The final follow-up plain radiograph is sufficient for clinical evaluation of polyethylene wear in total hip arthroplasty: A study of validity and reliability

Background and purpose

Radiostereometric analysis (RSA) is a highly accurate tool for assessment of polyethylene (PE) wear in total hip arthroplasty (THA); however, PE wear measurements in clinical studies are often limited to plain radiographs. We evaluated the agreement between PE wear measured with PolyWare software, which uses plain radiographs, and by model-based RSA, which uses stereo radiographs.

Methods

Measurements of PE wear postoperatively and at final follow-up (after mean 6 years) on plain radiographs of 12 patients after cementless THA were evaluated with PolyWare software and the results were compared with those from RSA as the gold standard (Model-based RSA using elementary geometrical shape models; EGS-RSA). With PolyWare, we either used the final radiographic follow-up (PW1) only or both the postoperative follow-up and the final follow-up (PW2).

Results

The 2D mean wear measured (in mm) was 0.80, 1.07, and 0.60 for the PW2, PW1, and RSA method. 2D intra-method repeatability was similar for PW1 and RSA with limits of agreement (LOAs, in mm) of ± 0.22, and ± 0.23, respectively. 2D inter-method concurrent validity was best between PW1 and EGS-RSA with LOAs of ± 0.55. For 2D linear wear measurements, the PW1 method had a clinical repeatability similar to that of RSA.

Interpretation

PW1 is sufficient for retrospective determination of 2D wear from medium-term wear measurements above 0.5 mm, It alleviates the need for baseline plain radiographs, has a clinical precision similar to that of RSA, and is easy and inexpensive to use.Wear of polyethylene (PE) components is widely regarded as the main factor limiting longevity of total hip arthroplasty (THA) (Cooper et al. 1992). Clinical studies have shown that periprosthetic osteolysis and aseptic loosening is strongly related to wear rates of above 0.2 mm/year (Sochart 1999, Dowd et al. 2000).Radiostereometric analysis (RSA) is the most accurate tool for in vivo assessment of PE wear (Kärrholm et al. 1997, von Schewelov et al. 2004, Bragdon et al. 2006), and it is regarded as the gold standard (Ilchmann et al. 1995). However, many radiographic in vivo studies of PE wear in THA are restricted to measurements on plain radiographs because the RSA set-up is expensive and not widely available. Measurement of PE wear on plain radiographs is often limited to 2D analysis because poor quality of cross-table lateral radiographs is a common problem (Sychterz et al. 1999b, 2001). Although PE wear is known to occur multidirectionally (Yamaguchi et al. 1997, Akisue et al. 1999), the bulk of the wear is detectable on the anterior-posterior radiographs alone (Sychterz et al. 1997, Hui et al. 2003, Martell et al. 2003). Based on the availability of radiographs and investigator preferences, some authors favor analysis of serial radiographs (Sychterz et al. 1997, Kim et al. 2001, Hernigou and Bahrami 2003) to describe the pattern of wear and the steady-state wear (Sychterz et al. 1999a, Bragdon et al. 2006), whereas others use 2 radiographic follow-ups (postoperative and latest) (Kraay et al. 2006), or only the latest radiographic follow-up with the assumption of zero wear at baseline (Norton et al. 2002)Little is known about the conformity between PE wear results measured with RSA and computerized methods using plain radiographs (Ilchmann et al. 1995, von Schewelov et al. 2004, Bragdon et al. 2006). Our group has questioned the conformity of 2D PE wear measurements based on serial, 2, or 1 radiographic follow-up (Stilling et al. 2009b). We determined that there was a statistically significant difference between all approaches, but we were unable to determine which strategy best reflected the true extent of wear (Stilling et al. 2009b). In addition, we recently showed that model-based RSA is an accurate tool for measurement of PE wear in good agreement with the true wear (Stilling 2009).We have now studied the intra-method repeatability and concurrent validity between 2 methods (PolyWare and EGS-RSA) for measurement of PE wear in THA, in a group of patients with an average follow-up of 6 years. We wanted to determine (1) whether there would be a difference in repeatability between the methods, (2) whether there would be a difference in wear measured using 1 or 2 radiographic follow-ups with the PolyWare method, and (3) whether either of the 2 PolyWare measurement strategies (1 or 2 radiographic follow-ups) would give results similar to the wear measured by RSA (concurrent validity).