CORR Synthesis: What Is the Most Effective Treatment for Rockwood Type III Acromioclavicular Joint Dislocations?

In the Beginning…

Physicians have written about acromioclavicular joint (ACJ) injuries since at least 400 BC, when Hippocrates warned it is possible to mistake an ACJ injury for a glenohumeral joint injury [23]. Galen, another ancient Greek physician, sustained an ACJ dislocation and ironically did not adhere to treatment because he could not bear upper extremity immobilization [23]. Since then, the understanding and management of ACJ dislocation has greatly improved. Nonetheless, the most effective treatment for this painful and often disabling injury remains controversial more than 2000 years later. An epidemiologic study estimated that ACJ dislocations account for 9% of all shoulder injuries [6]. They are much more common among young men who participate in contact sports or other athletic activities in which direct impact to the shoulder is possible, such as cycling, body surfing, and skiing [7, 13, 19].

In the 1980s, Rockwood [22] created a six-part classification system to grade ACJ dislocations (Table 1). Rockwood’s system became the most widely accepted schema for grading this injury, surpassing Allman’s [1] and Tossy et al.’s [26] methods of the 1960s, which were broader and did not describe all ACJ injury patterns.

Table 1.

Rockwood classification [22]

Rockwood type	Acromioclavicular ligament	Coracoclavicular ligament	Deltotrapezial fascia	Direction of clavicle
I	Sprain	Intact	Intact	Nondisplaced
II	Complete rupture	Sprain	Intact	25% superior
III	Complete rupture	Complete rupture	Detached	25% to 100% superior
IV	Complete rupture	Complete rupture	Detached	Posterior through trapezius
V	Complete rupture	Complete rupture	Detached	100% to 300% superior
VI	Complete rupture	Complete rupture	Detached	Inferior to acromion or coracoid

The Argument

Orthopaedic surgeons typically treat Rockwood Types I and II ACJ dislocations nonoperatively, and they typically operate on Types IV to VI. But despite more than 2000 years of experience in treating this injury, surgeons disagree about how to treat Rockwood Type III ACJ dislocations. Even in Major League Baseball pitchers, in whom scapular mechanics are critical, disagreement exists among physicians as to how best to treat these injuries in this setting, with most favoring nonoperative management [20]. In 2019, a Cochrane review concluded the best evidence was derived from only two studies despite decades of specific investigation on this topic [25]. Over time, more than 150 surgical interventions have been introduced to treat these Type III ACJ dislocations [2], including the use of K-wires, Bosworth screws, hook plates, and newer arthroscopic techniques using suspensory loop fixation devices and tendon grafting. But the fundamental question remains: Is surgery—regardless of approach—superior to nonoperative care for Rockwood Type III injuries? Confirming the severity of this controversy, the International Society of Arthroscopy, Knee Surgery, and Orthopaedic Sports Medicine recently recommended consideration of subdividing Rockwood Type III injuries into Types IIIA and IIIB, to guide physicians toward better management [3]. This nonvalidated approach, which involves reexamination of patients 3 to 6 weeks postinjury to evaluate for scapular dyskinesis, leaves significant room for judgement on the part of the surgeon, which introduces inconsistency in its application. This system has not been widely adopted. With ACJ separations accounting for some 9% of shoulder injuries, more than half of these ACJ separations have been shown to be Type III dislocations [6, 9]. For these reasons, a critical evaluation of the current evidence comparing treatments is important in determining the best practice for the treatment of this injury.

Understanding the ACJ’s complex functional anatomy directly affects the results after treatment. The shoulder girdle and arm are suspended from the clavicle through the ligamentous complex of the ACJ. These ligaments include the ACJ capsule, which confers horizontal stability between the scapula and clavicle, as well as the conoid and trapezoid ligaments, which provide vertical stability. With an intact ACJ, the clavicle behaves as a strut that keeps the shoulder girdle away from the center of the trunk and allows the arm to pivot in a wide arc of motion through actions of the pectoralis major on the humerus. When the ACJ ligamentous complex is disrupted, the shoulder girdle and arm are no longer suspended from the clavicle. Gravity pulls it inferiorly, causing tenting of the distal clavicle on the trapezius, which may lead to an aching pain and fatigue in the periscapular muscles. In addition, forward arm flexion is biomechanically disadvantageous because of the lack of the stabilizing force of the ACJ capsule, leaving the distal clavicle to be pulled forward by the trapezius. Both issues can lead to chronic aching of the periscapular muscles and arm weakness.

