Surgical Versus Non-Surgical Treatment of Humeral Shaft Fractures Compared By a Patient Reported Outcome: The Scandinavian Humeral Diaphyseal Fracture Trial (SHAFT) - A Study Protocol For a Pragmatic Randomized Clinical Trial

Background: The outcome of non-surgical treatment is generally good, but the treatment course can be long and painful with approximately a quarter of the patients acquiring a nonunion. Both surgical and non-surgical treatment can have disabling consequences such as nerve injury, infection and nonunion. The purpose of the study is to compare patient reported outcome after surgical and non-surgical treatment for humeral shaft fractures. Methods: A pragmatic randomized controlled trial is planned with two study groups (SHAFT-Y and SHAFT-E). A total of 287 eligible acute humeral shaft fractures are scheduled to be recruited and randomly allocated to surgical or non-surgical treatment with the option of early crossover due to delayed union. The surgical method within the allocation is decided by the surgeon. The primary outcome is the Disability of Arm, Shoulder and Hand (DASH) score at 52 weeks, and is assessor blinded. The secondary outcomes are DASH score, EQ-5D-5L, pain assessed by visual analogue score, Constant-Murley score including elbow range of motion and anchor-questions collected at all timepoints throughout the trial. All complications will be reported including; infection, nerve or vascular injury, surgical revisions (implant malpositioning, hardware failure, aseptic loosening and peri-implant fracture), major adverse cardiovascular events, and mortality. Patients declining randomization will be asked to enroll into an observational cohort with same outcomes and post-treatment regimes. Discussion: SHAFT will provide information on the effectiveness of two standard treatments for humeral shaft fractures, while taking the dilemmas within the population into account.

may suggest that the prolonged course for patients with union problems is unfavorable. However, the challenge is to identify the patients that will bene t from early xation. A way of determining early onset of union problems in a young cohort is by gently testing the fracture site at 6 weeks (14) and later for the elderly (15).
To our knowledge, three randomized controlled trials (RCT) have been completed and all trials compared plate osteosynthesis to non-surgical treatment without distinction of age (11,16,17). Four RCTs (18)(19)(20)(21) are registered in clinicaltrials.gov, WHO and ISRCTN registry. Two RCTs are comparing plate osteosynthesis to non-surgical treatment and the last two RCTs are comparing plate and nails to nonsurgical treatment. Adults of all ages are included, except in one RCT(20) that has an upper age limit at 65 years. Furthermore, in one already nished trial, the trial design had to be adjusted from an RCT to a prospective non-randomized comparative trial due to treatment allocation problems as there was a strong physician preference towards the surgical treatment option (22). A Cochrane review could not demonstrate any difference in union rate between different surgical procedures (intramedullary nail and plate osteosynthesis) (23).
This emphasizes the need for a pragmatic approach of interventions by including usual care of surgical treatment, considering the in uence of age as well as accepting early secondary surgery as a part of treatment to improve disability and function after 12 months.
The SHAFT protocol conform with the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) (24,25).

Methods: Participants, Interventions And Outcomes
Objectives To compare surgical xation of humeral shaft fracture to non-surgical treatment with early identi cation and treatment of delayed union by a patient-reported outcome after 52 weeks.
The trial population is divided in two groups by age due to the bimodal age distribution with more elderly than young and the treatment courses are different in terms of early crossover: 1. SHAFT-Y for the young with an age cut-off of 18 to 64 years. The early identi cation and treatment of delayed union is set to 6 to 12 weeks 2. SHAFT-E for the elder with an age cut-off + 65 years. The early identi cation and treatment of delayed union is set to 12 to 26 weeks Trial design A pragmatic multicenter, randomized, controlled, outcome assessor-blinded, clinical superiority trials. We have assessed the pragmatic design using PRECIS-2(26) which yielded 40 points out of possible 45 7. Open fracture 8. Floating elbow fractures 9. BMI > 40 10. Health conditions preventing either treatment Primary radial nervepalsy (RNP) is not an exclusion criterion as there has not been shown bene ts in recovery time with early exploration (28).

