Abstract
Background: Immobilization of the shoulder in external rotation has been shown to reduce the risk of recurrence after traumatic anteroinferior shoulder dislocation. It remains unclear how duration of immobilization affects labral coaptation.
Hypothesis: Immobilization of the shoulder in 30° of external rotation for 5 weeks allows better coaptation of the anteroinferior labrum than does an immobilization period of 3 weeks.
Study Design: Cohort study; Level of evidence, 2.
Methods: Twenty-two patients with traumatic anteroinferior dislocation of the glenohumeral joint were included in this study. Patients were divided into 2 groups. Group 1 consisted of the initial 11 patients (mean age, 37.4 years) immobilized for 3 weeks; group 2 consisted of the subsequent 11 patients (mean age, 29.7 years) immobilized for 5 weeks in 30° of external rotation. With use of magnetic resonance imaging, displacement and separation of the glenoid labrum and anterior joint effusion were assessed in different arm positions (internal rotation, neutral rotation, 30° of external rotation, maximum external rotation) within 3 days, 3 weeks, and 5 weeks after reduction.
Results: Displacement and separation of the labrum and anterior joint effusion were significantly less, particularly with maximum external rotation compared with neutral and internal rotation, during the acute magnetic resonance imaging evaluation in both groups (P < .05). No statistically significant differences were found in all parameters comparing internal rotation with neutral rotation, 30° of external rotation, and maximum external rotation in both groups after 5 weeks (P > .05). No statistically significant differences were found between both groups comparing the results of the measured variables during the acute, 3-week, and 5-week magnetic resonance imaging examinations (P > .05).
Conclusion: Immobilization of the shoulder in 30° of external rotation seems to allow a similar coaptation of the glenoid labrum, regardless of duration of immobilization (3 vs 5 weeks). Clinical trials are needed to evaluate the effect of these results on recurrence rates. The optimum position of immobilization in external rotation has yet to be determined.
Keywords
Dislocation of the glenohumeral joint is well known for its high recurrence rates, which vary from <25% to up to 100% according to different studies. ‖ The unacceptably high failure rates after conventional nonoperative treatment with sling immobilization of the shoulder in the adduction and internal rotation position, particularly in young adults, have been described. Today, arthroscopic and open shoulder stabilization procedures show better results regarding the recurrence rate than does nonoperative management. 1 – 3 , 11 – 13 , 22
References 5-7, 14, 15, 17, 19-21, 25, 28.
Using MRI, Itoi et al 10 found that immobilization of the arm in external rotation better approximates the Bankart lesion to the glenoid neck than does the conventional (traditional) position of internal rotation. Different cadaveric and MRI studies demonstrated that tensioning of the anteroinferior structures (subscapularis muscle, labroligamentous structures, capsule) in external rotation improved the tissue apposition and, therefore, probably the healing of the anteroinferior labral lesion in a more anatomical position. 10 , 16 , 23 Itoi et al 9 have recently presented the 2-year results of a prospective clinical trial comparing immobilization of the shoulder in internal versus 10° of external rotation for 3 weeks. The authors found that immobilization in external rotation significantly reduces the risk of recurrence compared with the conventional immobilization method. There are no studies currently available that have examined how different periods of immobilization in external rotation affect labral coaptation and the final reduction results of the labral lesion. The purpose of this study was to assess position, reduction, and mobility of the displaced anteroinferior labrum in patients after traumatic dislocation of the shoulder dependent on the duration of immobilization. We hypothesized that patients who underwent immobilization for 5 weeks would show a better coaptation of the disrupted anteroinferior labrum than would those immobilized for 3 weeks in 30° of external rotation.
