O2146: Does a Park-Harris Line of the Proximal Femur Predict Proximal Femoral Growth Disturbance in Developmental Hip Dislocation?

Background: O’Brien first described a Park-Harris line of the proximal femur in patients treated for developmental hip dislocation which reportedly predicted the absence of avascular necrosis/proximal femoral growth disturbance (PFGD). However, these findings have not been validated in a modern population of patients undergoing closed or open reduction with objective classification systems of PFGD. The purpose of this study is to evaluate the diagnostic utility of “O’Brien’s line” as a predictor of PFGD.

Methods: We identified patients treated with open or closed reduction for unilateral developmental hip dislocation between 1980 and 2011. All patients had radiographic analysis at 3 time points: within 1 year, 12-24 months, and skeletal maturity. Each postoperative radiograph was evaluated for presence and displacement of O’Brien’s line at the capital femoral physis (medial) and the trochanteric apophysis (lateral) and the ratio between them (Figure 1). PFGD was classified using Salter’s criteria at 12-24 months following initial reduction, and by Bucholz-Ogden (BO) classification at skeletal maturity. The association between O’Brien line and proximal femoral growth disturbance was then analyzed.

Results: 117 patients (83% female, 72% closed reductions) underwent closed or open reduction at a mean age 1.15±0.49 years. O’Brien’s line was identified in 81 (69.2%) patients at 4.4±1.6 months from reduction and persisted for 20.7±10.9 months with a mean growth ratio of 1.73±0.57. O’Brien’s line appeared less often in patients < 1 year of age (46.4% vs 81.6%, p = < 0.001) and patients with milder dislocations (20.0% IHDI II hips vs 67.5% IHDI III hips vs 86.8% IHDI IV hips, p < 0.001). When comparing the cohorts of hips with or without the development of O’Brien’s line, there was no difference in the incidence of PFGD by Salter’s criteria (19.8% vs 27.3%, p = 0.8) or BO classification 2-4 (23.4% vs 22.2% p=1.0). Similarly, there was no difference in the ratio of growth displacement for O’Brien’s line between those hips with and without PFGD (p>0.05 in each). The predictive value that a hip with O’Brien’s line will not develop PFGD (PPV) was superior for patients age >1 year by Salter’s criteria (83.9% vs 68.4%) and by BO classification (80.6% vs 63.2%). The predictive value that a hip without O'Brien's line would develop PFGD (NPV) for this cohort was 27.3% by Salter’s criteria and 22.2% by BO classification. The accuracy of O’Brien’s line as a test was 39% in patients < 1 year old and 72% after age 1.

Conclusion: An O’Brien’s Line is less likely to form in patients < 1 year of age and those with less severe dislocations. Overall, while the appearance of an O’Brien’s line may be a reassuring sign following reduction of developmental hip dislocation, the prognostic value of its absence for predicting PFGD was low.