A little over a century has passed since the first introduction of treatment for acromioclavicular (AC) joint instability by Cadenet in 1917 [1] . In contrast to the pale options in conservative treatment for AC joint instability, surgical or operative treatments seem to be derived in various forms. More than 150 techniques have been described for the surgical treatments of AC joint injury [2] . Wide range of complication rates have been reported from 5% to 30% of cases [3] . ACJ injuries are fairly common with an incidence of 12% of all shoulder girdle injuries [4] . Higher prevalence was seen in male in their 20 to 30 decades participating in high-contact sports [5] . High-impact injury to the shoulder results in disruption of the stabilizing structures around the AC joint. Falling from height, road traffic accidents and high-energy contact sports are among the common causes of AC joint injuries. The common occurring mechanism of injury is a direct impact to the acromion with an adducted shoulder. The force will drive the acromion inferiorly from the clavicle, which then starts the cascade of disruption of the AC ligaments, followed by the coracoclavicular (CC) ligaments. Indirect mechanism of injury describes an abducted arm causing the humeral head to impact onto acromion, forcing it to displace superiorly from the clavicle.

Tossy et al. developed a classification system for AC joint injuries in 1963 [21] , which was then expanded by Rockwood to become the 6-part classification which is widely utilized to date [22] .

Rockwood Types I and II injury are stable or low-grade injury which non-operative treatments are the first to be considered [10] . Conservative management has shown favorable outcome, higher return to activities and less complications [23] [24] ( Table 1 ).

The other side of the injury spectrum, which consists of high-grade Types IV, V and VI, require surgical treatments to restore the integrity and functionality of the AC joint [14] [24] . Type IV injuries involve the complete tear of the AC, CC ligaments and trapezial fascia injury with the posterior displacement of the distal clavicle. Type V AC joint separation is characterized by a bigger gap of CC space and the irreducible penetration of distal clavicle through a torn deltotrapezial fascia. In Type VI injuries, the mechanism of hyperabduction and external rotation causes the distal clavicle to be inferiorly dislocated into subacromial or subcoracoid with the detachment of trapezius and deltoid at the insertion sites.

Much debatable is the Type III injury where non-operative and operative treatments have their respective proponents. The optimal managements for Type III injury still remain controversial [25] . A consensus had been established that Type III injuries to be further divided into stable (IIIA) and unstable (IIIB) based on functionality assessment [24] . Initial conservative treatments are allowed for Type III injuries with reassessments after 3 - 6 weeks. The presence of persisting pain, shoulder dyskinesia and overriding distal clavicle on acromion on cross body adduction radiographic view during reassessment points towards an unstable

Type IIIB injuries which surgical intervention is recommended. Inability to pursue with daily activities, work or sport due to irretractable pain, is the valid reason to warrant for a subsequent surgical treatment.

Time of injury is among the determining factors for interventions and outcomes of treatments. AC joint ligaments lose the potential to heal after 3 weeks post injury [26] . A consensus among the ESA-ESSKA members recommended that 3 weeks from injury is the time point to distinguish between acute and chronic injury.

The surgical interventions for unstable AC joint injuries had seen much diversification, from the earlier open techniques to the current arthroscopic assisted method, rigid versus non-rigid fixation, anatomical versus non-anatomical reconstruction, and graft versus non-graft reconstruction. Most of the techniques that evolved to date were improvements or combinations of historically described techniques. In 1941, Bosworth described an anatomic fixation of the clavicle to the coracoid using a non-cannulated lag-screw fixation technique [27] . Rigid fixation using coracoclavicular screw fixation or hook plate fixation require hardware removal in 8 to 12 weeks post-operatively to prevent screw migration and hardware failure. The motions between coracoid and clavicle can lead to implant fatigue and failure over time [28] . Tsou reported a high percentage of failure (32%) in 17 of 53 patients who underwent percutaneous cannulated screw fixation [29] . Although hook plate fixation of the AC joint provides a stable fixation and higher American shoulder and elbow surgeons (ASES) score and lower VAS scores, it causes subacromial erosion and requires removal later [30] . They saw all patients’ post hook plate removal showed an increase in the CC distance. Yoon et al. reported as high as 9 of 24 patients treated with hook plate had subacromial erosion [31] . Hardware removal and implant failure were among the reasons for the decrease popularity of rigid fixation.

