BAlloon versus Stenting in severe Ischaemia of the Leg-3 (BASIL-3): study protocol for a randomised controlled trial

This is a clinician-initiated study. The sponsor (University of Birmingham) holds public liability (negligent harm) and clinical trial (negligent harm) insurance policies which apply to this trial. Full details of indemnity arrangements for this trial can be obtained from the sponsor.

Full inclusion/exclusion criteria are shown in Table 2. All patients with SLI referred to vascular units within participating NHS organisations may be considered for enrolment (a full list of participating centres can be obtained from the trial sponsor). Participants should have atherosclerotic FP disease and be judged by the vascular multidisciplinary team (consisting of two endovascular practitioners from either vascular surgery or interventional radiology) to require early endovascular revascularisation. Following diagnostic imaging, all participants should be deemed suitable to receive any of the three trial revascularisation strategies and should also have adequate inflow to support these strategies (either at the time of randomisation or achieved as part of a ‘hybrid’ procedure in which inflow is restored concurrently with the FP revascularisation). Participants should have an anticipated life expectancy of at least 6 months and be able to complete the trial quality of life (QoL) and resource use questionnaires. Patients will be excluded from the trial if they lack capacity to provide written informed consent, have received a previous intervention to the target vessel within the past 12 months, or speak insufficient English to guarantee informed consent where translation services are not available.

Randomisation will be provided by a computer-generated programme hosted online from Birmingham Clinical Trials Unit (BCTU), using a minimisation algorithm with a random element incorporated within it to ensure balance between the three arms with regard to important clinical variables. The following minimisation variables will be used:

Randomisation will be performed between the three endovascular interventions at a ratio of 1:1:1.

All participants will undergo a baseline assessment consisting of:

The randomised trial intervention should normally take place within 14 days of randomisation where clinically and logistically practical. The randomised endovascular procedure will normally be carried out under local anaesthesia with access via the common femoral artery. Technical success of the procedure will be assessed by pulse status, completion angiography, and haemodynamic measures such as ABPI. Data collected relating to the randomised procedure will include type and number of the devices used by artery segment as well as disease severity. Standard data relating to the hospital admission or day case will be collected and include summary of the admission, discharge destination, additional interventions, and medical complications.

Further intervention is possible in all arms of the trial, even when the trial endovascular intervention has been successful (for example, due to multilevel disease or following restenosis). This may either be with the same endovascular intervention (re-intervention), one of the alternative endovascular interventions (endovascular cross-over intervention) or surgical intervention (surgical cross-over intervention), each of which may be repeated more than once.

Based on clinical experience, and data from the original BASIL trial [18], we anticipate that further intervention:

Outcomes will be recorded at 1 month post intervention and 6, 12, 24, and 36 months post randomisation. Data collected will include peripheral pulses, ABPI, TBPI (where indicated), functional status, and current drug therapy. All participants will undergo a full review of hospitalisations, further interventions (vascular, nonvascular, and amputations), and (serious) adverse events (SAEs/AEs). The QoL assessments (previously described) will be completed at each follow-up time point as well as health resource usage and personal circumstance assessments as part of the economic analysis. A schedule of assessments for all time points is given in Fig. 2.

Appropriate strategies to promote participant retention will be considered and implemented by the Trial Management Group (TMG) as required.

Outcomes will be collected by research nurses at the clinical centres on paper Case Report Forms (CRFs). All data will be handled in accordance with the UK Data Protection Act, 1998. CRFs, other than the Patient Contact Form and Consent Form, will not bear the participant’s name. For all other forms the participant’s initials, date of birth and trial number will be used for identification.

The collection and reporting of AEs and SAEs will be in accordance with Good Clinical Practice (GCP) and the Research Governance Framework 2005. Safety will be assessed continuously throughout the trial. There are no investigational medicinal products being used as part of BASIL-3 and all of the surgical procedures being tested in this trial are part of current UK ‘standard of care’; therefore, no (S)AEs are anticipated as a unique consequence of participation in BASIL-3. In addition, at regular time points, the Trial Steering Committee (TSC) and Data Monitoring Committee (DMC) will be provided with details of all SAEs.

The sample size for this trial was calculated based on a time-to-event analysis making two key comparisons between standard care and the new treatments (PBA ± BMS versus DES; and PBA ± BMS versus DCB ± BMS). To maintain an overall 5% type I error rate, each comparison will be tested at a significance level of 2.5% to account for the increase in the risk of type I error associated with making two key comparisons. The total trial duration is 5 years with 3 years’ recruitment (20% of participants are to be recruited in year 1, and 40% in years 2 and 3) and 2 years’ follow-up resulting in a mean follow-up of 3.3 years per patient.

The sample size calculation is based on estimated event rates in the PBA ± BMS arm taken from the angioplasty arm of the original BASIL trial (observed to be 0.70, 0.64, 0.52, 0.46, and 0.36 at the end of years 1–5, respectively) [18]. The study is powered at 90% to detect a hazard ratio of 0.60 for both comparisons reducing the risk of the primary outcome (AFS). Across the three arms, a total of 342 events would be required to detect a hazard ratio of 0.60 (equivalent to an absolute difference in AFS of 13% at year 2) at the 2.5% significance level. Conservatively, allowing for 5% drop out for the primary outcome (equivalent to 1% dropout in each year for 5 years) a total of 861 participants is required. It is anticipated that around 50 UK clinical centres will be required to achieve this recruitment target. Appropriate strategies to meet recruitment targets will be discussed and implemented by the TMG as required.

