Correlations between ultrasonographic and subsequent radiographic findings of developmental dysplasia of the hips

Abstract

Purpose: The study aimed to investigate the utility of ultrasonographic (US) findings in predicting the subsequent radiographic parameters of developmental dysplasia of the hips. Methods: In this 12-year retrospective cohort study, all new-born infants with a positive clinical examination or risk factors were included. They were scheduled for hip ultrasonography in the first 3 months, and subsequent radiographs at 1 year of life. The US images were evaluated using the Graf classification, Harcke’s dynamic screening method, and Terjesen’s femoral head coverage method. The radiographic images were evaluated using the acetabular index and femoral head position. The overall US or radiographic findings were considered abnormal if they were classified as abnormal for any of their respective parameters. The overall US and radiographic parameters were correlated. Results: A total of 160 patients were included. The overall US and radiographic parameters showed no statistically significant difference (P=0.050). The sensitivity, specificity, and accuracy of the overall US parameters were 57.1%, 84.9%, and 81.3%, respectively. All three individual US parameters showed no statistically significant differences, with the overall radiographic findings and acetabular index (P>0.05). However, they showed a statistically significant difference, with the position of the femoral head (P<0.001), with the US parameters having an excellent negative predictive value of 100% for identifying an abnormal femoral head position. Conclusion: The current study suggests that US findings evaluated in the first 3 months of life showed no statistically significant difference with radiographic findings evaluated at 1 year of life. The US parameters showed an excellent negative predictive value for abnormal femoral head position on radiographs.