Types of studies

We considered all randomised controlled trials (RCTs) comparing wound cleansing with no wound cleansing, or RCTs comparing different wound cleansing solutions, or different wound cleansing techniques, for this review. We would have considered including controlled clinical trials (CCTs) in the absence of RCTs reporting any of the above comparisons but we did not find any CCTs that met the inclusion criteria. We included studies involving wound cleansing of different wounds if they presented data pertaining to venous leg ulcers separately and randomisation was stratified by wound type and it was possible to extract these data from the study.

Types of participants

We included adults (over 18 years of age) with a diagnosed venous leg ulcer, managed in any healthcare setting.

Types of interventions

For the purposes of this review, we defined cleansing as the application of fluid to a venous leg ulcer to aid removal of exudate, debris and contaminants, but not the use of dressings or mechanical debridement (Towler 2001). Studies investigating the following comparisons were eligible:

1. cleansing compared with no cleansing;

2. comparisons between different types of cleansing solutions;

3. comparisons between different cleansing techniques (e.g. irrigation, swabbing, soaking, immersion).

Types of outcome measures

Electronic searches

We searched the following electronic databases to identify reports of relevant RCTs:

• the Cochrane Wounds Specialised Register (searched 4 September 2019);

• the Cochrane Central Register of Controlled Trials (CENTRAL; 2019, Issue 8) in the Cochrane Library (searched 4 September 2019);

• Ovid MEDLINE including In‐Process & Other Non‐Indexed Citations (1946 to 4 September 2019);

• Ovid Embase (1974 to 4 September 2019);

• EBSCO CINAHL Plus (Cumulative Index to Nursing and Allied Health Literature; 1937 to 4 September 2019).

The search strategies for the Cochrane Wounds Specialised Register, CENTRAL, Ovid MEDLINE, Ovid Embase and EBSCO CINAHL Plus can be found in Appendix 1. We combined the Ovid MEDLINE search with the Cochrane Highly Sensitive Search Strategy for identifying randomised trials in MEDLINE: sensitivity‐ and precision‐maximising version (2008 revision) (Lefebvre 2011). We combined the Embase search with the Ovid Embase filter developed by the UK Cochrane Centre (Lefebvre 2011). We combined the CINAHL Plus search with the trial filter developed by Glanville 2019. There were no restrictions with respect to language, date of publication or study setting.

We also searched the following clinical trials registries:

• ClinicalTrials.gov (www.clinicaltrials.gov) (searched 10 September 2019);

• World Health Organization (WHO) International Clinical Trials Registry Platform (apps.who.int/trialsearch/Default.aspx) (searched 10 September 2019);

Search strategies for clinical trial registries can be found in Appendix 1.

Searching other resources

We searched the reference lists of all the included pertinent studies and other publications, such as guidelines and systematic reviews. We contacted 11 experts in the field, but did not identify any further study information. We planned to write to other authors of relevant publications to identify any completed or ongoing studies; however, we did not identify any.

Selection of studies

Two review authors independently assessed the titles and, where available, abstracts of the studies identified in the search strategy, to determine their eligibility for inclusion in the review. We obtained full versions of potentially relevant studies, and two review authors independently screened these against the inclusion criteria (see PRISMA flow chart in Figure 1; Liberati 2009). Review authors resolved any differences in opinion by discussion.

Figure 1

---

---

PRISMA Flowchart

Data extraction and management

The primary review author (NMcL) extracted the details of eligible studies and summarised them into a pre‐prepared data extraction table. The second and third review authors (ZM, PA) checked the extracted data. If we found that there were missing data in the studies, we contacted the study authors to obtain the missing information.

We extracted the following information for each study;

• author, title, source;

• date of study, study's geographical location;

• care setting;

• inclusion/exclusion criteria;

• duration of ulcer;

• participant characteristics;

• study design details;

• sample size calculation and sample size;

• intervention details, concurrent interventions (compression therapy: applied or not applied, elastic or inelastic, light (18 mmHg to 24 mmHg) or full compression (35 mmHg to 40 mmHg);

• use of additional dressing materials;

• outcome measures;

• length of follow‐up;

• results;

• conclusions, as reported by the study authors;

• funding source.

