Background: In vivo measurement of wear in the ball and socket articulation of total hip arthroplasties is of interest in the evaluation of both existing and new implants. Controversy reigns regarding the accuracy of different radiological measurement techniques and in particular how accuracy has been assessed.
Material and methods: We assessed the accuracy of 2 radiostereometric (RSA) techniques for wear measurement and 3 standard radiographic techniques, namely Imagika (image analyzing software), Imagika corrected for head center displacement, and the Charnley Duo method. 5 custom-made adjustable phantoms with different prosthetic components were used.
Results: In 20 measurements of all 5 phantoms at 3 levels of simulated wear (0.2 mm, 1.0 mm and 1.5 mm), the mean measurement error of the digital RSA examinations was 0.010 mm (accuracy 0.42). The corresponding error values for the three radiographic techniques were 0.19 (accuracy 1.3) for Charnley Duo, 0.13 (accuracy 1.3) for Imagika corrected, and 1.021 (accuracy 2.99) for Imagika. Measurement error decreased from 0.011 mm with ordinary RSA to 0.004 with RSA digital measurement. Head size, direction of wear in relation to the cup or type of prosthetic component did not influence the measurement error. The results of Charnley Duo and Imagika corrected were similar but the latter had an inexplicable systematic error in measuring one of the phantoms. Imagika had the worst results due to its inability to compensate for the out-of-head center effect. Alumina heads were difficult to analyze with all methods.
Interpretation: By using the ISO standard for assessing accuracy, RSA can be expected to measure wear with an accuracy of about 0.4 mm irrespective of prosthetic component studied or direction of wear, whereas the best technique, in our study, based on standard radiographs can be accurate to about 1.3 mm.