Materials and Methods
Twenty-two patients were prospectively included in this study. The inclusion criteria were a primary traumatic anteroinferior shoulder dislocation with an anteroinferior labral detachment and a corresponding Hill-Sachs lesion confirmed on MRI. Exclusion criteria were patients with atraumatic or recurrent shoulder instability, a humeral avulsion of the glenohumeral ligament lesion, a glenoid rim fracture (bony Bankart lesion), a fracture of the greater tuberosity, associated rotator cuff injury, or neurovascular injuries. Patients were also excluded if they did not finish the required immobilization period. All patients were seen in our emergency department with a dislocated shoulder. After radiologic confirmation of the diagnosis, the shoulder was reduced manually, and the arm was placed in an external rotation brace (Ultrasling ER 30°, Donjoy, Carlsbad, California) designed to immobilize the shoulder in slight abduction and 30° of external rotation (Figure 1). Twenty-two patients were prospectively included into 2 groups. Group 1 consisted of 11 patients (2 women and 9 men) with a mean age of 37.4 years (range, 23–54 years) who underwent a 3-week period of immobilization, and group 2 consisted of 11 patients (2 women and 9 men) with a mean age of 29.7 years (range, 19–43 years) who were immobilized for 5 weeks using the above mentioned orthosis. The patients were treated successively, meaning that the first 11 patients were included in group 1, whereas the following 11 patients were included in group 2. There was no significant (P = .107) difference in age between the groups. All patients gave informed consent to participate in this study. Elbow range of motion was permitted during the time of immobilization with or without the brace, and the patients were told to avoid any internal rotation movement while taking the brace on or off.
Immobilization of the shoulder in slight abduction and 30° of external rotation using the Ultrasling ER 30°.
Magnetic resonance imaging was performed within 3 days after reduction as well as 3 and 5 weeks after reduction. After the brace was removed, the patient began a standardized physical therapy program. After a passive free range of motion was achieved, the patients started an intensive strengthening program including deltoid, rotator cuff, and scapulothoracic muscle exercises.
Magnetic Resonance Imaging
All patients had MRI performed with a 1.5-T system (Signa Twin Speed, General Electric Medical Systems, Milwaukee, Wisconsin). The shoulder was placed in supine position in a commercially available shoulder coil. This shoulder coil (Large Shoulder Array Coil, MRI Devices Corporation, Mankesha, Wisconsin) contributed to noise reduction and improved spatial resolution.
The following sequences were applied in transaxial, parasagittal (perpendicular to the scapula), and paracoronal (parallel to the scapula) slice orientation:
localizer sequence in all 3 directions of space;
parasagittal T1-weighted spin echo (SE) sequence (repetition time [TR], 560 milliseconds; echo time [TE], 14 milliseconds; flip angle, 90°; field of view, 200 mm; slice thickness, 3 mm; matrix, 256 × 224);
paracoronal T1-weighted SE sequence (TR, 560 milliseconds; TE, 14 milliseconds; flip angle, 90°; field of view, 200 mm; slice thickness, 3 mm; matrix, 256 × 224);
paracoronal proton density weighted SE sequence (TR, 2420 milliseconds; TE, 28.2 milliseconds; flip angle, 90°; field of view, 200 mm; slice thickness, 3 mm; matrix, 256 × 224);
transaxial T1-weighted SE sequence (TR, 560 milliseconds; TE, 14 milliseconds; flip angle, 90°; field of view, 200 mm; slice thickness, 3 mm; matrix, 256 × 224);
transaxial proton density weighted SE sequences (TR, 2300 milliseconds; TE, 27.5 milliseconds; flip angle, 90°; field of view, 200 mm; slice thickness, 4 mm; matrix, 256 × 224).
The shoulder was placed in internal rotation first, simulating the conventional sling immobilization position. Next, the arm was positioned in neutral rotation, followed by 30° of external rotation resembling the current immobilization position, and finally in maximum external rotation to the extent that it was tolerated by the patient. For that, the arm rested at the side of the trunk and was stabilized with small sandbags to avoid movement of the shoulder during the examination.
For all sequences, 2D acquisitions were used. The phase codifying was done in anteroposterior orientation parallel to the y-axis. The acquired data were saved as DICOM files on compact disc. For data analysis, the work station Advantage Windows (General Electric) and the software program Functool 2.5.24 (General Electric) were used.
