Reliability of Tönnis classification in early hip arthritis: a useless reference for hip-preserving surgery.

Abstract

The Tönnis classification is widely accepted for grading hip arthritis, but its usefulness as a reference in hip-preserving surgery is yet to be demonstrated. We aimed to evaluate reproducibility of the Tönnis classification in early stages of hip osteoarthritis, and thus determine whether it is a reliable reference for hip-preserving surgery. Three orthopaedic surgeons with different levels of experience examined 117 hip X-rays that were randomly mixed of two groups: a group of 31 candidates for hip-preserving surgery and a control group of 30 patients that were asymptomatic with respect to the hip joint. The surgeons were asked to rate an eventual osteoarthritis according to the Tönnis classification. After 2 months, the surgeons were asked to re-evaluate the X-rays in a random order. Intra- and interobserver reliabilities were calculated by comparing the observers' two estimations using Kappa statistics. Kappa values for interobserver reliability were slight or fair (range 0.173-0.397). Kappa values for intraobserver reproducibility were fair (range 0.364-0.397). Variance in grading no and slight osteoarthritis was the most frequent cause for intra- and interobserver disagreements (76.3 and 73.01% of the non-concordant observations, respectively). The confidence interval analysis showed that the observers' experience did not affect reproducibility. The Tönnis classification is a poor method to assess early stages of hip osteoarthritis. These findings suggest that its routine use in therapeutic decision-making for conservative hip surgery should be reconsidered.