Shoulder joint dislocation, also known as glenohumeral joint dislocation, is the most common joint dislocation in clinic, which accounts for about 40% of all joint dislocation. It occurs mostly in young adult with more men being affected than women. Recurrent shoulder dislocation is one type of dislocation of shoulder joint, and there is a higher occurrence of anterior dislocation. This dislocation is mainly due to the anterior capsule tear after initial dislocation, the avulsion of the glenoid or the rim of it, and the malunion caused by un-repaired of them, as well as the posterior humeral fracture of humeral head. Consequently, repeated dislocation can occur. At present, there is no domestic and international consensus in the choice of surgical method for recurrent shoulder dislocation with severe bone defect. Previously, we took lower glenoid bony defect diameter ≥25% or deep Hill-sachs defect engaging with the glenoid with 90°of shoulder joint abduction and external rotation as giant bone defect. Some literatures report that arthroscopic Bankart repair surgery has a recurrence rate of up to 67% for these two types of shoulder dislocation combined with giant bone defect. The current treatment mostly uses Latarjet surgery or autogeneic or allogeneic bone grafting. This surgical method can regain the stability of shoulder joint and reduce the rate of redislocation.

1.Normal information: The retrospective study was performed in 12 patients (10 males and 2 females) who underwent Latarjet surgery and Bankart repair for the treatment of recurrent shoulder anterior dislocation with severe bone defect from April 2015 to February 2017, there were 7 cases with the left side affected and 5 cases with the right side affected, and the age ranged from 19-69 years with an average of 31.89 years. The average duration of disease was 25.12 days. The preoperative apprehension test was positive, and the average dislocation time was 14.5 (9 to 30 times) . As the glenoid was measured before the operation, the degree of bone defect was more than 25%, and there was no combined rotator cuff injury. The causes of primary shoulder dislocation injury included 9 cases of sports injury, 2 cases of high fall injury and 1 case of traffic injury. CT scan and three-dimensional reconstruction of shoulder joint were performed before and after the operation. All cases were recorded with the results of active flexion, internal rotation, external rotation and 90° external rotational exhibition of shoulder before the operation and during the last follow-up. 2.Inclusion and exclusion criteria: nclusion criteria: (1) The diagnosis of recurrent anterior dislocation of shoulder was clear, and there were at least two times of dislocation history. (2) The shoulder was extremely abducted, and the external rotation was painful with the positive result of apprehension test. (3) Preoperative CT evaluation showed severe glenoid bone defect (>25%) . (4) The patient required surgical treatment. (5) The follow-up cases and imaging data remained intact. Exclusion criteria: (1) Shoulder joint multi-directional instability. (2) The coracoid was too small. (3) Severe front joint capsule injury. (4) The ipsilateral shoulder joint was combined with other injuries such as rotator cuff injury, nerve injury, etc. 3. Surgical methods: The patient with brachial plexus block + tracheal intubation under general anesthesia was in beach chair position, and the affected limb was fixed by adjustable spider arm. With routine disinfection and drape, controlled hypotension was executed during the surgery. For the patient without hypertension or cardiovascular and cerebrovascular diseases, the systolic blood pressure was controlled at around 100 mmHg. The posterior observational approach, anterior and anterior lateral working portals were constructed routinely, and the joint was cleaned by radiofrequency and shaver. The gelnoid defect, Hill-Sachs lesion and the quality of labrum and joint capsule were re-evaluated under arthroscope. The anterior and inferior labrum and joint capsule tissue were released, and the anterior bone surface of the glenoid was freshened. After the opening of rotator interval and the exposure of coracoid, the soft tissue around coracoid was cleaned. The coracoacromial ligament, pectoralis minor insertion, coracohumeral ligament and other structures were cut off, and the conjoined tendon was protected. The subscapularis tendon and muscle abdomen were exposed, and the axillary nerve was explored. Then, the subscapularis was opened at the lower third part. The special guider of Mitek Latarjet was used to drill 2 holes on coracoid , and the lower surface of coracoid was grinded with grinding drill to make it decorticated. The traction wire was passed through, and the coracoid was cut off at the root of coracoid, made it attached and fixed with guiding sleeve. The lower surface of coracoid was further trimmed. With the holding and assisting of guiding sleeve, the