199. REPRODUCIBILITY OF HERRING’S CLASSIFICATION USING TWO DIFFERENT TECHNIQUES AND ITS ALTERATION DURING THE EVOLUTION OF THE DISEASE

Abstract

Purpose: The lateral pillar classification for Perthes disease described by Herring in 1992 has gained wide acceptance as a method of predicting outcome and planning treatment. Our purpose was to determine the reproducibility of Herring’s lateral pillar classification using visual estimation and by direct measurement and determine if the Herring’s grading alters as the child passes through the stage of fragmentation in Perthes’ disease.

Method: One hundred AP and frog lateral radiographs of children with unilateral Perthes’ disease in the stage of fragmentation were classified according to the Herrings classification by two investigators utilizing a visual and measurement technique. The change in Herring’s grading with progression of disease was evaluated in 86 patients with sequential radiographs in the stage of fragmentation.

Results: The level of intra-observer agreement by the measurement technique was excellent for both AP and lateral radiographs (Kappa = 0.92 and 0.98) as compared to the visual method for which the agreement was moderate (Kappa = 0.65 and 0.5). The inter-observer reproducibility was moderate by the visual method for both AP and lateral radiographs (Kappa = 0.51 and 0.43). The level of agreement for the measurement method was good for the AP radiographs (Kappa = 0.66) and was only moderate for the frog lateral radiographs (Kappa = 0.53). Of the total 86 cases that had sequential radiographs in the stage of fragmentation, 33 showed change in Herring’s grading. Among these 33 cases, 25 showed a change in the extent of epiphyseal collapse in the AP radiographs alone whereas 8 cases showed a change in lateral radiographs. Upgrading of Herring’s grade from A to B was seen in 11 cases and from B to C in 14 cases as observed in the AP radiographs. The clinical variables and radiological variables did not show any association with progression of Herring’s grade.

Conclusion: The measurement technique of assessing Herring’s classification is much more reliable than the originally described visual method. However, the Herring’s grade changes with the evolution of the disease even during the process of fragmentation and must be used with caution when predicting prognosis.