Rockwood’s classification of acromioclavicular dislocation is based on the degree and direction of clavicular displacement [¹]. Grades I and II are benign and are widely regarded as best managed conservatively [², ³]. There is a general consensus that type V and VI lesions should be treated operatively [², ⁴]. However, there remains controversy over the optimal management strategy for grade III and IV injuries [⁴–⁷]. Grade III is classified as a superior displacement of the lateral end of the clavicle of one clavicular diameter or 1 cm on the anteroposterior radiograph, whilst grade IV is described as a separation of the acromioclavicular joint with the distal clavicle displaced posterior into the trapezial fascia [⁶, ⁷]. In both grades the acromioclavicular and coracoclavicular ligaments are torn.

Advocates of non-operative treatment suggest that patients often regain excellent clinical results and painless shoulder function, although for some there is the potential for chronic instability and pain [⁸, ⁹]. Alternatively, operative treatment strategies are able to address these shortcomings, but occasionally compromise shoulder function [², ⁸].

Given this degree of equipoise, the purpose of this study was to compare the clinical outcomes of patients managed operatively and non-operatively following grade III acromioclavicular dislocation.

The principal finding of this study was that, for the majority of outcomes, there was no statistically significant difference in clinical or radiological outcomes between operative and non-operative management for this patient group. Nonetheless, there was some evidence to suggest that operative management provided a significantly better Constant score compared to non-operative following grade III acromioclavicular dislocation, but this was based on the results from a single study. Non-operative management was associated with significantly poorer cosmetic outcome but less sick leave compared to operative management (P < 0.001).

The current evidence base presented with a number of methodological limitations, including not randomising patients to group allocation, permitting allocation bias [²³], and not blinding assessors to subject groups, therefore increasing the risk of assessment bias [²⁴]. Finally, the studies did not base their sample sizes on power calculations, increasing the risk of a type II statistical error due to an insufficient sample size [²⁵]. Accordingly, future robust, well-designed RCTs are required to improve the currently poor evidence base in order to determine the optimal management strategy for patients following grade III acromioclavicular separation.

A previous meta-analysis by Philips et al. [⁶] ultimately advised against surgical treatment following grade III acromioclavicular separation. This review differed to the previous review as it specifically included only those studies with cohorts of predominantly grade III acromicoclavicular separation. Furthermore, with the advantage of time, we have been able to include a number of studies which have recently been published on this topic. Whilst there is agreement with some of Philips et al.’s [⁶] conclusions, this study concludes, with some reservations, that there is little difference in the outcome of operative and non-operative management for patients following grade III acromioclavicular separation, with the exception that non-operative management provides cosmetically poorer outcomes. A more recent paper [²²] has shown that maintenance of reduction is possible with the operative group having a statistically better outcome than the non-operative group. As operative techniques improve, there may be a paradigm shift from the historically poor results of fixation with K-wires.

The mechanism of injury appeared similar among the studies, with a combination of sporting, accidental and occupation trauma as the associated factor. Few studies distinguished whether upper limb dominance was a factor in outcome. This may have a been particularly important confounding variable for functional-based outcomes and return to sports measures, where those with a dominant limb injury may present with poorer outcomes—particularly during early review—compared to non-dominant limb injury. A further confounding factor which may have affected outcome was time from injury to surgery. Rolf et al. [²] reported that those patients who had an acute acromioclavicular reconstruction after trauma reported significantly better functional outcomes and patient satisfaction rates as well as lower complication rates compared to patients with delayed reconstruction. Whilst the four studies reported that all operations were acute, the duration from injury to surgical reconstruction was not clearly stated in the papers of Galphin et al. [²⁶] or Taft et al. [⁹]. Finally, to the study’s credit, the follow-up period of the evidence base was reasonable, providing some evidence for detecting late failures and longer-term outcomes.

