Malunion of the proximal humerus is common and variably tolerated. Classifications developed for proximal humerus malunion (PHM) rely on standard radiographs, which underestimate bone fragment displacement and lack accuracy. The clinical tolerance of PHM is subjective, and revision surgery is not always necessary. The primary objective of this study was to assess the reproducibility and relevance of four CT angle measurements for objectively quantifying the morphological disharmony caused by PHM in a control population then in a population with PHM. The secondary objectives were to identify angle cut-offs and to assess the correlations between angle values and the clinical tolerance of PHM.

Objective criteria for assessing proximal humerus malunion can be identified using CT scans.

Four angles were chosen to quantify proximal humerus disharmony: the angles between the humeral head and the glenoid in the coronal plane (HGCo) and axial plane (HGAx), the angle of tuberosity divergence in the axial plane (TDAx), and the centrum collum diaphyseal angle (CCD). The reproducibility of measurements of the four angles on computed tomography (CT) views was evaluated in a control population and in 46 patients with PHM. To this end, the reproducibility of reference slice selection was determined and intra- and interobserver reproducibility of the angle measurements was then assessed. Patients with PHM were divided into two groups based on clinical tolerance to allow testing for disharmony parameters associated with poor clinical tolerance, which was defined as functional impairment and surgical revision.

Slice selection was found to be reproducible. The Bland-Altman plot indicated that the angle measurements in both the controls and the patients were reproducible within ±2 SDs. Intraclass correlation coefficient values ranged from fair to excellent for all angles in both the controls and the patients. The mean TDAx was higher in the patients than in the controls (72.0° vs. 56.1°, P<0.05) and, within the PHM group, was higher in the subgroup with good vs. poor clinical tolerance (75.8° vs. 69.5°, P<0.05). The CCD angle was greater in the controls than in the patients (129.8° [range, 128.3°–131.3°] vs. 125.9° [range, 122.9°–128.9], respectively) and was significantly greater in the PHM subgroup with good vs poor clinical tolerance (131.4° vs. 122.3°, respectively; P=0.007). The HGCo and HGAx angles were significantly greater in the patients than in the controls (HGCo: 66.6° vs. 52.2°, respectively; HGAx: 17.5° vs. 13.3°, respectively, P=0.55).

The measurement method described here provides a quantitative assessment of postfracture disharmony based on four angles, the HGCo, HGAx, and TDAx. Measurement of these four angles on CT images was found to have good intra- and interobserver reproducibility. The angle values were significantly greater in the patients with PHM than in the controls. Within the patient group, the subgroup with poor clinical tolerance had smaller values of the TDAx, CCD, and HGAx angles and a greater value of the HGCo angle.

IV, retrospective observational study.