Study design

International multicentre open-label RCT with an adaptive (multi-arm, multi-stage (MAMS)) design, with a follow-up of 12 months, following the (extended) Consolidated Standards of Reporting Trials guidelines.12 We will have four study arms and two stages (one interim and one final analysis; see figure 1). Patients will be randomised through Castor to receive standard (endo)vascular surgical treatment and wound care with 0, 20, 30 or at least 40 HBOT treatments.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Flow of patient inclusion and analysis. HBOT, hyperbaric oxygen therapy.

The efficient MAMS design for clinical trials includes a planned interim analysis, based on which only the best performing study arms will be chosen to continue.13–15 This will reduce the total number of patients required and the number of patients receiving an ineffective treatment, because after the interim analysis no patients will be included in less effective treatment arms.

An open-label trial was chosen because a sham-controlled trial was not feasible. Only a few HBOT centres have sufficient time slots to provide sham treatments and HBOT treatments are typically reimbursed by the health insurance companies, whereas sham treatments are not.

Patient and public involvement

The relevant Dutch patient association (Harteraad) was involved during the development of the study protocol. They defended the best interests of the patients and gave written feedback on the protocol. Harteraad will also facilitate the dissemination and implementation of the outcomes of the trial.

Study population

Consecutive patients with type I or II diabetes, a DFU and leg ischaemia, presenting at the vascular surgery departments of the participating hospital. This international trial will be coordinated from the Netherlands, where 10 Dutch HBOT centres with their referring hospitals will participate to include patients for the trial. We will invite as many Dutch vascular surgical centres as possible. In addition, a growing number of centres from different countries, among which Australia, Curaçao, Spain and Bosnia have agreed to collaborate. All participating centres have consented to treat similar patient populations and apply HBOT based on a similar protocol. Patients will be screened for eligibility in the outpatient clinics of vascular surgeons and diabetologists or hospital wards of participating hospitals. Also, HBOT centres and dermatology clinics may refer patients to the vascular surgery departments for inclusion in the study. Patients have to be presented at, and included after, a multidisciplinary consultation to select suitable patients and avoid doctor’s delay. Hospitals are encouraged to discuss all patients with DFU in a multidisciplinary meeting to assess all eligible patients. Patients are eligible for inclusion irrespective of their suitability for revascularisation.

Inclusion criteria

In order to be eligible to participate in this study, a subject must meet all of the following criteria:

1. Type I or II diabetes.

2. One or more deep and clinically infected lower extremity ulcers. It is allowed to grade these ulcers by means of either the Meggit-Wagner, the Texas or the WIfI classification tools.16 17 Thus, patients with a Meggit-Wagner class 3 or 4, a Texas class 2C, 3C, 2D or 3D or a WIfI class W>1, I>1 and fI>0 ulcer are eligible for inclusion. Ulcers should be present and treated for at least 4 weeks. In case more than one ulcer is present, the largest will be regarded as target ulcer. Alternatively, patients with diabetes who have undergone a minor amputation because of a previously existing ischaemic DFU on a toe or forefoot are also eligible.

3. Leg ischaemia, characterised by a highest ankle systolic blood pressure <70 mm Hg, or a toe systolic pressure <50 mm Hg, or a transcutaneous oxygen pressure (TcpO₂) <40 mm Hg.

4. Complete assessment of peripheral arterial lesions from the aorta to the pedal arteries with duplex ultrasonography, magnetic resonance angiography, CT angiography and/or intra-arterial digital subtraction angiography of the ipsilateral leg.

5. Patients have to be discussed in, and included after a multidisciplinary consultation.

6. Age ≥18 years.

7. Written informed consent.

Exclusion criteria

A potential subject who meets any of the following criteria will be excluded from participation in this study:

1. Chronic obstructive pulmonary disease Global Initiative for Chronic Obstructive Lung Disease (GOLD) class IV.

2. Treatment with chemotherapy, immunosuppressive drugs or systemic corticosteroids within last 3 months.

3. End-stage renal disease requiring dialysis, because of logistic challenges of combining HBOT and dialysis, which has led to high drop-out rates in earlier studies.

4. Metastasised malignancy.

5. Left ventricular failure with ejection fraction <20% or external pacemaker.

6. Recent thoracic surgery or middle ear surgery.

7. Severe epilepsy.

8. Uncontrollable high fever.

9. Pregnancy.

10. Insufficient proficiency of local language, or inability to complete the questionnaires.

Intervention

Patients allocated to the intervention group will receive 20, 30 or a minimum of 40 sessions of HBOT adjunctive to standard care according to the International Working Group on the Diabetic Foot (IWGDF) guidelines.18 Although some differences in local best practices for DFU care are possible, this potential variation is accepted as part of a pragmatic trial. In the group with at least 40 treatments, the patients will initially receive 40 treatments. Next, the hyperbaric medicine physician may decide on clinical grounds whether it seems wise to continue the HBOT treatment up to a maximum of 60 sessions.

