- Email: [email protected]
A prospective multicenter clinical study of the Discovery elbow
J Shoulder Elbow Surg (2014) 23, e95-e107
www.elsevier.com/locate/ymse
A prospective multicenter clinical study of the Discovery elbow Hill Hastings II, MDa, Donald H. Lee, MDb, William S. Pietrzak, PhDc,d,* a
Indiana Hand to Shoulder Center, Indianapolis, IN, USA Vanderbilt University School of Medicine, Nashville, TN, USA c Biomet Inc, Warsaw, IN, USA d Department of Bioengineering, University of Illinois at Chicago, Chicago, IL, USA b
Background: Semiconstrained total elbow arthroplasty is used to improve elbow function and reduce pain. Although effective, high complication rates exist, with the polyethylene bushing especially susceptible to failure. The Discovery Elbow System (Biomet Inc, Warsaw, IN, USA) contains a spherical bearing designed to minimize polyethylene wear. This prospective, multicenter clinical study investigated the 4year (mean) outcomes of this elbow. Methods: From 2002 to 2009, 92 patients (71 women, 21 men; mean age, 63.9 years; range, 33.4-88.7 years) received 99 Discovery elbows at 4 centers. The study cohort was limited to 46 elbows with complete preoperative and minimum 2-year clinical (modified American Shoulder and Elbow Surgeons elbow score) and radiographic follow-up. Results: Mean follow-up was 4.1 years (range, 2-5.9 years). All American Shoulder and Elbow Surgeons elbow score components improved significantly (P < .001). Mean flexion-extension arcs increased from 81 to 121 and pronation-supination arcs from 134 to 163 (P < .001). Loose locking screws in 2 elbows (first-generation screws), a loose polyethylene bearing in 1 (history of falls), and a condyle/bearing in 1 (deep infection) were exchanged. Among the 46 elbows, gross survivorship was humeral/ulnar components, 100%; condyles, 97.8%; bearings, 95.7%; and screws, 95.7%. One humeral component (2.2%) was radiographically loose but not revised. An additional elbow (elbow 47) that did not meet the criteria for inclusion (<2 years of follow-up) was revised due to a loose humeral component and was reported separately. Conclusion: The Discovery elbow increased function and decreased pain with high survivorship at a mean of 4.1 years. Level of evidence: Level IV, Case Series, Treatment Study. Ó 2014 Journal of Shoulder and Elbow Surgery Board of Trustees. Keywords: Discovery elbow; outcomes; prostheses; semiconstrained; survivorship; total elbow arthroplasty
The following Investigational Review Board approvals are relevant to this study: St. Vincent, Indianapolis, IN, USA (R2002-014); Partners Healthcare, Boston, MA, USA (2001P001731/MGH); Vanderbilt University, Nashville, TN, USA (060092); and the University of South Florida, Tampa, FL, USA (102165). *Reprint requests: William S. Pietrzak, PhD, 1691 S Meadow Dr, Warsaw, IN 46581, USA. E-mail address: [email protected] (W.S. Pietrzak).
Proper elbow motion is more critical to upper-extremity function than that of any other associated joint.11 For instance, a 100 arc of motion is required to accomplish the activities of daily living.2,10 Since the 1970s, total elbow arthroplasty (TEA) has been increasingly used as a salvage technique to decrease pain, increase joint stability, and
1058-2746/$ - see front matter Ó 2014 Journal of Shoulder and Elbow Surgery Board of Trustees. http://dx.doi.org/10.1016/j.jse.2013.12.033
e96 improve overall range of motion.10 The first TEA prostheses were fully constrained and hinged, with metal-onmetal articulation and no significant varus or valgus laxity.3,10 Unfortunately, these led to metallic particulate synovitis, loosening rates as high as 45% due to force transmittance through the hinged articulation to the bone–cement interface, and implant failure.5,10,29 In response, semiconstrained and nonconstrained prostheses were developed and have been used exclusively during the past decade.29 Unlinked, or minimally constrained, TEA prostheses are elbow resurfacing devices that can allow for a more anatomic ulnohumeral articulation.29 Although such designs may reduce the risk of loosening, they may increase the risk of dislocation and have limited application in cases of bone loss or ligamentous deficiency.10 Semiconstrained TEA prostheses contain a hinged linkage to help resist dislocation and is ‘‘sloppy,’’ in that it allows for some varus/valgus motion, better replicating the kinematics of the natural elbow, and can be used in the presence of bone or ligamentous deficiency.10,27 Also, force transfer across the bone–cement interface is lower than for constrained devices,27,29 thus reducing the potential for loosening. TEA is effective, but the results generally have not been as good as those for total hip or knee replacement,17,29 although rheumatoid elbow patients are an exception because patients with diffuse disease place less demand on their elbow.6,8,17,27-29 TEA tends to be associated with a high rate of complications (eg, 14%-80%),17 partially due to the difficult nature of surgery in this complex joint with thin soft tissue coverage.2 In particular, deep infection and septic loosening are a concern, with rates as high as w10% reported.2,22 Ulnar neuropathy has also been reported to occur in up to 21% of patients within a few days of surgery, with the rate of permanent dysfunction ranging from 0% to 10%.26 High rates of polyethylene bushing failure/wear have been reported for semiconstrained prostheses.6,10,11,19,20,24,27 For instance, rates of 14% to 47% have been cited for the Coonrad/Morrey prosthesis (Zimmer, Warsaw, IN, USA)6,11,19,20 and 15% for the Solar Elbow System (Stryker, Mahwah, NJ, USA).24 Semiconstrained devices can also experience hinge failure, as noted by a systematic review by Little et al,17 to occur in 6% of patients. The Discovery Elbow System (Biomet Inc, Warsaw, IN, USA) was developed to address some of the issues associated with semiconstrained TEA prostheses, including 1 spherical bearing designed to reduce wear and allow for simple polyethylene exchange, if required.10,11 Previous clinical reports on this system have been limited by small numbers of patients (ie, 1-2),2,11 short mean follow-up of <2 years,30 or the provision of only a brief, overall clinical summary.9 As such, a need remains for a detailed, clinical assessment of this system. Our purpose was to (1) determine the short-term to midterm survivorship of the Discovery elbow and (2) report the clinical and radiographic outcomes
H. Hastings II et al. of a multicenter study, comparing these results with those reported for other semiconstrained TEA prostheses.
Materials and methods The Discovery Elbow System The Discovery Elbow System has been described by Hastings and Theng11 and Hastings.9,10 Lateral and posterior views of the prosthesis are shown in Figures 1 and 2, respectively. Two cobalt chromium molybdenum condylar hemispheres lock into the distal humeral component with a medial and lateral Ti6Al4V screw, respectively. These articulate with congruent ArCom polyethylene (Biomet Inc) bearing surfaces captured within the proximal ulnar component with a locking pin, providing 7 of varus/valgus laxity. The humeral stem is made of Ti6Al4V and contains a posterior bow, including a 5 lateral offset and 5 internal rotation to reproduce the anatomy. The distal humeral flange resides outside of the canal and hooks around the distal anterior humeral cortex to resist posteriorly directed forces of the implant in the distal humerus. The flange is designed to sit flush against the distal anterior humeral cortex to obviate the need for a bone graft and potential graft resorption. The ulnar stem, made of the same alloy, contains a lateral bow with a 23 anterior neck angle and a lateral offset to properly position the joint axis in an anatomic position. The humeral and ulnar components can be first cemented into place and later coupled together. Alternatively, the humeral and ulnar components can be preassembled and then implanted.
Clinical study design Four clinical centers participated in this prospective study. Written informed consent was obtained from each patient. Inclusion criteria were noninflammatory joint disease, including osteoarthritis and avascular necrosis; inflammatory arthritis; revision where other devices or treatments have failed; correction of functional deformity; and treatment of acute fractures or nonunion about the elbow. Exclusion criteria included patients aged <18 years, pregnancy, metabolic disorders that could impair bone formation, marked bone loss, active or suspected infection about the elbow or distant foci of infections that could spread to the implant site, unwillingness or inability to comply with the rehabilitation program, and factors that could limit the patient’s ability to conform to the prescribed follow-up schedule. Hastings9 described the technique for implanting the Discovery elbow using 3 surgical approach options: triceps-off (reflecting), triceps-off (splitting), and triceps-on (sparing). Postoperatively, patients with primary arthroplasty were immobilized in a bulky dressing and splint for 5 days. If the wound was sealed at that time, active, passive range of motion was initiated. An extension splint was used at night as needed to maintain full elbow extension. Activities of daily living were encouraged. In cases in which the triceps was detached for exposure, no extension against resistance was allowed for 6 weeks after surgery. Patients were given a lifting restriction of 5 lbs. Patients were clinically and radiographically evaluated preoperatively and at the postoperative follow-up intervals of 3 months, 6 months, 1 year, then annually thereafter up to 5 years. To the
Clinical study of the Discovery elbow
e97
Figure 1 Discovery elbow is seen in a lateral view, with the ulnar stem on the left and the humeral stem on the right. The inset shows detail of articulation.
Figure 2 Discovery elbow is seen in a posterior view, with the ulnar stem on the left and the humeral stem on the right. The inset shows detail of articulation.
extent practicable, the treating physician was directly involved with the clinical evaluations but was allowed to delegate this responsibility at his or her discretion. Radiographic views included anteroposterior (AP) and lateral. Clinical performance was assessed using the American Shoulder and Elbow Surgeons (ASES) Elbow Assessment Form.15 This scoring system includes a physician assessment–signs table that includes 22 signs with categoric scores. Six of the signs deal with tenderness and 2 with impingement pain (in flexion and in extension). These 8 signs are graded with scores of 0 ¼ none, 1 ¼ mild, 2 ¼ moderate, and 3 ¼ severe, with the remaining 16 signs scored yes or no. For the purposes of this study, the other 5 pain signs (pain on resisted wrist extension, resisted wrist flexion, resisted long finger extension, resisted wrist pronation, and resisted wrist supination) that would have ordinarily been graded as yes or no were, instead, graded on the 0- to 3-points system. Thus, the maximum potential pain/ tenderness score (worst pain/tenderness) was 39 rather than 24. Figures 3, 4, and 5 define the radiographic zones in the lateral view for the ulnar (zones 1-9) and humeral (zones 1-8) components, the lateral view in the vicinity of the humeral flange (zones 1-5), and the AP view of the ulnar (zones 1-7) and humeral (zones 1-7) components, respectively. Radiolucent line widths were measured and recorded by other than the treating surgeon for the cement–prosthesis (CP) and bone–cement (BC) interfaces for the humeral and ulnar components. Regarding the humeral flange region, the percent completeness of the linear contact at the CP interface was measured for zones 1 and 2 and for the bone–prosthesis (BP) interface for zones 4 and 5. Values were categorized as 0% to 25%, 26% to 50%, 51% to 75%, and 76% to 100%, as appropriate. For zone 3, the presence or absence of contact at the CP interface and the BC interface was noted.
Figure 3 Ulnar and humeral radiographic zones are seen in a lateral view.
Figure 4 Radiographic zones about the humeral flange region are seen in a lateral view.
Statistical analysis A paired Student t test was used to compare the preoperative ASES scores with those at the last follow-up using Excel software (Microsoft Corp, Redmond, WA, USA). Significance was taken for P < .05.
e98
H. Hastings II et al. Table I
Patient demographics
Parameter
No. or mean (range)
Patients Bilateral patients Sex Female Male Total elbows Female Male Age, y Surgery side Left Right Dominant side affected? Yes Ambidextrous No Body mass index, kg/m2 Follow-up, y
42 4 34 8 46 37 9 63.4 (33.4-88.7) 21 25 24 1 21 27.7 (19.2-40.3) 4.1 (2.0-5.9)
Figure 5 Ulnar and humeral radiographic zones are seen in an anteroposterior view.
