Visualizing the Effect of Arthroscopic Bankart Repair and Remplissage: A Pre- and Postoperative Ultrasonographic Evaluation

Abstract

Background:

Arthroscopic Bankart repair and the remplissage procedure are minimally invasive treatments for anterior shoulder instability. Cadaveric studies have demonstrated the biomechanical effectiveness of these treatments, and clinical studies have reported favorable outcomes; however, quantitative clinical data are required to confirm their in vivo biomechanical efficacy.

Hypothesis:

Bankart repair and remplissage, based on previous preclinical studies, could significantly reduce anterior humeral head translation (AHHT) in clinical patients.

Study Design:

Cross-sectional study; Level of evidence, 3.

Methods:

A total of 29 patients (23 male, 6 female; mean age, 21.7 ± 7.1 years), diagnosed with anterior shoulder instability, were studied between March 2020 and September 2023. Of these patients, 11 underwent only arthroscopic Bankart repair, whereas the remaining 18 underwent remplissage combined with arthroscopic Bankart repair. All procedures were performed under general anesthesia in the beach-chair position. Ultrasonographic evaluations were conducted by the same surgeon immediately pre- and postoperatively. In each shoulder abduction position at 0°, 45°, and 90°, the posterior aspect of the glenohumeral joint was visualized using ultrasonography before and after applying a 40-N anterior force to the upper arm using a handheld dynamometer. AHHT was quantified by measuring the change in distance from the posterior edge of the glenoid to the humeral head due to the traction force at each shoulder angle.

Results:

Postoperative assessments revealed a significant reduction in AHHT at 90° (from 8.8 ± 5.3 mm to 6.1 ± 2.8 mm; P = .02) of abduction. Subgroup analysis showed that this reduction was significant only when Bankart repair was combined with remplissage (from 10.1 ± 6.0 mm to 5.9 ± 2.9 mm; P = .01). Posterior shifts of the humeral head without traction were observed at 0° (2.7 ± 3.2 mm; P < .001), 45° (2.4 ± 2.7 mm; P < .001), and 90° (3.6 ± 2.5 mm; P < .001) of abduction. Subgroup analysis showed these shifts were significant for all abduction angles except 0° in the group that underwent Bankart repair with remplissage (for Bankart repair alone: [0°] 4.4 ± 3.3 mm, P = .005; [45°] 3.0 ± 3.2 mm, P = .04; and [90°] 4.6 ± 2.1 mm; P < .001; for Bankart repair with remplissage: [0°] 1.6 ± 2.6 mm, P = .06; [45°] 2.1 ± 2.3 mm, P = .006; and [90°] 3.0 ± 2.5 mm, P < .001).

Conclusion:

Arthroscopic Bankart repair combined with remplissage significantly reduced AHHT at 90° of abduction, whereas Bankart repair alone showed no significant decrease in AHHT at any abduction angle. However, both Bankart repair alone and Bankart repair with remplissage caused a posterior shift of the humeral head without load at 0°, 45°, and 90° of abduction. Measuring AHHT could be critical for assessing postoperative anterior shoulder instability, as it provides the surgeon with objective data to evaluate the procedure's success and guide postoperative management. Future research could investigate clinical differences based on the measured AHHT, potentially providing valuable insights for improving patient outcomes.

Anterior shoulder instability is the most common form of joint instability, affecting approximately 1.7% of the general population, and is more prevalent in male individuals.^33,47 This condition is predominantly observed in younger individuals and most commonly occurs after shoulder dislocation. ⁵³ It is particularly prevalent during adolescence, accounting for 80% of shoulder instability cases in young athletes.^3,25,53 Patients with anterior shoulder instability are at a high risk of recurrent dislocations, characterized by anterior movement of the humeral head, causing discomfort and instability. Clinically, physical examinations, such as the load-and-shift test, anterior drawer test, apprehension test, and relocation test, are performed.^39,50 Additionally, imaging modalities, such as computed tomography and magnetic resonance imaging, are also invaluable in the assessment of anterior shoulder instability.^38,44 In recent years, ultrasonographic evaluation has become widespread, providing a noninvasive and real-time imaging assessment. Quantitative assessment of anterior shoulder instability using ultrasonography, both pre- and postoperatively, has been explored in several studies.^26,27