coracoid bone was fed into joint cavity through the rupture of subscapularis muscle. The bone position was adjusted so that the center of bone was about 4 o'clock. Simultaneously, the lateral edge of bone was aligned with glenoid articular surface. After drilling, two AO hollow screws with the diameter of 4.5 mm were screwed in to fix the bone. If the bone block protruded to the outside of glenoid articular surface, the grinding was entered from posterior portal to grind flush the bone. Two double-loaded anchors were inserted into the anterior edge of glenoid, and the labrum and the anterior bundle of inferior glenohumeral ligament were repaired with anchor sutures. Also, the anterior and posterior joint capsules were tightened and transform the bone into extra-articular structure, which helped to restore the tension of joint capsule and ligament. The shoulder joint was moved at multiple angles to dynamically assess joint stability. Finally, the wound was sutured with absorbable suture. 4. Postoperative rehabilitation: After the operation, the affected shoulder was bandaged and fixed, and the anti-infection, swelling, and cleansing of the operation area were performed. The healing and pain of the incision were closely observed. The routine blood, biochemistry, erythrocyte sedimentation rate and C-reactive protein were regularly reviewed. The active activity training of the elbow, wrist and hand of the affected limb was carried out under the guidance of a professional physiotherapist for two days, but the resistive flexion of elbow training was avoided. Within 6 weeks, the shoulder abduction pillow was fixed at abduction 30° with a forward flexion of 15°, you can do mild "pendulum" and "shrub" exercises while training the muscles of the affected limbs. After 6 weeks, the abduction pillow was removed, and full-range motion and passive activity exercises were started, and active activities were gradually started. The patients underwent regular outpatient review, filming and follow-up at 3 weeks, 1 month, 3 months, 6 months, 1 year and 2 years.5.Observation indicators:Clinical indicators: (1) CMS (Constant-Murley-Score) : consists of four parts: pain, daily activities, joint mobility and muscle strength. The total score is 100 points, >90 is excellent, 80-90 is divided into good, 70-79 is divided into general, <70 is divided into poor, and shoulder function is evaluated before and at the last follow-up. (2) Pain visual analogue scale (VAS) : no pain, 0 points; slight pain, can endure, rated 3 points or less; pain and affect sleep, can still bear, comment 4-6 points; Strong pain, unbearable, rating 7-10 points. (3) American shoulder and elbow surgeons (ASES) : 100-90 are divided into excellent, 89-75 is divided into good, 74-51 is divided into general, and ≤ 50 is divided into poor. (4) Active joint activity: anterior flexion, internal rotation, body lateral external rotation, abduction 90° external rotation. Imaging evaluation of bone position and absorption: At the postoperative follow-up, a three-dimensional reconstruction of the shoulder joint CT was performed to evaluate the location and absorption of the graft. In this study, the coracoid graft bone was fixed with AO double hollow screws, and the screw tail was originally left outside the bone. Based on Professor Zhu Yiming's bone absorption grading system and bone position , combined with the actual situation of this study, we define the degree of bone absorption as follows: 0, the inner-outer diameter and the anterior-posterior diameter of the bone There is no observable change; grade I, the inner-outer diameter or anterior-posterior diameter of the graft is shorter than that immediately after surgery. The former shows that the graft is more inward than the front, and the latter shows an increase in the exposed end of the screw. (non-buried screw) ; Grade II, part of the screw body exposed, most of the bone absorption; Grade III, the nail head and the nail body are completely exposed, the coracoid graft bone is completely absorbed; and three-dimensional CT is performed for postoperative follow-up to evaluate the position and fixation of the bulge were examined by CT cross-sectional scan to examine the relationship between the outer edge of the displaced coracoid and the glenoid surface. If the two are completely flush, the position of the coracoid bone is flush with the glenoid. If the outer edge of the coracoid bone is more than 1 mm away from the glenoid surface, the position is abnormal. 6.Statistical methods: Collect data to build a database, and import the data into SPSS 22.0 statistical software for data analysis. Measurement data are expressed as mean ± standard deviation. The preoperative and final follow-up of the shoulder joint active activity ROM and ASES score, Constant score, VAS score and ROWE score were consistent with normal distribution, using paired t-test of quantitative data. The difference was statistically significant at P<0.05.

