Abstract
Background:
Glenohumeral instability can be addressed surgically with repair of the capsulolabral complex, the “Bankart” repair, or bone augmentation with coracoid autograft, the “Latarjet” procedure. Superior return-to-sport (RTS) rates with either Bankart repair or Latarjet have yet to be delineated, and it remains unclear which surgical procedure is optimal.
Purpose:
To review the current literature and report on RTS data for patients who received Bankart repair or Latarjet to address glenohumeral instability.
Study Design:
Systematic review; Level of evidence, 4.
Methods:
This systematic review and meta-analysis was conducted following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guidelines by an individual researcher screening through 3 databases (PubMed, Cochrane Library, Embase) for articles including RTS rates after Bankart repair and Latarjet. Eight studies that met inclusion and exclusion criteria were included for review.
Results:
Meta-analysis included 874 athletes who underwent shoulder stabilization: 479 Bankart repair and 395 Latarjet. An overall 94.4% of athletes returned to sport with no statistically significant difference between the surgical groups. Of 665 athletes, 68.1% returned to sport at the same level or higher with no statistically significant difference between the groups. RTS time among the 665 athletes averaged 6 months with no statistically significant difference between the groups. A total of 56 athletes did not RTS: 33 (7.4%) Bankart repair and 23 (6.3%) Latarjet. There was a statistically significant difference (P < .0001) in recurrent instability rates. Bankart repair (14.8%) was 4 times more likely to result in recurrent instability as compared with Latarjet (3.5%). A total of 85 athletes, 71 Bankart repair and 14 Latarjet, had recurrent instability with 53.5% (7.9% overall) and 100% (3.5% overall) requiring a revision procedure, respectively.
Conclusion:
Bankart repair and Latarjet have a high RTS rate with no significant difference in rate, rate of return to previous level of play, or time. One of 4 athletes will not return to the level played before surgery, independent of stabilization technique. Bankart repair is 4 times more likely to result in recurrent instability as compared with Latarjet.
The inherent mobility of the glenohumeral joint allows for a wide range of motion but also predisposes the joint to instability. 8 Approximately 96% of all glenohumeral instability is traumatic and unidirectional in nature, of which 85% to 98% is anterior based. 2
In the general population, 1% to 2% of people will experience glenohumeral instability at some point in their lifetimes; however, the prevalence of glenohumeral instability is much higher in military members and contact athletes.20,28 A recent collegiate athlete study showed a glenohumeral instability incidence of 0.12 per 1000 exposures and 0.40 injuries per 1000 men's football exposures, which was the sport with the highest injury rate. 23 Notably, 45% of the injuries in this study required at least 10 days away from sport, highlighting the significant impact that these events can have on athletes’ careers.
Nonsurgical treatment usually includes a course of physical therapy for dynamic shoulder strengthening. Common surgical treatments are repair of the anteroinferior capsulolabral complex, also known as the “Bankart” repair, and the Latarjet procedure.28,32 A systematic review and meta-analysis on adolescent athletes with glenohumeral instability found a 41.3% return-to-sport (RTS) rate and 72.3% recurrence rate for the nonsurgical group as compared with a 95.3% and 13.2% for the surgical group who had a single dislocation and 77.6% and 22.3% for the surgical group with recurrent instability, respectively. 19
When surgery is indicated, the decision regarding which surgical approach to take remains a topic of ongoing debate. There is a higher redislocation rate after Bankart repair as compared with a Latarjet procedure, although Bankart repairs are associated with lower complication rates when recurrent dislocations are excluded.3,31 Some surgeons, especially in Europe, recommend performing a Latarjet procedure in most primary cases of recurrent anterior shoulder dislocation, while other surgeons, such as those in the United States, recommend performing an arthroscopic Bankart repair for patients with <13.5% glenoid bone loss and/or an on-track Hill-Sachs lesion.2,7 Factors such as age, injury history, and physical demands of the athlete also play a role in surgical decision making.
Still, there is a need for further analysis focusing on the optimal surgical management of glenohumeral instability in athletes. The goal of this study was to review the current literature and report RTS rates in athletes who received either a Bankart repair or a Latarjet stabilization procedure.