Arthroscopic Bankart repair is a widely performed surgery for the treatment of traumatic anterior shoulder instability. ¹² This minimally invasive procedure is primarily conducted in patients with minor bone defects and on-track lesions. However, despite this, the reported risk of recurrent dislocation is 4% to 17%.^24,52 For patients with a high risk of recurrent dislocation, additional techniques, such as remplissage, are incorporated.^{10,15,35,45,54} While cadaveric studies have shown that Bankart repair can decrease humeral head translation, thus contributing to shoulder stability,^2,7,8,29,36 only few researchers have quantitatively evaluated the effectiveness of Bankart repair in actual clinical practice. Peltz et al ⁴³ reported that in patients with anterior shoulder instability, the humerus was significantly shifted posteriorly by 1.1 mm during the apprehension test after arthroscopic Bankart repair. Similarly, multiple in vitro biomechanical studies have demonstrated that remplissage can stabilize the glenohumeral joint.^{4,16,17,19-21,23,55} However, to the best of our knowledge, no studies have demonstrated the in vivo biomechanical effect of remplissage. To advance knowledge on dynamic effects after Bankart repair and remplissage, in this study we aimed to evaluate the effects of arthroscopic Bankart repair and remplissage on anterior instability using ultrasonography immediately pre- and postoperatively. On the basis of previous literature and preclinical studies, we hypothesized that Bankart repair and remplissage would result in (1) a decrease in the amount of anterior translation of the humeral head caused by anterior traction and (2) a posterior shift of the humeral head without traction.

Methods

This cross-sectional study was conducted with the approval of our institute's ethics committee and involved 29 patients diagnosed with posttraumatic anterior shoulder instability by a positive apprehension test between March 2020 and September 2023 (Figure 1). All patients were scheduled for arthroscopic Bankart repair at a facility specializing in arthroscopic surgery and sports medicine. This procedure was indicated for patients with <20% bone loss in the scapular glenoid and was evaluated using the Sugaya index, which measures the ratio of glenoid defect width to its total diameter. ⁴⁸ Various studies have shown that for patients with recurrent anterior shoulder instability and <20% glenoid defect, stability can be effectively regained after arthroscopic Bankart repair alone.^13,28,34,40 In conjunction with Bankart repair, remplissage was utilized in patients aged <20 years, with off-track lesions, external rotation >85°, ⁵ hyperabduction >105°, ¹⁸ or a Beighton score of >4 points. ⁶ Patients participating in collision sports were not included in this study because they all underwent an arthroscopic Bankart-Bristow procedure, ⁹ a form of coracoid transfer, due to the higher reported recurrence rates postrepair among collision athletes than among noncollision athletes.^1,42 Of the 53 patients who were assessed for eligibility, 24 were excluded from the study for the following reasons: glenoid fracture (n = 7), rotator cuff tear (n = 8), osteoarthritis in the glenohumeral joint (n = 0), unstable painful shoulder (n = 2), multidirectional instability (n = 2), previous unsuccessful surgeries on the affected shoulder (n = 4), unawareness of obvious dislocations (n = 0), or insufficient data (n = 1). Informed consent was obtained from all participants. A single surgeon (A.T.) with >20 years of experience in arthroscopic surgery and ultrasonographic examination conducted all surgical and sonographic imaging procedures.

Figure 1.

Flowchart of the inclusion, exclusion, and allocation of patients in the study on arthroscopic Bankart repair. Group N consists of patients who underwent Bankart repair without the addition of remplissage; group R consists of patients who underwent Bankart repair with the addition of remplissage.

Surgical Procedures

Arthroscopic surgery was performed with the patient in the beach-chair position under general anesthesia, supplemented by an interscalene nerve block using 10 mL of 0.25% levobupivacaine. For Bankart repair, standard posterior and anterior portals were established. The anteroinferior labral complex was detached from the glenoid rim, ranging from the 6-o’clock to the 1- or 2-o’clock position on the right shoulder and to the 10- or 11-o’clock position on the left shoulder. The repair of the anteroinferior labral complex involved using ≥4 anchors (Gryphon BR; DePuy Mitek) placed from the 6-o’clock to the 1- or 2-o’clock position on the right shoulder and to the 10- or 11-o’clock position on the left shoulder. For remplissage, an additional portal called the “remplissage portal”^10,45,54 was created above the Hill-Sachs lesion.

