Introduction

The aetiology of the anteroinferior glenohumeral instability is multifactorial [8, 45]. The satisfactory treatment of this condition requires that every surgical approach is flexible enough to allow the surgeon to identify and treat any anatomical lesion that could cause instability of the shoulder [42, 52]. Over the past years, numerous techniques have been proposed for arthroscopic stabilization. Despite the improvement of the techniques, the results of arthroscopic stabilization have not yet reached those of open techniques such as the Latarjet procedure or the anterior-inferior capsular shift. The failure rate of an isolated Bankart arthroscopic repair is reported to be 15–30 % when performed in unselected patient populations and when patients are followed for more than 2 years after surgery [30]. Glenoid erosion is a quite common occurrence associated with chronic anterior instability of the shoulder, along with Hills–Sachs lesion [56]. A relationship was found between the extension of glenoid bone defect and the result of treatment for anterior instability of the shoulder [14]. Anterior shoulder instability can be classified as recurrent dislocation, recurrent subluxation or unstable painful shoulder [72]. Dislocated shoulders show the highest incidence of bone lesions, while the painful shoulder group shows the lowest incidence. The prevalence of fractures or erosion of the anterior-inferior glenoid shoulders with recurrent dislocation was reported to be from 8 to 90 % [12, 41]. Lesions are caused by impact of the upper and rear portions of the humeral head on the anteroinferior glenoid region during the episode of dislocation. Severe bone lesions are associated with failure of soft tissue techniques and, until recently, represented the maximum real limit of the arthroscopic approach [14, 22]. Appropriate imaging of the unstable shoulder shows 90 % of bony lesions [20, 27, 64]. Many authors recommend the transfer of the coracoid if the glenoid defect affects 20 % or more of the anteroposterior diameter of the glenoid [7, 17, 40, 62, 65, 67]. The relationship between the width of the glenoid defect and the clinical result continues to be investigated. So far, the exact size of the glenoid deficit that contraindicates a soft tissue technique remains unknown. In the past, more than 100 different techniques have been proposed for stabilizing the shoulder, but only some are still used. A very popular soft tissue technique still in use has been described by Bankart [5] and includes an open [35, 38, 53] and arthroscopic approach [6, 57]. In cases with bone defects either in the glenoid or in the humeral head, bony procedures are indicated [46]. These procedures, which involve the placement of a bone graft (Eden–Hybinette) or the transfer of the tip of the coracoid (Bristow) or the whole coracoid (Latarjet) on the anterior-inferior glenoid, have traditionally been performed as open procedures [2, 31, 69, 71]. Only recently new arthroscopic techniques with placement of a bone block on the anterior-inferior region of the glenoid have been described [10, 37, 65, 66].

Glenoid bone loss

In 1998, Bigliani [9] defined prognostic features associated with glenoid deficit. The author described four types of glenoid defects: type 1: non-displaced anterior glenoid fragment; type 2: small detached anterior fragment; type 3a: anterior glenoid deficits <25 %; and type 3b: defects >25 %. For type 1, 2 and 3a, the author recommends a soft tissue reconstruction, while for type 3b a glenoid augmentation is suggested. Several biomechanics studies show how glenoid shape changes when the bone loss exceeds 20–25 %. Burkhart first called this glenoid appearance “inverted pear” [16] (Fig. 1). Gerber and Nyffeler [26] demonstrated how with subsequent loss of anterior-inferior glenoid arc the resistances to dislocation decreased exponentially. Recent biomechanical studies have also introduced the concept of “glenoid track” [18, 76]. This concept has shifted the previous paradigm from engaging defects to track-off track mismatch. On 9 cadaveric shoulders, Yamamoto showed how dislocation was most likely with disruption of the medial margin of this track [76].