Weaver-Dunn procedure received much attention during its introduction in 1972 as one of the important treatments for AC joint separation using coracoacromial ligament transfer [32] . However, this non-anatomical reconstruction technique has lost its enthrallment as anatomical restoration of the AC joint separation became the mainstream treatments.

Anatomical reconstruction of the AC joint had shown better postoperative outcomes compared to non-anatomical reconstruction [33] [34] . A cadaveric study by Mazzocca comparing between arthroscopic reconstruction, anatomical CC ligament reconstruction and modified Weaver-Dunn procedure, concluded that the anatomical CC reconstruction has less anterior and posterior translation and more closely approximates the intact state of the AC and CC ligaments, restoring their functions [35] . A series of anatomical reconstructions showed a good preliminary result with improvements in American shoulder and elbow surgeons (ASES) score and Constant-Murley score [36] .

The objective of this series was to identify the advantages of arthroscopic assisted anatomical reconstruction of chronic AC joint separation in the aspects of patient reported outcomes and the possible complications that would arise from the approach practiced.

We retrospectively reported 6 (5 male, 1 female) chronic high grade AC joint injury cases who underwent arthroscopically assisted AC joint reconstruction surgery between year 2018 to 2023. Patients who had completed follow-ups and were discharged from the clinic were included. The small number of cases was due to inadequate data as some patients did not attend follow-ups. The mean age was 25.7 years-old (range, 21 - 30 years-old) at the time of injury. The mechanism of injury was due to various degrees of severity of motor vehicle accidents. The mean duration of follow-up was 39.6 months (range, 24 - 56 months). The demographic data of the patients was shown in ( Table 2 ). The average time from injury to surgery was 194.3 days (range, 21 - 461 days).

All the 6 cases had arthroscopically assisted anatomical CC ligament reconstruction done with a knotless suture-button system (Arthrex, Naples, FL) and autologous hamstring (semitendinosus) graft. The AC joint was also reconstructed with FiberTape, and the AC joint capsule and deltotrapezial fascia were repaired. The surgeries were performed with the patients on beach-chair position under general anesthesia. A C-arm coming in from the contralateral side was used to confirm the AC joint anatomical reduction and fluoroscopy assessment. The operated arm was left free to rest on patient’s body without using any arm holder. The injured upper limb, shoulder and ipsilateral lower limb were cleaned and draped in a sterile manner.

A standard diagnostic shoulder arthroscopy was performed with a 30˚ arthroscope via a standard posterior viewing portal. An anterior portal just lateral to the coracoid was created, through the rotator interval, for subcoracoid debridement to expose the inferior surface of base of the coracoid. An anterolateral portal was created for the viewing of the coracoid undersurface. Simultaneously, the semitendinosus tendon was harvested from the ipsilateral lower limb through an anteromedial approach on the tibia. A horizontal incision along the axis of the clavicle was made 5 cm medial from the AC joint and extended 2 cm lateral to the AC joint. The superior, anterior and posterior borders of the clavicle which correspond to the attachment of the CC ligaments on the clavicle was adequately exposed for later drilling of tunnels. The AC joint was adequately debrided and assessed. Any invaginated tissue was released and debrided to facilitate the reduction of the joint.

