Trial design

This is a multi-centre, assessor-blind randomised controlled trial with a superiority comparison. The primary outcome is assessed blindly.

Study setting

This trial will take place in NHS hospitals and private clinics delivering endovenous varicose vein procedures under local anaesthesia. Recruitment centres will need to have a pre-existing practice prior to the trial to prevent any learning curve effects.

Eligibility criteria

Inclusion criteria are: adult patients (>18 years) scheduled to undergo endovenous intervention of truncal varicose veins under local anaesthesia. Treatment technologies include radiofrequency, laser, mechanochemical, foam sclerotherapy and cyanoacrylate glue.

Exclusion criteria are: clinical indication for therapeutic anticoagulation (eg, atrial fibrillation, previous personal or first-degree relative history of VTE, thrombophilia). Also female patients of childbearing potential with a positive pregnancy test, those with a history of allergy to heparins or DOACs, a history of heparin-induced thrombocytopaenia in the last year, inherited and acquired bleeding disorders, evidence of active bleeding, concomitant major health problems such as active cancer and chronic renal and/or liver impairment, known thrombocytopaenia (platelets known to be less than 50×10⁹/L), surgery or major trauma in the previous 90 days, recent ischaemic stroke in the previous 90 days, inability to provide consent.

Interventions

There are currently three thromboprophylaxis strategies used across the UK, with the trial arms of this application mirroring these practices.23 Intervention arms will consist of a single prophylactic dose of LMWH (eg, dalteparin sodium, tinzaparin sodium, enoxaparin sodium) plus compression as per local practice (eg, stockings, bandages, wraps, pads) and a single dose of LMWH plus extended thromboprophylaxis with LMWH or a DOAC (eg, rivaroxaban, apixaban, dabigatran etexilate) plus compression. The choice between LMWH and DOAC for the extended thromboprophylaxis arm will be site-specific and dependent on local practice. The duration of this must be at least 7 days, but can be in line with local practice, that is, between 7 and 14 days in duration. The control arm will consist of compression (bandages, stockings, wraps or padding) as per local practice alone. The participant flow in the trial is displayed in figure 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

THRIVE study flow chart. DHRA, Department of Health Risk Assessment; DOAC, direct-acting oral anticoagulant; DVT, deep vein thrombosis; ICER, incremental cost-effectiveness ratio; LMWH, low-molecular-weight heparin; VTE, venous thromboembolism.

Primary outcome

Imaging-confirmed lower limb DVT with or without symptoms, or PE with symptoms within 90 days of varicose vein treatment.

Secondary outcomes

The secondary outcomes include:

1. Individual components of the composite outcome

2. Comparisons of quality of life at 7 and 90 days post-procedure using the EQ-5D

3. Mortality rates in each group

4. Cost-effectiveness of providing pharmacological thromboprophylaxis

5. Exploratory analyses to assess how well the DHRA tool and Caprini score predict outcome

Safety outcomes

Safety monitoring includes any bleeding event. Major bleeding is defined as per the International Society on Thrombosis and Haemostasis standardised definition,27 which includes:

1. Bleeding into a critical organ

2. Bleeding into a surgical site requiring re-operation

3. Bleeding that leads to presentation to acute service

Sample size and study duration

The most comprehensive evidence from 52 studies suggests that the rate of VTE (encompassing DVT, PE, EHIT) post-endovenous great saphenous vein interventions is 1.7%.1 However, this analysis did not investigate treatment effects. To address this, we conducted an analysis of 229 study arms, pooling data from 480 581 participants undergoing endovenous interventions. Our approach involved subgroup analyses, comparing those receiving pharmacological thromboprophylaxis and compression versus compression alone, and distinguishing between asymptomatic screen-detected and symptomatic VTE.24

