Outcomes of radial head implants in total elbow arthroplasty.

Abstract

There is no consensus on the management of the radial head in total elbow arthroplasty (TEA). In 3-part TEA designs, options include radial head retention, excision, or arthroplasty. Biomechanical studies suggest improved varus-valgus stability with radial head implants in unlinked total elbows. Unfortunately, complications with radial head implants have been common with historical designs. The aim of this study was to evaluate the clinical and radiographic outcomes of radial head implants in a current 3-part TEA and identify risk factors for mechanical failure. We performed a retrospective review of radial head implants with a 3-part convertible TEA from 2001 to 2016. Clinical outcomes, functional scores, and radiographic outcomes were recorded. The preoperative radiocapitellar alignment was measured using the radiocapitellar ratio (RCR). Statistics include descriptive statistics, t tests, logistic regression, and Kaplan-Meier survival curves. We identified 44 TEAs in 40 patients, with a mean follow-up period of 7.2 years. The average age at surgery was 58 ± 11 years; 80% of the TEAs were performed in women. The indication for surgery was rheumatoid arthritis in 86%; of the implants, 61% were unlinked. The average preoperative RCR was 10.7 ± 17.9. Postoperatively, 2 radial head implants (5%) were subluxated, 6 (14%) were dissociated, and 2 (5%) were dissociated with implant dislocation on radiographic review. The revision rate for radial head subluxation, dissociation, or dislocation was 7% (n = 3). Univariate logistic regression showed that male sex (P = .002), abnormal preoperative RCR (P = .02), linked implant (P = .03), and older age (P = .04) were risk factors for radial head subluxation, dissociation, or implant dislocation. A multivariate model with all 4 variables did not demonstrate statistical significance. The incidence of radial head arthroplasty subluxation, dissociation, or implant dislocation was high (23%). In a univariate logistic regression model, male sex, abnormal preoperative RCR, and linked implants were all statistically significant risk factors for mechanical failure of the radial head implant. Our multivariate model did not show any statistically significant independent risk factors. Polyethylene wear or loosening of the radial head implants was not observed in this study; failure of the bipolar linkage was the principal mode of failure. Although further study is required, caution should be used when considering inserting a radial head implant in male patients with significant preoperative radiocapitellar malalignment. Radial head subluxation or dissociation is not an absolute indication for revision in an asymptomatic patient. Improvements in radial head implant designs in TEA are needed.