Recruitment
Patients admitted to the emergency department (ED) in any of the trial sites will be clinically examined and plain radiographs will be obtained to con rm the diagnosis. If the patient ful lls the eligibility criteria's, they will be informed of the trial by staff and receive written information with patient participation information and "Researchers rights in a health science research project". They will be given time to consider and will be scheduled for an appointment with research staff within 10 days. If written consent is obtained at the consultation, randomization will occur immediately after. If patients do not wish to participate in the RCT, they will be asked to consider enrollment in an observational cohort.
The informed consent gives the investigators permission to obtain information from the patients' health records including; age, sex, height, weight, arm dominance, American Society of Anesthesiologists grade (ASA classi cation), UCLA activity score (29), list of current diagnosis and medication, mechanism of injury, radiographic images, previous surgery to the arm, tobacco (cigarettes, smokeless tobacco) and alcohol habits, employment and educational status. Patients will be provided with transport allowance to cover the costs for consultations at 26 weeks and 52 weeks. They further accept telephone calls if they miss follow-up visits to promote participant retention and complete follow-up.

Interim analyses
We will carry out two interim analyses after trial completion of 25% and 50% of the planned number of included patients, separately for the two age groups. We will use the O'Brian-Fleming rule(30) for determining appropriate signi cance levels for each analysis, which results in a signi cance level of 0.0006 for the rst interim analysis, 0.0151 for the second interim analysis and 0.0471 for the nal analysis. Interim analyses will be carried out for both the primary endpoint (DASH) and for serious complications (iatrogenic nerve injury, deep infection, major adverse cardiac events (MACE) and death) and will be conducted by one-sided tests to determine if the improvement in DASH is signi cantly above 25 points, respectively if the complication rate is at least 20% higher.
A Data Monitoring Committee (DMC) has been organized to monitor and evaluate the data from the interim analyses. The DMC consist of seven members in total; three members of the steering committee and four independent members including the Patient Representative (PR), an orthopedic researcher, an orthopedic surgeon and a nurse.
In the event of one group having zero complications, the signi cance level cannot be computed by a Chisquare test, and the DMC will have to assess the data due to a statistical variation.
In the event of a patient sustaining several complications, the most severe complication will follow the patient. The complications are ranked in a hierarchy model describing the severity and is ranked from most severe to least severe: Death, MACE, deep infection with debridement, iatrogenic nerve injury.
The interim analyses will be conducted by a biostatistician, blinded to the treatment allocation. Data will then be presented to the DMC, in means and proportions and be accompanied with con dence intervals.
The DMC will have access to the data unblinded, and if at least one of those two conditions are ful lled or a statistical variation has occurred at an interim analysis the DMC will be asked to investigate the results in detail and present their recommendation to continue or stop the trial. In the event of disagreement, the steering committee will be involved in discussing the stopping of the trial.
Furthermore, the steering committee will be monitoring recruitment and drop-out. Any modi cation in design and recruitment will be registered in clinicaltrials.gov.
If it is observed that inclusion of patients in the two age groups is too slow, it can be decided to pool the age groups and analyze them as one total group instead. If this decision is made before the rst interim analysis, above signi cance levels will be used for the total group. If the decision is made between the rst and second interim analysis, O'Brian-Fleming signi cance levels for four analyses will be used, resulting in a signi cance level of 0.0184 for the second interim analysis and 0.0412 for the nal analysis. If the decision is made after the second interim analysis, a signi cance level of 0.0417, corresponding to O'Brian-Fleming with ve analyses, will be used for the nal analysis.

