One-year follow-up of 20 patients undergoing the Latarjet procedure

Abstract

Aims

The Latarjet procedure is the treatment of choice for patients who have recurrent anterior shoulder instability with glenoid bone loss. However, the stabilizing effect of the Latarjet procedure in patients is sparsely described. The aim of this study was to evaluate the glenohumeral joint (GHJ) kinematics during an apprehension-relocation test in patients with anterior shoulder instability before and after their Latarjet procedure, and in comparison with their contralateral healthy shoulder.

Methods

A total of 20 patients scheduled for the Latarjet procedure were enrolled. The patients were examined preoperatively with bilateral radiostereometric analysis (RSA) and one year after surgery on the operated shoulder with an apprehension-relocation test. Bone models were obtained from CT scans and aligned with the RSA images. The GHJ kinematics was evaluated with two methods: the humeral head centre location relative to the glenoid centre, and the GHJ contact point relative to the glenoid centre.

Results

No difference in GHJ kinematics was found between the healthy and the postoperative GHJ. Compared with the preoperative injured shoulder, the postoperative mean (95% CI) humeral head centre was 0.8 mm (0.1 to 1.4) more superior and 0.7 mm (-0.1. to 1.4) more posterior during the apprehension test, and 0.5 mm (0.0 to 1.1) more posterior during the relocation test. The postoperative contact point was posterior to the coracoid bone block and 0.9 mm (-0.2 to 2.0) more posterior than in the preoperative injured shoulder during the apprehension test. The articulating area of the coracoid bone block was decreased by 63.9% (75.5 to 114.6) one year after surgery.

Conclusion

The Latarjet procedure restored the humeral head centre location posterior and superior, and the contact point posterior, to the coracoid bone block. This suggests that the stabilizing effect of the GHJ following the Latarjet procedure is primarily due to the conjoined tendon rather than the coracoid bone block itself.

Cite this article: Bone Joint Res 2025;14(6):506–515.

Article focus

• The apprehension-relocation test is the most frequently applied clinical test to evaluate anterior shoulder instability.

• This prospective cohort study aimed to evaluate the glenohumeral joint kinematics during an apprehension-relocation test in patients before and after the Latarjet procedure, and to compare these findings to the kinematics of their healthy contralateral shoulder.

Key messages

• The Latarjet procedure was able to restore the glenohumeral joint kinematics compared to the healthy shoulder.

• The humeral head centre and contact point moved posteriorly compared to the position following injury, thus the coracoid bone block is not what establishes the glenohumeral joint stability.

• A large bone block resorption was observed one year postoperatively.

Strengths and limitations

• This is a prospective clinical study evaluating the effect before and after the Latarjet procedure.

• The glenohumeral joint kinematics was evaluated with the highly precise method of radiostereometry.

• Due to the static recordings, muscle reflexes may have caused a more central position of the humeral head when the glenohumeral joint was recorded, which may have decreased the differences between the pre- and postoperative measurements.

Introduction

The Latarjet procedure is often the first-choice surgery in the treatment of anterior shoulder instability (ASI) with glenoid bone loss, or after failed Bankart repair.^1,2 The procedure provides glenoid bone restoration by enhancing the width of the glenoid through a transposition of the coracoid process to the anteroinferior glenoid rim. Furthermore, the conjoined tendon is believed to provide dynamic stability through a “sling effect” when the shoulder is abducted and externally rotated (ABER).^3,4

The biomechanical properties, and therefore the stabilizing mechanisms, of the glenohumeral joint (GHJ) following the Latarjet procedure have been primarily evaluated in experimental studies.^5-8 These studies found that the Latarjet procedure reduced anterior humeral head translation compared to the translation observed in cases with a 15% to 30% induced glenoid bone defect. Evaluation of the GHJ kinematics in patients has been performed previously on the shoulder in the ABER position, with varying results.^9,10 ASI symptoms and GHJ dislocations are typically triggered in the combination of the ABER position and concomitant anterior-directed stress to the GHJ. This situation can be reproduced with the apprehension test and relieved during the relocation test, which is the preferred clinical examination to evaluate GHJ stability.¹¹