Because the normal biomechanics of the ACJ are disrupted, nonoperative treatment of ACJ injuries can lead to chronic aching and weakness [24]. However, residual pain after surgery can also occur, owing to complications such as coracoid or clavicular fracture, loss of reduction, nerve injury, ACJ arthrosis, infection, and stiffness [18]. Given that operative and nonoperative treatment can lead to pain and dysfunction, the guidance of randomized controlled trials (RCTs) would be beneficial to a surgeon in dealing with this traumatic injury.

Essential Elements

We therefore sought to identify all RCTs published in the past decade comparing nonoperative and operative treatment of Rockwood Type III ACJ dislocations. On October 17, 2022, we searched the PubMed and Embase databases to include published articles from October 17, 2012, to October 17, 2022. The Boolean-style search included the terms “acromioclavicular joint dislocation OR shoulder separation” AND “Type III OR Type 3” to capture relevant terminology describing this injury. The initial search returned 543 articles from PubMed and 346 from Embase. There were 11 RCTs from PubMed and four RCTs from Embase after removing duplicates and limiting the search to full-text articles available in English published within the past 10 years. A number of these articles were excluded because of a lack of comparison with nonoperative treatment.

All RCTs were scored by the three authors (RWD, XAD, and REM) using the Jadad scale from 0 to 5 (higher scores represent superior methodologic rigor). A score of 3 or higher was required for study inclusion in this review, with the understanding that appropriate double blinding required for a score of 5 is not possible with the treatment for this injury. The Jadad scoring system is very objective and there were no disagreements among the three authors. Four RCTs qualified for this synthesis. A total of 286 patients, predominately young to middle-aged men, were included in these RCTs (Fig. 1)

Fig. 1.

The search and study selection process is shown in this flowchart.

What We (Think) We Know

One small RCT [12] found no differences between operative and nonoperative treatment at 18 to 20 years after surgery in patients treated for Type III and Type V ACJ dislocations. This is, to our knowledge, the only long-term study on the topic, but it is of limited utility because the treatments being compared (K-wire fixation and Kenny-Howard splints) are no longer widely used, and because there were only 11 patients with Type III ACJ dislocations in this small trial.

More commonly used surgical treatment methods were used in the other RCTs included in this synthesis. In addition, these studies addressed more than shoulder function. Data regarding return to work or sport, quality of life, cost, and cosmetic satisfaction were collected.

A recent RCT [5] found few important differences between patients treated operatively (with hook plates) and those treated with a sling and physical therapy. The main between-group differences favored nonoperative treatment at earlier timepoints, with that group having better DASH and Constant scores in the months after injury than the operative group, and the operative group having more (and generally more severe) complications, although that endpoint was not tested statistically. Patients in the operative group were also more likely to be satisfied with the cosmetic appearance of their shoulder. Although this is not a validated endpoint, cosmetic appearance may have significant importance to women wearing apparel that exposes their shoulders.

Another RCT [4] that also compared hook plate fixation to sling and physical therapy treatment found no differences between groups in terms of Constant scores, the Subjective Shoulder Value, and the DASH score 3 months after surgery. In addition, there were no differences in quality-of-life metrics 1 month after surgery. Interestingly, there were no differences in satisfaction between patients treated nonoperatively and those who underwent hook plate fixation in terms of the cosmetic appearance of their shoulders. These results are consistent with those of the previous study [5] and further support nonoperative treatment for Rockwood Type III ACJ dislocations.

An RCT comparing open reduction and tunneled suspensory device fixation (TightRope, Arthrex) to sling and physical therapy treatment found few differences between the groups, and those they found (earlier return to work and superior early DASH scores and cost of care) favored nonoperative treatment [21]. This operative treatment provided better reduction but was not functionally superior to sling and physical therapy treatment. At 6 weeks, more nonoperative patients had returned to work and had higher DASH shoulder function measures. There were no other differences between treatment groups except for the higher cost of operative treatment. The results of this study continue to show that regardless of operative treatment, nonoperative treatment is less burdensome to the patient and results in a shorter recovery time and lower cost (Table 2).

