➢ Arthroscopic labral repair for the treatment of anterior shoulder instability associated with osseous defects has been shown to have variable results, and additional procedures involving transfer of the coracoid process can provide successful outcomes.
➢ Computed tomography (CT) scanning should be used for preoperative planning and the assessment of glenoid bone loss. Patients with ≥30% bone loss in the glenoid need bone graft for shoulder instability.
➢ Proper placement of the bone graft (1) flush with the surface of the glenoid and (2) within 10° perpendicular to the face of the glenoid decreases the potential for shoulder arthritis and suprascapular nerve injury, respectively.
➢ The rate of complications associated with glenoid bone-grafting, including recurrent instability or neurologic injury, has ranged from 5% to 25%. Traction injury to the musculocutaneous and axillary nerves is the most common neurologic complication.
Shoulder instability is most commonly seen in young athletes and may require operative treatment to prevent recurrent dislocations. Successful surgical treatment of shoulder instability depends on an awareness of various factors that can affect outcome, such as soft-tissue and osseous changes, the length of time since the injury, previous operations, and compliance with postoperative guidelines1.
The prevalence of recurrent instability after operative treatment has been reported to range from 8% to 18%2. One of the reasons for such a recurrence is glenoid bone loss3. Arthroscopic labral repair for the treatment of anterior shoulder instability associated with substantial osseous defects has been shown to have variable results, and additional procedures may be needed2,4.
Osseous defects can be categorized as either an acute glenoid fracture or a glenoid bone defect due to attrition3. Glenoid defects associated with recurrent instability can be divided into three groups: type I (displaced avulsion fracture), type II (malunited avulsion fracture), and type III (erosion of the glenoid rim)5. Loss of the anteroinferior aspect of the glenoid decreases the articular arc, leading to recurrent instability. An inverted-pear appearance of the glenoid suggests substantial glenoid bone loss6. Investigators have defined the critical limit beyond which glenoid bone loss may lead to recurrent instability as 25% to 30% bone loss at the level of the bare spot or >20% to 25% bone loss at the inferior glenoid surface7,8.
Structural bone-grafting to restore the anatomy of the glenoid has been associated with good results. Autogenous iliac crest bone, coracoid bone, or allograft may be used. Autogenous iliac crest bone may be associated with additional morbidity9. The Latarjet procedure, first described in 1954, involves transfer of a portion of the coracoid along with the attached conjoined tendon to the anteroinferior aspect of the glenoid rim10. This procedure has demonstrated reliable results when used for the treatment of shoulder instability associated with glenohumeral bone loss and capsular deficiency11-18. The coracoid process is affixed lengthwise and is held with two screws. However, there is some confusion in the literature as previous reports have described the results of the Bristow procedure. The Bristow procedure involves transfer of only the tip of the coracoid distal to the insertion of pectoralis minor along with the conjoined tendon19. In contrast, the Latarjet procedure involves transfer of most of the horizontal process of the coracoid distal to the coracoclavicular ligaments (Fig. 1). The Latarjet procedure is thought to restore shoulder stability through three factors: (1) extra-articular fixation of the coracoid segment next to the glenoid defect at the scapular neck in order to restore the anteroposterior diameter, (2) creation of a sling effect with the conjoined tendon to support the anteroinferior aspect of the capsule, and (3) dynamic contraction of the conjoined muscles. The purpose of this article is to review the surgical anatomy, biomechanical characteristics, outcomes, and complications of the Latarjet procedure.
The osseous dimensions and the soft-tissue attachments on the coracoid process have been described in the literature20,21. The ligaments attached to the coracoid are the coracoacromial ligament and the coracoclavicular ligaments. The conoid and trapezoid ligaments (the coracoclavicular ligaments) are attached laterally and superiorly to the coracoid, respectively. The coracobrachialis, biceps brachii, and pectoralis minor muscles attach inferiorly and medially. The mean coracoid transfer length (i.e., the length of the bone segment from the tip of the coracoid to the anterior edge of the coracoclavicular ligament) is 28.5 mm21. The mean coracoid height is 8.0 mm, and the mean coracoid width is 13.6 mm; thus, attaching the coracoid process with its inferior surface along the glenoid face imparts more surface area. The coracoacromial ligament occasionally may overlap the coracoclavicular ligament footprint on the coracoid; thus, care should be taken while releasing the coracoacromial ligament during the coracoid transfer22,23.
