Abstract
Background:
Total elbow arthroplasty for the treatment of posttraumatic arthritis is associated with a relatively high failure rate. An understanding of these failures can lead to improved implant design and surgical technique.
Methods:
Eighty-four consecutive patients underwent eighty-five semiconstrained total elbow arthroplasties for the treatment of posttraumatic arthritis. Sixty-nine elbows with a retained primary prosthesis were followed for an average of nine years. Clinical results were graded with use of the Mayo Elbow Performance Score. Radiographs were assessed for mechanical failure, and all complications were recorded.
Results:
Sixteen primary arthroplasties (19%) failed. Causes of failure included isolated bushing wear (seven), infection (four), component fracture (three), and component loosening (two). The most common cause of early failure (failure after less than five years) was infection, whereas intermediate-term failure (failure after five to ten years) typically was due to bushing wear. Late failure (failure after more than ten years) was uncommon and involved component loosening or fracture. Seventy-five percent of the failures were in patients who were less than sixty years old at the time of surgery (p = 0.03). Progressive radiolucent lines were noted around four implants, three of which had clinically important loosening. Total elbow arthroplasty was associated with significant improvements in terms of pain, motion, and the Mayo Elbow Performance Scores (p = 0.002). Sixty-eight percent of the patients achieved a good or excellent clinical result, and 74% were subjectively satisfied. Kaplan-Meier analysis demonstrated a fifteen-year survival rate of 70% with revision or resection for any reason as the end point.
Conclusions:
Semiconstrained total elbow arthroplasty in patients with posttraumatic arthritis places high demands on the implant and is associated with a relatively high failure rate. Seventy-five percent of failures occur in patients less than sixty years of age, and infection continues to represent a frequent mode of early failure. Bushing wear and component loosening or fracture are seen more commonly in the intermediate and late term, whereas aseptic loosening remains relatively uncommon.
Level of Evidence:
Therapeutic Level IV. See Instructions to Authors for a complete description of levels of evidence.
Total elbow arthroplasty is a treatment option for a number of degenerative and posttraumatic conditions of the elbow. The reported indications for total elbow arthroplasty in patients with a history of trauma include distal humeral fractures, posttraumatic instability, and posttraumatic ankylosis1-4. Some studies have documented the results of total elbow arthroplasty for the treatment of posttraumatic arthritis, most with relatively short follow-up5-7. In a previous short-term study of our experience with the use of semiconstrained arthroplasty for the treatment of posttraumatic arthritis of the elbow, we reported overall satisfactory results despite a relatively high complication rate8.
Most complications after total elbow arthroplasty are well documented in the literature and are most commonly noted in patients who have had previous surgery, which is common in those with posttraumatic conditions. Infection, periprosthetic fracture, polyethylene wear, aseptic loosening, and component malposition all have been reported following total elbow replacements performed for a variety of diagnoses8-12. However, while the mechanisms of failure of linked total elbow arthroplasty have been previously studied, little is known of their relative frequency or severity in the setting of posttraumatic reconstruction, particularly with longer follow-up13,14. In the present study, we update our experience with linked semiconstrained total elbow arthroplasty for the treatment of posttraumatic arthritis to analyze the modes of failure. We hypothesized that failure modes change over time. In particular, we theorized that infection is the primary failure mode in the first five years whereas bushing wear is the most common mode of failure in the intermediate period (five to ten years postoperatively). Finally, we hypothesized that implants were most likely to fail because of continued bushing wear as well as component fracture or loosening in the late postoperative period (more than ten years postoperatively).