Table 2.

Results of RCTs

Randomized controlled trial	Number of patients	Patient demographics	Operative treatment	Nonoperative treatment	Number of complications (with specifics)	Outcomes
Joukainen et al. [12] (18- to 20-year follow-up)	25	Mean age = 55 years -23 men -2 women	K-wire fixation (n = 16) Type III (n = 7) Type V (n = 9)	Kenny-Howard splint (n = 9) Type III (n = 4) Type V (n = 5)	Total = 7 Operative: loss of optimal K-wire positioning (n = 4), infection (n = 1), broken K-wire (n = 1) Nonoperative: persistent pain requiring ACJ resection (n = 1)	-No significant statistical differences in CS, UCLA, Larsen, and SST results between groups.
Canadian Orthopaedic Trauma Society [5] (2-year follow-up)	83	Mean age = 37 years -78 men -5 women	Hook plate fixation (n = 40) (did not differentiate among Types III, IV, V)	Sling and physical therapy (n = 43) (did not differentiate among Types III, IV, V)	Total = 16 Operative: plate loosening (n = 2), acromial erosion (n = 2), stiff shoulder (n = 1), clavicle fracture (n = 1), deep wound infection (n = 1), superficial wound infection (n = 3), incision site numbness (n = 4) Nonoperative: fall on shoulder treated with operative ACJ stabilization (n = 2)	-Nonoperative group had superior DASH scores at 6 weeks and 3 months. No differences between groups at 1-year and 2-year follow-up. -Nonoperative group had superior CS at 6 weeks, 3 months, and 6 months. No differences between groups at 1-year and 2-year follow-up.
Boström Windhamre et al. [4] (2-year follow-up)	118	Mean age = 40 years -111 men -7 women	Hook plate fixation (n = 61) Type III (n = 30) Type V (n = 31)	Sling and physical therapy (n = 60) Type III (n = 31) Type V (n = 29)	Total = 4 Operative: wound infection (n = 1), stiff shoulder (n = 1), redislocation after plate removal (n = 2) Nonoperative: none	-Nonoperative group had higher CS, QuickDASH, and SSV scores at 3 months, and higher quality of life measures at 1 month. There were no differences at other timepoints.
Murray et al. [21] (1-year follow-up)	60	Mean age = 30 years -56 men -4 women	TightRope, Arthex Tunneled Suspensory Device (n = 29) Type III (n = 18) Type IV (n = 11)	Sling and physical therapy (n = 31) Type III (n = 18) Type IV (n = 13)	Total = 5 Operative: none Nonoperative: persistent discomfort resulting in operative treatment (n = 4), concern about cosmetic appearance requesting operative treatment (n = 1)	-Nonoperative group had superior DASH and OSS scores at 6 weeks. There were no differences at other timepoints. -Nonoperative group returned to sport at 6 weeks and 1 year compared with operative group.

CS = Constant score; UCLA = University of California Los Angeles; SST = simple shoulder test; SSV = subjective shoulder value; OSS = Oxford shoulder score.

Knowledge Gaps and Unsupported Practices

It is apparent there are few RCTs comparing operative with nonoperative treatment for Rockwood Type III injuries and that further study, particularly focused on patients who did not do well with nonoperative treatment, is warranted. However, the best evidence consistently demonstrates that operative treatment provides no clinically important benefits. Although operative treatment can provide anatomic reduction, the long-term benefits associated with this, if any, have not been proven. Although there are many studies—mostly case series—advocating surgery using newer technologies, none have yet demonstrated superiority over nonoperative treatment; until they do, surgeons (and their patients) should be wary.

Further supporting a cautious approach when adopting new surgical approaches for ACJ surgery (and even when recommending established approaches) is the frequency with which patients experience complications after these interventions. This is especially important when high complication rates have been documented. An RCT conducted by the Canadian Orthopaedic Trauma Society had seven major and seven minor complications in the operative group of only 40 patients [5]. The best evidence at present is that nonoperative treatment is at least as effective, more cost effective, and allows for faster recovery than operative intervention [8].

