- Email: [email protected]
Glenosphere dissociation after reverse shoulder arthroplasty
J Shoulder Elbow Surg (2015) 24, 1061-1068
www.elsevier.com/locate/ymse
Glenosphere dissociation after reverse shoulder arthroplasty Michael C. Cusick, MDa, Michael M. Hussey, MDa, Brandon M. Steen, MDa, Robert U. Hartzler, MDa, Rachel E. Clark, BA, CCRPb, Derek J. Cuff, MDc, Andres F. Cabezas, MEb, Brandon G. Santoni, PhDb, Mark A. Frankle, MDa,* a
Florida Orthopaedic Institute, Shoulder and Elbow Service, Tampa, FL, USA Foundation for Orthopaedic Research and Education, Phillip Spiegel Orthopaedic Research Laboratory, Tampa, FL, USA c Suncoast Orthopaedic Surgery & Sports Medicine, Venice, FL, USA b
Background: Reverse shoulder arthroplasty (RSA) is gaining popularity for the treatment of debilitating shoulder disorders. Despite marked improvements in patient satisfaction and function, the RSA complication rate is high. Glenosphere dissociation has been reported and may result from multiple mechanisms. However, few RSA retrieval studies exist. Methods: We reviewed our RSA database and identified patients with glenosphere dissociation between 1999 and 2013. Prosthesis type, glenosphere size, and contributing factors to dissociation were noted. Five retrieved implants were available for analysis, and evidence of wear or corrosion on the Morse taper was documented. Further, we biomechanically investigated improper Morse taper engagement that may occur intraoperatively as a potential cause of acute dissociation. Results: Thirteen patients with glenosphere dissociation were identified (0.5 months to 7 years postoperatively). Glenosphere size distribution was as follows: 32 mm (n ¼ 1), 36 mm (n ¼ 4), 40 mm (n ¼ 6), and 44 mm (n ¼ 2). Incidence of dissociation was correlated to glenosphere size (P < .001). Taper damage was limited to fretting wear, and there was minimal evidence of taper corrosion. Biomechanically, improper taper engagement reduced the torsional capacity of the glenosphere-baseplate interface by 60% from 19.2 1.0 N-m to 7.5 1.5 N-m. Conclusion: We identified several mechanisms contributing to glenosphere dissociation after RSA, including trauma and improper taper engagement. Limited evidence of corrosive wear on the taper interface was identified. Although it is rare, the incidence of glenosphere dissociation was higher when 40- and 44-mm glenospheres were implanted compared with smaller glenospheres (32 and 36 mm), probably because of the larger exposed surface area for potential impingement. Level of evidence: Level IV, Case Series, Treatment Study. Ó 2015 Journal of Shoulder and Elbow Surgery Board of Trustees. Keywords: Reverse shoulder arthroplasty; Morse taper; dissociation
Funding: DJO Surgical supported this work in the form of a research grant to the Foundation for Orthopaedic Research and Education. DJO Surgical did not have input into the study as it was investigator initiated. This study was determined to be exempt from review by the Western Institutional Review Board.
*Reprint requests: Mark A. Frankle, MD, Florida Orthopaedic Institute, 13020 N Telecom Pkwy, Tampa, FL 33637, USA. E-mail address: [email protected] (M.A. Frankle).
1058-2746/$ - see front matter Ó 2015 Journal of Shoulder and Elbow Surgery Board of Trustees. http://dx.doi.org/10.1016/j.jse.2014.12.019
1062 Since Food and Drug Administration approval in 2004, reverse shoulder arthroplasty (RSA) has gained increasing acceptance as a surgical treatment in elderly patients with rotator cuff disease and accompanying shoulder pain and dysfunction.6 Owing to its success in addressing cuff tear arthropathy, indications for RSA are quickly expanding to include massive cuff tears without glenohumeral arthritis, proximal humerus fractures, and as a salvage procedure for failed hemiarthroplasty or anatomic shoulder replacement.5,8,18,25 Despite multiple reports that RSA alleviates pain while improving patient satisfaction and shoulder function, the procedure is associated with a relatively high complication rate, with literature reports being variable and ranging upward of 20% for primary surgeries to >40% for revision procedures.22 Reported postoperative complications include but are not limited to instability, acromial fracture, periprosthetic fracture, glenoid component loosening, and scapular notching.4,22,25 A design feature common to all commercially available RSA implant systems is a baseplate component with a Morse taper feature onto which the glenosphere is seated intraoperatively. A modular connection placed in a biologic environment with repetitive exposure to mechanical stresses during functional activities renders the interface subject to in-service damage that may be manifested as mechanical wear, corrosion, chemical degradation, fatigue, and mechanical overload. These damage mechanisms may potentiate glenoid component dissociation, which, although rare, has been reported as a postoperative complication in the shoulder arthroplasty literature.7,16,17 Whereas retrieval studies in the total hip and knee replacement literature have cited mechanical wear and corrosion as predominant and contributing agents to revision surgery,10,12,15 we are unaware of any RSA retrieval studies. Therefore, the goal of our study was to present our case series of acute and chronic RSA prosthesis dissociations and to discuss our examination of the Morse taper interface retrieved at the time of revision surgery. The mechanisms responsible for acute and chronic dissociation vary and are likely multifactorial. We explore, through biomechanical testing, improper Morse taper engagement and mechanical compromise of the glenosphere-baseplate interface as an explanation for acute prosthesis dissociation.
Materials and methods We retrospectively reviewed our RSA patient database and identified patients with glenosphere dissociation between May 1999 and July 2013. Operative notes were reviewed, confirming the type of prosthesis and implant sizes. A thorough chart review was performed to determine timing of the dissociation and the presence or absence of a traumatic event. Patients who met the inclusion criteria had plain film imaging demonstrating a reverse prosthesis with baseplate-glenosphere dissociation. Patients who had their index procedure performed at an outside institution or implant designs that were used before Food and Drug
M.C. Cusick et al. Table I Modified Goldberg Score10 for visual assessment of taper junction in retrieved RSA implants Damage
Score
Criteria
Minimal
1
Mild
2
Moderate
3
Severe
4
Fretting on <10% of the surface and no corrosion damage Fretting on >10% of the surface and/or corrosion attack confined to one or more small areas Fretting on >30% of the surface and/or aggressive local corrosion attack with corrosion debris Damage over the majority (>50%) of the surface with severe corrosion attack and abundant corrosion debris
Administration approval of the device were excluded from statistical analysis. Basic statistical analysis was performed by Fisher exact test to determine if any correlation exists between incidence of dissociation and glenosphere size.
Retrieval analysis The taper interfaces were evaluated with gross observation and macrophotography. For semiquantitative assessment, we used the Modified Goldberg Score,10 in which the taper junction was assessed and given a score of 1 to 4 based on the extent of fretting wear and corrosion observed (Table I).
Biomechanical study We measured the effect of improper Morse taper engagement using a series of axial distraction and torsional experiments (Fig. 1) of glenosphere-baseplate constructs (RSP; DJO Global, Vista, CA, USA) with simulated impingement between the glenosphere (size ¼ 36 mm neutral) and underlying bone. Sawbones test blocks (Model #1522; Pacific Research Laboratories, Vashon, WA, USA) were used to simulate the glenoid. The baseplate was assembled into the block in accordance with the manufacturer’s instructions and instrumentation kits in combination with four 22mm-long peripheral locked screws. Pilot experiments were undertaken to create a repeatable impingement test model (Fig. 2). Because of its circular design, our laboratory setup was modeled to replicate the clinical scenario of impingement between the underside of the glenosphere and underlying glenoid that may occur superiorly or inferiorly to the periphery of the RSP baseplate (Fig. 3). A single 22-mm-long locking screw (DJO Global) was inserted into the test block immediately adjacent to the rim of the circular baseplate. We chose to simulate 3 conditions of impingement (mild, moderate, and severe), which we equated to the number of screw threads visible and protruding from the top of the Sawbones test block. Specifically, 3 visible threads served as the severe impingement model, whereas a single visible thread was considered the mild impingement condition in our test setup. Initially, the screw was inserted such that all 3 locking threads were protruding from the top of the test block (severe impingement). Visually, in this condition, the top of the screw head was flush with the top side of the baseplate. Thereafter, proper seating of a 36mm neutral glenosphere onto the baseplate’s taper was attempted.
