 The errors measured in rotation would be expected to have a greater effect on the 2 anatomic implants than on the axisymmetric implant because the anatomic implants have off-centered dishes that rely on accurate placement to replicate native radial head kinematics. It is likely that up to 11 of malrotation altered the position of the dish of both the quasi-anatomic and patient-specific implants. This error in rotational orientation would not have had a significant effect on the axisymmetric implant because the dish is located centrally on this implant. The observation that the kinematics of the more anatomic implants and the axisymmetric implants were similar to the native articulation, despite rotational malpositioning, suggests other factors such as translational malpositioning or annular ligament and interosseous membrane tension may have a greater influence on radiocapitellar kinematics than implant shape. Further studies are needed to quantify the articular contact area, location, and pressure at the radiocapitellar joint and to further elucidate the influence of errors in implant positioning. The reverse-engineered patientspecific radial head implants used in the current study had kinematics that were similar to the more conventional offthe-shelf implant designs, suggesting that the increased costs associated with these custom devices may not be justified. The radial head has a complex elliptical shape. The articular dish is typically offset from the center of the head, and the head is offset from the center of the neck.11,12,18 Hence, changes in kinematics after radial head arthroplasty are expected unless the implant is perfectly positioned during surgery and the implant shape and size replicates the native anatomy. Proper tracking within the joint is essential to maintain the health of the articular cartilage. Abnormal contact between the radius and capitellum during elbow movement could lead to increased stress on parts of the articulation causing pain, premature cartilage, wear of the capitellum,17 and an early onset of osteoarthritis. This may lead to the need for removal of the radial head implant or revision to a unicompartmental radiocapitellar arthroplasty.10,16 The implant stem was cemented into the canal of the radius to ensure the quasi-anatomic and patient-specific radial head implants were both located in an optimal position. Because we used a custom stem to fit all implants, the axisymmetric implant was fixed in place as well. Although a fixed position is necessary for anatomically shaped implants and is used by some commercially available axisymmetric implants, other unipolar implants are designed to have a loose-fitting stem or have a bipolar articulation, such that the implant has some ability to selfalign the dish against the capitellum.2 During forearm motion, the radial heads of these implants may move
263 slightly with respect to the proximal radius such that they stay in optimal contact with the capitellum. Future studies should compare the radiocapitellar kinematics of fixedstem unipolar radial head implants with the kinematics of a loose-fitting stem or a bipolar articulation. The native radial head with an LCL repair was used as the control in this experiment so that repeated access to the radiocapitellar joint could be achieved. The LCL resection and repair technique described by Fraser et al.8 was successfully repeated in this study. By removing the LCL from the lateral epicondyle, the radial head could be accessed easily. The annular ligament, an important radial head stabilizer, was kept intact. By using strong braided sutures and tensioning the LCL repair to 20 N, the control closely replicated the radiocapitellar kinematics of the intact elbow during forearm rotation, as evidenced by the fact that no difference was observed in ML or AP measurements between intact and repaired testing groups. The kinematic pathways of the radial head with respect to the capitellum observed in the current study are similar to those reported by Galik et al.9 They found that the radial head translates an average of 2.1 mm in the AP direction and 1.6 mm in the ML direction during forearm rotation. Our data for mean AP and ML translations are within these ranges. In our study we kept the annular ligament intact, unlike in the study by Galik et al,9 where the annular ligament was sectioned. This important elbow stabilizer, as well as the concavity compression of the curved articular dish of the radial head with respect to the spherical capitellum, likely helped to control the motions of the native radial head and the various implants tested. The current study has several limitations. We used cadaveric specimens from elderly donors, whereas most patients undergoing radial head arthroplasty tend to be younger. The motion simulator may not precisely replicate muscle activation and forearm kinematics in vivo; however, the repeated-measures experimental design and precise kinematic measurements using an optical tracking system made it possible to quantify significant differences between radial head morphologies. A post hoc power analysis showed that our sample size of 8 specimens was sufficient to determine these differences. We only studied forearm rotation kinematics at 90 of flexion and did not quantify kinematics during elbow motion. Owing to the limitation of testing with fresh cadaveric specimens, we chose to focus on this position because it reflects the position of function for the upper extremity and thus is most often used in studies of radiocapitellar kinematics. Our computer and image-assisted implant positioning did not precisely replicate the target position of the implant stems as we had hoped. However, the error in positioning a radial head implant without navigation has not been quantified but is likely considerably higher than those achieved with the precise techniques used in the current study.14 Despite this degree of inaccuracy in implant positioning (<2 mm in translation and <11 in rotation), it
264 did not appear to affect the kinematics of the implants, which were similar to the native radial head. Further studies are needed to determine whether the favorable kinematics of the fixed-stem unipolar radial head implants used in this in vitro study can be achieved clinically with standard nonnavigated techniques of implant positioning.
Conclusions The kinematics of the radial head implants and the native articulation were similar except for a difference in ML translation between the axisymmetric and patientspecific implants. This study suggests that the kinematics of the radiocapitellar joint are not sensitive to radial head implant shape during simulated forearm rotation at 90 of flexion when using an optimally positioned fixed-stem unipolar device. The reverseengineered patient-specific radial head implants used in the current study had similar kinematics as the more conventional off-the-shelf implant designs.