Knowledge concerning GHJ kinematics before and after stabilizing surgical procedures is crucial for understanding and improving the surgical treatments of patients with ASI. GHJ kinematics during an apprehension test has been previously evaluated on patients with ASI, but to our knowledge it has not yet been assessed following the Latarjet procedure.^12,13

Radiostereometric analysis (RSA) is an accurate method for evaluating joint kinematics, and has previously been used to evaluate GHJ kinematics in patients with impingement and rotator cuff lesions.^14,15

This prospective cohort study aimed to evaluate the GHJ kinematics during an apprehension-relocation test in patients before and after the Latarjet procedure, and to compare these findings to the kinematics of their healthy contralateral shoulder. It was hypothesized that the Latarjet procedure would restore the GHJ kinematics toward the healthy contralateral shoulder.

Methods

The study was conducted in accordance with the Declaration of Helsinki,¹⁶ and patients signed an informed consent form before participation. This prospective cohort study was conducted in agreement with the STROBE guidelines.¹⁷

Study design and patients

A total of 20 patients with unilateral traumatic ASI scheduled for the Latarjet procedure were prospectively enrolled from February 2022 to August 2023 at Aarhus University Hospital, Denmark. Inclusion criteria were: 1) age 18 to 40 years, 2) a healthy contralateral shoulder, and 3) a CT scan verifying a glenoid bone lesion > 10% of the glenoid width measured by the best-fit circle method and/or recurrent ASI despite previous Bankart repair.^18,19 Exclusion criteria were patients with bilateral or multidirectional shoulder instability, and alcohol or drug abuse. Overall, 19 patients completed the one-year follow-up. The mean follow-up time was 13 months (95% CI 12.5 to 13.6). Table I presents the patients’ baseline demographics.

Table I.

Patient demographic data.

Characteristic	Value (n = 20)
Males/females, n (%)	17 (85)/3 (15)
Median age, yrs (IQR)	29 (25 to 38)
ASI shoulder = right, n (%)	8 (40)
Mean glenoid bone loss, % (SD)	17.5 (5.3)
Previous Bankart repair, n (%)	16 (80)
Mean number of dislocations (95% CI)	11.1 (7.4 to 14.8)
Engaging/non-engaging bipolar lesions, n (%)27	4 (20)/16 (80)

ASI, anterior shoulder instability.

Each patient case was its own control by comparing the kinematics of the ASI shoulder with the kinematics of the contralateral healthy shoulder. The patients were examined before their Latarjet procedure, and again one year after their surgery with RSA during an apprehension-relocation test to evaluate the GHJ kinematics as the primary outcome. Secondary outcomes were patient-reported outcomes and evaluation of bone block resorption.

Surgical technique

The open Latarjet procedure was performed by experienced shoulder surgeons (TFJ, TMT) as described by Young et al.²⁰ The coracoid process with the conjoined tendon was osteotomized (approximately 25 mm) and fixed to the anterior-inferior glenoid rim with two cannulated 4.5 mm screws (Arthrex, USA).

Preoperatively and one year postoperatively, an apprehension-relocation test was performed while static RSA recordings were captured at the end-range positions (Figure 1). The end-range position was defined as when the patient felt apprehension during the apprehension test, or maximal pressure was applied and relief during the relocation test. Preoperatively, the examination was performed bilaterally to use the healthy shoulder as a reference. At the one-year postoperative examination, only the shoulder that had undergone the Latarjet procedure was examined. All recordings were performed twice to examine the repeatability of the GHJ kinematic measures during the apprehension-relocation test.

Fig. 1.

Test setup during the apprehension-relocation test. The apprehension and relocation tests were performed by experienced shoulder surgeons. The patient was positioned supine with the shoulder approximately 90° abducted, and 90° externally rotated. During the apprehension test, anterior-directed stress was applied to the proximal part of the humerus until the patient felt the apprehension or maximal possible stress was applied (left image). During the relocation test, posterior-directed stress was applied to the proximal part of the humerus until the patient felt relieved of the apprehension or maximal possible stress (right image).

