Specimen preparation
Fresh-frozen human cadavers of this study were provided by Department of Anatomy in Chiang Mai University. Six shoulders from six fresh-frozen cadavers, comprising five male specimens and one female specimen, were used in this study, with a mean age of 68.7 (range, 51–87) years. The specimens were prepared by thawing overnight at room temperature one day before the experiment. The specimens were subsequently sectioned above the first cervical vertebrae, below the sternum, and on the right side of the sternum and below the 11th thoracic vertebrae. The left upper extremity was sectioned at the midshaft of the humerus, and a hole was drilled with a Kirschner wire with a diameter of 3 mm in the humerus near the attachment of the deltoid to pass a silk thread with a diameter of 2 mm that could be used to apply stress to the upper limb; tensile strength was approximately 400 N. All specimens were kept moist by spraying with normal saline during the experiment. Standard anteroposterior (AP) radiographs were obtained for each specimen, and no specimens had osteoarthritis at the sternoclavicular, AC, and glenohumeral joints.
The specimens were firmly fixed on a customized wooden jig with external fixators (Orthofix®; Japan Medicalnext Co., Ltd., Osaka, Japan). Five fully threaded stainless rods with a diameter of 6 mm were inserted into the 2nd and lower cervical vertebraes, upper and lower thoracic vertebraes, and lower sternum. The rods were connected to external fixators (Fig. 1). The displacement of the distal end of the clavicle relative to the acromion was measured using an electromagnetic tracking device (trakSTAR™; Ascension Technology Corporation, Shelburne, VT, USA). Sensors were inserted into the distal end of the clavicle and acromion. The proper location of the inserted sensors was verified with AP radiographs (BV Pulsera; Philips, Best, Netherlands) (Fig. 2).
Sectioning the distal clavicle stabilizers
We resected the AC, trapezoid and conoid ligaments sequentially, and simulated AC joint dislocation model (Fig. 3). Sectioning stages were defined as follows. Stage 0: the AC and CC ligaments and the AC joint capsule were intact; stage 1, the AC ligament, AC joint capsule, and disc were sectioned; stage 2, trapezoid ligament were sectioned; and stage 3, conoid ligaments were sectioned. The trapezius and deltoid muscle were incised parallel to the AC joint when the AC ligament and joint disc were removed. When resecting the trapezoid and conoid ligaments, the deltoid muscle was incised in the direction of the muscle fibers, and the ligaments were resected after clearly viewing them. Ligaments were sectioned according to previous biomechanical studies [16, 17, 19].
Loading testing and data acquisition
Stress was added to the AC joint by pulling the cord that was passed through the humerus. An examiner elevated the upper limb to 90° in the sagittal plane and adducted the upper limb manually until the maximum adduction angle was acquired in the horizontal plane, the cross-body adduction radiography was also performed in this position. Using this imaging technique, we assessed the degree to which the clavicle overlapped the acromion because of anteromedial scapula translation. Clavicle overriding on the cross-body adduction view was defined as the superior or lateral displacement of the inferior edge of the clavicle in the AC joint compared to the that of superior edge of the acromion in the AC joint (Fig. 4).
We defined the direction parallel to the AC joint is the X-axis, perpendicular to the AC joint is the Y-axis, and perpendicular to the ground is the Z-axis. The magnitude of the displacement was measured in each direction. In the horizontal adduction position, the direction of the AC joint changed, as demonstrated by the electromagnetic tracking device; yet, we could not measure the horizontal translation of the AC joint. We predicted posterior instability by determining whether the distal clavicle overrode the acromion [20]. Displacement was measured in the Z-axis by calculating the difference between the values of both sensors in the acromion and the distal end of the clavicle. Values in stage 0 were used as control values.
Statistical analysis
Displacement magnitudes between each stage were compared using a one-way analysis of variance, and post-hoc tests were performed using the Tukey–Kramer method. P-values of < 0.05 were considered statistically significant. Statistical analysis was performed using SPSS for Windows version 22.0 (IBM Corp., Armonk, NY, USA).