Musculocutaneous nerve palsies and brachial plexus traction injuries have been reported in association with the Latarjet procedure. The musculocutaneous nerve penetrates the coracobrachialis muscle an average of 56 mm from the tip of the coracoid process; however, small proximal twigs innervating the muscle penetrate higher, at an average of 31 mm. If the proximal twigs are counted, the nerve can be found to penetrate the coracobrachialis muscle within 5 cm of the coracoid 74% of the time. It also has been noted that the proximal twigs can form a complex binding tether to the muscle at several points of fixation and potentially can contribute to traction injuries24.
Suprascapular nerve injury, although rare, also has been reported25. The mean distance from the superior screw to the major branch of the suprascapular nerve is 4 mm19. It is recommended that screw placement be within 10° of perpendicular to the face of the glenoid in the axial plane to avoid potential for nerve injury26 (Fig. 2).
Standard shoulder imaging for patients with recurrent dislocations includes anteroposterior, scapular lateral, and axillary radiographs. These radiographs are not accurate for demonstrating glenoid bone loss but may show a displaced bone fragment or a large Hill-Sachs lesion27,28. Other views such as apical oblique20, Didiée29, or West Point27,30 radiographs provide better analysis of the glenoid. The bone loss noted on the West Point radiograph correlates positively with the loss of glenoid width on computed tomography (CT)31. However, it is often difficult to make standardized radiographs in a clinical setting27. Additional radiographic evaluation of the humerus for a Hill-Sachs lesion can be performed with use of the Stryker notch view or an anteroposterior view with the humerus internally rotated.
CT scanning is useful for evaluating and detecting rim fractures as well as for estimating bone loss. In a cadaveric study in which glenoid bone loss was correlated with the findings on CT scans, 50% loss of the glenoid width at the inferior one-fourth slice on the CT corresponded with an actual 21% glenoid osseous defect on the face. Three-dimensionally reconstructed CT scans provide enhanced quantification of glenoid deficiency and have been shown to accurately predict glenoid bone loss32,33. Currently, en face three-dimensional CT views are preferred for the evaluation of glenoid osseous defects34. Glenoid deficiency is quantified as a percentage of the normal inferior glenoid surface area35,36. A circle is drawn to best fit the inferior two-thirds of the three-dimensional glenoid image and the amount of bone missing from the circle is determined with use of digital measurements. The glenoid index is defined as the ratio of the maximum inferior diameter of the injured glenoid compared with the maximum inferior diameter of the uninjured, contralateral glenoid16. Another technique measures the bone loss as the ratio of the inferior width of the injured glenoid to the predicted preinjury width37. The preinjury width is determined by multiplying the length from the center of the glenoid to the posterior margin of the glenoid at the 3-o’clock position by 2. This technique eliminates the need for CT scanning of the contralateral shoulder.
Magnetic resonance imaging (MRI) arthrography may be performed in addition to CT scanning to assess the labrum and the status of the capsule.
A glenoid bone defect involving >21% of the glenoid width substantially decreases the force required to subluxate the humeral head in an abducted and externally rotated position7. Gerber and Nyffeler demonstrated that a loss of more than half of the anteroposterior diameter of the inferior portion of the glenoid decreases the resistance to anterior dislocation by >30%21. In another biomechanical study, Allain et al. compared the results of anterior translational testing under four different conditions: (1) anteroinferior capsulotomy, (2) anteroinferior glenoid defect, (3) transplantation of a contoured bone graft, and (4) the Latarjet procedure11. The Latarjet procedure increased glenohumeral stability by 350% and demonstrated maximum effectiveness in the 60° abducted and 30° externally rotated positions. In comparison, a transplanted structural bone graft restored stability by 179% and had the lowest effect in external rotation at 60° of abduction. Through sequential loading of muscles and tendons in cadaveric specimens, Halder et al. tested the effect of the dynamic contribution of the conjoined tendon in inhibiting anteroinferior translation of the humeral head and found that the conjoined tendon had the capability to translate the humeral head superiorly38. This finding is similar to those reported by Jiang et al., who also found that the transfer of the conjoined tendon along with the coracoid provided a substantial stabilizing effect on shoulder stability37.
