Comparison of Bankart repair with remplissage and Latarjet procedure for anterior shoulder instability: a systematic review and meta-analysis Authors: Kongying Nong, Gaoming Liu, Shide Jiang, Qilin Li, Bosomtwe Richmond Kwabena, Dongliang Yuan, Wenfeng Xiao, Kai Zhang, Shuguang Liu, Yusheng Li Published: 2026 ABSTRACT BackgroundAnterior shoulder instability is the most common type of shoulder instability. For cases with subcritical glenoid bone loss, both Bankart repair with remplissage (BR) and the Latarjet procedure are viable surgical options. The aim of this study was to compare the postoperative outcomes in patients undergoing BR and Latarjet procedure.MethodsThis review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline. The methodological index for nonrandomized studies (MINORS) was used to characterize the methodological quality. A literature search was conducted using PubMed, the Cochrane Library, Embase, Web of Science, and Scopus. The analyzed outcomes included postoperative Rowe score, visual analog scale (VAS) score, the Single Assessment Numeric Evaluation (SANE) score, range of motion (ROM), return-to-sport (RTS) rate, RTS time, recurrence rate, revision rate, and complications. Subgroup analysis was performed on the basis of follow-up duration.ResultsOverall, 12 studies involving 1186 patients were included. The meta-analysis showed that BR was associated with a significantly lower complication rate than the Latarjet procedure. No significant differences were observed in functional outcomes, ROM, recurrence, revision, RTS, or RTS time. Subgroup analysis showed a significant interaction between follow-up duration and outcomes, particularly for VAS and Rowe scores, with a trend toward better short-term results for BR and better longer-term results for Latarjet.ConclusionsBR and the Latarjet procedure provide comparable clinical outcomes for anterior shoulder instability, with BR demonstrating a significantly lower complication rate. Follow-up duration may influence VAS and Rowe outcomes, suggesting a potential temporal shift in treatment effects.Level of evidence: Level III.Trial registration: CRD42024582343.Supplementary InformationThe online version contains supplementary material available at 10.1186/s10195-026-00932-8. Introduction Anterior glenohumeral instability is the most common form of shoulder instability, accounting for over 95% of cases, with an estimated annual incidence of 0.219‰ in the general population [1–3]. The economic costs of shoulder instability are burdensome as research shows costs for surgical procedures ranging from US $13,672 to US $21,389 [4, 5]. It commonly occurs in a position of excessive abduction and external rotation, resulting in labral tears and capsuloligamentous injuries [6]. Given the high recurrence rate after conservative treatment, particularly in young and active patients, surgical management is often preferred [7, 8]. Available surgical techniques are broadly classified into soft-tissue repairs and bony reconstruction procedures. Bankart repair is a method of reconstructing and fixing the glenoid labrum–capsule ligament complex by using suture anchors or absorbable screws. It is traditionally indicated for glenoid bone loss (GBL) below the critical threshold (20–25%), although recent evidence suggest the cutoff may be lower [9, 10]. Bankart repair with remplissage (BR) is classically used for off-track Hill–Sachs lesions [11], but its indications have expanded to include on-track lesions with subcritical bone loss [12, 13]. The procedure involves arthroscopic posterior capsulodesis and infraspinatus tenodesis to fill the defect in the humeral head. This converts the defect into an extra-articular structure and removes its articular surface. These changes prevent engagement with the anterior glenoid rim and ultimately reduce the risk of recurrence compared with isolated Bankart repair [14, 15]. However, the increased posterior capsular tension may alter joint kinematics and limit external rotation [16]. The Latarjet procedure, performed via open or arthroscopic techniques, is the most commonly used bony procedure for patients with critical GBL or large bony defects. It restores anterior stability by augmenting glenoid width with a coracoid bone graft and providing a dynamic sling effect from the transferred conjoint tendon. In addition to its traditional indications, it is also used in selected patients with off-track Hill–Sachs lesions without critical bone loss [9, 17, 18]. Clinical outcomes appear comparable between the open and arthroscopic approaches. While the procedure effectively increases stability and reduces recurrence [19, 20], preliminary data indicate relatively high complication rates, with overall complication rates of 8.9% for the arthroscopic Latarjet and 4.9% for the open Latarjet [21]. Clinical overlap exists in patients with subcritical GBL and significant off-track Hill–Sachs lesions, for whom either BR or Latarjet may be considered. However, evidence directly comparing the two procedures in this context is limited. A prior review conducted by Haroun et al. [22] compared the two procedures but was limited by the scarcity of relevant literature and small sample sizes, making it difficult to detect significant differences in outcomes. Similarly, the descriptive review by Casper et al. [23], which was conducted by a single researcher, lacked a comprehensive meta-analysis and was prone to subjectivity and bias. These limitations highlight the need for a more comprehensive evaluation to determine the optimal treatment approach. The purpose of this study is to compare the clinical outcomes of BR and the Latarjet procedure, providing new insights into