Abstract
Summary
Comparison of accuracy and precision in measuring wear using 4 commonly used uncemented cup designs shows small differences in mean and data scatter for marker and model-based RSA.
Introduction
The disadvantage with conventional RSA is that implant has to be supplied with tantalum markers, which may be difficult to visualise. This problem can be resolved with model-based RSA, but it is uncertain if this method has the same precision as marker-based RSA to measure wear. We compared these methods and studied different prosthesis geometries represented by four different uncemented cup designs (Trilogy, TMT-Trabecular Metal, Zimmer, Warsaw, USA, Ringloc, Biomet, Inc., Warsaw, Indiana, and ABG, Howmedica International, Staines, UK).
Patients and Methods
Stereoradiographs of 75 patients (19 Trilogy, 17 TMT, 20 Ringloc, 19 ABG) were exposed postoperatively (2 examinations) and after 2 years. The patients were selected from prospective clinical studies. During operation tantalum markers had been inserted into the liner in all cases. The measurements and analysis of all radiographs were performed with UmRSA Digital Measure and UmRSA Analysis 6.0. We used the differences between the postoperative double-examinations to compute the precision for the two methods and for the different implant designs. The proximal and the total (vectorial sum of medial/lateral, proximal/distal and anterior posterior) femoral head penetration up to 2 years were compared.
Results
The mean differences and the standard deviation of mean obtained from calculations between the double examinations in the total material did not differ between the 2 methods for any of the designs studied. The mean values and SD for marker and model-based RSA were −0, 00 mm ± 0.09 mm and 0.02 mm ± 0.08 mm, respectively (p>0.05). The comparison between Classical marker-based RSA and Model-based RSA in measuring wear up to 2 years did not showed any statistically significant differences for the Trilogy, TMT and ABG cups (p>0.05). However the mean difference of the postoperative double examinations were slightly higher for the Ringloc design (p=0, 02) and the data scatter (SD) at 2 years was higher (p=0,004) with use of model-based RSA.
Conclusions
We found small differences between marker and model-based RSA for measurements of proximal and total wear (penetration). In 3 of the 4 cup designs studied the data scatter was about equal for the 2 methods. In the 4th design (Ringloc) the data scatter was higher when model-based RSA was used.
