The classification of fractures is necessary to ensure a reliable means of communication for clinical interaction, research, and education. Distal humeral coronal plane shear fractures (CSF) are very rare in the growing-age population, despite that multiple classification schemes exist. The aim of this study was to assess the inter- and intra-rater reliability of the 4 most commonly used CSF classifications and their role in guiding treatment decisions in everyday clinical practice. 51 patients with CSF were retrospectively analyzed in a multicenter study. Of these, X-rays in two planes of 20 randomly selected CSF were given to 8 independent raters for classification according to Dubberley, Bryan & Morgan, Murthy, and AO. Cohens Kappa statistic was used to assess inter-rater agreement and intra-rater consistency. Intra-class correlation coefficient (ICC) estimates and 95% confident intervals (CI) were calculated. Reliability values were classified as excellent (ICC ≥0.8), substantial (ICC 0.61-0.80), moderate (ICC 0.41-0.6), fair (ICC 0.2-0.4), slight (ICC 0.00-0.2), and poor (ICC <0.00). All patients received x-rays in 2 planes initially. Cross-sectional imaging was added in 76.5% of cases (CT 58,8%, MRI 11,7%, both 5,9%). Inter-rater reliabilities were classified as fair for Dubberley (ICC 0,354; 95% CI 0,198; 0,573) and Bryan & Morgan (ICC 0,357; 95% CI 0,200; 0,576), slight for AO (ICC 0,226; 95% CI 0,100; 0,434), and poor for Murthy (ICC -0,012; 95% CI -0,063; 0,102). The Dubberley subtype showed slight agreement at lower ICC values (ICC 0,024; 95% CI -0,041; 0,161). Intra-rater agreement was moderate to substantial for most of the 8 raters, and did not differ between the four classification systems. The rate of surgical treatment was 49/51 patients. The most commonly used classification schemes for CSF failed to achieve a substantial agreement among the raters. This is probably because the fracture patterns of CSF in adolescence and high adulthood differ significantly. An adaptation of the classification for pediatric and adolescent patients is necessary, but only for scientific purposes. Classification is not necessary for the clinical management of patients, as virtually all patients require surgical treatment.
Read full abstract