Following institutional review board approval, all patients undergoing semiconstrained total elbow arthroplasty because of trauma or a posttraumatic disorder of the elbow were identified with use of the joint registry database at our institution. Initially, 179 patients were identified and their medical records were reviewed. Exclusion criteria included a primary diagnosis of acute elbow trauma (defined as occurring less than three months previously), nonunion, flail elbow, nontraumatic osteoarthritis, posttraumatic instability without degenerative change, or rheumatoid arthritis as the primary cause of joint degeneration. Patients with a previous or active infection at the time of evaluation were not necessarily excluded. Those patients underwent irrigation and debridement procedures followed by staged total elbow arthroplasty on the basis of the judgment of the operating surgeon. Eighty-five elbows from eighty-four patients comprised the main study cohort; implant survival was evaluated for the entire cohort of eighty-five elbows. Clinical outcomes were assessed for a subset of sixty-nine elbows (81%) (sixty-nine patients) with at least two years of clinical follow-up and a retained primary prosthesis. The remaining sixteen patients either had insufficient follow-up (ten) or did not have a primary implant in place at the time of follow-up (six). Forty-four (64%) of the sixty-nine patients had clinical follow-up by means of correspondence with use of a mailed questionnaire. Ten (14%) had a final follow-up evaluation conducted at our institution. The follow-up for the remaining fifteen patients (22%) was done by telephone. A second subset of sixty-four elbows (75%) with at least two years of radiographic follow-up underwent radiographic evaluation. All radiographs either were made at our institution or were made by the patient's local physician and were mailed to us for evaluation. Patients without two-year radiographic follow-up were not included in this part of the analysis.
Clinical Data
Preoperative data, including demographic characteristics, associated injuries, previous operations, previous complications, pain, and functional status, were assessed for all patients. Information regarding the initial traumatic event also was recorded. Pain was recorded as none (1 point), mild (2 points), moderate (3 points), or severe (4 points). The Mayo Elbow Performance Score was used to assess functional status15. In this system, outcomes are scored on the basis of four parameters: pain (0 to 45 points), range of motion (5 to 20 points), stability (0 to 10 points), and function (0 to 25 points). Data regarding elbow stability were compiled with use of a combination of physical examination findings, when available, and patient self-reported assessment.
The surgical procedures were performed as previously described8. The Bryan-Morrey approach was typically employed with routine transposition of the ulnar nerve. Starting in 1983, the linked semiconstrained Coonrad/Morrey device (Zimmer, Warsaw, Indiana) was inserted with use of antibiotic-impregnated cement (1 g vancomycin/40 g cement) and with the placement of bone graft behind the anterior humeral flange. No custom-made devices were used, and the collateral ligaments were not reconstructed. Over the course of the study, no substantial changes were made in the design or materials of the humeral component. However, the ulnar component underwent several modifications. Because of concerns regarding ulnar stem fracture associated with the original bead-coated design, a polymethylmethacrylate precoating was used, starting in 1993. However, problems related to the precoating led to its abandonment, and it was not in use after 2000. The current ulnar component has a titanium plasma-sprayed design. One change was made in the axis pin between the ulnar and humeral components in 1993, with the material used in the pin being changed from titanium alloy to cobalt-chromium. Finally, the axis pin mechanism was changed from a C-ring to a snap-fit articulation.
Patients with a history of infection were carefully evaluated and staged according to the presence of ongoing involvement. Preoperative workup included an analysis of blood inflammatory markers and joint aspiration with preoperative culture. Intraoperative pathologic examination was also done. Any evidence of ongoing infection was treated with staged reconstruction with initial irrigation and debridement and postoperative antibiotic therapy in an effort to clear the infection, with total elbow arthroplasty being performed later if the treatment was successful. Patients in whom the infection could not be cleared were not offered total joint replacement.
Postoperatively, data were collected on pain, functional status, the number of subsequent operations, and complications. Follow-up data were collected with use of a combination of return visits to our institution, visits to the patient's local orthopaedic surgeon, and mailed questionnaires. The Mayo Elbow Performance Score was again used for the postoperative evaluation of function. Clinical results were defined as excellent (90 to 100), good (70 to 89), fair (50 to 69), and poor (<50). Patient satisfaction was reported as much better, somewhat better, the same, or worse as compared with the preoperative state.
All implant-related complications and reoperations were analyzed in detail. Mechanical failure was defined as an implant-related complication that required additional operative intervention for component revision or resection. Failures of the primary implant and their etiologies were then analyzed relative to their occurrence in the early term (less than five years), intermediate term (five to ten years), or late term (more than ten years) following total elbow replacement.