Fig. 1
figure 1

Radiographic examinations show the bone loss of the glenoid with characteristic deformation to “inverted pear”

Full size image

Clinical evaluation

Medical history

Patients with anterior glenohumeral instability with bone lesions report different traumatic mechanisms. They often include a forced external rotation with the arm abducted at 90°. A similar pattern might occur after a fall on the arm extended and abducted. The distraction forces directly on the capsule and ligaments may be the cause of the injury. In patients with history of dislocation, the force that causes the damage is generally greater and more probably produces bony lesions associated with breakage of the soft tissues. Patients who require a reduction after the first episode of dislocation are more likely to have insertional injuries of the ligaments, while the bone defects become more common as a result of subsequent episodes. Patients who spontaneously reduce the dislocation after the first episode are more prone to develop a capsular elongation, and a hyperlaxity of the rotator interval without bone lesions. Laxity must be distinguished from the instability.

Physical examination

The diagnosis of unstable shoulder is rarely performed without a clinical history reported by the patient. The symptomatic instability is often diagnosed by history and confirmed by physical examination. It is important to distinguish between laxity and instability. Anterior hyperlaxity is defined as an external rotation with the arm adduct >85°. The lower hyperlaxity is defined when Gagey’s test is positive [23]. The test is defined as positive when the abduction of the affected limb with the scapula blocked by the examiner is >20° with respect to the other side. Instability is defined as the symptomatic expression of excessive translation of the humeral head. The presence of generalized laxity is more likely in an individual with multidirectional instability. Rarely, these patients have bone lesions associated with the instability.

Clinical tests can be divided into those that assess the laxity of the shoulder and those that assess the anterior instability of the shoulder.

Tests to evaluate shoulder laxity:

  • Drawer test [60]

  • Push–pull test

  • Sulcus sign

Tests for anterior instability:

  • Rockwood test

  • Rowe test [54]

  • Apprehension test

  • Fulcrum test [33]

  • Relocation test [29, 61]

  • Load and shift test [60]

In the literature, there is no consensus about which tests should be used in the diagnosis of unstable shoulder [58, 75], but a study by Sciascia shows that the four tests most commonly used are: apprehension, relocation, load and shift and anterior drawer.

Diagnostic value of these tests is reported in Table 1 [70].

Table 1 Diagnostic value of clinical tests mainly used for anterior shoulder instability diagnosis
Full size table

Imaging

Plain radiographs for instability include anteroposterior view performed in three rotations (internal, neutral and external) [73]. The presence of a Hill–Sachs lesion may be noted for each rotation (absent or present). If present in external rotation, its location is higher than the humeral head. Lesions on the glenoid are distinguished in fracture avulsion and loss of anterior-inferior sclerotic outline using a Bernageau projection for the glenoid profile [47]. The interruption of the anterior bone triangle compared to the contralateral is classified into three groups. Fractures are defined as an anomaly of the anterior glenoid rim characterized by a visible bone fragment fracture. The “cliff” sign is defined as the loss of normal anterior triangle without a visible bone fragment. The sign called “blunted angle” is defined as a shape of blunt apex of the triangle [21]. If an anterior glenoid deficiency is visible with radiographs, it is important to determine the area and the percentage of bone deficit [24]. Computed tomography (CT) scanning allows for precise evaluation of anterior glenoid bone loss in multiple planes. In a sagittal view, inferior glenoid bone loss can be appreciated as a percentage of its normal area. A best-fit circle is used to approximate normal inferior glenoid surface area and observed bone loss can be calculated from this measurement [55, 63]. Glenoid bone loss will lead to an “inverted pear” appearance. The glenoid assumes an inverted pear shape when the bone deficit minimum is at least 25 %. Hill–Sachs lesions, especially those that are subtle, can also be evaluated on CT studies [39]. Armitage et al. [3] reported that Hill–Sachs lesion is oriented towards 6:46 at the top and 8:56 at the bottom of the lesion, with the average orientation towards 7:58 ± 0:48 on a clock face with 12:00 defined as the intertubercular sulcus. Magnetic resonance imaging (MRI) with or without arthrography is frequently used to evaluate the chronically unstable shoulder. In addition to providing useful information about soft tissue anatomy, including the glenoid labrum, chondral surfaces, glenohumeral capsuloligamentous structures and the rotator cuff, MRI can also demonstrate bone loss [43, 51].

