Acromioclavicular (AC) joint dislocation is a common shoulder injury and is reported to occur most frequently among the young and athletic populations [1]. Athletes involved in contact sports, skiing and cycling are at increased risk of this injury [2]. The incidence of acute AC joint disruptions is high, accounting for 9% of all shoulder girdle injuries. Up to 40% of acute AC joint disruptions occur in elite athletes participating in competitive contact sports [1,3,4]. The mechanism of injury is typically a direct blow to the adducted shoulder or an axially directed force to the ipsilateral extremity. The stability of the AC joint is mainly provided by the coracoclavicular (CC) and AC ligaments [3]. In 1984, Rockwood described his six-part classification system for AC joint separations, and this remains the most common classification system in use today [4].

The management of AC joint dislocations is highly variable both in terms of surgical indication and proposed surgical technique. Nonoperative management is accepted for low-grade injuries (type I and II injuries), whereas operative management is indicated for type IV to VI [5]. Significant controversy remains regarding proper diagnosis and management of type III injuries [1,3-6].

The main goal of surgical treatment of AC joint dislocation is an anatomical reduction of the AC joint and to reconstruct the biomechanical forces of the CC ligaments [1]. Standard surgery for AC joint dislocation requires open approaches (e.g., Bosworth screw fixation, Weaver-Dunn procedure, clavicle hook plate fixation, ligament reconstruction) and is associated with an up to 27% rate of postoperative complications (e.g., surgical site infection, deltoid detachment, neurovascular injury due to extensive soft tissue dissection, osseous fractures from drill holes, osteolysis and loss of reduction) [5,6]. Radiographic loss of reduction rates are as high as 30 to 47% for both open and arthroscopic techniques [7-9]. In order to reduce these complications, there has been increased interest in the use of minimally invasive surgical techniques (the arthroscopic approach) to reconstruct the CC ligament [10].

These arthroscopic procedures theoretically offer superior view to the coracoid base, minimal soft tissue dissection, smaller incisions, fewer deep infections and the capacity to treat concomitant shoulder injuries. Literature review revealed an elevated risk for clavicular/coracoid fractures in open procedures compared to arthroscopic procedures, possibly due to increased soft tissue dissection in open surgery and suboptimal view on the coracoid base [5]. Furthermore, it has been stated that up to 40% of AC joint dislocations have associated other shoulder pathology such as partial articular supraspinatus tendon avulsion (PASTA) lesions, superior labrum from anterior to posterior (SLAP) lesions, labral tears and chondral defects [6,11,12]. In open surgical procedures these injuries will often be overlooked or missed. Arthroscopy offers not only diagnosis of such lesions but also the potential to treat these injuries simultaneously if desired. Good short-term results have been reported on arthroscopic AC fixation, however, there is scarce documentation on long-term results to confirm whether the initial results stand the test of time.

The purpose of this study was to document our surgical treatment of acute AC joint dislocations with an arthroscopic technique using the ZipTight Fixation System for AC Joint Reconstruction (Zimmer Biomet), and to assess clinical and radiographic long-term outcome compared to similar arthroscopic and conventional open techniques.

This study is a single-center, prospective, non-randomized, clinical and radiographic follow-up study. Research approval was obtained by the Ethics Committee of AZ Delta Roeselare (approval No. 21069; clinical trial No. B1172021000024). All patients who participated signed the patient information and consent form approved by the Ethics Committee.

From July 2011 through August 2018, 32 consecutive patients diagnosed with an acute AC joint dislocation were treated in our department with an arthroscopic technique making use of the knotless ZipTight single tunnel, double-button fixation system. Our technique is used for repair of acute grade III, IV and some grade V AC joint dislocations in slim patients (Rockwood classification). Diagnosis was obtained clinically and radiographically. We considered injuries presenting within 3 weeks to be acute. Exclusion criteria for this study involved patients older than 65 years, chronic AC joint dislocations, obesity, infection and lateral clavicle fractures. Other patients with AC joint injuries type I, II, and III (low functional demand patients) and those who refused surgical intervention were treated nonoperatively. All surgical procedures were performed by a single senior shoulder surgeon (SM) using a standardized procedure.