Methods
This systematic review and meta-analysis was conducted with adherence to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guidelines.
Eligibility Criteria
Inclusion criteria were outcomes-based studies that assessed 2 surgical techniques for anterior shoulder instability: Bankart repair and Latarjet, defined as coracoid autograft transfer. Shoulder instability was defined as 1 or more instability events, and surgical techniques could be arthroscopic or open. Studies also needed to include data on RTS, follow-up, and outcomes. Open and arthroscopic techniques were included to attempt to assess for differences in outcomes based on surgical technique. Studies were excluded if they failed to report RTS data, there was no direct comparison of groups (Bankart and Latarjet), the population was adolescent (<16 years), and revision shoulder surgery or concomitant procedures (eg, Remplissage) were included in analyses. Concomitant procedures were excluded to reduce variation among athletes and confounding variables that may affect RTS and functional outcomes of the shoulder.
Search Strategy
PubMed, Cochrane, and Embase were searched until August 29, 2024. The search strategy involved use of the following keywords to capture all potential studies on this topic:
(‘shoulder joint’ OR shoulder* OR glenohumer* OR ‘anterior shoulder’) AND (‘joint instability’ OR dislocation OR hyperlaxity* OR laxit* OR unstab* OR instab* OR dislocate * OR luxat* OR sublax*) AND bankart* AND (laterjet* OR latarjet* OR bristow*) AND (surg* OR management OR treat* OR repair OR arthroscop*) AND (‘return to sport’ OR rts OR ‘return to play’ OR athlete* OR sport*).
Studies were screened by titles and abstracts by an independent reviewer (C.E.N.). A full-text review was then performed if a study matched the eligibility criteria. The search strategy was performed by 1 independent author (C.E.N.) under the supervision of the senior author (E.C.M.), who reviewed the final studies to ensure appropriate inclusion. Furthermore, the references of each eligible article were manually reviewed to ensure that no eligible studies were missed. Duplicate studies were not included.
Data Collection Process and Data Items
Data collection forms were used independently to collect the following information: authorship, study year, institution, level of evidence, study design, typification, age, sex, number of patients, and follow-up intervals. Postoperative outcomes data included time to RTS, percentage of athletes returning to sport, and percentage of athletes returning to sport at the same level.
Statistical Methods
Numeric study outcomes were originally reported as mean (standard deviation) or median (range). When dispersion statistics for RTS data were missing, a standard deviation was calculated by taking 25% of the mean.
Mixed-effects meta-analysis models and forest plots were constructed to evaluate differences and consistency of time to RTS and RTS at the same level. For studies that reported single proportions, Freeman-Tukey transformation was applied to those proportions to ensure stable estimation of effects. For those same single proportion models, the DerSimonian-Laird estimator was used to calculate τ2 and the Jackson method for confidence intervals of τ and τ2. For individual study confidence intervals, the Clopper-Pearson confidence interval was given. For studies that reported means for time to RTS, the restricted maximum likelihood estimator for τ2 was employed, and the Q-profile method was used with a confidence interval of τ and τ2.
Mixed-effects meta-regression models were built from the meta-analysis model, which included mean patient age (years), the proportion of females in each study, and various preoperative PROs (if possible) as fixed effects and the studies as random effects. For these models, the QM omnibus test of moderators and tests of heterogeneity are reported with the regression coefficient table for each outcome.
Summary statistics are reported as means and standard deviations for numeric data and counts with relative percentages for categorical data. Statistical significance is set at α = .05.
R Version 4.3.2 was used for analysis, and the meta and metafor packages were used for all meta-analysis modeling and plot construction.4,26,30
Results
This systematic review and meta-analysis was conducted with adherence to the PRISMA guidelines. 24 Screening was performed by an individual researcher (C.E.N.) through 3 databases: PubMed, Cochrane Library, and Embase. In total, 356 studies were identified from these databases, with 105 duplicates removed; 251 studies were then screened with 164 considered irrelevant, leaving 87 full-text studies to be assessed for eligibility (Figure 1). Overall, 71 studies were excluded for the following reasons: incorrect study design (n = 25), incorrect intervention (n = 22), incorrect outcomes (n = 19), and wrong patient population (n = 5). Eight studies were included with 874 patients who defined themselves as athletes, with a mean age of 31.87 ± 9.27 years. A total of 479 athletes were treated with Bankart repair and 395 with Latarjet (Table 1). There was no statistically significant difference in age between the Bankart repair and Latarjet stabilization cohorts.
