Limitations of the radiocapitellar line for assessment of pediatric elbow radiographs

Abstract

Background: The radiocapitellar line (RCL) is recommended for evaluating radiocapitellar alignment in skeletally immature elbows, yet its parameters have not been clearly defined. This study systematically assesses the RCL relationship in normal elbows, investigating the impacts of radiographic view, choice of anatomic landmarks, patient age, forearm position, and observer bias on the manner in which the RCL intersects the capitellum.

Methods: On radiographs of 20 normal elbows (age range, 1 to 8 y), 3 pediatric orthopaedic surgeons, blinded to clinical history, drew lines (RCLs) on anteroposterior and lateral projections, along the radial shaft and neck, and with and without the capitellum visible. Line placement was repeated 2 weeks later. The relationship of each RCL to the capitellum was assessed continuously using the perpendicular distance to the center of the capitellum, normalized to capitellar width [line-capitellar distance (LCD)], and categorically as passing through the middle third, outer two-thirds, or outside the capitellum.

Results: Of the 480 RCLs drawn, 23 (5%) missed the capitellum and 224 (47%) missed the middle third. More radial neck than shaft lines intersected the middle third on both anteroposterior and lateral views (P < 0.05, Fisher exact test), with the lowest LCD values for neck lines on the lateral view (P < 0.05, analysis of variance (ANOVA)). More RCLs intersected the middle third when the capitellum was visible than when it was obscured (P = 0.03, Fisher exact test), suggesting an effect of observer bias. Patient age correlated inversely with LCD (P < 0.001). The angle between the neck and shaft lines correlated positively with LCD (P < 0.001), suggesting an impact of forearm rotation position. Intraobserver and interobserver reliability was moderate-to-substantial (κ = 0.40-0.75).

Conclusions: The RCL best defines normal radiocapitellar alignment when the line is drawn along the radial neck on the lateral view, although this relationship is affected by bias, patient age, and forearm rotation position. The RCL does not reliably intersect the middle third of the capitellum, arguing against its sufficiency for assessing precise radiocapitellar alignment.

Level of evidence: Diagnostic Level 3.