Use of Anatomic Radiographic Horizons for the Lateral Center-Edge Angle in the Classification of Hip Dysplasia

Background: The lateral center-edge angle of Wiberg (LCEA) and Tönnis angle (TA) rely on a "horizon" that functions as a reference point for measurements of hip dysplasia on weightbearing anteroposterior pelvis radiographs. There are 3 different horizons that are currently utilized depending on surgeon preference: (1) a line parallel to the radiographic floor (F), (2) a line that connects the inferior portions of the acetabular teardrops (ATs), or (3) a line that connects the ischial tuberosities (ITs). It is imperative to accurately assess the degree of hip dysplasia on initial workup to select the appropriate surgical intervention.

Purpose: To assess how the choice of a radiographic horizon affects the measurements of the LCEA and TA. The secondary purpose was to assess how the horizon affected the classification of hips as either normal, borderline dysplastic, or dysplastic.

Study design: Cohort study (diagnosis); Level of evidence, 3.

Methods: The LCEA and TA were measured on all preoperative weightbearing anteroposterior pelvis radiographs for 186 consecutive patients who underwent hip preservation surgery between February 2016 and November 2020 (140 hip arthroscopic procedures, 46 combined hip arthroscopic procedures with periacetabular osteotomy), 3 times per hip, each time using an F, AT, and IT horizon. The Student t test was used to analyze the differences in the measurements of the LCEA and TA, and discordance rates in the dysplasia classification between horizons were quantified.

Results: For the entire study population, the mean LCEA_F (23.4°± 7.4°) was significantly greater than the mean LCEA_IT (23.1°± 7.2°) (95% CI, -0.634 to -0.003; P = .047) and mean LCEA_AT (23.0°± 7.1°) (95% CI, -0.723 to -0.139; P = .004). There was no difference between the LCEA_IT and LCEA_AT (95% CI, -0.305 to 0.080; P = .251). When stratified by hip arthroscopic surgery, the mean LCEA_F (26.3°± 5.6°) was significantly greater than the mean LCEA_AT (25.8°± 5.3°) (95% CI, -0.845 to -0.162; P = .004). Overall, there was a 17.7% and 18.3% discordance rate in the dysplasia classification using the LCEA_IT and LCEA_AT compared with the LCEA_F, respectively. There were no statistically significant differences between the mean TA_IT, TA_AT, and TA_F for the entire study population, the arthroscopic surgery group, and the combined arthroscopic surgery and periacetabular osteotomy group.

Conclusion: There was no statistical difference between the AT and IT horizons for LCEA measurements. The dysplasia classification was in better agreement between the 2 anatomic horizons compared with the F horizon. The TA was not affected by changes in the horizon.