Abstract

The interobserver variability and the ability of the Lauge-Hansen, A.O. and Broos-Bisschop classification systems to encompass all the ankle fracture patterns were investigated in a study of the radiographs of 293 patients with a total of 294 malleolar fractures. Three different orthopaedic surgeons independently evaluated the sets of ankle radiographs. The examiners classified the ankle fractures using the Lauge-Hansen, A.O. and Broos-Bisschop systems. The overall percentage of unclassified fracture patterns was 0.7% with the Broos-Bisschop system, 10% with the Lauge-Hansen system and 8.7% with the A.O. system. The concordance rate using Kappa coefficient ranged from 0.327 to 0.408 for the Broos-Bisschop system, from 0.174 to 0.476 for the Lauge-Hansen system and from 0.397 to 0.483 for the A.O. system. These results show that these three classification systems have in common a considerable interobserver variability deficiency which restricts their validity in selection of treatment options, prognosis and comparison between different materials.