Detailed flowchart of the literature search using PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) criteria.
Demographic Data of Included Studies
No. of glenoid bone defects (types 0, 1, 2, 3).
There were 830 male-identifying patients and 44 female-identifying patients. One study listed “men” and “women,” which we presumed was intended to be defining their sex rather than gender. These patients were treated surgically with either Bankart or Latarjet. The mean follow-up time was 32 ± 17 months. Four articles reported on dislocations before surgery, accounting for 401 patients: 260 Bankart and 141 Latarjet. Patients who received a Bankart repair and Latarjet had, on average, 6.13 and 8.25 dislocations before surgical intervention, respectively.
In the overall meta-analysis for proportion of athletes to RTS, there were 8 studies with a total of 874 athletes, where 818 returned to sport (Figure 2). The estimated mean proportion to RTS was 94.41% (95% CI, 0.9264-0.9599) using a fixed-effects model. The fixed-effects model may be more appropriate for interpreting the pooled estimate as there is low heterogeneity across studies (τ2 = 0.0006; I2 = 12.0%; P = .3159).
Forest plot for return to sport at any level per study.
The surgical subgroup analysis reveals that the most, albeit low, heterogeneity occurs in the Latarjet group (I2 = 20.8%), comprising 5 studies, which had an estimated RTS proportion of 95.04% (95% CI, 0.9243-0.9719) under the fixed-effects model and 94.96% (95% CI, 0.9189-0.9742) under the random-effects model. The Bankart group, comprising 5 studies, also shows low heterogeneity measures of I2 = 10.5%. While the Bankart group (93.86%) had slightly less heterogeneity and a lower estimated proportion to RTS than the Latarjet group, the difference is negligible and not statistically significant, as seen by the test for subgroup differences (P = .6041), the overlapping confidence intervals for heterogeneity and proportion to RTS, and the nonsignificant meta-regression.
In the overall meta-analysis for proportion to RTS at the same level or higher, there were 5 studies with 665 athletes where 444 returned to sport (Figure 3). The estimated mean proportion to RTS was 74.68% (95% CI, 0.6921-0.7981) using a random-effects model. The random-effects model is more appropriate for interpreting the pooled estimate as there is considerable heterogeneity across studies (τ2 = 0.0786; I2 = 95.4%; P < .0001).
Forest plot for return to sport at a preinjury level or higher per study.
The surgical subgroup analysis shows that the most substantial heterogeneity occurs in the Bankart group (I2 = 97.1%; τ2 = 0.1084), comprising 5 studies, which had an estimated proportion of 63.38% (95% CI, 0.5851-0.6812) for RTS at the same level or higher under the fixed-effects model and 71.92% (95% CI, 0.4287-0.9361) under the random-effects model. The Latarjet group, comprising 5 studies, had an estimated proportion of 74.68% (95% CI, 0.6921-0.7981) for RTS at the same level or higher under the fixed-effects model and 71.41% (95% CI, 0.5033-0.8866) under the random-effects model, as well as substantial heterogeneity measures of I2 = 91.7% and τ2 = 0.0536. While the Bankart group had more heterogeneity but a lower estimated proportion to RTS than the Latarjet group, the difference is negligible and not statistically significant, as seen by the test for subgroup differences (P = .9699), the overlapping confidence intervals for heterogeneity and proportion to RTS, and nonsignificant meta-regression.
In the overall meta-analysis for time to RTS, there were 6 studies with 665 athletes (Figure 4). The estimated mean time to RTS was 6.0580 months (95% CI, 5.6684-6.4476) using a random-effects model. The random-effects model is more appropriate than a fixed-effects model for interpreting the pooled estimate as there is substantial heterogeneity across studies (τ2 = 0.3649; I2 = 76.1%; P < .0001).
Forest plot for time to return to sport (months) per study.