The results of this meta-analysis indicate that there was no significant difference in respect to maintenance of anatomical reduction between operative and non-operative management of grade III acromioclavicular separation (P = 0.15). Calvo et al. [⁸] acknowledged that complete reduction may not necessarily be a pre-requisite for optimal functional outcome [⁶, ⁸, ²⁷]. They suggested that the rationale of surgical reconstruction to achieve anatomic alignment for full functional recovery may not always be achieved following grade III acromioclavicular separation [⁸]. Thus, anatomical reduction alone cannot justify operative intervention. However, the method of assessing anatomical alignment was unclear from the included studies. Previous authors have argued that only by assessing the acromioclavicular joint with stress radiography can anatomical position be determined [²⁸]. Accordingly, future study is recommended to determine the optimal method of radiographic evaluation of acromioclavicular displacement following operative and non-operative management strategies.

The current meta-analysis suggests that there was no difference in the incidence of OA or ossification of the coraclavicular ligament between the two management strategies. Authors such as Calvo et al. [⁸] have suggested that the incidence of OA changes may be related to the surgical manipulation and inability to maintain reduction, whilst ossification of the coracoclavicular ligaments has been associated with the manipulation of ligament tissue when attempting to repair it [⁸, ²⁹]. Fremerey et al. [¹⁵] and Taft et al. [⁹] suggested that post-traumatic OA in surgically managed patients is related to the unphysiological contact of traumatised joint surface and subsequent joint cartilage injury.

Several authors have suggested that surgical reconstruction should be advocated for those patients who have physically demanding occupations or sporting interests. However, since the mean age of each study’s cohort was under 28 years, and the mechanism of injury was largely sporting or occupationally related, there was little evidence to substantiate this claim based on clinical outcomes. Furthermore, since this study suggested that duration of sick leave was significantly higher following non-operative procedures, and that there was no significant difference in strength outcomes, then non-operative management may be seen as superior to manage this patient group. For those patients who carry heavy weights on their shoulders, such as soldiers carrying rucksacks, operative intervention may be indicated to prevent anatomical deformities from affecting return to normal activities.

The literature poorly described the non-operative management strategies used. Historically, various straps, harnesses, casting techniques and traction methods have been used as part of closed reduction [³⁰–³⁴]. Currently, there appears greater support for the use of internal rotation slings. Since non-operative management strategies were not clearly defined, it remains unclear as to whether there was a variation in these strategies between the studies. Furthermore, it also remains unclear as to whether clinical outcomes are affected by the type of rehabilitation programme adopted, immobilisation method or period of immobilisation.

As Gstettner et al. [²²] acknowledged, the disadvantage of all operative strategies is the risk of complications. This was mirrored by our study, which demonstrated a significantly higher risk of infection following surgical management compared to non-operative treatment (P = 0.03). However, the incidence of infection was relatively low following acromioclavicular surgery. There has been a paradigm shift in clinical practice. Earlier studies adopted Phemister fixation methods. This developed into a consensus of using Bosworth screw and then later Hook plate fixation methods [³⁵–³⁸]. Currently, TightRope fixation methods and biodegradable slings have been introduced [³⁹, ⁴⁰]. Whilst clinical differences between operative and non-operative strategies have evaluated previous surgical interventions, the comparison to biodegradable sling fixation is yet to be evaluated using a large, well-designed RCT.

Finally, no study compared cost-effectiveness with a formal economical evaluation. Since the meta-analysis indicated that whilst patients reported a shorter duration of sick leave following non-operative management, and that higher costs of hospitalisation, the operative procedure and prolonged rehabilitation are associated with this strategy, there initially appears to be greater support from an economic perspective for adopting a non-operative management strategy for this patient group. Formal health economical assessment is therefore imperative to assess the differences in this and clinical outcomes when developing the evidence base with well-designed, sufficiently powerful RCTs.

To conclude, based on the current evidence base, operative management of grade III acromioclavicular dislocations results in a better cosmetic outcome (P < 0.0001) but a greater duration of sick leave (P < 0.001) compared to non-operative management. There was no difference between the two interventions in terms of strength, pain and throwing ability (P > 0.05).

Study performed at the University of East Anglia and St George’s Hospital, UK.

We would like to thank the staff at the Sir Thomas Browne Library, Norfolk and Norwich University Hospital for their assistance in gathering the papers required for this review. We also thank Ms Leigh Davies, Senior Orthopaedic Physiotherapist, Norfolk and Norwich University Hospital, for her assistance in the preparation of this paper.

Conflict of interest None.

Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution and reproduction in any medium, provided the original author(s) and source are credited.