Reliability analysis of radiologic and intraoperative loosening in total elbow arthroplasty

Abstract

Background: Revisions of total elbow arthroplasties (TEAs) are commonly performed because of prosthetic loosening. UK National Joint Registry data show that TEA revisions are becoming increasingly common, with 123 TEA revisions performed in 2018 and 76 performed in 2017. TEA radiologic assessment is based on subjective interpretation with no published criteria. We defined TEA loosening by the presence of at least one of the following criteria: (1) progressive widening of the bone-cement, bone-prosthesis, or cement-prosthesis interface; (2) fragmentation or fracture of cement; (3) prosthetic component migration; and (4) bead shedding in porous-coated prostheses. Using this definition, we looked at interobserver and intraobserver agreement on radiologic loosening and compared this assessment with intraoperative findings.

Methods: In our tertiary care center, we conducted a retrospective review to identify TEA revisions performed between November 2008 and July 2018. Radiologic implant loosening was independently assessed by 5 orthopedic surgeons. Interobserver agreement (κ coefficient) was calculated. The majority's view of radiologic loosening was compared with intraoperative findings.

Results: We identified 93 sets of radiographs with implant stability clearly documented in their operative notes. The κ coefficient between assessors for humeral implant loosening was 0.87 (almost perfect). The κ coefficient for ulnar loosening was 0.75 (substantial). The κ coefficients for radiologic and intraoperative findings of humeral loosening and ulnar loosening were 0.67 and 0.71 (substantial), respectively. Intraobserver reliability was almost perfect for humeral loosening (κ = 0.86) and substantial for ulnar loosening (κ = 0.74).

Conclusion: Our definition of loosening provides reproducible interobserver and intraobserver agreement on radiographic component loosening. In our center's experience, radiologic findings may not translate to intraoperative findings, and we would advise that surgical strategies for TEA revision include the possibility of needing to perform a dual-implant exchange.

Keywords: Reliability; criteria; elbow arthroplasty; loosening; radiographs; revision.