What factors can reduce the need of repeat revision for humeral loosening in revision total elbow arthroplasty?

Abstract

This study aims to determine re-revision rate in a cohort who underwent revision total elbow arthroplasty (rTEA) for humeral loosening (HL) and identify factors contributing to re-revision. We hypothesize that proportional increases in the stem and flange lengths will stabilize the bone-implant interface significantly more than disproportional increases in stem or flange length alone. Additionally, we hypothesize that the indication for the index arthroplasty will impact the need for repeat revision for HL. The secondary objective was to describe the functional outcomes, complications, and radiographic loosening after rTEA. We retrospectively reviewed 181 rTEAs performed from 2000-2021. Forty rTEAs for HL performed on 40 elbows that either required a subsequent revision for humeral loosening (10 rTEAs) or had a minimum of two years of clinical or radiographic follow-up were included. One hundred thirty-one cases were excluded. Patients were grouped based on stem and flange length to determine the re-revision rate. Patients were divided based on re-revision status into a single revision group and re-revision group. The stem to flange lengths (S/F) ratio was calculated for each surgery. Mean clinical and radiographic follow-up was 71 months (range, 18-221 months; 3-221 months, respectively). Rheumatoid arthritis (RA) was statistically significant in predicting re-revision TEA for HL (p value = 0.024). The overall re-revision rate for HL was 25% at average 4.2 years (range, 1-19) from revision procedure. There was a significant increase in stem and flange lengths from index procedure to revision, on average by 70±47mm (p<0.001) and 28±39mm (p<0.001), respectively. In cases of re-revisions (n=10), four patients went into excisional procedure, and in the remaining six cases the size of re-revision implant increased on average by 37±40mm for stem and 73±70mm for flange (p=0.075 and p=0.046). Furthermore, average flange in these 6 cases was 7 times shorter than average stem (S/F=6.7±2.2). This was significantly different from cases were not re-revised (p=0.03; S/F=4.6±1.8 and 4.2±2, respectively). Mean range of motion was 16° (range: 0°-90°; SD: 20°) to 119° (range: 0°-160°; SD: 39°) at final follow-up. Complications included ulnar neuropathy (38%), radial neuropathy (10%), infection (14%), ulnar loosening (14%), and fracture (14%). None of the elbows were considered radiographically loose at final follow-up. We show that the primary diagnosis of RA and a humeral stem with a relatively short flange relative to the stem length significantly contribute to re-revision of TEA. The use of an implant where flange can be extended beyond one fourth of the stem length may increase implant longevity.