Arthroscopy evaluation

Glenoid bone loss can be also qualified and quantified with arthroscopy. The bare area has been shown to reliably mark the centre of the inferior glenoid [32, 44]. Using the bare area as a landmark, a calibrated probe can used to measure the distance from the bare spot to the posterior rim and compare it to the distance from the anterior rim. Assuming that the normal inferior glenoid is shaped as a nearly perfect circle [32], anterior-inferior glenoid deficiencies can then be quantified by the following [15]: glenoid deficiency = (distance from bare spot to posterior rim − distance from bare spot to anterior rim)/(2 × distance from bare area to posterior rim). Quantification of glenoid bone loss should be routinely performed preoperatively to determine the ideal anterior stabilization procedure.

Classification

Glenohumeral instability is classified according to four factors:

  • Time/frequency

  • Degree of instability

  • Direction

  • Aetiology

Time/frequency can be defined as acute or recurrent; degree of instability ranges from subclinical forms (grade 1) to real dislocations (grade 2/3); direction of instability can be anteroinferior, anterior direct, anterosuperior, posteroinferior and multidirectional; aetiology may be atraumatic, micro-traumatic (overuse) or traumatic. In 1989, Thomas and Matsen [68] classified instabilities in two groups: TUBS and AMBRI, which must be regarded as two extremes of a wide “range”. The first group, TUBS (traumatic unidirectional Bankart surgery), is an expression of unidirectional traumatic glenohumeral instability. The second group, AMBRI (atraumatic multidirectional bilateral rehabilitation inferior shift), includes atraumatic multidirectional bilateral instability.

Decision algorithm

The algorithms of treatment depend on many factors, but the size and type (fragment or erosion) of the glenoid bone deficit is the priority. If a mobile bone fragment is associated with a labral lesion, the possibility to perform an arthroscopic repair exists, despite the size of the fragment, especially in an acute or subacute condition. If there is a bone loss, to date there are no precise guidelines. If the bone deficit is >20 % with respect to the healthy contralateral glenoid, an open or arthroscopic bone grafting procedure is recommended by most authors to fill the defect and to reconstruct the anatomic glenoid arch [22, 28, 40–42]. If the missing area of the glenoid is <10 % and there are no soft tissue alterations, an arthroscopic soft tissue reconstruction is certainly a viable treatment option to restore the stability of the joint. If the bone loss is between 10 and 20 %, other factors should be considered, such as the presence of a Hill–Sachs lesion that could represent an indication for a bony procedure.

In addition to a classification of the possible presence of bone defects preoperatively, other risk factors that may preclude arthroscopic soft tissue stabilization should be considered. If the instability severity index score (ISIS) [4] (Table 2) is greater than 6 points, a reconstruction of the isolated soft tissues may be insufficient for the stability of the shoulder, especially in the long-term follow-up [13, 50] (Table 3).

Table 2 Scoring of the instability severity index score as described by Balg and Boileau [63]
Full size table
Table 3 Decision algorithm for the treatment of anterior traumatic instability with glenoid bone loss
Full size table

In conclusion, the preoperative evaluation of the bone loss, the clinical risk factors for recurrence of instability (ISIS scoring system), clinical examination and medical history of the patient can help the surgeon to better select patients for the appropriate surgical procedure.

Indications for surgery

Patients, who suffer a shoulder dislocation as a result of major trauma and who have no ligamentous laxity, benefit most from surgical treatment [48, 49]. Recurrence of instability is the main complication after anterior stabilization. Currently, most surgeons perform arthroscopic stabilization of the soft tissue with anchors and sutures due to more reproducible results. However, even after the recent technical developments, there is still a rate of recurrence of about 5–20 % [1, 8, 42, 59]. The best way is to identify preoperatively patients whose risk factors preclude arthroscopic soft tissue stabilization. The literature reports numerous prognostic factors. Athletes who practice contact sports have a higher incidence of recurrence after a classic Bankart stabilization. Patients with significant glenoid bone loss or with an unacceptable high clinical risk of recurrent dislocations are candidates for procedures with bone grafting, either with an open or arthroscopic approach [77].

