Pneumatic Tourniquet Versus no Tourniquet in Transfemoral Amputation

Hypothesis The hypothesis is that the use of tourniquet in Transfemoral amputation (TFA) surgery will reduce the total blood loss, the intraoperative blood loss and the need for blood transfusion post-operatively and thereby improve patient safety.

Aim Primary to investigate total blood loss and secondary to investigate differences in complications after TFA surgery between patients operated with use of pneumatic tourniquet versus no tourniquet.

Trial design overview Design: Randomized, prospective, blinded, two-arm, single center trial. Superiority-design. Group 1: Transfemoral amputation performed with tourniquet application Group 2: Transfemoral amputation performed without tourniquet application Population: Patients with indication for primary transfemoral amputation (intact femoral bone). Period: Expected trial inclusion period is October 2022- September 2024. Place: Odense University Hospital (OUH), Odense, Denmark Sample size: 62 in each group, a total of 124 participants inclusive a 20 % buffer.

The flow of TFA patients in the department is approximately 100 patients annually and it will therefore be possible to include the calculated sample size in the time frame available for this trial.

Tourniquet: The tourniquet that is used at OUH is FastAir® distributed and imported by Universal Medical Supplies Aps. Universal Medical Supplies Aps is not a part of this protocol.

This trial is not conducted in collaboration with Universal Medical Supplies Aps. Universal Medical supplies Aps has no right to unpublished study data and is neither financially nor personally involved in this trial. None in the study group is related to Universal Medical Supplies ApS. MDR classification for FastAir® tourniquet: class I, non-invasive device for temporary use.

The FastAir® tourniquet has a CE-certificate. The FastAir® tourniquet will be used according to instructions for use. Tourniquet application for transfemoral amputation at OUH is not standard of care.

Randomizing and database Randomization is performed internet-based using REDCap Randomize, allocation 1:1. The randomization itself takes place in the period 4 weeks prior to surgery to immediately before surgery. The randomization is performed as a block randomization and will be stratified for age (>70 years / ≤70 years) and hemoglobin value pre surgery (>6mmol/L / ≤6 mmol/L). Two stratifies is acceptable for the calculated sample size.

Blinding Participants and staff not attendant in the room will be blinded. The use of tourniquet will not be visible in the patient records but recorded directly in the REDCap Database. A standard phrase to describe the surgical procedure will be used in the patient record. Deviations and adverse events will be described. Participants will be able to be informed by the procedure after 6 months.

Feasibility In a pilot study, feasibility of tourniquet application was assessed. The pilot-study was carried out in the department in May-July 2021. Placement of tourniquet for TFA surgery is generally possible in >90%. The blood tests are not always drawn at third day post surgery, but this will be assessed in the new department algorithm for postoperative care for major lower extremity amputation patients, to ensure correct data collection in this study. The mean application time of the tourniquet was 12,15 minutes (5 minutes -21 minutes).

Sample Size From the pilot series there was observed a mean of 429 ml and standard deviation (SD) of 199 ml in 11 intervention procedures (with tourniquet) and a mean of 730 ml and SD of 446 ml in 12 control procedures (without tourniquet). The calculated blood loss in each group was normally distributed according to quantile-quantile plots (Appendix 1). It is assumed that a mean difference of 200 ml would be lower than any reasonable clinically important difference, and hence chose this as the difference for our sample size calculation.

From this,there was calculated the need of 49 participants in each group for a two-sample t-test fora superiority trial. To take into account for up to 20% drop-out (e.g. due to invalid data/protocol violation/intraoperative mortality/participant wish/investigator indication), it was decided to include 62 participants in each group (49/0,8=61,25). The sample size calculation was performed in Stata/IC 16.1with help from OPEN Statistics,OUH.

Screening log is performed on patients not included, to examine the cause of non-inclusion.

Patients that are not included in this study, no matter of cause, will be offered to participate in a concurrently prospective non-interventional cohort study where the main outcome is health related quality of life before and after surgery. These patients will only receive electronic surveys (in E-boks) regarding quality of life (EQ-5D), disease course, pain, mobility and daily life. This will serve as drop-out analysis. The same baseline data will be recorded on these patients. They will receive the department standard treatment.

Statistical analysis plan The investigators will report descriptive statistics of patient characteristic as mean and standard deviation (SD) for normally distributed, median and interquartile range (IQR) for non-normally distributed, and counts and proportions for categorical characteristics. The investigators will compare the primary outcome (calculated blood loss) by two-sample t-test with unequal variance and report mean and SD for each group as well as the mean difference with a 95% normal confidence interval (CI). The investigators will report the dichotomous secondary and exploratory outcomes as counts and proportions, compare these with chi-squared test (or Fisher's exact test, if any counts are below five) and report odds ratios with 95% CI from crude logistic regression. The investigators will report numerical secondary and exploratory outcomes either as means and SD and compare those by two-sample t-test, if deemed normally distributed by quantile-quantile plots, or otherwise, as medians and IQR and compared by Wilcoxon rank-sum test if not deemed normally distributed. All analyses will be carried out as a superiority design and as intention to treat analysis is performed for all included participants, and p-values below 0.05 will be considered statistically significant.