Radiographic Data
Radiographic analysis was performed by two elbow surgeons (T.T., J.S.-S.) who did not participate in the surgical procedures and were not otherwise involved in the patients’ care. Preoperative radiographs were graded according to consensus with regard to the severity of degenerative change as none (Grade 1), mild (Grade 2), moderate (Grade 3), or severe (Grade 4)16. Attention was also paid to any elbow malalignment. Postoperative analysis was done on the most recent available radiographs for all patients with at least two years of radiographic follow-up. Any evidence of component fracture, loosening, osteolysis, bushing wear, and periprosthetic fracture was recorded (Figs. 1-A and 1-B). Whenever radiolucent lines were identified, all available previous radiographs were evaluated to determine progression of the radiolucency. Type-0 radiographic change denoted either no radiolucent lines or lines measuring <1 mm in thickness and involving <50% of the interface. Type-I lines also involved <50% of the interface but were at least 1 mm thick. Type-II lines were >1 mm thick and involved >50% of the interface. Type-III lines were at least 2 mm thick and involved the entire interface. Type-IV change indicated gross loosening17 (Fig. 2). Bushing wear was defined as deterioration of the polyethylene bearing surface in the ulnohumeral articulation, resulting in an asymmetric tilt of the implant at the bearing surface, typically with a valgus posture, on an anteroposterior radiograph.
Statistical Methods
Descriptive statistics are reported as the mean (and range) for continuous measures and as the number (and percentage) for discrete assessments. Implant survival was estimated for all eighty-five elbows with the Kaplan-Meier method and was reported as the rate estimates and 95% confidence intervals. Implant survival was analyzed by examining four separate outcomes: (1) revision for any reason or implant removal, (2) revision because of mechanical failure, (3) revision because of clinical or radiographic loosening, and (4) revision because of mechanical failure or radiographic loosening. For each outcome, implant revision was the event of interest, and the time to revision was defined as the time from the date of the primary total elbow arthroplasty to the date of the revision or most recent follow-up; those with no record of revision were censored at the date of the latest follow-up. For clinical outcomes, paired t tests were used to compare preoperative and postoperative changes in extension, flexion, pronation, supination, and the Mayo Elbow Performance Score. A nonparametric signed-rank test was used to compare preoperative and postoperative changes in pain. Postoperative assessments were made at the time of the latest clinical contact. Sixty-nine patients with at least two years of clinical follow-up and a retained primary implant were included in the clinical outcomes assessment. The Mayo Elbow Performance Score was compared on the basis of sex (male or female), elbow deformity (no or yes), previous complications (no or yes), age group (less than sixty years of age or sixty years of age or more), and the number of previous procedures (fewer than four or four or more) with use of two-sample t tests assuming unequal variances. The alpha level was set at 0.05 for significance.
Source of Funding
No external funds were used for this research project or the preparation and submission of this manuscript. The senior author (B.M.) is a designer of the implant studied and receives royalties related to it.
Clinical Data
Eighty-five elbows in eighty-four patients met the inclusion criteria. The average age of the patients at the time of surgery was sixty-four years (range, thirty-two to eighty-six years). There were sixty-three women and twenty-one men. There were forty-eight right and thirty-seven left elbows. Forty-five arthroplasties involved the dominant elbow, thirty-one involved the nondominant elbow; this information was not known for nine patients.
The most common posttraumatic condition was arthritis following distal humeral fracture, which was seen in twenty-nine elbows (34%). Other injuries included elbow fracture/dislocation, olecranon fracture, radial head fracture, and isolated elbow dislocation. In fourteen elbows, the original injury was an uncharacterized periarticular elbow fracture (Table I). Five elbows had an open fracture; however, none of these elbows developed an infection before being seen at our institution or after total elbow arthroplasty. Associated neurovascular injuries were uncommon at the time of the index injury. There were two cases of traumatic laceration of the brachial artery, one case of transection of the ulnar nerve, and one case of transection of the median nerve.