Techniques

Arthroscopic Latarjet procedure

In this procedure, described the first time by Lafosse et al. [37], the stabilization mechanisms are as follows:

  • Increase the glenoid surface

  • Muscle-tendinous hammock effect created by the conjoint tendon passing through the subscapularis, thus creating a dynamic tension when the shoulder is in abduction and external rotation

  • Suturing the inferior glenohumeral ligament to the coracoacromial ligaments attached to the coracoid

In order to achieve a good stabilization of the joint, it is mandatory that the coracoid has a perfect alignment with the glenoid rim. A medial positioning of the graft would create unfavourable biomechanical conditions with higher recurrence rate of instability; a lateral position of the coracoid with an overlap of the graft in the joint leads to a premature osteoarthritis [37].

Surgical technique [37]

Patient is positioned in beach-chair position. A diagnostic arthroscopic evaluation is initially performed. Soft tissues between the two and five o’clock positions are resected with a radiofrequency. The coracoacromial ligament and pectoralis minor are detached from the coracoid. The coracoid is cleared of soft tissues circumferentially to its base while protecting the attachment of the conjoint tendon at the coracoid tip. Two holes are drilled and tapped over guide wires, through a portal superior to the coracoid. Then, the osteotomy of the coracoid is completed by the use of a curved osteotome. The surgeon splits the subscapularis at the junction of its inferior third and superior two-thirds. The split is completed medially via blunt dissection with a trocar and external rotation of the arm. The anterior glenoid face is then prepared with a bur used to create a flat bed of bleeding bone. The coracoid is then retrieved. The inferior surface of the coracoid is decorticated with a bur, creating a flat surface to match the anterior glenoid. The coracoid is then manipulated via a double-barrel cannula, through the subscapularis split, and onto the glenoid face between the two and five o’clock positions. Optimal positioning of the coracoid is parallel to the glenoid rim. Once graft positioning is verified, a 3.2-mm cannulated drill is used to predrill each hole before insertion of both 3.5-mm cannulated screws, beginning with the inferior screw. With the arm in abduction and external rotation, the sling effect of the conjoint tendon can be observed.

Indications

  • Anterior glenoid bone loss

  • Ligamentous tissues of poor quality

  • Revision surgery

  • Involvement in collision sports

Arthroscopic Bristow–Latarjet procedure

The arthroscopic procedure according to Bristow–Latarjet was described for the first time by Boileau et al. [10].

The effectiveness of the procedure is linked to a “triple lock”:

  • The effect of the bone block by the coracoid tip increases the extension of the glenoid surface

  • The sling effect caused by the passage of the conjoined tendon through the subscapularis

  • The reproduction of the glenoid concavity due to the labral repair

Surgical technique [10]

The patient is positioned in the beach-chair position with no traction. Shoulder arthroscopy is performed with a standard posterior and anterosuperior portal. The shoulder joint is visually assessed for lesions consistent with anterior instability. In order to harvest and prepare the tip of the coracoid, the scope is placed in the anterior subdeltoid space. An anterolateral portal located 2 cm lateral to the anterosuperior portal is established and is used as a viewing portal for this step of the procedure. The coracoid process and the conjoined tendon insertion are identified, and fibrous tissue from the neck of the scapula is removed. The coracoacromial ligament insertion and the pectoralis minor insertion are partially divided. A coracoid fragment measuring 15 mm in length is then removed. The bone fragment is then brought outside the incision for preparation. An adsorbable suture is passed through a drill hole in the coracoid fragment and through the coracobiceps tendon. Once the coracoid fragment is prepared, an arthroscopic cannula 8 mm in diameter is inserted in the anterosuperior portal, passing above the subscapularis in the rotator interval. An arthroscopic Bankart repair is then performed. Under arthroscopic vision, the glenoid neck is penetrated with a sharp-tipped awl, which prevents sliding of the guidewire along the cortical bone of the glenoid neck when drilling. A guidewire is then placed and oriented parallel to the glenoid articular surface. The guidewire is drilled until it just penetrates the posterior cortex of the glenoid. The guidewire is then overdrilled with a 10-mm cannulated reamer to a depth of 15 mm. The reamer and guidewire are then removed. A Beath pin pull-through technique is used for coracoid placement. This pin has an eyelet on its trailing end and serves as a suture passer. The Beath pin is placed into the glenoid socket and the previous hole created by the guidewire and is recovered behind the shoulder. Both ends of the suture placed previously through the coracoid fragment and the coracobiceps tendon are passed through the eyelet of the Beath pin and then recovered behind the shoulder. Progressive traction on the suture makes it possible to pull the coracoid fragment into the glenoid socket. Before the entire coracoid fragment enters the glenoid socket, a flexible guidewire for the interference screw is inserted to prevent screw divergence. The coracoid graft is then pulled inside the glenoid socket by traction on the posterior suture. The graft is then fixed in the hole by use of a 7 × 15 mm bioabsorbable interference screw, inserted over the flexible guidewire.

