Universal or Selective Hip Ultrasound in Newborns
A dislocated infant hip that is missed can need open surgery in childhood and a hip replacement in young adulthood, while a normal hip that is braced is exposed to avascular necrosis for no gain.
Key terminology
- Developmental dysplasia of the hip (DDH)
- The hip socket is too shallow or the femoral head sits loose, ranging from mild immaturity to full dislocation. It is the target of every newborn hip screen.
- Barlow maneuver
- This exam adducts the flexed hip and pushes it backward. A hip that slides out of the socket is dislocatable.
- Ortolani maneuver
- This exam abducts the flexed hip while lifting the greater trochanter. A palpable clunk means a dislocated hip has reduced.
- Acetabulum
- This cup on the side of the pelvis holds the femoral head. In DDH it is shallow and covers too little of the head.
- Graf method
- This ultrasound system measures the bony acetabular roof on a coronal image of the infant hip. Type I is mature and type IIa is immature but often normalizes.
- Selective ultrasound screening
- This strategy images only infants with an abnormal exam or a risk factor. Universal screening images every newborn instead.
- Pavlik harness
- This soft brace holds the infant hips flexed and abducted so the femoral head stays in the socket. It is the usual first treatment in young infants.
Takeaways
- A normal newborn hip exam does not rule out developmental dysplasia of the hip. Pooled across 8 studies (44,827 hips), the Barlow maneuver and Ortolani maneuver had a sensitivity of 46% and a specificity of 99.1% against ultrasound (positive LR 52). A positive test nearly confirms instability, and a negative test misses about half of dislocated hips. Limited abduction and a hip click had no clear diagnostic value in infants 3 months or younger. (Singh et al, JAMA 2024;331:1576-1585, PMID 38619828).
- Selective ultrasound screening images infants with an abnormal exam or a risk factor. Breech presentation is the strongest risk factor (RR 3.75), followed by female sex (RR 2.54), left hip (RR 1.54), first born (RR 1.44) and family history (RR 1.39). Norwegian trialists recommend selective ultrasound for these infants when clinical screening is of high quality. (Ortiz-Neira et al, Eur J Radiol 2012;81:e344-e351, PMID 22119556; Holen et al, J Bone Joint Surg Br 2002;84:886-890, PMID 12211684).
- Ultrasound is the imaging test for the young infant hip because the femoral head is still cartilage. A plain film cannot show an unossified femoral head, so ultrasound is the standard under 6 months of age. The Graf method grades the hip by how deep the bony acetabulum is on a coronal image. In 28,092 Austrian newborns screened this way, 90.2% of hips were normal at birth. (Ömeroğlu et al, Int Orthop 2024;48:1373-1380, PMID 38150007; Biedermann et al, Bone Joint J 2018;100-B:1399-1404, PMID 30295526).
- Universal ultrasound treats more infants without clearly lowering late diagnoses. In a Norwegian RCT of 11,925 newborns, universal ultrasound raised the treatment rate to 3.4% vs 2.0% (selective) and 1.8% (exam only), and late dislocation was 0.3, 0.7 and 1.3 per 1000 (P = .11). A second RCT of 15,529 newborns found late dysplasia in 0.13 vs 0.65 per 1000 (p = 0.22). (Rosendahl et al, Pediatrics 1994;94:47-52, PMID 8008537; Holen et al, J Bone Joint Surg Br 2002;84:886-890, PMID 12211684).
- Most immature hips on a newborn ultrasound mature without any treatment. Immature (Graf IIa) hips made up 8.9% of Austrian newborn hips and normalized at a median of 6 weeks. In a German series, 89% matured on their own. Delayed ultrasound with targeted splinting reduces treatment without raising late diagnosis. (Biedermann et al, Bone Joint J 2018;100-B:1399-1404, PMID 30295526; Falliner et al, J Pediatr Orthop B 1999;8:112-117, PMID 10218172; Shorter et al, Cochrane Database Syst Rev 2011;(9):CD004595, PMID 21901691).
- The screening trials are too small to detect a difference in late dislocation. Cochrane found neither ultrasound strategy improved late diagnosis or surgery and called the trials underpowered. Across 76 observational studies, late detection was 0.2 (universal), 0.6 (selective) and 0.5 (exam only) per 1000, with overlapping CIs. Settling the question by RCT would take about 61,000 newborns per arm. (Shorter et al, Cochrane Database Syst Rev 2011;(9):CD004595, PMID 21901691; Kuitunen et al, JAMA Netw Open 2022;5:e2227638, PMID 35980635).
- Guidelines split by region on whether every newborn gets an ultrasound. Austria has scanned every newborn since 1992 and reports 46% less pelvic surgery (retrospective, no concurrent control). An Eastern European and Middle Eastern expert panel calls universal screening at 3 to 6 weeks necessary. The AAP calls the evidence controversial, and Norwegian trialists favor selective screening. (Thallinger et al, J Child Orthop 2014;8:3-10, PMID 24488847; Ömeroğlu et al, Int Orthop 2024;48:1373-1380, PMID 38150007; Shaw & Segal, Pediatrics 2016;138(6), PMID 27940740).
- Overtreatment and late diagnosis both harm the patient. The Pavlik harness remains the first-line treatment, and AVN of the femoral head is its most serious complication (reported rates 0% to 28%). On the other side, hip dysplasia caused 26% of hip replacements in Norwegians aged 40 or younger, usually diagnosed in late childhood. (Gulati et al, World J Orthop 2013;4:32-41, PMID 23610749; Engesæter et al, Acta Orthop 2011;82:149-154, PMID 21434808).