HBOT will be administered in a hyperbaric chamber. This can be both a monoplace or multiplace chamber. In a monoplace chamber, oxygen can be administered by either by mask or by flushing the chamber with oxygen following local best practice. Monoplace chamber atmosphere compression on air is allowed. In multiplace chambers, both masks and hoods can be applied, depending on local practice and patient preference. All participating HBOT centres have to adhere to the range of time and pressure provided in the trial protocol. When flushing a monochamber with oxygen, air breaks will not be possible and sessions will be conducted conforming to best practice within the ranges of the protocol. A HBOT session will take 90–120 min at 2.2–2.5 atmospheres absolute. Besides the 90–120 min of treatment, 10–20 min are required for compression and decompression. During the treatment session, the patients will breathe 100% FiO₂ except for three blocks of 5 min during which atmospheric air will be administered to prevent oxygen intoxication. Patients will either lay in bed or sit in a chair during the session. At least the first time, cabin personnel will accompany patients in the cabin and instruct them about the use of the oxygen masks or hoods. Subsequently, the supervising staff member will supply patients with masks through which the oxygen will be administered. In case of adverse events requiring termination of the session, patients will be able to leave the cabin after a short decompression period. After the treatment, the cabin will be decompressed in about 10 min.

HBOT is given at least 5 times per week, until the allocated number of sessions has been reached. To minimise interruption of HBOT and masking of the effect due to additional (endo)vascular interventions, any (additional) vascular procedure will preferably be performed after randomisation but before the start of HBOT. In case a vascular intervention is required (due to progression of the disease) while the patient is undergoing the HBOT sessions, these are interrupted for the period of the vascular intervention and continued directly afterwards until completion of the sessions. There is no time limit before a vascular intervention should take place or when HBOT should be resumed. However, data will be collected regarding the delay from inclusion until vascular intervention and the delay in HBOT treatment.

Sample size calculation

Primary outcome is the major amputation rate after 12 months. No criteria have been specified in the protocol as to when to perform a major amputation. This is a clinical decision mostly, but not exclusively, made by the vascular surgeon, depending on the wishes of the patient. A meta-analysis showed a 15% difference (11% vs 26%) of major amputations between patients treated with HBOT and patients receiving standard wound care.8 The sample size calculation is based on a one-sided type I error of 5% and a power of 90% to control for a type 2 error. A generalisation of the power requirements for testing multiple active treatments to one control group was used as proposed by Dunnett et al.19

The critical values for continuation of a treatment arm are ORs of 0.76 and 0.35, which represent a drop in amputation rates from 26% to 21% and from 26% to 11%, respectively. If the difference between two arms is <5% the HBOT arm is deemed futile, while a difference of 15% or higher is deemed superior. Based on these critical values, the required number of patients in each group in each stage is 68, leading to a required maximum number of 544 patients. The sample size calculation is based on a generalisation of Whitehead and Jaki and Jaki et al.20 21 Taking into account a possible loss to follow-up of 5%, we will need to include up to 573 patients.

Objectives

The aim of the study is to confirm or refute the hypothesis that HBOT is effective as an adjunctive treatment to standard wound care for patients with an ischaemic DFU to prevent major amputations, and to establish the optimum number of hyperbaric sessions to obtain this purported effect.

Primary end point

Major (above-ankle) amputation rate after 12 months of follow-up.

Secondary end points

• (Major) AFS after 3 months, 12 months and 3 years after the start of HBOT.

• Health-related quality of life, measured using the EuroQol 5-Dimension 5-Level (EQ-5D-5L) and 12-item Short Form Health Survey (SF-12) questionnaires at the start, 3 months and 12 months after the start of HBOT.

• Complete wound healing, defined as complete epithelial coverage of the wound, assessed by a vascular surgeon after 3 months, 12 months and 3 years after the start of HBOT.

• Pain scores assessed weekly during HBOT by means of the Visual Analogue Scale (VAS) on a 0–10 scale.

• Number and type of additional (vascular) interventions during the 12-month follow-up period.

• Cost-effectiveness and budget impact with a time horizon of 12 months.

• Mortality during the 12-month follow-up period.

• Adverse effects of HBOT during the 12-month follow-up period.

• Patient’s perception of improvement, as measured by an anchor question at the end of the HBOT treatment.

• TcpO₂ before, during and after 3 months and 1 year of HBOT. This can be measured by either the hospital or the HBOT facility, depending on availability of the device.

Follow-up moments should be completed within 2 weeks.

Statistical analysis

All analyses will be conducted according to the intention-to-treat principle and per-protocol analysis. Baseline characteristics are summarised with descriptive statistics. Differences between dichotomous variables will be expressed as a risk difference and numbers needed to treat or harm with 95% CIs. Continuous variables are summarised as means and SD or as medians and IQRs in case of a skewed distribution.

At the interim analysis, differences in amputation rates between the study arms will be assessed with the χ² statistic or Fisher’s exact test, where appropriate. At the final analysis, major amputation rates and AFS will be estimated using Kaplan-Meier survival analysis and differences between the groups will be analysed using the log-rank test.