Glenosphere dissociation after RSA
Figure 1
1063
Biomechanical setups for (A) axial distraction and (B) torsional tests.
Figure 2 Illustrations of the control (A) and simulated impingement conditions (B, C) evaluated biomechanically. The simulated impingement served to prevent full taper engagement (C). After impaction, the integrity of the construct was manually evaluated and found to be mechanically unstable as the glenosphere dissociated from the baseplate with minimal load application. Resultantly, this condition was eliminated from consideration as our experimental impingement model. Thereafter, the locked screw was advanced such that 2 threads were visibly protruding from the test block surface. Again, the glenosphere was impacted and the construct was manually manipulated by the senior author and found to be mechanically secure, although visually it was apparent that full taper engagement was not achieved as the underside of the glenosphere was not parallel to the top surface of the test block (Fig. 2, B and C). The glenosphere was removed and the locked screw was advanced such that a single thread was visible above the test block surface. The glenosphere was impacted, and the senior author (M.A.F.) confirmed that the construct was mechanically secure. However, in the mild impingement model, the base of the glenosphere appeared parallel to the top surface of the test block, indicating that full taper engagement may have been achieved despite a single screw thread still protruding slightly from the block. Based on the visual perception that impingement was not achieved, we adopted the moderate impingement condition as our experimental impingement condition for all subsequent experiments. The senior author and 3 shoulder and elbow fellows performed all biomechanical experiments to evaluate the effect of simulated impingement on the mechanical integrity of the Morse taper interface between the glenosphere and baseplate of the RSP system. Test specimens were evaluated with and without simulated moderate impingement, and for each condition, the glenospherebaseplate construct was manually tested in axial distraction and torsion in succession. Each individual performed the biomechanical experiment with and without impingement once each, and all
tests were performed with the same baseplate and glenosphere combination. The constructs were mounted in a vise clamp (Fig. 1, A and B) and rigidly coupled to a 6-axis load cell (MC3A-1000; Advanced Medical Technologies, Inc., Watertown, MA, USA), which recorded force (N) and moment (N-m) data continuously during axial distraction and torsional testing at 30 Hz. For the axial distraction test, the surgeon was instructed to attempt to disengage the taper using a custom-designed T-handle that was threaded into the glenosphere. Further, the surgeon was instructed to load the construct to a maximum applied force of up to 400 N using visual feedback from the load cell’s output. This force, although likely larger than the loads applied during implant placement to gauge proper taper impaction and stability, is considerably less than literature-based reports of dissociation force magnitudes reported by Blevins et al.2 If the glenosphere withstood axial force application without dissociating from the taper, torsional loading of the glenosphere commenced. The surgeon was instructed to torque the glenosphere up to a maximum of 20 N-m, again using visual feedback from the load cell’s output, as torque magnitudes above this value were determined in pilot experiments to fracture the Sawbones test block with a properly seated glenosphere. Maximum applied axial load (N) and torsional moment (N-m) were recorded for the control (no impingement) and impingement conditions.
Results During the 14-year study period, we confirmed 13 glenosphere dissociations radiographically. Eleven of the dissociations occurred in patients operated on by the senior
1064
M.C. Cusick et al.
Figure 3 Sagittal view of a synthetic model scapula implanted with the RSP baseplate illustrating that after placement, there is exposed glenoid both superiorly and inferiorly that may contribute to glenosphere impingement and compromised mechanical integrity.
author of a total of 1800 reverse shoulder replacements performed with the RSP system in the study period, for an overall glenosphere-baseplate dissociation rate of <1%. Two additional dissociations were confirmed in patients initially treated at an outside institution. Both of these RSAs were performed with the Trabecular Metal Reverse Shoulder System (Zimmer, Warsaw, IN, USA) prosthesis. Glenosphere sizes identified included 32 mm (n ¼ 1, DJO), 36 mm (n ¼ 2, DJO; n ¼ 2, Zimmer), 40 mm (n ¼ 6, DJO), and 44 mm (n ¼ 2, DJO). Timing of dissociation ranged from <1 month after surgery to 7 years after surgery (Table II). Because of differences in the design features of the DJO and Zimmer RSA prostheses, which include but are not limited to the addition of a glenosphere locking screw in the DJO design that is lacking in the Zimmer prosthesis, and differences in taper geometry (length and angle) between both implants, we investigated whether glenosphere size correlated to the rate of dissociation for the DJO design only. This analysis was also performed because the DJO design constituted 85% (11 of 13) of our retrieved implants. The incidence of dissociation by glenosphere size for the DJO retrievals was 0.09% (32 mm), 0.25% (36 mm), 5.4%
(40 mm), and 8.3% (44 mm), and this correlation was significant (P < .001). In addition, glenospheres were divided into a small group (sphere sizes 32 mm and 36 mm) and a large group (sphere sizes 40 mm and 44 mm) on the basis of surface area overhang relative to the baseplate. The incidence of dissociation in the larger glenosphere group (2%) was significantly greater (P < .001) compared with the smaller glenosphere group (0.1%). Review of patient records revealed 1 incident of dissociation with a 32-mm sphere. The patient was an 82-yearold man who was treated with a primary RSA for cuff tear arthropathy. The dissociation occurred 27 months after surgery and was caused by trauma in the form of a fall on an adducted arm while playing tennis. At the time of revision surgery, there appeared to be an impacted area on the medial aspect of the polyethylene socket. In 4 cases of dissociation with a size 36-mm sphere, no trauma was identified. Time to dissociation in these cases ranged from 7 months to 7 years. Two cases involved a Zimmer prosthesis and 2 cases were with a DJO prosthesis, 1 of which was custom and performed in 1999. Dissociation of this device occurred 7 years after the index procedure in an 82-year-old woman who had undergone revision
Glenosphere dissociation after RSA
1065
Table II
Summary of dissociation findings
RSA make
Gender
Age at time of dissociation
Glenosphere size (mm)
Time to dissociation
Implant available for retrieval analysis?
Modified Goldberg Score of retrieval
Primary reason (if known) for dissociation
DJO DJO DJO
Male Female Female
83 82 78
32-4 36-4 36-4
27 months 86 months 28 months
Y N N
1 N/A N/A
DJO DJO DJO DJO DJO DJO DJO DJO Zimmer Zimmer
Female Female Male Male Female Female Male Male Male Male
48 72 64 75 74 69 69 60 75 82
40 40 40 40 40 40 44 44 36 36
2 weeks 48 months 8 months 7 months 5 months 5 months 6 months 3 months 24 months 7 months
N N Y Y N N N N Y Y
N/A N/A 1 1 N/A N/A N/A N/A 2 1
Fell while playing tennis Lifted granddaughter Overexerted shoulder while on crutches after foot surgery Not reported but atraumatic Not reported but atraumatic Not reported but atraumatic Not reported but atraumatic Not reported but atraumatic Not reported but atraumatic Fell out of bed Not reported but atraumatic Not reported but atraumatic Not reported but atraumatic
to RSA for failed hemiarthroplasty. Dissociation occurred when the patient lifted her grandchild. Review of immediate postoperative films revealed that the glenosphere was not fully impacted on the Morse taper, perhaps resulting in mechanical compromise of glenosphere fixation. The other DJO implant in this group failed more than 2 years after implantation. This patient had recent foot surgery and was noted to be ambulating with heavy use of her upper extremities before dissociation, but no specific traumatic episode was noted. One Zimmer prosthesis dissociated at 2 years after surgery; the other dissociated twice, once from trauma 6 weeks after surgery and a second time 7 months after the initial revision (which prompted referral to our office). Both cases were revised with use of the same size sphere and had no further complications. There were 6 size 40-mm spheres whose time to dissociation varied from 2 weeks to 4 years. No trauma was identified in any case. One case involved another custom prosthesis (late 1990s) that failed identically to the method previously described. Two cases of size 44-mm sphere dissociation were identified. One patient was a Vietnam War veteran and carried a diagnosis of post-traumatic stress disorder. Six months after surgery, the patient fell out of bed while sleeping and developed acute-onset shoulder pain and dysfunction. The second was in a case of documented recurrent instability and infection in a failed RSA. Glenosphere dissociation in this patient occurred 3 months after the index procedure.