A 1.6 mm Kirschner wire was temporarily inserted percutaneously across the AC joint to maintain the reduction with the C-arm guidance. An AC aiming guide (Arthrex, Naples, FL) was inserted through anterior portal with the aiming arm at the undersurface of the coracoid. The drilling sleeve was put at the superior of the clavicle, 3.5 cm medial to the AC joint, centered anterior to posterior. Drilling was performed with a 3.0 mm cannulated drill (Arthrex) at the superocentral part of the clavicle to avoid being too anterior or posterior to the rim that would increase the risk of clavicle fracture. Direct arthroscopic visualization of the undersurface of the coracoid was performed during drilling of the clavicle till the cannulated drill exited the designated center part of the coracoid undersurface ( Figure 1 ). By adjusting the aiming device under the coracoid and the sleeve above the clavicle, two tunnels for the semitendinosus graft passage were drilled 1.0 cm medial and 1.0 cm lateral to the first tunnel with 5.0 mm cannulated drill bit. The drilling was ensured at central position anterior to posterior on the superior clavicle. The drill bit’s exit on medial and lateral of the coracoid were directly visualized with arthroscope. Double loaded knotless CC fixation device (Knotless AC TightRope device, Arthrex) with cortical fixation button (Dog Bone, Arthrex) was then shuttled from coracoid to clavicle. The cortical fixation button was manipulated to sit snugly on the undersurface of the coracoid under direct arthroscopic visualization. Another cortical fixation button was attached to the tight rope on the superior part of the clavicle. The FiberTape was then pulled and

clipped temporarily to secure the AC joint reduction which was checked with C-arm fluoroscopy. Semitendinosus tendon graft was then shuttled through the 2 side tunnels to loop around coracoid undersurface. Further compression and reduction were done before knot was tied on the tight rope device. The graft was tied with a simple overhand throw on the superior part of the clavicle and further secured with a nonabsorbable suture. The temporary Kirschner wire was removed. The AC joint reduction was checked with the C-arm to ensure no reduction loss. Attention was then shifted to the AC joint where horizontal tunnels were drilled with 2.4 mm cannulated drill bit at the distal clavicle and the acromion, 1.0 cm and 1.5 cm from the AC joint respectively, in an anterior-posterior fashion. The remaining AC capsule was repaired with absorbable suture. FiberTape (Arthrex) was shuttled through the tunnels to make a figure of eight configuration to further compress and secure the AC joint reduction. C-arm fluoroscopy was performed to check final reduction of the AC joint. The deltotrapezial fascia was repaired securely with an absorbable suture before soft tissue closure in layers.

All the patients were put on abduction sling for 6 weeks. Cryocuff therapy was repeated every 2 hours on the operated shoulder. Post-operative isometric strength exercise was commenced as tolerated. Joint mobilization of the elbow, wrist and hand were encouraged. Pendulum exercise and gradual passive forward flexion of the shoulder from 0˚ to 90˚ and passive external rotation from 0˚ to 45˚ in adduction were started 2 weeks post-operatively. The abduction sling was removed after 6 weeks of surgery and progressive physiotherapy and rehabilitation with active range of motion and strength was started. Patients were being followed up post-operatively at 2 weeks, 6 weeks, 3 months and subsequently every 6 months.

The American shoulder and elbow surgeons (ASES) score for the 6 patients increased with the median of 36 (range 15 to 82) at pre-operative to 77.5 (range 35 to 88) at 6 months post-operative and 92.5 (range 82 to 100) at final follow-up ( Figure 2 ). All the patients returned to work 3 months post-operatively. All of them were able to carry out daily activities without many difficulties as evidenced by the good ASES score at final follow-up. The Visual Analogue Scale for pain showed improvements from the median of 5.5 (range 2 to 8) at pre-operative to 3 (range 1 to 5) at 6 months post-operative and 1 (range 0 to 3) at final follow-up ( Figure 3 ). The pain at their last follow-up was well tolerated and did not disturb their daily living and jobs.

Post-operatively, 3 patients experienced tolerable discomfort over operated side. One of the patients with shoulder discomfort showed clavicular tunnels widening (cTW). One of the 6 patients showed a loss of reduction (LOR) and cTW radiographically but without pain or discomfort ( Figures 4 - 6 ). A 6 mm superior displacement of the clavicle was deemed LOR [59] . The LOR was recognized at 4 months after surgery, with a CC distance of 24 mm. The LOR was seen as a result of failed AC joint reconstruction which was evidenced by the FiberTape cut out from the acromion. Two patients had no complication or complaint post operatively. No infection occurred on the patients post-operatively.