Seventy-six elbows (89%) had undergone a mean of three previous operations (range, zero to twenty operations) prior to total elbow arthroplasty. Thirty-one previous complications (range, zero to four complications) had occurred in twenty-five (30%) of the eighty-four patients prior to total elbow arthroplasty. Four patients had postoperative infections prior to evaluation at our institution. One of these four patients developed an infection following total elbow arthroplasty but subsequently was lost to follow-up.
Fifty-four elbows (64%) were reported to have severe pain at the time of presentation. Preoperative extension averaged 41° (range, 0° to 105°), whereas preoperative flexion averaged 115° (range, 70° to 145°). The mean preoperative Mayo Elbow Performance Score for the eighty-five elbows was 39 points (range, 5 to 75 points). The preoperative clinical data are summarized in Table II.
Of the eighty-five elbows that were included in the study, sixty-nine (81%) had adequate two-year clinical follow-up (average, 109 months; range, twenty-four to 246 months) and a retained primary prosthesis and underwent an analysis of clinical function. There was significant improvement in terms of pain (p < 0.001), with forty-seven patients (68%) reporting either no or mild pain (Table II). Twenty patients reported moderate pain, leaving two elbows (3%) with severe pain at the time of the latest follow-up. On the average, extension improved 11° (from 41° to 30°) whereas flexion improved 24° (from 115° to 139°). These increases in motion were both significant (p = 0.002 and p < 0.001, respectively). Average pronation improved from 58° to 75° (p < 0.001), and average supination increased from 59° to 75° (p = 0.08). Stability was not significantly changed, with forty-five patients reporting elbow stability and twenty-one patients reporting moderate instability. Three patients considered the elbow to be unstable.
The average Mayo Elbow Performance Score for these sixty-nine elbows improved significantly to 75 points (p < 0.001). Thirteen elbows had an excellent result, thirty-four had a good result, eighteen had a fair result, and four had a poor result. Thus, a total of forty-seven elbows (68%) had a good or excellent result and twenty-two (32%) had a fair or poor result. Poor results were often due to a combination of continued pain and limited motion with resultant functional deficits.
Subjective improvement was also assessed. Forty-one of the sixty-nine patients stated that the elbow was much better, and ten reported that the elbow was somewhat better. Eight elbows were unchanged, and three were worse. For seven patients, this information was not reported. Therefore, 74% of respondents believed that the elbow was improved at the time of the latest follow-up.
Complications were evaluated for all eighty-five elbows in the original cohort. In this group, there were a total of sixty-eight complications (range, zero to fifteen complications) in twenty-nine elbows (34%). Complications were divided into those that were related to the implant and those that were not. Thirty-nine complications (57%) did not relate directly to the total elbow arthroplasty components. These included nine extensor mechanism failures, seven deep infections, four wound-healing problems, four periprosthetic fractures, four episodes of arthrofibrosis, three cases of ulnar neuropathy, and one ulnar nerve rupture. The remaining twenty-nine complications (43%) were implant-related complications and included seventeen cases of bushing wear or exchange of bushings during a revision procedure, seven cases of loosening (four ulnar and three humeral), and five component fractures (three ulnar and two humeral). Of note, many patients had combinations of these complications (e.g., humeral and ulnar loosening or multiple bushing exchanges over the course of several revisions) and others had sequential complications (e.g., bushing exchange followed by humeral revision at a later date). These data are summarized in Table III.
Twenty-four elbows (28%) underwent a total of 122 additional operations (range, zero to thirty-eight operations). Procedures were again divided into implant-related and non-implant-related operations. Of the 122 additional operations, seventy-five (61%) were not related to the total elbow components. These operations were predominantly debridement procedures for the treatment of deep infection. The forty-seven implant-related operations included twenty bushing exchanges for various diagnoses including infection and as part of a revision procedure, six implant resections, five humeral component revisions, and five ulnar component revisions. In addition, there were four cases of bone-grafting and three cases of total elbow arthroplasty reimplantation (Table III).
