Background
Several classification systems have been used to describe early lesions of hip cartilage and the acetabular labrum in young adults with hip pain. Some of them were introduced before the concept of femoroacetabular impingement was proposed. Others were developed for other joints (such as the patellofemoral joint). However, these often demonstrate inadequate reliability, and they do not characterize all possible lesions. Therefore, we developed a novel classification system.
Question/purpose
We asked: What is the (1) intraobserver reliability, (2) interobserver reproducibility, and (3) percentage of nonclassifiable lesions of the new classification system for damage to the hip cartilage and labrum compared with six established classification systems for chondral lesions (Beck et al. [4], Konan et al. [10], Outerbridge et al. [14]) and labral lesions (Beck et al. [3], Lage et al. [12], Peters and Erickson [15])?
Methods
We performed a validation study of a new classification system of early chondrolabral degeneration lesions based on intraoperative video documentation taken during surgical hip dislocations for joint-preserving surgery in 57 hips (56 patients) performed by one surgeon with standard video documentation of intraarticular lesions. The exclusion criteria were low-quality videos, inadequate exposure angles, traumatic lesions, and incomplete radiographic documentation. This left 42 hips (41 patients) for the blinded and randomized analysis of six raters, including those with cam-pincer-type femoroacetabular impingement (FAI) (19 hips in 18 patients), isolated cam-type FAI (10 hips), extraarticular FAI due to femoral anteversion (seven hips), isolated pincer-type FAI (two hips), focal avascular necrosis (two hips), localized pigmented villonodular synovitis (one hip), and acetabular dysplasia as a sequelae of Perthes disease (one hip). The raters had various degrees of experience in hip surgery: Three were board-certified orthopaedic fellows and three were orthopaedic residents, in whom we chose to prove the general usability of the classification systems in less experienced readers. Every rater was given the original publication of all existing classification systems and a visual guide of the new Bern classification system. Every rater classified the lesions according the existing classifications (cartilage: Beck et al. [4], Konan et al. [10], and Outerbridge et al. [14]; labrum: Beck et al. [3], Peters and Erickson [15], and Lage et al. [12]) and our new Bern chondrolabral classification system. The intraclass correlation coefficient with 95% confidence interval was used to assess the intraobserver reliability and interobserver reproducibility. The percentage of nonclassifiable lesions was calculated as an absolute number and percentage.
Results
The intraobserver intercorrelation coefficients (ICCs) for cartilage lesions were as follows: the Bern classification system (0.68 [95% CI 0.61 to 0.70]), Beck (0.44 [95% CI 0.34 to 0.54]), Konan (0.39 [95% CI 0.29 to 0.49]), and the Outerbridge classification (0.57 [95% CI 0.48 to 0.65]). For labral lesions, the ICCs were as follows: the Bern classification (0.70 [95% CI 0.63 to 0.76]), Peters (0.42 [95% CI 0.31 to 0.51]), Lage (0.26 [95% CI 0.15 to 0.38]), and Beck (0.59 [95% CI 0.51 to 0.67]). The interobserver ICCs for cartilage were as follows: the Bern classification system (0.63 [95% CI 0.51 to 0.75), the Outerbridge (0.14 [95% CI 0.04 to 0.28]), Konan (0.58 [95% CI 0.40 to 0.76]), and Beck (0.52 [95% CI 0.39 to 0.66]). For labral lesions, the ICCs were as follows: the Bern classification (0.61 [95% CI 0.49 to 0.74]), Beck (0.31 [95% CI 0.19 to 0.46]), Peters (0.28 [95% CI 0.16 to 0.44]), and Lage (0.20 [95% CI 0.09 to 0.35]). The percentage of nonclassifiable cartilage lesions was 0% for the Bern, 0.04% for Beck, 17% for Konan, and 25% for the Outerbridge classification. The percentage of nonclassifiable labral lesions was 0% for Bern and Beck, 4% for Peters, and 25% for Lage.
Conclusion
We have observed some shortcomings with currently used classification systems for hip pathology, and the new classification system we developed seems to have improved the intraobserver reliability compared with the Beck and Konan classifications in cartilage lesions and with the Peters and Lage classifications in labral lesions. The interrater reproducibility of the Bern classification seems to have improved in cartilage lesions compared with the Outerbridge classification and in labral lesions compared with the Beck, Peters, and Lage classifications. The Bern classification identified all present cartilage and labral lesions. It provides a solid clinical basis for accurate descriptions of early degenerative hip lesions independent of etiology, and it is reproducible enough to use in the reporting of clinical research. Further studies need to replicate our findings in the hands of nondevelopers and should focus on the prognostic value of this classification and its utility in guiding surgical indications.
Level of Evidence
Level II, diagnostic study.