Article Text
PDF
Statistics from Altmetric.com
Commentary on: OpenUrlCrossRefPubMed
Context
The operative treatment of proximal humerus fractures with locking plates has substantially increased.1 Just recently, our study group published a meta-analysis showing no difference between non-operative treatment and surgery.2 After this meta-analysis, a novel pragmatic multicentre PROFHER study was published.
Methods
In the PROFHER trial, 32 UK hospitals screened patients over 16 years old (n=1250), who had a displaced proximal humerus fracture, for study eligibility between 2008 and 2011. Fractures of the proximal humerus that involved the surgical neck with ‘sufficient’ displacement to be considered for surgery were included. However, there were no displacement criteria (eg, Neer). The choice of surgical intervention (humeral head replacement with prosthesis or internal fixation with nail, plates, k-wires or screws) was left to the surgeon. There were various non-operative treatments performed based on the surgeon's preference.
Findings
Two hundred and fifty patients (20% of the total screened patients), aged 25–92 years old, were recruited. No clinically or statistically significant difference between groups was observed after 2 years follow-up, when measured with the Oxford Shoulder score (difference 0.75 points (p=0.48, 95% CI 1.33 to 2.84 points) in favour of surgery). Subgroup analysis by age (under and over 65 years) revealed no differences. There were complications (eg, metalwork problems, symptomatic malunions, stiffness, cuff tears, etc) in 24% (30/125) of the operative group and in 18% (23/125) of the non-operative group (p=0.28).
Commentary
Pragmatic trials aim to include a heterogeneous population, which results in higher external validity. Thus, the theoretical advantage of the pragmatic trials is that they measure effectiveness of the treatment in routine practice.3–5 However, PROFHER only included 20% of the screened population; while, some explanatory randomised controlled trials (RCTs) have included 48% and 29%, respectively, indicating that PROFHER had similar problems as these explanatory RCTs.6 ,7 In addition, the strength of the trial was diminished because of a lack of true comparative treatment groups—patients were treated in any fashion the surgeon preferred. We think this leads to a wide set of unnecessary and unfortunate interventions. Hence, the PROFHER trial described the current care in the UK in a cohort manner, but lacked the scope to help clinicians who struggle daily to make decisions for a new patient. Another grievance was the lack of any fracture classification. Included were the fractures with ‘sufficient’ displacement, but the Neer or any other classification criteria were not used. Thus, the morphology of the included patient's fractures is not known. The same problem applies for all RCTs on proximal humerus fractures.8
Implications for practice
Despite the criticism presented, the PROFHER trial provided valuable information on a large portion of the population. It showed that, in general, there are no differences between operative and non-operative treatments when considering all age groups and all treatments. One must not forget that the results of any trial apply only to the population included. The line between pragmatic and explanatory RCTs is not always clear and, as Schwartz and Lellouch3 declared, the attitude of design characterises the trial. Pragmatic trials are not here to replace explanatory trials; instead, they complement them. Large, explanatory RCTs are still warranted to assess the optimal treatment of displaced proximal humerus fractures.
Footnotes
Competing interests None declared.
Provenance and peer review Commissioned; internally peer reviewed.