Radiographic Data
Preoperatively, the mean grade for radiographic arthritis was 3.6 on a scale of 4, and moderate to severe degenerative changes were present in all elbows. Furthermore, nineteen (22%) of the eighty-five elbows had incongruity in the form of either subluxation or chronic dislocation. Two additional elbows were in moderate axial malalignment at the time of the initial presentation. Of the eighty-five elbows that were included in the study, sixty-four (75%) had at least two years of radiographic follow-up (average, ninety-six months; range, twenty-four to 243 months) after surgery. At the time of the latest follow-up, fifty-two (81%) of these sixty-four elbows had no radiographic evidence of loosening, seven elbows had Type-I changes, one elbow each had Type-II and Type-III changes, and three elbows had gross loosening (Fig. 2). Of the elbows with radiolucent lines, four had evidence of progressive radiolucency when compared with the immediate postoperative radiographs. Interestingly, twelve patients had evidence of bushing wear on the most recent radiographs but had no evidence of progressive radiolucent lines (Figs. 3-A and 3-B).
Survivorship Analysis
Thirty-four elbows survived for at least ten years and, of those, twenty remained in place at least fifteen years after surgery. Kaplan-Meier survivorship analysis was conducted with use of several end points for all eighty-five elbows that constituted the original study group. The survival rate with revision for any reason or implant resection as the end point was 92% (95% confidence interval, 87% to 98%) at five years, 78% (95% confidence interval, 67% to 90%) at ten years, and 70% (95% confidence interval, 57% to 86%) at fifteen years (Fig. 4). The survival rate with mechanical failure of the implant as the end point was 97% (95% confidence interval, 94% to 100%) at five years, 82% (95% confidence interval, 72% to 94%) at ten years, and 74% (95% confidence interval, 60% to 90%) at fifteen years. With revision because of mechanical failure or radiographic loosening as the end point, the survival rate was 96% (95% confidence interval, 91% to 100%) at five years, compared with 79% (95% confidence interval, 68% to 91%) at ten years and 71% (95% confidence interval, 58% to 87%) at fifteen years. Finally, with revision for clinical or radiographic loosening as the end point, the survival rate was 97% (95% confidence interval, 94% to 100%) at five years, 95% (95% confidence interval, 90% to 100%) at ten years, and 90% (95% confidence interval, 80% to 100%) at fifteen years.
Analysis of Implant Revisions
For this analysis, all eighty-five elbows were included. In the study group, a total of sixteen failures of the primary implant (19%) were documented. Implant-related complications were defined as those requiring additional operative intervention for component revision or resection. Some patients had additional mechanical failures following revision surgery, and those complications were noted previously. However, the purpose of this analysis was to determine the mode of failure of the original implants, rendering those additional complications beyond the scope of this aspect of the investigation.
Several demographic variables were analyzed relative to their impact on implant failure. We could not identify a significant association between total elbow arthroplasty failure and sex, the amount of preoperative elbow deformity, the number of previous operations, or the presence of previous complications (p > 0.11). However, there was a significantly higher rate of failure in patients younger than sixty years of age at the time of total elbow arthroplasty (p = 0.03), with 75% of the failures occurring in the younger cohort.
Four implants were resected because of infection, and seven implants failed as a result of bushing wear without loosening. Two humeral stems and one ulnar stem were revised because of component fracture. Two ulnar components and one humeral component were revised because of aseptic loosening. One patient had simultaneous humeral and ulnar loosening, resulting in a total of sixteen implant failures that led to additional surgery.
Finally, the modes of implant failure were assessed according to the time following implantation. Seven of the sixteen implant failures that resulted in additional surgery occurred in the first five years after total elbow arthroplasty. These complications included four infections and one instance each of ulnar loosening, ulnar stem fracture, and bushing wear. Another seven complications that resulted in revision occurred between five and ten years after total elbow arthroplasty. Six failures were related to bushing wear, and one failure was due to a humeral stem fracture. Finally, two failures occurred more than ten years postoperatively; one was due to combined humeral and ulnar loosening, and one was due to a humeral stem fracture (Table IV).
