To the best of our knowledge, there is no study comparing the inter and intraobserver reliability of current classifications for postseptic hip sequelae in children. The current study aims to assess the interobserver and intraobserver reliability of four current classifications and identify hips that could not be classified in each classification system. The hip radiographs of 148 consecutive children with sequelae of sepsis of the hip from 2 centers were assessed after a minimum of 2 years of follow-up after sepsis. All hips (affected and normal sides) were classified according to the 4 original descriptions of the authors of the respective classifications. If a hip did not fall into any subtype of the classification, the rater was asked to mark it as nonclassifiable and state the reason for being unable to classify the hip in the respective classification. The intraclass correlation coefficient was computed to assess the reproducibility of each classification. Interrater reliability and intrarater reliability were moderate (0.57 to 0.72) while including all hips. The reliability was poor (0.35 to 0.49) in all 4 classifications, with an evaluation of only affected 180 hips. A few sequelae of infection, including caput valgus (n = 7), acetabular dysplasia (4), joint space narrowing (2), and bony ankylosis (1), were not included in any of the 4 current existing classification systems. The reliability of all current classifications of sequelae of septic arthritis of the hip is moderate. A proportion of sequelae do not find a place in all current classifications. Level III.
Read full abstract