The effect of the rotator interval on glenohumeral kinematics during abduction

The rotator interval (RI) is a triangularly shaped space bounded by the supraspinatus superiorly, the subscapularis inferiorly and the coracoid process of the scapula medially. It represents a complex interaction of the coracohumeral ligament, the superior glenohumeral (GH) ligament, GH joint capsule, and the supraspinatus and subscapularis tendons. The lateral apex of this triangle is composed of the intertubercular sulcus and transverse ligament and the coracoid process composes its base. The RI also contains the intra-articular portion of the long head of the biceps tendon and is concealed by a fibrous capsule [1]. The RI functionally limits external rotation and resists humeral head translation in the adducted shoulder and posterior translation in the flexed or abducted and externally rotated shoulder [2, 3]. In patients with shoulder instability this triangular area of connective tissue has been implicated in the treatment of various shoulder condition. Since the rotator cuff tendons are not injured, the RI provides an easy site for accessing the GH joint for many common shoulder procedures and is an ideal approach for the open treatment of septic arthritis. Recent interest in less invasive techniques has illustrated how this potential space can be exploited for shoulder arthroplasty [4]. In this way, the RI is believed to provide a safe point of entry to the shoulder with minimal impact on the joint’s kinematics.

However, in patients with shoulder instability, tightening the RI has been shown to decrease external rotation and eliminate the sulcus sign associated with joint laxity. This finding suggests that the RI has a more complex role in normal motion.

Few studies have examined the effect of the RI on shoulder kinematics. Of note, studies have investigated GH and scapulothoracic motion in normal patients and resultant kinematics in patients with shoulder instability or rotator cuff (RC) tears; however, the kinematics associated with RI insult and subsequent repair is uncertain.

Differences were observed in GH translations in the coronal plane (x-axis) between Baseline and Opened conditions at 135° and 150° (both p < 0.05; Fig. 4a, *), and between both closed conditions (Repaired, and Tightened) and Baseline at 60°, 75°, 90°, 105°, 120°, 135°, and 150° (all p < 0.05; Fig. 4a, ^∆).

In the sagittal plane (y-axis), Opened and Baseline were significantly different at 135° and 150° (both p < 0.05; Fig. 4b, *). Moreover, differences in GH translation in the sagittal plane were observed between both closed conditions (Repaired and Tightened) and Baseline at 60°, 75°, 90°, 105°, 120°, 135°, and 150° (all p < 0.05; Fig. 4b, ^∆).

In the transverse plane (z-axis), GH translation was different between Opened and Baseline at 30°, 45°, 60°, 75°, 90°, 105°, 120°, 135°, and 150° (all p < 0.05; Fig. 4c, *). Finally, both closed conditions (Repaired and Tightened) demonstrated significant differences at 135° and 150° when compared to Baseline (both p < 0.05; Fig. 4c, ^∆).

GH AUC in the transverse plane demonstrated differences between Open and Baseline at 120°-150° (p < 0.05; Fig. 5a *). GH AUC in the transverse plane demonstrated differences between both Repaired and Tightened when compared to Baseline (⋄) and Opened (§) at 90°-120°, 120°-150°, and for the entire abduction 30°-150° range (all p < 0.05; Fig. 5a). GH AUC in the sagittal plane revealed differences between Opened and Baseline states at 30°-60°, 120°-150°, and for the entire abduction 30°-150 (p < 0.05; Fig. 5b, *). GH AUC was significantly different in the sagittal plane for Repaired and Tightened when compared to Baseline (⋄) and Opened (§) at 90°-120°, 120°-150°, and for the entire abduction 30°-150° (all p < 0.05; Fig. 5b). GH AUC in the transverse plane was different for Opened versus Baseline at 30°-60°, 60°-90°, 90°-120°, 120°-150°, and for the entire abduction 30°-150° (all p < 0.05; Fig. 5c). AUC in the transverse plane was also different for Repaired and Tightened when compared to Opened (§) at 90°-120°, 120°-150°, and for the entire abduction 30°-150° (all p > 0.05; Fig. 5b). AUC analysis did not show a difference between Repaired and Tightened at 90°-120°, 120°-150° (both p > 0.05; Fig. 5c).