Malpositioning or non-optimal positioning of the glenoid bone block has been recognized as a source of arthritis and shoulder pain18. The effect of malpositioned bone blocks (either proud or recessed) on glenohumeral contact pressures was evaluated by Ghodadra et al. in a study of cadaveric shoulders39. The authors found that glenohumeral contact pressure was optimally restored with a flush bone block. Bone grafts that were placed 2 mm proud increased peak pressure in the anteroinferior quadrant as well as the contact pressure in the posterosuperior quadrant of the glenoid. Grafts placed in a 2-mm-recessed position led to increased pressure and edge-loading in the anteroinferior quadrant. Also, the inferior surface of the coracoid process provided greater surface area and increased congruity with the native glenoid than the lateral surface did.
Technique for Open Latarjet Procedure
The patient is placed in the beach-chair position with a towel roll under the scapula. The deltopectoral approach is used to expose the coracoid process. The coracoacromial ligament, the trapezoid ligament, the pectoralis minor tendon, and the conjoined tendon are identified (Fig. 3). The pectoralis minor tendon is released on the lateral edge of the coracoid process. The coracoacromial ligament is released on the medial side, 1 cm from its attachment on the coracoid; this ligament will be sutured to the capsule later in the procedure. A Hohmann retractor is placed over the superior portion of the coracoid to expose the trapezoid ligament. The osteotomy is initiated distal to this ligament with a saw and is completed with use of an osteotome or a saw (Fig. 4).
The subscapularis muscle may be split at the middle or lower third of the muscle belly40 or, alternatively, a small portion of the tendon may be released from its insertion41. External rotation of the arm is useful for identifying the plane between the subscapularis and the underlying capsule. The humeral head is retracted with use of a Fukuda retractor. The glenoid is exposed, and a Hohmann retractor or an anterior glenoid retractor is placed. Proper placement of retractors is vital to the procedure as the exposure is limited. The anterior portion of the glenoid is then prepared with a burr or an osteotome to create a bleeding cancellous surface. The inferior or lateral surface of the coracoid may be placed against the glenoid bone. In the classic Latarjet technique, the inferior portion of the coracoid is placed against the glenoid. The modification of this procedure, called the “congruent-arc” method, was described by Burkhart et al.41. In that technique, the lateral surface of the coracoid is placed against the anterior neck of the glenoid so that the inferior surface of the coracoid is flush with the glenoid (Fig. 5). The congruent-arc modification of the Latarjet technique was first described by Burkhart et al.41.
Some biomechanical studies have indicated that the congruent-arc method may restore the contact surface better than the classic technique39,42. With the congruent arc method, two holes are drilled through the coracoid. Proprietary instruments are available for drilling these holes at a fixed distance and angle to aid placement of the graft flush with the glenoid. The graft is then held against the glenoid surface, and the screws are inserted for secure fixation. The screws should not be overtightened as there is a risk of fracture of the graft. If the graft is proud (lateral to the glenoid), it should be trimmed with a burr. The capsule may be plicated or simply closed40,41. It also may be attached to the remainder of the coracoacromial ligament on the coracoid process. Inserting anchors along the glenoid rim and suturing the capsule over them may impart additional stability. In such cases, the coracoid process is extra-articular, which reduces the “abrasive” effect of the humeral head over the coracoid process41.