That being so, we consider surgery as a first-line treatment for Type III ACJ dislocations to be an unsupported practice, because it involves subjecting patients to risks without any clinically important benefits.

Barriers and How to Overcome Them

The best evidence to support a treatment comes from RCTs. The problem with the current RCTs evaluating the treatment of Type III ACJ dislocations is that there are variations in how the diagnosis is defined radiologically. Adding to the complexity are the numerous surgical procedures, varied expertise levels among surgeons, and differences among patients. Additionally, it is difficult to enroll large numbers of patients to adequately power these types of studies. Unfortunately, case series about surgical techniques cannot tell surgeons whether those approaches are superior to nonoperative treatment, which is ultimately the key information we seek. We agree with Teppo Järvinen MD, PhD (who has coordinated many large, randomized surgical trials [16]) when he opined, “Procedures that carry greater risk (like shoulder surgery) should be superior to interventions with little or no risk (like shoulder exercises)” [11].

Although the Rockwood classification is not perfect, it is currently the best system to classify these injuries. There are two steps any treating physician can take to improve patient care for ACJ dislocations: The first step is to choose the most appropriate imaging techniques. The manner in which initial radiographs are taken for ACJ dislocations may lead to an incorrect diagnosis and alter treatment. One study found that bilateral weighted radiographs allow for a more precise Rockwood classification than a single AP image of the injured ACJ [10]. These findings highlight a barrier in treating these injuries, because injury severity can be underestimated [10]. We have cared for patients whose pain persisted after what appeared to be a Type III injury; bilateral weighted radiographs demonstrated these, in fact, to be Type V injuries. Careful history and physical examination supplemented by thoughtfully ordered and interpreted imaging can remove this barrier to good care. The other approach physicians should employ is to set expectations carefully [17]. In particular, we need to educate patients that some persistent discomfort and a mild deformity after a severe shoulder injury is not a failure of treatment. However, the largest barrier to good care is established practice patterns driven by false impressions of our efficacy as surgeons; we cannot fix everything and should not try [13]. We feel as though we are helping more than we really are; our happy patients return, those with complications after our interventions often go elsewhere [15]. We believe surgeons should not offer surgery to patients with Type III ACJ injuries because the best-available evidence from several high-quality RCTs points in the same direction: This treatment is no better than nonsurgical approaches. We hope to see strong leadership from orthopaedic societies to help make this point more broadly.

5-year Forecast

We believe that in the next 5 years, there will be more effective improvements in nonoperative treatments. A clinical trial from Switzerland compared clinical outcomes between using a sling versus the Acromion 2.0 brace (OrthoService) and was completed in the past year (https://clinicaltrials.gov/ct2/show/NCT03261778?id=NCT03261778&draw=2&rank=1). We predict the results from this trial will further substantiate nonoperative treatment. Physical therapy programs that are functionally based and focused on restoring dynamic acromioclavicular, scapular, and shoulder motion could further improve the results of nonoperative treatment [24].

Still, we acknowledge there is a subgroup of patients in the nonoperative arm of some RCTs whose pain or shoulder dysfunction persisted. Some of these patients crossed over to operative treatment and unfortunately continued to have lower outcome scores. Specific features of the patients who present in this fashion have not been identified in the acute postinjury period. Further study of these patients who did not do well with nonoperative treatment may reveal important information regarding treatment implications. We wonder whether there is some specific anatomic problem in these patients who might benefit from delayed surgical intervention, which has been shown to have no difference in outcomes compared with early surgical intervention [14].

To provide the most effective care for patients, it is essential for surgeons to support their clinical decisions with high-quality trial-based evidence when it is available, as is the case for Rockwood Type III ACJ dislocations. Recommending costly and possibly harmful surgical treatments for patients with these injuries when those treatments have not been shown to be better or less harmful than nonoperative treatment can result in needless surgical complications, is unlikely to improve patients’ pain or function, and erodes critical trust in doctor-patient relationships [8].