Results of the search

The electronic search identified 62 studies; no studies were found through additional searching strategies (such as contact with experts and checking the references of studies​​). After the analysis of title and abstracts, we excluded 59 studies as they did not meet the inclusion criteria.Three RCTs were selected for full‐text analysis: Layton 1994; Ibbotson 1995 and del Río Solá 2012. After the full‐text analysis, two RCTs were included in the review (Layton 1994; del Río Solá 2012). Ibbotson 1995 used the same people and data as the Layton RCT but compared them with a control group of healthy people, so we excluded this trial (Figure 2).

Figure 2

---

---

Flow diagram of included and excluded studies

Included studies

Details of the included studies are summarised in Characteristics of included studiesTable 1.

Layton 1994 reported a single‐centre parallel RCT (20 participants) that evaluated daily administration of oral aspirin (300 mg) in addition to compression, compared to placebo and compression. Participants were recruited from the Academic Unit of Dermatology, General Infirmary at Leeds, West Yorkshire, UK. del Río Solá 2012 reported a single‐centre parallel RCT (51 participants) that evaluated daily administration of oral aspirin (300 mg) in addition to compression compared to compression only and recruited from the Department of Angiology and Vascular Surgery, University Hospital of Valladolid, Spain, from December 2001 to September 2005.

Layton 1994 was a prospective trial that included 20 people with venous leg ulcers 2 cm or larger, divided into two parallel groups: 10 participants received oral aspirin (300 mg/day) plus compression therapy compared with 10 who received placebo plus compression therapy.

The del Río Solá 2012 study was a prospective trial that included 51 people with venous leg ulcers 2 cm or larger, divided into two parallel groups: 23 participants received oral aspirin (300 mg/day) plus compression therapy compared with 28 who received compression therapy alone.

The source of funding was not reported for either trial.

Both RCTs used the same intervention (oral aspirin 300 mg/day). The Layton 1994 study randomized participants to a placebo as an adjunct to compression therapy in the control group, while the del Río Solá 2012 study randomized those in the control group to compression therapy alone (no placebo). Compression therapy consisted of high compression (Setopress) in the Layton 1994 RCT. The del Río Solá 2012 RCT did not specify what type of compression was used, although the trialists reported the application of a two‐layer system consisting of one layer of padding and one layer of compression bandage.

The Layton 1994 trial followed participants for four months and then stopped the trial. The results were expressed in terms of reduction of the ulcers' size (surface area cm²) and proportions of ulcers completely healed in the trial period.

The del Río Solá 2012 trial followed participants until their ulcers had healed completely. The results were expressed in terms of proportion of healed ulcers in each group and average time for ulcer healing. . After healing, they continued to follow participants to analyse the proportion of ulcer recurrence and time until ulcer recurrence.

Excluded studies

We excluded the Ibbotson 1995 trial. This trial used the same participants reported in Layton 1994 and compared them with a control group of healthy people without venous leg ulcers. The objective was to evaluate some haemostatic parameters in people with venous leg ulcers taking oral aspirin (see Characteristics of excluded studies).

Allocation

The Layton 1994 trial was described as 'randomized' but did not report the method of randomizations or the concealment of allocation. In del Río Solá 2012, randomization was undertaken by an independent researcher using a computer program, but allocation concealment was not described.

Blinding

The Layton 1994 trial compared oral aspirin with placebo, and the trial was described as 'double‐blinded', but the trialists did not describe their methods. The evaluation of ulcer size was conducted by planimetry of photographs taken in standardized conditions, so we judged this as low risk of bias.

The del Río Solá 2012 trial did not use a placebo but compared intervention (aspirin) with no intervention. In this scenario, blinding of participants and personnel was not possible. The study reported information about ulcer development (size, epithelization) weekly in a specific data sheet, but they did not describe who did what and how this work was done, so we judged this study as high risk of bias.