Retrieval analysis Five of our 13 dissociated glenospheres were available for analysis (Table II). We attribute the loss of 8 RSA retrieved implants to the retrospective nature of this study and, until
Figure 4 Image of a retrieved taper illustrating evidence of fretting wear and highly localized corrosion.
recently, the lack of an implant storage facility as well as a dedicated implant retrieval and analysis laboratory. Time to dissociation of the retrievals ranged from 7 to 27 months. Qualitatively, visual inspection of all implants revealed minor evidence of fretting wear regardless of the time of revision surgery. One instance of highly localized corrosion in combination with fretting wear was identified on 1 implant (Fig. 4). No evidence of permanent deformation of the taper was noted on any implant. Based on the Modified Goldberg Score,10 4 of 5 retrieved implants received a score of 1, with only 1 taper surface receiving a score of 2
1066 Table III
M.C. Cusick et al. Raw data from biomechanical experiments
Test condition Trial 1 2 3 4 Average Standard deviation
Without impingement
Simulated impingement
Distraction force (N)
Applied torque (N-m)
Distraction force (N)
Applied torque (N-m)
390 455 407 358 402.5 40.5
20.5 19.0 18.0 19.2 19.2 1.0
345 480 352 386 390.8 62.1
9.3 5.9 8.0 7.0 7.5 1.5
(Table II). No implant received a score of higher than 2 based on this scoring system.
Biomechanical results Maximum measured axial force (N) and torsion (N-m) were recorded for both phases of testing. Our biomechanical results indicated that our laboratory model of impingement that resulted in less than optimal taper engagement did not negatively affect the mechanical integrity of the taper when axial distraction forces were applied to the glenosphere (Table III). Specifically, the average axial distraction forces for the control and simulated impingement conditions were 402.5 40.5 N and 390.8 62.1 N, respectively, and no glenosphere in either test condition dissociated from the underlying baseplate. However, torsional testing revealed that simulated interference and less than optimal taper engagement negatively affected the mechanical integrity of the glenospherebaseplate interface. Specifically, each construct tested without impingement withstood the set maximum applied torque of 20 N-m without the interface failing (19.2 1.0 N-m). However, in the impingement group, dissociation failure was observed in all test constructs at an average applied torque of 7.5 1.5 N-m. Thus, simulated impingement reduced the biomechanical integrity of the taper interface by approximately 60%.
Discussion Every reverse arthroplasty system has a baseplate coupled to a sphere by a Morse taper. Morse tapers are not standardized in the orthopedic industry and vary in design between implant manufacturers, notably with regard to taper angles that generally fall in the range of 5 to 18 .11 In addition, RSA prostheses differ regarding whether, after proper taper engagement, the glenosphere sits flush with the underlying glenoid or is elevated slightly from the bone surface. An advantage of the latter is that a fully seated taper is less likely to impinge on surrounding glenoid soft tissue or bone. A potential disadvantage is the greater possibility of fluid contamination of the taper itself, which can lead to corrosive
wear of the taper. Advantages of RSA designs in which the glenosphere sits flush with the glenoid include load sharing between the taper and the surrounding flush rim of glenoid bone. In addition, it may be less likely to be subjected to corrosive wear. The disadvantage of this design is that it is subject to ‘‘impingement’’ or improper taper impaction that puts the device at risk for early mechanical failure. In addition, some systems have different methods of engaging the glenosphere on the glenoid baseplate. Some have an instrument to aid in the engagement of the glenosphere to the baseplate, whereas others use manual placement. Ultimately, the surgeon should be aware of advantages and disadvantages of the taper and design features of his or her preferred system. In the system biomechanically tested in this study (RSP, DJO Global), the glenosphere is manufactured with a threaded hole that allows attachment of a T-handle to aid glenosphere insertion. The design rationale for this threaded insertion handle (as tested mechanically in this experiment) assumes that a manual distraction force can be applied during surgery with the threaded handle to ensure proper taper impaction and stability, thus providing additional evidence of full taper engagement. Unfortunately, based on our biomechanical testing, it is likely that a false taper was perceived as fully engaged intraoperatively. Despite the biomechanical data reported here that confirm Morse tapers are weaker in torsion than in axial distraction in conditions in which the taper is improperly seated, we are unable to conclude that torsional testing will improve our intraoperative ability to avoid dissociation. This is because the laboratory-measured torques required to dissociate falsely engaged tapers (7 N-m) are likely not practical to apply intraoperatively without markedly increasing the risk of fracturing the patient’s glenoid or scapula. Nevertheless, our current surgical technique includes the application of both manual axial distraction and torsion in an attempt to avoid early dissociation. Multiple factors are thought to contribute to possible dissociation of Morse tapers, and much of this information is derived from modular hip prosthesis designs in which Morse tapers are an integral design component.1,3,10,14,19-21,23,24 Improper taper impaction, contamination of the taper socket with fluid or solid material, improper taper design or manufacturing, micromotion-inducing rotational forces
Glenosphere dissociation after RSA across the joint, and material failure by corrosion have been shown to negatively affect Morse tapers. Glenosphere dissociation after documented trauma appears to be a feasible mechanism. To our knowledge, no studies have evaluated the force necessary to dissociate a glenosphere in vivo. Biomechanical axial pullout tests of the Morse taper in 1 implant show that forces greater than 4448 N are required to dissociate the taper. Biomechanical evaluation of modular total shoulder humeral components required forces of around 3000 N to dissociate the humeral head.2 In the absence of trauma, we speculate that early glenosphere dissociation (<6 months) is likely due to technical error, which can take multiple forms and negatively affect taper strength, including low impaction force,20 inadvertent fluid contamination of the taper,2,13,20 and impingement between the underside of the glenosphere and glenoid as demonstrated here in our biomechanical model. Blevins et al2 reported that as little as 0.4 mL of water or blood could prevent a Morse taper from engaging properly. These authors also demonstrated that if the taper socket is contaminated with water or blood during impaction, the force required to dissociate the taper decreased by 24.3% and 76.9%, respectively. Every attempt should be made to dry the taper at the time of glenosphere placement. More than two thirds of our dissociations involved glenospheres size 40 mm or larger. We believe impingement may be responsible for these early failures in larger sphere sizes. Soft tissue or bone impingement due to inadequate glenoid exposure or preparation may lead to an improperly impacted taper or ‘‘impaction default.’’ When the source of impingement is later removed in vivo, the taper interface becomes loose and can lead to dissociation. Therefore, full glenoid exposure with 360 of soft tissue release in combination with adequate glenoid reaming is essential to mitigate the potential for impingement. Adequate glenoid exposure can be difficult to attain in revision cases or in stiff shoulders from preexisting deformity or advanced disease processes. Late dissociation (>6 months) may be due to material properties of the implant or implant design. Corrosion at the Morse taper has been postulated as a potential source for weakening the connection between modular implants.9,10,14 The baseplate in our reported cases is titanium articulating with a cobalt chrome glenosphere. Corrosion between titanium and cobalt chrome taper articulations has been previously reported in the literature.9,26 Gilbert et al9 reported that approximately 16% to 35% of 148 retrieved total hip implants (titanium femoral stem and cobalt-alloy femoral head) showed signs of moderate to severe corrosive attack in the head-neck taper connection. This corrosion may lead to relative motion (fretting) at modular connections and ultimately to abrasive loss of the passivating oxide layer between the implants, leading to poor mechanical stability of the connection. Gross evaluation of our late dissociation (7 years) taper appeared to show some evidence of fretting along the implant. However, in our
1067 limited sample of retrievals, we identified minor and limited evidence of taper corrosion. This observation of prosthesis damage in the form of fretting wear without significant evidence of metal corrosion is inconsistent with prior literature describing total hip and knee retrievals, in which corrosion is commonly seen.10,12,14,26 However, our observations are similar to those of Day and Abboud (personal communication, 2014), who report observing fretting damage with little or no corrosion on 18 RSA taper retrievals from multiple manufacturers.
Conclusion In our series, we identified several potential contributing mechanisms to the postoperative complication of glenosphere-baseplate dissociation after RSA, including trauma and improper taper engagement. Although it is a rare complication after RSA, we identified a significantly higher likelihood of glenosphere dissociation when larger diameter glenospheres (40 and 44 mm) were implanted, likely attributable to the larger exposed surface area for potential soft tissue or bone impingement. Finally, in contrast to retrieval studies of modular total hip arthroplasty designs in which taper corrosion is commonly observed, we identified limited evidence of corrosive wear at the Morse taper interface in the RSA prosthesis designs analyzed in our series, and the primary evidence of implant damage was in the form of mild fretting wear.