For individuals receiving pharmacological thromboprophylaxis (either single dose of LMWH or extended duration), rates of DVT, PE and EHIT 3–4 were 0.521%, 0.216% and 0.354%, respectively, resulting in a maximum summary VTE rate of 1.09%. Those receiving compression alone showed rates of 2.264% for DVT, 0.058% for PE and 0.878% for EHIT 3–4, resulting in a maximum summary VTE rate of 3.20%. Rates of EHIT and DVT are distinct in this calculation, thus the impact of double counting contributing events will be negligible. In the subset data from RCT arms alone, rates of DVT were similar. When interpreting these figures in the context of the wider literature, balancing this with confounding by indication and acknowledging that VTE rate will likely be lower in the compression alone arm due to the exclusion of individuals at the highest risk of VTE, we anticipate that the true value lies nearer to 1.0% in the pharmacological thromboprophylaxis arm and 2.7% in the compression alone arm.

At 90% power and 2.5% alpha (to approximately control overall alpha to 5%, with two active drugs being compared with a common control), the study could detect a significant change of 1.7% in 90-day VTE. This base case would require 1554 participants per group. Allowing for 10% crossover (which can only be control patients, compression only, receiving pharmacological thromboprophylaxis, since everyone receives compression) increases this to 1919 per group. Allowing for a single interim analysis at half time (50% randomised with 90-day follow-up) for early stopping, for either futility or overwhelming evidence of efficacy analysis, inflates this to 1998 per group, under a two-sided, asymmetric group sequential design implementing the non-binding Hwang-Shih-DeCani spending function (lower bound (futility) with gamma-2; upper bound (efficacy) with gamma= −4).28 If we then allow 10% for loss to follow-up, the total sample, randomised 1:1:1 between all groups, becomes 6660 (or 2220 per group).

We will more accurately estimate the required sample size by simulation in a sample size re-estimation step at around 20% mature data, using Dunnett’s three-arm design with a common control, with correction, but for simplicity retaining the 1:1:1 equal randomisation and inputting the observed missing data proportion at that stage.

Interim analysis

There will be a formal interim analysis with the possibility of stopping early for futility (no prospect of a clinically meaningful treatment effect), at the point of 50% mature primary outcome data. Full details of the stopping boundaries and analysis will be detailed in a Statistical Analysis Plan (SAP), which the independent Data Monitoring Committee (iDMC) will approve prior to seeing unblinded data. The unblinded statistician will have no other role in the study while it is ongoing. The stopping rules are statistically non-binding. The iDMC may recommend early stopping of the study if the boundaries are crossed. They would make a recommendation to the independent Trial Steering Committee (TSC) who may or may not endorse that recommendation. The trial may stop at any time for safety if there is an excess of events in the intervention groups that is considered to generate avoidable harm. This decision would not be based on any statistical criterion and would be taken by the iDMC, then endorsed by the TSC.

Recruitment and randomisation

Adults scheduled to undergo SET will be pre-screened by a member of the direct care team and invited to discuss the trial with a member of the research team. Informed consent will be obtained from each participant prior to participation in the trial, following adequate explanation of the aims, methods, anticipated benefits and potential hazards of the trial. The individual obtaining consent will be a registered healthcare professional and will have been delegated this duty by the principal investigator (PI) on the delegation log. Participants may provide their consent to participate in the trial electronically or in written form. Participants will initial alongside each statement on the consent form to confirm agreement with each clause. A copy of the consent form will be provided to the participant for their records. There is no defined time frame between initial consent and the baseline visit, however, confirmation that the participant’s consent is still valid prior to randomisation will be required in the baseline case report form (CRF).

Participants (n=6660) will undergo 1:1:1 web-based randomisation to one of three thromboprophylaxis strategies prior to undergoing SET. Randomisation will be conducted by an automated system linked to the Research Electronic Data Capture (REDCap) database setup via the Study Data Centre at the Edinburgh Clinical Trials Unit (ECTU), University of Edinburgh (a fully registered UKCRC Clinical Trials Unit, registration number 15). Participants will be assigned a study identification number by REDCap in sequential numerical order. Recruitment will commence on 1 January 2024 for 27 months. The study will close on 31 December 2026.