One or 2 anchors (Gryphon BR) were inserted into the Hill-Sachs lesion, and the number of anchors was determined by the lesion size. One of the 2 sutures for each anchor was removed, and the remaining suture was drawn through the posterior capsule and infraspinatus muscle. After repairing the anteroinferior labral complex, mattress stitches were secured on the posterior aspect of the infraspinatus muscle.^10,45,54

Sonographic Measurements

Ultrasonographic examinations were conducted preoperatively (immediately after general anesthesia induction) and postoperatively (immediately after surgical procedure completion). The patient was placed in the beach-chair position with the torso at a 45° angle, and the forearm was then secured with a positioning device (TRIMANO FORTIS Support Arm; Arthrex) that was attached to the bed. The shoulder was kept in a position with no flexion or rotation, and the elbow was bent at 90°, also in neutral position. ²⁷ All sonographic examinations were conducted by the same surgeon (A.T.) using an ultrasound imaging system (SONIMAGE HS1; Konica Minolta) equipped with a 3- to 11-MHz linear probe.

To visualize the glenohumeral joint, particularly between the infraspinatus and teres minor muscles, the examiner aligned the transducer parallel to the longitudinal body axis. This alignment allowed us to view the short axes of both muscles. For comprehensive imaging, the transducer was angled approximately 90° between the 2 muscles, aligning it parallel to the scapular spine. Special attention was paid to maintaining a consistent probe positioning and angle during each examination to minimize variability. This technique enabled visualization of both the posterior aspect of the glenoid and the humeral head.

During the examination, one clinician (A.T.) maintained the transducer's position, while another (N.O.) applied a 40-N constant anterior traction to the humerus, measured with a digital handheld dynamometer (Ergo FET; Nihon Medix Co). Force was applied using a traction band placed on the upper arm (Figure 2). Ultrasonographic imaging was performed at shoulder abduction angles of 0°, 45°, and 90°, all with the shoulder in a neutral rotation position. Images were acquired with and without the application of anterior force. After each anterior translation test, the examiner (A.T.) performed a manual compression maneuver to recenter the humeral head within the glenoid (Figure 3). An independent observer (T.O.), an orthopaedic surgeon with 8 years of musculoskeletal sonography experience, reviewed the uploaded ultrasonographic images on the picture archiving and communication system (RapideyeCore; Canon Medical Systems). This observer was blinded to all of the examinations and surgeries. To measure the shortest distance between the posterior edges of the glenoid and humeral head, 2 parallel lines were delineated, both without (D1) and with a 40-N force (D2) applied. The assignment of negative values occurred if the posterior edge of the humeral head was positioned anteriorly to the glenoid. The AHHT was calculated by subtracting D1 from D2 (Figures 4 and 5).

Figure 2.

Imaging of the posterior glenohumeral joint. During the examination, one examiner held the ultrasound transducer steady, while another examiner applied a 40-N force to the humerus in the anterior direction using a dynamometer.

Figure 3.

Flowchart of the sonographic measurement protocol.

Figure 4.

Ultrasonographic images of the posterior part of the glenohumeral joint. (A) Joint without any applied traction. (B) Joint under a 40-N anterior traction. A parallel line along the infraspinatus (ISP) muscle was drawn adjacent to the posterior edge of the glenoid (white line) and the humeral head (yellow dashed line). The minimal distance between these lines was measured in both scenarios (indicated by a vertical yellow line with arrows): without any traction (D1) and with a 40-N traction (D2). Negative values were assigned if the yellow dashed line was below the white line in the images. Anterior humeral head translation was calculated as the difference between D1 and D2.

Figure 5.

Schematic representation of the measurements of the position of the posterior edge of the humeral head relative to the posterior edge of the glenoid and its translation due to anterior traction. A baseline was drawn adjacent to the posterior edge of the glenoid. Each line parallel to the baseline was drawn adjacent to the posterior edge of the humeral head without any traction (D1) and with a 40-N anterior traction (D2). D1 represents the distance between the baseline and the parallel line without any force, and D2 represents the distance with a 40-N force. Negative values were assigned if the posterior edge of the humeral head was anterior to that of the glenoid. Anterior humeral head translation (AHHT) was calculated by subtracting D1 from D2.

To calculate the interobserver reliability of AHHT measurement, the pre- and postoperative measurements were performed by another independent observer (K.K.), an orthopaedic surgeon with 5 years of musculoskeletal sonography experience, across 10 consecutive patients. Intraobserver reliability was assessed by the primary observer (T.O.), who repeated the measurements on the same 10 patients.