Incomplete outcome data

The Layton 1994 trial did not have any withdrawals. The del Río Solá 2012 trial reported four withdrawals, two from each group. Two of these four participants were withdrawn because they needed to be in hospital, but the cause and group assignments were not reported.

Selective reporting

As there were several inconsistencies in the del Río Solá 2012 report, we assigned the study a low risk of reporting bias. Time for complete ulcer healing in the control group, which was stated as 22 weeks in the text but as 16.5 weeks in the table (Table I of the study); errors with the numbers of people in each group, which were reported as 23 for the aspirin group and 28 for the control group in the main text, but appeared as 20 and 31 in the table; and the proportion of people completely healed was reported as 0.73% rather than 75% (21 out of 28 people) for the control group and 0.75% rather than 74% (17 out of 23 people) for the treatment group.

As the results of both prespecified primary outcomes were reported in Layton 1994, we judged it to be free of selective reporting.

Other potential sources of bias

The Layton 1994 trial was stopped in the fourth month. In that time, 38% of ulcers had healed in the aspirin group, but no ulcers had healed in the placebo group. However, the time period could be considered too short, as the assumption that all ulcers could heal in this period may lead to misinterpretation (i.e. some ulcers may have healed in the control group given more time).

In the del Río Solá 2012 trial, despite an appropriate method of randomizations (an independent researcher and computer program), the generated groups were different in relation to the length of time that people had ulcers before treatment. 'Young' ulcers predominated in the aspirin group, and may have had a tendency to heal faster than chronic or 'older' ulcers. Also, the trial authors did not determine a specific trial period. To calculate the size of the study, they used an expected difference of 45% between groups, but did not specify whether the difference was in the proportion of people with healed ulcers, the area of the ulcers or the time for healing.

Primary outcomes

Time to complete ulcer healing

The del Río Solá 2012 trial followed the people until the healing of their ulcers and reported the average time for healing was 12 weeks in the aspirin group and 22 weeks in the control group, P values and confidence intervals were not reported Table 1.

Rate of changes in the area of the ulcer in the trial period

The Layton 1994 stopped the trial in the fourth month and reported this outcome: by four months, ulcer area had reduced (by 6.5 cm², a 39.4% reduction) in the aspirin group compared with no reduction in ulcer area in the control group (P value < 0.002; summary of findings Table for the main comparison).

Proportion of ulcers healed in the trial period

The Layton 1994 trial reported this outcome: a higher proportion of the ulcers (38%) in the aspirin group had completely healed in the trial period compared with none in the control group (P value < 0.007; summary of findings Table for the main comparison).

Proportion of people with ulcers healed in the trial period

The del Río Solá 2012 trial reported that there was no real difference between the proportion of people with ulcers healed in the aspirin group (74%, 17 out of 23 people) and in the control group (75%, 21 out of 28 people; Table 1), (RR 0.99, 95% CI 0.71 to 1.36; Analysis 1.1).

Major bleeding (haemorrhagic stroke, gastric bleeding, any bleeding requiring blood transfusion, any bleeding causing hospitalisation)

The del Río Solá 2012 trial reported two hospitalisations, in addition to two other withdrawals, but did not specify the cause or the group assignment of the participants who withdrew (Table 1).

The Layton 1994 trial reported that no side effects were experienced (Table 1).

Secondary outcomes

Pain relief

Neither study reported pain relief.

Mortality

The Layton 1994 trial reported that no side effects were experienced.

The del Río Solá 2012 trial reported that two people needed hospitalisation and had to be withdrawn from the trial (in addition to two other withdrawals), but they didn't report the reason for hospitalisation or their group assignment. The authors did not report other adverse events.

Adverse events

The Layton 1994 trial reported that participants had no adverse effects, but did not explain how this was established. The del Río Solá 2012 trial did not report adverse events.

Ulcer recurrence

The del Río Solá 2012 trial reported the average time of ulcer recurrence was 16.33 days (SD: 7.5) in the control group and 39 days (SD: 6.0) in the aspirin group (P = 0.007).

Layton 1994 did not report recurrence.