All patients were contacted for an additional clinical and radiographic assessment in the long term. Clinical assessment at last follow-up included the following data: AC stability, AC tenderness, AC compression test, shrug test, active and passive shoulder range of motion (ROM). Functional outcome was evaluated with retrospective preoperative and postoperative American Shoulder and Elbow Surgeons (ASES) scores, Simple Shoulder Test (SST), Specific AC Score (SACS) and Quick Disabilities of the Arm, Shoulder and Hand (DASH) scores. The subjective perception of pain was evaluated by a retrospective preoperative and postoperative visual analog scale (VAS). Additional anteroposterior (AP) radiographs of the AC joints were obtained in every patient at last follow-up to assess maintenance of CC reduction. Using these radiographs, CC distances were determined and compared with previously taken short-term postoperative radiographs by two independent researchers (EVE and SM). Since other studies revealed interobserver variability to be approximately 5 mm, a value of 5-mm increase was chosen to represent loss of reduction [6,13]. Possible AC osteoarthritis and CC ligament calcification/ossification were also documented. Postoperative complications and overall revision rates were documented. Finally, global satisfaction of the patients with the results of the surgery was assessed with “satisfied” or “not satisfied.”

Thirty-two patients met the inclusion criteria and participated in the study. Of this cohort there were 30 male and 2 female patients. The mean age of the patients was 37.4 years (range, 24–60 years; standard deviation [SD], 10.4). According to Rockwood's classification there were 21 patients (65.6%) with grade III, 6 (18.7%) with grade IV and 5 (15.6%) with grade V injuries. Mean follow-up time was 67.6 months (range, 14–123 months; SD, 17.6) (Table 1).

All patients obtained a full ROM at final follow-up. Two patients (6.3%) experienced mild AC tenderness upon clinical examination. These same two patients had a positive resisted AC compression test. Two other patients had a positive shrug sign at last follow-up. No clinical AC instability was observed. There were no complaints about irritation related to the clavicular button.

Pain was evaluated by retrograde preoperative and postoperative VAS. Mean VAS scores fell from 7.2 (range, 3–10; SD, 1.8) preoperatively to 1.0 (range, 0–7; SD, 1.5) at final follow-up. Mean preoperative SST and ASES scores were 0.3 (range, 0–2; SD, 0.6) and 25.1 (range, 0–80; SD, 18.1), respectively. At final follow-up, mean postoperative SST and ASES scores were 11.7 (range, 10–12; SD, 0.6) and 96.3 (range, 68–100; SD, 6.5), respectively. SACS decreased from 80.5 (range, 43–100; SD, 15.3) to 6.1 (range, 0–26; SD, 7.4). Quick DASH scores improved from 79.8 (range, 31–100; SD, 20.1) preoperatively to 3.7 (range, 0–22; SD, 6.4) at final follow-up. All differences in preoperative and postoperative outcomes scores were statistically significant (P<0.001) (Table 2). Thirty-one patients (96.9%) were satisfied with the procedure.

Radiographic evaluation using AP radiographs of the AC joint revealed a mean CC distance of 10.9 mm (range, 5.6–16.8 mm; SD, 2.6) at one day postoperatively. At final follow-up, mean CC distance was 12.2 mm (range, 8.4–21.1 mm; SD, 2.9). A mean difference of 1.3 mm was seen. Since other studies revealed interobserver variability to be approximately 5 mm, this suggested that there is no significant loss of reduction between short-term and long-term postoperative CC distance. In our series, two patients (6.3%) had a loss of reduction of >5 mm between short term and final follow-up (5.1 and 8.1 mm); neither patient showed an increase of >10 mm. These two patients had a radiographic relapse to grade II, but without clinical implications.

Twenty-three patients (71.8%) showed ossifications between the clavicle and the coracoid (Fig. 4). Ossifications did not impair ROM/strength/patient reported outcomes. Two patients (6.3%) showed radiographic signs of AC osteoarthritis. No signs of clavicular osteolysis were observed (Table 3).

During the diagnostic glenohumeral arthroscopy we found 31.3% of patients had concomitant lesions. Five SLAP type 1 lesions and four PASTA lesions were observed, and all were treated with arthroscopic debridement. There was one bucket handle tear of the superior labrum with loose biceps tendon (SLAP type 4). This lesion was treated with an arthroscopic biceps tenodesis.

In this study we sought to illustrate the advantages and reliability of an arthroscopic technique for acute grade III, IV and some grade V AC dislocations as compared to other available open techniques. We also want to stress the long-term reliability, which has rarely been reported in other studies. The first limitation of the study is the relatively small number of included patients. The most relevant limitation of this study; however, is the lack of a control group treated non-arthroscopically with the same implant. Therefore, our clinical and radiographic results will be compared with other open and arthroscopic reports in the literature. This technique is not appropriate for chronic dislocation or for several grade V injuries in obese or muscular patients and thus cannot be proposed as a single universal technique for all types of AC dislocation. However, the same goes for open procedures where the choice of procedure is personalized according to timing, surgeon, patient and injury variables.