Statistical Analysis

The primary endpoint was the AHHT change from pre- to postoperative measurements. To perform a detailed analysis, the patients were categorized into 2 subgroups: group N, which consisted of patients who underwent Bankart repair without the addition of remplissage; and group R, which contained patients who underwent the repair with remplissage. The secondary endpoints were established to further dissect the outcomes, which involved assessing the differences in D1 across all patients from the pre- to the postoperative phases. Moreover, within each subgroup, variations in AHHT and D1 were analyzed pre- and postoperatively. All statistical analyses were performed using EZR (Saitama Medical Center; Jichi Medical University), a graphical user interface for R (Version 2.13.0; The R Foundation for Statistical Computing). ³⁰ EZR is a modified version of R Commander (Version 1.6-3) enhanced with additional statistical functions frequently used in biostatistics.

D1 and AHHT measurements were compared pre- and postoperatively using a Bonferroni-adjusted paired t test at each shoulder abduction angle (0°, 45°, and 90°). All tests were 2-sided, and P < .05 denoted statistical significance. The intra- and interobserver reliability of AHHT measurements was calculated using the intraclass correlation coefficient (ICC[1,1] and ICC[2,1], respectively). ICC values were interpreted as follows: <0.50, poor agreement; 0.50 to <0.75, moderate agreement; 0.75 to <0.90, good agreement; and ≥0.90, excellent agreement. ³¹ A priori power analysis was conducted using G* Power (Version 3.1.9; Heinrich Heine University) to calculate the necessary sample size required to detect changes in measurement pre- and postoperatively. Our analysis revealed that ≥28 cases were required to attain a power (1-beta) of 0.80, assuming a significance level (alpha) of .05 and an effect size of 0.5. The effect size was derived from the data obtained from measurements taken in the 10 consecutive patients in this study.

Results

Patient Background

In total, 29 Japanese patients were evaluated (sex, 79.3% male; mean age, 21.7 ± 7.1 years; mean height, 167.6 ± 14.3 cm; mean weight, 63.7 ± 16.4 kg; mean body mass index, 22.7 ± 5.2 kg/m²). The median number of dislocations or subluxations was 3 (IQR, 2-7), and the median duration of symptoms preoperatively was 1 (IQR, 0.5-7.0) year. Among these patients, 8 engaged in baseball, 4 in soccer, 2 in basketball, 2 in handball, 1 in badminton, 1 in bouldering, 1 in dance, 1 in gymnastics, 1 in tennis, and 1 in volleyball. The remaining 7 patients did not engage in any sports. Of the 29 patients, 11 underwent only arthroscopic Bankart repair (group N), whereas the remaining 18 underwent remplissage combined with arthroscopic Bankart repair (group R). Several patient characteristics differed significantly between the 2 groups. Patients in group N were significantly older than those in group R (25.0 ± 9.6 vs 19.7 ± 3.7 years; P = .03). Additionally, patients in group N had a significantly higher body weight (67.6 ± 8.2 vs 61.3 ± 7.6 kg; P = .03) and body mass index (23.8 ± 2.9 vs 22.0 ± 2.2 kg/m²; P = .04) than those in group R. Imaging findings revealed that the glenoid defect was significantly smaller in group N than in group R (6.7% ± 5.7% vs 10.6% ± 4.3%; P = .03). Furthermore, intraoperative findings showed a significantly higher incidence (P = .03) of capsular tears in group N (54.5%) than in group R (16.7%). Table 1 provides a comprehensive description of the patient characteristics.

Table 1.