the treatment of anterior shoulder instability. Materials and methods The study was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline [24]. The study protocol was registered on The International Prospective Register of Systematic Reviews (PROSPERO). Search strategy and study selection Related studies were comprehensively searched using the search terms “shoulder dislocation,” “Bankart repair,” “remplissage,” and “Latarjet” through PubMed, Embase, the Cochrane Library, Web of Science, and Scopus from 1 January 2014 to 5 April 2026. Two investigators independently screened the retrieved literature on the basis of titles and abstracts. A full-text search of literature was conducted whenever one of the investigators considered it potentially relevant. Any disagreements were resolved through discussion with a third investigator. Eligibility criteria Eligible studies encompassed original clinical research publications in English, including randomized controlled trials, cohort studies, case–control studies, and case series. Studies were required to report at least one relevant clinical outcome. Studies that did not include any clinical outcomes, or reported them without providing extractable data, were excluded. Outcomes The primary outcome was recurrence rate. Secondary outcomes included the Rowe score, the Single Assessment Numeric Evaluation (SANE) score, range of motion (ROM), return-to-sport (RTS) rate, revision rate, and complications. The visual analogue scale (VAS) and time to RTS were also collected when available but were not considered major comparative endpoints owing to heterogeneity in reporting. Data extraction Two investigators independently extracted all pertinent data related to study characteristics using a standardized form, and any discrepancies were resolved through discussion with a third investigator. General study information collected included first author, publication year, study design, surgical technique, sample size, participant characteristics, and mean follow-up duration (months). Preoperative lesion details, such as Hill–Sachs lesion depth and GBL, were recorded. Quality assessment The methodological index for nonrandomized studies (MINORS) was used to characterize the quality of included studies [25]. The MINORS scale has 12 items, the first 8 for noncomparative and comparative studies and the last 4 for comparative studies only. Each item is scored as 0, 1, or 2. We divided methodological quality into three levels as follows: a score of 0–9 or 0–12 was considered poor quality, 10–12 or 13–18 was considered moderate quality, and 13–16 or 19–24 was considered excellent quality for noncomparative and comparative studies, respectively [26]. Two researchers independently assessed the quality of the literature. The intraclass correlation coefficient (ICC) was used to evaluate interrater reliability [27]. Statistical analysis All statistical analyses were performed using Stata (version 18) and R (version 4.5.1). Continuous variable data were described as mean with its standard deviation. Baseline differences between the BR and Latarjet groups were assessed descriptively using standardized differences, with standardized differences in proportions used for categorical baseline variables. For analysis of categorical variables, the effect of treatment was quantified by calculating the odds ratio (OR) and its associated 95% confidence interval (CI). A random-effects model was used for combining data, considering the inherent diversity among the studies included [28, 29]. The I2 statistic and Cochran’s Q test were conducted to estimate the heterogeneity between studies. Sensitivity analyses were conducted by sequentially omitting each study to identify influential studies and potential sources of heterogeneity. In addition, to explore the potential impact of indication bias related to GBL, a sensitivity analysis restricted to studies involving patients with mean GBL < 15% was performed. Subgroup analyses were performed on the basis of follow-up duration, with the requirement that each subgroup contain at least one study. Studies with a significant difference in follow-up period between the two groups were excluded. Statistical significance was set at P < 0.05. Results Literature selection A comprehensive search of the literature on BR and Latarjet procedures for anterior shoulder instability was conducted in the following databases: PubMed (523), Web of Science (935), Embase (588), Cochrane (23), and Scopus (701). After removing 1009 duplicate records, the retrieved articles were screened according to the predefined inclusion and exclusion criteria, and 36 articles passed the preliminary screening. Subsequently, articles were further evaluated by full-text review, which led to the exclusion of three studies owing to incomplete outcome data [30–32]. Other studies were excluded for not providing comparison between the BR and Latarjet. Two studies [33, 34] were published by the same group and had overlapping patient cohorts; therefore, only the study with the more comprehensive dataset was included in the quantitative synthesis. Consequently, a total of 12 articles were included for qualitative analysis (Fig. 1).Fig. 1Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart General information and characteristics of included studies Of the included studies, 645 patients underwent the Latarjet procedure, and 541 patients underwent the BR procedure. The calculated mean age was 27.1 years for the BR group and 28.5 years for the Latarjet group. The mean follow-up periods for the BR and Latarjet procedure were 42.4 and 46.7 months, respectively. Baseline characteristics showed some differences between the two groups. The