The RSA images were acquired using the digital Adora RSA system (NRT X-ray, Denmark). The system consists of two ceiling-mounted x-ray tubes, each angled at 20°, and two digital Canon CXDI-70C image detectors were positioned behind the carbon calibration box (Carbon box 24; Medis Specials, Netherlands). The resolution of the images was 2,800 × 3,408 pixels (0.125 × 0.125 mm/pixel).

Prior to surgery, the patients’ shoulders and elbows were CT scanned (SOMATOM Definition Flash; Siemens Healthcare, Germany). At the one-year postoperative examination, the operated shoulder was CT-scanned again. The exposure settings were 120 kV, 260 mAs, slice thickness 0.3, and pixel spacing 0.49 × 0.49 mm. The patient-specific bone models, without cartilage, were generated from the CT scans by automated graph cut segmentation and marching cubes algorithm.

The RSA images were calibrated using model-based RSA (RSAcore, the Netherlands) and manually initialized in AutoRSA software (Orthopaedic Research Unit, Aarhus, Denmark), using digitally reconstructed radiographs generated from the patient-specific bone models.²¹ The preoperative bone models were matched on both the preoperative and postoperative RSA images. The analysis was first done in half resolution using a global optimizer, followed by a refined local optimization in full resolution while excluding the models’ surroundings.²² The precision of the bone model-based RSA analysis has been reported to be up to 0.095 mm for the humerus and 0.130 mm for the scapula.¹⁴ All recordings were analyzed by one person (JOK), who was blinded to the results.

Anatomical landmarks were manually applied to the preoperative bone models of the humerus and the scapula (Figure 2).^23,24 Since the preoperative bone models were used to analyze both the pre- and postoperative recordings, the coordinate system was identical for all data within each patient.

Fig. 2.

Anatomical landmarks and anatomical coordinate systems with kinematics in six degrees of freedom. a) The superior-inferior axis (green arrow, pointing up) was defined by the midpoint between the medial and lateral epicondyle, and the glenohumeral centre was defined by the centre of a best-fit sphere to the articulating part of the humeral head. The anterior-posterior axis (blue arrow, pointing outwards) was the line perpendicular to the plane formed by the epicondyles and the glenohumeral centre. The medial-lateral axis (orange arrow, pointing east) was the cross-product of the anterior-posterior and superior-inferior axes. b) The medial-lateral axis (orange arrow, pointing outwards) was defined as the line between the trigonum spinae and the glenoid centre, defined by a best-fit circle to the inferior part of the glenoid. The superior-inferior axis (green arrow, pointing up) was defined by an orthogonal projection from the medial-lateral axis towards the inferior angle of the scapula. The cross-product of the medial-lateral axis and superior-inferior axis defined the anterior-posterior axis (blue arrow, pointing outwards).

The GHJ kinematics during the apprehension-relocation test were calculated using two different methods: the ‘Centre method’ and the ‘Contact point method’ (Figure 3).¹³ The Centre method is defined as the location of the humeral head centre in the glenoid coordinate system. The Contact point method is defined as the weighted average proximity of the GHJ joint.²⁵ The postoperative glenoid surface model was extracted from the scapula and transformed with surface matching into the preoperative bone coordinate system. Both measures were evaluated for each test relative to the glenoid in anterior (+)/posterior (-) and superior (+)/inferior (-) directions.

Fig. 3.

Glenohumeral joint kinematic outcomes. a) The Centre method, marked with a red dot, was defined by the best-fit sphere on the humerus and the best-fit circle on the glenoid. b) The Contact point method, marked with a black dot, was defined by the weighted average proximity of the humeral head surface and the glenoid surface. The blue arrow represents the anterior-posterior direction, and the green arrow the superior-inferior direction.

As a secondary outcome, the area of the articular side (mm²) of the coracoid bone block was estimated by comparison of the pre- and postoperative scans, and again one year thereafter to evaluate the degree of osteolysis. The length and width of the resected coracoid processes were evaluated from surface-aligned preoperative and postoperative CT bone models to define the articular side area of the bone block when the Latarjet procedure was performed. The remaining articular side area of the bone block was assessed one year after the Latarjet procedure using the one-year postoperative CT bone model.

