Arthroscopic Bankart Repair With Hill‐Sachs Remplissage Shows Low Recurrence and Excellent Outcomes for Primary and Revision Surgery at Midterm Follow‐Up Authors: Yash Sewpaul, Joseph DiCecco, Mustafa S. Rashid, Stephen S. Burkhart, Robert U. Hartzler Published: 2026 ABSTRACT PurposeTo assess midterm outcomes, including recurrence and patient‐reported measures, after arthroscopic Bankart repair with remplissage in both primary and revision cases.MethodsA retrospective, single‐surgeon review was performed on patients who underwent arthroscopic Bankart repair with remplissage between 2012 and 2016 with a minimum 7‐year follow‐up. The primary outcome was self‐reported recurrence of anterior instability (subluxation or dislocation). Secondary outcomes included Western Ontario Shoulder Instability Index, range of motion, American Shoulder and Elbow Surgeons score, Single Assessment Numeric Evaluation, Simple Shoulder Test, and pain rating.ResultsOf the 22 eligible patients, 13 (59%) with a mean of 13 preoperative instability events were included (10 primary, 3 revision). Mean age was 32 years, with an average follow‐up of 9.5 years. One patient (33%) in the revision group experienced a traumatic subluxation; no patients required revision surgery. Compared with primary cases, revision cases had significantly smaller glenoid track size (19.7 vs 23.9 mm, P = .02) and deeper Hill‐Sachs defects (4.5 vs 3 mm, P = .03). Final Western Ontario Shoulder Instability Index scores were similar between groups (192.2 vs 193.3, P = .99). American Shoulder and Elbow Surgeons scores improved from 67.2 to 93.2 overall (P < .001), with higher scores in revisions (98.0 vs 91.7, P = .02). Single Assessment Numeric Evaluation (89 vs 86, P = .62), Simple Shoulder Test (11.5 vs 11.6, P = .69), and visual analog scale (0.8 vs 0, P = .01) were not significantly different. Nine athletes returned to sport. No complications were reported.ConclusionsIn this study, we found significant improvements in clinical and functional outcomes in both the primary and revision arthroscopic Bankart repair with remplissage groups, with only 1 patient in the revision group experiencing a traumatic instability event at the final follow‐up. No patients in the primary group experienced further instability events. No patients underwent further instability surgery.Level of EvidenceLevel IV, retrospective therapeutic case series. Arthroscopic Bankart repair (ABR) is a common procedure for the treatment of anterior shoulder instability; however, recurrence rates are high if follow‐up beyond 2 years is assessed. 1 , 2 The addition of a posterior musculocapsulotenodesis “Hill‐Sachs remplissage” is often added (ABR with remplissage [ABR + R]) in instability cases with an associated Hill‐Sachs defect because it has been shown to reduce recurrent instability. 3 Yet, indications for ABR + R versus alternative instability procedures remain undefined, 4 , 5 especially given varying failure rates of primary and revision procedures. 6 , 7 Although there is a paucity of published mid‐ to long‐term outcomes, existing results suggest a varying level of clinical success with the procedure. 8 , 9 Recurrence after arthroscopic soft‐tissue stabilization continues to rise for up to 7 years postoperatively. 10 , 11 Despite the growing interest in arthroscopic Hill‐Sachs remplissage, there is a further scarcity of published mid‐ to long‐term outcomes on the procedure as an adjunct to ABR. 11 , 12 Thus, the success of the procedure with respect to preventing recurrent dislocation remains relatively unknown. The purpose of this study was to assess midterm outcomes, including recurrence and patient‐reported measures, after ABR + R in both primary and revision cases. We hypothesized that ABR + R shows durable clinical results with low rates of instability recurrence and no significant differences between primary and revision groups. METHODS Institutional Review Board approval was obtained prior to study commencement (WCG 20233706). The study is a single‐center, single‐surgeon retrospective case series of patients who underwent ABR with arthroscopic Hill‐Sachs remplissage between July 2012 and July 2016. The study period start date corresponds to the time point at which remplissage was routinely used with bipolar bone loss by the senior author as guided by the glenoid track concept of the Hill‐Sachs lesion engagement. The study period end date was selected to include patients with a minimum 7‐year follow‐up after study initiation. Patients were identified by a search of a large private orthopaedic practice's surgical database. Inclusion criteria were primary or revision surgery with ABR and Hill‐Sachs remplissage for anterior shoulder instability