Patient and public involvement
Patients were involved in the planning and development of the study protocol. A series of in-depth semistructured interviews were conducted with the qualitative purpose of exploring the experiences acquired during the treatment course of a humeral shaft fracture. Beyond the qualitative study aim, a discussion of the most meaningful primary outcome measures was undertaken. All patients completed two questionnaires (DASH and QuickDASH) to solicit their feedback and to determine which PROM they found most appropriate when considering relevance, comprehensibility, comprehensiveness and length. The interviews further revealed complaints that were discussed with the patients and were subsequently priorities to be implemented as outcome measures of the trial.
One of the patients from the interviews accepted to be involved in the trial as the patient representative (PR). The PR is included in the steering committee and in the data monitoring committee (DMC). The study protocol was discussed with the PR in layman terms to facilitate a common understanding of the trial and to solicit feedback that could minimize patient burden and risk of missing data, as well as providing insights from a patient perspective to optimize the communication between physicians and patients during the trial course and by written information. The feedback resulted in minor revisions of words in the patient information sheet to conform with patient concerns.

Interventions
Treatment will be performed within 14 days after injury. Eligible patients will be randomly allocated to one of two treatment options.

Surgical treatment
Non-surgical treatment with the option of early crossover surgery at 6-12 weeks (14) for SHAFT-Y and at 12-26 weeks for SHAFT-E We anticipate that surgical treatment will include plate osteosynthesis (MIPO and ORIF), intramedullary nailing (antegrade and retrograde) and external xation. Plate and nail types, screw con gurations and surgical approaches will be decided by the surgeons. The procedure will be conducted or supervised by a senior consultant.
Non-surgical treatment will include sugar tong splint, plaster splints, hanging casts, or functional bracing such as the Sarmiento brace, and will be worn until a surgeon removes it.
All patients will be advised to follow this rehabilitation protocol.

Criterias for early crossover
Patients can be offered to undergo early crossover xation with a surgical procedure of the surgeon's choice, if one of these criteria are met: Unacceptable pain experienced by the patient Severe pain with gross instability of the fracture site assessed by: Unable to en bloc elevate the arm due to clear fracture instability Gentle manipulation of the fracture site. Gentle manipulation should respect the risk of callus breakage Severe problems tolerating the brace, e.g. discomfort, skin irritation, wounds, hygiene problems.
The patients that undergo early crossover surgery will have the reason for crossover thoroughly noted. We anticipate the surgical procedures will be similar to the ones previously mentioned with the possible addition of bone graft.

Randomization
A computerized database software, Research Electronic Data Capture (REDCap)©(31) will be used to generate an irreversible random allocation sequence and perform block randomization with selected block sized of 2 and 4, which will be strati ed on site and age (18-64 and + 65). Patients will be assigned to the trial with an allocation of 1:1 to either surgical treatment or non-surgical treatment. The trial worker acquires the allocated treatment from the central coordinator with randomization rights to REDCap. The trial worker then initiates the treatment, either by scheduling the surgery date or applying the chosen nonsurgical method.

Protocol violation
Lost to follow-up Treatment crossover outside the pre-de ned interval for early crossover surgery Patients that meet any of these criteria will remain in the study and be included in the intention-to-treat analysis but omitted from a per-protocol analysis.

Participant withdrawal
If the patient withdraws the consent, patients will be included in the statistical analysis through multiple imputation, if baseline data is obtained, otherwise imputation is not possible, and the patient will be replaced to meet the calculated sample size.

Blinding
The trial will consist of several levels of blinding: The primary outcome will be blinded to everyone involved in the trial, apart from the patients and a central trial worker (non-physician), who will only review the questionnaire for completion in REDCap The statistical analysis will be conducted by a blinded biostatistician