Table II
Primary diagnosis
Incidence (%) (n ¼ 46 elbows)
Results
Rheumatoid arthritis Traumatic arthritis Humeral fracture Osteoarthritis Recurrent dislocation/instability Revision total elbow arthroplasty Ulnar fracture Distal humeral nonunion Radial head/distal humeral fracture
23 6 5 5 2 2 1 1 1
Patient demographics TEAs were performed from January 2002 to August 2011 on 92 patients (99 elbows) by 6 surgeons at 4 surgery centers (2 surgeons at each of 2 sites and 1 surgeon at each of the remaining sites). The mean number of TEAs performed at each center was 25 (range, 10-61). Five patients (6 elbows) died, all with their total elbow prostheses in place at the time of death, and 15 patients (15 elbows) rescinded consent. Forty-three patients (46 elbows) had a minimum of 2 years of follow-up (mean, 4.1 years; median, 4.2 years; range, 2.0-5.9 years), with complete preoperative and postoperative evaluation, and formed the study population. These included 2 of the patients who died (3 elbows) and 3 of the patients who rescinded consent (3 elbows) after a minimum of 2 years of follow-up. The mean number of TEAs performed within this subset per site was 12 (range, 1-32). Patient demographics are summarized in Table I. Approximately 80% were women and about half of the operated elbows were on the dominant side. Table II lists the primary diagnoses. Rheumatoid arthritis was present in 50% of elbows, and post-traumatic arthritis, osteoarthritis, and humeral fracture were the primary diagnoses for 35%. As reported in Table III, the surgical approaches used for 96% of elbows were tricepsoff (reflecting), triceps-off (splitting), or triceps-on (sparing).
Primary diagnoses
(50.0) (13.0) (10.9) (10.9) (4.3) (4.3) (2.2) (2.2) (2.2)
Humeral component stem lengths were 100 to 150 mm and diameters were 3 to 6 mm. Ulnar component stem lengths were 75 or 115 mm and diameters were 3 to 5 mm; however, 1 elbow required a custom 84 mm 2.5 mm ulnar component. All elbows received cemented fixation, and a bone graft was placed under the humeral flange in 3 (6.5%).
Survivorship Among the 46 elbows, 2 with loose locking screws underwent screw exchange at 2.0 and 2.1 years, respectively; the pin and bearing were replaced in 1 elbow due to loosening after multiple falls; and treatment of 1 elbow with a deep infection included condyle and bearing exchange at 6.2 years. Gross survivorship in this series as determined by revision, by component, was humeral/ulnar components, 100% (46 of 46); condyles, 97.8% (45 of 46); bearing, 95.7% (44 of 46); and locking screws, 95.7% (44 of 46) at a
Clinical study of the Discovery elbow Table III
Operative information
Variable
No. (%) or average (range) (n ¼ 46 elbows)
Surgical approach Triceps-on (sparing) 18 (39.1) Triceps-off (reflecting) 19 (41.3) Triceps-off (splitting) 7 (15.2) Other 2 (4.3) Tourniquet time, min 113.6 (57-209) Removing 3 outliers) 112.8 (57-209) Operative time, min 149.7 (90-386) Removing 3 outliers) 138.6 (90.0-234.0) Humeral component Stem diameter 3 mm 1 4 mm 21 5 mm 11 6 mm 13 Stem length 100 mm 41 150 mm 4 Not available 1 Ulnar component Stem diameter 2.5 mm (custom) 1 3 mm 24 4 mm 14 5 mm 7 Stem length 75 mm 40 84 mm (custom) 1 115 mm 5 Bone graft used under humeral flange (No. of elbows) Yes 3 No 43 Cemented fixation (No. of 46 elbows) ) Outliers: (1) concurrent wrist reconstruction procedure, (2) revision elbow with preoperative radial nerve palsy, and (3) fractured lateral condyle.
mean of 4.1 years. Details of these 4 revisions are presented below. As described, the study cohort was limited to elbows with a minimum of 2 years of follow-up. One elbow that was excluded because it had follow-up of <2 years is separately reported here because it was revised. The male patient, treated with TEA at age 70 years for post-traumatic arthritis with humeral fracture, presented with pain and underwent revision of the loose 100-mm 4-mm humeral component, along with the condyles and bushing, at 4.6 years.
Clinical outcomes There was significant improvement (P < .001) in all patient, physician, and range-of-motion components of the
e99 ASES Elbow Assessment Form (Table IV). In particular, mean patient assessments of pain improved from 36.5 to 4.0 points and function from 14.6 to 29.6 points (P < .001). Also, mean patient satisfaction with the procedure was high (9.1 points of a maximum of 10). Mean physician assessments of signs (pain/tenderness), stability, and strength improved from 16.2 to 1.2 points, 1.9 to 0.1 points, and 14.9 to 19.8 points, respectively (P < .001). Finally, the mean flexion-extension arc increased from 81 to 121 and the pronation-supination arc increased from 134 to 163 (P < .001).
Radiographic outcomes Figure 3 and Figure 5 define the lateral view and AP view radiographic zones that were tracked during follow-up. The clinically relevant zones are those responsible for cement fixation; that is, the stem regions of the humeral and ulnar components (humeral lateral view zones 1-8 and AP view zones 2-6, and ulnar lateral view zones 2-8 and AP view zones 2-6). Radiolucencies that developed after 3 months were observed in 7 of 46 elbows (15.2%) and were generally nonprogressive. Only 4 (8.7%) had relevant radiolucencies at last follow-up: (1) humeral AP view CP zone 3, (2) ulnar AP view BC and CP zone 2, (3) humeral lateral view BC zone 1, and (4) ulnar lateral view BC zones 6 and 7 as well as ulnar AP view BC zones 4 and 6. Thus, of the 7 radiolucencies present at last follow-up, 5 were associated with the ulnar component and 2 with the humeral component. One of the 46 humeral components (2.2%) was radiographically loose at 5 years but was not revised (see below). By contrast, complete cement fill does not occur in the humerus distal to the stem or in the ulna proximal to the stem. Radiolucencies in these regions are unassociated with implant loosening but are of interest because they can show how the bone responds to stress shielding here. For instance, Figure 6 shows 4.4-year radiographs of a female rheumatoid patient who underwent TEA at age 72 years, with 4-mm radiolucencies evident in zones 1 and 7 (AP view) of the distal 100-mm 6-mm humeral component. This well functioning elbow was stable, with normal strength and no pain and with a flexion/extension arc of 105 and a pronation/supination arc of 145 . The completeness of the CP (zones 1 and 2) and BP (zones 4 and 5) interfaces by the humeral component anterior flange at last follow-up are listed in Table V. The CP interface was 76% to 100% complete in >95% of elbows, whereas the BP interface was 76% to 100% complete in only about 54% to 61%. The shape of the humerus does not allow for, in most instances, continuous complete contact of the anterior flange to the bone. The CP and BC contact at the radius junction where the anterior flange meets the body of the humeral component (zone 3) was evident in 46% and in 94% of the elbows, respectively.
e100 Table IV
H. Hastings II et al. American Shoulder and Elbow Surgeons Elbow Evaluation Form results
Measurement Patient ASES Pain Function Satisfaction Physician ASES Signs) Stability Strength Range of motion, Flexion Extension Flexion-extension arc Pronation Supination Pronation-supination arc
Preoperative
Postoperative
Average (range)
Average (range)
36.5 (6-50) 14.6 (0-30) N/A
4.0 (0-28) 29.6 (7-36) 9.1 (2-10)
16.2 (1-39) 1.9 (0-6) 14.9 (0-20)
1.2 (0-12) 0.1 (0-3) 19.8 (16-20)
<.001 <.001 <.001
120 39 81 72 63 134
143 22 121 84 80 163
<.001 <.001 <.001 <.001 <.001 <.001
(50-150) (0-80) (20-145) (10-95) (0-95) (15-190)
(105-155) (5 to 90) (45-155) (55-95) (50-90) (125-180)
P
<.001 <.001 N/A
ASES, American Shoulder and Elbow Surgeons. ) All 13 types of pain/tenderness signs were rated on scale of 0 to 3 points for a total worst pain/tenderness score of 39 points.
Figure 6 (Left) Lateral and (Right) anteroposterior radiographs at 4.4 years of a rheumatoid patient with bone resoprtion/remodelling in the intercondylar uncemented portion of the humeral component.
Complications
Device-related complications
At least 1 complication occurred in 20 of the 46 elbows (43.5%). These complications were categorized as devicerelated and non–device-related.
In general, device-related complications included the exchange, revision, or loosening of any of the elbow components, including stems, ulnar bearing, and condylar
Clinical study of the Discovery elbow Table V
Radiographic flange contact in zones 1 to 5
Completeness of contact
0%-25% 26%-50% 51%-75% 76%-100%
e101
Prosthesis-Cement
Prosthesis-Cement
Bone-Cement
Bone-Prosthesis
Zone 1
Zone 2
Zone 3
Zone 3
Zone 4
Zone 5
No. (%)
No. (%)
No. (%)
No. (%)
No. (%)
No. (%)
Yes: 21 (45.7) No: 25 (54.3)
Yes: 43 (93.5) No: 3 (6.5)
12 4 2 28
18 2 1 25
1 (2.2) 1 (2.2) 44 (95.6)
1 (2.2) 45 (97.8)
locking screws. Condylar screw loosening occurred in 3 elbows (6.5%): 1. A loose screw was found in a patient undergoing surgery at 2.0 years after TEA for triceps insufficiency. The screw was exchanged, with reactive bursitis in response to the loose screw noted. 2. The medial locking screw in 1 elbow backed out without loosening of the humeral condyle. Both screws were exchanged at 2.1 years. The bearing had mild wear but was not replaced. 3. The radiograph at 5 years showed that a locking screw was beginning to back out, but no intervention was performed. Bearing or condyle-only exchange, or both, was performed in 2 elbows (4.3%). One patient had received her Discovery elbow during revision of a failed Mayo-modified Coonrad implant. No condyles remained to support the implant. After multiple falls, the locking pin disengaged and became floating, resulting in grinding wear on the loose polyethylene bearing. The pin and the worn bearing were exchanged at 4.5 years. In the second patient, a deep infection related to a hematogenous spread from an infected spinal cord stimulator was debrided at 6.2 years, with exchange of the condyles and bearing. No gross wear pattern on the bearing was noted. One elbow (2.2%) developed loosening of the humeral component. A female patient, who originally underwent TEA due to post-traumatic arthritis at the age of 72 years, had persistent loosening of the 100-mm 6-mm humeral component at 5.0 years; however, revision was not performed, with the patient rating high satisfaction (10 points) and an ASES score that remained improved compared with its preoperative value at this time. At 6.5 years, a periprosthetic humeral fracture was repaired with plate and screw fixation. One patient (2.2%) described occasional ‘‘squeaking’’ of the elbow; nevertheless, giving a 5-year satisfaction rating of 10 points.
Non–device-related complications Complications that were not directly related to the device included 6 (13.0%) wound-healing complications, 3 cases (6.5%) of bursitis, 2 superficial infections (4.3%), 2 (4.3%)
(26.1) (8.7) (4.3) (60.9)
(39.1) (4.3) (2.2) (54.3)
periprosthetic fractures, 1 (2.2%) deep infection (see above), and in 1 elbow (2.2%) each, posteromedial olecranon impingement, elbow contracture, posterior interosseous nerve palsy, ulnar neuropathy, ruptured triceps tendon, intraoperative fracture, and intraoperative perforation.