mean GBL was 10.3 ± 7.6% in the BR group and 11.9 ± 8.1% in the Latarjet group. The mean Hill–Sachs lesion depth was 7.9 ± 2.6 mm in the BR group and 6.0 ± 3.2 mm in the Latarjet group. The proportion of patients with off-track lesions was higher in the BR group than in the Latarjet group. The patient characteristics are summarized in Tables 1 and 2. Across the included studies, BR was consistently performed arthroscopically, whereas Latarjet was primarily performed using an open approach, with occasional use of arthroscopic techniques. Details of the surgical techniques used in each included study are summarized in Supplementary Material 1. Table 1Demographic information and characteristics of the included studiesAuthor (year)Study designInterventionSample sizeSex (M/F)Mean age, yearsGlenoid bone loss (%)Hills–Sachs lesion depth (mm)Follow-up (months)Main outcomeLesion classificationAbouelsoud M.M., 2015 [37]Prospective comparative studyBR1629/328.2––31.31 ± 3.87Rowe, VAS, recurrence rateAll engagingL16––All engagingCho N.S., 2015 [38]Retrospective comparative studyBR3734/324.8 ± 9.08.5 ± 5.86.8 ± 1.724.7 ± 9.5Rowe, FF, ERs ERa, IR, recurrence rate, complicationsAll engagingL3532/327.8 ± 7.99.8 ± 6.16.4 ± 2.430.4 ± 11.2All engagingBah A., 2017 [35]Retrospective comparative studyBR438/3524.25 ± 6.45–7.7 ± 1.247.3 ± 9.07Rowe, VAS, ERs, ERa, IR, recurrence rate, revision rate–L437/36–7.9 ± 1.5–Yang J.S., 2018 [36]Retrospective comparative studyBR9894/428.3 ± 1010.4 ± 6.8–38.4 ± 22.09VAS, SANE, FF, ERs, ERa, IR, recurrence rate, revision rate, complicationsAll off-trackL9186/530 ± 12.112.3 ± 8.79–All off-trackSaremi H., 2021 [41]Retrospective comparative studyBR68–27 ± 6.33––74.4 ± 12.11VAS–L26–––Horinek J.L., 2022 [33]Retrospective comparative studyBR7060/1026.7 ± 7.312.3 ± 10.98.7 ± 3.830 ± 5.32SANE, VAS, FF, ERs, RTS, recurrence rate, complications16/70 off-trackL188157/3126.8 ± 9.97.6 ± 9.03.9 ± 2.624/188 off-trackHurley E.T., 2022 [42]Retrospective comparative studyBR4128/1329.6 ± 10.87.3 ± 7.8–58.5 ± 24VAS, RTS, RTS time, recurrence rate, complicationsAll engagingL2621/532.3 ± 12.719.1 ± 4.7–52 ± 23.4All engagingMaiotti M., 2023 [39]Retrospective comparative studyBR6556/926.7 ± 8.1––25 ± 22.3Rowe, ERs, ERa, recurrence rate, complications–L10490/1430.7 ± 10––49.9 ± 21.2–Paul R.W., 2023 [43]Retrospective comparative studyBR2825/328.2 ± 8.811 ± 4–38.4 ± 22.8SANE, FF, ERa, RTS, RTS time, recurrence rate, revision rate18/22 off-trackL4336/729.8 ± 12.119 ± 6–40.8 ± 22.820/43 off-trackGereli A., 2024 [45]Retrospective comparative studyBR1612/429.6 ± 8.47.2 ± 2.310.1 ± 2.652 ± 12Rowe, SANE, VAS, FF, ERs, ERa, IR, RTS timeAll off-trackL1413/130.9 ± 5.111.4 ± 1.713.9 ± 4.152 ± 12All off-trackNelson B.A., 2025 [44]Retrospective comparative studyBR2523/223.8 ± 6.316.1 ± 3.251.3 ± 5.2SANE, recurrence rate, complications1/25 off-trackL2524/125.2 ± 5.119.3 ± 2.135.6 ± 12.43/25 off-trackLimam K., 2026 [40]Retrospective comparative studyBR3425/925 ± 7.55.72 ± 4.855 ± 2.4987 ± 21Rowe, FF, ERs, ERa, RTS, recurrence rate, revision rate, complications–L3422/1225 ± 4.85.37 ± 6.474.53 ± 2.3693 ± 30–BR Bankart repair combined with remplissage procedure, L Latarjet procedure, M male, F female, VAS visual analogue scale, SANE Single Assessment Numeric Evaluation, FF forward flexion, ERs external rotation with the arm at the side, ERa external rotation of the arm in abduction, IR internal rotation, RTS return to sport, Engaging engaging Hill–Sachs lesionTable 2Baseline characteristics of patients in the BR and Latarjet groups across included studiesVariableBR (n = 541)Latarjet (n = 645)Standardized differenceBR reported (N)Latarjet reported (N)Average age (years)27.1 ± 8.828.5 ± 10.2−0.15414560Average follow-up (months)42.4 ± 21.946.7 ± 22.5−0.13246281Male patients, n (%)369 (80.6)498 (85.4)−0.19458583Average GBL (%)10.3 ± 7.611.9 ± 8.1−0.20349456HSL depth (mm)7.9 ± 2.66.0 ± 3.2+0.66200314Patients with off-track lesion, n (%)149 (64.5%)152 (42.1%)+0.46231361Standardized differences were used to describe baseline imbalance between groups; for categorical variables, standardized differences in proportions were calculated. BR Bankart repair combined with remplissage procedure, GBL glenoid bone loss, HSL Hill–Sachs lesion Table 1Demographic information and characteristics of the included studiesAuthor (year)Study designInterventionSample sizeSex (M/F)Mean age, yearsGlenoid bone loss (%)Hills–Sachs lesion depth (mm)Follow-up (months)Main outcomeLesion classificationAbouelsoud M.M., 2015 [37]Prospective comparative studyBR1629/328.2––31.31 ± 3.87Rowe, VAS, recurrence rateAll engagingL16––All engagingCho N.S., 2015 [38]Retrospective comparative studyBR3734/324.8 ± 9.08.5 ± 5.86.8 ± 1.724.7 ± 9.5Rowe, FF, ERs ERa, IR, recurrence rate, complicationsAll engagingL3532/327.8 ± 7.99.8 ± 6.16.4 ± 2.430.4 ± 11.2All engagingBah A., 2017 [35]Retrospective comparative studyBR438/3524.25 ± 6.45–7.7 ± 1.247.3 ± 9.07Rowe, VAS, ERs, ERa, IR, recurrence rate, revision rate–L437/36–7.9 ± 1.5–Yang J.S., 2018 [36]Retrospective comparative studyBR9894/428.3 ± 1010.4 ± 6.8–38.4 ± 22.09VAS, SANE, FF, ERs, ERa, IR, recurrence rate, revision rate, complicationsAll off-trackL9186/530 ± 12.112.3 ± 8.79–All off-trackSaremi H., 2021 [41]Retrospective comparative studyBR68–27 ± 6.33––74.4 ± 12.11VAS–L26–––Horinek J.L., 2022 [33]Retrospective comparative studyBR7060/1026.7 ± 7.312.3 ± 10.98.7 ± 3.830 ± 5.32SANE, VAS, FF, ERs, RTS, recurrence rate, complications16/70 off-trackL188157/3126.8 ± 9.97.6 ± 9.03.9 ± 2.624/188 off-trackHurley E.T., 2022 [42]Retrospective comparative studyBR4128/1329.6 ± 10.87.3 ± 7.8–58.5 ± 24VAS, RTS, RTS time, recurrence rate, complicationsAll engagingL2621/532.3 ± 12.719.1 ± 4.7–52 ± 23.4All engagingMaiotti M., 2023 [39]Retrospective comparative studyBR6556/926.7 ± 8.1––25 ± 22.3Rowe, ERs, ERa, recurrence rate, complications–L10490/1430.7 ± 10––49.9 ± 21.2–Paul R.W., 2023 [43]Retrospective comparative studyBR2825/328.2 ± 8.811 ± 4–38.4 ± 22.8SANE, FF, ERa, RTS, RTS time, recurrence rate, revision rate18/22 off-trackL4336/729.8 ± 12.119 ± 6–40.8 ± 22.820/43 off-trackGereli A., 2024 [45]Retrospective comparative studyBR1612/429.6 ± 8.47.2 ± 2.310.1 ± 2.652 ± 12Rowe, SANE, VAS, FF, ERs, ERa, IR, RTS timeAll off-trackL1413/130.9 ± 5.111.4 ± 1.713.9 ± 4.152 ± 12All off-trackNelson B.A., 2025 [44]Retrospective comparative studyBR2523/223.8 ± 6.316.1 ± 3.251.3 ± 5.2SANE, recurrence rate, complications1/25 off-trackL2524/125.2 ± 5.119.3 ± 2.135.6 ± 12.43/25 off-trackLimam K., 2026 [40]Retrospective comparative studyBR3425/925 ± 7.55.72 ± 4.855 ± 2.4987 ± 21Rowe, FF, ERs, ERa, RTS, recurrence rate, revision rate, complications–L3422/1225 ± 4.85.37 ± 6.474.53 ± 2.3693 ± 30–BR Bankart repair combined with remplissage procedure, L Latarjet procedure, M male, F female, VAS visual analogue scale, SANE Single Assessment Numeric Evaluation, FF forward flexion, ERs external rotation with the arm at the side, ERa external rotation of the arm in abduction, IR internal rotation, RTS return to sport, Engaging engaging Hill–Sachs lesion Demographic information and characteristics of the included studies BR Bankart repair combined with remplissage procedure, L Latarjet procedure, M male, F female, VAS visual analogue scale, SANE Single Assessment Numeric Evaluation, FF forward flexion, ERs external rotation with the arm at the side, ERa external rotation of the arm in abduction, IR internal rotation, RTS return to sport, Engaging engaging Hill–Sachs lesion Table 2Baseline characteristics of patients in the BR and Latarjet groups across included studiesVariableBR (n = 541)Latarjet (n = 645)Standardized differenceBR reported (N)Latarjet reported (N)Average age (years)27.1 ± 8.828.5 ± 10.2−0.15414560Average follow-up (months)42.4 ± 21.946.7 ± 22.5−0.13246281Male patients, n (%)369 (80.6)498 (85.4)−0.19458583Average GBL (%)10.3 ± 7.611.9 ± 8.1−0.20349456HSL depth (mm)7.9 ± 2.66.0 ± 3.2+0.66200314Patients with off-track lesion, n (%)149 (64.5%)152 (42.1%)+0.46231361Standardized differences were used to describe baseline imbalance between groups; for categorical variables, standardized differences in proportions were calculated. BR Bankart repair combined with remplissage procedure, GBL glenoid bone loss, HSL Hill–Sachs lesion Baseline characteristics of patients in the BR and Latarjet groups across included studies Standardized differences were used to describe baseline imbalance between groups; for categorical variables, standardized differences in proportions were calculated. BR Bankart repair combined with remplissage procedure, GBL glenoid bone loss, HSL Hill–Sachs lesion Quality assessment The results of heterogeneity evaluation using the MINORS indicated that the mean scores of the included studies were 17.08 ± 1.14, reflecting a moderate quality. All studies scored high for the follow-up period, since all studies had a follow-up period longer than 24 months. As shoulder rehabilitation after surgery is completed in 4–6 months, 2 years or longer is an adequate follow-up duration. However, two studies [35, 36] reported a loss to follow-up more than 5%. Inter-rater reliability was excellent (ICC = 0.81; 95% CI 0.478–0.941). The detailed results of the MINORS assessment are presented in Supplementary Material 2 (Table 3). Table 3Summary of pooled clinical outcomes comparing BR and Latarjet proceduresOutcomeBRLatarjetEffect size (MD or OR)95% CII2 (%)P-valueRowe91.1993.60−0.70−3.52 to 2.1267.60.629VAS1.421.47−0.07−0.59 to 0.4574.80.794SANE89.3687.553.17−1.54 to 7.8876.00.187FF167.78169.800.29−2.71 to 3.2849.90.850ERs56.6665.19−3.22−10.74 to 4.3093.30.401ERa70.4976.31−3.91−14.65 to 6.8395.80.476IR52.0957.32−1.02−7.70 to 5.6679.30.765RTS76.15%70.53%1.530.86 to 2.730.00.147RTS time8.967.481.46−0.49 to 3.400.00.143Recurrence rate8.67%4.52%1.710.99 to 2.930.00.052Revision rate5.91%2.84%2.300.82 to 6.410.00.112Complication rate1.08%8.75%0.170.07 to 0.400.0< 0.001BR Bankart repair combined with remplissage procedure, VAS visual analogue scale, SANE Single Assessment Numeric Evaluation, FF forward flexion, ERs external rotation with the arm at the side, ERa external rotation of the arm in abduction, IR internal rotation, RTS return to sport Table 3Summary of pooled clinical outcomes comparing BR and Latarjet proceduresOutcomeBRLatarjetEffect size (MD or OR)95% CII2 (%)P-valueRowe91.1993.60−0.70−3.52 to 2.1267.60.629VAS1.421.47−0.07−0.59 to 0.4574.80.794SANE89.3687.553.17−1.54 to 7.8876.00.187FF167.78169.800.29−2.71 to 3.2849.90.850ERs56.6665.19−3.22−10.74 to 4.3093.30.401ERa70.4976.31−3.91−14.65 to 6.8395.80.476IR52.0957.32−1.02−7.70 to 5.6679.30.765RTS76.15%70.53%1.530.86 to 2.730.00.147RTS time8.967.481.46−0.49 to 3.400.00.143Recurrence rate8.67%4.52%1.710.99 to 2.930.00.052Revision rate5.91%2.84%2.300.82 to 6.410.00.112Complication rate1.08%8.75%0.170.07 to 0.400.0< 0.001BR Bankart repair combined with remplissage procedure, VAS visual analogue scale, SANE Single Assessment Numeric Evaluation, FF forward flexion, ERs external rotation with the arm at the side, ERa external rotation of the arm in abduction, IR internal rotation, RTS return to sport Summary of pooled clinical outcomes comparing BR and Latarjet procedures BR Bankart repair combined with remplissage procedure, VAS visual analogue scale, SANE Single Assessment Numeric Evaluation, FF forward flexion, ERs external rotation with the arm at the side, ERa external