Influence of x-ray direction on measuring shortening of the fractured clavicle

Abstract

Midshaft clavicle fractures are often associated with a certain degree of shortening. There is great variety in the imaging techniques and methods to quantify this shortening. This study aims to quantify the difference in measurements of shortening and length of fracture elements between 5 views of the fractured clavicle. Furthermore, the interobserver and intraobserver agreement between these views using a standardized method is evaluated. Digitally reconstructed radiographs were created for 40 computed tomography datasets in the anteroposterior (AP), 15° and 30° craniocaudal, and 15° and 30° caudocranial views. A standardized method for measuring the length of fracture elements and the amount of shortening was used. Interobserver and intraobserver agreement for each of the 5 views was calculated. The interobserver and intraobserver agreement was excellent for all 5 views, with all intraclass correlation coefficient values greater than 0.75. The measured differences in relative and absolute shortening between views were statistically significant between the 30° caudocranial view and all other views. The increase in median shortening measured between the commonly used 30° caudocranial view (2.7 mm) and the AP view (8.5 mm) was 5.8 mm (P < .001). The relative median shortening between these views increased by 3.5% (P < .001). The length of fracture elements and the amount of shortening in the fractured clavicle can be reliably measured using a standardized method. The increase in absolute and relative shortening when comparing the caudocranial measurements with the AP and craniocaudal measurements may indicate that the AP and craniocaudal views provide a more accurate representation of the degree of shortening.