A prospective randomised controlled trial of operative versus non-operative management of fractures of the humeral diaphysis: the HUmeral Shaft Fracture FIXation (HU-FIX) Study protocol

Potential HU-FIX study participants will be identified at the point of referral to the Orthopaedic Service, either directly (via the on-call team) or via the Trauma Triage Clinic system [26]. Patients satisfying the inclusion and exclusion criteria will be introduced to the study by the treating clinical team and provided with a patient information sheet to read before being asked to participate. Patients will be predominantly enrolled into the study either on the Orthopaedic Ward (if admitted) or at their first outpatient clinic appointment. A study investigator will review the study protocol in detail with the patient and address any of their questions. If the patient is willing to participate, a study investigator will complete the informed consent process.

Once informed consent has been obtained, patients will be enrolled into the study by an investigator and randomly allocated to either operative or non-operative management of their humeral shaft fracture.

The study will use a parallel assignment model, with 1:1 allocation of patients into each treatment arm (operative or non-operative). Randomisation will be stratified by patient age (< 65 years old or ≥ 65 years old), to ensure there are approximately equal numbers of younger and older patients in each treatment arm. Using block randomisation with a mixed block size and a 1:1 ratio of operative to non-operative, a computer-generated randomisation schedule was produced by a senior statistician from the local clinical research facility using nQuery Advisor v7.0 software (Statsols, Cork, Ireland). A member of staff independent of the study used this schedule to produce a series of opaque sealed envelopes, each containing a sticker bearing the words ‘operative’ or ‘non-operative’. Once the randomisation envelope is opened, the sticker is placed onto the already-signed consent form to indicate the treatment group allocation.

All patients will be initially treated in the Emergency Department with closed reduction of their humeral shaft fracture, and application of a either a U-slab or above-elbow hanging plaster cast.

Patients allocated to operative management of their humeral shaft fracture will undergo surgical fixation, using a standard technique of plating and screw fixation. The exact surgical approach and fixation technique utilised will be at the discretion of the treating surgeon. Similarly, post-operative immobilisation and range-of-motion restrictions will be at the discretion of the treating surgeon. This is determined by a number of factors including the injury and fracture pattern, bone quality, co-morbidities and patient compliance. The approach, fixation technique and range-of-motion restrictions will be recorded for all patients and presented as a sub-group analysis. Although the study sample will not be specifically powered on this basis (see Power calculation and statistical analysis, below), the investigators hope that a sub-group analysis will nonetheless allow any differences between the groups to be detected.

All follow-up assessments will take place during outpatient clinic visits at 2 weeks, 6 weeks, 3 months, 6 months and 1 year. At each clinic visit, a physical examination will be performed, complications noted and the need for any further surgery recorded. Patients will be given time to complete a self-reported questionnaire, which will generate clinical outcome scores. The 1-year follow-up will enable study investigators to assess whether any differences between the operative and non-operative groups (if any) are sustained for 1 year post-intervention and to detect any important complications that may occur within the first year (e.g. nonunion).

The Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) [27] checklist is given in Additional file 1. The SPIRIT schedule of enrolment, interventions and assessments is shown in Fig. 3.

The primary outcome measure is the DASH score at 3 months post-intervention. The DASH score is a functional rating scale that was developed in 1996. It has been used extensively as a research tool in upper limb surgery [28]. It consists of a 30-item self-reported questionnaire, and two optional modules, one assessing work and the other sports and performing arts. Patient symptoms (including pain, weakness, stiffness and tingling or numbness) and functional status (including physical, social and psychological aspects) during the week prior to survey completion are assessed. The final score is from 0 to 100, with 0 representing no disability and 100 the worst possible disability. A recent study demonstrated the DASH score to be both valid and reliable in evaluating outcomes in patients following a humeral shaft fracture [29].

The null hypothesis is that there is no difference in early functional outcomes (DASH score at 3 months post-intervention) between the operative and non-operative management of patients following a humeral shaft fracture.

Secondary outcome measures will include:

Outcome measures 1 to 9 have been collated into a single data collection form (one form per study participant).

Prior to study commencement, a prospective power calculation was used to determine the number of patients required in each treatment group. The minimal clinically important difference in the DASH score is reported to be 10 points [32], and the assumed standard deviation is 12 points for both the operative and non-operative groups [33]. Thus, using a two-sided t-test with 5% significance, a total of 70 patients (35 in each group) are required to produce a 90% power to detect a meaningful difference (in terms of the DASH score at 3 months post-intervention) between the two groups, assuming a 10% loss to follow-up.