The postoperative regimen consists of two to four weeks of treatment with the arm in a sling. On the first postoperative day, the patient is allowed isometric scapular shrugs, pendulum exercises, and elbow flexion and extension and is encouraged to use the hand for activities of daily living such as eating, drinking, dressing, and holding a book to read. At three to four weeks, the patient may begin formal therapy with a physiotherapist focusing on passive-assisted external rotation and forward elevation. Strengthening exercises are allowed at three months postoperatively. Contact sports and heavy labor are allowed at three to six months, depending on the progress of the patient and on radiographic evidence of satisfactory healing of the graft3,17,41,43-45. A CT scan may be repeated to confirm healing before the patient is permitted to perform all activities (Fig. 6).
Allain et al.11 reviewed the records for fifty-six patients (fifty-eight shoulders) who had undergone the Latarjet procedure (Table I). The mean duration of follow-up was 14.3 years. Glenoid bone loss was seen in 96% of the shoulders preoperatively. Good to excellent results were reported in 88% of the shoulders. No episodes of recurrent dislocation were noted; however, 12% of the shoulders did have symptoms of instability. Sixty-two percent of the shoulders developed glenohumeral arthritis, with an average loss of 21° of external rotation. Improper placement of the coracoid graft (proud to the native glenoid) and a coexistent rotator cuff tear were thought to be associated with glenohumeral arthritis.
Burkhart et al.41 reported on 102 patients who had been managed with the Latarjet technique over a period of six years. Five (4.9%) of the 102 patients reported recurrent dislocation or subluxation. Only forty-seven patients were available for follow-up. There were no recurrent dislocations or subluxations among those forty-seven patients, although one patient had a positive apprehension sign. An average 5.1° loss of external rotation was noted. Osteoarthritis was not evaluated. Shoulder outcome scores showed substantial improvements after the procedure.
Lafosse et al.46 were the first investigators to describe an all-arthroscopic Latarjet procedure, to our knowledge. Their initial study evaluated the results for forty-four patients. Lafosse and Boyle17 recently reported the results for ninety-eight patients (100 shoulders) at eighteen and twenty-six months. All patients had returned to work at a mean of two months, and 98% described results as excellent or good. At twenty-six months, only thirty-five patients were available for clinical review; however, all reported excellent or good scores. The average loss of external rotation was 18°. Eleven percent of the patients had progressed by one stage of arthrosis. CT scanning revealed that the vertical positioning of the graft was perfect (3 to 5 o’clock) in 78% of the patients and that the graft was flush with the glenoid in 80%.
Schmid et al.47 recently reported on the use of the Latarjet procedure for the treatment of recurrent shoulder instability in a study of forty-nine patients with a mean duration of follow-up of thirty-eight months. Eighty-eight percent of the patients reported excellent results. No patient reported redislocation, two reported subluxation, and five reported subjective unspecific shoulder symptoms, corresponding with an overall failure rate of 14%. The mean subjective shoulder value significantly increased from 53% to 79% (p < 0.001). Seventy-six percent of the patients were pain-free at the time of the latest follow-up. Suboptimal graft placement was associated with the formation and progression of osteoarthritis in 27% of the patients.
Hovelius et al. presented a series of studies examining the long-term outcomes for patients who were managed with the Bristow procedure (Table I). The first study evaluated the outcomes for 118 patients who had had surgery between 1980 and 200114. The patients were followed for a mean of fifteen years. Ninety-eight percent of the patients were satisfied with the result. Sixteen patients (13.6%) reported recurrent instability during the follow-up period; however, only one required revision surgery. The second study analyzed 115 of the same patients radiographically to determine the prevalence of glenohumeral arthritis13. Forty percent of the shoulders demonstrated glenohumeral arthritis, and an average 11° loss of external rotation was noted. However, the authors were unable to identify a correlation between the loss of rotation and the development of glenohumeral arthritis. The prevalence of postoperative glenohumeral arthritis in that series was similar to that reported by Rahme et al.48, although in the latter series glenoid reconstruction was performed with use of iliac crest grafts. The third study was based on three series of Bristow repairs that were performed from 1980 to 200449. A total of 319 shoulders were included in the analysis: the first series comprised the previously mentioned 118 shoulders that were treated from 1980 to 1985, the second series comprised 167 shoulders that were treated from 1986 to 1999, and the third series comprised thirty-four shoulders that were treated from 2000 to 2004. Results with respect to capsular shift, coracoid position, and coracoid union were evaluated. Ninety-six percent of the patients were satisfied or very satisfied at the time of the latest follow-up. Recurrent instability was reported in 13% of the patients. However, when a horizontal capsular shift was added to the transfer, the instability rate decreased to 4%. The average loss of external rotation in the second and third series was found to be 10° for 115 shoulders in “very satisfied” patients, compared with 21° for thirty-nine shoulders in “satisfied” patients. Eighty-three percent of patients achieved osseous union on radiographs. The prevalence of glenohumeral arthritis was not mentioned. Hovelius et al. also compared the results of the Bristow procedure (ninety-seven patients) with those of the Bankart repair (eighty-eight patients) in a study of 185 patients who were followed for an average of seventeen years15. Patient satisfaction and outcome scores were significantly better in the Bristow procedure group than in the Bankart repair group (p = 0.01). In addition, patients in the Bankart repair group had greater loss of external rotation than those in the Bristow procedure group (19° compared with 11°; p = 0.12).