1. postoperative functional assessment of patients: All patients were followed up for 13 to 24 months with an average of 16.6 months. 12 patients did not suffer from re-dislocation after the operation. The flexion and lifting before the operation and during the last follow-up were (142.86±4.88) ° and (169.57±5.56) ° respectively, and the difference was statistically significant (t=12.50, P<0.05) . There was no significant difference of internal rotation between the preoperative and the last follow-up (t=2.29, P>0.05) . The difference of mean side-lateral external rotation between the one before the operation (59.00±4.20) °and the one during the last follow-up (52.57±2.99) °was statistically significant (t=6.97, P<0.05) ; the difference of mean external rotation between the one before the operation (87.29±4.72) ° and the one during the last follow up (79.00±7.53) ° was statistically significant (t=6.35, P<0.05) . The ASES scores were (69.20±3.42) preoperatively and 97.90 during the last follow-up, and the difference was statistically significant (t=779.24, P<0.05) ; the difference in Constant-Murley score was statistically significant between the preoperative (90.86±1.57) and the last follow-up (96.57±2.99) (t= 8.40, P<0.05) ; the difference in Rowe score between the preoperative (39.29±7.32) and last follow-up (92.86±2.67) was statistically significant. (t= 17.68, P<0.05) ; the VAS score before the operation and during the last follow-up were (7.21±1.54) and (4.31±2.34) , and the difference was statistically significant (t= 17.68, P<0.05) . 2. Imaging evaluation of bone position and absorption: Preoperative routine CT examination measured that all cases had >25% of glenoid bone defect. The anteroposterior diameter of coracoid bone was (8.83±2.81) mm, and the preoperative trajectory G size was (14.41±1.50) mm. The width H between Hill-Sachs lesion medial edge to rotator cuff foot print medial margin was (21.61±3.41) mm. All the 12 patients were evaluated as "Off-Track" . CT scan was performed on the second day after the operation, and the measured trajectory width was (21.78±1.86) mm. All patients were restored with humeral head-shoulder "On-Track" trajectory. 9 cases had the coracoid graft completely aligned with glenoid, and the other 3 cases had the coracoid graft bulging outward from glenoid with an average external displacement of (1.63±0.42) mm. In sagittal position, all patients had the transplanted bone located at a good position of lower one-third of glenoid. During the final follow-up, 9 patients achieved bone healing; 2 patients had grade II bone absorption; 7 patients had grade I absorption at the upper part of bone; 3 patients had no observable bone resorption. Among them, CT showed that the bone mass absorption for number 4 male patient (19 years old) was the most serious during follow-up. Immediately after the operation, the transplanted bone was moved outward from glenoid by 1 mm. After 8 months, CT found the absorption of the outer surface of bone, which made it moved 3 mm inward relative to glenoid articular surface. However, the screw body was still embedded in bone, and the absorption level was I.

For the patients with anterior shoulder dislocation and severe glenoid bone loss, arthroscopic Latarjet procedure combined with Bankart repair technique can restore the stability of shoulder joint. The trauma is small, and the early and mid-term outcomes are good. Due to long learning curve of this type of surgery and massive structural damage, however, the long-term side effect to shoulder function requires further follow-up observation.