Discussion TEA, although effective, has not enjoyed the same overall success as its hip and knee counterparts, partly because of the challenging anatomy of this joint.17,29 Although semiconstrained designs can replicate normal elbow kinematics, be used in the presence of bone or ligamentous deficiency, and limit force transfer across the BC interface to mitigate loosening,2,10,27 high complication and revision rates, or both,6,10,11,17,24,25,27 remain. Thus, TEA designs continue to evolve to improve patient outcomes. The Discovery elbow is a semiconstrained device with a design that includes a unique bearing and anatomic modifications to the humeral and ulnar stems. Our purpose was to report on the 4-year (mean) results of a prospective, multicenter study investigating this device. Patient self-evaluation of pain and function and surgeonassessed signs, stability, and strength significantly improved (P < .001). Patient satisfaction was high, with a mean score of 9.1 (10 ¼ very satisfied). The mean flexionextension arc improved by 40 (P < .001). By way of comparison, Van der Lugt and Rozing29 performed a systematic review of the literature, finding mean increase of flexion-extension arc in the range of 12 to 39 for a variety of nonconstrained and semiconstrained TEA prostheses. Implant survivorship is fundamental to the performance of any joint replacement device. Techniques such as the Kaplan-Meier method to estimate survivorship13 take into account the frequency and timing of failures as well loss to follow-up, deaths, and other variables to obtain an estimate of the cumulative survivorship over time. This helps quantify not only the proportion of devices that have failed but also whether these failures tend to occur early or late, knowledge of which can help to determine the cause. The Kaplan-Meier method is best suited for relatively large populations due to the statistical instability that can result when too few patients remain at risk.4 In general, TEA study populations are not large, which may help explain why most cite gross survivorship, or the proportion of
e102 devices at the beginning of the study that have not failed by final follow-up, with ‘‘failure’’ defined as the investigator sees fit to reflect the nature or purpose of the study. Accordingly, care must be taken to ensure that survivorship comparisons among studies are valid. In our study of 46 elbows with complete preoperative evaluation and a minimum of 2 years of postoperative follow-up, gross survivorship based on revision of individual components was 100% for the humeral and ulnar components, 97.8% for the condyles, 95.7% for the bearing, and 95.7% for the locking screws at a mean of 4.1 years. Four of the 46 elbows underwent revision of at least 1 component, for an overall gross survivorship of 91.3%. Table VI lists semiconstrained elbow survivorships for several short-term (2-4 years), medium-term (5-9 years), and long-term (10 years) studies, including reasons for revision. All are gross survivorships, with the exception of Gill and Morrey,6 Shi et al,26 and Mansat et al,19,20 who computed Kaplan-Meier estimates. Short-, mid-, and longterm survivorship studies involving mainly primary Coonrad/Morrey elbows, generally using revision of any component (humeral, ulnar, and bushing components) as the end point, were 70.5% to 96%,2,11,16,21 63.2% to 100%,1,6,18,19,23,26 and 92%,6,20 respectively. Similarly, midterm and long-term survivorships for primary GSB III Elbow System (Zimmer, Warsaw, IN) were 71% to 92.8%12,14,25 and 87.7%,7 respectively. Finally, midterm survivorship for primary Solar elbows was 46.2%.24 Thus, the 4.1-year (mean) survivorship of the Discovery elbow was within the upper end of the short- to midterm range cited by others for semiconstrained devices.2,11,16,17,21 It is important, however, to consider variation in patient demographics and primary diagnoses before drawing firm conclusions because this can influence survivorship.1 Comparing overall complication rates among studies can be difficult because some cite all reported complications, including those of a general surgical nature, whereas others limit focus to major complications or only those that are relevant to the unique aspects of TEA. Table VI summarizes the complications reported in 16 studies for 595 semiconstrained TEA devices (561 primary, 34 revision), including 449 Coonrad/Morrey,1,2,6,11,16,18-21,23,26 130 GSB III,7,12,14,25 13 Solar,24 and 3 Discovery2,11 total elbows. Reported complication rates ranged from 14% to 72%. Overall, there were 44 (8.9%) instances of loosening, 26 (4.4%) of ulnar neuropathy, 11 (1.8%) of disassembly, 25 (4.2%) of bushing wear/failure, and 3 (0.5%) of component fracture. Deep infections required revisions in 17 (2.9%) elbows. Kelly et al,14 summarized the complication rates from 6 semiconstrained TEA series, citing the frequency of aseptic loosening (4%-50%), dislocation (0%-14%), triceps avulsion (2%-11%), ulnar neuropathy (0%-11%), and deep infection (3%-8%). In our series, although at least 1 complication occurred in 20 elbows (43.5%), many were general surgical in nature, such as wound complications (12.8%), superficial
H. Hastings II et al. infections (4.3%), and deep infections (2.2%). Aseptic loosening of the humeral component within the study population occurred in 1 patient who has remained asymptomatic and has not required revision. However, revision of a loose humeral component, along with its associated condyles and bearing, was performed in 1 elbow that did not fit the study population inclusion criteria but is nevertheless reported here due to the severity of this complication. Osteolysis and aseptic loosening are a concern within the first few years after TEA, but this does not necessarily result in significant symptoms because most of these patients are low demand and do not always proceed to severe loosening or dislocation.18 In our series, radiolucencies were minor in all but 1 elbow and only evident in small areas of the reporting grid zones. Other complications involving implant hardware included 3 cases (6.4%) of locking screw loosening, 1 (2.2%) bearing failure associated with excessive use, and 1 deep infection (2.2%) resulting in condyle/bushing exchange. There were no cases of disassembly and only 1 case (2.1%) of ulnar neuropathy that was secondary to surgeries that preceded TEA. It is important to note that 43 of 46 elbows, including the 3 that incurred loose locking screws, were implanted during 2003 to 2007. This was before a 2008 change in the design of the screw-drive mechanism from a Phillips-type to a startype that allows greater torque application. Screw loosening has not been observed with the second-generation screw design. Thus, the complication profile for the Discovery elbow from the current study is well within the range cited by other clinical studies of semiconstrained TEA. The surgical approach used in 57% of the elbows in this series was a triceps-off variant, which we continue to use. The hinge components appear to be robust in the light of the 3 cases of locking screw loosening being associated with the first-generation screw-drive design and the exchanges of condyles and/or bushings being associated with deep infection, a history of falling, or a revised loose humeral component. One of this study’s strengths was that it was prospective and multicenter, with 4 sites and 6 surgeons contributing cases. Another was that it included a study population of 46 elbows, which avoids some of the limitations of markedly smaller populations characteristic of some TEA studies.2,11,12,14,16,18,19,24,25 There were, however, some limitations that warrant mention. First, 32 of 47 of the TEAs reported here were performed by the developing surgeon (H.H.), which could conceivably reduce the study’s general usefulness. (This study received industry funding.) However, of the 20 elbows that experienced complications, 13 were performed by the developing surgeon and 7 were performed by the group of 5 other surgeons, for rates of elbows experiencing complications of 41% (13 of 32) and 50% (7 of 14), respectively. Also, the means of the flexion-extension arc increases for the developing surgeon and the group of other
Summary of survivorship analyses for semiconstrained elbow prostheses
Ref.
Elbow (Qty)
Condition
Survivorship
Complications [reported overall rate] (Qty)
Revision reasons (Qty)
11
C/M) (17); Discoveryy (2)
Various
2
C/M) (17); Discoveryy (1)
RA (14), OA (2), PA (1), GA (1)
74% at mean 40.4 mon 70.5% for C/M only 94.4% at mean 55 mon 94.1% for C/M only
Five C/M due to bushing failure (4) and loosening (1) One C/M due to aseptic loosening
6
C/M) (78)
RA
94.4% at 5 y)) 92.4% at 10 y))
23
C/M) (49)
Distal humeral fx s
90% at mean w7 y
[26%] Bushing failure (4) and loosening (1) [22%] Stiffness (1), transient ulnar nerve palsy (1), wound infection (1), aseptic loosening (1) [14%] Triceps avulsion (3), deep infection (2), component fx (1), post-op ulna fx (2), aseptic loosening (2) [35%] Hematoma (3), wound dehiscence (2), infection (1), injury from fall (3), plus several general surgical complications
21
C/M) (25)
Displaced, intra-articular distal humeral fx s in patients >65 y
96% at 2 y
18
C/M) (31)
RA (27), fracture (3), OA (1)
90.6% survival at 124 mon (mean, 5.5 mon)
1
C/M) (55)
JRA (21 pts) RA (15) PTA (19 pts) in patients <40 y
JRA: 90.5% at mean 7.6 y RA: 80% at mean 7.3 y PTAO: 63.2% at mean 8.5 y
[72%] Ulnar nerve sensory (3) and motor deficit (2), radial nerve palsy (1), wound complication (4), posttraumatic stiffness (3), malalignment (1), deep (1) and superficial infection (1), HO (3). [20%] Ulnar nerve palsy (1), ulnar neurapraxia (1), loose hinge (1), infection (1), loosening (2) [25%] Superficial infection (2), deep infection (2), aseptic loosening (4), triceps insufficiency (3), bushing wear (3)
Reimplantation due to infection (1), component fx (1), aseptic loosening (2)
Clinical study of the Discovery elbow
Table VI
Five due to aseptic loosening (1), septic loosening (1), component fx from fall (2), and 1 revised twice, first due to a fall, then due to aseptic loosening One for deep infection
Two for loosening
JRA: triceps insufficiency (2). RA: loosening (1), bushing wear (1), deep infect (1). PTA: triceps insufficiency (1), bushing wear (2), loosening (3), deep infect (1) (continued on next page)
e103
e104
Table VI (continued ) Ref.
Elbow (Qty)
Condition
Survivorship
Complications [reported overall rate] (Qty)
Revision reasons (Qty)
26
C/M) (67)
RA, PTA, trauma, OA, HA 37 primary 30 revision
Primary: 72% at 5 y)) Revision: 64% at 5 y)) based on 11 failed knees in each group
Primary: 10 revised/resected for periprosthetic fx (2), loosening (5), infect (3) Revision: 5 re-revised for loosening (3), worn bushing (1), periprosthetic fx (1)
16
C/M) (17)
PTA
94.1% at mean of 32 m
19
C/M) (15)
RA (8), PTA (5), PA (1), JA (1) 11 primary 4 revision
100% at 5 y)), 90% at 10 and 13 y))
20
C/M) (78)
IA (45), trauma (33)
97.7% at 5 y)) 91.0% at 10 y)) based on aseptic loosening 92% based on component revision
Primary: [30%] flexion contracture (6), ulnar neuropathy (2), triceps avulsion (1), loosening (7), periprosthetic fx (1) Revision: [37%] flexion contracture (3), ulnar neuropathy (4), intra-op humeral fx (2), triceps avulsion (1), loosening (8), periprosthetic fx (2), worn bushing/dislocation (1) [29%]: Minor: stiffness (1), superficial infection (1) Major: ulnar nerve dysesthesia (1), component loosening (3), including loose humeral component after deep infection, 1 loose humeral component after a fall, and 1 aseptic loosening of ulnar component. No bushing wear observed Overall: [67%] triceps insufficiency (3), ulnar nerve (2) and radial nerve (1) complications, deep infection (2), severe bushing wear (1), bone fxs (2) [34%] ulnar nerve (7), radial nerve (3), triceps weakness (5), humeral looseness (1), ulnar looseness, humeral and ulnar looseness (2), deep infection (3), extra small ulnar component fx (1), bone fx (1), locking pin failure (1). Bushing wear observed in 14
One patient scheduled for revision for aseptic loosening of ulnar component
Deep infection (1), failed hinge (2 revisions same patient)
6 prostheses were revised for deep infection (3), loosening (2), broken pin (1)
