ABSTRACT
Peri-acetabular radiolucent lines (RLLs) seen on "early" post-operative radiographs have been identified as a potential predictor of long-term implant performance. This study examines the inter- and intra-observer variation encountered when assessing such radiographs. Four consultant orthopaedic surgeons assessed the presence, extent and width of RLLs in 220 radiographs performed on 50 patients taken one to two weeks, six weeks, six months and one year following surgery. Inter-observer agreement was fair at 7-14 days but improved to moderate to good in films at six and 12 months. Intra-observer agreement was moderate to good at 7-10 days but again improved to good at 6 and 12 months. When only the presence or absence of RLLs was considered, both inter-observer and intra-observer agreement improved for both the six-month and one-year radiographs. This experiment shows that caution must be used for the interpretation of RLLs on hip radiographs taken during the very early post-operative period. We recommend that films taken at least six weeks to six months following surgery should be used for assessment to reduce observer variation. For optimum results, a single experienced observer should do the assessment with a simple classification.