Data will be analysed on an intention-to-treat basis. Data analysis will be performed by a senior statistician, employed by the local clinical research facility and independent of the study team. The primary outcome measure (DASH score at 3 months post-intervention) will be compared between the two treatment arms, using a two-sided independent samples t-test or non-parametric equivalent (as indicated by the normality of the data). This method will also be used to compare the other continuous outcome measures between the two treatment arms (e.g. pain score, satisfaction scores, shoulder and elbow range of motion, EQ-5D score and SF-12 score). The pattern of change in continuous outcome measures over the study period will be presented graphically, broken down by treatment allocation. A linear regression model will be generated for each participant by plotting the DASH score from 6 weeks to 1 year post-intervention and the gradient of the regression line (i.e. the rate of change in the score) will be compared between treatment groups. Multivariate linear regression, including analysis of covariance (ANCOVA), will also be used to adjust for the potential effect of pre-injury DASH score, age and other prognostic factors affecting the DASH score. Binary outcomes (e.g. treatment complications, re-operation and nonunion or malunion) will be compared between the two treatment arms using a binomial test for the comparison of proportions. Two-tailed p values will be reported where possible, and statistical significance will be set at p = 0.05.

The study sample size has been increased to provide for up to 10% loss to follow-up at the primary outcome timepoint (3 months post-intervention). All clinical and outcome data will be collected in the outpatient clinic, where clinical reviews and radiographs (where required) are also performed. These appointments reflect routine follow-ups of a humeral shaft fracture in the study centre, and the investigators anticipate this arrangement will limit missing data considerably. The participant dropout rate for each treatment arm will be reported and compared. If the participant dropout rate at the primary outcome timepoint is substantial, a sensitivity analysis may be performed. Missing data at other timepoints will be accounted for by plotting a line of best fit for the data points for each participant and comparing the gradient of the regression line (i.e. the rate of change in the variable of interest) between treatment groups.

Direct costs of the injury and its treatment will be assessed, including the costs for any of the following:

Secondary costs will also be assessed, including:

These data will allow an analysis of the overall economic impact of humeral shaft fractures and a comparison of the cost implications of each treatment strategy. A cost per quality-adjusted life year (QALY) gained will be calculated (total cost difference between treatment groups / difference in QALYs between treatment groups). QALYs will be calculated at 1 year, using the difference between the EQ-5D scores at the follow-up intervals.

Patient and public involvement was not sought for this small, single-centre trial.

The maximum possible radiation dose for study participants is equivalent to 2 weeks of the dose to the UK population from natural sources of background radiation. The risk of fatal cancer arising from this level of exposure is negligible, at 1 in 230,000 [34]. This proposed radiation exposure has been approved by a medical physics expert and a consultant radiologist.

Complication rates will be monitored by study investigators throughout the trial, but a formal data monitoring committee will not be convened. Both treatment options (operative and non-operative) are regularly employed in the study centre, and any additional risk to patient safety is low. The study sponsor will be regularly updated with study progress and any patient safety issues that may arise, including adverse or serious adverse events. The trial steering committee will be consulted if required. The procedure for identifying, recording and reporting adverse events and urgent safety measures will adhere to relevant guidelines from the study sponsor, which will also advise upon any action required to protect patient safety.

Data collection forms will be kept in a secure filing cabinet in the research office on the study site, both of which are locked when not in use by the study investigators. These personal data will then be entered and stored electronically, under a password-protected file on a password-protected NHS computer system and will be accessible only by the study investigators. Once data collection is complete and analysis begins, non-personal data (stored using unique participant numbers rather than names or other identifiable information) will be made available to other authorised members of the research team. All electronic data will be handled according to study sponsor guidelines on data protection and confidentiality.

When finalised, study results will be disseminated via an internal report, conference presentations and peer-reviewed scientific journals. A summary of the results will be made available to study participants upon request.

This manuscript is based on HU-FIX study protocol version 1.3 (29 August 2018). The first patient was enrolled on 19 September 2018. The investigators estimate enrolment will be complete by July 2020, with the study follow-up complete by July 2021.