The RI resists inferior and posterior translation in the abducted shoulder, particularly when the arm is externally rotated [3, 11‐17]. This constraint results from the RI’s stabilizing effect on GH motion, preventing displacement of the humeral head while connecting the supraspinatus to the subscapularis [3]. Generalized tissue laxity can negate this effect, resulting in multidirectional instability (MDI). For MDI, arthroscopic RI closure, combined with capsular plication, has been used to increase shoulder stability while reducing the range of motion [18‐20].

However, the role of the RI in GH kinematics is still debated. Provencher et al. found that arthroscopic RI closure added little support to posterior and inferior stability of the shoulder during abduction [21]. Harryman et al. described how a medial-to-lateral open RI imbrication decreased inferior translation in an abducted shoulder and decreased posterior translation in a flexed shoulder [3]. Decreased anterior GH translation occurs consistently following RI closure [21‐23].

At the same time, RI closure can overtighten the shoulder, constraining motion at some arm positions [20]. Though inconsistent, the data gathered from cadaveric RI closure suggest that this tissue has an effect on shoulder kinematics.

Few studies have delineated the effect of opening the RI on GH kinematics. This study sought to characterize how opening the RI may uncouple the stabilizing effect of the anterior and superior tissues, and whether it’s subsequent closure would restore the shoulder’s kinematics in this cadaveric model.

Four conditions were compared and the GH displacements during continuous abduction were recorded using a validated, automated upper extremity testing system with seven degrees of freedom. Glenohumeral kinematics was disturbed throughout abduction following RI manipulation. Both side-to-side closure and tightening altered the shoulder’s kinematics equivalently throughout all angles of abduction and resembled the kinematics of the baseline profile from 30–90° of abduction. This effect suggests that the RI plays an important role in regulating the shoulder’s motion as a mechanical support.

The equivalence of the two repair techniques on GH motion from 30 to 90° of abduction is clinically insightful. In treating shoulder instability, tightening the RI may be the ideal approach because it does not negatively affect the shoulder’s motion as this experiment demonstrates. Additionally, it produces a more widespread stabilizing effect on the shoulder joint than the side-to-side repair.

With the exception of medial-lateral displacement, the AUC of GH translation varied significantly from 90–150° of abduction when the baseline and repair conditions were compared. These data suggest that RI’s contribution to shoulder kinematics may be altered when the arm is abducted more than 90°. Large changes in shoulder kinematics were observed after both methods of RI closure, and questions the use of RI closure as the sole method of treatment for shoulder instability.

This study is not without limitations. First, we examined passive abduction in a cadaveric model. Dynamic forces central to glenohumeral stability and shoulder motion, including various degrees of rotation relating to physiologic GH motion, were not simulated [24]. Secondly, the size of the RI opening size was not studied. It was widely incised to separate the supraspinatus and infraspinatus from the base of the coracoid to the insertion on the humerus. Varying the length of the RI opening could influence the extent of change from baseline. Simply, a smaller incision may produce less change. Thirdly, GH translation was calculated using a regression analysis to determine the instant center of rotation using defined anatomical landmarks. The precision of this characterization depends on the quality of the anatomic calibration [6, 8] As a result of this estimation, there is inherent variability, as anatomical landmarks are areas rather than discrete points [25]. This effect may partly explain the inter-variability among specimens. The speed of the abduction from motion from 30° to 120° was simulated at a reduced speed than what may be considered clinically relevant. While Bergmann et al. [26] have demonstrated that reduced speeds change GH peak forces and corresponding moments, the directions of GH forces remain constant. Furthermore, glenohumeral forces were not calculated to further distinguish between the methods of RI closure. Due to the fact that this is a cadaveric study, it is impossible to predict the exact strength and functional outcome of repaired tissues. This simulated parametric biomechanical study can most assertively investigate the strength and functional outcome of suture materials and suture technique in the groups tested and serves as a physiologic approximation of shoulder kinematics. Moreover, the physiological organization of collagen and other matrix elements will change the characteristics of the tissues during the healing process whereas our results communicate affect immediately following RI insult and repair.

Changes in the rotator cuff interval have a significant effect on glenohumeral kinematics during abduction. This investigation provides an improved perspective on the role of the RI on GH kinematics during abduction. This study demonstrates that GH translation is decreased with RI closure solely during abduction without rotation in a cadaveric torso specimen. The impact on clinical situations needs addition study measuring translation with abduction and rotation of the humerus. When managing shoulder pathology, surgeons should consider how these different methods of RI closure affect the joint’s motion. Suture closure may offer the same benefit as tightening the rotator interval. These findings may be relevant with regard to the treatment of shoulder laxity.