In the early 2000s Nowak et al.  shattered the belief that mid-shaft clavicle fractures always healed fine; even after 10 years almost half of the patients had remaining symptoms*. This common injury had also by in others  (with shorter follow-ups) been hinted as problematic, and within short we were showered with fancy new implants. After ≥ 7 RCTs ,  on the subjects it seems that these new implants failed do deliver. Can we finally start questioning if surgery is the solution?
* answered no to “Are you fully recovered from your clavicular injury?”
The dreaded non-union
It is true that once there is a bony unions between the fracture parts, the fracture no longer causes pain. Unfortunately the opposite is not always true; a fracture can heal with only fibrous tissue, i.e. without a bony union, but still not be painful. There is clearly a higher risk of pain with non-unions and in mid-shaft fractures they seem to occur in about 10-20 % of the cases.
Why not surgery then?
We know that surgery almost entirely removes the risk of non-union, all the RCTs are pretty clear on this point. Unfortunately surgery has its own risks; see this thread on HealthBoards, the initial post excellently pinpoints one of these risks:
… I still have a lot of nerve pain including numbness, BURNING, tingling and TIGHTNESS up and down my shoulder where the incision was made and goes as far down as my elbow and up my neck making it hard to turn my neck in certain directions. I try not to wear shirts when I can because the rubbing from the fabric really gets painful…
Many of the studies support that this is not a trivial issue as many report that a 1/3 of the patients remove the hardware due to symptoms. While the above is hopefully an outlier, the number of patients with local issues is troublesome. One can also wonder if the standard DASH adequately identifies the above symptoms or if they’re drowned in the noise from other questions.
The case for delayed surgery
The current evidence seems to show that surgery does not change the final outcome. It is important to remember that the outcome is a mean, i.e. a mix of poor and good outcomes. My guess is that most do well, and a few do poorly†. If we compare two cases that do poorly, one after surgery and one without surgery, it it is most likely easier to help the latter by trying the surgical option.
† it is surprisingly hard to find a histogram of the outcomes in any of the studies.
Why not be smarter about which patients to operate?
There have been attempts at identifying patients that do poorly with predictor variables such as shortening, Z-fractures, scapula position… the evidence that I’ve seen on the subject is, at best, a weak hint for which patients we should follow more closely . If you believe that you are a surgeon that is better at this than your peers, or that you are a patients that does not fit the RCTs mentioned, then please consider that you may suffer from better-than-the-average syndrome.
Prediction is hard and our intuition is often wrong. This does not translate into that if you choose treatment A instead of B everything will be fine, you can only increase your chances of being fine. Life is about accepting risk and that bad things may happen even though the risk estimates were on our side.
Why is it so hard to fix a simple clavicle!?
It is sometimes easy to forget that we’re new to surgery. We are trying to beat evolution that has had millions of years to figure out how to heal bones. The clavicle fracture has furthermore been put under a lot of selection pressure since:
- it is common
- often occurs at a young age
- it is hard to survive with only one arm
Our bodies have therefore most likely found a good way of dealing with this injury. We may not be fully restored afterwards, but this doesn’t mean that a new fancy plate/nail will make it better.
- A mid-shaft fracture will most likely fare equally well without surgery
- Some patients may suffer residual issues
- Non-unions are rarer with surgery but surgery has its own down-sides
…even Nowak and his team seem to have come to the same conclusion, see their paper titled “Overtreatment of displaced midshaft clavicle fractures” .
 J. Nowak, M. Holgersson, and S. Larsson, “Can we predict long-term sequelae after fractures of the clavicle based on initial findings? A prospective study with nine to ten years of follow-up”, J. Shoulder Elbow Surg., vol. 13, no. 5, pp. 479–486, Sep. 2004.
 J. M. Hill, M. H. McGuire, and L. A. Crosby, “Closed Treatment of Displaced Middle-Third Fractures of the Clavicle Gives Poor Results”, J Bone Jt. Surg Br, vol. 79–B, no. 4, pp. 537–538, Jul. 1997.
 S. Woltz, P. Krijnen, and I. B. Schipper, “Plate Fixation Versus Nonoperative Treatment for Displaced Midshaft Clavicular Fractures: A Meta-Analysis of Randomized Controlled Trials”, J. Bone Jt. Surg., vol. 99, no. 12, pp. 1051–1057, Jun. 2017.
 M. J. S. Tamaoki, F. T. Matsunaga, A. R. F. da Costa, N. A. Netto, M. H. Matsumoto, and J. C. Belloti, “Treatment of Displaced Midshaft Clavicle Fractures: Figure-of-Eight Harness Versus Anterior Plate Osteosynthesis”, J. Bone Jt. Surg., vol. 99, no. 14, pp. 1159–1165, Jul. 2017.
 A. Jørgensen, A. Troelsen, and I. Ban, “Predictors associated with nonunion and symptomatic malunion following non-operative treatment of displaced midshaft clavicle fractures–a systematic review of the literature”, Int Orthop, vol. 38, no. 12, pp. 2543–2549, Dec. 2014.
 I. Ban, J. Nowak, K. Virtanen, and A. Troelsen, “Overtreatment of displaced midshaft clavicle fractures”, Acta Orthop., vol. 87, no. 6, pp. 541–545, Nov. 2016.