We are indebted to the surgeons who were curious enough to test their own biases. We are even more indebted to the patients who understand the value of clinical research enough to be randomly assigned to operative or nonoperative treatment so that future patients can make a more informed decision.
As the authors point out, the key to interpreting this study is that the primary outcomes were the Disabilities of the Arm, Shoulder and Hand (DASH) and Constant scores rather than fracture union. In spite of the fact that displaced clavicular fractures have difficulty healing and heal out of place without operative treatment (substantial disease or pathophysiology), patients treated with or without surgery have comparable arm-specific disability (comparable illness). Six (24%) of twenty-five nonoperatively treated fractures did not heal, but none of these patients had sufficient symptoms or disability to find surgery appealing. Three patients had symptoms related to malalignment of the fracture, and one underwent surgery.
I can hear some of my American colleagues saying “Yeah. Maybe in Finland.” Finns may have exceptional adaptation and resilience—and if they do, we should make it a priority to figure out how to make these exceptional coping strategies accessible to all—but what I’ve always wondered in the United States is, “Where were all of the clavicular nonunions before?” I know that some will say that in the past, surgeons told patients with a clavicular nonunion that nothing could be done and left them to suffer. But I’ve met patients with clavicular nonunions that they were not aware of, and patients with diagnosed nonunions that were not very bothersome. On one visit to the Cleveland Museum of Natural History, 200 clavicles from about 100 years ago had been laid out, and I was impressed that about fifteen of them had fractures and there were two or three nonunions. I think that clavicular nonunion may pass my “cave person” rule. What happened to people with displaced clavicular fractures that failed to heal when there were no doctors to see them, no x-rays to image them, and no implants to fix them? I think it’s safe to say that “cave people” with clavicular nonunions were able to care for themselves well without any effect on life span.
That’s not to say that clavicular nonunions don’t affect upper extremity use. Even though patients with nonunion did not request surgery, they did have greater symptoms and disability as measured by the DASH score. The problems associated with a clavicular nonunion seem too subtle on average to be measured by the Constant score, which primarily addresses motion and strength.
I’d like to highlight a few other things. First, the handling of missing data is important as it can introduce bias. An initial analysis of the results from a recent Canadian Orthopaedic Trauma Society trial appeared to show a significant difference in union rate between treatment groups, but a reanalysis using last-carried-forward data showed no significant difference1. Researchers should always remember to specify how they will handle missing data prior to enrolling the first patient.
Second, the clavicular surgery in this study was performed with subperiosteal stripping and fixation using a non-locked 3.5-mm reconstruction plate placed in the anterior position with at least three screws in each fragment. One plate bent and another broke, but both of these fractures healed. More data are needed to determine the degree to which muscle and periosteal attachments should be preserved and the optimal type and position of the plate, but the evidence to date suggests that technical details are relatively unimportant.
Finally, substantial fracture displacement (dichotomized as a displacement of >1.5 bone widths) was the only risk factor for nonunion among nonoperatively treated fractures.
Patients and surgeons should decide together how to treat a displaced clavicular fracture. I recommend the development of a decision aid in the form of a video or an interactive web site that presents the current best evidence to patients in a way that they can understand2. Patients can use this aid to clarify their treatment goals and preferences and come to a decision that suits them. Boiled down to one sentence, the best evidence to date is that fracture-healing will not occur in about one in four patients (one in four in this study) with a displaced diaphyseal fracture of the clavicle that is treated nonoperatively (the greater the displacement, the greater the risk), but—at least in Finland—most don’t have enough symptoms or disability to request later operative treatment.