Recent literature has confirmed that an arthroscopic single-tunnel CC stabilization is not sufficient for grade V injuries. Therefore, in all grade V injury cases, we perform an open procedure with allograft augmentation. One could state that an arthroscopic assisted procedure with a CC single tunnel, double-button construct and a self-locking ZipTight technology might not be as strong as other open procedures with the same indication. This could suggest our technique might lose reduction or fail altogether more easily. For this reason, we emphasized a long-term cohort of results in order to demonstrate the long-term reliability and healing potential. Furthermore, the application of this minimally invasive technique is restricted to acute injury because it relies on intrinsic, biological healing potential of the native CC ligaments. The double-button construct is merely there to keep the AC joint reduced in the early healing stage, bringing together the stumps of the conoid and trapezoid ligaments that in the early posttraumatic phase still have biologic healing potential. We consider injuries beyond 3 weeks not suitable for our technique as one expects that the torn ligaments have scarred and stumped by then with associated fading away of the initially liberated local growth factors. Beyond 3 weeks we perform an open technique with AC capsule reinforcement and CC tendinous allograft augmentation. The single arthroscopic technique of double-button construct would not be strong enough to resist the forces of a chronic dislocation without further healing potential.

Gowd et al. [5] systematically reviewed the outcomes and complications of different techniques of AC joint reconstruction. They report an overall complication rate of 14.2% in a heterogeneous group of 1,704 AC joint reconstructions with a mean follow-up of 34.3 months. The most common complications include infection, fracture of the coracoid or clavicle and hardware failure. Their mean infection rate is reported at 6.3% for both open and arthroscopic techniques, significantly higher than the 3.1% in our series. Another systematic review by Woodmass et al. [14] about complications in arthroscopic AC reconstruction reveals infection rates of 3.8%. They report no cases of deep infection following arthroscopic AC reconstruction. These findings are consistent with our results and confirm the theoretical advantage of arthroscopic surgery in reducing infection rate. This is particularly relevant when treating controversial grade III lesions, where the risk of infection associated with surgery may outweigh the benefits of AC joint reconstruction [14].

A recent meta-analysis by Gowd et al. [5] reveals a mean loss of reduction rate of 18.9% (19.3% in arthroscopic vs 18.5% in open techniques) and overall revision rate of 9.5% for all techniques [5]. There was no statistically significant difference between loss of reduction or revision rates for arthroscopic and open procedures. Our findings surpass these rates with a loss of reduction rate of only 6.3% and overall revision rate of 9.4% in the long term with a mean follow-up of 67 months. Our study thus illustrates a superior reduction maintenance and similar revision rate comparing to other open and arthroscopic techniques.

Some might argue against the arthroscopic approach as the constant lavage of physiological fluid during the procedure might wash away the remaining growth factors and biological healing properties. However, in open surgery the shoulder is rinsed as well during the procedure and the constant bleeding is drained by suction, sucking away the same biologicals. Furthermore, this rinsing argument is seldom used when discussing arthroscopic rotator cuff or Bankart surgery; now that these arthroscopic procedures have become the gold standard as opposed to their open alternatives, the washing away of biological healing factors is no longer seen as an active concern. Finally, the fact that the postoperative x-rays rather consistently show CC ossifications (71.8%) in the CC space suggests that bone marrow from the drill tunnel adds multipotent cells to the ligamentous healing area, regardless of an arthroscopic lavage effect. This can also be found in open procedures [15]. Scheibel et al. [16] described similar rates of CC ossifications (67.9%) and mention that these CC ossifications may be regarded as positive in terms of AC joint stability. CC ossifications did correlate inversely with the CC distance and did not impair ROM. Our results confirm these findings. Motta et al. [17] reported that CC ossifications increase the stiffness of the surgical reconstruction and are, therefore, positively correlated with a normal AC joint motion. In their findings, patients without CC ossifications had a more hypermobile/unstable AC joint.

