Supracondylar fractures of the humerus are frequent paediatric injuries. The aims of this study were to evaluate the applicability and reproducibility of the Gartland and Wilkins classification, the Baumann angle (BA) and the Anterior Humeral Line (AHL). This retrospective monocentric observational study was conducted on 217 patients. Four observers assessed the pre-operative radiographs by applying the Gartland and Wilkins classification and the post-operative X-rays by measuring the BA and AHL. The kappa coefficient (K) and the Cohen's kappa were used for the reliability of the Gartland classification; the Intraclass Correlation Coefficient (ICC) for that of the BA. The AHL was evaluated in a double manner by using first the K and the Cohen's kappa and then the ICC. A total of 186 patients were eligible. Inter-observer reliability for the Gartland classification was K = 0.73-0.61 for type III, 0.65-0.61 for type Ia and 0.43-0.26 for type IIb. The Baumann angle mean value in the first data collection was 73.5 ± 6.85 (inter-observer ICC 0.74) and 72.9 ± 6.83 (inter-observer ICC 0.77) for the second data collection; AHL: inter-observer ICC 0.87 for the first evaluation and 0.80 for the second one. Gartland's classification modified by Wilkins has a high degree of reliability. BA and AHL appear reproducible and reliable.
Keywords: Baumann angle; Gartland classification; anterior humeral line; children; supracondylar fractures.