Disclaimer Grants from the Natural Sciences and Engineering Research Council of Canada (grant no. 153063-2013) and the Canadian Institutes of Health Research (grant no. MOP-106500) supported the acquisition of cadaveric specimens, surgical supplies, imaging, and personnel required for this study. Graham King is a consultant and receives royalties from Tornier Inc. and Wright Medical Technology, which are bith related to the subject of this article. The other authors, their immediate families, and any research foundations with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.
References 1. Beingessner DM, Dunning CE, Gordon KD, Johnson JA, King GJ. The effect of radial head excision and arthroplasty on elbow kinematics and stability. J Bone Joint Surg Am 2004;86:1730-9. 2. Calfee R, Madom I, Weiss AP. Radial head arthroplasty. J Hand Surg Am 2006;31:314-21. http://dx.doi.org/10.1016/j.jhsa.2005.12.005.
H.L. Shannon et al. 3. Chanlalit C, Shukla DR, Fitzsimmons JS, An KN, O’Driscoll SW. Influence of prosthetic design on radiocapitellar concavitycompression stability. J Shoulder Elbow Surg 2011;20:885-90. http:// dx.doi.org/10.1016/j.jse.2011.03.009 4. Deluce SR, Shannon H, Lalone E, Ferreira LM, King GJ, Johnson JA. Comparison of population-based and patient-specific radial head implants using surface matching. Presented at the Orthopaedic Research Society 2013 Annual Meeting, San Antonio, TX; Jan 26-29, 2013. 5. Deluce SR, Shannon H, Lalone E, Ferreira LM, King GJ, Johnson JA. Development of a computer- and image-assisted guidance system for radial head arthroplasty. Presented at the Orthopaedic Research Society 2013 Annual Meeting, San Antonio, TX; Jan 26-29, 2013. 6. Ferreira LM, Johnson JA, King GJ. Development of an active elbow flexion simulator to evaluate joint kinematics with the humerus in the horizontal position. J Biomech 2010;43:2114-9. http://dx.doi.org/10. 1016/j.jbiomech.2010.04.007 7. Ferreira LM, King GJ, Johnson JA. Motion-derived coordinate systems reduce inter-subject variability of elbow flexion kinematics. J Orthop Res 2011;29:596-601. http://dx.doi.org/10.1002/jor.21278 8. Fraser GS, Pichora JE, Ferreira LM, Brownhill JR, Johnson JA, King GJ. Lateral collateral ligament repair restores the initial varus stability of the elbow: an in vitro biomechanical study. J Orthop Trauma 2008;22:615-23. http://dx.doi.org/10.1097/BOT.0b013e31818 86f37 9. Galik K, Baratz ME, Butler AL, Dougherty J, Cohen MS, Miller MC. The effect of the annular ligament on kinematics of the radial head. J Hand Surg Am 2007;32:1218-24. http://dx.doi.org/10.1016/j.jhsa. 2007.06.008. 10. Heijink A, Morrey BF, Cooney WP III. Radiocapitellar hemiarthroplasty for radiocapitellar arthritis: a report of three cases. J Shoulder Elbow Surg 2008;17:e12-5. http://dx.doi.org/10.1016/j.jse. 2007.04.009. 11. King GJ, Zarzour ZD, Patterson SD, Johnson JA. An anthropometric study of the radial head: implications in the design of a prosthesis. J Arthroplasty 2001;16:112-6. 12. Koslowsky TC, Germund I, Beyer F, Mader K, Krieglstein CF, Koebke J. Morphometric parameters of the radial head: an anatomical study. Surg Radiol Anat 2007;29:225-30. http://dx.doi.org/10.1007/ s00276-007-0197-1 13. Mason ML. Some observations on fractures of the head of the radius with a review of one hundred cases. Br J Surg 1954;42:123-32. 14. McDonald CP, Johnson JA, Peters TM, King GJ. Image-based navigation improves the positioning of the humeral component in total elbow arthroplasty. J Shoulder Elbow Surg 2010;19:533-43. http://dx. doi.org/10.1016/j.jse.2009.10.010 15. Morrey BF. The elbow and its disorders. Philadelphia, PA: Saunders Elsevier, ISBN 1416029028; 2008. p. 419. 16. Sabo MT, Shannon H, De Luce S, Lalone E, Ferreira LM, Johnson JA, et al. Elbow kinematics after radiocapitellar arthroplasty. J Hand Surg Am 2012;37:1024-32. http://dx.doi.org/10.1016/j.jhsa.2012.02.021 17. van Riet RP, Van Glabbeek F, Neale PG, Bimmel R, Bortier H, Morrey BF, et al. Anatomical considerations of the radius. Clin Anat 2004;17:564-9. http://dx.doi.org/10.1002/ca.10256 18. van Riet RP, Van Glabbeek F, Neale PG, Bortier H, An KN, O’Driscoll SW. The noncircular shape of the radial head. J Hand Surg Am 2003;28:972-8. http://dx.doi.org/10.1016/S0363-5023(03) 00426-X.