Background: There is no evidence that shoulder stabilization effectively corrects the glenohumeral translation in unstable shoulders, explaining residual apprehension in certain patients. The purpose of this study was to analyze the effect of surgical stabilization on glenohumeral translation.

Methods: Anteroposterior and superoinferior translations were assessed in patients, before and after shoulder stabilization, through a dedicated patient-specific measurement technique based on optical motion capture and computed tomography.

Introduction The anterior stabilizing system for the glenohumeral joint is quite complex and may be altered by variable factors: anatomy, anatomic variants, overload and trauma. The latter mechanisms affects 1.7% of the general population, making glenohumeral instability the most frequent type of all joint instabilities.1 Shoulder apprehension is defined as anxiety and resistance in patients with a history of anterior glenohumeral instability. After an open or arthroscopic stabilization, 3% to 51% of patients will keep apprehension or will avoid any shoulder movement because of fear of dislocation.2,3 This can lead to increased morbidity for patients: increased pain, decreased activity level, prolonged absence from work and sports, and a general decrease in quality of life.4,5 Currently, the origin of persistent apprehension is unknown. Although a bony defect has been recognized as a major cause of residual instability,6 some patients remain apprehensive without any proven recurrence of dislocation and a clinically stable shoulder. Theoretically, such apprehension after glenohumeral stabilization could be related to (1) central nervous system sequelae subsequent to a learned negative stimulus,7,8, (2) peripheral neurological lesion consecutively to dislocation affecting proprioception,9 or (3) persistent mechanical instability consisting in micro-movements (i.e., postoperative form of unstable painful shoulder as described by Patte et al.10).

Using a dedicated and non-invasive patient-specific measurement technique11 based on optical motion capture and computed tomography, the purpose of this article was to describe the glenohumeral translation in patients suffering from anteroinferior instability, to analyze the effect of glenohumeral stabilization on this translation, and consequently determine if shoulder stabilization effectively stabilizes shoulders or solely prevents further dislocations. The hypothesis was that shoulder stabilizations only partially correct the glenohumeral translation in unstable shoulders explaining residual apprehension in certain patients.

Methods Patient Selection Between October 2014 and January 2015, a consecutive series of patients evaluated in a shoulder clinic who had a primary anteroinferior shoulder stabilization performed by the senior author were considered potentially eligible for inclusion in this prospective study. Institutional ethics committee approval was obtained before the study began (AMG 12-18), and the subjects signed a written informed consent form before participation.

Operative Technique All operations were performed in the usual semi-beach chair position under general anaesthesia with an interscalenic block or catheter. Open Latarjet was performed as the standard and well-described Latarjet-Patte procedure with subscapularis split and triple locking mechanism.14 The graft was intra-articular in every case, the capsule was systematically reattached to glenoid according to Favard's modification,15 and a capsular shift was added. Arthroscopic Latarjet was carried out in one case according to a modified Lafosse technique.16 In the latter treatment option, no reattachment of the capsule was realized. In both arthroscopic and open techniques, the patients were postoperatively protected with a sling for ten days and were able to immediately start full active range of motion. Return to low-risk sports was allowed at six weeks, and high-risk (throwing and collision) sports at three months. The arthroscopic Bankart repair consisted in a mobilization of the anteroinferior capsule and the labrum with an arthroscopic elevator. The glenoid rim and neck were then prepared with a mechanical shaver device. Two loaded anchors were inserted at the 5 and 3 o'clock position, and sutures were shuttled across the inferior glenohumeral ligament and labrum, starting at the inferior position and progressing in a superior direction. Postoperatively, the arm was protected during four weeks. Return to low-risk sports was allowed at ten weeks, and high-risk (throwing and collision) sports at 4.5 months.

Radiographic Evaluation and Motion Capture All volunteers underwent a computed tomography of both arms and shoulders. The computed tomography examinations were conducted with a LightSpeed (LS) VCT 64 rows (General Electric Healthcare, Milwaukee WI, USA). Images were acquired at 0.63 mm slice resolution. Based on the computed tomography images, patient-specific 3D models of the shoulder bones (humerus, scapula, clavicle and sternum) were reconstructed for each patient using Mimics software (Materialize NV, Leuven, Belgium).