Patient Demographics, Functions, and Findings ^a

Variables	All (N = 29)	Group N (n = 11)	Group R (n = 18)	P
Age, y	21.7 ± 7.1	25.0 ± 9.6	19.7 ± 3.7	.03 a
Sex
Male	23 (79.3)	9 (81.8)	14 (77.8)	.79 b
Female	6 (20.7)	2 (18.2)	4 (22.2)
Height, cm	167.6 ± 14.3	168.9 ± 9.2	166.8 ± 5.7	.23 a
Body weight, kg	63.7 ± 16.4	67.6 ± 8.2	61.3 ± 7.6	.03 a
Body mass index, kg/m2	22.7 ± 5.2	23.8 ± 2.9	22.0 ± 2.2	.04 a
No. of dislocations/subluxations	3 [2-7]	3 [1-6]	3.5 [2-8]	.48 c
Duration of symptoms, y	1 [0.5-7.0]	3 [1-11]	1 [0.5-7.0]	.08 c
Range of motion, deg
External rotation	68.8 ± 18.6	67.2 ± 21.9	69.7 ± 16.4	.73 a
Hyperabduction	115.0 ± 15.3	113.1 ± 13.2	116.0 ± 16.2	.62 a
General joint laxity
Beighton score	4.5 ± 2.4 (n = 26)	4.5 ± 2.4 (n = 8)	4.5 ± 2.4 (n = 17)	.48 a
Functional outcome score
Rowe score	32.3 ± 18.3 (n = 28)	38.2 ± 22.9	28.5 ± 13.4 (n = 17)	.09 a
Imaging findings
Glenoid defect, %	9.1 ± 5.2	6.7 ± 5.7	10.6 ± 4.3	.03 a
Hill-Sachs lesion	27 (93.1)	9 (81.8)	18 (100)	.06 b
Width, mm	11.6 ± 2.3	11.4 ± 2.3	11.7 ± 2.3	.36 a
Length, mm	20.8 ± 5.4	21.7 ± 4.3	20.4 ± 5.9	.30 a
Depth, mm	4.8 ± 2.2	4.5 ± 1.7	4.9 ± 2.4	.32 a
Glenoid track
On-track	28 (96.6)	11 (100)	17 (94.4)	.43 b
Off-track	1 (3.4)	0 (0)	1 (5.6)
Intraoperative finding
Capsular tear	9 (31.0)	6 (54.5)	3 (16.7)	.03 b

Data are presented as mean ± SD, No. (%), or median [IQR]. Missing Beighton and Rowe scores were addressed by excluding cases with absent data from the analysis under the assumption that data were missing completely at random. Bold indicates P < .05.

^aStatistical analyses were conducted using unpaired t test.

^bChi-square analysis.

^cMann-Whitney U test.

Reliability of Measurements

In evaluating AHHT, the intraobserver reliability (ICC[1,1]) was high at 0.914, with a 95% CI of 0.861 to 0.948, indicating excellent agreement. The interobserver reliability (ICC[2,1]) was determined to be 0.857 (95% CI, 0.754-0.916), indicating good agreement.

AHHT Due to Anterior Traction

In all patients, postoperative AHHT was significantly decreased compared with preoperative AHHT at 90° of abduction (from 8.8 ± 5.3 mm to 6.1 ± 2.8 mm; P = .02) (Figure 6). In the subgroup analysis, group N showed no significant changes in AHHT at all positions. However, in group R, AHHT significantly decreased at 90° of abduction (from 10.1 ± 6.0 mm to 5.9 ± 2.9 mm; P = .01) (Table 2). No significant differences were observed between the groups at any shoulder angle.

Figure 6.

Anterior humeral head translation (AHHT) at each shoulder abduction angle pre- and postoperatively. Error bars indicate SD. *Significant difference (P < .05).

Table 2.

Pre- to Postoperative Comparison of AHHT and D1 Measurements Across Group N and Group R ^a