with associated Hill‐Sachs defect in adult (≥18 years of age) subjects. Patients who underwent concomitant glenoid bone augmentation via coracoid or free bone block transfer, those unable to be contacted, and those who refused to participate in the retrospective follow‐up were excluded. All patients underwent preoperative computed tomography (CT) or magnetic resonance imaging (MRI) assessment. Images were assessed for glenoid bone loss (%), glenoid track size (mm), Hill‐Sachs interval (mm), Hill‐Sachs depth (mm), and Hill‐Sachs lesion status (on/off track). To calculate glenoid bone loss in patients who underwent MRI (n = 3), the linear method based on the perfect circle technique was used. 13 The linear method based on uninjured glenoid width was used in patients who underwent bilateral CT (n = 10). 14 Imaging measurements were performed by the senior author (S.S.B.), who routinely performed these in clinical practice. The final decision to perform ABR + R was always made intraoperatively but guided by the glenoid track concept of the Hill‐Sachs lesion engagement risk and considering specific patient risk factors. 14 Surgical Technique All procedures were performed by a single fellowship‐trained shoulder surgeon (S.S.B.). Surgery was performed under general anesthesia with a suprascapular nerve block in a lateral decubitus position. Examination under anesthesia was then performed to assess the severity of instability. Diagnostic arthroscopy was performed through a standard posterior portal using a 30° and 70° arthroscopes and through an anterosuperolateral portal (30° arthroscope). Glenoid bone loss, glenoid track size, Hill‐Sachs interval, Hill‐Sachs depth, and Hill‐Sachs lesion status (on/off track) were measured arthroscopically. The Bankart lesion was mobilized and bone bed prepared. The inferior glenoid anchors were placed and sutures passed but left untied. A mean of 3.3 ± 0.5 anchors were used for the Bankart repair. Subacromial bursoscopy was performed. The Hill‐Sachs lesion was prepped to healthy bone, and 2 anchors were placed at the medial margin, superior and inferior. The remplissage sutures were passed through the cuff retrograde but left untied until after completion of the Bankart repair. Finally, the remplissage sutures were visualized in the subacromial space and tied between anchors (double pulley technique) to bring the musculocapsulotendinous tissue down into the Hill‐Sachs defect. 15 , 16 All anchors used were 3.0 mm BioComposite SutureTak anchors (Arthrex, Naples, FL). Postoperative rehabilitation followed a standard rotator cuff repair protocol: 6‐week sling immobilization and delay of active overhead range of motion, strengthening, and external rotation beyond half‐normal (contralateral side) until 3 months after surgery. 17 Clinical Evaluation Data collection was performed by the first author (Y.S.), independent of the operating surgeon (S.S.B.). The primary outcome was the rate of self‐reported recurrence of instability (dislocation/subluxation) via an electronic survey instrument. Secondary outcome measures included self‐reported bilateral active shoulder range of motion and the following patient‐reported outcome measures: Western Ontario Shoulder Instability Index (WOSI), American Shoulder and Elbow Surgeons (ASES) score, Simple Shoulder Test (SST) score, Single Assessment Numeric Evaluation (SANE), numeric pain rating (0‐10), reoperations, and return to sport. Statistical Methods Statistical analysis was performed using Jamovi software (The jamovi project, 2024; jamovi, Version 2.5 [Computer Software]; retrieved from https://www.jamovi.org.). Mean and standard deviations were calculated for all continuous outcomes. The primary analysis was measuring the rate of self‐reported instability recurrence. Secondary analysis included comparison of pre‐ and final follow‐up patient reported outcome data. Paired Student's t‐tests were used to compare means of the patient reported outcomes between pre‐ and postoperative outcome data. Independent t‐tests were used to compare outcomes between primary versus revision groups. Statistical significance was defined as P < .05. Categorical data was compared using the Fisher exact test. Cohort‐specific minimal clinically important difference (MCID) values were defined as one‐half of the standard deviation of the change score (Δ = postoperative − preoperative). MCID achievement was assessed at the patient level and reported as the proportion of patients meeting or exceeding the threshold. Between‐group comparisons were performed using nonparametric testing. Patient acceptable symptom state was assessed using a specific binary postoperative satisfaction question. Substantial clinical benefit analysis was