In 1997, we reported an 83% rate of good and excellent results in a study of forty-one patients who were managed with total elbow arthroplasty for the treatment of posttraumatic arthritis and were followed for an average of approximately six years8. The rate of subjective satisfaction was 95% for patients with a functioning implant at the time of the latest follow-up. At that point, the complication rate was 27% and 22% of the patients had required an additional operation. Modes of failure were primarily ulnar component fracture and bushing wear. These failures were often attributed to heavy use of the elbow postoperatively. Importantly, none of these failures occurred in patients with radiographic evidence of loosening.
The current report is an update of our experience with total elbow arthroplasty for the treatment of posttraumatic arthritis and was performed in an effort to better understand the causes of long-term failure and failure rates. In the present series, the clinical follow-up of the primary implant was extended to more than nine years (range, two to twenty years). We found this to be a challenging population, with nearly 90% of the patients having undergone multiple (average, three) previous operations and almost one-third having had a previous complication. Although substantial improvements in patient function and Mayo Elbow Performance Scores persisted, the rate of good and excellent results in the present study was only 68% after a longer duration of follow-up, compared with 83% in the initial report after an average of sixty-eight months of follow-up8. Interestingly, twenty-four patients reported moderate or severe instability after the implantation of this semiconstrained device. We attribute this finding to differing patient perceptions of the 7° of play built into the device. We found the complication and reoperation rates to be similar to those in the original series. Thirty-four percent of the elbows experienced a complication, and 28% underwent additional surgical treatment following the initial total elbow arthroplasty. While complications were slightly more frequent than they were in the original series, we did not see a large increase in the rate of complications with longer follow-up. Along with our finding of only two late primary implant failures among thirty-four patients with at least ten years of follow-up, these data suggest that late complications are uncommon in this population.
However, despite a relatively stable late-term rate of mechanical failure, we found total elbow arthroplasty to be a high-risk operation in this population. Implant-related complications such as bushing wear were seen in twenty-nine cases (34%), and other complications, including triceps insufficiency, deep infection, periprosthetic fracture, arthrofibrosis, and ulnar neuropathy, were seen in thirty-nine cases (46%), sometimes concurrently. In addition, a total of 122 additional operative interventions were required for the treatment of these complications. On the basis of these findings, we recommend that surgeons undertaking total elbow arthroplasty for the treatment of posttraumatic arthritis be comfortable with the techniques and pitfalls regarding revision elbow arthroplasty.
Other studies have quantified the long-term survival of semiconstrained total elbow replacements. Aldridge et al. evaluated the Coonrad-Morrey prosthesis in a study of forty-one elbows that had various diagnoses and that were followed for ten to thirty-one years18. They reported fourteen complications and thirteen revisions but no cases of acute infection or permanent removal of any implant. Kraay et al. reported a rate of implant survival of 73% at three years and of only 53% at five years for patients with posttraumatic conditions including arthritis, fractures, and nonunion19. In contrast, the present study documented an implant survival rate of approximately 70% at fifteen years in patients with posttraumatic arthritis. Yet, there was deterioration of the survival rate with time. Furthermore, these results indicate that total elbow arthroplasty is not as durable for the treatment of posttraumatic arthritis as it is for the treatment of the rheumatoid elbow20.