Variable	Preoperative	Postoperative	Delta	Effect Size	P
AHHT, mm
Total (N = 29)
0° of Abduction	5.5 ± 3.3	4.3 ± 2.5	−1.2 ± 3.4	0.42 (–0.11 to 0.95)	.20
45° of Abduction	8.3 ± 5.8	6.3 ± 2.9	−2.0 ± 5.3	0.44 (–0.09 to 0.97)	.16
90° of Abduction	8.8 ± 5.3	6.1 ± 2.8	−2.6 ± 4.8	0.62 (0.09 to 1.15)	.02
Group N (n = 11)
0° of Abduction	5.2 ± 2.6	4.3 ± 1.6	−0.9 ± 3.1	0.42 (–0.44 to 1.28)	>.99
45° of Abduction	7.0 ± 3.3	6.3 ± 2.7	−0.7 ± 4.4	0.25 (–0.59 to 1.09)	>.99
90° of Abduction	6.6 ± 2.8	6.5 ± 2.6	−0.1 ± 3.1	0.04 (–0.80 to 0.88)	>.99
Group R (n = 18)
0° of Abduction	5.7 ± 3.6	4.3 ± 2.9	−1.4 ± 3.6	0.43 (–0.24 to 1.10)	.36
45° of Abduction	9.8 ± 5.6	6.3 ± 3.1	−3.5 ± 4.6	0.78 (0.07 to 1.49)	.17
90° of Abduction	10.1 ± 6.0	5.9 ± 2.9	−4.2 ± 5.1	0.89 (0.18 to 1.60)	.01
P value for the comparison of AHHT between group N and group R
0° of Abduction	>.99	>.99	>.99
45° of Abduction	.48	>.99	.41
90° of Abduction	.29	>.99	.08
D1, mm
Total (N = 29)
0° of Abduction	5.3 ± 2.8	8.0 ± 3.4	2.7 ± 3.2	0.86 (0.29 to 1.43)	<.001
45° of Abduction	5.7 ± 2.7	8.1 ± 3.0	2.4 ± 2.7	0.85 (0.28 to 1.42)	<.001
90° of Abduction	4.3 ± 2.3	7.9 ± 2.8	3.6 ± 2.5	1.41 (0.78 to 2.04)	<.001
Group N (n = 11)
0° of Abduction	5.9 ± 4.0	10.3 ± 3.4	4.4 ± 3.3	1.18 (0.20 to 2.16)	.005
45° of Abduction	6.5 ± 3.2	9.5 ± 3.3	3.0 ± 3.2	0.93 (0.01 to 1.85)	.04
90° of Abduction	4.8 ± 2.6	9.4 ± 2.3	4.6 ± 2.1	1.88 (0.72 to 3.04)	<.001
Group R (n = 18)
0° of Abduction	4.9 ± 1.6	6.5 ± 2.5	1.6 ± 2.6	0.78 (0.07 to 1.49)	.06
45° of Abduction	5.2 ± 2.3	7.2 ± 2.4	2.1 ± 2.3	0.88 (0.17 to 1.59)	.006
90° of Abduction	4.0 ± 2.1	7.0 ± 2.7	3.0 ± 2.5	1.26 (0.50 to 2.02)	<.001
P value for the comparison of D1 between group N and group R
0° of Abduction	>.99	.008	.07
45° of Abduction	.66	.14	>.99
90° of Abduction	>.99	.08	.29

^aData are presented as mean ± SD. Delta values are determined by subtracting the preoperative measurements from the postoperative measurements. Effect sizes are presented as Cohen d with 95% CI. Statistical analyses were conducted using Bonferroni-adjusted paired t tests for pre- and postoperative comparisons and Bonferroni-adjusted unpaired t tests for group comparisons. Bold indicates P < .05. AHHD, anterior humeral head translation; D1, anterior–posterior distance between the posterior edge of the glenoid and the posterior aspect of the humeral head without traction.

Position of the Posterior Edge of the Humeral Head

In all patients, postoperative D1 showed a significant increase compared with preoperative D1 in all positions, indicating a posterior shift of the humeral head (Figure 7). At 0° of abduction, it changed from 5.3 ± 2.8 mm to 8.0 ± 3.4 mm (P < .001); at 45° of abduction, it changed from 5.7 ± 2.7 mm to 8.1 ± 3.0 mm (P < .001); and at 90° of abduction, it changed from 4.3 ± 2.3 mm to 7.9 ± 2.8 mm (P < .001). In the subgroup analysis, both group N and group R showed significant increases at all positions, except at 0° of abduction in group R. Postoperative D1 in group N was significantly greater than that in group R at 0° of abduction (group N, 10.3 ± 3.4 mm; group R, 6.5 ± 2.5 mm; P = .008). Table 2 presents a detailed analysis.

Figure 7.

Glenoid posterior edge (D1) at each shoulder abduction angle pre- and postoperatively. Error bars indicate SD. **Significant difference (P < .001).

Discussion

In this study, 2 clinically significant changes were identified. First, AHHT significantly decreased at 90° of abduction after arthroscopic Bankart repair with remplissage but not with Bankart repair alone. Second, the humeral head was significantly shifted posteriorly at 0°, 45°, and 90° of abduction after arthroscopic Bankart repair, except at 0° of abduction after repair combined with remplissage.

Concerning assessment of joint stability, several methods have been developed, such as stress radiography, dynamic sonography, motion capture, 3-dimensional magnetic resonance imaging reconstruction, computed tomographic model–based tracking, and 4-dimensional computed tomography.^{11,14,32,37,41,43,49,51} Recently, the application of sonographic evaluation in clinical settings has been explored. In previous research, the use of AHHT measurements employed to quantitatively evaluate joint stability via ultrasonography using a dynamometer, which exerted a 40-N anterior force, demonstrated good intraobserver reliability (ICC[1,1] = 0.810) and moderate interobserver reliability (ICC[2,1] = 0.724). ²⁷ Additionally, a previous clinical study that conducted ultrasonographic measurements before and after the arthroscopic Bankart-Bristow procedure demonstrated a significant decrease in AHHT and indicated excellent intra- and interobserver reliability (ICC[1,1] = 0.952, ICC[2,1] = 0.912). ²⁶ In the current study, measurements taken before and after arthroscopic Bankart repair and remplissage confirmed the previously established reliability of the ultrasonographic assessment method, demonstrating excellent intraobserver reliability (ICC[1,1] = 0.914) and good interobserver reliability (ICC[2,1] = 0.857). Therefore, this method remains reliable for evaluating joint stability pre- and postoperatively.