not performed because of the absence of an appropriate anchor. RESULTS Twenty‐two patients were identified as eligible for the study. Follow‐up was available for 13 patients (58%). The mean follow‐up period was 9.5 years (7‐11). Mean age at remplissage was 32 years (range 20‐45). Revision ABR with Hill‐Sachs remplissage was performed in 3 (23%) patients. Baseline characteristics were similar between groups (Table 1). Mean number of prior instability events (13) did not differ between groups (P = .74). TABLE 1Study Population Demographics Parameter All Patients Primary Revision P Total13103Mean age31.5 ± 7.930.1 ± 7.736.3 ± 8.1.27Sex (M/F)13/010/03/0.05Overhead/contact sports4 (30.7%)4 (40%)0 (0%).49Mean number of instability episodes12.5 ± 15.213 ± 17.411 ± 4.58.74Preoperative patient reported outcomes ASES67.2 ± 20.274.1 ± 15.544 ± 18.1.07 SST7.5 ± 2.58.2 ± 2.55.3 ± 2.5<.05 SANE55.1 ± 25.063.6 ± 19.726.7 ± 20.8.08 Numeric pain score (0‐10)4 ± 3.13.6 ± 3.25.3 ± 3.2.43Preoperative range of motion Forward elevation, °179.2 ± 2.8179 ± 3.2180 ± 0.34 External rotation, °60.4 ± 22.862.8 ± 24.153.3 ± 20.8.62 Internal rotation (spinal levels)T11 ± 2T11 ± 2T12 ± 0.5.19ASES, American Shoulder and Elbow Surgeons; F, female; M, male; SANE, Single Assessment Numeric Evaluation; SST, Simple Shoulder Test.© 2026 WILEY‐VCH GmbH Study Population Demographics ASES, American Shoulder and Elbow Surgeons; F, female; M, male; SANE, Single Assessment Numeric Evaluation; SST, Simple Shoulder Test. Three patients had preoperative MRI and 10 patients had preoperative bilateral CT. Mean preoperative glenoid track size and Hill‐Sachs lesion depths were significantly different between primary and revision groups when measured on preoperative imaging (Table 2). There was no significant difference between mean preoperative glenoid bone loss (15% vs 10%, P = .2) or off‐track lesions (40% vs 67%, P = .56). Off‐track lesions were different by CT measurement versus intraoperative measurement (30.7% vs 46.2%), particularly for revision cases (0% vs 66.7%) (Table 2). Five patients (50%) in the primary group underwent concomitant SLAP repair, 1 patient (10%) underwent biceps tenodesis for a long head of the biceps tear, and 1 patient (10%) had additional posterior labral repair. In the revision group, 1 patient underwent a concomitant SLAP repair, and 1 patient underwent a biceps tenodesis for long head of biceps tear. TABLE 2Preoperative and Intraoperative Imaging Data Parameter All Patients Primary Revision P Preoperative imaging Glenoid bone loss, %13.3 ± 8.114.5 ± 8.89.5 ± 1.2.2 Glenoid track size, mm20.7 ± 3.119.7 ± 2.923.9 ± 1.1.02 Hill‐Sachs interval on axial cut, mm17.5 ± 4.117.6 ± 4.517.3 ± 2.1.89 Hill‐Sachs depth, mm4.4 ± 1.94.5 ± 2.03 ± 0.04 Off‐track Hill‐Sachs lesion4 (30.7%)4 (40%)0 (0%).49Intraoperative findings Glenoid bone loss, %12.3 ± 5.112.6 ± 5.511 ± 1.4.45 Glenoid track size, mm18.6 ± 3.518.3 ± 3.223 ± 1.4.03 Hill‐Sachs interval on axial cut, mm20.6 ± 4.520.6 ± 5.220.3 ± 1.4.82 Hill‐Sachs depth, mm3.5 ± 0.94 ± 13 ± 0.23 Off‐track Hill‐Sachs lesion6 (46.2%)4 (40%)2 (66.7%).56 Engaging Hill‐Sachs lesion5 (38.5%)4 (40%)1 (33.3).99 Number of anchors6 ± 1.66.2 ± 1.85.6 ± 1.2.56© 2026 WILEY‐VCH GmbH Preoperative and Intraoperative Imaging Data One patient (33%, revision group) reported a single instability event (traumatic subluxation), and no other patients experienced a further redislocation or subluxation event. One patient (10%) in the primary group underwent subsequent surgery for long thoracic nerve decompression. No revision instability surgery was performed. Analysis of patient‐reported outcome measures shows improvements in WOSI, ASES, SST, and visual analog scale (VAS) scores from preoperative to postoperative scores (Table 3). There was no significant difference between mean WOSI scores between groups (192 vs 193, P = .98). Mean ASES scores increased from 67 preoperatively to 93 postoperatively (P < .001). A significant difference was seen between final ASES scores for primary versus revision surgery (92 vs 98.0, P = .02). No differences were seen for SANE (89 vs 87, P = .71), SST (11.5 vs 11.6, P = .71), or VAS (0.8 vs 0, P = .12) scores. Using a cohort‐specific MCID threshold, values calculated were 1.53 for VAS, 1.17 for SST, 13.34 for SANE, and 11.0 points for ASES. MCID was achieved by 10 out of 13 patients for both VAS and SST (76.9%, 95% confidence interval: 49.7%‐91.8%). Nine out of 13 patients achieved MCID for both SANE and ASES (69.2%, 95% confidence interval: 42.4%‐87.3%). Postoperative MCID achievements for SST, ASES, VAS, and SANE scores did not differ significantly between primary and revision cohorts (all P > .1). Patient acceptable symptom state was achieved