Acknowledgment The authors would like to acknowledge Frederick A. Matsen III, MD, for bringing this problem to our attention and stimulating the completion of this project. We would also like to acknowledge Kathleen O’Brien, BS, and Ben Cottrell, BS, for assistance with manuscript preparation.
Disclaimer The research foundation with which Rachel E. Clark is affiliated received research support from DJO Surgical, a designer and manufacturer of orthopedic surgical products related to the subject of this work. Derek J. Cuff receives consulting fees from and is a member of the speaker’s bureau for DJO Surgical and Arthrex, both designers and manufacturers of orthopedic surgical products related to the subject of this work. The research foundation with which Brandon G. Santoni is affiliated received research support from DJO
1068 Surgical, a designer and manufacturer of orthopedic surgical products related to the subject of this work. Mark A. Frankle receives royalties from and is a paid consultant for DJO Surgical, a designer and manufacturer of orthopedic surgical products related to the subject of this work. 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.
M.C. Cusick et al.
13.
14.
15.
16.
References 17. 1. Atwood SA, Patten EW, Bozic KJ, Pruitt LA, Ries MD. Corrosioninduced fracture of a double-modular hip prosthesis: a case report. J Bone Joint Surg Am 2010;92:1522-5. http://dx.doi.org/10.2106/JBJS. I.00980 2. Blevins FT, Deng X, Torzilli PA, Dines D, Warren RF. Dissociation of modular humeral head components: a biomechanical and implant retrieval study. J Shoulder Elbow Surg 1997;6:113-24. 3. Bobyn JD, Tanzer M, Krygier JJ, Dujovne AR, Brooks CE. Concerns with modularity in total hip arthroplasty. Clin Orthop Relat Res 1994: 27-36. 4. Cheung E, Willis M, Walker M, Clark R, Frankle MA. Complications in reverse total shoulder arthroplasty. J Am Acad Orthop Surg 2011; 19:439-49. 5. Cuff DJ, Pupello DR. Comparison of hemiarthroplasty and reverse shoulder arthroplasty for the treatment of proximal humeral fractures in elderly patients. J Bone Joint Surg Am 2013;95:2050-5. http://dx. doi.org/10.2106/JBJS.L.01637 6. Drake GN, O’Connor DP, Edwards TB. Indications for reverse total shoulder arthroplasty in rotator cuff disease. Clin Orthop Relat Res 2010;468:1526-33. http://dx.doi.org/10.1007/s11999-009-1188-9 7. Farshad M, Gerber C. Reverse total shoulder arthroplastydfrom the most to the least common complication. Int Orthop 2010;34:1075-82. http://dx.doi.org/10.1007/s00264-010-1125-2 8. Flury MP, Frey P, Goldhahn J, Schwyzer HK, Simmen BR. Reverse shoulder arthroplasty as a salvage procedure for failed conventional shoulder replacement due to cuff failuredmidterm results. Int Orthop 2011;35:53-60. http://dx.doi.org/10.1007/s00264-010-0990-z 9. Gilbert JL, Buckley CA, Jacobs JJ. In vivo corrosion of modular hip prosthesis components in mixed and similar metal combinations. The effect of crevice, stress, motion, and alloy coupling. J Biomed Mater Res 1993;27:1533-44. 10. Goldberg JR, Gilbert JL, Jacobs JJ, Bauer TW, Paprosky W, Leurgans S. A multicenter retrieval study of the taper interfaces of modular hip prostheses. Clin Orthop Relat Res 2002:149-61. 11. Hernigou P, Queinnec S, Flouzat Lachaniette CH. One hundred and fifty years of history of the Morse taper: from Stephen A. Morse in 1864 to complications related to modularity in hip arthroplasty. Int Orthop 2013;37:2081-8. http://dx.doi.org/10.1007/s00264-0131927-0 12. Higgs GB, Hanzlik JA, MacDonald DW, Gilbert JL, Rimnac CM, Kurtz SM, et al. Is increased modularity associated with increased
18.
19.
20.
21.
22.
23.
24.
25.
26.
fretting and corrosion damage in metal-on-metal total hip arthroplasty devices?: a retrieval study. J Arthroplasty 2013;28:2-6. http://dx.doi. org/10.1016/j.arth.2013.05.040 Jauch SY, Huber G, Hoenig E, Baxmann M, Grupp TM, Morlock MM. Influence of material coupling and assembly condition on the magnitude of micromotion at the stem-neck interface of a modular hip endoprosthesis. J Biomech 2011;44:1747-51. http://dx.doi.org/10. 1016/j.jbiomech.2011.04.007 Kop AM, Swarts E. Corrosion of a hip stem with a modular neck taper junction: a retrieval study of 16 cases. J Arthroplasty 2009;24:1019-23. http://dx.doi.org/10.1016/j.arth.2008.09.009 Kurtz SM, Kocagoz SB, Hanzlik JA, Underwood RJ, Gilbert JL, MacDonald DW, et al. Do ceramic femoral heads reduce taper fretting corrosion in hip arthroplasty? A retrieval study. Clin Orthop Relat Res 2013;471:3270-82. http://dx.doi.org/10.1007/s11999-013-3096-2 Middernacht B, De Wilde L, Mole D, Favard L, Debeer P. Glenosphere disengagement: a potentially serious default in reverse shoulder surgery. Clin Orthop Reatl Res 2008;466:892-8. http://dx.doi.org/10. 1007/s11999-007-0090-6 Mole D, Navez P, Garaud P. Reverse shoulder prosthesis: problems related to the glenoid. In: Walch G, Boileau P, Mole D, Favard L, Levigne C, Sivreau F, editors. Reverse shoulder arthroplasty: clinical results, complications, revision. Montpellier, France: Sauramps Medical; 2006. p. 289-301. Mulieri P, Dunning P, Klein S, Pupello D, Frankle M. Reverse shoulder arthroplasty for the treatment of irreparable rotator cuff tear without glenohumeral arthritis. J Bone Joint Surg Am 2010;92:254456. http://dx.doi.org/10.2106/JBJS.I.00912 Pansard E, Fouilleron N, Dereudre G, Migaud H, Girard J. Severe corrosion after malpositioning of a metallic head over the Morse taper of a cementless hip arthroplasty. A case report. Orthop Traumatol Surg Res 2012;98:247-50. http://dx.doi.org/10.1016/j.otsr.2011.05.018 Pennock AT, Schmidt AH, Bourgeault CA. Morse-type tapers: factors that may influence taper strength during total hip arthroplasty. J Arthroplasty 2002;17:773-8. http://dx.doi.org/10.1054/arth.2002. 33565 Rehmer A, Bishop NE, Morlock MM. Influence of assembly procedure and material combination on the strength of the taper connection at the head-neck junction of modular hip endoprostheses. Clin Biomech 2012;27:77-83. http://dx.doi.org/10.1016/j.clinbiomech.2011. 08.002 Saltzman BM, Chalmers PN, Gupta AK, Romeo AA, Nicholson GP. Complication rates comparing primary with revision reverse total shoulder arthroplsty. J Shoulder Elbow Surg 2014;23:1647-54. http:// dx.doi.org/10.1016/j.jse.2014.04.015 Shareef N, Levine D. Effect of manufacturing tolerances on the micromotion at the Morse taper interface in modular hip implants using the finite element technique. Biomaterials 1996;17:623-30. Unnanuntana A, Chen DX, Unnanuntana A, Wright TM. Trunnion fracture of the anatomic medullary locking A Plus femoral component. J Arthroplasty 2011;26:504.e13-e16. http://dx.doi.org/10.1016/j.arth. 2010.03.028 Walker M, Willis MP, Brooks JP, Pupello D, Mulieri PJ, Frankle MA. The use of the reverse shoulder arthroplasty for treatment of failed total shoulder arthroplasty. J Shoulder Elbow Surg 2012;21:514-22. http://dx.doi.org/10.1016/j.jse.2011.03.006 Witt F, Bosker BH, Bishop NE, Ettema HB, Verheyen CC, Morlock MM. The relation between titanium taper corrosion and cobalt-chromium bearing wear in large-head metal-on-metal total hip prostheses: a retrieval study. J Bone Joint Surg Am 2014;96:e157. http://dx.doi.org/10.2106/JBJS.M.01199
www.elsevier.com/locate/ymse
Glenosphere dissociation after reverse shoulder arthroplasty Michael C. Cusick, MDa, Michael M. Hussey, MDa, Brandon M. Steen, MDa, Robert U. Hartzler, MDa, Rachel E. Clark, BA, CCRPb, Derek J. Cuff, MDc, Andres F. Cabezas, MEb, Brandon G. Santoni, PhDb, Mark A. Frankle, MDa,* a
Florida Orthopaedic Institute, Shoulder and Elbow Service, Tampa, FL, USA Foundation for Orthopaedic Research and Education, Phillip Spiegel Orthopaedic Research Laboratory, Tampa, FL, USA c Suncoast Orthopaedic Surgery & Sports Medicine, Venice, FL, USA b
