A new pilon fracture classification system based on CT scan data was recently published, showing almost perfect interobserver and intraobserver agreement among the authors who developed it. However, an independent assessment has not been done.
To do an independent agreement evaluation of the new pilon fracture classification system with physicians with different levels of expertise in the management of pilon fractures.
Seventy-one cases of acute pilon fracture were retrospectively collected. Fractures were classified by six evaluators (three foot and ankle surgeons and three orthopaedic surgery residents) using CT scans according to the morphological grading of the new pilon fracture classification system developed by Leonetti et al. Cases were presented to the same evaluators in a random sequence after a 6-week interval to determine intraobserver agreement. The kappa coefficient (κ) was used to determine agreement among evaluators.
The interobserver agreement was substantial regarding the main fracture type (I, II, III, or IV), with an overall κ value of 0.69 (0.65 to 0.72). When including the II and III subtypes, the overall agreement was still substantial, with a κ value of 0.61 (95% confidence interval: 0.58 to 0.64). The intraobserver agreement was substantial when considering the main fracture categories (I, II, III, or IV), with a κ value of 0.78 (confidence interval: 0.72 to 0.84), and full agreement at the type level was observed in 76% (324/426) of evaluations. There was no notable difference between the foot and ankle surgeons and orthopaedic surgery residents in the interobserver and intraobserver agreement.
The new classification system demonstrated substantial interobserver and intraobserver agreement between evaluators with different levels of expertise in the management of pilon fractures. Prospective studies should be done to evaluate its prognostic value and utility in clinical practice.