Aspirin versus placebo for the treatment of venous leg ulcers—a phase II, pilot, randomised trial (AVURT)

AVURT was a multi-centred, pilot, phase II, randomised, double-blind, parallel-group, placebo-controlled, efficacy trial. Ten recruiting centres participated, which included leg ulcer hospital outpatient clinics (n = 5), community leg ulcer clinics or community caseloads (n = 3), a wounds clinic within a university (n = 1) and a primary care leg ulcer clinic (n = 1). The full details of the inclusion and exclusion criteria for trial participants can be found in the study protocol, which has been published [15]. Briefly, the trial recruited adults (aged 18 years or over with no upper limit) with at least one chronic VLU, ulcer area > 1 cm², ankle branchial pressure index (ABPI) ≥ 0.8 taken within the previous 3 months or, if the ABPI was incompressible, other forms of clinical assessment to exclude peripheral arterial disease (e.g. peripheral pulse examination, toe pressure, duplex ultrasound and clinical judgement). For patients with more than one ulcer, the largest ulcer was chosen as the reference ulcer for the purposes of the trial.

Before approaching patients, study research nurses pre-screened on three criteria (concomitant aspirin, wound size and ulcer duration or history of venous ulceration) to determine those potentially eligible for the study. Two pre-screening logs were issued. The first, used during the first 4 months of recruitment, was non-mandatory as stipulated by the trial Sponsor. The second pre-screening, implemented in the fifth month of recruitment, was mandatory in accordance with a recommendation by the Data Monitoring Committee (DMC). Therefore, the number pre-screened as shown in the pre-screening study flow diagram (Fig. 1) is an under-representation of the number pre-screened as not all sites completed the non-mandatory log.

Potentially eligible patients visiting the clinics for a routine appointment were invited to take part in the trial by a designated health care professional, including research nurses. A patient information leaflet was provided and patients were given a minimum of 24 h to consider participation. A full eligibility check was conducted for patients who gave consent and an eligibility case report form (CRF) was completed by the health professional. Patients were informed during consent that their eligibility would be subject to confirmation by a doctor. Following review and confirmation of a patients’ eligibility by a doctor, the doctor signed off the prescription for the investigational medicinal product (IMP).

Participants were randomised on a 1:1 basis to receive either aspirin (300 mg) or placebo, in addition to standard care. The aspirin and placebo manufacturer, Sharp Clinical Services UK Limited, generated the randomisation schedule in advance. The schedule was stratified by ulcer area (≤ 5 cm² or > 5 cm²) and mirrored top to bottom–bottom to top in order to facilitate participant allocation according to stratification. One randomisation list was provided to St George’s Hospital Research Pharmacy and a copy to the Senior Trial Statistician in the York Trials Unit (YTU). The Research Pharmacy was responsible for the randomisation, which was performed upon receipt of a valid prescription for a participant, and provided the 24-h code-breaking service in case emergency un-blinding was needed. Participants, investigators, research and treating nurses and other attending clinicians were unaware of the trial drug allocation throughout the trial.

Participants in the intervention group were allocated 300 mg of daily oral aspirin (single dose) for 24 weeks whereas participants in the control group were allocated daily oral placebo. Placebo capsules were identical in weight, colour and size to the aspirin capsules and contained the same filler.

Participants were offered an evidence-based standardised approach to the management of their leg ulcers in accordance with Scottish Intercollegiate Guideline Network (SIGN) guidance [16]. This consisted of multi-component compression therapy aiming to deliver 40 mmHg of pressure at the ankle, where possible. The type of dressing used was at the discretion of the healthcare professionals managing the participants.

After consent and confirmation by a doctor that the patient was eligible, but prior to randomisation, a nurse conducted a clinical assessment of the participant and wound. Baseline demographic and clinical details were recorded, including ulcer size, on a paper CRF. Medication that the patient was currently taking was recorded in a diary to be retained by the patient.

Following randomisation, participants continued in the normal care pathway of weekly or fortnightly clinical assessments at community ulcer clinics, hospital outpatient clinics or home visits with no additional visits required for the study. Nurses completed follow-up paper CRFs at each visit to record assessments of healing outcomes, treatment concordance with IMP and compression bandaging, AEs or side effects, change of concomitant medication and resource use (e.g. number of visits, number of changes to dressings and changes to compression therapy). Digital photographs, or leg ulcer tracings, were also taken by the treating or research nurse at each visit.

Participants kept a medication diary to record any changes to their concomitant medication. The diary was reviewed by clinical staff at follow-up visits to monitor and record changes in medication. Changes in medication were subsequently reviewed by a doctor to ensure safety.

All participants were to be followed for 25 weeks.