Indications

  • Anteroinferior glenoid bone loss

  • Ligamentous tissues of poor quality

  • Revision surgery

Arthroscopic bone graft procedure

The arthroscopic bone block technique was originally described by Taverna [66] (Fig. 2) and has been recently modified [65]. In the new technique, the bone graft is fixed with four endobuttons. The technique combines the Bankart repair with the transfer of the graft, harvested from the iliac crest, which is inserted through the cannula in the rotator interval and fixed on the glenoid rim. The effectiveness of this procedure is related to the effect of the bone block produced by the tricortical graft, which increases the area of the glenoid and the reproduction of the glenoid concavity produced by the insertion of the labrum on the glenoid rim and the anterior-inferior shift of the capsule and the anterior-inferior glenohumeral ligament. The purpose of this procedure is to restore the normal anatomy of the unstable shoulder.

Fig. 2
figure 2

Radiographic results at 2-year follow-up of previous surgical technique for arthroscopic bone block. The two images on the left show a CT-scan 3D reconstruction of the shoulder, and the image on the right shows an anteroposterior standard radiograph. The graft is correctly positioned, and there are no signs of resorption, non-union or breaking of the fixation devices

Full size image

Surgical technique [65]

Patient is placed in the beach-chair position. A standard posterior portal is created for the insertion of the arthroscope. Viewing from the posterior portal, the surgeon creates an anterior superior portal and a midglenoid portal. The labrum is detached, and all soft tissues are removed from the anterior glenoid neck. Then, the arthroscope is introduced through the anterosuperior portal, the anterior glenoid rim is further decorticated and the glenoid neck is prepared with a motorized bur to create a flat and bleeding bony surface (Fig. 3). A spinal needle is inserted from posterior to anterior along the face of the glenoid and centred on the anterior glenoid bone defect. Then, a more posteromedial portal is made to provide access to the glenoid guide. The hook end of the glenoid guide is inserted and passed along the glenoid parallel to the glenoid face to avoid damaging the articular surface, and then, it is passed over the anterior edge. Then, the guide is rotated to capture the anterior edge of the glenoid under the hook. The hook should be placed at the centre of the anterior glenoid defect, usually between the three and four o’clock positions. It is mandatory that the glenoid guide be aligned with the posterior and anterior glenoid rim.

Fig. 3
figure 3

Arthroscopic view from anterior portal. The anterior glenoid rim is decorticated and flattened with a motorized burr to create a flat and bleeding bony surface to accommodate the graft

Full size image

Then, a bullet is placed in the inferior hole of the guide and is advanced until it firmly contacts the posterior aspect of the glenoid neck. The ratchet teeth of the bullet should be aligned with the screws adjacent to the guide handle. The process is repeated for the superior bullet. A 2.8-mm sleeved drill is placed in each bullet and advanced under power until exiting from the anterior aspect of the glenoid. The drills are placed 5 mm on the centre below the cortical edge of the glenoid face, parallel to one another and 10 mm apart. The inner drill is removed, leaving the cannulated outer sleeve. Once drilling is completed, the bullets can be removed. The guide can be removed at this stage. Flexible looped guidewires are then introduced into the joint by passing one wire through each sleeve posterior to anterior. Each guidewire is retrieved using a loop grasper. The wires are separated and stored. The drill sleeves should be removed after this step is completed.

The tricortical bone graft could be an autograft (harvested from the ipsilateral anterior iliac crest) or an allograft tricortical iliac crest graft. Two 2.8-mm drill holes are made 10 mm apart and 5 mm from each edge. The drill enters through the cortex and exits the cancellous side of the bone graft. The holes created correspond to the distance of the cannulated drill sleeves previously placed in the glenoid neck. With a marking pen, we colour the superior aspect of the bone graft.

