Accounting for 12% of all shoulder injuries, acromioclavicular (AC) joint dislocation is a commonly occurred injury of the upper extremity. It is commonly seen in young active patients, and its incidence in athletes is even up to 91%. According to Rockwood classification, type I and type II are incomplete injuries that usually require non-operative treatment. TypeⅣ-Ⅵinjuries are complete AC joint dislocations, and most scholars advocate early surgical treatment. However, the treatment of type Ⅲ injury still remains controversial. Numerous surgical procedures have been reported in previous literatures, and the traditional surgical method includes Kirschner wire fixation, screw fixation, clavicular hook plate fixation, Weaver-Dunn technology, etc. Currently, button plate technology is one of the methods commonly used in the clinical treatment of acute AC joint dislocation, among which Triple Endobutton technique is the anatomical reconstruction of coracoclavicular (CC) ligament using 3 Endobutton plates, a closed loop ring and Ethibond non-absorbable sutures. The evidence of biomechanical and clinical researches shows that the treatment of acute AC joint dislocation with Triple Endobutton technique has significant effect. However, it is also reported that this technique has caused related complications such as loss of reduction, AC joint redislocation, coracoid process fracture, etc. For these reasons, many scholars who adopted this technique have analyzed the risk factors that affect the reconstruction of CC ligament and believed that the factors such as age, indication, surgical technique error, the location and size of clavicular and coracoid bone tunnels, the placement of button plate, etc. have an influence on the ultimate therapeutic efficacy. Among these, the location of coracoid bone tunnel is the most essential factor.Correct positioning of button plate at the base of coracoid process is extremely important for the avoidances of severe complication and operation failure. Due to the deep position of coracoid process base, the exact positions of coracoid bone tunnel and button plate are required to be confirmed under the guidance of fluoroscopy. Therefore, the purpose of this study is to investigate the relationship between the position of coracoid button plate and the therapeutic effect in treating AC joint dislocation with Triple Endobutton technology.

(1) General information. From February 2010 to December 2015, 67 patients (52 males and 15 females) with Rockwood type Ⅲ-V AC joint dislocations were treated with Triple Endobutton technique and reviewed retrospectively. All patients signed the informed consent and were approved by the ethics committee of hospital. The age ranged from 18 to 60 years with an average of 37 years. There were 40 cases of Rockwood type Ⅲ injury, 19 cases of Rockwood type Ⅳ injury and 8 cases of Rockwood type Ⅴ injury. The time from injury to operation was 2-7 days with an average of 4.5 days. 35 cases had the dominant hand affected, and 32 cases had the opposite hand affected. (2) Inclusive and exclusive criteria. Inclusive criteria: ① Simple acute complete AC joint dislocation, including Rockwood type Ⅲ, IV and V; ② No severe osteoporosis; ③ Complete follow-up. Exclusive criteria: ① Combination of coracoid process or clavicular fracture; ②Combination of neurovascular injury; ③ Active infection on the local or other parts;④Combination of other shoulder joint dysfunction;⑤ Previous shoulder joint history;⑥ Incomplete follow-up or information. (3) Surgical procedure.After general anesthesia or brachial plexus block, the patient was in beach-chair position with cushion under the affected shoulder. A transverse incision of 5 cm was made at the distal end of clavicle. The skin, subcutaneous tissue and deep fascia were cut open layer by layer to explore AC joint, and the embedded soft tissue and blood clot were cleaned. The AC joint was then reduced manually by assistant and fixed temporarily with 2.0 mm Kirschner wire. An incision of 3 cm was made with the center located on the tip of coracoid process, and the skin and subcutaneous tissue were cut open layer by layer. The base of coracoid process was bluntly split via the gap between pectoralis minor muscle and coracobrachialis muscle, which was inserted with the tip of locator. The other end was placed posterior to the middle of clavicle (30 mm from the distal end) . In this way, the first bone tunnel was formed. After guide pin positioning, the center of coracoid base was drilled through with 4.5 mm hollow bit. The second clavicular bone tunnel was placed anterior to the middle of clavicle (20 mm from the distal end) . The anatomical positions of conoid ligament and trapezoid ligament were replaced by the two bone tunnels respectively. The loop and Endobutton plate of proper size were selected. The first and fourth holes of plate were passed with 2 Ethibond sutures, and the loop was passed with another suture for traction. Then, the Endobutton plate was pulled from the clavicular bone tunnel to the base of coracoid process. The two sutures were extracted from the medial and lateral clavicular bone tunnels respectively, and the sutures and loop were tightened to locate the plate on the lower surface of coracoid base. Another plate without loop was selected, and its two lateral holes were inserted with the suture from medial clavicular bone tunnel. Later, the plate was made through the supraclavicular loop and was placed close to the clavicular surface, which was perpendicular to the loop as well. Afterward, the loop was locked with tightening of suture. One more plate without loop was chose, and its two lateral holes were inserted with the suture from lateral clavicular bone tunnel. The suture was then tied and fixed. The incision was sutured layer by layer and bandaged once the internal fixators were in good position under C-arm fluoroscopy. (4) Postoperative management.Routine anti-infection, analgesia, etc. were given for symptomatic support treatment after operation. Forearm sling was used for 4 weeks. Active movements of hand, wrist and elbow were conducted based on the pain tolerance 24 hours after operation, and the pendular exercise of shoulder joint was allowed as well. The forearm sling was removed for active anteflexion, upward lifting and abduction of shoulder joint 4 weeks after operation. Six weeks later, the shoulder was restored to allow normal activities, and strength training was gradually carried out. (5) Radiographic analysis and therapeutic effectiveness evaluation. The follow ups were conducted in the 2nd, 6th, 12th, 24th, 48th week and the 24th months with radiographs of AC joint series (including anteroposterior view) , and the position of coracoid Endobutton plate (lateral, central or medial) was determined by the anteroposterior view of shoulder joint radiograph. The base of the coracoid process in the anterior-posterior view is elliptical, and a line was drawn parallel to the axis of the clavicle. The positon of central button was the contact point of the parallel line and the oval. A perpendicular line was drawn through the contact point, which was vertical to the clavicle. If the Endobutton plate was located medial to this line, it would be divided into the medial group, and if the plate was located lateral to this line, it would be divided into the lateral group. According to the above criteria, the patients were divided into group A, B and C (the corresponding lateral group, central group and medial group) . The Constant score, visual analogue scale (VAS) and simple shoulder test (SST) of the last follow-up were used for shoulder joint functional and pain evaluation. There was no statistical difference in the general data of age, gender and Rockwood type among the three groups (P>0.05) , which was comparable. (6) Statistical analysis. SPSS 11.0 software was used for statistical analysis. The gender, Rockwood type and postoperative complications were compared using Pearson χ² test and Fisher's exact test of nonparametric categorical variables. The age difference, the interval of injury and operation time, the Constant score, the VAS score, and the SST score were all single factor analysis of variance, and conformed to the law of normal distribution. A P value < 0.05 was regarded as statistically significant.