Three of the 6 patients presented with limited abduction and forward flexion which were below 90˚. The other 3 patients able to achieve full or near to full range in abduction and forward flexion. These 2 motions are essential in overhead activities which are greatly impaired in AC joint injuries. There was 1 patient who showed a reduced range of motion (ROM) in abduction, forward flexion, external rotation and adducted internal rotation post-operatively at 3 months. The poor compliance with the physiotherapy programme was identified as the cause of the poor range of motion. The other 5 patients displayed progressive improvements in the ROM. Eventually, all the patients restored full functional range of motion at final follow-up without difficulty in carrying out their daily activities and jobs ( Table 3 ). All the 6 patients returned to work at 3 months post-operatively.

ROM: range of motion; ADIR: adducted internal rotation; T: thoracic; S: sacral; L: lumbar.

Injuries to the AC joint impaired the function and productivity of the patients. High grade AC joint separation involved the injury to the AC joint capsule, the CC ligaments and to some extent the rupture of deltotrapezial fascia. These circumstances required surgical intervention to restore the integrity of AC joint that formed the superior shoulder suspensory complex (SSSC) which is the essential link between the upper extremity and the trunk [60] . Various surgical techniques had been preferred with diverse outcomes. Biological augmentation with autologous or allogenous tendon grafts had been recommended for chronic injury reconstruction [61] . Complications of AC joint and CC ligaments reconstruction had been identified with CC ligament calcification, LOR and shoulder pain ranked as the commonest [56] . Coracoid and clavicular fractures occurred in reconstruction that involved tunnel drilling [57] . In our case series, discomfort at the operated site was the main issue the patients had. It was due to the soft tissue irritation by the prominent suture knots over the clavicle and around the AC joint. However, the discomfort was tolerable and none of the patients required second surgery.

The LOR noted on the patient was due to the FiberTape cut out from the acromion. Telltale sign was noted on radiograph during follow-up 2 months post-operatively, with a CC distance of 20mm. However, the patient did not have any discomfort or pain. Protected rehabilitation was carried out. Eventually, full range of motion of the shoulder was achieved. At the final follow-up, the CC distance was 26mm with cTW. No revision surgery was performed as the patient did not show impairment from the reduction loss. The shallow tunnel drilled on the acromion was identified as the cause for the thin tunnel wall to break during AC joint motion. Subsequently, it led to the failure of the AC joint capsule repair and CC ligament reconstruction.

Another patient with cTW and shoulder discomfort did not show LOR radiographically. It was known that cTW was associated with recurrent vertical and horizontal instability [62] . However, the AC joint remained intact and stable clinically even cTW was evident radiographically. Osteophytes were observed on the undersurface of the clavicle adjacent to the tunnels.

Tunnels made on the coracoid and clavicle were the known risk of fractures to occur during or after CC ligament reconstruction [57] [63] . The risk of fracture was correlated to the number of tunnel and the diameter created. The strength reductions correlated highly with the ratio of the tunnel width relative to overall clavicle width [64] . A 6 mm diameter tunnel posed a higher risk for fracture than a tunnel with 2.4 mm diameter. Multiple fixation points and drill holes may increase the risk of coracoid and clavicle fractures (4% - 11%) [3] . Milewski et al. found a pooled complications of 80% in their coracoid tunnel group with 2 coracoid fractures which had tunnels of 5 mm and 6 mm diameter. In their coracoid loop group, complications of 35% were reported but without coracoid fracture [57] . LOR was relatively high in the coracoid tunnel group (50%) compared to the coracoid loop group (12%).

Chernchujit et al. found that anatomical AC joint reconstruction had a better outcome than AC joint fixation in respect of specific AC score (SACS) and posterior translation [65] . Complications rate of 17% was found in their study. In our series, the CC ligaments were anatomically reconstructed with semitendinosus graft in addition to FiberTape AC joint stabilization for that the torn CC ligaments in chronic high-grade injuries had lost the ability to heal over time. Reconstructing the CC ligaments with biologic graft option enables the AC joint reduction to be maintained by the restoration of the CC ligaments function as close as possible. This non-rigid reconstruction allowed some mobility of the clavicle at the AC joint which mimicked its native characteristics.