Measurements on MRI
On the basis of the work of Itoi et al, 10 the inferior one-third level of the glenoid was used to assess position, reduction, and mobility of the anteroinferior labrum. The following parameters were assessed: (1) displacement (defined as the distance in millimeters between the tip of the labrum and the tip of the glenoid rim), (2) separation (defined as the distance in millimeters between the inner margin of the labrum and the anterior aspect of the glenoid neck), and (3) anterior joint effusion (defined as the area in square millimeters posterior to the anterior capsule and anterior to the glenoid and humeral head) (Figure 2). Changes in displacement and separation of the labrum were used as the main predictors of labral coaptation. Displacement was given a positive value when the labrum was displaced medially to the anteroinferior rim of the glenoid and a negative value when it was found laterally toward the humeral head. Separation was given a positive value when the labrum was found anterior to the tip of the glenoid, and it was given a negative value when it was found posterior to the tip of the glenoid rim. All measurements were performed by an experienced radiologist in a blinded fashion such that the examiner did not know if the patient was immobilized for 3 or 5 weeks in 30° of external rotation.
Measurements of displacement (1) and separation (2) of the labrum (in millimeters) and anterior joint effusion (3) (in square millimeters)
Statistical Methods
Statistics were carried out using the Mann-Whitney U test and the Wilcoxon signed rank test.
A post hoc analysis was performed for the calculation of the statistical power of the study. All statistical analyses were performed with SPSS, version 13.0 for Windows (SPSS Inc, Chicago, Illinois).
Results
The mean values and ranges for displacement, separation, and anterior joint effusion with internal rotation, neutral rotation, 30° of external rotation, and maximum external rotation are displayed in Table 1. The post hoc analysis revealed an achieved power of the study between 0.6 and 0.76 with an effect size between 1 and 1.2 for the analyzed parameters.
Mean Values and Ranges for Displacement and Separation of the Labrum and Anterior Joint Effusion With Different Arm Positions
Results of the MRI Measurements within 3 Days after Reduction (Acute MRI)
No statistically significant differences were found between groups comparing the results of the measured variables during the acute MRI examination. When internal rotation was compared with neutral rotation, no statistically significant differences were found in all parameters. When 30° of external rotation was compared with internal rotation, displacement (P = .028) and anterior joint effusion (P = .021) decreased significantly, whereas separation was significantly less only in group 2 (P = .015). When maximum external rotation was compared with internal rotation, all parameters decreased significantly in group 1 (displacement, P = .001; separation, P = .006; anterior joint effusion, P = .005) and in group 2 (displacement, P = .002; separation, P = .001; anterior joint effusion, P = .008).
Results of the MRI Measurements at 3 Weeks after Reduction (3-Week MRI)
No statistically significant differences were found between groups comparing the results of the measured variables at the 3-week MRI examination. In addition, no significant differences were found in displacement and separation comparing internal rotation with neutral rotation and 30° of external rotation in both groups. When maximum external rotation was compared with internal rotation, anterior joint effusion was significantly less in group 1 (P = .001) and in group 2 (P = .005). In addition, displacement was decreased with maximum external rotation compared with internal rotation in group 1 (P = .01) and in group 2 (P = .02), and separation decreased significantly in group 1 (P = .028) and group 2 (P = .031).
Results of the MRI Measurements at 5 Weeks after Reduction (5-Week MRI)
No statistically significant differences were found between groups comparing the results of the measured variables at the 5-week MRI measurements. In addition, no statistically significant differences were found for displacement and separation comparing internal rotation with neutral rotation, 30° of external rotation, and maximum external rotation in both groups. However, anterior joint effusion still decreased significantly with maximum external rotation compared with internal rotation in group 1 (P = .041) and group 2 (P = .038).
Comparison of Results between the Acute and 3-Week MRI Measurements
In internal rotation, all measured parameters decreased significantly from the acute to the 3-week MRI measurements in group 1 (displacement, P = .028; separation, P = .02; anterior joint effusion, P = .038) and in group 2 (displacement, P = .038; separation, P = .02; anterior joint effusion, P = .021). In addition, all measured parameters decreased significantly in neutral rotation in group 1 (displacement, P = .018; separation, P = .038; anterior joint effusion, P = .021) and in group 2 (displacement, P = .038; separation, P = .015; anterior joint effusion, P = .036). However, no significant differences were found in displacement and separation of the labrum with 30° of external rotation and maximum external rotation in both groups.
Comparison of Results between the 3- and 5-Week MRI Measurements
All measured parameters still decreased significantly in both groups from the 3-week to the 5-week MRI measurements in internal rotation (displacement, P = .025; separation, P = .031; anterior joint effusion, P = .002), neutral rotation (displacement, P = .031; separation, P = .027; anterior joint effusion, P = .001), and 30° of external rotation (displacement, P = .041; separation, P = .038; anterior joint effusion, P = .008).
