Ultrasound measurements in the management of unstable hips treated with the pavlik harness: reliability and correlation with outcome.

The purposes of this study were to determine the interobserver and intraobserver reliability of ultrasound measurements in unstable neonatal hips treated with the Pavlik harness and to determine whether ultrasound measurements correlate with radiological outcome at 6 months. Sixty-four babies treated from birth with the Pavlik harness for neonatal hip instability were scanned at 2 and 6 weeks. The alpha and beta angles of Graf, the combined (H) angle of Hosny, and the femoral head coverage (FHC) were measured by 3 observers and remeasured by each observer on a minimum of 50 scans. From 248 scans, 792 sets of measurements were made. Hips were categorized as normal, abnormal, or borderline for each parameter; and interobserver and intraobserver repeatability coefficients and Kappa values were calculated. The alpha angle had the smallest interobserver range (17 degrees), the H angle range was 21 degrees, and the beta angle 28 degrees. Kappa values were best for the FHC and beta angle (0.66-0.8). The mean acetabular index (AI) of all hips at 6 months was 26 degrees (SD, 4.9). The AI was 30 degrees or greater in 24 hips (18 babies) despite prolonged splintage in 9 hips (6 babies). A stepwise linear regression analysis showed that the FHC at 6 weeks was predictive of AI at 6 months (regression coefficient -0.27; 95% confidence interval -0.42 to -0.12; P<0.001). We recommend the FHC as being reproducible, useful, and predictive of outcome in neonatal hips treated for instability.