Each looped guidewire is fed through the prepared bone graft and exits on the cortical side. The graft is oriented so that the cancellous surface is facing the anterior neck of the glenoid. The anterior implant is fed with the preassembled suture through the end of the looped guidewire with a classic slip knot. This can be achieved by passing the lead suture through the looped guidewire and feeding the implant through the lead suture. The graft is slid towards the end of the guidewires to lodge the implants. Anterior round endobuttons are advanced until they lie flat on the bone graft. Sutures should be taut to allow smooth movement down the cannula. The bone graft is tipped to be inserted into the 10-mm cannula, and care is taken to ensure that the superior end of the bone graft enters the cannula first. The graft is advanced by pulling the guidewires out posteriorly. Slight tension should be maintained on the sutures throughout this step. The sutures should advance the implant until the bone graft sits flush on the anterior neck of the glenoid, with each implant’s lead suture exiting the skin posteriorly. The posterior implants are placed on the transporter by advancing the instrument through the eyelet of a posterior round endobutton. The suture is passed through the transporter. The transporter is retracted to allow the suture to pass through the eyelet of the posterior round endobutton. The same steps must be performed for the second eyelet with the other side of the suture. The posterior round endobuttons are advanced until they sit flush against the posterior face of the glenoid. The knot pusher is used to secure the posterior round endobuttons. The knot pusher will provide tactile feedback when the posterior round endobuttons are properly seated.

Once the implant has been tensioned, the posterior knots are secured with half hitches, and the remaining sutures are cut using a blind knot cutter (Fig. 4).

Fig. 4
figure 4

Arthroscopic view from anterior portal. The bone block is perfectly flush with the anterior glenoid rim

Full size image

The anterior labrum, capsule and ligaments are repaired to the glenoid rim with suture anchors and a standard arthroscopic soft tissue repair technique (Fig. 5a–b).

Fig. 5
figure 5

a CT-scan 3D reconstruction and b anteroposterior radiograph at 2 years of follow-up. The graft is correctly positioned, there are no signs of bone resorption or non-union, and the patient reported no further dislocation of the shoulder

Full size image

Indications (Table 4)

Table 4 Indications for arthroscopic bone graft procedure
Full size table

  • Anteroinferior glenoid bone defect

  • Revision surgery

  • Involvement in extreme sports

Complications

The three surgical techniques have some common complications, which can be divided into intra- or post-operative complications (Table 5).

Table 5 Complications related to arthroscopic surgical procedure for anterior traumatic instability with glenoid bone loss
Full size table

Rehabilitation protocol

Rehabilitation protocol is the same in the three above-mentioned surgical procedures. After surgery, the shoulder is immobilized in a brace in abduction of 10° for 3–4 weeks. There are no limitations regarding passive movements after immobilization, and patients are subsequently allowed to regain full elevation and external rotation. After complete healing of the wound, pool exercises and return to work activities are authorized. Progressive strengthening exercises are started after 6–8 weeks. Return to contact sports and overhead mobility are generally allowed 4–6 months after surgery [25, 74].