Primary outcome measure
The primary outcome is the Disability of Arm, Shoulder and Hand (DASH) score at 52 weeks(32). The DASH score is a 30-item self-reporting patient-reported outcome measure speci c for physical function and symptoms of the upper limb. Scores range from 0 (no disability) to 100 (most severe disability). The DASH score is validated for the target population(33) and has undergone cross-cultural adaptation in Danish, Swedish and Norwegian(34). Complications after treatment will be recorded and include local complications, early general complication and mortality: Local complications: Infection (needing antibiotic treatment with or without debridement), nerve or vascular injury, surgical revision (due to implant malpositioning, hardware failure, aseptic loosening or peri-implant fracture) and tolerance problems with brace (discomfort resulting in non-compliance of wearing the brace Early general complications needing hospitalization within 12 weeks(38,39): Major adverse cardiac events (MACE) including myocardial infarction, heart failure, thromboembolism, cardiomyopathy and cardiac arrythmias.
Other major adverse events including pneumonia, renal failure, electrolyte abnormality and de ciency anemia and other unforeseen reasons for hospitalization. The DASH, EQ-5D-5L and CA questions will be sent by email or mail and patients are asked to complete the questionnaires before visits. The questionnaires will be reviewed for missing data by a trial worker and patients will be assisted with completing the questionnaires if any data is missing, without interference from the medical staff. The trial worker collects the questionnaires before consultation with the physician. The physician then collects the additional outcome measures at the consultation and enters data directly into REDCap. Gross instability of the fracture site is assessed by a physician through gentle manipulation. This is rst tested at 6 weeks follow-up for SHAFT-Y and 12 weeks for SHAFT-E Return to work will be assessed by direct questioning at clinical visits. Five answers will be available: Unemployed, not returned, partially returned, fully returned and retired. If patients are still on sick leave they have 'not returned'. If they are back at work but not in full capacity as prior to the fracture, they are 'partially returned'; if they have returned as prior to the fracture they are 'fully returned'; and if the patients are a pensioner, they are "retired" Mechanism of injury measured as low-energy fractures e.g. fall from own height and high-energy fractures e.g. fall from heights, sports, motor vehicle accidents and other miscellaneous accidents Nonunion is de ned as gross mobility at the fracture site after 12 weeks for SHAFT-Y and 26 weeks for SHAFT-E. Pain from the fracture site will also be noted to determine if the nonunion is symptomatic or asymptomatic. Radiological assessment of type of nonunion will also be assessed

QALY-based cost analysis
Successful primary treatment vs early crossover surgery

Length of rehabilitation
Level of activity by the UCLA activity score (42) Other outcome measures Any ancillary outcome measure or analysis will be reported.

Hypothesis
The null-hypothesis is: The DASH score at 52 weeks after surgical treatment is not superior to non-surgical treatment with the option of early crossover surgery in patients with humeral shaft fractures Sample size The two groups (SHAFT-Y and SHAFT-E) require individual sample size calculations due to our presumption that these groups cannot be compared in DASH scores. Based on the following assumptions, the total sample size is estimated to 163 patients for SHAFT-Y and 124 patients for SHAFT-E. The calculations are powered to detect a minimal important change (MIC) of 7 points in the young and 10 points in the elderly group in DASH, respectively. The MIC is determined on a population with the average age of 59 years (35). Our SHAFT-E will have a considerably higher average age; thus we assume a higher MIC is needed for this group. Two standard deviations were obtained from the data of the FISH trial(11) and were separated in age groups of 18-64 years and 65 years and above. The standard deviations 14.91 and 18.59, respectively. We used an α-level of 0.05, a power of 0.80 and the allowance of attrition is set to 15%.
Patients will be included consecutively to an observational group until the sample size for the RCT is completed. The trial management committee can decide to pool data from both RCTs (SHAFT-Y, SHAFT-E) if recruitment is prolonged.

Statistical methods
The data will be analyzed using computerized statistical software and all data will be entered into REDCap.

Primary analysis
Descriptive statistics will be used to report demographic data. Demographic data and outcome measures will be tested visually and statistically (i.e. Shapiro Wilks test). Numeric variables will be summarized by means, standard deviations and 95% con dence intervals (95% CI). Median and interquartile ranges will be used when normal distribution is not met. Categorical variables will be summarized by frequency and proportion. For group comparison with numerical data, a student's t-test will be used if data is normally distributed, otherwise a non-parametric test will be used. For categorical data a Chi-square test will be used for group comparison. An intention-to-treat (ITT) analysis of the primary outcome will be conducted by univariable linear regression, including all patients that do not meet the withdrawal criteria and will be conducted to minimize bias within results. A sensitivity analysis will test the effects of non-adherence to protocol by conducting a per-protocol analysis and includes only patients who comply with the protocol. For missing data points in an outcome measure, a multiple imputation analysis using predictive covariates (sex, smoking, alcohol, UCLA activity, ASA score)(43-45) will be conducted to deal with nonresponse bias. For comparison we will carry out a sensitivity analysis excluding all the missing values.
Data will be considered statistically signi cant if p-values < 0.0471.