Background: Reverse shoulder arthroplasty (RSA) is gaining popularity for the treatment of debilitating shoulder disorders. Despite marked improvements in patient satisfaction and function, the RSA complication rate is high. Glenosphere dissociation has been reported and may result from multiple mechanisms. However, few RSA retrieval studies exist. Methods: We reviewed our RSA database and identified patients with glenosphere dissociation between 1999 and 2013. Prosthesis type, glenosphere size, and contributing factors to dissociation were noted. Five retrieved implants were available for analysis, and evidence of wear or corrosion on the Morse taper was documented. Further, we biomechanically investigated improper Morse taper engagement that may occur intraoperatively as a potential cause of acute dissociation. Results: Thirteen patients with glenosphere dissociation were identified (0.5 months to 7 years postoperatively). Glenosphere size distribution was as follows: 32 mm (n ¼ 1), 36 mm (n ¼ 4), 40 mm (n ¼ 6), and 44 mm (n ¼ 2). Incidence of dissociation was correlated to glenosphere size (P < .001). Taper damage was limited to fretting wear, and there was minimal evidence of taper corrosion. Biomechanically, improper taper engagement reduced the torsional capacity of the glenosphere-baseplate interface by 60% from 19.2 1.0 N-m to 7.5 1.5 N-m. Conclusion: We identified several mechanisms contributing to glenosphere dissociation after RSA, including trauma and improper taper engagement. Limited evidence of corrosive wear on the taper interface was identified. Although it is rare, the incidence of glenosphere dissociation was higher when 40- and 44-mm glenospheres were implanted compared with smaller glenospheres (32 and 36 mm), probably because of the larger exposed surface area for potential impingement. Level of evidence: Level IV, Case Series, Treatment Study. Ó 2015 Journal of Shoulder and Elbow Surgery Board of Trustees. Keywords: Reverse shoulder arthroplasty; Morse taper; dissociation
Funding: DJO Surgical supported this work in the form of a research grant to the Foundation for Orthopaedic Research and Education. DJO Surgical did not have input into the study as it was investigator initiated. This study was determined to be exempt from review by the Western Institutional Review Board.
*Reprint requests: Mark A. Frankle, MD, Florida Orthopaedic Institute, 13020 N Telecom Pkwy, Tampa, FL 33637, USA. E-mail address: [email protected] (M.A. Frankle).
1058-2746/$ - see front matter Ó 2015 Journal of Shoulder and Elbow Surgery Board of Trustees. http://dx.doi.org/10.1016/j.jse.2014.12.019
1062 Since Food and Drug Administration approval in 2004, reverse shoulder arthroplasty (RSA) has gained increasing acceptance as a surgical treatment in elderly patients with rotator cuff disease and accompanying shoulder pain and dysfunction.6 Owing to its success in addressing cuff tear arthropathy, indications for RSA are quickly expanding to include massive cuff tears without glenohumeral arthritis, proximal humerus fractures, and as a salvage procedure for failed hemiarthroplasty or anatomic shoulder replacement.5,8,18,25 Despite multiple reports that RSA alleviates pain while improving patient satisfaction and shoulder function, the procedure is associated with a relatively high complication rate, with literature reports being variable and ranging upward of 20% for primary surgeries to >40% for revision procedures.22 Reported postoperative complications include but are not limited to instability, acromial fracture, periprosthetic fracture, glenoid component loosening, and scapular notching.4,22,25 A design feature common to all commercially available RSA implant systems is a baseplate component with a Morse taper feature onto which the glenosphere is seated intraoperatively. A modular connection placed in a biologic environment with repetitive exposure to mechanical stresses during functional activities renders the interface subject to in-service damage that may be manifested as mechanical wear, corrosion, chemical degradation, fatigue, and mechanical overload. These damage mechanisms may potentiate glenoid component dissociation, which, although rare, has been reported as a postoperative complication in the shoulder arthroplasty literature.7,16,17 Whereas retrieval studies in the total hip and knee replacement literature have cited mechanical wear and corrosion as predominant and contributing agents to revision surgery,10,12,15 we are unaware of any RSA retrieval studies. Therefore, the goal of our study was to present our case series of acute and chronic RSA prosthesis dissociations and to discuss our examination of the Morse taper interface retrieved at the time of revision surgery. The mechanisms responsible for acute and chronic dissociation vary and are likely multifactorial. We explore, through biomechanical testing, improper Morse taper engagement and mechanical compromise of the glenosphere-baseplate interface as an explanation for acute prosthesis dissociation.
Materials and methods We retrospectively reviewed our RSA patient database and identified patients with glenosphere dissociation between May 1999 and July 2013. Operative notes were reviewed, confirming the type of prosthesis and implant sizes. A thorough chart review was performed to determine timing of the dissociation and the presence or absence of a traumatic event. Patients who met the inclusion criteria had plain film imaging demonstrating a reverse prosthesis with baseplate-glenosphere dissociation. Patients who had their index procedure performed at an outside institution or implant designs that were used before Food and Drug
M.C. Cusick et al. Table I Modified Goldberg Score10 for visual assessment of taper junction in retrieved RSA implants Damage
Score
Criteria
Minimal
1
Mild
2
Moderate
3
Severe
4
Fretting on <10% of the surface and no corrosion damage Fretting on >10% of the surface and/or corrosion attack confined to one or more small areas Fretting on >30% of the surface and/or aggressive local corrosion attack with corrosion debris Damage over the majority (>50%) of the surface with severe corrosion attack and abundant corrosion debris
Administration approval of the device were excluded from statistical analysis. Basic statistical analysis was performed by Fisher exact test to determine if any correlation exists between incidence of dissociation and glenosphere size.
Retrieval analysis The taper interfaces were evaluated with gross observation and macrophotography. For semiquantitative assessment, we used the Modified Goldberg Score,10 in which the taper junction was assessed and given a score of 1 to 4 based on the extent of fretting wear and corrosion observed (Table I).
Biomechanical study We measured the effect of improper Morse taper engagement using a series of axial distraction and torsional experiments (Fig. 1) of glenosphere-baseplate constructs (RSP; DJO Global, Vista, CA, USA) with simulated impingement between the glenosphere (size ¼ 36 mm neutral) and underlying bone. Sawbones test blocks (Model #1522; Pacific Research Laboratories, Vashon, WA, USA) were used to simulate the glenoid. The baseplate was assembled into the block in accordance with the manufacturer’s instructions and instrumentation kits in combination with four 22mm-long peripheral locked screws. Pilot experiments were undertaken to create a repeatable impingement test model (Fig. 2). Because of its circular design, our laboratory setup was modeled to replicate the clinical scenario of impingement between the underside of the glenosphere and underlying glenoid that may occur superiorly or inferiorly to the periphery of the RSP baseplate (Fig. 3). A single 22-mm-long locking screw (DJO Global) was inserted into the test block immediately adjacent to the rim of the circular baseplate. We chose to simulate 3 conditions of impingement (mild, moderate, and severe), which we equated to the number of screw threads visible and protruding from the top of the Sawbones test block. Specifically, 3 visible threads served as the severe impingement model, whereas a single visible thread was considered the mild impingement condition in our test setup. Initially, the screw was inserted such that all 3 locking threads were protruding from the top of the test block (severe impingement). Visually, in this condition, the top of the screw head was flush with the top side of the baseplate. Thereafter, proper seating of a 36mm neutral glenosphere onto the baseplate’s taper was attempted.
Glenosphere dissociation after RSA
Figure 1
1063
Biomechanical setups for (A) axial distraction and (B) torsional tests.
