SCIENTIFIC ARTICLE
Effect of Radial Head Implant Shape on Joint Contact Area and Location During Static Loading Hannah L. Shannon, MSc, Simon R. Deluce, MSc, Emily A. Lalone, PhD, Ryan Willing, PhD, Graham J. W. King, MD, MSc, James A. Johnson, PhD
Purpose To examine the effect of implant shape on radiocapitellar joint contact area and location in vitro. Methods We used 8 fresh-frozen cadaveric upper extremities. An elbow loading simulator examined joint contact in pronation, neutral rotation, and supination with the elbow at 90 flexion. Muscle tendons were attached to pneumatic actuators to allow for computercontrolled loading to achieve the desired forearm rotation. We performed testing with the native radial head, an axisymmetric implant, a reverse-engineered patient-specific implant, and a population-based quasi-anatomic implant. Implants were inserted using computer navigation. Contact area and location were quantified using a casting technique. Results We found no significant difference between contact locations for the native radial head and the 3 implants. All of the implants had a contact area lower than the native radial head; however, only the axisymmetric implant was significantly different. There was no significant difference in contact area between implant shapes. Conclusions The similar contact areas and locations of the 3 implant designs suggest that the shape of the implant may not be important with respect to radiocapitellar joint contact mechanics when placed optimally using computer navigation. Further work is needed to explore the sensitivity of radial head implant malpositioning on articular contact. The lower contact area of the radial head implants relative to the native radial head is similar to previous benchtop studies and is likely the result of the greater stiffness of the implant. Clinical relevance Radial head implant shape does not appear to have a pronounced influence on articular contact, and both axisymmetric and anatomic metal designs result in elevated cartilage stress relative to the intact state. (J Hand Surg Am. 2015;40(4):716e722. Copyright Ó 2015 by the American Society for Surgery of the Hand. All rights reserved.) Key words Biomechanics, elbow, radial head arthroplasty, computer-assisted orthopedic surgery, joint contact.
From the Hand and Upper Limb Centre, Western University, London, Ontario, Canada. Received for publication August 21, 2013; accepted in revised form December 9, 2014. G.J.W.K. receives royalties from and is a consultant for Wright Medical Technologies and Tornier, Inc. Corresponding author: Graham J. W. King, MD, MSc, Roth j McFarlane Hand and Upper Limb Centre, St. Joseph’s Health Centre, 268 Grosvenor Street, London, Ontario N6A 4L6, Canada; e-mail: [email protected]. 0363-5023/15/4004-0012$36.00/0 http://dx.doi.org/10.1016/j.jhsa.2014.12.017
716
r
Ó 2015 ASSH
r
Published by Elsevier, Inc. All rights reserved.
T
and variable shape. Numerous elbow morphology studies have reported that the radial head is elliptical3e5; however, most commercially available radial head implants are symmetric about a central axis, or axisymmetric. Only one anatomical asymmetric design is currently commercially available.6 In some systems, the implant stem is smooth and purposely placed loosely, because small amounts of stem movement in the radial neck may compensate for the non-anatomic shape.7 Other axisymmetric implants have a bipolar HE RADIAL HEAD HAS A COMPLEX 1,2
EFFECT OF RADIAL HEAD SHAPE ON JOINT CONTACT
articulation, containing a joint between the stem and the radial head to optimize joint contact, but have a potential risk of polyethylene wear and provide less contribution to radiocapitellar stability.6,8,9 Another group of axisymmetric implants aims for secure stem fixation, most commonly with uncemented ingrowth stems.6 When anatomic asymmetrical designs are used, it is essential that they be positioned and fixed in the correct location to ensure proper joint alignment and hence optimize radiocapitellar contact.10 The contact of a metallic radial head on articular cartilage can be expected to alter joint contact patterns owing to the stiffness of the implant.11,12 Changes in implant alignment with respect to the capitellum owing to incorrect positioning or differences in the implant shape relative to the native radial head may also contribute to changes in contact patterns and hence alter articular cartilage loading. Collectively, like any hemiarthroplasty, these changes in stiffness, alignment, and shape have a potential to cause degenerative changes in the opposing cartilaginous surface.11,12 The focus of the current study was on evaluating the effect of radial head implant shape on radiocapitellar contact using computer-assisted surgical techniques to ensure optimal implant positioning and a whole elbow model to mimic a clinically relevant loading environment. The objective of this study was to compare the radiocapitellar contact patterns of 3 radial head implant designs that included axisymmetric, population-based quasi-anatomic, and reverse-engineered patient-specific devices. We hypothesized that anatomically shaped radial head implants would have greater contact area than the axisymmetric radial head implants and demonstrate radiocapitellar contact patterns similar to the native radial heads.