H. Hastings II et al.
GSB IIIz (14)
RA (9) PTA (5)
71% at mean 6 y
14
GSB IIIz (28)
RA (22 pts) PA (1 pt)
92.8% at mean 7.6 y
7
GSB IIIz (65)
RA (51 pts) PTA (8 pts)
87.7% at mean w13.5 y, but last f/u interval not provided for 29 elbows
12
GSB IIIz (23)
RA
82.6% at mean w5 y
24
Solarx (13)
PA (6) RA (6) GA (1)
46.2% at mean 8.4 y
(18%) [43%] Aseptic loosening (4), disassembly (2), device fx (1), infection (1), triceps tendon rupture (1), post-op stiffness (2) [39%] Deep infection (1), PE wear/osteolysis (1), ulnar nerve palsy (1), acute pain (1), partial triceps avulsion (3), superficial infect (2), median nerve paresthesia (1), post-op distal humeral fx (1) [32%] Aseptic loosening (3), deep infection (3), disassembly (9), triceps disruption (1), extensor deficit (1), hyperesthesia (2), intra-op humeral fx (1), postop ulnar condyle fx [35%] Aseptic loosening (1), malposition (2), intra-op fx (4) [54%] Motor ulnar nerve palsy (1), loosening (4), bushing failure (1), deep infection (1)
4 elbows revised for aseptic loosening
2 revisions for infect (1) and PE wear with osteolysis (1)
8 revisions for aseptic loosening (1), disassembly (4), deep infection (3)
Clinical study of the Discovery elbow
25
4 revisions for aseptic loosening (1), malposition (2), and disassembly (1) 7 revisions for loosening (4), bushing failure (2), deep infection (1)
fx, fracture; C/M, Coonrad/Morrey; GA, gouty arthritis; HA, hemophiliac arthropathy; HO, heterotopic ossification; IA, inflammatory arthritis; JA, juvenile arthritis; JRA, juvenile rheumatoid arthritis; OA, osteoarthritis; PA, psoriatic arthritis; PE, polyethylene; PTO, post-traumatic arthritis; RA, rheumatoid arthritis. ) Coonrad/Morrey Total Elbow, Zimmer, Warsaw, IN, USA. y Discovery Elbow System, Biomet, Warsaw, IN, USA. z Zimmer, Warsaw, IN, USA. x Solar Elbow System, Stryker, Mahwah, NJ, USA. )) Kaplan-Meier survivorship estimate.
e105
e106 surgeons were 39 and 44 , respectively, whereas the means of the pronation-supination arc increases were 28 and 30 , respectively. Thus, there was no overt evidence of surgeon bias, nor did there appear to be any influence of a learning curve, because all surgeons were very experienced with TEA before beginning the study. Second, there were several primary diagnoses among the patients treated, with nearly half of the elbows presenting with rheumatoid arthritis. However, only a small number of elbows presented with each of the other diagnoses, so the functional dependence of outcomes on diagnosis could not be meaningfully performed. Third, follow-up was relatively short at a mean of 4 years, which was sufficient to characterize functional outcome but was not adequate to investigate late complications or longer-term survivorship. Future work should focus on longer-term follow-up of a larger population of patients receiving the Discovery elbow using locking screws having the second-generation star-drive mechanism.
Conclusion Historically, the hinge has been one of the most problematic aspects of semiconstrained total elbow designs. This multicenter, prospective clinical study was performed to examine the outcomes and complication profile of a semiconstrained total elbow prosthesis with a spherical bearing. Forty-six Discovery elbows with 4.1year mean follow-up (minimum, 2 years) showed significant improvement in the ASES elbow score, high patient satisfaction, and a gross survivorship of 91.3% (42 of 46) that considers revision or exchange of any component for any reason. Revisions in 2 of the 4 elbows were limited to the exchange of first-generation condylar locking screws that had loosened. This has not been observed with the second-generation screw design. No bearings were revised due to wear with the exception of one that had become damaged secondary to interaction with a locking pin that had become free. Within this series, no humeral or ulnar components were revised. However, an additional patient was reported who did not qualify for inclusion in the study cohort but did undergo revision of a loose humeral component along with the bearing and condyles. Within this context, the spherical design of the condyles and bearing of the Discovery elbow appears to be relatively robust; however, longerterm studies will be required to verify this.
Disclaimer Funding for this study was provided by Biomet Inc, Warsaw, IN, USA, for purposes of data collection.
H. Hastings II et al. Hill Hastings II receives a consulting fee/honorarium, travel support, and royalties from Biomet, and his institution receives grant money from Biomet. Donald H. Lee receives speaker fees, consulting fees, and royalties from Biomet and his institution received funding for participation in the Discovery elbow clinical study. William S. Pietrzak is an employee of Biomet.
References 1. Celli A, Morrey BF. Total elbow arthroplasty in patients forty years of age or less. J Bone Joint Surg Am 2009;91:1414-8. http://dx.doi.org/ 10.2106/JBJS.G.00329 2. Corradi M, Frattini M, Panno B, Tocco S, Pogliacomi F. Linked semiconstrained total elbow prosthesis in chronic arthritis: results of 18 cases. Musculoskelet Surg 2010;94(Suppl 1):S11-23. http://dx.doi.org/ 10.1007/s12306-010-0070-y 3. Dee R, Sweetnam DR. Total replacement arthroplasty of the elbow joint for rheumatoid arthritis: two cases. Proc R Soc Med 1970;63:654-5. 4. Dorey F, Amstutz HC. Survivorship analysis in the evaluation of joint replacement. J Arthroplasty 1986;1:63-9. 5. Garrett JC, Ewald FC, Thomas WH, Sledge BS. Loosening associated with G.S.B. hinge total elbow replacement in patients with rheumatoid arthritis. Clin Orthop Rel Res 1977;127:170-4. 6. Gill DR, Morrey BF. The Coonrad-Morrey total elbow arthroplasty in patients who have rheumatoid arthritis. A ten to fifteen-year follow-up study. J Bone Joint Surg Am 1998;80:1327-35. 7. Gschwend N, Scheier NH, Baehler AR. Long-term results of the GSB III elbow arthroplasty. J Bone Joint Surg Br 1999;81:1005-12. 8. Gschwend N, Simmen BR, Matejovsky Z. Late complications in elbow arthroplasty. J Shoulder Elbow Surg 1996;5:86-96. 9. Hastings H 2nd. Linked total elbow arthroplasty. The Discovery prosthesis. In: Yamaguchi K, King GJW, McKee M, O’Driscoll SW, editors. Advanced reconstruction: elbow. Rosemont, IL: American Academy of Orthopaedic Surgeons; 2007. p. 229-40. ISBN No. 9780892033911. 10. Hastings H 2nd. Minimally constrained elbow implant arthroplasty: the Discovery elbow system. Tech Hand Extrem Surg 2004;8:34-50. 11. Hastings H 2nd, Theng CS. Total elbow replacement for distal humerus fractures and traumatic deformity: results and complications of semiconstrained implants and design rationale for the Discovery elbow system. Am J Orthop (Belle Mead NJ) 2003;32(9 Suppl):20-8. 12. Jensen CH, Jacobson S, Ratchke M, Sonne-Holm S. The GSB III elbow prosthesis in rheumatoid arthritis. A 2- to 9-year follow-up. Acta Orthop 2006;77:143-8. http://dx.doi.org/10.1080/17453670610045830 13. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Amer Statist Assn 1958;53:457-81. 14. Kelly EW, Coghlan J, Bell S. Five- to thirteen-year follow-up of the GSB III total elbow arthroplasty. J Shoulder Elbow Surg 2004;13:43440. http://dx.doi.org/10.1016/j.jse.2004.01.028 15. King GJ, Richards RR, Zuckerman JD, Blasier R, Dillman C, Friedman RJ, et al. A standardized method of assessment of elbow function. Research Committee, American Shoulder and Elbow Surgeons. J Shoulder Elbow Surg 1999;8:351-4. 16. Kodde IF, van Riet RP, Eygendaal D. Semiconstrained total elbow arthroplasty for posttraumatic arthritis or deformities of the elbow: a prospective study. J Hand Surg Am 2013;38:1377-82. http://dx.doi. org/10.1016/j.jhsa.2013.03.051 17. Little CP, Graham AJ, Carr AJ. Total elbow arthroplasty. A systematic review of the literature in the English language until the end of 2003. J
Clinical study of the Discovery elbow
18.
19.
20.
21.
22.
Bone Joint Surg Br 2005;87:437-44. http://dx.doi.org/10.1302/0301620X.87B4.15692 Maheshwari R, Vaziri S, Helm RH. Total elbow replacement with the Coonrad-Morrey prosthesis: our medium to long-term results. Ann R Coll Surg Engl 2012;94:189-92. http://dx.doi.org/10.1308/ 003588412X13171221589775 Mansat P, Bonnevialle N, Rongieres M, Mansat M, Bonnevialle P. Results with a minimum of 10 years follow-up of the Coonrad/Morrey total elbow arthroplasty. Orthop Traumatol Surg Res 2013;99:S33743. http://dx.doi.org/10.1016/j.otsr.2013.07.002 Mansat P, Bonnevialle N, Rongieres M, Mansat M, Bonnevialle P. Experience with the Coonrad-Morrey total elbow arthroplasty: 78 consecutive total elbow arthroplasties reviewed with an average 5 years of follow-up. J Shoulder Elbow Surg 2013;22:1461-8. http://dx. doi.org/10.1016/j.jse.2013.07.042 McKee MD, Veillette CJH, Hall JA, Schemitsch EH, Wils LM, McCormack R, et al. A multicenter, prospective, randomized controlled trial of open reduction-internal fixation versus total elbow arthroplasty for displaced intra-articular distal humeral fractures in elderly patients. J Shoulder Elbow Surg 2009;18:3-12. http://dx.doi. org/10.1016/j.jse.2008.06.005 Morrey BF. Complications of elbow replacement surgery. In: Morrey BF, editor. The elbow and its disorders. 3rd ed. Philadelphia: WB Saunders; 2000. p. 667-76. ISBN-10 No. 07216775725 ISBN-13 No. 978–0721677521.
