Background: Ultrasonographic evaluation of the hip in infants is considered both reliable and reproducible in the diagnosis of developmental dysplasia of the hip. Ultrasonographic evaluation of the shoulder in infants has been reported as a valuable diagnostic aid in dysplastic development following neonatal brachial plexus palsy. To our knowledge, there has been no study of the intraobserver reproducibility and interobserver reliability of sonography of the shoulder in infants with and without suspected posterior shoulder dislocation.

Methods: Two identical but randomly ordered sets of the same deidentified sonographic images of shoulders in infants were given to radiologists, pediatric orthopaedists and orthopaedic residents, and fellows with varying degrees of experience in the evaluation of shoulder pathology in infants, who measured the position of the humeral head relative to the axis of the scapula. Intraobserver reproducibility and interobserver reliability of the measurements were assessed.

Results: For the position of the humeral head with respect to the glenoid in both normal and abnormal conditions, the Pearson correlation coefficient for intraobserver reproducibility was 0.91 and the intraclass correlation coefficient for interobserver reliability was 0.875. For estimating the percentage of the humeral head posterior to the axis of the scapula, the Pearson correlation was 0.85 and the intraclass correlation coefficient was 0.77.

Conclusions: Ultrasonographic examination of the shoulder in infants to assess for the position of the humeral head with respect to the scapula showed high intraobserver reproducibility and interobserver reliability. It is recommended as a reliable technique for evaluating shoulder position in infants with neonatal brachial plexus palsy.