FIGURE 1: Customized stem. View of the stem inside the implant.13
We used custom-made axisymmetric radial head implants in this study in lieu of a commercially available implant so that it would match our custom stem. In this case, the axisymmetric implant was modeled after the Evolve Proline Radial Head System (Wright Medical Technology, Inc, Arlington, TN) and included 20-, 22-, 24-, and 26-mm implant sizes. An experienced upper extremity orthopedic surgeon compared the size of the excised native radial head with the axisymmetric implants as is performed clinically. The minor diameter of the elliptical native radial head was used to select the diameter of the axisymmetric implant. If the radial head was judged to be between sizes, the smaller sized prosthesis was selected. To design a series of population-based, quasianatomic radial head implants, we measured computed tomographic scans of 34 male elbows. These specimens were sorted by maximum diameter into 3 groups representing the specimens within 1 SD of the mean (QM) (n ¼ 24), above 1 SD (Qþ) (n ¼ 5), and below it (Qe) (n ¼ 5). All specimens were within 3 SD of the mean. Implants were then generated for each of these groups by averaging a large number of measured parameters in these specimens.13 The size of the excised native radial head was compared with the population-based elliptical implants. The populationbased implant that most closely matched the maximum diameter of the native radial head was used. If the radial head was judged to be between sizes, the smaller sized prosthesis was selected. We applied the same measurement techniques to the scans of each
MATERIALS AND METHODS Design of implants The radial head implant system designed for this protocol consisted of 2 components: a custom-made generic stem and the radial head (Figs. 1, 2).13 All radial head implants were formed out of acrylonitrile butadiene styrene (ABS) M30 plastic (Stratasys, Eden Prairie, MN) using a rapid prototyping machine with an accuracy of 0.127 mm (400MC, Stratasys). We performed a benchtop study to determine whether the plastic implants would produce a contact area similar to that of traditional metal implants. Three casts were made using a cadaveric humerus with one plastic axisymmetric implant and one metal axisymmetric implant of the same size. J Hand Surg Am.
717
r
Vol. 40, April 2015
718
EFFECT OF RADIAL HEAD SHAPE ON JOINT CONTACT
FIGURE 2: Implants. Example subset of implants that were fabricated.13
maintained in 90 flexion by resting the ulna on a smooth support bar.
individual specimen tested in this protocol to produce 8 reverse-engineered, patient-specific implants.13
Joint contact measurements We quantified radiocapitellar joint contact using a previously reported casting technique (Fig. 4).11,16,17 After mixing, 2 mL silicone impression material (Reprosil Medium Body Vinyl Polysiloxane Impression Material, DENTSPLY International, Inc, York, PA) was injected onto the radial head using a syringe. The radiocapitellar joint was reduced and the LCL actuator was loaded to 20 N to simulate a clinical ligament repair.15 Static (muscle) loading was applied across the elbow to mimic normal muscle tone in pronation, neutral rotation, and supination and a cast was taken in each of these positions. The casting process was first conducted with the native radial head. The radial head was then resected and the custom stem was navigated into position using image-based computer navigation.13 Casting was then repeated for all 3 implant designs in randomized order. The cast was left to cure for 25 minutes before removal. Once the cast had set, the LCL tension was released to allow the joint to be subluxated for cast removal.
Specimen testing Eight fresh-frozen cadaveric specimens (male; average age, 75 8 y; right upper extremity) were mounted in a forearm motion simulator (Fig. 3).14 To gain access to the radial head, the extensor muscles were elevated off the lateral collateral ligament (LCL) and the humeral origin of the LCL was sectioned from the lateral epicondyle. The LCL was repaired using heavy nonabsorbable locking Krakow sutures that were then passed through an isometric drill hole at its anatomic origin and transosseous tunnels.15 The triceps (T) and biceps (B) and the pronator teres (PT) were sutured with a braided suture. The sutures of these muscles and the LCL were then attached to pneumatic actuators (Airpel, Airpot Corp, Norwalk, CT) via a cable. A 20-N constant load was applied to the LCL actuator to simulate a clinical LCL repair.15 Static loads (based on joint angle and muscle load data from a simulated active forearm rotation) were applied to the T, B, and PT to maintain full pronation (T ¼ 50 N, B ¼ 15 N, and PT ¼ 80 N), neutral rotation (T ¼ 50 N, B ¼ 35 N, and PT ¼ 25 N), and full supination (T ¼ 50 N, B ¼ 40 N, and PT ¼ 30 N). The elbows were J Hand Surg Am.
Analysis of casts The lateral side of the cast was marked before it was removed from the radiocapitellar joint. The excised r
Vol. 40, April 2015
EFFECT OF RADIAL HEAD SHAPE ON JOINT CONTACT
719
FIGURE 3: Elbow motion simulator. The forearm is in neutral rotation at 90 flexion in the elbow motion simulator. Static loads were applied to the 3 muscles: biceps (BIC), triceps (TRI), and pronator teres (PT), and also the LCL, using pneumatic actuators.
tracked stylus (Optotrak Certus, NDI, Waterloo, Ontario, Canada) relative to the digitization block (Fig. 5C). The stylus had an accuracy of 0.1 mm.18 The point cloud digitizations were then converted to 3-dimensional surfaces (Fig. 5D, E) using a radial basis function program in MATLAB (MathWorks, Natick, MA). Custom Visualization Toolkit (Kitware, Clifton Park, NY) software was used to measure the contact area (Fig. 5F). To determine the location of the contact area on the radial head, we calculated the centroid of the contact area patch. The radial basis function surfaces that were previously calculated were converted back into points to achieve an evenly distributed point cloud. Custom MATLAB code was used to fit a plane to the points. These were given indices and projected onto the plane such that a 2-dimensional surface was created. The mean of the points was calculated to determine the point closest to the center of the surface. The points were then projected back onto the original surface. The 3-dimensional coordinates of the center point (centroid) were determined. To measure the location of the contact area, the center of the radial head was determined. Twenty points were selected around the rim on the radial head. A circle was then fit to these points and the center and radius of the circle were calculated. The distance between
FIGURE 4: Example of casts. A set of casts, all taken in the neutral position.