Differences in (semi-)continuous variables between the two treatment groups (such as the VAS, TcpO₂ and quality of life scores) will be analysed with the Student’s t-test in case of normally distributed data, and otherwise with the non-parametric Mann-Whitney U test, after correcting for baseline differences, if any. Time to complete wound healing will be estimated using Kaplan-Meier survival analysis and differences between the groups will be analysed using the log-rank test. Changes over time of TcpO₂ and quality of life scores will be analysed using the Wilcoxon test. Potential confounding factors will be adjusted for by means of Cox proportional hazards regression modelling. For all analyses, a p value <0.05 will be used to indicate statistical significance.

The economic evaluation of additional HBOT will be performed as a cost-effectiveness analysis from a societal perspective. Additionally, a cost-utility analysis will be performed with the costs per quality-adjusted life-year (QALY) as outcome. As the time horizon is restricted to 12 months, no discounting (of costs and effects) will be done.

The EQ-5D-5L is used to generate health status scoring profiles over time, which will subsequently be translated in QALYs by applying time trade-off based health utility algorithms and assuming that interpolation between successive measurements best reflects a patient’s health status during the year. Sensitivity analyses will be performed to account for sampling variability, for plausible ranges in unit costs and for different health utility algorithms. In addition, a budget impact analysis will be undertaken to assess the possible cost savings of HBOT if applied to all eligible patients on a national scale. Subgroup analyses per country will be performed for the primary and secondary outcome measures.

Interim analysis

After inclusion of 68 patients in each group and 3 months of follow-up of these patients, an interim analysis will be conducted (figures 1 and 2). The data safety monitoring board (DSMB), including a vascular surgeon, diabetologist and statistician, will conduct the interim analysis. From our previous trial, we know that the vast majority of amputations occurs within 3 months of follow-up.11 Whether and how the trial will continue will depend on the outcome of the interim analysis, which is the main feature of the MAMS study design.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Possible scenarios at the interim analysis. HBOT, hyperbaric oxygen therapy.

The stopping rules and possible scenarios after the interim analysis apply:

1. If one of the HBOT regimens leads to a <5% improvement in amputation rate as compared with controls (based on the expected 26% vs 21% difference, ie, an OR of 0.76), then this treatment group will be discontinued as it is futile.

2. If only one HBOT regimen proves to decrease amputation rate significantly by >15% (26% vs 11%; OR 0.35) as compared with controls, then superiority of this treatment arm is proven and the trial will be discontinued as it has reached its aim.

3. If more than one HBOT treatment arm shows this superiority, the trial will continue with these groups. In this scenario, no more patients will be included in the other HBOT group.

4. If one or more HBOT group(s) show(s) minor improvements in amputation rate (between 5% and 15%) over control, then these groups will continue in the second stage of the trial.

Ethics and dissemination

The DIONYSIUS trial will be conducted according to the principles of the Declaration of Helsinki (version of Fortaleza, 2013) and in accordance with the Medical Research Involving, Human Subjects Act (WMO) and other guidelines, regulations and Acts. The medical ethics review board of the Amsterdam University Medical Centres has approved the trial protocol NL72855.018.20.

Recruitment and consent

Patients will be recruited by their vascular surgeon in the participating hospitals and will be informed both verbally and through the patient information form. They will be asked for written informed consent by a local investigator, or a co-worker in the contributing centre who is not the treating physician.

Benefits and risks assessment

All patients enrolled in this trial will receive the optimal (endo)vascular or conservative (antibiotics, anticoagulants, antihypertensives, lipid-lowering drugs, glycaemic control) and local wound treatment according to best practice and (inter)national guidelines. The possible effect of adding HBOT to the standard treatment may result in major benefits such as an improved limb salvage and higher wound healing rates. HBOT is regarded as a low-risk to moderate-risk therapy. Described adverse effects are middle ear barotrauma (up to 2%), myopia and sinus barotrauma.22

Data storage

All patient data will be stored in an electronic case report form in Castor, including a unique study number allocated to each patient. This number identifies the patient and must be reported on all web-based case forms and questionnaires. The handling of personal data will be according to the EU General Data Protection Regulation and the Dutch Act on Implementation of the General Data Protection Regulation. Patient data will only be available to the principal and coordinating investigators. The data will be stored for 15 years. The clinical research unit of the Amsterdam University Medical Centres will monitor the study following current guidelines. In addition, the DSMB will monitor the safety and efficacy of the intervention. The DSMB will assess the safety of the trial after 25%, 50% and 75% of the trial inclusion and perform the interim analysis after 50% of the inclusions with 3-month follow-up.

Public disclosure and publication policy

At least one full scientific article will be prepared by the coordinating and principal investigators. All local investigators will be acknowledged in the final manuscript when they include any number of patients with a complete follow-up. Coauthorship will be granted if complying with the International Committee of Medical Journal Editors guidelines. Study results will be presented at national and international meetings by the coordinating and/or principal investigator. The subsets of data from each individual contributing centre may be used for quality improvement purposes by the local investigator(s).