A total of 67 patients were included in the study with 19 cases in group A, 27 cases in group B and 21 cases in group C. All patients were followed up after operation, and the follow-up time was 21-36 months. Among them, the mean follow-up times were (30.1±5.2) months for group A, (27.3±4.9) months for group B and (28.2±5.4) months for group C. There was no significant difference in general factors such as age, gender, dislocation type, injured side and the time from injury to operation. The difference between the postoperative complication rate of three groups was statistically significant (12/19 vs. 6/27 vs. 10/21, P=0.017) . In group A, 12 patients had complications of different level. During rehabilitation exercise, two cases had shoulder joint pain and activity limitation in the 6th and 8th weeks respectively. The X-ray examinations suggested coracoid process fracture, and the patients subsequently underwent surgical treatment (clavicular hook plate and hollow nail fixation) . Two cases had redislocation, and eight cases had loss of reduction. The complication rate of group B was the lowest, and there were 2 cases of loss of reduction, 1 case of redislocation and 3 cases of heterotopic ossification. 10 cases had complications in group C, including 2 cases of redislocation, 5 cases of loss of reduction and 3 cases of heterotopic ossification. The difference of the incidence rate of loss of reduction between three groups was statistical significant (P=0.020) , but there was no statistical difference in the comparisons of coracoid fracture, redislocation and heterotopic ossification between three groups. No patient had intraoperative neurovascular injury, postoperative infection, osteolysis and osteoarthritis.The mean Constant scores of group A, B and C were (91.2±4.2) points (85-100 points) , (94.3±3.6) points (88-100 points) and (93.9±4.2) points (85-100 points) respectively. The mean Constant scores of groups B and C were remarkably higher than that of group A, and the difference was statistical significant (P=0.036) . The mean VAS scores of groups A, B and C were (0.4±0.5) point (0-1 point) , (0.3±0.6) point (0-2 points) and (0.4±0.5) points (0-1 point) respectively. The mean SST scores of groups A, B and C were (11.4±0.7) points (10-12 points) , (11.7±0.6) points (10-12 points) and (11.5±0.6) points (10-12 points) respectively. The differences of the mean VAS score and SST score between three groups were not statistically significant (P>0.05) .

In the treatment of acute AC joint dislocation with Triple Endobutton technique, there was close relationship between the position of coracoid bone tunnel and the therapeutic efficacy. The central bone tunnel had significant effect and few complications. If button malposition, especially the lateral migration, was found under fluoroscopy, it should be corrected immediately.