native radial head was cemented to the digitization block. Infrared light-emitting diode markers were also attached so that the position of the contact patch would be known. The orientation markers on the native radial head were digitized and the entire radial head was traced. The cast was then placed onto the radial head in the correct orientation such that the lateral marks on the cast and a lateral hole on the radial head (marked before removal) lined up. For the casts of the radial head implants, the implant was removed from the stem with the cast still in situ and then placed on a stem fixed to the digitization block (Fig. 5AeC). A technique described by Lalone et al16 was used to determine the radiocapitellar contact area. This was quantified by digitizing the casts using an optically J Hand Surg Am.
r
Vol. 40, April 2015
720
EFFECT OF RADIAL HEAD SHAPE ON JOINT CONTACT
FIGURE 5: Process of quantifying the contact area. A Space where the silicone impression material was injected. B Cast removed from the joint after it has cured. The void space represents the contact area. C The stylus was used to trace the contact area and was tracked by the Optotrak Certus system. D Point cloud produced from digitization. E Surface made from the point cloud. F Area calculation.
FIGURE 6: Contact area for all radial head conditions. Mean contact area ( 1 SD) is shown for all radial head conditions at all 3 angles of rotation. There was a significant effect of radial head condition on contact area (P ¼ .005). There was no difference in contact area between implants; however, the axisymmetric implant had significantly lower contact area than the native radial head (P ¼ .008).
FIGURE 7: Distance from the center of the radial head to the contact area centroid. Mean distance ( 1 SD) is shown for all radial head conditions. The data were normalized against the radius of the radial head. There was no significant difference in contact location (P ¼ .220).
radial head radius such that the center was 0 and the rim was 1.
the center of the radial head and the contact centroid was compared with the distance from the centroid to the rim of the radial head. This was done to quantify whether the contact patch lay closer to the center or the rim for the different radial head conditions. Values for distances were normalized against the J Hand Surg Am.
Statistical analyses Two-way repeated-measures ANOVA was performed with radial head condition (native radial head, axisymmetric, quasi-anatomic, and patient-specific) r
Vol. 40, April 2015
EFFECT OF RADIAL HEAD SHAPE ON JOINT CONTACT
and rotation angle (full pronation, neutral, and full supination) as the 2 factors to determine whether there was a statistically significant difference in contact area and location. We used a post hoc pairwise comparison (a ¼ .05) to determine potential differences in contact area and location between each implant and the native radial head. This same technique was used to compare the 3 implant morphologies with each other.
40% reduction in the study presented here. In the experimental setup used by Liew et al,11 the radiocapitellar joint was aligned by the experimenter, and concavity-compression was created as loading was applied, whereas our model allowed the muscles, articulation, and ligaments to align the joint while applying muscle tones through actuators. Therefore, the contact area measured in the current study was likely more clinically relevant than the contact area determined from previous benchtop studies. In both studies, the implant materials used were much stiffer than the cartilage of the native radiocapitellar joint. In the current study, the modulus of elasticity of the ABS plastic used to make the implants was 2.3 GPa whereas in the study by Liew et al the implants were made of cobalt chromium with a modulus of 230 GPa. The effect of the different moduli of materials was a potential concern; therefore, we compared the contact area of ABS plastic implants relative to metal implants in a benchtop study.19 Three casts were made using a cadaveric humerus with one ABS plastic axisymmetric implant and one metal axisymmetric implant of the same size. The ABS plastic and metal implants showed no significant differences in contact area (P ¼ .920). It is likely that the high stiffness of the implants and not the implant shape resulted in the general decrease in contact area. Future work is needed to find an implant material that behaves similarly to cartilage. The decrease in contact area seen after radial head arthroplasty was performed could also have resulted from small errors (less than 2-mm translation and less than 11 rotation) during the image-based computer navigation of the stem.13 It is likely that rotational error would have a more noticeable effect on the anatomical implants because the dish of these implants is located eccentrically and there is an optimal position for the dish to ensure that it can track accurately. Because the dish is located centrally on the axisymmetric implant, rotational error would not have had a significant effect. Factors such as translational malpositioning or annular ligament and interosseous membrane tension also may have an influence on radiocapitellar contact area. Further studies are needed to better understand the sensitivity of implant malpositioning on contact area at the radiocapitellar joint. A potential weakness of this study was that we used an LCL repair technique to gain access to the joint. On assessment there was no difference in motion pathways before and after ligament sectioning and repair.19 Therefore, repair of the LCL using this technique restores the normal kinematics of the elbow.
RESULTS This benchtop study to compare the plastic and metal implants showed no significant differences in contact area (P ¼ .92). There was a significant effect of radial head condition (P ¼ .005) (Fig. 6). All of the implants had a lower contact area than the native radial head but only the axisymmetric implant was significantly different (P ¼ .008). There was no significant difference in contact area among the 3 implant shapes (P > .050). There was no significant effect of forearm rotation position on contact area (P ¼ .240). There was no significant difference in contact location for the native radial head, axisymmetric, quasianatomic, or patient-specific implants (P ¼ .220) (Fig. 7). DISCUSSION Previous investigations have shown that the articulating surface of the radial head is variably elliptical and is typically offset from the center of the radial neck.1e3 Differences in radial head implant shape relative to the native articulation were expected to affect both the magnitude and location of contact within the radiocapitellar joint. Three radial head implant shapes, either axisymmetric, population-based quasi-anatomic, or reverse-engineered patient-specific, were implanted and the radiocapitellar contact was quantified in 3 positions of forearm rotation. The results showed no significant differences among the 3 implant morphologies for contact area or contact location. This suggests that the shape of the radial head implant may not be important with respect to altering radiocapitellar joint contact mechanics. A previous study11 reported significantly less contact with axisymmetric metal implants compared with the native radial head. The results from the current study agree with those earlier findings in which the axisymmetric radial head displayed a significantly smaller contact area than the intact radial head. However, the previous authors found an average 60% decrease in contact area with the axisymmetric implant, compared with an average J Hand Surg Am.