Discussion
The new concept of immobilization of the arm in external rotation after anteroinferior dislocation of the glenohumeral joint has gained increasing interest. Itoi et al 10 were the first who found that immobilization of the arm in external rotation better approximates the Bankart lesion to the glenoid neck than does the conventional position of internal rotation. The authors recently published the results of a prospective clinical trial in which patients were immobilized either in conventional internal rotation or in 10° of external rotation for 3 weeks. 9 A recurrence rate of 42% for the internal rotation group and a recurrence rate of 26% for the external rotation group were reported after a mean follow-up of 25.6 months. The goal of our study was not to present the clinical results of our patients in terms of recurrence because the currently available clinical follow-up is too short. We believe, even with the new immobilization techniques, at least 2 years of follow-up are necessary to provide sufficient clinical data because the risk of recurrence is highest within the first 2 years after the initial dislocation and then reaches a plateau. 18
The aim of this study was to evaluate position, reduction, and mobility of the anteroinferior labrum after different durations of immobilization in external rotation. In contrast to other studies, we immobilized the patients using the Ultrasling ER with a 30° wedge, which represents a soft cushion–type brace. Sullivan et al 24 have shown that the commercially available shoulder external rotation braces vary in their ability to achieve and maintain a desired position of external rotation and comfort ratings. When the authors examined the Ultrasling ER with a 15° wedge, they found that the actual immobilization position was only 9.4°. We believe that using the 30° wedge, a higher degree of external rotation can be achieved compared with other orthoses. This seems to be beneficial in terms of labral coaptation and perhaps soft tissue healing. Miller et al 16 measured the contact force between the detached labrum and the glenoid in a cadaveric model. They found that there was no detectable contact force when the arm was placed in internal rotation, the contact force increased as the arm passed through neutral rotation, and the contact force reached a maximum at 45° of external rotation. We compared 3 versus 5 weeks of immobilization because the period of 3 weeks was recommended by Itoi et al 8 in their initial study. Our hypothesis was that immobilization of the shoulder in 30° of external rotation for 5 weeks allows a better position, reduction, and stability of the anteroinferior labrum compared with an immobilization period of 3 weeks owing to improved soft tissue healing. In agreement with Itoi et al, we found significantly less displacement and separation of the labrum and significantly less anterior joint effusion, particularly with maximum external rotation, compared with neutral and internal rotation during the acute MRI evaluation (Figure 3). These findings were similar in both groups and suggest an improved labral coaptation due to tensioning of the anterior soft tissue structures. Although when 30° of external rotation was compared with internal rotation during the acute evaluation, displacement and anterior joint effusion decreased significantly in both groups, whereas separation was significantly less only in group 2. However, after 3 weeks of immobilization, no differences were found between groups in displacement and separation comparing internal rotation with neutral rotation and 30° of external rotation, suggesting a stable position of the anteroinferior labrum between internal rotation and 30° of external rotation range of motion. However, when maximum external rotation was compared with internal rotation displacement, separation and anterior joint effusion still decreased significantly in both groups. Although these findings after 3 weeks suggest that a further immobilization, particularly in maximum external rotation, seems beneficial, no statistically significant differences were found in all parameters after 5 weeks comparing internal rotation with all other arm positions in both groups (Figure 4).
Acute MRI showing the anteroinferior labral lesion in internal rotation (A), neutral rotation (B), 30° of external rotation (C), and maximum external rotation (D). With the arm in internal and neutral rotation (A and B, respectively), the joint cavity anterior to the glenoid is wide open and the labrum displaced. With advanced external rotation, the anterior joint cavity is closed, the hematoma is shifted posteriorly, and the labrum is reduced on the glenoid rim (C and D).
Magnetic resonance imaging 5 weeks after reduction of the same patient as in Figure 3 showing the anteroinferior labrum in internal rotation (A), neutral rotation (B), 30° of external rotation (C), and maximum external rotation (D). An anatomical and stable reduction of the anteroinferior labrum is seen with all different arm positions.