Kinematic data was recorded using a Vicon MX T-Series motion capture system (Vicon, Oxford Metrics, UK) consisting of twenty-four cameras (24 × T40S) sampling at 120 Hz. The patients were equipped with a dedicated shoulder markers protocol,11 including sixty-nine spherical retroreflective markers placed directly onto the skin using double sided adhesive tape. The setup included four markers (Ø 14 mm) on the thorax (sternal notch, xyphoid process, C7 and T8 vertebra), four markers (Ø 6.5 mm) on the clavicle, four markers (Ø 14 mm) on the upper arm - two placed on the lateral and medial epicondyles and two as far as possible from the deltoid - and fifty-seven markers on the scapula (1x Ø 14 mm on the acromion and a 7x8 grid of Ø 6.5 mm). Finally, additional markers were distributed over the body (non-dominant arm and legs) to provide a global visualization of the motion.

Patients participated in two motion capture sessions: a first session before surgery and a second one year after shoulder stabilization. During each session, they were asked to perform the following motor tasks (three trials each): (1) internal-external rotation of the arm with 90° abduction and the elbow flexed 90°, (2) internal-external rotation of the arm with elbow at side, (3) flexion of the arm from neutral to maximum flexion, and (4) empty-can abduction from neutral to maximum abduction in the scapular plane. Both shoulders (ipsilateral and contralateral) were measured during the first session, whereas only the operated shoulder was assessed after surgery (second session). The same investigators attached all markers and performed all measurements.

Kinematic Analysis Shoulder kinematics were computed from the recorded markers' trajectories using a validated biomechanical model which accounted for skin motion artifacts.11,21 The model was based on a patient-specific kinematic chain using the shoulder 3D models reconstructed from computed tomography data and a global optimization algorithm with loose constraints on joint translations (accuracy: translational error <3 mm, rotational error <4°).

Glenohumeral range of motion was quantified for flexion, abduction and internal-external rotations at the maximal range of motion and expressed in clinical terms.22 This was achieved by calculating the relative orientation between two local coordinates systems, one for the scapula and one for the humerus, based on the definitions suggested by the International Society of Biomechanics.23 The local systems were created using anatomical landmarks identified on the patient's bony 3D models. The glenohumeral joint center was calculated based on a sphere fitting method.24 To facilitate clinical comprehension and comparison, motion of the humerus with respect to the thorax was also calculated. This was obtained with the same method but using the thorax and humerus coordinate systems.

Glenohumeral translations were assessed at maximal range of motion during all tested movements. Glenohumeral translation was defined as anterior-posterior and superior-inferior motion of the humeral head center relative to the glenoid coordinate system.25 This coordinate system was determined by an anterior-posterior X-axis and a superior-inferior Y-axis with origin placed at the intersection of the anteroposterior aspects and superoinferior aspects of the glenoid rim. Subluxation was defined as the ratio (in %) between the translation of the humeral head center and the radius of width (anteroposterior subluxation) or height (superoinferior subluxation) of the glenoid surface. Instability was defined as subluxation > 50%.26

Statistical Analysis Glenohumeral range of motion, humerus motion relative to the thorax, as well as glenohumeral translations were computed at maximal range of motion for all patients and for all movements recorded during the two motion capture sessions (before and after surgery). Paired Student's t-tests were used to determine if the kinematic data differed between the contralateral and ipsilateral pre- and postoperative arms, and between the pre- and postoperative pain scores. A significance level was chosen at p < 0.05. Descriptive statistics are presented as mean and standard deviations. The statistical software package R, v3.1.2 Portable (Free Software Foundation Inc, Vienna, Austria) was employed.

Inclusion Criteria:

• Anteroinferior shoulder stabilization

Exclusion Criteria:

• Incomplete documentation
• Follow-up of less than twelve months
• History of bilateral instability
• Previous shoulder surgery
• Contraindications for computed tomography
• Non-traumatic onset
• Hyperlaxity. The latter was defined as more than 85° of external rotation at the elbow against the waist{Coudane, 2000 #3124} or hyperabduction by more than 105°.{Gagey, 2001 #1915}