721
r
Vol. 40, April 2015
722
EFFECT OF RADIAL HEAD SHAPE ON JOINT CONTACT
ACKNOWLEDGMENT Support was received from the Canadian Institute for Health Research. National Sciences and Engineering Research Council.
This study used a sample size of 8, which has been adequate for various in vitro biomechanical studies completed in our laboratory. A post hoc power analysis determined that although there was sufficient power (0.90) for the contact area analysis with 8 specimens, 15 specimens would be needed to achieve sufficient power for the contact location analysis (observed power, 0.25). The native, quasi-anatomic, and patient-specific radial head implants varied in dish position and shape (from more circular to more elliptical in nature), which may have contributed to the variability in our contact location data and the lack of detection of a significant difference between groups. The radiocapitellar joint experiences concavee convex surface contact mechanics. This relationship between the capitellum and the radial head relies on proper joint alignment such that the concave dish of the radial head centers on the convex surface of the capitellum. A few options exist to ensure that the joint mechanics are restored. Bipolar radial heads contain a joint between the head and stem of the implant, allowing some tilting of the dish.9 Some axisymmetric implants employ a smooth stem that is purposely placed loose so that the head can move small amounts and essentially self-align with the capitellum.7 Anatomic implants are equipped with markings to try to aid in the best possible alignment.10 In this study, all of the implants were fixed in place. Although this was essential for anatomically designed implants and is commonly used for most axisymmetric implants, this process likely does not represent the behavior of axisymmetric implants, which employ a loose-fitting smooth stem, or those that incorporate a bipolar articulation. By cementing the axisymmetric implants we removed their potential self-alignment property. The fact that there were no differences in contact area or location among the 3 implants suggests that the concavity compression of the radiocapitellar articulation together with the annular ligament (which was left intact in the current study) kept the implants aligned, making implant shape less important. To increase joint contact and hence reduce cartilage loading, future work should focus on optimally aligning these implants and perhaps investigating less stiff materials. Further studies are needed to evaluate the importance of implant positioning, which we did not address. Clinical studies are also required to determine whether these contact patterns will influence the outcomes of radial head arthroplasty.
J Hand Surg Am.
REFERENCES 1. Popovic N, Djekic J, Lemaire R, Gillet P. A comparative study between proximal radial morphology and the floating radial head prosthesis. J Shoulder Elbow Surg. 2005;14(4):433e440. 2. Swieszkowski W, Skalski K, Pomianowski S, Kedzior K. The anatomic features of the radial head and their implication for prosthesis design. Clin Biomech (Bristol, Avon.). 2001;16(10):880e887. 3. 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(1):112e116. 4. 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(3):225e230. 5. 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(6):972e978. 6. Calfee R, Madom I, Weiss AP. Radial head arthroplasty. J Hand Surg Am. 2006;31(2):314e321. 7. van Riet RP, Van Glabbeek F, Baumfeld JA, et al. The effect of the orientation of the radial head on the kinematics of the ulnohumeral joint and force transmission through the radiocapitellar joint. Clin Biomech (Bristol, Avon.). 2006;21(6):554e559. 8. 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(6):885e890. 9. Dotzis A, Cochu G, Mabit C, Charissoux JL, Arnaud JP. Comminuted fractures of the radial head treated by the Judet floating radial head prosthesis. J Bone Joint Surg Br. 2006;88(6):760e764. 10. Acumed. Anatomic Radial Head System. Available at: http://www. acumed.net/sites/default/files/literature/brochure-surgical-technique/ ARH_Brochure.pdf. Accessed March 3, 2014. 11. Liew VS, Cooper IC, Ferreira LM, Johnson JA, King GJ. The effect of metallic radial head arthroplasty on radiocapitellar joint contact area. Clin Biomech (Bristol, Avon.). 2003;18(2):115e118. 12. Sabo MT, Shannon H, Ng J, Ferreira LM, Johnson JA, King GJ. The impact of capitellar arthroplasty on elbow contact mechanics: Implications for implant design. Clin Biomech (Bristol, Avon.). 2011;26(5):458e463. 13. Deluce SR. Design, Validation and Navigation of Anatomic Population-Based and Patient-Specific Radial Head Implants [Master’s thesis]. London, Ontario, Canada: Western University; 2011. 14. 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(11):2114e2119. 15. 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(9):615e623. 16. Lalone EA, Peters TM, King GW, Johnson JA. Accuracy assessment of an imaging technique to examine ulnohumeral joint congruency during elbow flexion. Comput Aided Surg. 2012;17(3):142e152. 17. Stormont TJ, An KN, Morrey BF, Chao EY. Elbow joint contact study: comparison of techniques. J Biomech. 1985;18(5):329e336. 18. Nothern Digital, Inc. Optotrak Certus Motion Capture System. Available at: http://www.ndigital.com/lifesciences/certus-techspecs. php. Accessed March 3, 2014. 19. Shannon HL. The Contact Mechanics and Kinematics of Radial Head Implants [Master’s thesis]. London, Ontario, Canada: Western University; 2012.