Figure 2 Illustrations of the control (A) and simulated impingement conditions (B, C) evaluated biomechanically. The simulated impingement served to prevent full taper engagement (C). After impaction, the integrity of the construct was manually evaluated and found to be mechanically unstable as the glenosphere dissociated from the baseplate with minimal load application. Resultantly, this condition was eliminated from consideration as our experimental impingement model. Thereafter, the locked screw was advanced such that 2 threads were visibly protruding from the test block surface. Again, the glenosphere was impacted and the construct was manually manipulated by the senior author and found to be mechanically secure, although visually it was apparent that full taper engagement was not achieved as the underside of the glenosphere was not parallel to the top surface of the test block (Fig. 2, B and C). The glenosphere was removed and the locked screw was advanced such that a single thread was visible above the test block surface. The glenosphere was impacted, and the senior author (M.A.F.) confirmed that the construct was mechanically secure. However, in the mild impingement model, the base of the glenosphere appeared parallel to the top surface of the test block, indicating that full taper engagement may have been achieved despite a single screw thread still protruding slightly from the block. Based on the visual perception that impingement was not achieved, we adopted the moderate impingement condition as our experimental impingement condition for all subsequent experiments. The senior author and 3 shoulder and elbow fellows performed all biomechanical experiments to evaluate the effect of simulated impingement on the mechanical integrity of the Morse taper interface between the glenosphere and baseplate of the RSP system. Test specimens were evaluated with and without simulated moderate impingement, and for each condition, the glenospherebaseplate construct was manually tested in axial distraction and torsion in succession. Each individual performed the biomechanical experiment with and without impingement once each, and all
tests were performed with the same baseplate and glenosphere combination. The constructs were mounted in a vise clamp (Fig. 1, A and B) and rigidly coupled to a 6-axis load cell (MC3A-1000; Advanced Medical Technologies, Inc., Watertown, MA, USA), which recorded force (N) and moment (N-m) data continuously during axial distraction and torsional testing at 30 Hz. For the axial distraction test, the surgeon was instructed to attempt to disengage the taper using a custom-designed T-handle that was threaded into the glenosphere. Further, the surgeon was instructed to load the construct to a maximum applied force of up to 400 N using visual feedback from the load cell’s output. This force, although likely larger than the loads applied during implant placement to gauge proper taper impaction and stability, is considerably less than literature-based reports of dissociation force magnitudes reported by Blevins et al.2 If the glenosphere withstood axial force application without dissociating from the taper, torsional loading of the glenosphere commenced. The surgeon was instructed to torque the glenosphere up to a maximum of 20 N-m, again using visual feedback from the load cell’s output, as torque magnitudes above this value were determined in pilot experiments to fracture the Sawbones test block with a properly seated glenosphere. Maximum applied axial load (N) and torsional moment (N-m) were recorded for the control (no impingement) and impingement conditions.
Results During the 14-year study period, we confirmed 13 glenosphere dissociations radiographically. Eleven of the dissociations occurred in patients operated on by the senior
1064
M.C. Cusick et al.
Figure 3 Sagittal view of a synthetic model scapula implanted with the RSP baseplate illustrating that after placement, there is exposed glenoid both superiorly and inferiorly that may contribute to glenosphere impingement and compromised mechanical integrity.
author of a total of 1800 reverse shoulder replacements performed with the RSP system in the study period, for an overall glenosphere-baseplate dissociation rate of <1%. Two additional dissociations were confirmed in patients initially treated at an outside institution. Both of these RSAs were performed with the Trabecular Metal Reverse Shoulder System (Zimmer, Warsaw, IN, USA) prosthesis. Glenosphere sizes identified included 32 mm (n ¼ 1, DJO), 36 mm (n ¼ 2, DJO; n ¼ 2, Zimmer), 40 mm (n ¼ 6, DJO), and 44 mm (n ¼ 2, DJO). Timing of dissociation ranged from <1 month after surgery to 7 years after surgery (Table II). Because of differences in the design features of the DJO and Zimmer RSA prostheses, which include but are not limited to the addition of a glenosphere locking screw in the DJO design that is lacking in the Zimmer prosthesis, and differences in taper geometry (length and angle) between both implants, we investigated whether glenosphere size correlated to the rate of dissociation for the DJO design only. This analysis was also performed because the DJO design constituted 85% (11 of 13) of our retrieved implants. The incidence of dissociation by glenosphere size for the DJO retrievals was 0.09% (32 mm), 0.25% (36 mm), 5.4%
(40 mm), and 8.3% (44 mm), and this correlation was significant (P < .001). In addition, glenospheres were divided into a small group (sphere sizes 32 mm and 36 mm) and a large group (sphere sizes 40 mm and 44 mm) on the basis of surface area overhang relative to the baseplate. The incidence of dissociation in the larger glenosphere group (2%) was significantly greater (P < .001) compared with the smaller glenosphere group (0.1%). Review of patient records revealed 1 incident of dissociation with a 32-mm sphere. The patient was an 82-yearold man who was treated with a primary RSA for cuff tear arthropathy. The dissociation occurred 27 months after surgery and was caused by trauma in the form of a fall on an adducted arm while playing tennis. At the time of revision surgery, there appeared to be an impacted area on the medial aspect of the polyethylene socket. In 4 cases of dissociation with a size 36-mm sphere, no trauma was identified. Time to dissociation in these cases ranged from 7 months to 7 years. Two cases involved a Zimmer prosthesis and 2 cases were with a DJO prosthesis, 1 of which was custom and performed in 1999. Dissociation of this device occurred 7 years after the index procedure in an 82-year-old woman who had undergone revision
Glenosphere dissociation after RSA
1065
Table II
Summary of dissociation findings
RSA make
Gender
Age at time of dissociation
Glenosphere size (mm)
Time to dissociation
Implant available for retrieval analysis?
Modified Goldberg Score of retrieval
Primary reason (if known) for dissociation
DJO DJO DJO
Male Female Female
83 82 78
32-4 36-4 36-4
27 months 86 months 28 months
Y N N
1 N/A N/A
DJO DJO DJO DJO DJO DJO DJO DJO Zimmer Zimmer
Female Female Male Male Female Female Male Male Male Male
48 72 64 75 74 69 69 60 75 82
40 40 40 40 40 40 44 44 36 36
2 weeks 48 months 8 months 7 months 5 months 5 months 6 months 3 months 24 months 7 months
N N Y Y N N N N Y Y
N/A N/A 1 1 N/A N/A N/A N/A 2 1
Fell while playing tennis Lifted granddaughter Overexerted shoulder while on crutches after foot surgery Not reported but atraumatic Not reported but atraumatic Not reported but atraumatic Not reported but atraumatic Not reported but atraumatic Not reported but atraumatic Fell out of bed Not reported but atraumatic Not reported but atraumatic Not reported but atraumatic
to RSA for failed hemiarthroplasty. Dissociation occurred when the patient lifted her grandchild. Review of immediate postoperative films revealed that the glenosphere was not fully impacted on the Morse taper, perhaps resulting in mechanical compromise of glenosphere fixation. The other DJO implant in this group failed more than 2 years after implantation. This patient had recent foot surgery and was noted to be ambulating with heavy use of her upper extremities before dissociation, but no specific traumatic episode was noted. One Zimmer prosthesis dissociated at 2 years after surgery; the other dissociated twice, once from trauma 6 weeks after surgery and a second time 7 months after the initial revision (which prompted referral to our office). Both cases were revised with use of the same size sphere and had no further complications. There were 6 size 40-mm spheres whose time to dissociation varied from 2 weeks to 4 years. No trauma was identified in any case. One case involved another custom prosthesis (late 1990s) that failed identically to the method previously described. Two cases of size 44-mm sphere dissociation were identified. One patient was a Vietnam War veteran and carried a diagnosis of post-traumatic stress disorder. Six months after surgery, the patient fell out of bed while sleeping and developed acute-onset shoulder pain and dysfunction. The second was in a case of documented recurrent instability and infection in a failed RSA. Glenosphere dissociation in this patient occurred 3 months after the index procedure.
