Inter-observer and intra-observer reliability in the radiographic diagnosis of avascular necrosis of the femoral head following reconstructive hip surgery in children with cerebral palsy

Abstract

PurposeThe incidence of avascular necrosis (AVN) following reconstructive hip surgery in cerebral palsy (CP) ranges from 0 to 69 % in the current literature. The purpose of this study was to determine the inter- and intra-observer reliability of radiographically diagnosing AVN in children with CP after hip surgery.MethodsA retrospective review of 65 children with CP who had reconstructive hip surgery between 2009 and 2012 at BC Children’s Hospital was completed. Anterior–posterior and lateral radiographs were presented to four pediatric orthopaedic surgeons over two rounds. Surgeons were asked to review the set of unidentified radiographs and comment ‘yes’ or ‘no’ for the presence of AVN. Two weeks later the same set of radiographs was sent in a different order and the surgeons were again asked to comment on AVN. Inter- and intra-observer reliability was determined using kappa statistics.ResultsThe intra-observer reliability ranged from 0.65 to 0.88 with an average score of 0.76. Inter-observer reliability showed greater variability, ranging from 0.41 to 0.77 with an average score of 0.56 across all surgeons.ConclusionsAlthough the intra-rater reliability produced a strength of “good” and the inter-rater reliability a strength of “moderate” agreement, the variability within these scores is clinically important as it demonstrates the difficulty in identifying AVN. This may explain the variability in AVN that is reported in the literature. The need for further education and research in the diagnosis of AVN in children with CP who have undergone reconstructive hip surgery is clinically necessary.

Osteonecrosis of the Femoral Head: A Multidisciplinary Approach in Diagnostic Accuracy

“…Normal plain radiographs and physical exams can be falsely reassuring and delay appropriate referral [ 11 , 26 ]. In this context, several studies have found suboptimal diagnostic accuracy, reproducibility, and reliability [ 13 , 14 , 25 , 27 , 28 , 29 ]. The aim of this paper is to specifically clarify the association and accuracy of ONFH diagnosis between orthopedic surgeons, radiologists, and pathologists, considering imaging modes, microscopic and macroscopic bone features, and disease staging.…”

“…Twenty-one of the twenty-four diagnoses agreed (interobserver kappa reliability coefficient of 0.58; percentage agreement of 85.71%), expressing moderate agreement. Literature regarding ONFH interobserver reliability has focused mainly on imaging staging correlation, with average kappa values ranging from 0.39 to 0.56, thus evidencing a poor interobserver reliability, especially among the intermediate stages [ 25 , 29 , 34 , 35 ]. In the present work, interobserver diagnosis reproducibility did not show a statistically significant difference when considering the different Ficat and Arlet stages.…”

“…Reported evaluations of intra-observer diagnosis variation in ONFH, based mainly in disease imaging staging, have evidenced fair reproducibility, with mean kappa values ranging from 0.43 to 0.88 [ 25 , 29 , 34 , 35 ]. Regarding histology, we are not aware of any studies that have looked at intra- or inter-observer variability in the pathological diagnosis of avascular necrosis of the femoral head.…”

“…These samples were preserved and subsequently used for the anatomopathological analyses. For preservation, the cylindrical biopsies were immediately placed in 10% neutral buffered formalin at room temperature for 24 h and subsequently decalcified in 10% ethylenediaminetetraacetic acid at room temperature following standard procedures [ 24 , 25 ]. Decalcified specimens were embedded in paraffin and paraffin sections (4 mm thick) were cut using conventional methods.…”

The incidence of avascular necrosis in children with cerebral palsy after hip containment surgery

“…This is in range with the currently published data. [4][5][6][7][8]12 Thirty-three children developed unilateral AVN and 5 children developed bilateral AVN. All 38 hips that developed AVN had femoral procedures While more children with GMFCS level V required a hip containment procedure, the rate of AVN development was similar for GMFCS levels III, IV, and V. While the unstandardized beta value was low, higher Reimers migration percentage was found to be a predictable risk factor in developing AVN, with children who developed AVN having a significantly greater (61.3 vs 51.1) Reimers migration percentage.…”