There might be some indication that a single CC button construct is not sufficient to withstand both vertical and horizontal AP forces. To provide more AP stability to replace the function of the AC ligament, which is not repaired in a standard, across the board way in our technique except for according to the indications stated above, one can choose alternative procedures or implants. First, AC capsule reinforcement can be performed. Second, some studies have proposed using two implants to recreate the different directions of the trapezoid and conoid ligaments. Alternately, a single implant augmented with CC allograft ligament reconstruction through two extra clavicular bone tunnels can be used, or some companies also provide similar implants with a double tail on the clavicular part (e.g., Arthrex Twintail Tightrope). Gu et al. [18] compared double and single tunnel constructs and found no difference in clinical outcomes. However, radiological outcomes revealed the CC distance was significantly increased after 6 weeks in the single tunnel group compared to the double tunnel group. An important disadvantage of these double tunnel implants/techniques is the increased risk of fractures/tunnel break out [19].

Advantages of the arthroscopic approach thus lie in the minimally invasive exposure, better cosmetics, avoidance of partial deltoid detachment that might pose problems, and especially the superior viewing for adequate placement of the guide pin and drill tunnel. By this one advantage, some of the worst complications like misplacement of the tunnel with cutting out sideways or fracture of the coracoid (reported at rates of up to 20% in literature) can be limited [6]. These adverse events will lead to loosening of the construct with proximal migration, possible intraosseous migration of the distal button into the coracoid with loss of reduction, and the need to convert the procedure to another type of fixation either immediately or in a second stage. A recent literature review by Gowd et al. [5] revealed combined fracture rates in 5% of cases in open techniques compared to 3.4% in all-arthroscopic or arthroscopic-assisted techniques. Milewski et al. [20] described a combined clavicle/coracoid perioperative fracture rate of 18.5% with open and 10% with arthroscopic technique. Woodmass et al. [14] described a pooled fracture rate of 5.3% in arthroscopic fixation of AC separation, mainly at the level of the coracoid (7/1). Martetschläger et al. [6] also reported both open and arthroscopic techniques and found a mean rate of 20% with no statistical difference between both groups. In our study, we only had one such occasion with an intraoperative tunnel breakthrough (at the level of the coracoid) with need to convert to a hook plate (3.1%), which supports a low degree of fracture for arthroscopic techniques.

Another factor that may influence the fracture risk is the diameter of the bone tunnels used in AC joint repair. Limiting the diameter of the drill tunnel is an extra option to limit the risk of tunnel breakout, since biomechanical testing shows that smaller bone tunnels have an increased load to failure [19]. Newer techniques now allow bone tunnels of only 3 mm with retrograde passage of the implant, which will probably decrease the fracture risk in future. Besides increased diameter of the tunnel, the fracture risk should also increase with the number of tunnels [5].

The specific implant used in our study provides the opportunity to lock the system in the final step without the need for knot tying on top of the clavicular button (Fig. 4). In other implants this is often necessary, and this knot can cause hardware irritation or granuloma formation that requires removal in a later stage. Some other studies using a tightrope/endobutton fixation technique for AC joint reconstruction report hardware irritation rates of more than 25% [14,16,21,22]. Scheibel et al. [16] reported tenderness on the superior aspect of the implants in 39% of patients. Despite the high rate of hardware irritation, revision surgery for hardware removal was not routinely performed in any study [14,16]. This is in contrast to patients treated with open surgery with hook plates leading to a routine second surgery for removal.

Our findings match the earlier described high incidence of concomitant shoulder injuries (31.3%). Markel et al. [12] reported an incidence of almost 40% of co-existing shoulder injuries. However, in their study only 21.9% of all patients with concomitant pathology needed additional reconstructive surgery, since 70% of these injuries could be attributed to a mostly degenerative etiology. In our series, 31.3% concomitant injuries were seen, and only one patient with a SLAP type 4 injury required treatment (with arthroscopic biceps tenodesis). The main limitations of this study are the lack of a control group treated with open surgery and the small cohort size.

Arthroscopic stabilization for acute grade III, IV and some grade V AC joint dislocations offers satisfactory clinical and radiographic outcomes and our results demonstrate reliability of the arthroscopic technique in the long run. After a mean follow-up of 67.6 months, our clinical results showed a 100% clinical enforcement of AC joint reduction and a 93.7% rate of radiographic maintenance of reduction, exceeding reported rates of existing techniques. Additionally, we report high patient satisfaction rate, fewer complications and similar re-intervention rates compared to other techniques. We recommend this arthroscopic reconstruction as a readily reproducible technique and our findings refute the possible arguments against sufficient strength of a single tunnel arthroscopic technique.