r
Vol. 40, April 2015
Effect of Radial Head Implant Shape on Joint Contact Area and Location During Static Loading Hannah L. Shannon, MSc, Simon R. Deluce, MSc, Emily A. Lalone, PhD, Ryan Willing, PhD, Graham J. W. King, MD, MSc, James A. Johnson, PhD
Purpose To examine the effect of implant shape on radiocapitellar joint contact area and location in vitro. Methods We used 8 fresh-frozen cadaveric upper extremities. An elbow loading simulator examined joint contact in pronation, neutral rotation, and supination with the elbow at 90 flexion. Muscle tendons were attached to pneumatic actuators to allow for computercontrolled loading to achieve the desired forearm rotation. We performed testing with the native radial head, an axisymmetric implant, a reverse-engineered patient-specific implant, and a population-based quasi-anatomic implant. Implants were inserted using computer navigation. Contact area and location were quantified using a casting technique. Results We found no significant difference between contact locations for the native radial head and the 3 implants. All of the implants had a contact area lower than the native radial head; however, only the axisymmetric implant was significantly different. There was no significant difference in contact area between implant shapes. Conclusions The similar contact areas and locations of the 3 implant designs suggest that the shape of the implant may not be important with respect to radiocapitellar joint contact mechanics when placed optimally using computer navigation. Further work is needed to explore the sensitivity of radial head implant malpositioning on articular contact. The lower contact area of the radial head implants relative to the native radial head is similar to previous benchtop studies and is likely the result of the greater stiffness of the implant. Clinical relevance Radial head implant shape does not appear to have a pronounced influence on articular contact, and both axisymmetric and anatomic metal designs result in elevated cartilage stress relative to the intact state. (J Hand Surg Am. 2015;40(4):716e722. Copyright Ó 2015 by the American Society for Surgery of the Hand. All rights reserved.) Key words Biomechanics, elbow, radial head arthroplasty, computer-assisted orthopedic surgery, joint contact.
From the Hand and Upper Limb Centre, Western University, London, Ontario, Canada. Received for publication August 21, 2013; accepted in revised form December 9, 2014. G.J.W.K. receives royalties from and is a consultant for Wright Medical Technologies and Tornier, Inc. Corresponding author: Graham J. W. King, MD, MSc, Roth j McFarlane Hand and Upper Limb Centre, St. Joseph’s Health Centre, 268 Grosvenor Street, London, Ontario N6A 4L6, Canada; e-mail: [email protected]. 0363-5023/15/4004-0012$36.00/0 http://dx.doi.org/10.1016/j.jhsa.2014.12.017
716
r
Ó 2015 ASSH
r
Published by Elsevier, Inc. All rights reserved.
T
and variable shape. Numerous elbow morphology studies have reported that the radial head is elliptical3e5; however, most commercially available radial head implants are symmetric about a central axis, or axisymmetric. Only one anatomical asymmetric design is currently commercially available.6 In some systems, the implant stem is smooth and purposely placed loosely, because small amounts of stem movement in the radial neck may compensate for the non-anatomic shape.7 Other axisymmetric implants have a bipolar HE RADIAL HEAD HAS A COMPLEX 1,2
EFFECT OF RADIAL HEAD SHAPE ON JOINT CONTACT
articulation, containing a joint between the stem and the radial head to optimize joint contact, but have a potential risk of polyethylene wear and provide less contribution to radiocapitellar stability.6,8,9 Another group of axisymmetric implants aims for secure stem fixation, most commonly with uncemented ingrowth stems.6 When anatomic asymmetrical designs are used, it is essential that they be positioned and fixed in the correct location to ensure proper joint alignment and hence optimize radiocapitellar contact.10 The contact of a metallic radial head on articular cartilage can be expected to alter joint contact patterns owing to the stiffness of the implant.11,12 Changes in implant alignment with respect to the capitellum owing to incorrect positioning or differences in the implant shape relative to the native radial head may also contribute to changes in contact patterns and hence alter articular cartilage loading. Collectively, like any hemiarthroplasty, these changes in stiffness, alignment, and shape have a potential to cause degenerative changes in the opposing cartilaginous surface.11,12 The focus of the current study was on evaluating the effect of radial head implant shape on radiocapitellar contact using computer-assisted surgical techniques to ensure optimal implant positioning and a whole elbow model to mimic a clinically relevant loading environment. The objective of this study was to compare the radiocapitellar contact patterns of 3 radial head implant designs that included axisymmetric, population-based quasi-anatomic, and reverse-engineered patient-specific devices. We hypothesized that anatomically shaped radial head implants would have greater contact area than the axisymmetric radial head implants and demonstrate radiocapitellar contact patterns similar to the native radial heads.