The main goal of the present study was to delineate the primary modes of implant failure after long-term follow-up in patients with posttraumatic arthritis. In a recent study of implant failure, Goldberg et al. studied the mechanisms of wear after semiconstrained total elbow arthroplasty13. In a group of sixteen elbows with a variety of diagnoses, wear mechanisms and subsequent histopathological findings were found to be similar to those noted in association with osteolysis following total hip and knee arthroplasty. Of note, almost all (fourteen) of the sixteen ulnar component failures involved the polymethylmethacrylate-precoated design, which is no longer in use. Additionally, it has been reported that polymethylmethacrylate-precoated ulnar components can loosen by means of a pistoning mechanism within the cement mantle when the depth of implant insertion is such to cause anterior impingement14. We found ulnar component loosening to be uncommon overall, but we did note that one of the periprosthetic fractures in the present series was due to osteolysis around a precoated ulnar component. Additionally, Lee et al. studied the rate of bushing exchange because of polyethylene wear following semiconstrained total elbow arthroplasty and documented a 1% incidence after >900 procedures11. Posttraumatic arthritis and prosthetic angular deformity were documented risk factors for requiring this procedure. In addition, Wright and Hastings, reporting on bushing wear as a failure mechanism in a series of ten patients, also found it to be associated with posttraumatic arthritis21. They suggested that this mode of failure may be more common than previously thought. Finally, two additional reports also documented increased wear secondary to articular malalignment in unlinked devices as well12,22.
Observations from the survival analysis reveal differences in failure mechanisms over time. In the early postoperative period (less than five years postoperatively), we found infection to be the most common cause of failure. In the intermediate term (five to ten years postoperatively), bushing wear, usually with associated metal debris, was responsible for six of seven failures. In patients with more than ten years of follow-up, implant loosening and fracture accounted for two additional failures. However, implant loosening as a cause of failure was relatively uncommon, accounting for only two (13%) of the sixteen failures.
A significantly increased risk of implant failure was noted in patients younger than sixty years of age at the time of total elbow arthroplasty. We attribute this finding to the higher loads placed on these elbows over time in younger patients. While offering younger patients a total elbow replacement demands caution in the decision-making process, we do not dismiss prosthetic reconstruction purely for this reason. We recognize that, in the setting of posttraumatic arthritis, total elbow arthroplasty may be the only tenable solution for the patient. However, we proceed only after a thorough discussion of risks with the patient and careful evaluation of whether total elbow arthroplasty is the only appropriate alternative.
The present study had several limitations. It was a retrospective study, with all of the disadvantages inherent in this design. In addition, although we attempted to standardize follow-up data with use of the Mayo Elbow Performance Score, this information was procured from several sources, including mailed questionnaires, outside physician assessments, and personal evaluations at our institution. This creates the potential for bias in the follow-up data. Furthermore, only 14% of the patients had the latest follow-up evaluation at our institution. The inability to attain current radiographs for the entire cohort was also disappointing. Also, the inclusion of failures due to infection makes this report more comprehensive but makes the ability to precisely compare it with other studies that exclude infection more difficult. Finally, we noted discrepancies between the present study and our earlier study8. For example, in the previous series, five ulnar component fractures were noted, but we found only one additional ulnar component fracture with longer follow-up. This is because the previous and current reports employed different inclusion criteria. Thus, these two papers reflect similar, but not identical, patient populations.
In summary, the present study defined the failure mechanisms and long-term results of a linked semiconstrained total elbow replacement for the treatment of posttraumatic arthritis. Although these patients presented a particularly difficult therapeutic challenge, most were subjectively satisfied with their results and had a satisfactory functional outcome. However, the long-term failure rate in this series is high (19%), and the survivorship analysis demonstrated approximately 70% retention at fifteen years of follow-up, which is less than that reported following treatment of the rheumatoid elbow.
Mansat
P;
Morrey
BF. Semiconstrained total elbow arthroplasty for ankylosed and stiff elbows. J Bone Joint Surg Am.
2000;82:1260-8.[PubMed]
Ramsey
ML;
Adams
RA;
Morrey
BF. Instability of the elbow treated with semiconstrained total elbow arthroplasty. J Bone Joint Surg Am.
1999;81:38-47.[PubMed]
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.[PubMed]
Garcia
JA;
Mykula
R;
Stanley
D. Complex fractures of the distal humerus in the elderly. The role of total elbow replacement as primary treatment. J Bone Joint Surg Br.
2002;84:812-6.[PubMed][CrossRef]
Moro
JK;
King
GJ. Total elbow arthroplasty in the treatment of posttraumatic conditions of the elbow. Clin Orthop Relat Res.