Retrieval analysis Five of our 13 dissociated glenospheres were available for analysis (Table II). We attribute the loss of 8 RSA retrieved implants to the retrospective nature of this study and, until
Figure 4 Image of a retrieved taper illustrating evidence of fretting wear and highly localized corrosion.
recently, the lack of an implant storage facility as well as a dedicated implant retrieval and analysis laboratory. Time to dissociation of the retrievals ranged from 7 to 27 months. Qualitatively, visual inspection of all implants revealed minor evidence of fretting wear regardless of the time of revision surgery. One instance of highly localized corrosion in combination with fretting wear was identified on 1 implant (Fig. 4). No evidence of permanent deformation of the taper was noted on any implant. Based on the Modified Goldberg Score,10 4 of 5 retrieved implants received a score of 1, with only 1 taper surface receiving a score of 2
1066 Table III
M.C. Cusick et al. Raw data from biomechanical experiments
Test condition Trial 1 2 3 4 Average Standard deviation
Without impingement
Simulated impingement
Distraction force (N)
Applied torque (N-m)
Distraction force (N)
Applied torque (N-m)
390 455 407 358 402.5 40.5
20.5 19.0 18.0 19.2 19.2 1.0
345 480 352 386 390.8 62.1
9.3 5.9 8.0 7.0 7.5 1.5
(Table II). No implant received a score of higher than 2 based on this scoring system.
Biomechanical results Maximum measured axial force (N) and torsion (N-m) were recorded for both phases of testing. Our biomechanical results indicated that our laboratory model of impingement that resulted in less than optimal taper engagement did not negatively affect the mechanical integrity of the taper when axial distraction forces were applied to the glenosphere (Table III). Specifically, the average axial distraction forces for the control and simulated impingement conditions were 402.5 40.5 N and 390.8 62.1 N, respectively, and no glenosphere in either test condition dissociated from the underlying baseplate. However, torsional testing revealed that simulated interference and less than optimal taper engagement negatively affected the mechanical integrity of the glenospherebaseplate interface. Specifically, each construct tested without impingement withstood the set maximum applied torque of 20 N-m without the interface failing (19.2 1.0 N-m). However, in the impingement group, dissociation failure was observed in all test constructs at an average applied torque of 7.5 1.5 N-m. Thus, simulated impingement reduced the biomechanical integrity of the taper interface by approximately 60%.
Discussion Every reverse arthroplasty system has a baseplate coupled to a sphere by a Morse taper. Morse tapers are not standardized in the orthopedic industry and vary in design between implant manufacturers, notably with regard to taper angles that generally fall in the range of 5 to 18 .11 In addition, RSA prostheses differ regarding whether, after proper taper engagement, the glenosphere sits flush with the underlying glenoid or is elevated slightly from the bone surface. An advantage of the latter is that a fully seated taper is less likely to impinge on surrounding glenoid soft tissue or bone. A potential disadvantage is the greater possibility of fluid contamination of the taper itself, which can lead to corrosive
wear of the taper. Advantages of RSA designs in which the glenosphere sits flush with the glenoid include load sharing between the taper and the surrounding flush rim of glenoid bone. In addition, it may be less likely to be subjected to corrosive wear. The disadvantage of this design is that it is subject to ‘‘impingement’’ or improper taper impaction that puts the device at risk for early mechanical failure. In addition, some systems have different methods of engaging the glenosphere on the glenoid baseplate. Some have an instrument to aid in the engagement of the glenosphere to the baseplate, whereas others use manual placement. Ultimately, the surgeon should be aware of advantages and disadvantages of the taper and design features of his or her preferred system. In the system biomechanically tested in this study (RSP, DJO Global), the glenosphere is manufactured with a threaded hole that allows attachment of a T-handle to aid glenosphere insertion. The design rationale for this threaded insertion handle (as tested mechanically in this experiment) assumes that a manual distraction force can be applied during surgery with the threaded handle to ensure proper taper impaction and stability, thus providing additional evidence of full taper engagement. Unfortunately, based on our biomechanical testing, it is likely that a false taper was perceived as fully engaged intraoperatively. Despite the biomechanical data reported here that confirm Morse tapers are weaker in torsion than in axial distraction in conditions in which the taper is improperly seated, we are unable to conclude that torsional testing will improve our intraoperative ability to avoid dissociation. This is because the laboratory-measured torques required to dissociate falsely engaged tapers (7 N-m) are likely not practical to apply intraoperatively without markedly increasing the risk of fracturing the patient’s glenoid or scapula. Nevertheless, our current surgical technique includes the application of both manual axial distraction and torsion in an attempt to avoid early dissociation. Multiple factors are thought to contribute to possible dissociation of Morse tapers, and much of this information is derived from modular hip prosthesis designs in which Morse tapers are an integral design component.1,3,10,14,19-21,23,24 Improper taper impaction, contamination of the taper socket with fluid or solid material, improper taper design or manufacturing, micromotion-inducing rotational forces
Glenosphere dissociation after RSA across the joint, and material failure by corrosion have been shown to negatively affect Morse tapers. Glenosphere dissociation after documented trauma appears to be a feasible mechanism. To our knowledge, no studies have evaluated the force necessary to dissociate a glenosphere in vivo. Biomechanical axial pullout tests of the Morse taper in 1 implant show that forces greater than 4448 N are required to dissociate the taper. Biomechanical evaluation of modular total shoulder humeral components required forces of around 3000 N to dissociate the humeral head.2 In the absence of trauma, we speculate that early glenosphere dissociation (<6 months) is likely due to technical error, which can take multiple forms and negatively affect taper strength, including low impaction force,20 inadvertent fluid contamination of the taper,2,13,20 and impingement between the underside of the glenosphere and glenoid as demonstrated here in our biomechanical model. Blevins et al2 reported that as little as 0.4 mL of water or blood could prevent a Morse taper from engaging properly. These authors also demonstrated that if the taper socket is contaminated with water or blood during impaction, the force required to dissociate the taper decreased by 24.3% and 76.9%, respectively. Every attempt should be made to dry the taper at the time of glenosphere placement. More than two thirds of our dissociations involved glenospheres size 40 mm or larger. We believe impingement may be responsible for these early failures in larger sphere sizes. Soft tissue or bone impingement due to inadequate glenoid exposure or preparation may lead to an improperly impacted taper or ‘‘impaction default.’’ When the source of impingement is later removed in vivo, the taper interface becomes loose and can lead to dissociation. Therefore, full glenoid exposure with 360 of soft tissue release in combination with adequate glenoid reaming is essential to mitigate the potential for impingement. Adequate glenoid exposure can be difficult to attain in revision cases or in stiff shoulders from preexisting deformity or advanced disease processes. Late dissociation (>6 months) may be due to material properties of the implant or implant design. Corrosion at the Morse taper has been postulated as a potential source for weakening the connection between modular implants.9,10,14 The baseplate in our reported cases is titanium articulating with a cobalt chrome glenosphere. Corrosion between titanium and cobalt chrome taper articulations has been previously reported in the literature.9,26 Gilbert et al9 reported that approximately 16% to 35% of 148 retrieved total hip implants (titanium femoral stem and cobalt-alloy femoral head) showed signs of moderate to severe corrosive attack in the head-neck taper connection. This corrosion may lead to relative motion (fretting) at modular connections and ultimately to abrasive loss of the passivating oxide layer between the implants, leading to poor mechanical stability of the connection. Gross evaluation of our late dissociation (7 years) taper appeared to show some evidence of fretting along the implant. However, in our
1067 limited sample of retrievals, we identified minor and limited evidence of taper corrosion. This observation of prosthesis damage in the form of fretting wear without significant evidence of metal corrosion is inconsistent with prior literature describing total hip and knee retrievals, in which corrosion is commonly seen.10,12,14,26 However, our observations are similar to those of Day and Abboud (personal communication, 2014), who report observing fretting damage with little or no corrosion on 18 RSA taper retrievals from multiple manufacturers.
Conclusion In our series, we identified several potential contributing mechanisms to the postoperative complication of glenosphere-baseplate dissociation after RSA, including trauma and improper taper engagement. Although it is a rare complication after RSA, we identified a significantly higher likelihood of glenosphere dissociation when larger diameter glenospheres (40 and 44 mm) were implanted, likely attributable to the larger exposed surface area for potential soft tissue or bone impingement. Finally, in contrast to retrieval studies of modular total hip arthroplasty designs in which taper corrosion is commonly observed, we identified limited evidence of corrosive wear at the Morse taper interface in the RSA prosthesis designs analyzed in our series, and the primary evidence of implant damage was in the form of mild fretting wear.