At inclusion and one year postoperatively, the patients completed the Western Ontario Shoulder Instability Index (WOSI) questionnaire, with a score range of 0 to 2,100, with 0 representing the best possible outcome. The minimal clinically important difference (MCID) is 220 points in patients with shoulder instability.²⁶ The WOSI score was transformed into a percentage score, resulting in a MCID of 10.4%.

Sample size

The sample size calculation was based on a study by Di Giacomo et al,⁹ which compared the kinematics between the Latarjet patients and healthy controls. Assuming a similar humeral head translation difference of 2.69 mm between healthy shoulders and in shoulders after the Latarjet procedure, a SD of 1.97, an alpha of 0.05, and a power of 0.8, ten patients in each group were required. The sample size was increased to 20 patients to accommodate for dropouts and technical problems.

Statistical analysis

Continuous data were checked for normal distribution by QQ plots and reported with 95% CI. The mean value of the two recordings in each position was calculated and used in the kinematic outcomes. A paired t-test was used for statistical comparison of the paired normally distributed data (Stata v18.0; StataCorp, USA). The level of significance was set at p < 0.05 in all analyses. The repeatability was evaluated across all kinematic examinations by a mean difference, SD, and 95% limits of agreement (95% LoA).

Results

Patient-reported outcomes

At one-year follow-up, all 19 patients completed the WOSI. The mean WOSI score was improved by 18.2 percentage points (95% CI 7.4 to 29.1) compared with their preoperative score. Six patients did not achieve a clinically relevant improvement (MCID < 10.4 percentage points).

Repeatability of the apprehension-relocation test

The repeatability between the two tests was high. For the Centre method, the mean difference in the anterior-posterior direction was 0.1 mm (SD 0.7, 95% LoA -1.4 to 1.5) and 0.0 mm (SD 0.4, 95% LoA -0.7 to 0.8) in the superior-inferior direction. For the Contact point method, the mean difference was 0.0 mm (SD 0.6, 95% LoA -1.2 to 1.2) in the anterior-posterior direction and 0.2 mm (SD 0.6, 95% LoA -1.1 to 1.4) in the superior-inferior direction.

GHJ kinematics during the apprehension-relocation test

Preoperatively, all patients reported apprehension feelings during the apprehension test. None of the patients reported the apprehension feeling when testing the contralateral healthy shoulder. Postoperatively, two patients reported positive apprehension and relocation during testing; one of the patients had GHJ kinematics outside the 95% CI of the cohort with a more anterior contact point and inferior humeral head centre. Data are presented in Figure 4 and Table II for the anterior-posterior kinematics, and in Table III for the superior-inferior kinematics.

Fig. 4.

Illustration of the main pre- and postoperative kinematic findings measured with the Centre and Contact point methods in the anterior-posterior and superior-inferior directions. Mean location with 95% CIs of the humeral head centre and contact point are shown. The column “Difference” is the difference between the humeral head centre or contact point locations in the apprehension-relocation test. The blue dots represent the location of the postoperative Latarjet shoulder, while the red dots represent the location of the preoperative anterior shoulder instability.

Table II.

Anterior-posterior kinematics for the Centre and Contact point methods.

