INTRA- AND INTEROBSERVER VARIABLITY OF WEAR MEASUREMENTS FOR TOTAL HIP ARTHROPLASTY: THE IMAGIKA(R) SYSTEM

Abstract

Purpose: We tested the reliability of a digitalised x-ray reading system, Imagika(r), used to measure linear wear of total hip arthroplasy on the AP view of the pelvis.

Material and methods: Wear measurements were taken for total hip arthroplasties without cement (n=20) and with cement (n=19) using the distance between the centre of the acetabular cup and the femoral ball. The system delivered measures in hundredths of millimetres that were rounded off to the nearest tenth millimetre. For non-cemented implants, the centre of the acetabular cup was found automatically on the digitalised radiograms using the contour of the metal socket. For cemented cups, the centre of the cup was determined from five points situated on the metallic ellipse included in the polyethylene circumference. The software placed the point clicked by the reader on the adjacent intermediary zone showing the greatest contrast. Five observers read the radiograms twice at 15 day intervals. The observers were a young resident, a senior traumatology surgeon,and a senior surgeon specialised in hip surgery. Results were compared to determine inter- and intra- observer variability.

Results: Intra-observer variability was low since the standard deviation (at alpha error set at 5%) ranged from one tenth of a millimetre to six-tenths of a millimetre for four observers. It was higher (2 millimetres) for the fourth observer. The younger observers achieved the best reproducibility, to the order of a tenth of millimetre. Conversely, interobserver variability was high with standard deviation of several millimetres for an alpha risk of 5%. Comparing the two observers who achieved the best performances, the standard deviation of the measures was in the 3 to 4 millimetre range.

Discussion: Measurement precision was greater for cemented cups. Conversely, for press-fit cups, the contour of the head was sometimes difficult to distinguish even with optimal contrast and measurement deviations were to the order of one millimetre.

Conclusion: The reproducibility of the Imagika(r) system is insufficient to measure wear of total hip arthroplasty where the precision must be to the order of a tenth of a millimetre.