FIGURE 1: Customized stem. View of the stem inside the implant.13
We used custom-made axisymmetric radial head implants in this study in lieu of a commercially available implant so that it would match our custom stem. In this case, the axisymmetric implant was modeled after the Evolve Proline Radial Head System (Wright Medical Technology, Inc, Arlington, TN) and included 20-, 22-, 24-, and 26-mm implant sizes. An experienced upper extremity orthopedic surgeon compared the size of the excised native radial head with the axisymmetric implants as is performed clinically. The minor diameter of the elliptical native radial head was used to select the diameter of the axisymmetric implant. If the radial head was judged to be between sizes, the smaller sized prosthesis was selected. To design a series of population-based, quasianatomic radial head implants, we measured computed tomographic scans of 34 male elbows. These specimens were sorted by maximum diameter into 3 groups representing the specimens within 1 SD of the mean (QM) (n ¼ 24), above 1 SD (Qþ) (n ¼ 5), and below it (Qe) (n ¼ 5). All specimens were within 3 SD of the mean. Implants were then generated for each of these groups by averaging a large number of measured parameters in these specimens.13 The size of the excised native radial head was compared with the population-based elliptical implants. The populationbased implant that most closely matched the maximum diameter of the native radial head was used. If the radial head was judged to be between sizes, the smaller sized prosthesis was selected. We applied the same measurement techniques to the scans of each
MATERIALS AND METHODS Design of implants The radial head implant system designed for this protocol consisted of 2 components: a custom-made generic stem and the radial head (Figs. 1, 2).13 All radial head implants were formed out of acrylonitrile butadiene styrene (ABS) M30 plastic (Stratasys, Eden Prairie, MN) using a rapid prototyping machine with an accuracy of 0.127 mm (400MC, Stratasys). We performed a benchtop study to determine whether the plastic implants would produce a contact area similar to that of traditional metal implants. Three casts were made using a cadaveric humerus with one plastic axisymmetric implant and one metal axisymmetric implant of the same size. J Hand Surg Am.
717
r
Vol. 40, April 2015
718
EFFECT OF RADIAL HEAD SHAPE ON JOINT CONTACT
FIGURE 2: Implants. Example subset of implants that were fabricated.13
maintained in 90 flexion by resting the ulna on a smooth support bar.
individual specimen tested in this protocol to produce 8 reverse-engineered, patient-specific implants.13
Joint contact measurements We quantified radiocapitellar joint contact using a previously reported casting technique (Fig. 4).11,16,17 After mixing, 2 mL silicone impression material (Reprosil Medium Body Vinyl Polysiloxane Impression Material, DENTSPLY International, Inc, York, PA) was injected onto the radial head using a syringe. The radiocapitellar joint was reduced and the LCL actuator was loaded to 20 N to simulate a clinical ligament repair.15 Static (muscle) loading was applied across the elbow to mimic normal muscle tone in pronation, neutral rotation, and supination and a cast was taken in each of these positions. The casting process was first conducted with the native radial head. The radial head was then resected and the custom stem was navigated into position using image-based computer navigation.13 Casting was then repeated for all 3 implant designs in randomized order. The cast was left to cure for 25 minutes before removal. Once the cast had set, the LCL tension was released to allow the joint to be subluxated for cast removal.
Specimen testing Eight fresh-frozen cadaveric specimens (male; average age, 75 8 y; right upper extremity) were mounted in a forearm motion simulator (Fig. 3).14 To gain access to the radial head, the extensor muscles were elevated off the lateral collateral ligament (LCL) and the humeral origin of the LCL was sectioned from the lateral epicondyle. The LCL was repaired using heavy nonabsorbable locking Krakow sutures that were then passed through an isometric drill hole at its anatomic origin and transosseous tunnels.15 The triceps (T) and biceps (B) and the pronator teres (PT) were sutured with a braided suture. The sutures of these muscles and the LCL were then attached to pneumatic actuators (Airpel, Airpot Corp, Norwalk, CT) via a cable. A 20-N constant load was applied to the LCL actuator to simulate a clinical LCL repair.15 Static loads (based on joint angle and muscle load data from a simulated active forearm rotation) were applied to the T, B, and PT to maintain full pronation (T ¼ 50 N, B ¼ 15 N, and PT ¼ 80 N), neutral rotation (T ¼ 50 N, B ¼ 35 N, and PT ¼ 25 N), and full supination (T ¼ 50 N, B ¼ 40 N, and PT ¼ 30 N). The elbows were J Hand Surg Am.
Analysis of casts The lateral side of the cast was marked before it was removed from the radiocapitellar joint. The excised r
Vol. 40, April 2015
EFFECT OF RADIAL HEAD SHAPE ON JOINT CONTACT
719
FIGURE 3: Elbow motion simulator. The forearm is in neutral rotation at 90 flexion in the elbow motion simulator. Static loads were applied to the 3 muscles: biceps (BIC), triceps (TRI), and pronator teres (PT), and also the LCL, using pneumatic actuators.
tracked stylus (Optotrak Certus, NDI, Waterloo, Ontario, Canada) relative to the digitization block (Fig. 5C). The stylus had an accuracy of 0.1 mm.18 The point cloud digitizations were then converted to 3-dimensional surfaces (Fig. 5D, E) using a radial basis function program in MATLAB (MathWorks, Natick, MA). Custom Visualization Toolkit (Kitware, Clifton Park, NY) software was used to measure the contact area (Fig. 5F). To determine the location of the contact area on the radial head, we calculated the centroid of the contact area patch. The radial basis function surfaces that were previously calculated were converted back into points to achieve an evenly distributed point cloud. Custom MATLAB code was used to fit a plane to the points. These were given indices and projected onto the plane such that a 2-dimensional surface was created. The mean of the points was calculated to determine the point closest to the center of the surface. The points were then projected back onto the original surface. The 3-dimensional coordinates of the center point (centroid) were determined. To measure the location of the contact area, the center of the radial head was determined. Twenty points were selected around the rim on the radial head. A circle was then fit to these points and the center and radius of the circle were calculated. The distance between
FIGURE 4: Example of casts. A set of casts, all taken in the neutral position.
