The elbow joint has high morbidity, followed by joints in the finger, the wrist, the toe, and the knee. The incidence of elbow joint destruction in RA patients is 20–50% (Souter 1990, Mansat 2001). Continuous synovitis may cause progressive joint destruction, which results in severe joint instability or contracture.

The local treatment of continuous synovitis includes physiotherapy and intraarticular injections of steroids. However, when the effects of these treatments are insufficient, surgical treatment is often recommended: synovectomy or total elbow arthroplasty (TEA). Synovectomy is often reported as a successful procedure with relatively few complications for the treatment of rheumatoid elbows at radiographically early stages (Larsen grades 1 and 2) (Stein et al. 1975, Brumfield and Resnick 1985, Saito et al. 1986, Ferlic et al. 1987, Tulp and Winia 1989, Koshino et al. 1991, Gendi et al. 1997, Fuerst et al. 2006). TEA is usually performed for joints at radiographically advanced stages (Larsen grades 3–5) (Figgie et al. 1989, Connor and Morrey 1998, Mansat and Morrey 2000). However, for cases with good ligament stability, we have even performed synovectomies for Larsen grades 3–5, which, however, is controversial (Nestor 1998).

In this study, we investigated the long-term outcome of synovectomy that was performed for the treatment of rheumatoid elbows at radiographically advanced stages.

Functional outcome was assessed using the Mayo elbow performance score (MEPS) (Morrey 1991). Conventional anteroposterior and lateral radiographs were taken before surgery and at the final follow-up examination. The grades of Larsen’s classification and the elbow valgus angle before surgery and at the final follow-up examination were evaluated (Larsen et al. 1977).

Medium-term results of elbow synovectomy with an average follow-up period of up to 10 years have been published by several authors. Ferlic et al. (1987) reported excellent results with an average follow-up of 7 years and a maximum follow-up of 20 years. In their report, 44/57 elbows had excellent results, and they had better clinical results in patients who underwent surgery at earlier stages of the disease. Brumfield and Resnick (1985) stated that synovectomy was not contraindicated even for joints at radiographically advanced stages, such as Larsen grade 3 or 4, because an improvement in the range of motion can be expected. We also performed elbow synovectomy at advanced stages and this was not followed by substantial joint instability. In the present study, 12/16 elbows required conversion to TEA after a minimum of 10 years and an average of 15 years of follow-up. We consider this result to be more favorable than those in previous reports that included synovectomies for elbows at radiographically advanced stages (Ferlic et al. 1987, Fuerst et al. 2006). Recurrence of severe pain from progressive joint destruction was the most common cause of the conversion to TEA: 4 of 7 elbows of Larsen grade 5 underwent conversion.

During an elbow synovectomy, the radial head is often resected. Resection of the radial head enhances the operative visual field and the performance of an adequate resection of the joint synovium. Additionally, resection of the radial head improves flexion, especially in cases with anterior subluxation of the radial head. However, resection of the radial head during an elbow synovectomy is controversial. Ferlic et al. (1987) reported that there was no difference in clinical results between cases that had resection of the radial head and those with resection and radial head replacement. Lehtinen et al. (2001) reported that the elbow seemed to turn into valgus during rheumatoid destruction and resection of the radial head. Rymaszewski et al. (1984) suggested that resection of the radial head caused joint instability in the long term, and recommended radial head replacement rather than radial head resection. In addition, Taylor et al. (1976) and Copeland and Taylor (1979) reported that resection of the radial head caused progression of valgus deformities of the elbow in more than half of their cases. They also found that excessive axial pressure to the ulna by resection of the radial head caused pain on the ulnar side of the wrist. We resected the radial head in all our cases with only a slight increase in the valgus angle at the final follow-up examination (mean 2 degrees). At the final examination, only 2 elbows showed marked valgus deformities with severe instability of the elbow joint. Thus, we believe that resection of the radial head is appropriate.

One of the limitations of our study was the evaluation of joint instability, which was assessed clinically, not radiographically. Radiographic examination is a more precise method.