rotation of the arm in abduction, IR internal rotation, RTS return to sport Patient-reported outcome measures The Rowe score was reported in five studies [35, 37–40], including 195 BR procedures and 232 Latarjet procedures, with mean scores of 91.19 and 93.60, respectively. The VAS score was reported in six studies [33, 35, 36, 38, 41, 42], including 357 BR procedures and 409 Latarjet procedures, with mean scores of 1.42 and 1.47, respectively. The SANE score was reported in four studies [33, 36, 43, 44], including 221 BR procedures and 347 Latarjet procedures, with mean scores of 89.36 and 87.55, respectively. No significant differences were observed between the BR and Latarjet groups in Rowe score (MD, −0.70; 95% CI −3.52 to 2.12; I2 = 67.6%; P = 0.629), VAS score (MD, −0.07; 95% CI −0.59 to 0.45; I2 = 74.8%; P = 0.794), or SANE score (MD, 3.17; 95% CI −1.54 to 7.88; I2 = 76.0%; P = 0.187) (Fig. 2).Fig. 2Forest plots demonstrating the weighted mean difference for patient-reported outcome measures (PROMs) in BR group versus the Latarjet group (A forest plot for Rowe; B forest plot for VAS; C forest plot for SANE) Forest plots demonstrating the weighted mean difference for patient-reported outcome measures (PROMs) in BR group versus the Latarjet group (A forest plot for Rowe; B forest plot for VAS; C forest plot for SANE) ROM Forward flexion (FF) was reported in seven studies [33, 35, 36, 38, 40, 43, 45], including 326 BR procedures and 448 Latarjet procedures, with mean values of 167.78° and 169.80°, respectively. External rotation with the arm at the side (ERs) was reported in eight studies [33, 35, 36, 38–40, 43, 45], including 391 BR procedures and 552 Latarjet procedures, with mean values of 56.66° in the BR group and 65.19° in the Latarjet group. External rotation of the arm in abduction (ERa) was reported in six studies [35, 36, 38–40, 45], including 293 BR procedures and 321 Latarjet procedures, with mean values of 70.49° and 76.31°, respectively. Internal rotation (IR) was reported in four studies [35, 36, 38, 45], including 194 BR procedures and 183 Latarjet procedures, with mean values of 52.09° and 57.32°, respectively. No significant differences were observed between the BR and Latarjet groups in FF (MD, 0.29; 95% CI −2.71 to 3.28; I2 = 49.9%; P = 0.850), ERs (MD, −3.22; 95% CI −10.74 to 4.30; I2 = 93.3%; P = 0.401), ERa (MD, −3.91; 95% CI −14.65 to 6.83; I2 = 95.8%; P = 0.476), or IR (MD, −1.02; 95% CI −7.70 to 5.66; I2 = 79.3%; P = 0.765) (Fig. 3).Fig. 3Forest plots demonstrating the weighted mean difference for ROM in the BR group versus the Latarjet group (A forest plot for FF; B forest plot for ERs; C forest plot for ERa; D forest plot for IR) Forest plots demonstrating the weighted mean difference for ROM in the BR group versus the Latarjet group (A forest plot for FF; B forest plot for ERs; C forest plot for ERa; D forest plot for IR) RTS and return-to-sport time Return to sport (RTS) was reported in three studies [33, 42, 43], including 139 BR procedures and 257 Latarjet procedures. The RTS rate was 76.15% in the BR group (n = 83) and 70.53% in the Latarjet group (n = 134). RTS time was also reported in three studies [42, 43, 45], including 85 BR procedures and 83 Latarjet procedures, with mean times of 8.96 months and 7.48 months in the BR and Latarjet groups, respectively. No significant differences were observed between the two procedures in RTS rate (OR, 1.53; 95% CI 0.86 to 2.73; I2 = 0%; P = 0.147) or RTS time (MD, 1.46; 95% CI −0.49 to 3.40; I2 = 0%; P = 0.143) (Fig. 4).Fig. 4Forest plot demonstrating the weighted mean difference for RTS and RTS time in BR group versus the Latarjet group (A forest plot for RTS; B forest plot for RTS time) Forest plot demonstrating the weighted mean difference for RTS and RTS time in BR group versus the Latarjet group (A forest plot for RTS; B forest plot for RTS time) Recurrence rate and revision rate The recurrence rate was reported in ten studies [33, 35–40, 42–44], including 457 BR procedures and 605 Latarjet procedures. Recurrent instability occurred in 8.67% of patients in the BR group (n = 41) and 4.52% in the Latarjet group (n = 28). The revision rate was reported in four studies [35, 36, 40, 43], including 203 BR procedures and 211 Latarjet procedures, with revision occurring in 5.91% of patients in the BR group (n = 12) and 2.84% in the Latarjet group (n = 6). No significant differences were observed between the two procedures in recurrence rate (OR, 1.71; 95% CI 0.99 to 2.93; I2 = 0%; P = 0.052) or revision rate (OR, 2.30; 95% CI 0.82 to 6.41; I2 = 0%; P = 0.112) (Fig. 5).Fig. 5Forest plot demonstrating the weighted mean difference for rates of recurrence, revision, and complication in the BR group versus the Latarjet group (A forest plot for recurrence rate; B forest plot for revision rate; C forest plot for complication rate) Forest plot demonstrating the weighted mean difference for rates of recurrence, revision, and complication in the BR group versus the Latarjet group (A forest plot for recurrence rate; B forest plot for revision rate; C forest plot for complication rate) Complications The complication rate was reported in seven studies [33, 36, 38–40, 42, 44], comprising 370 BR procedures and 503 Latarjet procedures. The BR procedure resulted in complications in 1.08% of patients (n = 4), whereas the Latarjet procedure resulted in complications in 8.75% of patients (n = 40). Among studies reporting detailed complication types, only 2 events were observed in the BR group (1 infection and 1 hardware-related complication), compared with 27 events in the Latarjet group, mainly hardware-related, bone-related, and infections. There was a significantly lower