Fractures of the humerus are discussed in four groups. diaphyseal, proximal, distal and combined. All can be challenging to treat. Best results occur by means which achieve early motion while maintaining satisfactory alignment so that union can be reasonably expected while preserving functional motion at the shoulder and elbow. Tolerance for angular malalignment of the diaphyseal component of the fracture is more liberal than in the lower extremity for the obvious reason that malangulation in the lower extremity interferes with gait. Secondly the multiplanar range of motion of the shoulder and the elbow/forearm articulations affords ample opportunity for funtional compensation for diaphyseal malangulations of as much as 30 degrees.
Beginning with the diaphyseal fracture, a parusal of abstracts over the last decade can easily yield a broad range of treatment options. As Osman puts it; "The indications for treatment should be eclectic"(Osman 1998). http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=10855286
In this abstract the displacement of a diaphyseal fracture is sufficient indication for operative treatment. The choice between rod, flexible rod, and plate fixation is guided by the pattern and degree of comminution. Could this perhaps reflect a different tradition in Europe vs North America? Seemingly even within Europe there is wide latitude taken as to indications for surgery.
Contrast this with the experience in Brug 1994. "Closed fractures without concomitant injuries can generally be managed conservatively." This seems closer to the North American standard if one could be said to exist, though there is no specific mention as to the criteria for an acceptable alignment. Furthermore the presence of a complete radial nerve palsy is taken as an indication to explore and internally fix the fracture, not generally taken as an indication for operative intervention in the US , with the caveate that some fracture patterns and perhaps, with disproportionate pain, or if the nerve goes out during splinting there is an indication to explore the radial nerve when it's function cannot be demonstrated.
Consensus it seems is lacking with regard to the optimal treatment for diaphyseal humerus fractures. It appears that a given patient and surgeon will in many cases opt for operative treatment where previously the patient would be managed with a coaptation splint and an off the shelf abduction pillow brace followed within a week or so with a functional orthosis, and continued use of the abduction pillow as needed to control varus alignment. Operative intervention would be reserved for non-unions or special circumstances, including open fractures, polytrauma, floating elbow.
We can speculate as to the reasons for the wide variation in treatment decisions for an otherwise uncomplicated diaphyseal humerus fracture. The traditional assumption is that closed treatment should be used when there is a reasonable expection of a successful outcome, so why then do we opt for surgery instead?
It is not that the results from the more conservative regimen have been found lacking, in any long term sense. Unions are generally achieved with satisfactory alignment and with preservation of motion at the shoulder and elbow. Indeed no bridge is really burned in the effort though the overall treatment course might arguably be prolonged if surgery is later required due to nonunion or inability to maintain a satisfactory reduction.
On the other hand it does not necessarily reflect favorably on the quality of the care provided when the intial closed treatment pursued does not work out, especially when 6 weeks of gainful employment is lost and the patient has endured weeks of discomfort and lack of sleep as the fracture became a non-union. Increasingly it seems surgeons must consider this a negative reflection on their care.