e107 23. Muller LP, Kamineni S, Rommens PM, Morrey BF. Primary total elbow replacement for fractures of the distal humerus. Oper Orthop Traumatol 2005;17:119-42. http://dx.doi.org/10.1007/s00064-0051125-3 24. Patil N, Cheung EV, Mow CS. High revision rate after total elbow arthroplasty with a linked semiconstrained device. Orthopedics 2009; 32:321. 25. Schneeberger AG, Hertel R, Gerber C. Total elbow replacement with the GSB III prosthesis. J Shoulder Elbow Surg 2000;9:135-9. 26. Shi LL, Zurakowski D, Jones DG, Koris MJ, Thornhill TS. Semiconstrained primary and revision total elbow arthroplasty with use of the Coonrad-Morrey prosthesis. J Bone Joint Surg Am 2007;89:146775. http://dx.doi.org/10.2106/JBJS.F.00715 27. Szekeres M, King GJ. Total elbow arthroplasty. J Hand Ther 2006;19: 245-54. http://dx.doi.org/10.1197/j.jht.2006.02.010 28. Trail IA, Nuttall D, Stanley JK. Survivorship and radiological analysis of the standard Souter-Strathclyde total elbow arthroplasty. J Bone Joint Surg Br 1999;81:80-4. 29. Van der Lugt JCT, Rozing PM. Systematic review of primary total elbow prostheses used for the rheumatoid elbow. Clin Rheumatol 2004;23:291-8. http://dx.doi.org/10.1007/s10067-004-0884-9 30. Vishwanathan K, Alizadehkaiyat O, Kemp GJ, Forstick SP. Responsiveness of the Liverpool Elbow Score in elbow arthroplasty. J Shoulder Elbow Surg 2013;22:312-7. http://dx.doi.org/10.1016/j.jse. 2012.09.003
www.elsevier.com/locate/ymse
A prospective multicenter clinical study of the Discovery elbow Hill Hastings II, MDa, Donald H. Lee, MDb, William S. Pietrzak, PhDc,d,* a
Indiana Hand to Shoulder Center, Indianapolis, IN, USA Vanderbilt University School of Medicine, Nashville, TN, USA c Biomet Inc, Warsaw, IN, USA d Department of Bioengineering, University of Illinois at Chicago, Chicago, IL, USA b
Background: Semiconstrained total elbow arthroplasty is used to improve elbow function and reduce pain. Although effective, high complication rates exist, with the polyethylene bushing especially susceptible to failure. The Discovery Elbow System (Biomet Inc, Warsaw, IN, USA) contains a spherical bearing designed to minimize polyethylene wear. This prospective, multicenter clinical study investigated the 4year (mean) outcomes of this elbow. Methods: From 2002 to 2009, 92 patients (71 women, 21 men; mean age, 63.9 years; range, 33.4-88.7 years) received 99 Discovery elbows at 4 centers. The study cohort was limited to 46 elbows with complete preoperative and minimum 2-year clinical (modified American Shoulder and Elbow Surgeons elbow score) and radiographic follow-up. Results: Mean follow-up was 4.1 years (range, 2-5.9 years). All American Shoulder and Elbow Surgeons elbow score components improved significantly (P < .001). Mean flexion-extension arcs increased from 81 to 121 and pronation-supination arcs from 134 to 163 (P < .001). Loose locking screws in 2 elbows (first-generation screws), a loose polyethylene bearing in 1 (history of falls), and a condyle/bearing in 1 (deep infection) were exchanged. Among the 46 elbows, gross survivorship was humeral/ulnar components, 100%; condyles, 97.8%; bearings, 95.7%; and screws, 95.7%. One humeral component (2.2%) was radiographically loose but not revised. An additional elbow (elbow 47) that did not meet the criteria for inclusion (<2 years of follow-up) was revised due to a loose humeral component and was reported separately. Conclusion: The Discovery elbow increased function and decreased pain with high survivorship at a mean of 4.1 years. Level of evidence: Level IV, Case Series, Treatment Study. Ó 2014 Journal of Shoulder and Elbow Surgery Board of Trustees. Keywords: Discovery elbow; outcomes; prostheses; semiconstrained; survivorship; total elbow arthroplasty
The following Investigational Review Board approvals are relevant to this study: St. Vincent, Indianapolis, IN, USA (R2002-014); Partners Healthcare, Boston, MA, USA (2001P001731/MGH); Vanderbilt University, Nashville, TN, USA (060092); and the University of South Florida, Tampa, FL, USA (102165). *Reprint requests: William S. Pietrzak, PhD, 1691 S Meadow Dr, Warsaw, IN 46581, USA. E-mail address: [email protected] (W.S. Pietrzak).
Proper elbow motion is more critical to upper-extremity function than that of any other associated joint.11 For instance, a 100 arc of motion is required to accomplish the activities of daily living.2,10 Since the 1970s, total elbow arthroplasty (TEA) has been increasingly used as a salvage technique to decrease pain, increase joint stability, and
1058-2746/$ - see front matter Ó 2014 Journal of Shoulder and Elbow Surgery Board of Trustees. http://dx.doi.org/10.1016/j.jse.2013.12.033
e96 improve overall range of motion.10 The first TEA prostheses were fully constrained and hinged, with metal-onmetal articulation and no significant varus or valgus laxity.3,10 Unfortunately, these led to metallic particulate synovitis, loosening rates as high as 45% due to force transmittance through the hinged articulation to the bone–cement interface, and implant failure.5,10,29 In response, semiconstrained and nonconstrained prostheses were developed and have been used exclusively during the past decade.29 Unlinked, or minimally constrained, TEA prostheses are elbow resurfacing devices that can allow for a more anatomic ulnohumeral articulation.29 Although such designs may reduce the risk of loosening, they may increase the risk of dislocation and have limited application in cases of bone loss or ligamentous deficiency.10 Semiconstrained TEA prostheses contain a hinged linkage to help resist dislocation and is ‘‘sloppy,’’ in that it allows for some varus/valgus motion, better replicating the kinematics of the natural elbow, and can be used in the presence of bone or ligamentous deficiency.10,27 Also, force transfer across the bone–cement interface is lower than for constrained devices,27,29 thus reducing the potential for loosening. TEA is effective, but the results generally have not been as good as those for total hip or knee replacement,17,29 although rheumatoid elbow patients are an exception because patients with diffuse disease place less demand on their elbow.6,8,17,27-29 TEA tends to be associated with a high rate of complications (eg, 14%-80%),17 partially due to the difficult nature of surgery in this complex joint with thin soft tissue coverage.2 In particular, deep infection and septic loosening are a concern, with rates as high as w10% reported.2,22 Ulnar neuropathy has also been reported to occur in up to 21% of patients within a few days of surgery, with the rate of permanent dysfunction ranging from 0% to 10%.26 High rates of polyethylene bushing failure/wear have been reported for semiconstrained prostheses.6,10,11,19,20,24,27 For instance, rates of 14% to 47% have been cited for the Coonrad/Morrey prosthesis (Zimmer, Warsaw, IN, USA)6,11,19,20 and 15% for the Solar Elbow System (Stryker, Mahwah, NJ, USA).24 Semiconstrained devices can also experience hinge failure, as noted by a systematic review by Little et al,17 to occur in 6% of patients. The Discovery Elbow System (Biomet Inc, Warsaw, IN, USA) was developed to address some of the issues associated with semiconstrained TEA prostheses, including 1 spherical bearing designed to reduce wear and allow for simple polyethylene exchange, if required.10,11 Previous clinical reports on this system have been limited by small numbers of patients (ie, 1-2),2,11 short mean follow-up of <2 years,30 or the provision of only a brief, overall clinical summary.9 As such, a need remains for a detailed, clinical assessment of this system. Our purpose was to (1) determine the short-term to midterm survivorship of the Discovery elbow and (2) report the clinical and radiographic outcomes
H. Hastings II et al. of a multicenter study, comparing these results with those reported for other semiconstrained TEA prostheses.
Materials and methods The Discovery Elbow System The Discovery Elbow System has been described by Hastings and Theng11 and Hastings.9,10 Lateral and posterior views of the prosthesis are shown in Figures 1 and 2, respectively. Two cobalt chromium molybdenum condylar hemispheres lock into the distal humeral component with a medial and lateral Ti6Al4V screw, respectively. These articulate with congruent ArCom polyethylene (Biomet Inc) bearing surfaces captured within the proximal ulnar component with a locking pin, providing 7 of varus/valgus laxity. The humeral stem is made of Ti6Al4V and contains a posterior bow, including a 5 lateral offset and 5 internal rotation to reproduce the anatomy. The distal humeral flange resides outside of the canal and hooks around the distal anterior humeral cortex to resist posteriorly directed forces of the implant in the distal humerus. The flange is designed to sit flush against the distal anterior humeral cortex to obviate the need for a bone graft and potential graft resorption. The ulnar stem, made of the same alloy, contains a lateral bow with a 23 anterior neck angle and a lateral offset to properly position the joint axis in an anatomic position. The humeral and ulnar components can be first cemented into place and later coupled together. Alternatively, the humeral and ulnar components can be preassembled and then implanted.
Clinical study design Four clinical centers participated in this prospective study. Written informed consent was obtained from each patient. Inclusion criteria were noninflammatory joint disease, including osteoarthritis and avascular necrosis; inflammatory arthritis; revision where other devices or treatments have failed; correction of functional deformity; and treatment of acute fractures or nonunion about the elbow. Exclusion criteria included patients aged <18 years, pregnancy, metabolic disorders that could impair bone formation, marked bone loss, active or suspected infection about the elbow or distant foci of infections that could spread to the implant site, unwillingness or inability to comply with the rehabilitation program, and factors that could limit the patient’s ability to conform to the prescribed follow-up schedule. Hastings9 described the technique for implanting the Discovery elbow using 3 surgical approach options: triceps-off (reflecting), triceps-off (splitting), and triceps-on (sparing). Postoperatively, patients with primary arthroplasty were immobilized in a bulky dressing and splint for 5 days. If the wound was sealed at that time, active, passive range of motion was initiated. An extension splint was used at night as needed to maintain full elbow extension. Activities of daily living were encouraged. In cases in which the triceps was detached for exposure, no extension against resistance was allowed for 6 weeks after surgery. Patients were given a lifting restriction of 5 lbs. Patients were clinically and radiographically evaluated preoperatively and at the postoperative follow-up intervals of 3 months, 6 months, 1 year, then annually thereafter up to 5 years. To the
Clinical study of the Discovery elbow
e97
Figure 1 Discovery elbow is seen in a lateral view, with the ulnar stem on the left and the humeral stem on the right. The inset shows detail of articulation.
Figure 2 Discovery elbow is seen in a posterior view, with the ulnar stem on the left and the humeral stem on the right. The inset shows detail of articulation.
extent practicable, the treating physician was directly involved with the clinical evaluations but was allowed to delegate this responsibility at his or her discretion. Radiographic views included anteroposterior (AP) and lateral. Clinical performance was assessed using the American Shoulder and Elbow Surgeons (ASES) Elbow Assessment Form.15 This scoring system includes a physician assessment–signs table that includes 22 signs with categoric scores. Six of the signs deal with tenderness and 2 with impingement pain (in flexion and in extension). These 8 signs are graded with scores of 0 ¼ none, 1 ¼ mild, 2 ¼ moderate, and 3 ¼ severe, with the remaining 16 signs scored yes or no. For the purposes of this study, the other 5 pain signs (pain on resisted wrist extension, resisted wrist flexion, resisted long finger extension, resisted wrist pronation, and resisted wrist supination) that would have ordinarily been graded as yes or no were, instead, graded on the 0- to 3-points system. Thus, the maximum potential pain/ tenderness score (worst pain/tenderness) was 39 rather than 24. Figures 3, 4, and 5 define the radiographic zones in the lateral view for the ulnar (zones 1-9) and humeral (zones 1-8) components, the lateral view in the vicinity of the humeral flange (zones 1-5), and the AP view of the ulnar (zones 1-7) and humeral (zones 1-7) components, respectively. Radiolucent line widths were measured and recorded by other than the treating surgeon for the cement–prosthesis (CP) and bone–cement (BC) interfaces for the humeral and ulnar components. Regarding the humeral flange region, the percent completeness of the linear contact at the CP interface was measured for zones 1 and 2 and for the bone–prosthesis (BP) interface for zones 4 and 5. Values were categorized as 0% to 25%, 26% to 50%, 51% to 75%, and 76% to 100%, as appropriate. For zone 3, the presence or absence of contact at the CP interface and the BC interface was noted.
Figure 3 Ulnar and humeral radiographic zones are seen in a lateral view.
Figure 4 Radiographic zones about the humeral flange region are seen in a lateral view.
Statistical analysis A paired Student t test was used to compare the preoperative ASES scores with those at the last follow-up using Excel software (Microsoft Corp, Redmond, WA, USA). Significance was taken for P < .05.
e98
H. Hastings II et al. Table I
Patient demographics
Parameter
No. or mean (range)
Patients Bilateral patients Sex Female Male Total elbows Female Male Age, y Surgery side Left Right Dominant side affected? Yes Ambidextrous No Body mass index, kg/m2 Follow-up, y
42 4 34 8 46 37 9 63.4 (33.4-88.7) 21 25 24 1 21 27.7 (19.2-40.3) 4.1 (2.0-5.9)
Figure 5 Ulnar and humeral radiographic zones are seen in an anteroposterior view.