by all 13 (100%) patients in both the primary and revision cohorts. All athletes (N = 9) were able to return to sports, with 89% of patients able to return at the same level as their preinjury state. TABLE 3Self‐Reported Postoperative Outcomes Outcome All Patients Primary Revision P WOSI192.5 ± 130.2192.2 ± 144.4193.3 ± 80.7.98ASES93.2 ± 6.191.7 ± 6.398 ± 1.7.02SST11.5 ± 0.711.5 ± 0.711.6 ± 0.6.69SANE88 ± 0.989 ± 9.986 ± 5.8.62Numeric pain score (0‐10)0.6 ± 0.80.8 ± 0.80 ± 0.01Return to sport9 (69.2%)7 (70%)2 (66.7%).99Postoperative range of motion Forward elevation176.9 ± 4.8178 ± 4.2173.3 ± 5.8.29 External rotation60.8 ± 23.665 ± 23.246.6 ± 23.1.31 Internal rotationT9 ± 0.9T10 ± 0.7T8 ± 1.7.73ASES, American Shoulder and Elbow Surgeons; SANE, Single Assessment Numeric Evaluation; SST, Simple Shoulder Test; WOSI, Western Ontario Shoulder Instability Index.© 2026 WILEY‐VCH GmbH Self‐Reported Postoperative Outcomes ASES, American Shoulder and Elbow Surgeons; SANE, Single Assessment Numeric Evaluation; SST, Simple Shoulder Test; WOSI, Western Ontario Shoulder Instability Index. Self‐reported postoperative range of motion is summarized in Table 3. No significant difference was observed between surgeon‐assessed preoperative range of motion and self‐assessed postoperative range of motion in forward elevation (P = .19), internal (P = .73), and external rotation (P = .86). At the final follow‐up, self‐reported external rotation was reduced by a mean of 18.5° on the operated shoulder compared with the contralateral shoulder. Postoperative range of motion did not differ between the primary and revision groups. DISCUSSION Our findings show that postoperative clinical outcomes in patients who underwent both primary and revision ABR + R improved significantly from preoperative measures. Because only 1 patient (33%) in the revision group experienced a single further traumatic instability event of minimal clinical significance, these findings support that ABR with arthroscopic Hill‐Sachs remplissage can provide satisfactory mid‐ to long‐term stability even in the face of bipolar bone loss and prior failed surgery. Additionally, at the final follow‐up, patients rated their shoulders as being in excellent functional condition (WOSI score < 500). One of the most consistent findings of the current shoulder instability literature is that recurrence rates continue to increase for arthroscopic soft tissue procedures until around 7 years postoperatively. 10 Thus, the current study design included minimum follow‐up at this time point with overall follow‐up close to the longer‐term time point of 10 years after surgery. At longer term follow‐up, the reported rate of recurrent instability after ABR is 13.3% to 82.2%. 18 The recurrence rate and clinical outcomes found in the systematic review and meta‐analysis by Meyer et al. contrast the findings of our study. We reveal a negligible revision rate (0%) compared with 4.8% to 42.9% and lower instability recurrence rates. Patient‐reported outcome measures in the systematic review included Rowe, Constant, and Subjective Shoulder Value scores, in which increases were seen in all domains postoperatively, similar to the findings of this current study. The Latarjet procedure is an alternative procedure with a lower instability recurrence rate than ABR (0%‐22.6%). However, it is associated with high complication and reoperation rates. 18 A systematic review of 89 studies and 7175 shoulders showed a 6.1% failure rate for an open procedure and 6.8% for the arthroscopic procedure. 19 Graft complications were the most common cause of failure in both the open (1.9%, nonunion) and arthroscopic (3.2%, graft fracture) groups. Reoperation rates, especially for screw‐related complications, are reported in the 3% to 7% range. 20 , 21 , 22 , 23 Thus, ABR + R emerges as a compelling alternative in both the primary and revision setting, given the excellent clinical outcomes and low recurrence rate as seen in this study without the risk of complications and reoperations seen in the Latarjet procedure. Variable rates of instability recurrence after remplissage are reported in the literature. Long‐term data remains limited. Brejuin et al. revealed a failure rate of 15% in their series of 51 patients with a minimum follow‐up of 5 years. 9 Younger patients and deeper Hill‐Sachs lesions were predictive of redislocation. In the setting of bipolar defects where glenoid bone loss was ≤20%, Martinez‐Catalan et al. found a recurrence rate of 9.3% and a revision surgery rate of 13.9%. 24 Lower age and higher ISIS scores were associated with instability recurrence. Garcia et al. observed an 11.8% redislocation rate in a series of 50 patients who underwent Bankart repair with remplissage; 81% of patients were able to return to their previous level of sport. 