Secondary analysis
In order to re-ensure the validity of data a linear regression analysis will be computed with DASH score as the dependent variable and treatment modality as the independent variable. Additional regression analysis will be carried out between the early crossover group and the primary treatments. A multivariate regression analysis will be conducted to adjust for potential confounders. Variables adjusted for are: sex, smoking, alcohol, UCLA activity, ASA score. Data will be summarized as coe cients with 95% CIs and variance will be summarized as r-squares, adjusted r-squares, predicted r-squares, standard errors.
Coe cients will be considered statistically signi cant if p-values < 0.05.

Observational cohort
An observational cohort will be established and followed simultaneously with the RCT. Choice of treatment will be based on local decisions. Outcome measures, data collection and statistical plan will be similar. Follow-up can vary depending on each case, but all patients will be invited to follow the same timeline as the RCT.
The observational cohort will be subdivided into two groups of populations when statistical analysis will be conducted.
1) Patients who ful l all eligibility criteria but decline randomization 2) Patients who ful ll all inclusion criteria but presents one or more exclusion criteria

Discussion
High level evidence on treatment of humeral shaft fracture treatment is sparse, but several RCT's are planned or ongoing (18)(19)(20)(21). The current discussion on surgical treatment versus non-surgical treatment have been ongoing for more than a decade (3). The discussion has been intensi ed with recent RCTs showing surprisingly high nonunion rates in the non-surgical groups (11,46). The orthopedic research community has demanded the need of more high quality trials (47,48). Thus it is our aim that SHAFT will contribute to the increasing quality of evidence published for the decision-making of treatment of humeral shaft fractures by conducting a pragmatic, two-arm, multicenter, superiority, randomized controlled trial.
The trial is planned pragmatic by introducing real-life treatment courses, to comply with the trial objective of comparing the effectiveness of treatments and furthermore increase external validity.
The distribution of the study population is bimodal, but with a steep increase in incidences from the fth decade (2,49). Our study group were humble recipients of demographic data from the FISH trial (11), which paradoxically, showed a clear difference in the number of recruited patients in the age-groups of 18-64 years versus + 65 years, favoring the "young" group. To protect the trial against imbalance, age is de ned as a strati ed variable to ensure equal distribution of age (18-64 and + 65) in the treatment groups. Furthermore, normative data from the general population suggests inferior DASH scores with increasing age and suggests that DASH scores of age-groups should not be compared (50,51). We further expect that the younger population can discriminate a smaller change in DASH than the elderly. The available data on MIC of DASH scores for humeral shaft fractures do not differentiate in age (35). Consequently, we have chosen to use two MICs, thereby having to recruit 163 young adults (SHAFT-Y) and 124 elderly (SHAFT-E) and to analyze data independently for each group.
Recruitment of 287 humeral shaft fractures present a challenge in a randomized setting as the annual incidence rate is 14.5 per 100.000 (49). To overcome this, we will be recruiting from multiple sites across 3 Scandinavian countries with comparable health-care systems and with cultural and demographic similarities. Despite these overall similarities, surgical management can differ from site to site. To prevent against bias in the surgical group, the randomization will be strati ed by study site. One drawback of this type of strati cation can be the risk of open blocks in multiple sites, which can give rise to allocation bias.
To limit this the randomization will be conducted in small permuted blocks.
One RCT demonstrated a statistical signi cant difference in a subgroup analysis between late surgical crossover and primary surgical treatment after 12 months, favoring early surgery, with respect to possible confounding (11). The research group has furthermore in a recent study showed same difference between late surgical crossover and primary successful treatment after 2 years (13). To avoid prolonged failed treatment courses, and to minimize overall nonunion rates, the trial has implemented an option of early crossover of non-surgical treatment, if certain criteria for delayed union are met. The criteria for early crossover surgery are identical for each age-group but differ regarding the time of crossover. The term "early crossover" and "delayed union" were discussed intensively by the study group. "When has the treatment failed?" Crossover to surgical treatment after non-surgical treatment is not uncommonly performed within 12 weeks for the young and described due to lack of healing and early non-surgical failure (52,53). One can argue that a portion of these patients will unite without intervention (54), but multiple studies have currently shown association between gross mobility from the fracture site to nonunion in a young population after 6 weeks (14,43,54). For the elderly, our clinical experience is that healing rst can be anticipated after 12 weeks. We therefore de ned delayed union in SHAFT-Y between 6 and 12 weeks and SHAFT-E between 12 and 26 weeks. Furthermore endpoints were established to create a clear timepoint of when a treatment has failed, which we believe the orthopedic society needs, since there is no consensus in nonunion for humeral shaft fractures (7,10,12,55).
The DASH questionnaire is adopted by the study group as primary outcome as it is the best psychometrically validated and cross-culturally adapted PROM for humeral shaft fractures (35,56). The full DASH questionnaire is preferred over the Quick-DASH by patients through simple semi-structured interviews. The interviews were simple in a sense that the objective was to determine the "most preferred" questionnaire rather than asses the content validity of each questionnaires (57). The moderator was instructed in the conduct of qualitative interviews by an expert in qualitative research, but not proper trained. This could give challenges with bias, if the moderator preferred one questionnaire. This was sought to be reduced by the semi-structured guide, which was intended to make patients re ect, rstly on their treatment course and secondly on the questionnaires and its relevance, comprehensibility and comprehensiveness, to induce a thought-through answer to the question: "Which questionnaire do you prefer?".
Through the interviews, it was furthermore recognized that around half of the patients failed to complete one or more questions, but none more than three, which is the cut-off for the score-calculation to be valid (32). To provide a margin of security for missing data, an unveiling trial worker will review the questionnaire for missing answers. Finally, to prevent performance bias, patients will complete the DASH questionnaire prior to visits with the physician, thereby blinding the physician.
The SHAFT trial is a pragmatic multicenter RCT, that will assess if there are a difference between surgical treatment versus non-surgical treatment in humeral shaft fractures, while taking the dilemmas within the population into account.