Table II
Primary diagnosis
Incidence (%) (n ¼ 46 elbows)
Results
Rheumatoid arthritis Traumatic arthritis Humeral fracture Osteoarthritis Recurrent dislocation/instability Revision total elbow arthroplasty Ulnar fracture Distal humeral nonunion Radial head/distal humeral fracture
23 6 5 5 2 2 1 1 1
Patient demographics TEAs were performed from January 2002 to August 2011 on 92 patients (99 elbows) by 6 surgeons at 4 surgery centers (2 surgeons at each of 2 sites and 1 surgeon at each of the remaining sites). The mean number of TEAs performed at each center was 25 (range, 10-61). Five patients (6 elbows) died, all with their total elbow prostheses in place at the time of death, and 15 patients (15 elbows) rescinded consent. Forty-three patients (46 elbows) had a minimum of 2 years of follow-up (mean, 4.1 years; median, 4.2 years; range, 2.0-5.9 years), with complete preoperative and postoperative evaluation, and formed the study population. These included 2 of the patients who died (3 elbows) and 3 of the patients who rescinded consent (3 elbows) after a minimum of 2 years of follow-up. The mean number of TEAs performed within this subset per site was 12 (range, 1-32). Patient demographics are summarized in Table I. Approximately 80% were women and about half of the operated elbows were on the dominant side. Table II lists the primary diagnoses. Rheumatoid arthritis was present in 50% of elbows, and post-traumatic arthritis, osteoarthritis, and humeral fracture were the primary diagnoses for 35%. As reported in Table III, the surgical approaches used for 96% of elbows were tricepsoff (reflecting), triceps-off (splitting), or triceps-on (sparing).
Primary diagnoses
(50.0) (13.0) (10.9) (10.9) (4.3) (4.3) (2.2) (2.2) (2.2)
Humeral component stem lengths were 100 to 150 mm and diameters were 3 to 6 mm. Ulnar component stem lengths were 75 or 115 mm and diameters were 3 to 5 mm; however, 1 elbow required a custom 84 mm 2.5 mm ulnar component. All elbows received cemented fixation, and a bone graft was placed under the humeral flange in 3 (6.5%).
Survivorship Among the 46 elbows, 2 with loose locking screws underwent screw exchange at 2.0 and 2.1 years, respectively; the pin and bearing were replaced in 1 elbow due to loosening after multiple falls; and treatment of 1 elbow with a deep infection included condyle and bearing exchange at 6.2 years. Gross survivorship in this series as determined by revision, by component, was humeral/ulnar components, 100% (46 of 46); condyles, 97.8% (45 of 46); bearing, 95.7% (44 of 46); and locking screws, 95.7% (44 of 46) at a
Clinical study of the Discovery elbow Table III
Operative information
Variable
No. (%) or average (range) (n ¼ 46 elbows)
Surgical approach Triceps-on (sparing) 18 (39.1) Triceps-off (reflecting) 19 (41.3) Triceps-off (splitting) 7 (15.2) Other 2 (4.3) Tourniquet time, min 113.6 (57-209) Removing 3 outliers) 112.8 (57-209) Operative time, min 149.7 (90-386) Removing 3 outliers) 138.6 (90.0-234.0) Humeral component Stem diameter 3 mm 1 4 mm 21 5 mm 11 6 mm 13 Stem length 100 mm 41 150 mm 4 Not available 1 Ulnar component Stem diameter 2.5 mm (custom) 1 3 mm 24 4 mm 14 5 mm 7 Stem length 75 mm 40 84 mm (custom) 1 115 mm 5 Bone graft used under humeral flange (No. of elbows) Yes 3 No 43 Cemented fixation (No. of 46 elbows) ) Outliers: (1) concurrent wrist reconstruction procedure, (2) revision elbow with preoperative radial nerve palsy, and (3) fractured lateral condyle.
mean of 4.1 years. Details of these 4 revisions are presented below. As described, the study cohort was limited to elbows with a minimum of 2 years of follow-up. One elbow that was excluded because it had follow-up of <2 years is separately reported here because it was revised. The male patient, treated with TEA at age 70 years for post-traumatic arthritis with humeral fracture, presented with pain and underwent revision of the loose 100-mm 4-mm humeral component, along with the condyles and bushing, at 4.6 years.
Clinical outcomes There was significant improvement (P < .001) in all patient, physician, and range-of-motion components of the
e99 ASES Elbow Assessment Form (Table IV). In particular, mean patient assessments of pain improved from 36.5 to 4.0 points and function from 14.6 to 29.6 points (P < .001). Also, mean patient satisfaction with the procedure was high (9.1 points of a maximum of 10). Mean physician assessments of signs (pain/tenderness), stability, and strength improved from 16.2 to 1.2 points, 1.9 to 0.1 points, and 14.9 to 19.8 points, respectively (P < .001). Finally, the mean flexion-extension arc increased from 81 to 121 and the pronation-supination arc increased from 134 to 163 (P < .001).
Radiographic outcomes Figure 3 and Figure 5 define the lateral view and AP view radiographic zones that were tracked during follow-up. The clinically relevant zones are those responsible for cement fixation; that is, the stem regions of the humeral and ulnar components (humeral lateral view zones 1-8 and AP view zones 2-6, and ulnar lateral view zones 2-8 and AP view zones 2-6). Radiolucencies that developed after 3 months were observed in 7 of 46 elbows (15.2%) and were generally nonprogressive. Only 4 (8.7%) had relevant radiolucencies at last follow-up: (1) humeral AP view CP zone 3, (2) ulnar AP view BC and CP zone 2, (3) humeral lateral view BC zone 1, and (4) ulnar lateral view BC zones 6 and 7 as well as ulnar AP view BC zones 4 and 6. Thus, of the 7 radiolucencies present at last follow-up, 5 were associated with the ulnar component and 2 with the humeral component. One of the 46 humeral components (2.2%) was radiographically loose at 5 years but was not revised (see below). By contrast, complete cement fill does not occur in the humerus distal to the stem or in the ulna proximal to the stem. Radiolucencies in these regions are unassociated with implant loosening but are of interest because they can show how the bone responds to stress shielding here. For instance, Figure 6 shows 4.4-year radiographs of a female rheumatoid patient who underwent TEA at age 72 years, with 4-mm radiolucencies evident in zones 1 and 7 (AP view) of the distal 100-mm 6-mm humeral component. This well functioning elbow was stable, with normal strength and no pain and with a flexion/extension arc of 105 and a pronation/supination arc of 145 . The completeness of the CP (zones 1 and 2) and BP (zones 4 and 5) interfaces by the humeral component anterior flange at last follow-up are listed in Table V. The CP interface was 76% to 100% complete in >95% of elbows, whereas the BP interface was 76% to 100% complete in only about 54% to 61%. The shape of the humerus does not allow for, in most instances, continuous complete contact of the anterior flange to the bone. The CP and BC contact at the radius junction where the anterior flange meets the body of the humeral component (zone 3) was evident in 46% and in 94% of the elbows, respectively.
e100 Table IV
H. Hastings II et al. American Shoulder and Elbow Surgeons Elbow Evaluation Form results
Measurement Patient ASES Pain Function Satisfaction Physician ASES Signs) Stability Strength Range of motion, Flexion Extension Flexion-extension arc Pronation Supination Pronation-supination arc
Preoperative
Postoperative
Average (range)
Average (range)
36.5 (6-50) 14.6 (0-30) N/A
4.0 (0-28) 29.6 (7-36) 9.1 (2-10)
16.2 (1-39) 1.9 (0-6) 14.9 (0-20)
1.2 (0-12) 0.1 (0-3) 19.8 (16-20)
<.001 <.001 <.001
120 39 81 72 63 134
143 22 121 84 80 163
<.001 <.001 <.001 <.001 <.001 <.001
(50-150) (0-80) (20-145) (10-95) (0-95) (15-190)
(105-155) (5 to 90) (45-155) (55-95) (50-90) (125-180)
P
<.001 <.001 N/A
ASES, American Shoulder and Elbow Surgeons. ) All 13 types of pain/tenderness signs were rated on scale of 0 to 3 points for a total worst pain/tenderness score of 39 points.
Figure 6 (Left) Lateral and (Right) anteroposterior radiographs at 4.4 years of a rheumatoid patient with bone resoprtion/remodelling in the intercondylar uncemented portion of the humeral component.
Complications
Device-related complications
At least 1 complication occurred in 20 of the 46 elbows (43.5%). These complications were categorized as devicerelated and non–device-related.
In general, device-related complications included the exchange, revision, or loosening of any of the elbow components, including stems, ulnar bearing, and condylar
Clinical study of the Discovery elbow Table V
Radiographic flange contact in zones 1 to 5
Completeness of contact
0%-25% 26%-50% 51%-75% 76%-100%
e101
Prosthesis-Cement
Prosthesis-Cement
Bone-Cement
Bone-Prosthesis
Zone 1
Zone 2
Zone 3
Zone 3
Zone 4
Zone 5
No. (%)
No. (%)
No. (%)
No. (%)
No. (%)
No. (%)
Yes: 21 (45.7) No: 25 (54.3)
Yes: 43 (93.5) No: 3 (6.5)
12 4 2 28
18 2 1 25
1 (2.2) 1 (2.2) 44 (95.6)
1 (2.2) 45 (97.8)
locking screws. Condylar screw loosening occurred in 3 elbows (6.5%): 1. A loose screw was found in a patient undergoing surgery at 2.0 years after TEA for triceps insufficiency. The screw was exchanged, with reactive bursitis in response to the loose screw noted. 2. The medial locking screw in 1 elbow backed out without loosening of the humeral condyle. Both screws were exchanged at 2.1 years. The bearing had mild wear but was not replaced. 3. The radiograph at 5 years showed that a locking screw was beginning to back out, but no intervention was performed. Bearing or condyle-only exchange, or both, was performed in 2 elbows (4.3%). One patient had received her Discovery elbow during revision of a failed Mayo-modified Coonrad implant. No condyles remained to support the implant. After multiple falls, the locking pin disengaged and became floating, resulting in grinding wear on the loose polyethylene bearing. The pin and the worn bearing were exchanged at 4.5 years. In the second patient, a deep infection related to a hematogenous spread from an infected spinal cord stimulator was debrided at 6.2 years, with exchange of the condyles and bearing. No gross wear pattern on the bearing was noted. One elbow (2.2%) developed loosening of the humeral component. A female patient, who originally underwent TEA due to post-traumatic arthritis at the age of 72 years, had persistent loosening of the 100-mm 6-mm humeral component at 5.0 years; however, revision was not performed, with the patient rating high satisfaction (10 points) and an ASES score that remained improved compared with its preoperative value at this time. At 6.5 years, a periprosthetic humeral fracture was repaired with plate and screw fixation. One patient (2.2%) described occasional ‘‘squeaking’’ of the elbow; nevertheless, giving a 5-year satisfaction rating of 10 points.
Non–device-related complications Complications that were not directly related to the device included 6 (13.0%) wound-healing complications, 3 cases (6.5%) of bursitis, 2 superficial infections (4.3%), 2 (4.3%)
(26.1) (8.7) (4.3) (60.9)
(39.1) (4.3) (2.2) (54.3)
periprosthetic fractures, 1 (2.2%) deep infection (see above), and in 1 elbow (2.2%) each, posteromedial olecranon impingement, elbow contracture, posterior interosseous nerve palsy, ulnar neuropathy, ruptured triceps tendon, intraoperative fracture, and intraoperative perforation.