Regarding the stabilizing effect on the glenohumeral joint after arthroscopic Bankart repair, a previous cadaveric study was conducted to investigate glenohumeral translation by applying a 20-N anterior load at 0° and 90° of abduction after anterior shoulder dislocation and arthroscopic Bankart repair. ³⁶ In that study, anterior translation decreased by 2.6 mm at 0° and 2.9 mm at 90° of shoulder abduction after arthroscopic Bankart repair. In our clinical study, arthroscopic Bankart repair significantly reduced AHHT by the application of a 40-N anterior traction at 90° of shoulder abduction. However, the subgroup analysis indicated that Bankart repair alone did not significantly reduce anterior translation at any of the examined shoulder abduction angles, with all changes being <1 mm. Differences between the aforementioned cadaveric study and our clinical study, such as in the methodology, sample characteristics, and the presence of muscle tonus in living participants, may explain the discrepancy.

To enhance shoulder joint stability, remplissage has been considered one of the effective additional surgical options. Multiple biomechanical studies over the past decade demonstrated that remplissage increased glenohumeral joint stiffness, thereby enhancing shoulder joint stability, at 90° of abduction with external rotation.^16,17,19-21 Elkinson et al ¹⁷ showed that the addition of remplissage to Bankart repair significantly increased joint stiffness (3.9 ± 3.2 N/mm) in specimens with a moderate (30%) Hill-Sachs defect. They reported that, while Bankart repair alone did not restore the level of joint stiffness to that of an intact shoulder, joint stiffness significantly exceeded that of an intact shoulder after Bankart repair with the addition of remplissage. However, no quantitative clinical studies have compared Bankart repair alone with Bankart repair combined with remplissage. Meanwhile, our subgroup analysis revealed that combining Bankart repair with remplissage significantly reduced anterior translation at 90° of shoulder abduction, resulting in reductions of 4.2 mm. The comparison of AHHT between Bankart repair alone and Bankart repair with remplissage found no significant differences. Although the difference did not attain significance, the preoperative AHHT tended to be greater in Bankart repair plus remplissage than that in Bankart repair alone (group N: 5.2, 7.0, and 6.6 mm; group R: 5.7, 9.8, and 10.1 mm, at 0°, 45°, and 90° of abduction, respectively). The higher trend of preoperative AHHT in group R can be explained by the discrepancy in glenoid bone loss between the groups. Patient data revealed that glenoid bone loss was significantly greater in group R than that in group N (group N, 6.7%; group R, 10.6%). Meanwhile, the postoperative AHHT was highly similar between the 2 groups (group N: 4.3, 6.3, and 6.5 mm; group R: 4.3, 6.3, and 5.9 mm, at 0°, 45°, and 90° of abduction, respectively). According to these findings, while an additional stabilizing effect beyond Bankart repair alone was not observed with the addition of remplissage, the remplissage procedure appears to have restored anterior stability. Our observation corroborates these biomechanical outcomes for the first time in clinical studies.

According to a cadaveric study related to humeral head position, ⁷ Bankart repair reduced both anterior and inferior shifts observed at 0°, 45°, and 90° of humeral abduction. These findings align with the humeral head position outcomes observed in our study. Additionally, a clinical study showed posterior translation of the humeral head during the apprehension test after arthroscopic Bankart repair, as assessed using computed tomography. ⁴³ However, this study was only implemented at 90° of abduction and maximal external rotation by comparing the humeral head positions before and at 6 months after surgery. In our current study, a posterior shift of the humeral head was observed immediately postoperatively at 0°, 45°, and 90° of abduction combined with neutral rotation via ultrasonography. These results suggest that Bankart repair effectively shifts the humeral head posteriorly at these abduction angles.

