How Is Variability in Femoral and Acetabular Version Associated With Presentation Among Young Adults With Hip Pain?

Background Acetabular and femoral version contribute to hip pain in patients with femoroacetabular impingement (FAI) or dysplasia. However, definitions and measurement methods of femoral version have varied in different studies, resulting in different “normal” values being used by clinicians for what should be the same anatomic measurement. This could result in discrepant or even inappropriate treatment recommendations. Questions/purposes In patients undergoing hip preservation surgery, (1) what is the range of acetabular and femoral version at presentation, and how much do two commonly used measurement techniques (those of Murphy and Reikerås) differ? (2) How are differences in acetabular and femoral version associated with clinical factors and outcomes scores at the time of presentation? Methods This was a retrospective analysis of data gathered in a longitudinally maintained database of patients undergoing hip preservation at a tertiary care referral center. Between June 2020 and December 2021, 282 hips in 258 patients were treated for an isolated labral tear (9% [26 hips]), hip dysplasia (21% [59 hips]), FAI (52% [147 hips]), mixed FAI and dysplasia (17% [47 hips]), or pediatric deformity (slipped capital femoral head epiphysis or Perthes disease; 1% [3 hips]) with hip arthroscopy (71% [200 hips]), periacetabular osteotomy (26% [74 hips]), surgical hip dislocation (2.5% [7 hips]), or femoral derotation osteotomy (0.5% [1 hip]). We considered those with complete radiographic data (CT including the pelvis and distal femur) and patient-reported outcome scores as potentially eligible. Exclusion criteria were age younger than 18 or older than 55 years (5 hips, 3 patients), signs of hip osteoarthritis (Tönnis grade ≥ 2; 0), pediatric deformity (slipped capital femoral head epiphysis or Perthes disease; 3 hips, 3 patients), previous femoral or acetabular osteotomy (2 hips, 2 patients), avascular necrosis of the femoral head (0), history of neuromuscular disorder (Ehlers-Danlos syndrome; 3 hips, 3 patients) or rheumatoid disease (ankylosing spondylitis; 1 hip, 1 patient), and when CT did not include the knees (19 hips, 19 patients). Based on these criteria, 249 hips in 227 patients were included. Of patients with bilateral symptomatic hips, one side was randomly selected for inclusion, leaving 227 hips in 227 patients for further analysis. The patients’ median age (range) was 34 years (19 to 55 years), the median BMI (range) was 27 kg/m2 (16 to 55 kg/m2), and 63% (144) were female; they were treated with hip arthroscopy (in 74% [168]) or periacetabular osteotomy (in 23% [52]). Patients underwent a CT scan to measure acetabular version and femoral version using the Murphy (low < 10°; normal: 10° to 25°; high > 25°) or Reikerås (low < 5°; normal: 5° to 20°; high > 20°) technique. The McKibbin index was calculated (low: < 20°; normal: 20° to 50°; high > 50°). Based on the central acetabular version and femoral version as measured by Murphy, hips were grouped according to their rotational profile into four groups: unstable rotational profile: high (high acetabular version with high femoral version) or moderate (high acetabular version with normal femoral version or normal acetabular version with high femoral version); normal rotational profile (normal acetabular version with femoral version); compensatory rotational profile (low acetabular version with high femoral version or high acetabular version with low femoral version); and impingement rotational profile (low acetabular version with low femoral version): high (low acetabular version with low femoral version) or moderate (low acetabular version with normal femoral version or normal acetabular version with low femoral version). Radiographic assessments were manually performed on digitized images by two orthopaedic residents, and 25% of randomly selected measurements were repeated by the senior author, a fellowship-trained hip preservation and arthroplasty surgeon. Interobserver and intraobserver reliabilities were calculated using the correlation coefficient with a two-way mixed model, showing excellent agreement for Murphy technique measurements (intraclass correlation coefficient 0.908 [95% confidence interval 0.80 to 0.97]) and Reikerås technique measurements (ICC 0.938 [95% CI 0.81 to 0.97]). Patient-reported measures were recorded using the International Hip Outcome Tool (iHOT-33) (0 to 100; worse to best). Results The mean acetabular version was 18° ± 6°, and mean femoral version was 24° ± 12° using the Murphy technique and 12° ± 11° with the Reikerås method. Eighty percent (181 of 227) of hips had normal acetabular version, 42% (96 of 227) to 63% (142 to 227) had normal femoral version per Murphy and Reikerås, respectively, and 67% (152 to 227) had a normal McKibbin index. Patients with an impingement profile (low acetabular version or femoral version) were older (39 ± 9 years) than patients with an unstable (high acetabular version or femoral version; 33 ± 9 years; p = 0.004), normal (33 ± 9 years; p = 0.02), or compensatory (high acetabular version with low femoral version or vice versa; 33 ± 7 years; p = 0.08) rotational profile. Using the Murphy technique, femoral version was 12° greater than with the Reikerås method (R2 0.85; p < 0.001). There were no differences in iHOT-33 score between different groups (impingement: 32 ± 17 versus normal 35 ± 21 versus compensated: 34 ± 20 versus unstable: 31 ± 17; p = 0.40). Conclusion Variability in femoral version is twice as large as acetabular version. Patients with an impingement rotational profile were older than patients with a normal, compensatory, or unstable profile, indicating there are other variables not yet fully accounted for that lead to earlier pain and presentation in these groups. Important differences exist between measurement methods. This study shows that different measurement methods for femoral anteversion result in different numbers; if other authors compare their results to those of other studies, they should use equations such as the one suggested in this study. Level of Evidence Level III, prognostic study.