Discussion TEA, although effective, has not enjoyed the same overall success as its hip and knee counterparts, partly because of the challenging anatomy of this joint.17,29 Although semiconstrained designs can replicate normal elbow kinematics, be used in the presence of bone or ligamentous deficiency, and limit force transfer across the BC interface to mitigate loosening,2,10,27 high complication and revision rates, or both,6,10,11,17,24,25,27 remain. Thus, TEA designs continue to evolve to improve patient outcomes. The Discovery elbow is a semiconstrained device with a design that includes a unique bearing and anatomic modifications to the humeral and ulnar stems. Our purpose was to report on the 4-year (mean) results of a prospective, multicenter study investigating this device. Patient self-evaluation of pain and function and surgeonassessed signs, stability, and strength significantly improved (P < .001). Patient satisfaction was high, with a mean score of 9.1 (10 ¼ very satisfied). The mean flexionextension arc improved by 40 (P < .001). By way of comparison, Van der Lugt and Rozing29 performed a systematic review of the literature, finding mean increase of flexion-extension arc in the range of 12 to 39 for a variety of nonconstrained and semiconstrained TEA prostheses. Implant survivorship is fundamental to the performance of any joint replacement device. Techniques such as the Kaplan-Meier method to estimate survivorship13 take into account the frequency and timing of failures as well loss to follow-up, deaths, and other variables to obtain an estimate of the cumulative survivorship over time. This helps quantify not only the proportion of devices that have failed but also whether these failures tend to occur early or late, knowledge of which can help to determine the cause. The Kaplan-Meier method is best suited for relatively large populations due to the statistical instability that can result when too few patients remain at risk.4 In general, TEA study populations are not large, which may help explain why most cite gross survivorship, or the proportion of
e102 devices at the beginning of the study that have not failed by final follow-up, with ‘‘failure’’ defined as the investigator sees fit to reflect the nature or purpose of the study. Accordingly, care must be taken to ensure that survivorship comparisons among studies are valid. In our study of 46 elbows with complete preoperative evaluation and a minimum of 2 years of postoperative follow-up, gross survivorship based on revision of individual components was 100% for the humeral and ulnar components, 97.8% for the condyles, 95.7% for the bearing, and 95.7% for the locking screws at a mean of 4.1 years. Four of the 46 elbows underwent revision of at least 1 component, for an overall gross survivorship of 91.3%. Table VI lists semiconstrained elbow survivorships for several short-term (2-4 years), medium-term (5-9 years), and long-term (10 years) studies, including reasons for revision. All are gross survivorships, with the exception of Gill and Morrey,6 Shi et al,26 and Mansat et al,19,20 who computed Kaplan-Meier estimates. Short-, mid-, and longterm survivorship studies involving mainly primary Coonrad/Morrey elbows, generally using revision of any component (humeral, ulnar, and bushing components) as the end point, were 70.5% to 96%,2,11,16,21 63.2% to 100%,1,6,18,19,23,26 and 92%,6,20 respectively. Similarly, midterm and long-term survivorships for primary GSB III Elbow System (Zimmer, Warsaw, IN) were 71% to 92.8%12,14,25 and 87.7%,7 respectively. Finally, midterm survivorship for primary Solar elbows was 46.2%.24 Thus, the 4.1-year (mean) survivorship of the Discovery elbow was within the upper end of the short- to midterm range cited by others for semiconstrained devices.2,11,16,17,21 It is important, however, to consider variation in patient demographics and primary diagnoses before drawing firm conclusions because this can influence survivorship.1 Comparing overall complication rates among studies can be difficult because some cite all reported complications, including those of a general surgical nature, whereas others limit focus to major complications or only those that are relevant to the unique aspects of TEA. Table VI summarizes the complications reported in 16 studies for 595 semiconstrained TEA devices (561 primary, 34 revision), including 449 Coonrad/Morrey,1,2,6,11,16,18-21,23,26 130 GSB III,7,12,14,25 13 Solar,24 and 3 Discovery2,11 total elbows. Reported complication rates ranged from 14% to 72%. Overall, there were 44 (8.9%) instances of loosening, 26 (4.4%) of ulnar neuropathy, 11 (1.8%) of disassembly, 25 (4.2%) of bushing wear/failure, and 3 (0.5%) of component fracture. Deep infections required revisions in 17 (2.9%) elbows. Kelly et al,14 summarized the complication rates from 6 semiconstrained TEA series, citing the frequency of aseptic loosening (4%-50%), dislocation (0%-14%), triceps avulsion (2%-11%), ulnar neuropathy (0%-11%), and deep infection (3%-8%). In our series, although at least 1 complication occurred in 20 elbows (43.5%), many were general surgical in nature, such as wound complications (12.8%), superficial
H. Hastings II et al. infections (4.3%), and deep infections (2.2%). Aseptic loosening of the humeral component within the study population occurred in 1 patient who has remained asymptomatic and has not required revision. However, revision of a loose humeral component, along with its associated condyles and bearing, was performed in 1 elbow that did not fit the study population inclusion criteria but is nevertheless reported here due to the severity of this complication. Osteolysis and aseptic loosening are a concern within the first few years after TEA, but this does not necessarily result in significant symptoms because most of these patients are low demand and do not always proceed to severe loosening or dislocation.18 In our series, radiolucencies were minor in all but 1 elbow and only evident in small areas of the reporting grid zones. Other complications involving implant hardware included 3 cases (6.4%) of locking screw loosening, 1 (2.2%) bearing failure associated with excessive use, and 1 deep infection (2.2%) resulting in condyle/bushing exchange. There were no cases of disassembly and only 1 case (2.1%) of ulnar neuropathy that was secondary to surgeries that preceded TEA. It is important to note that 43 of 46 elbows, including the 3 that incurred loose locking screws, were implanted during 2003 to 2007. This was before a 2008 change in the design of the screw-drive mechanism from a Phillips-type to a startype that allows greater torque application. Screw loosening has not been observed with the second-generation screw design. Thus, the complication profile for the Discovery elbow from the current study is well within the range cited by other clinical studies of semiconstrained TEA. The surgical approach used in 57% of the elbows in this series was a triceps-off variant, which we continue to use. The hinge components appear to be robust in the light of the 3 cases of locking screw loosening being associated with the first-generation screw-drive design and the exchanges of condyles and/or bushings being associated with deep infection, a history of falling, or a revised loose humeral component. One of this study’s strengths was that it was prospective and multicenter, with 4 sites and 6 surgeons contributing cases. Another was that it included a study population of 46 elbows, which avoids some of the limitations of markedly smaller populations characteristic of some TEA studies.2,11,12,14,16,18,19,24,25 There were, however, some limitations that warrant mention. First, 32 of 47 of the TEAs reported here were performed by the developing surgeon (H.H.), which could conceivably reduce the study’s general usefulness. (This study received industry funding.) However, of the 20 elbows that experienced complications, 13 were performed by the developing surgeon and 7 were performed by the group of 5 other surgeons, for rates of elbows experiencing complications of 41% (13 of 32) and 50% (7 of 14), respectively. Also, the means of the flexion-extension arc increases for the developing surgeon and the group of other
Summary of survivorship analyses for semiconstrained elbow prostheses
Ref.
Elbow (Qty)
Condition
Survivorship
Complications [reported overall rate] (Qty)
Revision reasons (Qty)
11
C/M) (17); Discoveryy (2)
Various
2
C/M) (17); Discoveryy (1)
RA (14), OA (2), PA (1), GA (1)
74% at mean 40.4 mon 70.5% for C/M only 94.4% at mean 55 mon 94.1% for C/M only
Five C/M due to bushing failure (4) and loosening (1) One C/M due to aseptic loosening
6
C/M) (78)
RA
94.4% at 5 y)) 92.4% at 10 y))
23
C/M) (49)
Distal humeral fx s
90% at mean w7 y
[26%] Bushing failure (4) and loosening (1) [22%] Stiffness (1), transient ulnar nerve palsy (1), wound infection (1), aseptic loosening (1) [14%] Triceps avulsion (3), deep infection (2), component fx (1), post-op ulna fx (2), aseptic loosening (2) [35%] Hematoma (3), wound dehiscence (2), infection (1), injury from fall (3), plus several general surgical complications
21
C/M) (25)
Displaced, intra-articular distal humeral fx s in patients >65 y
96% at 2 y
18
C/M) (31)
RA (27), fracture (3), OA (1)
90.6% survival at 124 mon (mean, 5.5 mon)
1
C/M) (55)
JRA (21 pts) RA (15) PTA (19 pts) in patients <40 y
JRA: 90.5% at mean 7.6 y RA: 80% at mean 7.3 y PTAO: 63.2% at mean 8.5 y
[72%] Ulnar nerve sensory (3) and motor deficit (2), radial nerve palsy (1), wound complication (4), posttraumatic stiffness (3), malalignment (1), deep (1) and superficial infection (1), HO (3). [20%] Ulnar nerve palsy (1), ulnar neurapraxia (1), loose hinge (1), infection (1), loosening (2) [25%] Superficial infection (2), deep infection (2), aseptic loosening (4), triceps insufficiency (3), bushing wear (3)
Reimplantation due to infection (1), component fx (1), aseptic loosening (2)
Clinical study of the Discovery elbow
Table VI
Five due to aseptic loosening (1), septic loosening (1), component fx from fall (2), and 1 revised twice, first due to a fall, then due to aseptic loosening One for deep infection
Two for loosening
JRA: triceps insufficiency (2). RA: loosening (1), bushing wear (1), deep infect (1). PTA: triceps insufficiency (1), bushing wear (2), loosening (3), deep infect (1) (continued on next page)
e103
e104
Table VI (continued ) Ref.
Elbow (Qty)
Condition
Survivorship
Complications [reported overall rate] (Qty)
Revision reasons (Qty)
26
C/M) (67)
RA, PTA, trauma, OA, HA 37 primary 30 revision
Primary: 72% at 5 y)) Revision: 64% at 5 y)) based on 11 failed knees in each group
Primary: 10 revised/resected for periprosthetic fx (2), loosening (5), infect (3) Revision: 5 re-revised for loosening (3), worn bushing (1), periprosthetic fx (1)
16
C/M) (17)
PTA
94.1% at mean of 32 m
19
C/M) (15)
RA (8), PTA (5), PA (1), JA (1) 11 primary 4 revision
100% at 5 y)), 90% at 10 and 13 y))
20
C/M) (78)
IA (45), trauma (33)
97.7% at 5 y)) 91.0% at 10 y)) based on aseptic loosening 92% based on component revision
Primary: [30%] flexion contracture (6), ulnar neuropathy (2), triceps avulsion (1), loosening (7), periprosthetic fx (1) Revision: [37%] flexion contracture (3), ulnar neuropathy (4), intra-op humeral fx (2), triceps avulsion (1), loosening (8), periprosthetic fx (2), worn bushing/dislocation (1) [29%]: Minor: stiffness (1), superficial infection (1) Major: ulnar nerve dysesthesia (1), component loosening (3), including loose humeral component after deep infection, 1 loose humeral component after a fall, and 1 aseptic loosening of ulnar component. No bushing wear observed Overall: [67%] triceps insufficiency (3), ulnar nerve (2) and radial nerve (1) complications, deep infection (2), severe bushing wear (1), bone fxs (2) [34%] ulnar nerve (7), radial nerve (3), triceps weakness (5), humeral looseness (1), ulnar looseness, humeral and ulnar looseness (2), deep infection (3), extra small ulnar component fx (1), bone fx (1), locking pin failure (1). Bushing wear observed in 14
One patient scheduled for revision for aseptic loosening of ulnar component
Deep infection (1), failed hinge (2 revisions same patient)
6 prostheses were revised for deep infection (3), loosening (2), broken pin (1)