complication rate in the BR group compared with the Latarjet group (OR, 0.17; 95% CI 0.07 to 0.40; I2 = 0%; P < 0.001) (Fig. 5). The pooled clinical outcomes comparing BR and the Latarjet procedure are summarized in Table 3. Subgroup analysis and sensitivity analysis Subgroup analysis indicated that follow-up duration may influence clinical outcomes, including both VAS and Rowe scores. In the short-term subgroup (2–3 years), BR demonstrated a trend toward better outcomes, with lower VAS scores (MD, −0.56; 95% CI −1.24 to 0.13) and higher Rowe scores (MD: 2.00; 95% CI −0.02 to 4.02). In contrast, in the longer-term subgroup (> 3 years), the Latarjet procedure showed a trend toward improved outcomes, with lower VAS scores (MD, 0.29; 95% CI −0.13 to 0.71) and higher Rowe scores (MD: −2.70; 95% CI −6.45 to 1.06). Although these differences within subgroups did not reach statistical significance, significant differences were observed between subgroups for both VAS (p = 0.039) and Rowe scores (p = 0.031), suggesting that follow-up duration may act as a potential effect modifier. Subgroup analyses are summarized in Table 4, and the corresponding forest plots are provided in Supplementary Material 3. Table 4Summary of subgroup analyses according to follow-up durationOutcomeStudies, nShort-termLonger-termP-value for subgroup differenceVAS6 −0.56 (−1.24 to 0.13)0.29 (−0.13 to 0.71)0.039Rowe score42.00 (−0.02 to 4.02)−2.70 (−6.45 to 1.06)0.031FF72.08 (−3.52 to 7.67)−0.54 (−4.76 to 3.68)0.465ERs7−9.90 (−30.46 to 10.67)−1.41 (−8.09 to 5.28)0.442Recurrence rate80.58 (0.15 to 2.22)1.97 (1.03 to 3.77)0.107Complication rate50.09 (0.01 to 0.70)0.14 (0.04 to 0.59)0.716Values are presented as pooled effect estimates with 95% CIs. MDs were used for continuous outcomes, and ORs were used for categorical outcomes. Short-term and longer-term subgroups were defined as follow-up durations of 2–3 years and > 3 years, respectively. VAS visual analogue scale, FF forward flexion, ERs external rotation with the arm at the side Table 4Summary of subgroup analyses according to follow-up durationOutcomeStudies, nShort-termLonger-termP-value for subgroup differenceVAS6 −0.56 (−1.24 to 0.13)0.29 (−0.13 to 0.71)0.039Rowe score42.00 (−0.02 to 4.02)−2.70 (−6.45 to 1.06)0.031FF72.08 (−3.52 to 7.67)−0.54 (−4.76 to 3.68)0.465ERs7−9.90 (−30.46 to 10.67)−1.41 (−8.09 to 5.28)0.442Recurrence rate80.58 (0.15 to 2.22)1.97 (1.03 to 3.77)0.107Complication rate50.09 (0.01 to 0.70)0.14 (0.04 to 0.59)0.716Values are presented as pooled effect estimates with 95% CIs. MDs were used for continuous outcomes, and ORs were used for categorical outcomes. Short-term and longer-term subgroups were defined as follow-up durations of 2–3 years and > 3 years, respectively. VAS visual analogue scale, FF forward flexion, ERs external rotation with the arm at the side Summary of subgroup analyses according to follow-up duration Values are presented as pooled effect estimates with 95% CIs. MDs were used for continuous outcomes, and ORs were used for categorical outcomes. Short-term and longer-term subgroups were defined as follow-up durations of 2–3 years and > 3 years, respectively. VAS visual analogue scale, FF forward flexion, ERs external rotation with the arm at the side To assess the robustness of the pooled data for the primary outcome, we conducted a leave-one-out sensitivity analysis by sequentially excluding individual studies. The results demonstrated that the pooled estimates for all outcomes remained stable. Exclusion of any single study did not materially alter the effect size or the statistical significance of the results, indicating that the findings were robust and not driven by any individual study (Supplementary Material 4). A sensitivity analysis including five studies [33, 36, 38, 40, 45] with a mean baseline GBL < 15% yielded results consistent with the primary analysis, with no changes in the direction or statistical significance of the pooled outcomes. The comprehensive results of the sensitivity analysis are presented in Supplementary Material 5. Discussion The present systematic review and meta-analysis quantitatively compared clinical outcomes between the BR and Latarjet procedures on the basis of the current literature. Overall, BR was associated with a lower complication rate, whereas no significant differences were observed in PROMS, ROM, recurrence, revision, RTS rates, or RTS time. Subgroup analyses suggested a potential influence of follow-up duration on postoperative outcomes, with a temporal trend favoring BR in the short term and the Latarjet procedure at longer-term follow-up. However, as no statistically significant differences were observed within individual subgroups, these findings should be interpreted with caution and do not indicate definitive superiority of either procedure. Previous studies have reported lower long-term VAS scores after Latarjet and more frequent residual pain following BR [35, 46]. The residual discomfort after BR may be related to impingement between the posterior labrum and the newly established footprint of the posterior cuff [47]. In addition, the Latarjet procedure restores glenoid bone anatomy and provides a durable bony block against anterior humeral head translation, which may contribute to sustained joint stability and functional recovery over time [48]. By contrast, BR depends primarily on soft-tissue healing, and its effect may be more vulnerable to gradual stretching or loosening [16], potentially influencing both pain and functional outcomes during longer follow-up. This temporal pattern was also observed in functional outcomes, as reflected by the Rowe score. Given that the Rowe score incorporates stability, motion, and