Trial Status
Protocol version 2.
Issue date: 30th of August 2021, Author DK.
Recruitment is anticipated to start on 1/1 2022 and is expected to be completed on 1/1 2024. If the sample size of 287 patients have not been meet, the extension of the study period will be discussed by the steering committee. participants. The management of personal data has been approved by the Data Protection Agency (reference number 20/60034). Any important protocol modi cation will be advertised to The Regional Committee on Health Research Ethics for East Denmark, and registered at ClinicalTrials.gov. Participants are covered by the insurance policies that apply to ordinary treatment. Separate ethical applications are sent and pending in Norway and Sweden. All important changes to the trial will be reported to the scienti c ethics committee and be registered in clinicaltrials.gov.

Consent for publication
Not applicable.

Availability of data and materials
Study-related information about patients are secured in REDCap (31). REDCap is a secure web application with authentication and data logging for managing databases. Access to the database will be limited to trial workers currently working on the trial. Their rights will be limited to data entry only. Principal investigator will keep a list of persons with access. All published data will be fully anonymized of patient identi ers. Furthermore, the trial has been approved by the Danish Data Protection Agency and will conform to the Act on Processing Personal Data. Trial results are planned to be published in an international peer-review journal.

Competing interests
The authors declare that they have no competing interests.

SHAFT authorship agreement
The rst authorship will be granted after the rst patient has completed the trial endpoint at 52 weeks. A recruiting site will then be granted an additional authorship for every 5 patients that completes the trial endpoint at 52 weeks. All authors are expected to provide substantially to the data collection. The authorships are granted the departments and not individuals.    Trial ow