25 Brilakis et al. showed a 5.6% failure rate over a mean follow‐up of 8.1 years in a cohort of 51 patients, with 71% of patients able to return to sport. 26 Horinek et al. found a 91.5% return to sport rate in overhead or contact athletes, compared with 72.7% in Latarjet patients (P = .007). 7 Postoperative loss of external rotation has been explored previously in the literature. Biomechanical studies suggest that remplissage reduces external rotation by approximately 15°, with the degree of restriction proportional to the size of the Hill‐Sachs lesion. 27 , 28 , 29 Elkinson et al. found that in small Hill‐Sachs defects (≤15%), adding remplissage to a Bankart repair significantly decreased external rotation by 15° compared with the intact condition. 28 In moderate Hill‐Sachs defects, remplissage decreased external rotation by 19.5° compared with Bankart repair alone. Omi et al. found that the decrease in range of movement was proportional to the size of the Hill‐Sachs lesion when filled with remplissage. 27 Clinical studies have also shown a loss of external rotation postoperatively; limitations in external rotation have been associated with larger Hill‐Sachs interval, shorter postoperative follow‐up, and use of 2 anchors for musculocapsulotenodesis. 30 No significant difference between preoperative and postoperative external rotation was seen in our study (60.4 vs 60.8, P = .86); however, the current series shows an average self‐reported loss of external rotation of nearly 20°. Postoperative osteoarthritis, capsulotenodesis healing and fatty infiltration of the rotator cuff are further variables that could be visualized on postoperative MRI in future studies to complement the clinical findings of this study There is a paucity of data of mid‐to‐long‐term outcomes of remplissage when performed as a revision procedure. In a case series of 31 shoulders, McCabe et al. report a 36% failure rate in the revision remplissage group compared with 0% in the primary remplissage group (P = .01). 6 All failures in the revision group (4 patients) were because of trauma, and no further surgery was performed. Postoperative Rowe scores were significantly less in the revision group than in the primary group (P = .009). Their study is limited by a mean follow‐up period of 41 months. Sinha et al. performed revision Bankart repair with remplissage in 55 patients who had failed previous Bankart repair. 31 Four cases remained unstable after revision surgery. They found a significant increase in clinical outcome scores and an 83% return to sport rate within 24 months. Our study complements the existing literature of mid‐to‐long term outcomes of remplissage as a revision procedure. Although our revision cohort is small (3 patients) and a 33% recurrent instability rate is notable, no dislocation or revision instability surgery was required with excellent final outcomes in this limited group. Revision remplissage may yet provide durable long‐term outcomes; however, further research is necessary to validate these findings. Limitations This study has several limitations. The small cohort size with high loss to follow‐up, the subjects only being men, and the retrospective design limit the generalizability of the findings. As noted, the revision group is so small that type 2 errors may be present in the comparative statistical analysis. Additionally, a single‐surgeon cohort further limits generalizability. As only 1 patient experienced postoperative instability, we were not able to perform any analysis of associated risk factors. Lastly, the lack of postoperative imaging to complement the clinical findings further reduces the analysis of this study. CONCLUSIONS In this study, we found significant improvements in clinical and functional outcomes in both the primary and revision ABR + R groups, with only 1 patient in the revision group experiencing a traumatic instability event at final follow‐up. No patients in the primary group experienced further instability events. No patients underwent further instability surgery. DISCLOSURES The authors (M.S.R., S.S.B., R.U.H.) declare the following financial interests/personal relationships which may be considered as potential competing interests: M.S.R. reports a relationship with Arthrex GmbH, Smith & Nephew, and Enovis that includes speaking and lecture fees. S.S.B. reports a relationship with Arthrex GmbH that includes consulting or advisory and Wolters Kluwer that includes book royalties. R.U.H. reports a relationship with Stryker that includes consulting or advisory. The other authors (Y.S., J.D.) declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this article.