In the subgroup analysis, a significant posterior shift of the humeral head was also observed after surgery at each shoulder angle in both group N and group R, except at 0° of abduction in group R. However, in group R, the 95% CI of the effect size for the posterior shift at 0° of abduction ranged from 0.07 to 1.49, suggesting a potential posterior humeral head shift even at 0° of abduction after performing Bankart repair with remplissage. These outcomes indicate that the humeral head is posteriorly shifted postoperatively, with or without the addition of remplissage to the Bankart repair. Comparison of the 2 groups revealed that the posterior aspect of the humeral head was significantly more posterior after surgery in group N than in group R at 0° of abduction (group N, 10.3 mm; group R, 6.5 mm). At 45° and 90° of abduction, it tended to be more posterior in group N than in group R, although the difference lacked statistical significance (group N, 9.5 and 9.4 mm; group R, 7.2 and 7.0 mm, at 45° and 90°, respectively). These findings can be explained by the concept of obligate glenohumeral translation. ²² As described by Harryman et al, ²² obligate glenohumeral translation refers to anterior glenohumeral translation caused by posterior capsular tightness. The remplissage procedure tightens the posterior part of the glenohumeral joint, potentially introducing obligate translation in the anterior direction. This mechanism may explain the reduced posterior shift observed in group R.

Limitations

This study had several limitations. First, while the sample size was sufficient to compare pre- and postoperative conditions using a paired t test, it was insufficient to adequately investigate the differences between arthroscopic Bankart repair alone and its combination with remplissage using an unpaired t test. Additionally, there was heterogeneity among the patients, including those who underwent only Bankart repair and those who underwent Bankart repair with remplissage. A larger cohort could provide more robust evidence and improve the statistical power of this study. However, the effect sizes of comparing pre- with postoperative measurements were evaluated. The post hoc analysis revealed that all variables with P < .05 achieved a power (1-beta) of >0.8, except for D1 in group N at 45° of abduction, which had a power of 0.79. This suggests that, with the exception of this variable, the sample size was adequate for evaluating the differences between pre- and postoperative measurements. Second, the reliability of sonographic examinations was not fully assessed in this study. Although we minimized variability by having all examinations conducted by the same highly experienced surgeon, the repeatability of the imaging results when performed by different examiners remains uncertain. However, previous studies have found that the test-retest reliability of sonographic examination ranges from moderate to good.^11,27,46 Third, sonographic examinations were performed exclusively during neutral rotation of the shoulder. While most instability events occur during marked external rotation, applying anterior force at high-risk shoulder positions can result in glenohumeral dislocation. To mitigate this risk, our measurement protocol excluded external rotation. Fourth, although the pushing force applied to the humerus via the ultrasound probe from the posterior side may have been slight, this force could not be entirely controlled and may have introduced some degree of variability in the measurements. This is recognized as a potential limitation, as differences in the probe force could affect the relative position of the humeral head and AHHT measurements. Fifth, this study did not investigate the use of specific anesthetic agents, including antinociceptive agents, sedatives, and muscle relaxants, during general anesthesia. While the collaboration of anesthesiologists ensured adequate anesthesia during both pre- and postoperative measurements, variations in the depth of anesthesia between these time points may have influenced the measurement values. Sixth, measurements were obtained immediately pre- and postoperatively without a long-term follow-up. While this design allows for a direct comparison of surgical effects, it limits the ability to assess the durability of surgical outcomes and long-term effects on shoulder stability and function. Additionally, the measurements were conducted under general anesthesia and an interscalene block, which limited their postoperative applicability during follow-up. To fully understand the implications of surgical interventions, further research is warranted to investigate the correlation between ultrasonographic assessments during surgery and long-term clinical outcomes postoperatively. Finally, the inclusion criteria introduced selection bias, as the study only included patients with <20% bone loss in the scapular glenoid and who were not participating in collision sports. This may limit the applicability of our findings to all patients with anterior shoulder instability.

Conclusion

Arthroscopic Bankart repair combined with remplissage significantly reduced AHHT at 90° of abduction, whereas Bankart repair alone showed no significant decrease in AHHT at any abduction angle. However, both Bankart repair alone and Bankart repair with remplissage caused a posterior shift of the humeral head without load at 0°, 45°, and 90° of abduction. Measuring AHHT could be critical for assessing postoperative anterior shoulder instability, as it provides the surgeon with objective data to evaluate the procedure's success and guide postoperative management. Future research could investigate clinical differences based on the measured AHHT, potentially providing valuable insights for improving patient outcomes.