A 3.0 mm diameter tunnel was made as the central tunnel for the FiberTape and Dog Bone device fixation. The central tunnel was 3.5 cm medial to the AC joint. It was located halfway in between the native conoid and trapezoid ligaments insertion at the clavicular tubercles. The centers of conoid and trapezoid ligaments insertion on clavicles were at around 4.5 cm and 2.5 cm respectively, from the lateral edge of the clavicle [66] . Thus, the 5.0 mm diameter tunnels for the semitendinosus autograft passage were created according to the center of the footprints. The distance between tunnels and the proximity to the clavicle anterior posterior rim were adequately spaced to ensure no weak links and reduce the risk of fracture. Geaney et al. showed that tunnel placement in the clavicle corresponding to the attachment of the CC ligaments has the highest bone marrow density and correlates to higher loads to failure experimentally [67] . Single tunnel on the coracoid for the FiberTape and graft loop around coracoid technique was performed to prevent the coracoid fracture risk. Figure of 8 FiberTape reconstruction was performed to reinforce the AC joint in addition to the capsule repair. It was shown that combined stabilization of the AC joint capsule and CC ligaments could restore the stability against rotational and translational loads [68] .

The clavicular vertical tunnel graft fixation technique was preferred over the graft loop around clavicle or the horizontal tunnel graft fixation methods as vertical tunnel provides better graft fixation and smaller tunnel diameter to bone width ratio thus reduced the clavicle fracture risk. One study had found there were comparable average clavicle length between Chinese population and the Western population [69] . However, extra care must be taken when tunnels are made on the clavicle, especially in shorter clavicle where the midshaft has the lowest thickness compared to the sternal and the acromial part [17] [69] .

Studies had shown that high grade AC joint injuries are associated with higher incidence of intraarticular lesions. Rotator cuff injuries, SLAP lesion and chondral lesion were the frequently encountered pathologies in patients with AC joint dislocations [37] [38] [40] . These intraarticular pathologies are consensually agreed upon that surgical intervention is required to prevent subsequent morbidity such as persistent pain and poor clinical functions. It is made possible to be detected through arthroscopic examination of the shoulder without excessive soft tissue dissection. Arthroscopic assisted AC joint reconstruction has the advantages of minimally invasive procedure, which allows the direct visualization of the coracoid and the glenohumeral joint [70] . In our series, we found no intraarticular lesions in all the 6 patients. The direct arthroscopic visualization of the coracoid undersurface enabled the accurate position of coracoid drill hole hence reduced the risk of coracoid process fracture or implant migration that could lead to fixation and reduction failures.

This paper reported the series of chronic AC joint separation cases that were treated at our center. The report of this series was to share the experience of arthroscopic assisted anatomic reconstruction for chronic AC joint disruption at our center. First and foremost, the number of high-grade AC joint injuries presented to our center was not vast. Majority of AC joint injuries were treated acutely within 3 weeks, hence the small number in this series. Furthermore, some of the post-operative patients were lost in follow up and inadequate data was available to be included in this series. There was no objective measurement of the strength of the operated limbs to be compared and reported.

The surgical treatments of AC joint separations will still be debatable in the recent future as there is still no best method to treat it. There were mixed clinical outcomes in regard to the various types of surgical methods. The complications that arise from AC joint surgeries are multifactorial and generally technical based. Larger-size studies on surgical techniques and outcomes are of high value for the betterment of AC joint injury treatments. From this small series of patients, we learned that a better surgical approach to treating AC joint injuries could be tailored with the combination of existing proven techniques. Although complications in anatomical reconstruction of chronic AC joint separation are significant, it is imperative to uphold established technical principles with stepwise approach to ensure good outcomes and minimize complications.

Heterogenous studies comparing different approaches to chronic AC joint disruption should be encouraged to refine the reliable treatments with reproducible good clinical outcomes. Homogenous longitudinal studies should be conducted to recognize a better surgical intervention. Arthroscopic-based AC joint reconstruction has shown its advantages of being minimally invasive, enabling simultaneous solution to intraarticular pathologies and direct visualization of the accurate placement of implant in CC ligaments reconstruction. Advancement of the future path in AC joint and CC ligaments reconstruction should be centered on arthroscopic or arthroscopically assisted approach.