function, these findings suggest that follow-up duration may influence not only pain perception but also broader aspects of shoulder recovery. ROM is an essential parameter in assessing the outcomes of instability procedures. A potential loss in postoperative ROM has been a concern among orthopedic surgeons. Stiffness and the loss of ER represent significant complications of the Latarjet procedure, with a mean ER loss of 10–15° [49]. For remplissage, an average ERa loss of 5.26° has been documented [50]. In our analysis, BR and Latarjet showed no statistically significant difference in ROM, which is consistent with previous studies [22, 51]. The heterogeneity observed in ERs and ERa may be attributed to differences in surgical techniques. The included Latarjet procedures were performed via open, arthroscopic, or modified approaches. Although our study found that both the BR and Latarjet procedures resulted in similar losses of ROM, it is important to note that even minor changes in ROM can adversely affect some patients [50]. Therefore, further research is necessary to elucidate the impact of different surgical approaches on ROM outcomes. When comparing the two procedures, the RTS rates were equally high, with 76.15% for BR and 70.53% for the Latarjet procedure, aligning with previous studies [50, 52]. Horinek et al. [33] further reported that, among overhead or contact sports athletes, RTS was higher after BR (91.5%) than after Latarjet procedure (72.7%), suggesting a particular benefit for this subgroup. Similarly, Davis et al. [12] found that BR was associated with higher odds of RTS at both any level and at the preinjury level compared with alternative procedures, particularly in athletes involved in collision sports. This advantage may be related to the ability of remplissage to convert an off-track Hill–Sachs lesion into an extra-articular defect, thereby enhancing stability and enabling earlier, safer return to play [53]. These findings highlight the importance of considering sport-specific demands when selecting the surgical approach and underscore the need for further research comparing long-term RTS outcomes across different athletic populations. This review observed higher recurrence and revision rates in BR, although the differences were not statistically significant. Overall, both procedures demonstrated comparable rates, indicating similar effectiveness in restoring shoulder stability, consistent with previous systematic reviews [54, 55]. However, most studies define recurrence as subluxation or recurrent instability, which may not fully reflect patients’ perceived shoulder stability. To provide a more comprehensive assessment of surgical success, further studies are needed that incorporate patients’ own perceptions and beliefs regarding the stability of their shoulders. In this study, we observed a higher incidence of complications in the Latarjet procedure group compared with the BR group (1.2% versus 8.7%). This is in accordance with previous reviews reporting a complication rate of 6–7% after Latarjet procedure [42]. In addition, the assessment of complications after BR in multiple studies showed complications ranging from 0.4% to 0.6% [55–57]. The high incidence of complications following the Latarjet procedure may be related to a variety of factors, including surgical complexity, graft-related issues, and hardware complications. GBL may influence surgical outcomes and is an important factor in the selection of surgical technique for anterior shoulder instability. In the present study, baseline GBL was comparable between the BR and Latarjet groups (10.3 ± 7.6% versus 11.9 ± 8.1%; standardized mean differences [SMD] = −0.20). To further explore the potential impact of indication bias, we performed a sensitivity analysis restricted to studies with a mean GBL < 15%. The results remained consistent with the primary analysis, with no changes in the direction or statistical significance of the pooled outcomes. Evidence from previous studies indicates that the relationship between GBL and surgical outcomes may vary across different levels of bone loss. Yang et al. [36] reported comparable overall outcomes between BR and Latarjet in patients with subcritical GBL (< 25%); however, in patients with GBL > 15%, the Latarjet procedure was associated with lower recurrence and revision rates. In contrast, Horinek et al. [34] demonstrated comparable clinical outcomes, return-to-sport, and recurrence rates between BR and Latarjet in patients with GBL > 15%. Taken together, these findings suggest that the comparative effectiveness of BR and Latarjet may differ across varying degrees of bone loss, which may partially contribute to the heterogeneity observed across studies. Further studies are needed to clarify how different levels of GBL may affect surgical outcomes. In addition to GBL, lesion characteristics such as Hill–Sachs lesion tracking status may further influence surgical outcomes. The incidence of Hill–Sachs lesions in anterior shoulder instability ranges from 38% to 88%, and these lesions have been associated with a higher risk of recurrent instability [58]. The glenoid track concept further classifies these lesions as on-track or off-track, with off-track lesions corresponding to engaging lesions and representing a more unstable lesion pattern. In the present analysis, a higher proportion of off-track lesions was observed in the BR group compared with the Latarjet group, suggesting that the BR cohort may have included more challenging lesion patterns. Despite this imbalance, recurrence and revision rates were comparable between the two procedures. Previous studies focusing specifically on off-track lesions have reported