In a long-term study of fifty-two United States Naval Academy midshipmen who were managed with the Bristow procedure, 70% of the patients had a good or excellent result after an average duration of follow-up of twenty-six years50. Multiple scoring systems such as the Rowe score51, Single Assessment Numeric Evaluation52, and Western Ontario Shoulder Instability Index were used50,53. Fifteen percent of the patients had recurrent instability, with five requiring revision surgery.
An arthroscopic technique for the Latarjet46 and Bristow procedures also has been described3. The arthroscopic Bristow procedure involves performing a Bankart repair followed by transfer of the tip of the coracoid with the conjoined tendon attachment. The osseous fragment is secured to the neck of the glenoid with use of one bioabsorbable interference screw. Thirty-six patients were evaluated at a minimum of twelve months in this series3. Eight percent of the patients had recurrent instability. The loss of external rotation ranged from 9° to 15°. The arthroscopic Latarjet procedure involves the transfer of a bigger bone block and requires the use of special instrumentation. Both of these techniques have a very steep learning curve, requiring hands-on surgical experience before the surgeon is comfortable with the technique.
Bone graft alternatives to the coracoid process include iliac crest autograft and allograft. Warner et al. described the autogenous iliac crest bone graft technique and reported the results of the procedure in a case series of eleven patients9. Most of the patients had substantial bone loss (as measured with the Gerber and Nyffeler technique21), with the length of the bone defect being more than the radius of the maximum anteroposterior diameter. Also, many of the patients had multiple shoulder operations for the treatment of instability. Nevertheless, excellent results were obtained in all patients, with no recurrence of instability, at minimum of two years of follow-up. A repeat CT scan at four to six months showed complete union of the bone graft. The authors cautioned that the long-term effects are not known and hypothesized that the ensuing arthritis may be due to either the recurrent injuries or the new glenoid surface articulation.
Provencher et al. described the use of distal tibial allograft in a small series of three patients and reported that all three had an excellent result56. Follow-up CT scanning demonstrated complete healing in all patients. A recent cadaveric study confirmed that distal tibial allografts and grafts from the inferior surface of coracoid have excellent conformity but iliac crest grafts do not57.
A posterolateral humeral head defect, also known as the Hill-Sachs lesion, is commonly seen in patients with shoulder instability58. A number of cadaveric biomechanical studies have shown that substantial bone loss in the posterolateral aspect of the humeral head may contribute to shoulder instability59-61. Those studies showed that a 25% defect may not be substantial but a lesion more than 37.5% may require an allograft. An engaging Hill-Sachs lesion is one in which the humeral head defect aligns with the anterior edge of the glenoid in abduction and external rotation, leading to shoulder dislocation4,62. This type of lesion can be viewed during arthroscopy and may be an indication for addressing the Hill-Sachs lesion along with the glenoid lesion. This concept may be more important than the size of the defect. A detailed review of humeral head defects is beyond the scope of the present article.
