Proximal humeral fractures in children - controversies in decision making

Proximal humeral fractures in children are rare and usually treated non-operatively, especially in children younger than ten. The decision between operative and non-operative treatment is mostly based on age and fracture angulation. In the current literature, diverging recommendations regarding fracture angulation that is still tolerable for non-operative treatment can be found. Besides, there is no consensus on how fracture angulation should be determined. This study aimed to determine whether leading experts in pediatric trauma surgery in Germany showed agreement concerning the measurement of fracture angulation, deciding between operative and non-operative treatment, and choosing a treatment modality.

Twenty radiographs showing a proximal humeral fracture and the patient’s age were assessed twice by twenty-two senior members of the “Section of Pediatric Traumatology of the German Association for Trauma Surgery”. Experts determined the fracture angulation and chose between several operative and non-operative treatment modalities. The mean of individual standard deviations was calculated to estimate the accuracy of single measurements for fracture angulation. Besides Intra-Class Correlation and Fleiss’ Kappa coefficients were determined.

For fracture angulation, experts showed moderate (ICC = 0.60) interobserver and excellent (ICC = 0.90) intraobserver agreement. For the treatment decision, there was fair (Kappa = 0.38) interobserver and substantial (Kappa = 0.77) intraobserver agreement. Finally, experts preferred ESIN over K-wires for operative and a Gilchrist over a Cuff/Collar for non-operative treatment.

Firstly, there is a need for consensus among experts on how fracture angulation in PHFs in children should be reliably determined. Our data indicate that choosing one method everybody agrees to use could be more helpful than using the most sophisticated. However, the overall importance of fracture angulation should also be critically discussed. Finally, experts should agree on treatment algorithms that could translate into guidelines to standardize the care and perform reliable outcome studies.

III.