Blinding

Clinicians and participants will be aware of their treatment allocation. Assessors, being those who perform the venous duplex ultrasound scan and those responsible for collecting follow-up data at 7- and 90 days post-procedure, will be blinded to the treatment allocation.

Follow-up periods

Participants will undergo a lower limb venous duplex ultrasound scan at 21–28 days post-intervention to identify asymptomatic DVT. This is timed to capture the peak onset of events which is at 3 weeks.29 Participants will be further followed up remotely by telephone, online or short message service at 7 and 90 days with an expected VTE capture >95%.29 Longer term follow-up may be considered with award from a subsequent project grant to assess long-term efficacy of the intervention, hence contact at a later point will be included in the consenting process.

Data collection and confidentiality

Participant data will be entered into the REDCap database by the local research teams. Source data stored at study centres will be archived locally as per local study operating procedures. Data and all appropriate documentation will be stored for a minimum of 10 years after the completion of the study, including the follow-up period. Details of procedures for CRF completion will be provided in a separate study manual. A formal data management plan will be constructed to describe the procedures involved in the data management activities and processes for the study so that it is managed and maintained in accordance with the Good Clinical Practice (GCP) guidelines, local Research and Governance Integrity Team (RGIT) standard operating procedures, appropriate regulatory requirements and the study protocol.

Statistical analyses

All statistical analyses will be governed by a comprehensive SAP, written by the study statistician and agreed by the TSC and iDMC before any unblinded data is seen. The main analysis will be according to the intention-to-treat principle where all consented participants will be included in the analysis retained in the group to which they were allocated (ie, ‘as randomised’) and for whom outcome data are available. All results will be presented as point estimates, CIs at the appropriate level and associated p values. Absolute measures of effect will be presented alongside relative measures. The primary analysis will compare the incidence of VTE at 90 days, using a repeated measures analysis of variance, adjusting for any pre-specified strongly prognostic baseline covariates using a mixed effects logistic regression. Study site will be included in the model as a random effect, and pre-specified baseline covariates strongly related to outcome will be included to adjust the estimated treatment effect. The findings will be assessed for robustness against any missing data, first using multiple imputation assuming this data is missing at random and, if appropriate and the data permits, further sensitivity analyses will be attempted under any plausible missing data mechanisms not missing at random. Secondary outcomes will be analysed in a similar fashion with generalised linear models appropriate to the distribution of the outcome. Safety data will be summarised descriptively.

Internal pilot

There will be an internal pilot to assess feasibility of recruitment over 9 months of recruitment to the trial, in which we will start recruiting from the (minimum) 40 centres. Site setup will be staggered over 9 months, that is, five centres per month. The target number of participants by the end of the 9-month internal pilot is 1450 participants. For the internal pilot, we will use stop–go criteria based on a Green–Amber–Red statistical approach (table 1).

Cost-effectiveness analysis

Two health economic analyses will be conducted, and a separate Health Economics Analysis Plan (HEAP) will be developed by the health economist detailing the proposed analyses. The main analyses will be performed from the perspective of the NHS and Personal Social Services, with secondary analyses from a societal perspective.

A within-trial analysis will compare the two pharmacological thromboprophylaxis strategies to compression alone over the 90 days of the study. Resource use items associated with treatments in primary and secondary care will be collected using case notes and self-completed patient diaries and costed using manufacturer list prices, previous literature and national reference costs. Days off work and normal activities and other patient-related costs will be collected for a secondary analysis. EQ-5D will be collected at baseline and follow-up, analysed using the NICE-approved tariff.

If the trial indicates that pharmacological thromboprophylaxis could be an effective therapy, a Markov (state-transition) decision model will be constructed to compare the cost-effectiveness of the two pharmacological thromboprophylaxis strategies and compression alone over a longer time horizon. The time horizon of the model will be 2 years, allowing extrapolation of sequelae of VTE events (such as PTS) over the longer term to quantify the impact of VTE on patient health via quality-adjusted life years (QALYs) and resource use. A preliminary model has been constructed based on published literature to identify the key variables that would need to be collected during the clinical study and to estimate the number needed to treat (NNT) to avoid one VTE, above which pharmacological thromboprophylaxis would not be considered cost-effective at NICE thresholds. This model conservatively assumes 30% of patients with VTE develop PTS, with 3% of those having severe PTS.