Short-term complication rates after the Latarjet procedure have ranged from 5% to 25% (Table II). Most of the complications resolved within a few months. Allain et al. reported a complication rate of 7%11. Three infections developed after surgery. Two resolved after operative drainage, and one resolved after antibiotic therapy. A frozen shoulder developed in a patient who refused postoperative rehabilitation. In another case, a humeral fracture occurred two months postoperatively during manipulation of the shoulder with the patient under anesthesia.
Burkhart et al. reported a complication rate of 5%; the five complications involved two hematomas (one of which required drainage), two loose implants, and one fibrous union of the bone graft41. The loose screws and fibrous union were diagnosed radiographically; however, they were asymptomatic and revision was not performed.
Lafosse and Boyle reported a complication rate of 10%17. Perioperative complications included two hematomas, one intraoperative graft fracture, and one transient musculocutaneous nerve palsy that resolved. Late complications included screw loosening in three shoulders and nonunion of the graft in four. Four of the patients with late complications underwent late arthroscopic screw removal. The authors reported no cases of recurrent dislocation. Di Giacomo et al. investigated coracoid bone graft osteolysis as one of the causes of failure of the Latarjet procedure63. After analyzing twenty-six patients with CT scanning, they found that most of the osteolysis was on the proximal part of the coracoid and concluded that the bone block may not be the principal factor in how the procedure works.
Paladini et al. assessed postoperative subscapularis strength in 376 patients, including 264 patients who underwent L-shaped tenotomy and 112 who underwent subscapularis muscle split44. Isometric peak torque was significantly lower in patients who underwent the L-shaped tenotomy, and the authors recommended that all patients should be managed with the muscle-split approach for optimal subscapularis function postoperatively.
Neurologic complications include brachial plexus traction injuries and musculocutaneous nerve palsies11,31,64-69. These complications may be caused by traction, patient malpositioning, or direct trauma to the nerves. Maquieira et al. reported suprascapular nerve palsy due to screw impingement25. The palsy resolved after removal of the hardware. Burge et al. demonstrated that the shortest nerves (the axillary and musculocutaneous nerves) are at the highest risk for traction injury70.
Shah et al. analyzed complications after the Latarjet procedure in a study of forty-seven patients71. The overall complication rate was 25%. Complications were divided into three groups: infection, recurrent instability, and neurologic injury. Three shoulders (6%) developed infection that resolved after irrigation and debridement along with antibiotic therapy. Four shoulders (9%) had recurrent instability. Five shoulders (11%) sustained a neurologic injury involving the musculocutaneous nerve (two shoulders), the radial nerve (one), or the axillary nerve (two). The musculocutaneous and radial nerve injuries involved sensory neurapraxia and resolved in subsequent months. The patients with axillary nerve dysfunction continued to have residual symptoms at the time of the latest follow-up.
Arthritis after the Latarjet procedure is a concern. There is controversy in the literature with regard to the prevalence of arthritis after this procedure. As most studies have been short to intermediate-term studies, the incidence of this complication cannot be predicted. Two studies indicated that the incidence may be as high as 40%16,48, whereas other studies have suggested that the incidence is no higher than that in the contralateral shoulder18. It is well known on the basis of biomechanical and clinical studies that lateral placement of the graft may lead to arthritis11,39. Arthritis also may be a result of repeated dislocations either before or after surgery18. On the basis of biomechanical evidence, it is imperative to place the bone graft flush with the glenoid and to avoid projection beyond the rim.
Osseous defects of the glenoid are seen in patients with recurrent anterior shoulder instability. The size, location, and acuity of such defects are important factors to evaluate when determining possible treatment options. Substantial anterior glenoid deficiency can be successfully treated with coracoid bone transfer; however, potential complications should be considered. The recommendations for care are summarized in Table III.
Source of Funding: No external funds were received in support of this study.
Investigation performed at Children’s Hospital of Michigan, Detroit, Michigan
Disclosure: None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any aspect of this work. One or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.
- Copyright © 2014 by The Journal of Bone and Joint Surgery, Incorporated