If the reference ulcer was assessed as healed during the follow-up period, a photograph was taken and the participant continued to take the aspirin or placebo for two further weeks. At this latter time point, they were re-assessed for healing (as per FDA guidelines on wound healing [17]). If the ulcer was confirmed as healed at the reassessment visit, the participant was advised to stop taking the IMP or placebo and the date of ulcer healing was recorded as the date the ulcer was first assessed as healed (i.e. 2 weeks earlier). Participants whose ulcer had healed no longer continued with regular follow-ups but were given a card with contact details of the research nurse at the site and the date of their 25-week follow-up. Participants were asked to contact the research nurse if a new ulcer occurred on the reference leg before the end of the study. If the ulcer was assessed as ‘not healed’ at the second assessment, the participant was advised to continue with the trial medication and continued to receive weekly or fortnightly follow-ups.

Final follow-up was 25 weeks post-randomisation. However, for participants whose leg ulcer was first assessed as healed in weeks 24–25, follow-up was extended to weeks 26–27, respectively, to allow a 2-week period of assessment to confirm whether the ulcer had healed. For all participants, no trial medication was taken after week 25.

For participants whose leg ulcer had healed before week 25, the research nurse or other nursing staff phoned participants in week 25 to ask if the reference ulcer had reoccurred and to check for adverse events. At the end of the study, participants were asked to return the bottle containing all the remaining capsules in order to assess compliance. St George’s Research Pharmacy conducted the pill count.

The primary outcome was time to ulcer healing, which was defined as ‘complete epithelial healing in the absence of scab (eschar) with no dressing required’. Time to healing was measured in days from the date of randomisation until the first date that healing was recorded. Patients who did not heal or fully withdrew were censored at the end of the study or at their last visit, respectively. Secondary outcomes were: ulcer size (area) measured in cm² using image analysis by SigmaScan (Systat Software Inc., San Jose, CA, USA) and/or wound tracings; recurrence of reference ulcer; adverse events; ulcer-related pain using a visual analogue scale; treatment compliance (patient self-assessment, capsule counting and nurse assessment of compression concordance); and resource use (i.e. number of wound consultations and types of dressing used).

Compliance with study capsules was assessed in two ways. The first was by analysing the answer given to the question in the CRF, which was completed at each visit: “How often has the participant taken their AVURT capsules (300mg Aspirin/placebo per day) this week?”. Numerical values were assigned to the responses available as follows: every day = 1, most days = 2, some days = 3, and not at all = 4. To calculate mean level of compliance with study capsules for each participant, the responses across all weeks following delivery of the capsules up to healing/trial exit were summed and divided by the number of visits attended. This was then categorised as fully compliant if the mean value was 1, partially compliant if the value was 2 or 3 and not at all compliant if the value was equal to 4. The second way of assessing capsule compliance was by the count of returned capsules at the end of the study. Each participant was given 190 capsules and by subtracting the number of returned pills it was possible to obtain an estimate of the number of capsules actually taken. The number of capsules which should have been taken was calculated starting from the date of first dose until 2 weeks after healing (for those who had healed) or the date of the last visit (for those who did not heal). From this the percentage of capsules each participant took (of those they should have taken) was calculated.

The target sample size of 100 participants was considered to be sufficient to demonstrate whether there was evidence of an effect of aspirin to treat VLUs. The primary outcome was time to healing of the reference ulcer. Assuming a standard error for the hazard ratio (HR) of 0.105 following adjustment for log area and log duration of ulcer (as in a previous leg ulcer study with data on 448 participants, VenUS IV [18]) and applying this to the smaller sample size of 100 in this study implied that the standard error would be 0.22. A 95% confidence interval (CI) for the log hazard ratio would thus be log(HR) ± 0.435. Hence, if the hazard ratio for this study were the same as that suggested by previous studies (around 1.5 [8, 9]), the CI would be (0.97, 2.31) which just includes 1.00. It would be unlikely that if the HR is as suggested in the two previous studies that we would observe an overall HR below 1.00. Analyses were conducted using the principles of intention to treat. Time to healing was summarised using Kaplan-Meier curves and investigated using a stratified log-rank test. The primary analysis was a Cox proportional hazard regression adjusted for ulcer area (cm²) and duration at baseline, both logarithmically transformed; an unadjusted Cox PH regression analysis was also conducted. Adverse events (AEs) were reported overall and by trial arm in terms of number of participants with at least one event, total number of events and number of non-serious and serious AEs. Differences in total number of events by trial arm were compared using negative binomial regression adjusted for size and duration of ulcer (both log transformed). Ulcer recurrence was analysed descriptively reporting the number of patients who experienced a recurrence and the time to recurrence from healing. Compliance with study capsules (self-reported level of compliance and percentage of capsules taken) were presented descriptively.

It was estimated that 100 patients were sufficient to demonstrate whether there was evidence of an effect of aspirin to treat VLUs. As this trial recruited only 27 patients we were unable to determine the effectiveness of aspirin for this condition.