The minimum cost of purchasing 10 days of thromboprophylaxis and providing allocated time for administration training equates to ~£63.13.30 31 Our model assumes cost of treatment of VTE, non-severe and severe PTS as £451, £872 and £1547, respectively, and estimates of the respective utility decrement associated with symptomatic VTE and PTS are 0.8628, 0.7745 and 0.6752, respectively.32 Different durations of thromboprophylaxis will be modelled. These estimates will be reviewed at the time of the cost-effectiveness analysis, and any changes to these estimates will be updated.

Using a 2-year time horizon, incremental cost-effectiveness ratio of pharmacological thromboprophylaxis in comparison to no pharmacological therapy would be £13 339 per QALY if the NNT were 59 participants (1/0.017). For pharmacological thromboprophylaxis to be cost-effective at an NICE willingness to pay threshold of £20 000 per QALY, the NNT would need to be below 80.

Main analyses will be undertaken from perspective of the NHS and Personal Social Services, with secondary analyses from a societal perspective. Health economic analysis will be conducted according to NICE reference case and Consolidated Health Economic Evaluation Reporting Standards (CHEERS) guidelines, including sensitivity analyses and probabilistic sensitivity analyses.33 34 Results will be presented as estimates of mean incremental costs, effects and incremental cost per QALY.

Data monitoring, safety and quality control

An independent TSC and iDMC have been convened. The role of the TSC is to provide overall supervision of trial conduct and progress and ensure the study adheres to GCP principles. The role of the iDMC is to oversee the safety of the trial participants, and the iDMC will be the only oversight committee that sees unblinded data as the trial progresses, which they will keep in strict confidence. Details of membership, responsibilities and frequency of meetings for the TSC and iDMC have been defined in separate charters.

The Trial Management Group will oversee trial progress. All adverse events (AEs) will be collected and recorded on the REDCap database. Causality and relatedness will be assigned by the local PI or other appropriately trained and delegated individual. Research sites will inform the coordinating centre of all serious adverse events (SAEs) within 24 hours of knowledge of the event.

A study-specific risk assessment has been prepared in preparation for the study by the trial manager and study sponsor, which will be updated as required during the course of the trial. The frequency, type and intensity of monitoring visits will be detailed in a separate Data Monitoring Plan (DMP). The DMP will also detail the procedures for completion and sign-off of monitoring reports. In the event of a request for a trial site inspection by any regulatory authority, the study coordinating team must be notified as soon as possible. Participating investigators must agree to allow trial-related monitoring, including audits, Research Ethics Committee (REC) review and regulatory inspections, by providing access to source data and other trial-related documentation as required. Participant consent for this must be obtained as part of the informed consent process for the trial. The trial coordinating centre will centrally review eCRF data for errors and missing key data points on an ongoing basis.

Quality control will be performed according to Imperial College internal procedures. The study may be audited by a quality assurance representative of the sponsor. All necessary data and documents will be made available for inspection. The study may be subject to inspection and audit by regulatory bodies to ensure adherence to GCP and the UK Policy Framework for Health and Social Care Research.

Patient and public involvement

Regular sustained patient and public involvement (PPI) has been undertaken to design and guide the study. Successful focus groups and interviews were held with patient representatives longitudinally throughout the trial development to gain insights into patients’ lived experiences and gather feedback on potential trial designs. Online surveys were also conducted with the wider venous community to inform the plain English summary. The patient-facing documentation has been designed in collaboration with three patient representatives. Two patient advisers have agreed to sit on the TSC, as well as review updated patient-facing documentation and assist with the dissemination of study results. An online survey of key stakeholders was also undertaken at the UK’s largest national conference to ascertain acceptability among clinicians of the proposed trial design.