The surgical group subanalysis demonstrates that most of the heterogeneity occurs in the Bankart group (I2 = 84.9%), comprising 6 studies, which had an estimated RTS time of 6.3797 months (95% CI, 5.5738-7.1855) under the random-effects model. The Latarjet group, comprising 6 studies, had an estimated RTS time of 5.87 months (95% CI, 5.5372-6.1973) under the random-effects model and a moderate heterogeneity measure of I = 60.8%. While the Latarjet group had less heterogeneity and a lower estimated time to RTS than the Bankart group, the difference is negligible and not statistically significant, as seen by the test for subgroup differences (P = .2488), the overlapping confidence intervals for heterogeneity and time to RTS, and nonsignificant meta-regression.
There is statistically significant evidence for a trend between the mean age of patients and the mean percentage to RTS. For every 1-unit increase in mean age, the mean percentage to RTS decreases by 0.021% (Figure 5). Likewise, there is statistically significant evidence for a trend between the sex of a patient and the percentage to RTS (P < .001). When controlling for age, the mean proportion to RTS decreases by 4.27% for each 1% increase in the proportion of females in a study (95% CI, –6.4092% to −2.1364%) (Figure 6). However, there is no statistically significant evidence of a trend in time to RTS and age or sex in conjunction.
Forest plot for percentage to return to sport (months) and age per study.
Forest plot for percentage to return to sport (months) and mean female proportion per study.
Of the 874 patients who identified as athletes, there was a mixture of contact and overhead sports: rugby, Gaelic football, judo, hurling, handball, soccer, and skiing. Of these athletes, 392 self-identified as “competitive,” and the rest were a mixture of professional and recreational (Table 1).
Recurrence
There were 85 shoulders, 71 Bankart and 14 Latarjet, that had recurrent instability, defined as more than 1 event. Bankart repair had a statistically significant increased risk for recurrence (P < .00001). Bankart repair (14.8%) was >4 times more likely to cause recurrence than Latarjet (3.5%). Recurrences from subluxation occurred 7 times: 6 times with Bankart and 1 time with Latarjet. Recurrences from a dislocation occurred 32 times: 28 times with Bankart and 4 times with Latarjet.
Complications and Failure
Of the 8 studies, there were 814 athletes with 10 complications and 56 revisions. Complications included 6 hematomas, 2 cases of biceps tendonitis, 1 deep infection, 1 surgical site infection, and 1 case of subacromial bursitis. Of the 56 revisions, 45 were due to instability: 38 in the Bankart group and 7 in the Latarjet group.
Discussion
The results of this review demonstrate that Bankart repair and Latarjet have a high RTS rate, with both stabilization techniques returning >93% of athletes to sport with no significant difference in RTS between them. These findings are in contrast to previous data that report higher percentages of athletes not being able to RTS after Bankart and Latarjet stabilizations.1,16 These studies, however, may demonstrate that a failure of RTS can be multifactorial and include reasons independent of shoulder function, such as fear of reinjury, retirement, or personal reasons.16,25
Among the athletes who do RTS, there is a lower return to the same level of play (RTSP) after anterior shoulder stabilization surgery. This study found that about 1 of every 4 athletes will not return to the level of play before surgery, with no significant difference observed between Bankart and Latarjet stabilization. This study is, again, in contrast to a previously reported systematic review that suggested a large difference when comparing Bankart and Latarjet stabilization; specifically, up to 69% and 90% of patients who underwent Latarjet and Bankart returned to their previous levels of play, respectively. 1 It is important to note that this prior review did comment on limitations and heterogeneity in the available literature. Notably, the Latarjet studies consisted of predominantly competitive collision athletes, which could have led to selection bias and a lower published RTSP. More recent research has suggested results similar to this study with no significant difference in overall RTS rate and return at the same level. 13 The primary aim of this study, however, was not limited to athletes and had limited data. 13 In addition, there was no stratification of collision sport versus other sports, which could have affected the RTS rates in this study.
Furthermore, time to RTS was not significantly different among the groups of athletes in this review. A mean 6 months was required for a RTS with Bankart repair and Latarjet. Hurley et al found similar results with comparable RTS rates and time, reported as ranges. 14 However, they did not report on recurrence or complications.