Assessment of risk of bias in included studies

For this review two review authors independently assessed each included study using the Cochrane tool for assessing risk of bias (Higgins 2011b). This tool addresses six specific domains, namely random sequence generation, allocation concealment, blinding, incomplete outcome data, selective outcome reporting and other issues (Appendix 2). We assessed blinding and completeness of outcome data for each outcome separately. We completed a 'Risk of bias' table for each eligible study, and judged each component of the study to be at low, unclear or high risk of bias. We present an assessment of risk of bias using a 'Risk of bias' summary figure, which reports all of the judgements in a cross‐tabulation of study by domain.

Measures of treatment effect

We conducted data analysis according to Cochrane guidelines (Higgins 2011a). We entered the data into Review Manager 2014. For dichotomous outcomes (e.g. ulcer healed during time period) we calculated the risk ratio (RR) with 95% confidence intervals (CIs). Risk ratio is the ratio of the risk of the event of interest (e.g. leg ulcers healed) in the experimental group divided by the risk of this event in the control group and indicates the chances of healing for people in the experimental group compared with the control group (Higgins 2011a). For continuous outcome data (e.g. absolute or percentage changes in the ulcer area) we used the mean difference (MD) with 95% CIs. If studies had used different assessment scales, we would have used the standardised mean difference (SMD) with 95% CIs. We would have reported time‐to‐event data (e.g. time to healing) as hazard ratios (HRs) where possible, in accordance with the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2011). For statistically significant effects in binary outcomes, we planned to calculate the number needed to treat for an additional beneficial outcome (NNTB), or number needed to treat for an additional harmful outcome (NNTH). If data appeared to be skewed, and if scale data had finite upper and lower limits, we would have used the easy ’rule of thumb’ calculation to test for skewness. That is, if the standard deviation (SD), when doubled, is greater than the mean, it is unlikely that the mean is the centre of the distribution (Altman 1996). In such a case, we would not have entered the data into any meta‐analysis. If we had found relevant data that was skewed, we would have presented the data in 'Additional tables'.

Unit of analysis issues

Some studies randomise by participant, but analyse outcomes by wound. If this had occurred, and the numbers of participants and wounds were equal (i.e. one wound per participant), we would have treated the participant as the unit of analysis. Where we identified clustered data outside of the context of a cluster‐randomised trial, that is, where a proportion of randomised trial participants had outcome data collected and reported on multiple wounds, we planned to record this in the 'Risk of bias' assessment. We would have extracted and presented the data, but it would not have been the subject of any further analyses.

For cluster‐randomised trials that used analysis methods to account for the clustering, we planned to extract effect sizes and standard errors from the appropriate analysis. Where cluster‐randomised trials had been analysed incorrectly, we planned to recalculate the results using the appropriate unit of analysis if required (Higgins 2011a), and if sufficient data were available. The included studies did not randomise multiple wounds, neither did we include cluster‐randomised trials, therefore we did not identify any unit of analysis issues.

Dealing with missing data

Where there were missing data, we attempted to contact the study authors to request the missing information. However in some cases the data remained missing despite our best efforts, and our analysis was restricted to the published data because specific figures for each group were not available. We were also unable to perform sensitivity analysis to assess how sensitive the results were to reasonable changes. We review the potential impact of the missing data in the discussion section.

Assessment of heterogeneity

We planned to consider both clinical and statistical heterogeneity. We planned to assess clinical heterogeneity using key study features: participant characteristics, intervention type, study duration and outcome types. We planned to assess statistical heterogeneity using the Chi² test with significance set at P < 0.10, in conjunction with the I² measure (Higgins 2003). The I² measure evaluates the percentage of total variation across the studies due to heterogeneity rather than due to chance (Higgins 2003). Where the I² measure is 0% to 25% this indicates a low level of heterogeneity, and where the I² value exceeds 75% this represents a very high level of heterogeneity (Deeks 2011). We planned to carry out statistical pooling on groups of studies which we considered to be sufficiently similar. Where heterogeneity was absent or low (I² = 0% to 25%), we planned to use a fixed‐effect model; if there was evidence of heterogeneity (I² > 25%), we would have used a random‐effects model. If heterogeneity was very high (I² > 75%) we would not have pooled the data (Higgins 2003).