Shoulder	Centre method			Contact point method
	Apprehension	Relocation	Difference	Apprehension	Relocation	Difference
Healthy shoulder	-5.4 (-6.7 to -4.2)	-4.9 (-6.3 to -3.5)	-0.6 (-1.8 to 0.6)	-1.2 (-2.1 to -0.3)	-1.2 (-2.0 to -0.5)	0.1 (-0.9 to 1.1)
Preoperative ASI shoulder	-5.0 (-6.2 to -3.8)	-5.3 (-6.7 to -3.9)	0.3 (-0.5 to 1.1)	0.3 (-0.5 to 1.1)	-0.9 (-2.0 to 0.1)	1.2 (0.2 to 2.2)*
Postoperative Latarjet shoulder	-5.7 (-7.2 to -4.2)	-5.8 (-7.4 to -4.2)	0.1 (-0.4 to 0.6)	-0.6 (-1.6 to 0.4)	-1.4 (-2.6 to -0.3)	0.8 (0.3 to 1.3)*
Difference between postoperative Latarjet shoulder, preoperative ASI shoulder, and healthy shoulder
Postoperative Latarjet shoulder vs healthy shoulder	-0.3 (-1.7 to -1.2)	-0.9 (-2.9 to 1.0)		0.6 (-0.8 to 2.0)	-0.2 (-1.6 to 1.2)
Postoperative Latarjet shoulder vs preoperative ASI shoulder	-0.7 (-1.4 to 0.1)	-0.5 (-1.1 to -0.0)*		-0.9 (-2.0 to 0.2)	-0.5 (-1.1 to 0.1)

Mean location of the centre of the humeral head and contact point in mm (95% CI) of the humeral head in the anterior (+) - posterior (-) direction relative to the centre of the glenoid.

^*Statistically significant differences.

ASI, anterior shoulder instability.

Table III.

Superior-inferior kinematics for the Centre and Contact point methods.

Shoulder	Centre method			Contact point method
	Apprehension	Relocation	Difference	Apprehension	Relocation	Difference
Healthy shoulder	7.6 (6.5 to 8.8)	7.8 (6.7 to 9.0)	-0.2 (0.7 to 0.2)	6.1 (4.5 to 7.6)	5.7 (4.5 to 6.9)	0.3 (-0.6 to 1.2)
Preoperative ASI shoulder	6.7 (5.6 to 7.9)	7.1 (5.8 to 8.3)	-0.3 (-0.8 to 0.1)	4.1 (2.5 to 5.6)*	4.4 (2.8 to 6.0)	-0.3 (-1.0 to 0.4)
Postoperative Latarjet shoulder	7.5 (6.4 to 8.7)	7.2 (6.0 to 8.5)	0.3 (0.1 to 0.5)†	4.7 (2.8 to 6.5)	4.0 (2.2 to 5.7)	0.7 (0.1 to 1.3)†
Difference between postoperative Latarjet shoulder, preoperative ASI shoulder, and healthy shoulder
Postoperative Latarjet shoulder vs healthy shoulder	-0.1 (-1.2 to 1.1)	-0.6 (-1.8 to 0.6)		-1.4 (-3.6 to 0.8)	-1.8 (-3.5 to 0.0)†
Postoperative Latarjet shoulder vs preoperative ASI shoulder	0.8 (0.1 to 1.4)	0.2 (-0.5 to 0.8)†		0.6 (-1.1 to 2.3)	-0.4 (-1.9 to 1.1)

Mean location of the centre of the humeral head and contact point in mm (95% CI) of the humeral head in the superior (+) - inferior (-) direction relative to the centre of the glenoid.

^*Statistically significant difference in the mean position for the patients with an engaging bone lesion (2.0 mm (95% CI 1.2 to 2.8)) and non-engaging bone lesion (4.6 mm (95% CI 2.8 to 6.5)).

^†Statistically significant differences.

ASI, anterior shoulder instability.

For the Centre method, no differences were found between the healthy and postoperative Latarjet shoulder either during the apprehension or the relocation test. During the apprehension test, the postoperative Latarjet shoulder was located 0.7 mm (95% CI -0.1 to 1.4) more posterior and 0.8 mm (95% CI 0.1 to 1.4) more superior than the preoperative ASI shoulder. During the relocation test, the postoperative Latarjet shoulder was located 0.5 mm (95% CI 0.0 to 1.1) more posterior than the preoperative ASI shoulder. The humeral head centre of the postoperative Latarjet shoulder was located 0.3 mm (95% CI 0.1 to 0.5) more superior during the apprehension test than the relocation test. No difference was found between the humeral head centre translation of the patients with engaging and non-engaging bipolar bone lesions.

