CORR Insights®: What Is the Best Evidence for Management of Displaced Midshaft Clavicle Fractures? A Systematic Review and Network Meta-analysis of 22 Randomized Controlled Trials

Where Are We Now? Clavicle fractures generally affect young, active patients, and so when treating these injuries, our goal is to obtain early union and to return the patient to his or her prior function. For a long time, non-surgical management had been the preferred treatment option for this patient population among orthopaedic surgeons, because clavicle fractures generally heal well, and historically, surgery had been associated with a high risk of complications and sometimes-unpredictable results [4, 13]. Just over 10 years ago, a multi-center randomized trial began a shift in what had been a stable practice pattern towards the more-frequent use of surgery [2]. In this study, the authors randomized 132 patients with a displaced midshaft clavicle fracture to either operative treatment or non-operative treatment and found that early plate fixation of displaced midshaft clavicle fracture allowed patients to return to function earlier and decreased rates of nonunion and malunion. Patients treated surgically were also more satisfied with their shoulder compared to patients who had non-surgical management. Subsequent to this, a number of other studies provided supportive evidence for surgical treatment over non-surgical treatment [5, 6, 8, 9]. In the current study, Axelrod and colleagues [1] performed a network meta-analysis of randomized, controlled trials and found that while surgical management resulted in a higher chance of union, the rest of their evaluated parameters did not show major differences when compared to non-surgical treatment. In particular, surgery did not result in clinically important improvements in validated Constant-Murley and DASH scores. According to the results of the current study, patients should not assume that they will achieve better function after surgery. Moreover, most patients can avoid surgery with little risk of a nonunion. Where Do We Need To Go? Although the current study shows that surgical treatment is unlikely to show meaningful differences in functional scores at 1 year compared to non-operative treatment [1], several questions remain. Although the current study and another recent meta-analysis [12] found no differences in outcomes scores, the absolute risk of nonunions in non-surgically treated patients is low (to avoid one nonunion in the current study, 10 patients would have to undergo surgery), and surgery itself adds risk and cost, surgeons seem enthusiastic about offering it; even the other meta-analysis that showed so few differences favoring surgery focused its conclusion on fracture healing and “appearance” rather than on what seem like the things that matter most to patients [12]. Why is this? Are the high-quality studies being analyzed in these meta-analyses [1, 12] somehow missing something important that surgeons believe they are seeing? Are we measuring results with the right instruments? Beyond answering the questions above, we also need fill several gaps in our knowledge. For example, we need to improve the way we measure results. Many outcomes tools we use to assess patients with clavicle fractures, such as the Constant-Murley Score, are joint-specific scores, but not injury-specific scores [3]; some others, like the DASH are not even joint-specific. While for many uncommon diagnoses, these should suffice, perhaps clavicle fractures are common enough that this injury deserves its own scoring system? Certainly others have raised concerns about the appropriateness of many commonly used instruments [7]. Additionally, studies on this topic need to go a bit deeper in terms of what we mean by revision surgery, and whether different kinds of revision really are comparable to one another; it is possible that patients may feel that some re-operations—particularly if they are anticipated or likely, such as plate removal—may be part of the “price of admission” and not all that troubling. By contrast, other patients may feel that any risk of reoperation would cause them to prefer non-surgical management. We also need to develop: (1) a clear definition of irritation, (2) improved classifications, and (3) better measurements to accurately gauge its potential severity. We also need to categorize complications. For example, a primary pseudoarthrosis does not have the same impact as a pseudoarthrosis after a failed osteosynthesis, since reoperations for non-union after prior attempts at internal fixation can be longer, riskier, and more complicated. How Do We Get There? In order to determine why there seems to be such large differences between results from meta-analysis [12] and network meta-analysis [1], which seem to favor non-surgical approaches, and typical practice patterns, which generally involve a great deal of surgery, we need to consider important factors not evaluated in most studies. For example, patients with a fractured clavicle have pain, functional impairment, difficulties dressing, moving, and working for weeks. Most of these problems disappear and patients return to their usual tasks shortly after surgery. Surgeons and researchers need to improve the objects of investigation and how to communicate their results. We can do this by being more precise in the way we measure results through the use of specific tools. To develop outcomes tools specific to patients with clavicle fractures, we need to include scores that evaluate the clavicle itself in addition to analyzing the function of the shoulder and upper limb. An evaluation after a clavicle fracture no matter the treatment should include: (1) the ability to carry weight on the clavicle, (2) local irritation and paraesthesia, (3) scars and deformities, and (4) pain in the area of the fracture and adjoining joints.