Assessment of reporting biases

We assessed reporting bias using guidelines in the Cochrane Handbook for Systematic Reviews of Interventions (Sterne 2011). If we had identified sufficient studies for a meaningful assessment of publication bias, we would have constructed a funnel plot of primary outcomes to test for asymmetry. If asymmetry was present, we planned to explore possible causes, including reporting and publication bias. However, as we only included four studies in this review, this was not possible.

Data synthesis

We have combined details of included studies in a narrative review according to type of comparator. We were unable to pool data for meta‐analysis. If we had identified more studies, we would have considered clinical and methodological heterogeneity and pooled data if studies appeared to be appropriately similar in terms of wound type, intervention type, duration of follow‐up and outcome type.

We were unable to predict the amount of clinical, methodological and statistical heterogeneity between included studies, but it might have been extensive. Thus, we anticipated using a random‐effects approach for meta‐analysis. Conducting meta‐analysis with a fixed‐effect model in the presence of even minor heterogeneity may provide overly narrow confidence intervals. We would only have used a fixed‐effect approach if we had assessed clinical and methodological heterogeneity to be minimal, and the assumption that a single underlying treatment effect is being estimated held. We would have used I² and Chi² statistics to quantify heterogeneity, but would not have used these to guide our choice of model for meta‐analysis. We would have exercised caution when meta‐analysed data were at risk of small study effects, because a random‐effects model may be unsuitable in such a situation. In this case, or where there were other reasons to question the selection of a fixed‐effect or random‐effects model, we would have assessed the impact of the approach using sensitivity analyses to compare results from alternate models (Thompson 1999).

We have presented the data using forest plots where possible. For dichotomous outcomes, we have presented the summary estimate as a risk ratio (RR) with 95% CI. We have used the mean difference (MD) with 95% CI for continuous outcomes. If we had included more than one study that reported continuous outcomes, and the studies measured these in the same way, we would have presented a pooled MD with 95% CI; we planned to pool standardised mean difference (SMD) estimates where studies measured the same outcome using different methods. For time‐to‐event data, we planned to plot (and, if appropriate, pool) estimates of hazard ratios and 95% CIs as presented in the study reports using the generic inverse variance method in Revman 5 (Review Manager 2014). Where a study analysed time‐to‐healing as a continuous measure but it was not clear if all wounds healed, we would have documented the study's use of the outcome but would not have summarised the data or used it in any meta‐analysis.

We would have obtained all pooled estimates of treatment effect using Revman 5 (Review Manager 2014).

Subgroup analysis and investigation of heterogeneity

If substantial heterogeneity had existed between studies for the primary outcomes (that is, when the I² measure exceeded 50%), we would have explored reasons for heterogeneity. We planned to include the following subgroup analyses:

• presence or absence of compression therapy;

• type of compression therapy (elastic or inelastic);

• level of compression therapy (light (exerting 14 to 17 mmHg of compression) or full (exerting up to 60 mmHg of compression)).

Sensitivity analysis

We planned to perform a sensitivity analysis by including only those studies assessed as low risk bias in the key domains of adequate generation of the randomisation sequence, adequate allocation concealment and blinding of outcome assessor for the estimates of treatment effect.

Summary of findings and assessment of the certainty of the evidence

We present the main results of the review in four 'Summary of findings' tables which provide key information concerning the quality of evidence, the magnitude of effect of the interventions examined, and the sum of available data on the main outcomes, as recommended by Cochrane (Schünemann 2011a). The 'Summary of findings' tables include an overall grading of the evidence related to each of the main outcomes, using the GRADE approach (Schünemann 2011b). We include the following main outcomes in the 'Summary of findings' tables:

• time to healing

• number of wounds completely healed during follow‐up (frequency of complete healing)

• change in ulcer size over time, a reduction in the original size of the wound area within the duration of the study

• procedural pain (using validated scales, where reported)

• cost

• health‐related quality of life (using valid scales, where reported)

• adverse events (e.g. infection, pain, maceration (wet, white, boggy tissue), excoriation (red, hot, shiny tissue) or bleeding.