When the acute MRI measurements were compared with the 3-week measurements, displacement and separation of the labrum decreased significantly in both groups with internal and neutral rotation, suggesting an improved labral coaptation after 3 weeks of immobilization in 30° of external rotation. No differences in labral parameters were found with 30° of external rotation and maximum external rotation in both groups after 3 weeks of immobilization, although the values for anterior joint effusion still decreased significantly. However, when the 3-week MRI measurements were compared with the 5-week measurements, all measured parameters still decreased significantly. Again, these findings were similar in both groups, suggesting that a prolonged immobilization up to 5 weeks is perhaps not necessary to achieve stable and more anatomical position of the labrum at 5-week follow-up. Seybold et al 23 demonstrated a significantly better position of the disrupted labrum in external rotation when the initial investigation was compared with the results 6 weeks after reduction after an immobilization period of 3 weeks in 10° of external rotation. The initial immobilization period of 3 weeks seems to be enough to improve labral coaptation.
We agree with Itoi et al 10 that the presence of hematoma or joint effusion has a significant effect on labral coaptation and may even compromise the healing by pushing the anterior capsulolabral structures of the glenoid. Most of our patients had a large amount of hematoma, which has been reported to resolve within 3 to 7 weeks after the initial dislocation. 27 Our study has confirmed that with internal rotation, the effusion was mainly located anteriorly, leading to a distension of the lax anterior capsule. With 30° of external rotation, and even more with maximum external rotation, a posterior shift of the joint effusion was noted owing to the tightening of the anterior soft tissue structures. Although the definite role of the joint effusion has yet to be determined, Wintzell et al 26 have shown that arthroscopic lavage and evacuation of the interposing hematoma reduce the risk for recurrent dislocation when compared with the conventional nonoperative treatment. Because arthroscopic lavage represents an invasive procedure performed after anesthesia, we have now started ultrasound-controlled aspiration in patients with a significant amount of anterior joint effusion to decrease the amount of capsular distension and improve labral coaptation before immobilization of the shoulder in external rotation is applied.
Our study implies that immobilization in maximum external rotation may improve labral coaptation to a higher extent than does 30° of external rotation. As mentioned previously, Miller et al 16 have shown that the higher the angle of external rotation, the higher the contact pressure between the detached labrum and the glenoid in a cadaveric model. In addition, Hart and Kelly 4 observed during arthroscopy that the best reduction was achieved with 30° of abduction and 60° of external rotation. Further clinical and basic science studies are needed to evaluate the optimum position of the arm in terms of labral coaptation after anterior shoulder dislocation, particularly with regard to the presence or absence of hematoma or joint effusion.
Our study has some limitations. First, although it is a prospective study, it is not randomized. However, because there were no statistically significant differences regarding the measured parameters during the acute MRI between groups, we believe a meaningful comparison is possible. We are also aware of the different mean ages in the groups studied; however, this was not statistically significantly different. A second potential limitation is the lack of arthro-MRI performance at the 5-week follow-up that may have detected a greater amount of persistent labrum lesion. Because the goal of this study was to evaluate mobility and position of the labrum in relation to the amount and changes of anterior joint effusion over time, MRIs were performed and allowed evaluation of the anteroinferior labrum in all of our patients. In addition, all patients had MRI evaluation performed at 3-week follow-up. One might argue that this would have negated the potential advantage of the additional 2 weeks of immobilization in group 2 because the shoulders in this group were briefly internally rotated to carry out the 3-week MRI study. The radiologic parameters did not show any significant differences in displacement and separation comparing internal rotation with neutral rotation and 30° of external rotation in either group at 3-week follow-up. We therefore believe that briefly placing the patient in internal rotation after 3 weeks did not have any negative effect on the final labral position. Finally, this study does not include any data concerning intraobserver or interobserver reliability, particularly with regard to the radiologic evaluation. However, all measurements had been performed by a blinded experienced radiologist in a standardized manner.
Immobilization of the shoulder in 30° of external rotation seems to allow a similar coaptation of the glenoid labrum regardless of the duration of immobilization (3 vs 5 weeks). A prospective, randomized controlled trial has been initiated to evaluate the effect of these results on the recurrence rates. In addition, the optimum position of immobilization in external rotation has yet to be determined.