Clinical results

Arthroscopic Latarjet procedure

In the first study in 2007, Lafosse et al. [37] reported the preliminary results in 44 patients. No complications were found, and he reported excellent clinical results. In a more recent study, the same author reported results of 100 cases [36]: 80 % of patients had excellent clinical results, 18 % good and 2 % a bad clinical outcome. Most patients returned to work at 2 months and returned to sports at 10 weeks after surgery. The graft positioning, evaluated with CT scanning, was flush with the glenoid in 80 %, medially placed in 8 %, and there was lateral overhang in 12 % of cases. On a vertical plane, the positioning was defined perfect (graft placed between three and five o’clock) in 78 % of cases, too high in 7 % and too low in 5 % of cases. Screw angle in relation to the glenoid face was on average 29°. Perioperative complications included two haematomas: one intraoperative fracture of the graft, and one transient musculocutaneous nerve palsy that fully recovered. Late complications included four cases of coracoid non-union. A further three shoulders were found to have lysis around the screws leading to prominence of the head of the screw. In total, four patients required a subsequent arthroscopic screw removal. At 26 months, an average loss of 18° of external rotation was reported. There were no cases of recurrent dislocation. In a recent study, Dumont et al. [19] evaluated patients who had undergone the arthroscopic Latarjet procedure at a minimum 5-year follow-up. Mean follow-up time was 76.4 months. One patient (1.59 %) had recurrent instability (subluxations) after the procedure. A recent prospective study evaluated with a multiplanar bidimensional computed tomography scan analysis 95 shoulders that had undergone the arthroscopic Latarjet procedure [34]. The coracoid graft resulted accurately positioned relative to the equator of the glenoid surface in 87 of 95 shoulders (91.5 %). Accurate bone block positioning on the axial view with “circle” evaluation was obtained for 77 of 95 shoulders (81 %). The mean screw angulation with the glenoid surface was 21°. One patient had transient axillary nerve palsy. Of the initial 104 patients, 3 (2.8 %) underwent revision. The author concluded that accurate positioning of the bone block onto the anterior aspect of the glenoid is possible, safe and reproducible with the arthroscopic Latarjet procedure without additional complications compared with open surgery.

Arthroscopic Bristow–Latarjet procedure

In 2010, Boileau et al. [10] treated 47 patients with glenoid bone loss and capsular deficiency with the arthroscopic Bristow–Latarjet technique. The procedure was performed entirely arthroscopically in 41 of 47 patients (88 %); a conversion to open surgery was needed in 6 patients (12 %). The axillary nerve was identified in all cases, and no neurologic injuries were observed. No patient had any recurrence of instability at the most recent follow-up (mean, 16 months). The mean Rowe score was 88, and the mean Walch–Duplay score was 87.6. The Subjective Shoulder Value was 87.5 %. The bone block was subequatorial in 98 % of the cases (46 of 47) and flush to the glenoid surface in 92 % (43 of 47); it was too lateral in 1 (2 %) and too medial (>5 mm) in 3 (6 %). There was 1 bone block fracture and 7 migrations. In a recent study, the same authors [11] evaluated 70 patients who had undergone arthroscopic Bristow–Latarjet at a mean follow-up of 35 months; 69 of 70 (98 %) patients had a stable shoulder, external rotation with arm at the side was 9° less than the non-operated side, and 58 (83 %) returned to sports at preinjury level. On the latest radiographs, 64 (91 %) had no osteoarthritis, and bone block positioning was accurate, with 63 (90 %) being below the equator, and 65 (93 %) flush to the glenoid surface. The coracoid graft healed in 51 (73 %), it failed to unite in 14 (20 %) and graft osteolysis was seen in 5 (7 %). Bone block non-union/migration did not compromise shoulder stability but was associated with persistent apprehension and less return to sports. The authors concluded that the arthroscopic Bristow–Latarjet procedure combined with Bankart repair for anterior instability with severe glenoid bone loss restored shoulder stability, maintained ROM and allowed return to sports at preinjury level. Adequate healing of the transferred coracoid process to the glenoid neck is an important factor of avoiding persistent anterior apprehension.

Conclusions

Anterior glenoid bone loss can be considered a difficult orthopaedic problem that leads to recurrent shoulder instability. Biomechanical studies have found an inverse relationship between the size of the glenoid defect and the stability of the shoulder: the larger the defect, the less stable the shoulder [42, 66].

Glenoid bony reconstruction is certainly needed for defects >20 %, but several other factors could indicate the need for an anterior glenoid bone grafting as a coexisting Hill–Sachs lesion or clinical conditions that increase the risk of recurrent instability (patients with an ISIS score higher than 6). To date, there is no consensus when and how bony procedures are needed to stabilize a glenohumeral joint. Further investigations are needed to determine when an isolated soft tissue repair in an unstable shoulder is not indicated. The rationale for a bony procedure is to provide an increase in the glenoid arc. Many surgical procedures exist such as coracoid transfer, tibial autograft, iliac crest autograft or allograft transfer, or osteochondral allograft.