native radial head was cemented to the digitization block. Infrared light-emitting diode markers were also attached so that the position of the contact patch would be known. The orientation markers on the native radial head were digitized and the entire radial head was traced. The cast was then placed onto the radial head in the correct orientation such that the lateral marks on the cast and a lateral hole on the radial head (marked before removal) lined up. For the casts of the radial head implants, the implant was removed from the stem with the cast still in situ and then placed on a stem fixed to the digitization block (Fig. 5AeC). A technique described by Lalone et al16 was used to determine the radiocapitellar contact area. This was quantified by digitizing the casts using an optically J Hand Surg Am.
r
Vol. 40, April 2015
720
EFFECT OF RADIAL HEAD SHAPE ON JOINT CONTACT
FIGURE 5: Process of quantifying the contact area. A Space where the silicone impression material was injected. B Cast removed from the joint after it has cured. The void space represents the contact area. C The stylus was used to trace the contact area and was tracked by the Optotrak Certus system. D Point cloud produced from digitization. E Surface made from the point cloud. F Area calculation.
FIGURE 6: Contact area for all radial head conditions. Mean contact area ( 1 SD) is shown for all radial head conditions at all 3 angles of rotation. There was a significant effect of radial head condition on contact area (P ¼ .005). There was no difference in contact area between implants; however, the axisymmetric implant had significantly lower contact area than the native radial head (P ¼ .008).
FIGURE 7: Distance from the center of the radial head to the contact area centroid. Mean distance ( 1 SD) is shown for all radial head conditions. The data were normalized against the radius of the radial head. There was no significant difference in contact location (P ¼ .220).
radial head radius such that the center was 0 and the rim was 1.
the center of the radial head and the contact centroid was compared with the distance from the centroid to the rim of the radial head. This was done to quantify whether the contact patch lay closer to the center or the rim for the different radial head conditions. Values for distances were normalized against the J Hand Surg Am.
Statistical analyses Two-way repeated-measures ANOVA was performed with radial head condition (native radial head, axisymmetric, quasi-anatomic, and patient-specific) r
Vol. 40, April 2015
EFFECT OF RADIAL HEAD SHAPE ON JOINT CONTACT
and rotation angle (full pronation, neutral, and full supination) as the 2 factors to determine whether there was a statistically significant difference in contact area and location. We used a post hoc pairwise comparison (a ¼ .05) to determine potential differences in contact area and location between each implant and the native radial head. This same technique was used to compare the 3 implant morphologies with each other.
40% reduction in the study presented here. In the experimental setup used by Liew et al,11 the radiocapitellar joint was aligned by the experimenter, and concavity-compression was created as loading was applied, whereas our model allowed the muscles, articulation, and ligaments to align the joint while applying muscle tones through actuators. Therefore, the contact area measured in the current study was likely more clinically relevant than the contact area determined from previous benchtop studies. In both studies, the implant materials used were much stiffer than the cartilage of the native radiocapitellar joint. In the current study, the modulus of elasticity of the ABS plastic used to make the implants was 2.3 GPa whereas in the study by Liew et al the implants were made of cobalt chromium with a modulus of 230 GPa. The effect of the different moduli of materials was a potential concern; therefore, we compared the contact area of ABS plastic implants relative to metal implants in a benchtop study.19 Three casts were made using a cadaveric humerus with one ABS plastic axisymmetric implant and one metal axisymmetric implant of the same size. The ABS plastic and metal implants showed no significant differences in contact area (P ¼ .920). It is likely that the high stiffness of the implants and not the implant shape resulted in the general decrease in contact area. Future work is needed to find an implant material that behaves similarly to cartilage. The decrease in contact area seen after radial head arthroplasty was performed could also have resulted from small errors (less than 2-mm translation and less than 11 rotation) during the image-based computer navigation of the stem.13 It is likely that rotational error would have a more noticeable effect on the anatomical implants because the dish of these implants is located eccentrically and there is an optimal position for the dish to ensure that it can track accurately. Because the dish is located centrally on the axisymmetric implant, rotational error would not have had a significant effect. Factors such as translational malpositioning or annular ligament and interosseous membrane tension also may have an influence on radiocapitellar contact area. Further studies are needed to better understand the sensitivity of implant malpositioning on contact area at the radiocapitellar joint. A potential weakness of this study was that we used an LCL repair technique to gain access to the joint. On assessment there was no difference in motion pathways before and after ligament sectioning and repair.19 Therefore, repair of the LCL using this technique restores the normal kinematics of the elbow.
RESULTS This benchtop study to compare the plastic and metal implants showed no significant differences in contact area (P ¼ .92). There was a significant effect of radial head condition (P ¼ .005) (Fig. 6). All of the implants had a lower contact area than the native radial head but only the axisymmetric implant was significantly different (P ¼ .008). There was no significant difference in contact area among the 3 implant shapes (P > .050). There was no significant effect of forearm rotation position on contact area (P ¼ .240). There was no significant difference in contact location for the native radial head, axisymmetric, quasianatomic, or patient-specific implants (P ¼ .220) (Fig. 7). DISCUSSION Previous investigations have shown that the articulating surface of the radial head is variably elliptical and is typically offset from the center of the radial neck.1e3 Differences in radial head implant shape relative to the native articulation were expected to affect both the magnitude and location of contact within the radiocapitellar joint. Three radial head implant shapes, either axisymmetric, population-based quasi-anatomic, or reverse-engineered patient-specific, were implanted and the radiocapitellar contact was quantified in 3 positions of forearm rotation. The results showed no significant differences among the 3 implant morphologies for contact area or contact location. This suggests that the shape of the radial head implant may not be important with respect to altering radiocapitellar joint contact mechanics. A previous study11 reported significantly less contact with axisymmetric metal implants compared with the native radial head. The results from the current study agree with those earlier findings in which the axisymmetric radial head displayed a significantly smaller contact area than the intact radial head. However, the previous authors found an average 60% decrease in contact area with the axisymmetric implant, compared with an average J Hand Surg Am.