For the Contact point method, no differences were found between the healthy and postoperative Latarjet shoulder in the anterior/posterior direction. However, during the relocation test, the contact point of the postoperative Latarjet shoulder was 1.8 mm (95% CI 0.0 to 3.5) more inferior than the healthy shoulder. During the apprehension test, the contact point of the postoperative Latarjet shoulder was 0.9 mm (95% CI -0.2 to 2.0) more posterior than the preoperative ASI shoulder. During the relocation test, the contact point of the postoperative Latarjet shoulder was located 0.5 mm (95% CI -0.1 to 1.1) more posterior than the preoperative ASI shoulder. Thus, the contact point of the postoperative Latarjet shoulder was not located on the coracoid bone graft in either the apprehension or the relocation test. The contact point of the preoperative ASI shoulder was 1.2 mm (95% CI 0.2 to 2.2) more anterior during the apprehension test than during the relocation test. This difference was reduced to 0.8 mm (95% CI 0.3 to 1.3) postoperatively. Furthermore, the contact point of the postoperative Latarjet shoulder was 0.7 mm (95% CI 0.1 to 1.3) inferior during the relocation test compared to the apprehension test. For the preoperative ASI shoulder during the apprehension test, the contact point was 2.7 mm (95% CI 0.7 to 4.6) more inferior for the patients with engaging bipolar bone lesions than for the patients with non-engaging bone lesions.

Bone block resorption

One year after the Latarjet procedure, the area of the articular side of the coracoid bone block was decreased by 63.9% (95% CI 57.9 to 70.0) from 259.5 mm² to 95.1 mm². The resorption was primarily seen at the superior part of the bone block (Figure 5).

Fig. 5.

CT bone models of the glenoid one year after the Latarjet procedure (left sides are mirrored for easy comparison).

Discussion

This study investigated the GHJ kinematics in patients undergoing the Latarjet procedure during an apprehension-relocation test. Overall, 17 out of 19 patients did not report apprehension during testing one year after the Latarjet procedure. The main finding was that the procedure restored the GHJ kinematics during the apprehension-relocation test in the anterior-posterior and superior-inferior direction, compared to the healthy shoulder. Compared to the preoperative ASI shoulder, the Latarjet procedure resulted in a more posterior and superior humeral head centre location, and a more posterior location of the GHJ contact point. During relocation, both the humeral head centre and the contact point were located more posteriorly than for the preoperative ASI shoulder.

This study is the first to examine the GHJ kinematics during an apprehension test in patients undergoing the Latarjet procedure. However, the GHJ kinematics following the Latarjet procedure have previously been examined in two studies with the shoulder in the ABER position.^9,10 Park et al¹⁰ used fluoroscopy and found that the humeral head centre was 5.39 mm more posterior in the operated shoulder than the contralateral healthy shoulder in the end-range ABER position. In the study by Di Giacomo et al,⁹ patients following the Latarjet procedure and controls were examined in an open MRI scanner. The patients were examined without and with a 5 kg load on the hand, where the 5 kg load resulted in an isometric contraction of the shoulder in the ABER position. During loading, the Latarjet group displayed a 2.69 mm more posterior location of the humeral head than the healthy controls, but no difference in the humeral head location in the superior-inferior direction. In the same study, without loading, no difference was found. In our study, the apprehension test was performed on the passive, relaxed shoulder. Similarly, we did not find any difference in the kinematics between the operated shoulders and the contralateral healthy shoulders. This may suggest that, in the ABER position without muscle activation, the GHJ kinematics following the Latarjet procedure is restored towards that of a healthy joint. With muscle activation through isometric contraction, the anterior stabilizing effect appears to increase beyond that of a healthy shoulder.

To our knowledge, this study is the first to examine the GHJ during a relocation test. During the relocation test, we found that the postoperative humeral head centre and the GHJ contact point were more posteriorly located than in the preoperative ASI shoulders. This observation is likely the result of the ‘sling effect’ from the conjoined tendon, which maintains the humeral head in a more posterior position.