Recurrence of instability after stabilization is a common but devastating complication of a Bankart repair and a Latarjet procedure and can occur at unacceptably high levels. 1 Recurrent instability events occur at rates up to 17.8% in the athlete after stabilization. 22 Therefore, evaluating reoccurrence is an important factor in recovery and outcomes. This is especially true in the athlete, as recurrence is likely to prolong recovery or increase time out of sport when stabilization fails and sport practice can not be maintained. This review observed a statistically significant difference in recurrence after Bankart repair (14.8%), which was >4 times more likely to occur as compared with the Latarjet procedure (3.5%). This is in agreement with previous reviews suggesting that the Latarjet resulted in fewer recurrent instability events after stabilization.1,5,13 However, these findings must be taken with discretion given the heterogeneity of studies and specifically the nature of sport (ie, collision) and competition level, which were inconsistently reported.
Addressing factors beyond functional rehabilitation of the shoulder, such as fear of reinjury, should be included in an RTS protocol to help more athletes RTS.12,29 This may be supported by the redislocation rate of 6.5% without an overall delay in timeline after inclusion of a psychological readiness assessment in the RTS protocol for shoulder stabilization. 18
An athlete's priority when choosing to undergo surgery is minimizing time away from the sport and preserving one's level of play. While some factors in an athlete's recovery are uncontrollable, the procedure and recovery protocol are within control. This study suggests that in the setting of shoulder instability in athletes, Bankart repair and Latarjet yield similar RTS rates. No significant differences in timeline to RTS, rate of RTS, and RTS at the same level were observed in this review. Therefore, choosing the proper stabilization procedure that is most appropriate for each patient should be done considering the aforementioned factors, such as age, presence of bone loss, participation in collision sports, and so on. 10 Educating the athlete on the outcomes and risks of stabilization is also crucial owing to the risk of complication and failure between the procedures, particularly the Bankart.
Limitations
This review has several limitations. First, it focused on Bankart repairs and Latarjet procedures without stratifying between open and arthroscopic procedures, and it did not consider technical variation in these procedures. No observations on the impact of these differences in procedure can be determined in this review. The indication for Bankart repair versus Latarjet can rely on multiple patient-specific factors, such as degree of bone loss, age, and type of sport, and can lead to heterogeneity among studies as well. This review did not stratify long- versus short-term follow-up rates, with a mean follow-up of 32 months, so it is possible that complications such as failure rates were missed within these limited time frames. Inherent to any systematic review is selection bias within the studies and patients, which can lead to significant heterogeneity of the pooled data. Furthermore, this systematic review included only 1 reviewer, eliminating agreement on study inclusion by multiple authors, which may also introduce bias. In addition, there can be publication bias of the studies, which may limit generalizability to different patient populations. Last, this review focused on the general athlete and was not stratified by type of sport, as heterogeneity in reporting limited the ability to identify RTS/RTSP rates in specific collision sports such as football/rugby, which could affect recurrence rates and RTS/RTSP.
Conclusion
Athletes with anterior shoulder instability often undergo surgical stabilization via a Bankart repair or Latarjet with the goal of minimizing recurrent instability and achieving high rates of RTS. This systematic review and meta-analysis of the available literature demonstrated that RTS after Bankart repair and Latarjet showed no significant difference in RTS time, rate of RTS, or rate of return to previous level of play between the surgical groups. Bankart repair and Latarjet have a high RTS rate, with both stabilization techniques returning >93% of athletes, which is in contrast with previous data that report higher percentages of athletes not being able to RTS. However, about 1 of every 4 athletes will not return to the level of play before surgery independent of stabilization technique. Furthermore, Bankart repair is 4 times more likely to result in recurrent instability as compared with stabilization with a Latarjet.
Footnotes
Final revision submitted May 19, 2025; accepted July 15, 2025.
One or more of the authors has declared the following potential conflict of interest or source of funding: T.J.N., J.W.G., and B.M. have received consulting fees from for Styker. J.T.B. has received consulting fees from Smith and Nephew and DJ Orthopaedics and research support from Stryker, Mitek, and Biomet. E.C.M. has received royalties from Zimmer and Biomet. AOSSM checks author disclosures against the Open Payments Database (OPD). AOSSM has not conducted an independent investigation on the OPD and disclaims any liability or responsibility relating thereto.
