Objective: To investigate the long-term follow-up results of total elbow arthroplasty (TEA) with a non-constrained elbow prosthesis with a solid ceramic trochlea (Stemmed Kyocera type I, SKC-I) on patients who have rheumatoid arthritis (RA).

Materials and Methods: Between May 1988 and February 1999, 84 patients of RA underwent TEA with cement fixation (108 elbows). Fourteen patients (14 elbows) were lost due to change of address. One patient (1 elbow) required revision surgery due to aseptic loosening, and 2 elbow implants in 2 individuals were removed due to deep infection, and 18 patients (24 elbows) have died within 5 years after the surgery due to causes unrelated to TEA. Of the remaining 52 patients (70 elbows), 38 patients (53 elbows, 75.7%) were available for the over-5-year detailed clinical and radiographic review at a mean period of 9.2 years (range, 5–17 years). The clinical condition of each elbow before and after operation was assessed according to the Japanese Orthopaedic Association (JOA) elbow scoring system (up to 100 points).

Results: The average postoperative JOA score improved from 45.1 to 83.3 points with marked pain relief in all but 2 cases. The mean range of motion (ROM) of extension / flexion before the surgery was −35.6 / 117.1 and at last follow-up was −17.6 / 136.9 degrees. The mean ROM of pronation / supination improved from 49.6 / 55.8 to 80.0 / 82.1 degrees. Revision surgery was required in 2 patients (2 elbows) due to humeral fracture and ulnar component fracture, respectively. One elbow implant was removed due to deep infection. An ulnar component was removed from one patient with olecranon fracture. During the follow-up, 3 elbow joints had dislocated in 3 patients, and loosening was seen in 5 elbows in 5 patients. Of the 53 elbows, 45 elbows (84.9 %) were judged to have excellent (90–100) or good (75–89) results, and 2 elbows to have poor (< 60) results (0.04%). With loosening and revision or removal of the implant defined as the end point, the likelihood of survival of the prosthesis was 92.2 and 88.3%, respectively, for as long as 10 years by Kaplan-Meier analysis.

Conclusion: The results of the current study showed a high reliability over a long period of the SKC-I when implanted with cement. However, good results in the use of non-constrained devices are limited by the amount of bone and by the need for the ligamentous stability, which can be problematic in RA cases.