Shoulder instability and impingement are common in tennis players. During tennis, several impingements could occur: subcoracoid and anterosuperior impingements at the follow-through phase of forehand and the backhand preparation phase; subacromial and postero-superior impingements at the cocking phase of serve. The precise causes for these impingements remain unclear, but it is believed that repetitive contact, glenohumeral instability may play a role. Impingement and glenohumeral instability at critical tennis positions have never been dynamically evaluated in-vivo. The purpose of this study was to develop a patient-specific measurement technique based on motion capture and MRI to accurately determine glenohumeral kinematics (rotations and translations). The second objective was to evaluate impingements and stability in tennis. Shoulder MR arthrography and motion capture were performed in 10 tennis players. Motion data were recorded during tennis movements. Glenohumeral kinematics was computed from the markers trajectories using a global optimisation algorithm with loose constraints on joint translations (accuracy: translational error ≈3mm, rotational error <4°). The translations patterns computed with the model were in good agreement with previous works. The resulting computed motions were applied to the subject's shoulder 3D bony models reconstructed from MRI data. While simulating the shoulder joint, minimum humero-acromial, humero-coracoid and humero-glenoid distances were measured at critical tennis positions. Given the thickness of the potential impinged tissues, impingement was considered when the computed distance was <5 mm (<6 mm for the humero-acromial distance). During serve, glenohumeral stability was assessed at the cocking, deceleration and finish phases. Glenohumeral translation was defined as the anterior-posterior and superior-inferior motion of the humeral head centre relative to a glenoid coordinate system. Subluxation was defined as the ratio (in %) between the humeral head centre translation and the radius of the width (antero-posterior subluxation) or height (supero-inferior subluxation) of the glenoid surface. Instability was thus considered when the subluxation was >50%, corresponding to a loss of congruence superior to half the radius of the width (or height) of the glenoid. No subcoracoid impingement was detected. Antero-superior impingements were observed in two subjects (29%) during forehand. Anterior and lateral subacromial impingements occurred during the cocking phase of serve in three (29%) and four subjects (42%), respectively. Postero-superior impingements during the cocking phase of serve were the most frequent (7 subjects, 75%). In this position, glenohumeral translation was anterior (mean: 34%) and superior (mean: 13%). During the deceleration phase, anterior and superior translation varied between 8–57% and between 5–34%, respectively. During the finish phase, anterior translation was slightly more intense (mean: 44%), while superior translation remained low (mean: 1%). MRI revealed eleven rotator cuff lesions in five subjects, and six labral lesions in five subjects. Postero-superior impingement was frequent when serving. No instability could be noted. Tennis players presented frequent radiographic signs of structural lesions which seem to be mainly related to postero-superior impingement due to repetitive abnormal motion contacts. Our findings are consistent with this hypothesis. To our knowledge, this is the first study demonstrating that a dynamic and precise motion analysis of the shoulder is feasible using an external measurement system, such as motion capture.