broadly similar findings. Yang et al. [36] demonstrated comparable PROMS, recurrence rates, and revision rates between BR and Latarjet in patients with off-track lesions, although the Latarjet group had a higher complication rate. Similarly, Gereli et al. [45] included a cohort composed entirely of patients with off-track lesions and found no significant differences between the two procedures in range of motion. Although the present analysis provides a comprehensive comparison, several limitations should be considered when interpreting the findings. Most included studies were nonrandomized, predominantly level III designs, making them susceptible to selection bias and residual confounding. Surgical decision-making was often based on surgeon preference, and variations in surgical technique and rehabilitation protocols across studies may have influenced outcomes. In addition, the number of included studies was relatively small, and no randomized controlled trials were available, precluding formal assessment of publication bias. Heterogeneity in baseline characteristics and outcome reporting remains a concern. Although GBL was relatively comparable between groups, greater imbalances were observed in the proportion of off-track lesions. To partially address potential indication bias related to bone loss, a sensitivity analysis restricted to studies with a mean GBL < 15% was performed. This threshold was selected as a conservative cutoff within the commonly described subcritical range; however, more granular stratification was not feasible because of limited and inconsistently reported data across studies. In particular, analyses based on lesion tracking status (off-track versus on-track) could not be performed owing to the inconsistent reporting of these variables across studies. Future studies should focus on well-designed prospective or randomized trials with standardized reporting of bone loss and lesion characteristics. In particular, stratified analyses based on GBL and lesion severity, including off-track status, are needed to better define the indications and comparative effectiveness of BR and Latarjet. Conclusions For anterior shoulder instability, BR was associated with a lower complication rate than the Latarjet procedure, with otherwise comparable PROMs, ROM, recurrence, revision, and RTS. Subgroup analysis indicated a temporal effect, with better short-term VAS and Rowe outcomes for BR and better long-term outcomes for Latarjet. Supplementary Information Supplementary material 1: Table S1. Surgical Techniques Used in Included Studies.Supplementary material 2: Table S1. Methodological Index for Non-Randomized Studies (MINORS) Scores of Included Studies; Table S2. Intraclass Correlation Coefficient Analysis of Interobserver Reliability for MINORS Quality Assessment; Figure S1. Interobserver Agreement of MINORS Quality Assessment.Supplementary material 3: Forest plots of subgroup analyses according to follow-up duration: Figure S1. Visual Analog Scale (VAS); Figure S2. Rowe Score; Figure S3. Forward Flexion (FF); Figure S4. External Rotation at the Side (ERs); Figure S5. Recurrence Rate; and Figure S6. Complication Rate.Supplementary material 4: Sensitivity analyses for the pooled outcomes: Figure S1. Rowe Score; Figure S2. Visual Analog Scale (VAS); Figure S3. Single Assessment Numeric Evaluation (SANE); Figure S4. Forward Flexion (FF); Figure S5. External Rotation at the Side (ERs); Figure S6. External Rotation at 90° Abduction (ERa); Figure S7. Internal Rotation (IR); Figure S8. Return to Sport (RTS); Figure S9. RTS Time; Figure S10. Recurrence Rate; Figure S11. Revision Rate; and Figure S12. Complication Rate.Supplementary material 5: Sensitivity analyses restricted to studies with mean baseline glenoid bone loss (GBL) <15%: Figure S1. Rowe Score; Figure S2. Visual Analog Scale (VAS); Figure S3. Forward Flexion (FF); Figure S4. External Rotation at the Side (ERs); Figure S5. External Rotation at 90° Abduction (ERa); Figure S6. Internal Rotation (IR); Figure S7. Recurrence Rate; Figure S8. Revision Rate; and Figure S9. Complication Rate. Supplementary material 1: Table S1. Surgical Techniques Used in Included Studies. Supplementary material 2: Table S1. Methodological Index for Non-Randomized Studies (MINORS) Scores of Included Studies; Table S2. Intraclass Correlation Coefficient Analysis of Interobserver Reliability for MINORS Quality Assessment; Figure S1. Interobserver Agreement of MINORS Quality Assessment. Supplementary material 3: Forest plots of subgroup analyses according to follow-up duration: Figure S1. Visual Analog Scale (VAS); Figure S2. Rowe Score; Figure S3. Forward Flexion (FF); Figure S4. External Rotation at the Side (ERs); Figure S5. Recurrence Rate; and Figure S6. Complication Rate. Supplementary material 4: Sensitivity analyses for the pooled outcomes: Figure S1. Rowe Score; Figure S2. Visual Analog Scale (VAS); Figure S3. Single Assessment Numeric Evaluation (SANE); Figure S4. Forward Flexion (FF); Figure S5. External Rotation at the Side (ERs); Figure S6. External Rotation at 90° Abduction (ERa); Figure S7. Internal Rotation (IR); Figure S8. Return to Sport (RTS); Figure S9. RTS Time; Figure S10. Recurrence Rate; Figure S11. Revision Rate; and Figure S12. Complication Rate. Supplementary material 5: Sensitivity analyses restricted to studies with mean baseline glenoid bone loss (GBL) <15%: Figure S1. Rowe Score; Figure S2. Visual Analog Scale (VAS); Figure S3. Forward Flexion (FF); Figure S4. External Rotation at the Side (ERs); Figure S5. External Rotation at 90° Abduction (ERa); Figure S6. Internal Rotation (IR); Figure S7. Recurrence Rate; Figure S8. Revision Rate; and Figure S9. Complication Rate.