H. Hastings II et al.
GSB IIIz (14)
RA (9) PTA (5)
71% at mean 6 y
14
GSB IIIz (28)
RA (22 pts) PA (1 pt)
92.8% at mean 7.6 y
7
GSB IIIz (65)
RA (51 pts) PTA (8 pts)
87.7% at mean w13.5 y, but last f/u interval not provided for 29 elbows
12
GSB IIIz (23)
RA
82.6% at mean w5 y
24
Solarx (13)
PA (6) RA (6) GA (1)
46.2% at mean 8.4 y
(18%) [43%] Aseptic loosening (4), disassembly (2), device fx (1), infection (1), triceps tendon rupture (1), post-op stiffness (2) [39%] Deep infection (1), PE wear/osteolysis (1), ulnar nerve palsy (1), acute pain (1), partial triceps avulsion (3), superficial infect (2), median nerve paresthesia (1), post-op distal humeral fx (1) [32%] Aseptic loosening (3), deep infection (3), disassembly (9), triceps disruption (1), extensor deficit (1), hyperesthesia (2), intra-op humeral fx (1), postop ulnar condyle fx [35%] Aseptic loosening (1), malposition (2), intra-op fx (4) [54%] Motor ulnar nerve palsy (1), loosening (4), bushing failure (1), deep infection (1)
4 elbows revised for aseptic loosening
2 revisions for infect (1) and PE wear with osteolysis (1)
8 revisions for aseptic loosening (1), disassembly (4), deep infection (3)
Clinical study of the Discovery elbow
25
4 revisions for aseptic loosening (1), malposition (2), and disassembly (1) 7 revisions for loosening (4), bushing failure (2), deep infection (1)
fx, fracture; C/M, Coonrad/Morrey; GA, gouty arthritis; HA, hemophiliac arthropathy; HO, heterotopic ossification; IA, inflammatory arthritis; JA, juvenile arthritis; JRA, juvenile rheumatoid arthritis; OA, osteoarthritis; PA, psoriatic arthritis; PE, polyethylene; PTO, post-traumatic arthritis; RA, rheumatoid arthritis. ) Coonrad/Morrey Total Elbow, Zimmer, Warsaw, IN, USA. y Discovery Elbow System, Biomet, Warsaw, IN, USA. z Zimmer, Warsaw, IN, USA. x Solar Elbow System, Stryker, Mahwah, NJ, USA. )) Kaplan-Meier survivorship estimate.
e105
e106 surgeons were 39 and 44 , respectively, whereas the means of the pronation-supination arc increases were 28 and 30 , respectively. Thus, there was no overt evidence of surgeon bias, nor did there appear to be any influence of a learning curve, because all surgeons were very experienced with TEA before beginning the study. Second, there were several primary diagnoses among the patients treated, with nearly half of the elbows presenting with rheumatoid arthritis. However, only a small number of elbows presented with each of the other diagnoses, so the functional dependence of outcomes on diagnosis could not be meaningfully performed. Third, follow-up was relatively short at a mean of 4 years, which was sufficient to characterize functional outcome but was not adequate to investigate late complications or longer-term survivorship. Future work should focus on longer-term follow-up of a larger population of patients receiving the Discovery elbow using locking screws having the second-generation star-drive mechanism.
Conclusion Historically, the hinge has been one of the most problematic aspects of semiconstrained total elbow designs. This multicenter, prospective clinical study was performed to examine the outcomes and complication profile of a semiconstrained total elbow prosthesis with a spherical bearing. Forty-six Discovery elbows with 4.1year mean follow-up (minimum, 2 years) showed significant improvement in the ASES elbow score, high patient satisfaction, and a gross survivorship of 91.3% (42 of 46) that considers revision or exchange of any component for any reason. Revisions in 2 of the 4 elbows were limited to the exchange of first-generation condylar locking screws that had loosened. This has not been observed with the second-generation screw design. No bearings were revised due to wear with the exception of one that had become damaged secondary to interaction with a locking pin that had become free. Within this series, no humeral or ulnar components were revised. However, an additional patient was reported who did not qualify for inclusion in the study cohort but did undergo revision of a loose humeral component along with the bearing and condyles. Within this context, the spherical design of the condyles and bearing of the Discovery elbow appears to be relatively robust; however, longerterm studies will be required to verify this.
Disclaimer Funding for this study was provided by Biomet Inc, Warsaw, IN, USA, for purposes of data collection.
H. Hastings II et al. Hill Hastings II receives a consulting fee/honorarium, travel support, and royalties from Biomet, and his institution receives grant money from Biomet. Donald H. Lee receives speaker fees, consulting fees, and royalties from Biomet and his institution received funding for participation in the Discovery elbow clinical study. William S. Pietrzak is an employee of Biomet.
References 1. Celli A, Morrey BF. Total elbow arthroplasty in patients forty years of age or less. J Bone Joint Surg Am 2009;91:1414-8. http://dx.doi.org/ 10.2106/JBJS.G.00329 2. Corradi M, Frattini M, Panno B, Tocco S, Pogliacomi F. Linked semiconstrained total elbow prosthesis in chronic arthritis: results of 18 cases. Musculoskelet Surg 2010;94(Suppl 1):S11-23. http://dx.doi.org/ 10.1007/s12306-010-0070-y 3. Dee R, Sweetnam DR. Total replacement arthroplasty of the elbow joint for rheumatoid arthritis: two cases. Proc R Soc Med 1970;63:654-5. 4. Dorey F, Amstutz HC. Survivorship analysis in the evaluation of joint replacement. J Arthroplasty 1986;1:63-9. 5. Garrett JC, Ewald FC, Thomas WH, Sledge BS. Loosening associated with G.S.B. hinge total elbow replacement in patients with rheumatoid arthritis. Clin Orthop Rel Res 1977;127:170-4. 6. Gill DR, Morrey BF. The Coonrad-Morrey total elbow arthroplasty in patients who have rheumatoid arthritis. A ten to fifteen-year follow-up study. J Bone Joint Surg Am 1998;80:1327-35. 7. Gschwend N, Scheier NH, Baehler AR. Long-term results of the GSB III elbow arthroplasty. J Bone Joint Surg Br 1999;81:1005-12. 8. Gschwend N, Simmen BR, Matejovsky Z. Late complications in elbow arthroplasty. J Shoulder Elbow Surg 1996;5:86-96. 9. Hastings H 2nd. Linked total elbow arthroplasty. The Discovery prosthesis. In: Yamaguchi K, King GJW, McKee M, O’Driscoll SW, editors. Advanced reconstruction: elbow. Rosemont, IL: American Academy of Orthopaedic Surgeons; 2007. p. 229-40. ISBN No. 9780892033911. 10. Hastings H 2nd. Minimally constrained elbow implant arthroplasty: the Discovery elbow system. Tech Hand Extrem Surg 2004;8:34-50. 11. Hastings H 2nd, Theng CS. Total elbow replacement for distal humerus fractures and traumatic deformity: results and complications of semiconstrained implants and design rationale for the Discovery elbow system. Am J Orthop (Belle Mead NJ) 2003;32(9 Suppl):20-8. 12. Jensen CH, Jacobson S, Ratchke M, Sonne-Holm S. The GSB III elbow prosthesis in rheumatoid arthritis. A 2- to 9-year follow-up. Acta Orthop 2006;77:143-8. http://dx.doi.org/10.1080/17453670610045830 13. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Amer Statist Assn 1958;53:457-81. 14. Kelly EW, Coghlan J, Bell S. Five- to thirteen-year follow-up of the GSB III total elbow arthroplasty. J Shoulder Elbow Surg 2004;13:43440. http://dx.doi.org/10.1016/j.jse.2004.01.028 15. King GJ, Richards RR, Zuckerman JD, Blasier R, Dillman C, Friedman RJ, et al. A standardized method of assessment of elbow function. Research Committee, American Shoulder and Elbow Surgeons. J Shoulder Elbow Surg 1999;8:351-4. 16. Kodde IF, van Riet RP, Eygendaal D. Semiconstrained total elbow arthroplasty for posttraumatic arthritis or deformities of the elbow: a prospective study. J Hand Surg Am 2013;38:1377-82. http://dx.doi. org/10.1016/j.jhsa.2013.03.051 17. Little CP, Graham AJ, Carr AJ. Total elbow arthroplasty. A systematic review of the literature in the English language until the end of 2003. J
Clinical study of the Discovery elbow
18.
19.
20.
21.
22.
Bone Joint Surg Br 2005;87:437-44. http://dx.doi.org/10.1302/0301620X.87B4.15692 Maheshwari R, Vaziri S, Helm RH. Total elbow replacement with the Coonrad-Morrey prosthesis: our medium to long-term results. Ann R Coll Surg Engl 2012;94:189-92. http://dx.doi.org/10.1308/ 003588412X13171221589775 Mansat P, Bonnevialle N, Rongieres M, Mansat M, Bonnevialle P. Results with a minimum of 10 years follow-up of the Coonrad/Morrey total elbow arthroplasty. Orthop Traumatol Surg Res 2013;99:S33743. http://dx.doi.org/10.1016/j.otsr.2013.07.002 Mansat P, Bonnevialle N, Rongieres M, Mansat M, Bonnevialle P. Experience with the Coonrad-Morrey total elbow arthroplasty: 78 consecutive total elbow arthroplasties reviewed with an average 5 years of follow-up. J Shoulder Elbow Surg 2013;22:1461-8. http://dx. doi.org/10.1016/j.jse.2013.07.042 McKee MD, Veillette CJH, Hall JA, Schemitsch EH, Wils LM, McCormack R, et al. A multicenter, prospective, randomized controlled trial of open reduction-internal fixation versus total elbow arthroplasty for displaced intra-articular distal humeral fractures in elderly patients. J Shoulder Elbow Surg 2009;18:3-12. http://dx.doi. org/10.1016/j.jse.2008.06.005 Morrey BF. Complications of elbow replacement surgery. In: Morrey BF, editor. The elbow and its disorders. 3rd ed. Philadelphia: WB Saunders; 2000. p. 667-76. ISBN-10 No. 07216775725 ISBN-13 No. 978–0721677521.
e107 23. Muller LP, Kamineni S, Rommens PM, Morrey BF. Primary total elbow replacement for fractures of the distal humerus. Oper Orthop Traumatol 2005;17:119-42. http://dx.doi.org/10.1007/s00064-0051125-3 24. Patil N, Cheung EV, Mow CS. High revision rate after total elbow arthroplasty with a linked semiconstrained device. Orthopedics 2009; 32:321. 25. Schneeberger AG, Hertel R, Gerber C. Total elbow replacement with the GSB III prosthesis. J Shoulder Elbow Surg 2000;9:135-9. 26. Shi LL, Zurakowski D, Jones DG, Koris MJ, Thornhill TS. Semiconstrained primary and revision total elbow arthroplasty with use of the Coonrad-Morrey prosthesis. J Bone Joint Surg Am 2007;89:146775. http://dx.doi.org/10.2106/JBJS.F.00715 27. Szekeres M, King GJ. Total elbow arthroplasty. J Hand Ther 2006;19: 245-54. http://dx.doi.org/10.1197/j.jht.2006.02.010 28. Trail IA, Nuttall D, Stanley JK. Survivorship and radiological analysis of the standard Souter-Strathclyde total elbow arthroplasty. J Bone Joint Surg Br 1999;81:80-4. 29. Van der Lugt JCT, Rozing PM. Systematic review of primary total elbow prostheses used for the rheumatoid elbow. Clin Rheumatol 2004;23:291-8. http://dx.doi.org/10.1007/s10067-004-0884-9 30. Vishwanathan K, Alizadehkaiyat O, Kemp GJ, Forstick SP. Responsiveness of the Liverpool Elbow Score in elbow arthroplasty. J Shoulder Elbow Surg 2013;22:312-7. http://dx.doi.org/10.1016/j.jse. 2012.09.003