Acknowledgment The authors would like to acknowledge Frederick A. Matsen III, MD, for bringing this problem to our attention and stimulating the completion of this project. We would also like to acknowledge Kathleen O’Brien, BS, and Ben Cottrell, BS, for assistance with manuscript preparation.
Disclaimer The research foundation with which Rachel E. Clark is affiliated received research support from DJO Surgical, a designer and manufacturer of orthopedic surgical products related to the subject of this work. Derek J. Cuff receives consulting fees from and is a member of the speaker’s bureau for DJO Surgical and Arthrex, both designers and manufacturers of orthopedic surgical products related to the subject of this work. The research foundation with which Brandon G. Santoni is affiliated received research support from DJO
1068 Surgical, a designer and manufacturer of orthopedic surgical products related to the subject of this work. Mark A. Frankle receives royalties from and is a paid consultant for DJO Surgical, a designer and manufacturer of orthopedic surgical products related to the subject of this work. 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.
M.C. Cusick et al.
13.
14.
15.
16.
References 17. 1. Atwood SA, Patten EW, Bozic KJ, Pruitt LA, Ries MD. Corrosioninduced fracture of a double-modular hip prosthesis: a case report. J Bone Joint Surg Am 2010;92:1522-5. http://dx.doi.org/10.2106/JBJS. I.00980 2. Blevins FT, Deng X, Torzilli PA, Dines D, Warren RF. Dissociation of modular humeral head components: a biomechanical and implant retrieval study. J Shoulder Elbow Surg 1997;6:113-24. 3. Bobyn JD, Tanzer M, Krygier JJ, Dujovne AR, Brooks CE. Concerns with modularity in total hip arthroplasty. Clin Orthop Relat Res 1994: 27-36. 4. Cheung E, Willis M, Walker M, Clark R, Frankle MA. Complications in reverse total shoulder arthroplasty. J Am Acad Orthop Surg 2011; 19:439-49. 5. Cuff DJ, Pupello DR. Comparison of hemiarthroplasty and reverse shoulder arthroplasty for the treatment of proximal humeral fractures in elderly patients. J Bone Joint Surg Am 2013;95:2050-5. http://dx. doi.org/10.2106/JBJS.L.01637 6. Drake GN, O’Connor DP, Edwards TB. Indications for reverse total shoulder arthroplasty in rotator cuff disease. Clin Orthop Relat Res 2010;468:1526-33. http://dx.doi.org/10.1007/s11999-009-1188-9 7. Farshad M, Gerber C. Reverse total shoulder arthroplastydfrom the most to the least common complication. Int Orthop 2010;34:1075-82. http://dx.doi.org/10.1007/s00264-010-1125-2 8. Flury MP, Frey P, Goldhahn J, Schwyzer HK, Simmen BR. Reverse shoulder arthroplasty as a salvage procedure for failed conventional shoulder replacement due to cuff failuredmidterm results. Int Orthop 2011;35:53-60. http://dx.doi.org/10.1007/s00264-010-0990-z 9. Gilbert JL, Buckley CA, Jacobs JJ. In vivo corrosion of modular hip prosthesis components in mixed and similar metal combinations. The effect of crevice, stress, motion, and alloy coupling. J Biomed Mater Res 1993;27:1533-44. 10. Goldberg JR, Gilbert JL, Jacobs JJ, Bauer TW, Paprosky W, Leurgans S. A multicenter retrieval study of the taper interfaces of modular hip prostheses. Clin Orthop Relat Res 2002:149-61. 11. Hernigou P, Queinnec S, Flouzat Lachaniette CH. One hundred and fifty years of history of the Morse taper: from Stephen A. Morse in 1864 to complications related to modularity in hip arthroplasty. Int Orthop 2013;37:2081-8. http://dx.doi.org/10.1007/s00264-0131927-0 12. Higgs GB, Hanzlik JA, MacDonald DW, Gilbert JL, Rimnac CM, Kurtz SM, et al. Is increased modularity associated with increased
18.
19.
20.
21.
22.
23.
24.
25.
26.
fretting and corrosion damage in metal-on-metal total hip arthroplasty devices?: a retrieval study. J Arthroplasty 2013;28:2-6. http://dx.doi. org/10.1016/j.arth.2013.05.040 Jauch SY, Huber G, Hoenig E, Baxmann M, Grupp TM, Morlock MM. Influence of material coupling and assembly condition on the magnitude of micromotion at the stem-neck interface of a modular hip endoprosthesis. J Biomech 2011;44:1747-51. http://dx.doi.org/10. 1016/j.jbiomech.2011.04.007 Kop AM, Swarts E. Corrosion of a hip stem with a modular neck taper junction: a retrieval study of 16 cases. J Arthroplasty 2009;24:1019-23. http://dx.doi.org/10.1016/j.arth.2008.09.009 Kurtz SM, Kocagoz SB, Hanzlik JA, Underwood RJ, Gilbert JL, MacDonald DW, et al. Do ceramic femoral heads reduce taper fretting corrosion in hip arthroplasty? A retrieval study. Clin Orthop Relat Res 2013;471:3270-82. http://dx.doi.org/10.1007/s11999-013-3096-2 Middernacht B, De Wilde L, Mole D, Favard L, Debeer P. Glenosphere disengagement: a potentially serious default in reverse shoulder surgery. Clin Orthop Reatl Res 2008;466:892-8. http://dx.doi.org/10. 1007/s11999-007-0090-6 Mole D, Navez P, Garaud P. Reverse shoulder prosthesis: problems related to the glenoid. In: Walch G, Boileau P, Mole D, Favard L, Levigne C, Sivreau F, editors. Reverse shoulder arthroplasty: clinical results, complications, revision. Montpellier, France: Sauramps Medical; 2006. p. 289-301. Mulieri P, Dunning P, Klein S, Pupello D, Frankle M. Reverse shoulder arthroplasty for the treatment of irreparable rotator cuff tear without glenohumeral arthritis. J Bone Joint Surg Am 2010;92:254456. http://dx.doi.org/10.2106/JBJS.I.00912 Pansard E, Fouilleron N, Dereudre G, Migaud H, Girard J. Severe corrosion after malpositioning of a metallic head over the Morse taper of a cementless hip arthroplasty. A case report. Orthop Traumatol Surg Res 2012;98:247-50. http://dx.doi.org/10.1016/j.otsr.2011.05.018 Pennock AT, Schmidt AH, Bourgeault CA. Morse-type tapers: factors that may influence taper strength during total hip arthroplasty. J Arthroplasty 2002;17:773-8. http://dx.doi.org/10.1054/arth.2002. 33565 Rehmer A, Bishop NE, Morlock MM. Influence of assembly procedure and material combination on the strength of the taper connection at the head-neck junction of modular hip endoprostheses. Clin Biomech 2012;27:77-83. http://dx.doi.org/10.1016/j.clinbiomech.2011. 08.002 Saltzman BM, Chalmers PN, Gupta AK, Romeo AA, Nicholson GP. Complication rates comparing primary with revision reverse total shoulder arthroplsty. J Shoulder Elbow Surg 2014;23:1647-54. http:// dx.doi.org/10.1016/j.jse.2014.04.015 Shareef N, Levine D. Effect of manufacturing tolerances on the micromotion at the Morse taper interface in modular hip implants using the finite element technique. Biomaterials 1996;17:623-30. Unnanuntana A, Chen DX, Unnanuntana A, Wright TM. Trunnion fracture of the anatomic medullary locking A Plus femoral component. J Arthroplasty 2011;26:504.e13-e16. http://dx.doi.org/10.1016/j.arth. 2010.03.028 Walker M, Willis MP, Brooks JP, Pupello D, Mulieri PJ, Frankle MA. The use of the reverse shoulder arthroplasty for treatment of failed total shoulder arthroplasty. J Shoulder Elbow Surg 2012;21:514-22. http://dx.doi.org/10.1016/j.jse.2011.03.006 Witt F, Bosker BH, Bishop NE, Ettema HB, Verheyen CC, Morlock MM. The relation between titanium taper corrosion and cobalt-chromium bearing wear in large-head metal-on-metal total hip prostheses: a retrieval study. J Bone Joint Surg Am 2014;96:e157. http://dx.doi.org/10.2106/JBJS.M.01199