The contact point of the GHJ in this study was determined by calculating the weighted average proximity between the humerus and glenoid. Given that the Latarjet procedure extends the anterior glenoid surface with a bone block, it could be expected that the contact point would shift more anteriorly. However, we found that the contact point was more posteriorly located than preoperatively during both the apprehension and relocation test, indicating limited contact between the humeral head and the coracoid bone block. This indicates that with the shoulder in the ABER position, the bone block may not contribute to the stabilization, even despite an anterior-directed load being applied during the apprehension test. The contribution of the different stabilizing mechanisms resulting from the Latarjet procedure has previously been examined in an experimental study measured with a potentiometer.⁸ Those authors showed that, in the end-range ABER position of the GHJ, the ‘sling effect’ contributed up to 77% of the stability, with the remaining 23% derived from the capsule. However, in the midrange position (60° abduction and neutral rotation), the ‘sling effect’ was estimated to be 51% to 62%, and the remaining part was due to glenoid reconstruction by the bone block. Our study found no correlation between the GHJ position and the location of the humeral head centre or contact point.

Studies have recommended that in the case of bipolar-engaging bone lesions, anterior glenoid bone reconstructions should be considered.^27,28 In this study, a sensitivity analysis revealed no statistically significant differences between preoperative and postoperative GHJ kinematics between the patients with engaging and non-engaging bone lesions.

Bone remodelling follows Wolff’s law, whereby bone adapts to the loads to which it is exposed.²⁹ The limited contact between the humeral head and the bone graft likely results in minimal loading of the graft, leading to stress shielding, which may contribute to the substantial bone block resorption of 63.9% observed one year after the Latarjet procedure. Additionally, transposition of the coracoid process may compromise vascularization, further promoting bone resorption. This finding is consistent with the findings by Di Giacomo et al,³⁰ who found a mean bone block resorption of 59.5% 17.6 months (SD 6.9) after the Latarjet procedure. Similarly, the largest resorption was found in the superior part of the bone block. Despite the large bone block resorption, the patients reported being satisfied with a Simple Shoulder Test score of 92.3% in the study by Di Giacomo et al,³⁰ and a WOSI improvement of 18.2 percentage points in our study. Other free bone grafting procedures have reported lower bone resorption rates.³¹ As such, the ‘sling effect’ of the conjoint tendon seems to provide the GHJ stability rather than the bony extension of the glenoid.

This prospective study design has allowed us to compare the effect of the Latarjet procedure with the preoperative instability within the same patient, thereby allowing for the investigation of interindividual kinematic differences. The GHJ kinematics were assessed using RSA, allowing us to identify both anterior-posterior and superior-inferior measures with high precision and repeatability. A limitation of this study is the lack of cartilage on the bone model for the calculation of the GHJ contact point. However, studies have shown a homogenous cartilage distribution of the humerus and glenoid, and therefore we assume that the addition of the cartilage would not change the contact point.³² Another limitation is that the RSA recording was only captured at the end-range position where the patient may, especially in the preoperative stage, have felt the apprehension and a reflexive muscle guard could potentially have caused the measured humeral head position to be in a more central position. This may have decreased the differences between the pre- and postoperative measurements. To evaluate the restoration of the GHJ, we have assumed that the contralateral healthy shoulder resembles the kinematics of the ASI shoulder before an injury. However, whether there are side-to-side differences between, for example, the dominant and non-dominant arms or different muscle guard and strength is unknown.

In summary, the Latarjet procedure restored the humeral head centre compared to the healthy contralateral shoulder. Compared to the preoperative ASI shoulder, the Latarjet procedure resulted in a more posterior and superior humeral head centre location and a more posterior location of the GHJ contact point. The contact point was posterior to the transferred coracoid bone block, indicating that the coracoid bone block itself is not used to increase glenoid support, but rather that the stabilizing effect of the Latarjet procedure relies on the ‘sling effect’. The stabilizing effect of the Latarjet procedure is affected by muscle activation, which is why further kinematic studies in an active dynamic clinical setting, enabling evaluation in several positions, are warranted.

Data Availability

The datasets generated and analyzed in the current study are not publicly available due to data protection regulations. Access to data is limited to the researchers who have obtained permission for data processing. Further inquiries can be made to the corresponding author.