Does using the medial or lateral humeral line improve reliability of Baumann angle measurement on plain x-ray? The effect of humeral length visualized on the x-ray

• Published in: Journal of pediatric orthopaedics Vol. 32; no. 4; p. 373
• Main Authors: Krengel, 3rd, Walter F, Wiater, Brett P, Pace, James Lee, Jinguji, Thomas M, Bompadre, Viviana, Stults, Janna K, Schmale, Gregory A
• Format: Journal Article
• Language: English
• Published: United States 01.06.2012
• Subjects: Analysis of Variance; Child; Child, Preschool; Female; Humans; Humeral Fractures - diagnostic imaging; Humerus - diagnostic imaging; Infant; Male; Observer Variation; Radiography; Reproducibility of Results; Sex Factors
• ISSN: 1539-2570
• DOI: 10.1097/BPO.0b013e318251969f

Baumann angle (BA) is a common measure of coronal plane alignment of the distal humerus. We hypothesize that the reliability of measuring BA would be improved by using the medial and lateral cortical margins of the humerus seen on plain x-ray, rather than the estimated central humeral line, which is the "standard" technique. Further, we analyze whether the amount of humerus visible on the film improves the reliability of the method. A total of 71 anteroposterior elbow digital radiographs from patients aged 0 to 12 were measured 3 times by 5 qualified observers. Each digital measurement included (1) BA using the estimated central humeral line; (2) BA using the medial humeral line (BA-MHL); and (3) BA using the lateral humeral line (BA-LHL). Inadequate radiographs or those showing any indication of current or previous fracture were excluded. Intraobserver reliability was estimated for each rater using a 1-way analysis of variance model and interobserver reliability of each set of measurements was estimated using a 2-way analysis of variance. The mean and SD for the BA, BA-MHL, and BA-LHL in females were 70.0 (6.73), 68.0 (6.84), and 72.3 (7.93), respectively, and for males 73.0 (5.22), 70.0 (5.56), and 76.0 (6.18), respectively. Intraobserver reliability (intraclass correlation coefficient) for BA, BA-MHL, and BA-LHL averaged 0.85, 0.92, and 0.90, respectively. Average interobserver reliability (intraclass correlation coefficient) for BA, BA-MHL, and BA-LHL were 0.79, 0.71, and 0.76, respectively. Intraobserver and interobserver reliability of BA and BA-LHL were significantly improved when at least 7 cm of humerus was visible on the x-ray, as compared with <7 cm visible. Intrarater reliability is better for both BA-LHL and BA-MHL than for standard BA (using the estimated central humeral axis). Interrater reliability was best using standard BA. Reliability of all methods is improved when >7 cm of the distal humerus is visible on the radiograph. In addition, at least 7 cm of the distal humerus on anteroposterior radiographs improves reliability of measuring BA. Diagnostic study-Level II.