721
r
Vol. 40, April 2015
722
EFFECT OF RADIAL HEAD SHAPE ON JOINT CONTACT
ACKNOWLEDGMENT Support was received from the Canadian Institute for Health Research. National Sciences and Engineering Research Council.
This study used a sample size of 8, which has been adequate for various in vitro biomechanical studies completed in our laboratory. A post hoc power analysis determined that although there was sufficient power (0.90) for the contact area analysis with 8 specimens, 15 specimens would be needed to achieve sufficient power for the contact location analysis (observed power, 0.25). The native, quasi-anatomic, and patient-specific radial head implants varied in dish position and shape (from more circular to more elliptical in nature), which may have contributed to the variability in our contact location data and the lack of detection of a significant difference between groups. The radiocapitellar joint experiences concavee convex surface contact mechanics. This relationship between the capitellum and the radial head relies on proper joint alignment such that the concave dish of the radial head centers on the convex surface of the capitellum. A few options exist to ensure that the joint mechanics are restored. Bipolar radial heads contain a joint between the head and stem of the implant, allowing some tilting of the dish.9 Some axisymmetric implants employ a smooth stem that is purposely placed loose so that the head can move small amounts and essentially self-align with the capitellum.7 Anatomic implants are equipped with markings to try to aid in the best possible alignment.10 In this study, all of the implants were fixed in place. Although this was essential for anatomically designed implants and is commonly used for most axisymmetric implants, this process likely does not represent the behavior of axisymmetric implants, which employ a loose-fitting smooth stem, or those that incorporate a bipolar articulation. By cementing the axisymmetric implants we removed their potential self-alignment property. The fact that there were no differences in contact area or location among the 3 implants suggests that the concavity compression of the radiocapitellar articulation together with the annular ligament (which was left intact in the current study) kept the implants aligned, making implant shape less important. To increase joint contact and hence reduce cartilage loading, future work should focus on optimally aligning these implants and perhaps investigating less stiff materials. Further studies are needed to evaluate the importance of implant positioning, which we did not address. Clinical studies are also required to determine whether these contact patterns will influence the outcomes of radial head arthroplasty.
J Hand Surg Am.
REFERENCES 1. Popovic N, Djekic J, Lemaire R, Gillet P. A comparative study between proximal radial morphology and the floating radial head prosthesis. J Shoulder Elbow Surg. 2005;14(4):433e440. 2. Swieszkowski W, Skalski K, Pomianowski S, Kedzior K. The anatomic features of the radial head and their implication for prosthesis design. Clin Biomech (Bristol, Avon.). 2001;16(10):880e887. 3. 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(1):112e116. 4. 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(3):225e230. 5. 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(6):972e978. 6. Calfee R, Madom I, Weiss AP. Radial head arthroplasty. J Hand Surg Am. 2006;31(2):314e321. 7. van Riet RP, Van Glabbeek F, Baumfeld JA, et al. The effect of the orientation of the radial head on the kinematics of the ulnohumeral joint and force transmission through the radiocapitellar joint. Clin Biomech (Bristol, Avon.). 2006;21(6):554e559. 8. 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(6):885e890. 9. Dotzis A, Cochu G, Mabit C, Charissoux JL, Arnaud JP. Comminuted fractures of the radial head treated by the Judet floating radial head prosthesis. J Bone Joint Surg Br. 2006;88(6):760e764. 10. Acumed. Anatomic Radial Head System. Available at: http://www. acumed.net/sites/default/files/literature/brochure-surgical-technique/ ARH_Brochure.pdf. Accessed March 3, 2014. 11. Liew VS, Cooper IC, Ferreira LM, Johnson JA, King GJ. The effect of metallic radial head arthroplasty on radiocapitellar joint contact area. Clin Biomech (Bristol, Avon.). 2003;18(2):115e118. 12. Sabo MT, Shannon H, Ng J, Ferreira LM, Johnson JA, King GJ. The impact of capitellar arthroplasty on elbow contact mechanics: Implications for implant design. Clin Biomech (Bristol, Avon.). 2011;26(5):458e463. 13. Deluce SR. Design, Validation and Navigation of Anatomic Population-Based and Patient-Specific Radial Head Implants [Master’s thesis]. London, Ontario, Canada: Western University; 2011. 14. 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(11):2114e2119. 15. 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(9):615e623. 16. Lalone EA, Peters TM, King GW, Johnson JA. Accuracy assessment of an imaging technique to examine ulnohumeral joint congruency during elbow flexion. Comput Aided Surg. 2012;17(3):142e152. 17. Stormont TJ, An KN, Morrey BF, Chao EY. Elbow joint contact study: comparison of techniques. J Biomech. 1985;18(5):329e336. 18. Nothern Digital, Inc. Optotrak Certus Motion Capture System. Available at: http://www.ndigital.com/lifesciences/certus-techspecs. php. Accessed March 3, 2014. 19. Shannon HL. The Contact Mechanics and Kinematics of Radial Head Implants [Master’s thesis]. London, Ontario, Canada: Western University; 2012.
r
Vol. 40, April 2015