2000;370:102-14.[PubMed][CrossRef]
Schneeberger
AG;
Meyer
DC;
Yian
EH. Coonrad-Morrey total elbow replacement for primary and revision surgery: a 2- to 7.5-year follow-up study. J Shoulder Elbow Surg.
2007;16(
3 Suppl):S47-54.[PubMed][CrossRef]
Little
CP;
Graham
AJ;
Karatzas
G;
Woods
DA;
Carr
AJ. Outcomes of total elbow arthroplasty for rheumatoid arthritis: comparative study of three implants. J Bone Joint Surg Am.
2005;87:2439-48.[PubMed][CrossRef]
Schneeberger
AG;
Adams
R;
Morrey
BF. Semiconstrained total elbow replacement for the treatment of post-traumatic osteoarthrosis. J Bone Joint Surg Am.
1997;79:1211-22.[PubMed]
Marra
G;
Morrey
BF;
Gallay
SH;
McKee
MD;
O'Driscoll
SW. Fracture and nonunion of the olecranon in total elbow arthroplasty. J Shoulder Elbow Surg.
2006;15:486-94.[PubMed][CrossRef]
Athwal
GS;
Morrey
BF. Revision total elbow arthroplasty for prosthetic fractures. J Bone Joint Surg Am.
2006;88:2017-26.[PubMed][CrossRef]
Lee
BP;
Adams
RA;
Morrey
BF. Polyethylene wear after total elbow arthroplasty. J Bone Joint Surg Am.
2005;87:1080-7.[PubMed][CrossRef]
Thillemann
TM;
Olsen
BS;
Johannsen
HV;
Søjbjerg
JO. Long-term results with the Kudo type 3 total elbow arthroplasty. J Shoulder Elbow Surg.
2006;15:495-9.[PubMed][CrossRef]
Goldberg
SH;
Urban
RM;
Jacobs
JJ;
King
GJ;
O'Driscoll
SW;
Cohen
MS. Modes of wear after semiconstrained total elbow arthroplasty. J Bone Joint Surg Am.
2008;90:609-19.[PubMed][CrossRef]
Cheung
EV;
O'Driscoll
SW. Total elbow prosthesis loosening caused by ulnar component pistoning. J Bone Joint Surg Am.
2007;89:1269-74.[PubMed][CrossRef]
Broberg
MA;
Morrey
BF. Results of treatment of fracture-dislocations of the elbow. Clin Orthop Relat Res.
1987;216:109-19.[PubMed]
Huang
TL;
Chiu
FY;
Chuang
TY;
Chen
TH. The results of open reduction and internal fixation in elderly patients with severe fractures of the distal humerus: a critical analysis of the results. J Trauma.
2005;58:62-9.[PubMed][CrossRef]
Morrey
BF;
Bryan
RS;
Dobyns
JH;
Linscheid
RL. Total elbow arthroplasty. A five-year experience at the Mayo Clinic. J Bone Joint Surg Am.
1981;63:1050-63.[PubMed]
Aldridge
JM
3rd;
Lightdale
NR;
Mallon
WJ;
Coonrad
RW. Total elbow arthroplasty with the Coonrad/Coonrad-Morrey prosthesis: a 10- to 31-year survival analysis. J Bone Joint Surg Br.
2006;88:509-14.[PubMed][CrossRef]
Kraay
MJ;
Figgie
MP;
Inglis
AE;
Wolfe
SW;
Ranawat
CS. Primary semiconstrained total elbow arthroplasty: survival analysis of 113 consecutive cases. J Bone Joint Surg Br.
1994;76:636-40.[PubMed]
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.[PubMed][CrossRef]
Wright
TW;
Hastings
H. Total elbow arthroplasty failure due to overuse, C-ring failure, and/or bushing wear. J Shoulder Elbow Surg.
2005;14:65-72.[PubMed][CrossRef]
van der Lugt
JC;
Geskus
RB;
Rozing
PM. Limited influence of prosthetic position on aseptic loosening of elbow replacements: 125 elbows followed for an average period of 5.6 years. Acta Orthop.
2005;76:654-61.[PubMed] [CrossRef]