Recruitment to BASIL-3 will be reviewed at the end of a 12-month pilot phase and discontinuation of the trial will be considered if:

The completeness of the QoL measures will also be evaluated at the end of the pilot phase. If completion rates are poor, consideration will be given to discontinuing some of these tools.

After the first year of recruitment, we aim to assess recruitment, retention, patient burden, and completeness of data. A full efficacy and safety analysis report will be reviewed by the DMC on an annual basis or more frequently if required by the DMC or TMG. The DMC will outline and agree the stopping rules for the trial which will be documented in the DMC charter. It is likely that the Haybittle-Peto boundary will be used. This approach states that if an interim analysis of the primary outcome shows, with a p value of less than 0.001, that the treatments are different, then the trial should be stopped early. This Haybittle-Peto approach will be used as stopping guide, alongside data on important secondary endpoints and all other relevant evidence. A DMC report and charter outlining the terms of reference (including information on stopping rules) will be agreed with the DMC. The report will specify which endpoints are to be included in the reports to the TSC. The membership of the DMC is stated within the current trial protocol which is published online [19].

Monitoring of BASIL-3 will ensure compliance with GCP. A risk proportionate approach to the initiation, management and monitoring of BASIL-3 will be adopted. Clinical centres will permit trial-related monitoring, audits, and Main Research Ethics Committee review; providing direct access to source data/documents. Trial participants will be informed of this during the informed consent discussion and will consent to provide access to their medical notes for these purposes. Important changes to the trial conduct or trial protocol will be communicated to all relevant interested parties and the trial sponsor assumes responsibility for this.

The final analysis for the BASIL-3 trial will occur once the last randomised patient reaches the 24-month follow-up assessment. Differences in the primary outcome (AFS) will be assessed by comparing time from randomisation to major limb amputation or death from any cause between randomised groups, assessed up until the end of the follow-up period which will be between 24 and 60 months.

The primary unadjusted analysis will use Kaplan-Meier plots and test the difference between groups using the log-rank test. Data will be censored when individuals reach the end of follow-up or are lost to follow-up prior to reaching the primary outcome. Further analysis of the primary outcome will involve fitting flexible parametric survival models to estimate both the relative and absolute differences in the hazard of the primary outcome, to model the underlying differences in hazard, and to allow for nonproportional hazards. Addition of covariates to this model will allow adjustment for any baseline differences, and the addition of their interactions with the treatment allocation variable will test for subgroup effects. These models will allow examination of differences in effect for short-, medium-, and longer-term follow-up. The primary analysis of AFS will be undertaken on an intention-to-treat basis according to allocated first intervention, regardless of whether the intervention was delivered and whether repeat and cross-over interventions were subsequently undertaken.

Secondary outcome measures that are based on a continuous scale will be analysed using a repeated measures, multilevel model to examine any differential effect over time. Where necessary, data transformations will be made to fulfil modelling assumptions. Treatment effects from the repeated measures model will be reported at the 1-month, 12-month, and ‘end of follow-up’ time points.

Other outcome measures will be explored using standard methods (Fisher’s Exact Test for dichotomous outcomes, log-rank test for time to event data) and will also be reported at 1 month, 12 months, and at the end of follow-up.

Variation in the treatment effect between subgroups will be limited to prespecified variables and investigated using appropriate tests for interaction in survival and repeated measures models. Variables likely to be considered will include, but will not necessarily be restricted to, ischaemic rest/night pain only versus tissue loss, or both rest pain and tissue loss, presence of DM, and presence of CKD.

The economic analysis will comprise both ‘within-study’ and ‘model-based’ analyses. The ‘within-study’ analysis will be carried out to determine the cost-effectiveness of all three trial-mandated interventions on the basis of the patient-level data obtained during the study period.

Costs will take into account the use of NHS resources related to the primary interventions and any secondary procedures, readmissions, SAEs, hospital stays, day-case admissions, outpatient visits, and appointments with general practitioners and nurses. Changes in health-related QoL and QALYs will be calculated according to patients’ responses to the EQ-5D-5L instrument. Results of the analysis will be presented in terms of cost per year of AFS and cost per additional QALY gained. In line with existing recommendations, the base-case analysis will adopt a health care system (payer’s) perspective by considering costs incurred by the NHS and Personal Social Services [20].

The model-based analysis will be conducted to extend the within-study evaluation by considering costs and benefits likely to accrue beyond the study follow-up period. This will be based on a decision analytic model which will be built to serve as a framework for quantifying long-term costs and outcomes.

The chief investigator will coordinate dissemination of data from BASIL-3. All publications and presentations, including abstracts, relating to the main trial will be authorised by the BASIL-3 TMG. The results of the analyses will be published in the name of the BASIL-3 Collaborative Group in peer-reviewed journals (provided that this does not conflict with the journal’s policy). All contributors to the trial will be listed, with their contribution identified. Trial participants will be sent a summary of the final results of the trial which will contain a reference to the full papers. All applications from groups wanting to use BASIL-3 data to undertake original analyses will be submitted to the TMG for consideration before release. To safeguard the scientific integrity of BASIL-3, trial data will not be presented in public before the main results are published without the prior consent of the TMG.

At the time of writing, the BASIL-3 trial was halfway through the study’s pilot phase with 15 clinical centres open to recruitment and over 50 participants randomised.