Introduction

Diabetic foot ulcers (DFU) affect about 2–5% of patients with diabetes mellitus [1] and represent a challenging complication due to the higher associated risk of major adverse lower-limb events (MALE; i.e.: artery revascularization, sepsis, amputation, disability, and all-cause mortality) in comparison with diabetic patients without DFU [2].

DFU often caused by underlying conditions, such as peripheral artery disease, neuropathy, foot deformities, required timely treatments to reduce the risk of adverse events, recurrence, and ulcer healing failure [3, 4]. Gold standard treatments of DFU are therefore peripheral artery revascularization, management of infections, and plantar pressure relief. However, together with these therapeutic approaches, managed by a multidisciplinary team [5], local treatments, such as an appropriate debridement, exudate and local infection/inflammation management, and local active dressings, are equally important for promoting and accelerating ulcer healing [6]. Common local dressings have differential actions for managing exudate, infection, non-viable tissues, and promoting tissue granulation [6,7,8].

Several commonly used specialized wound dressings are available for DFU local treatment, including: alginates; hydrocolloids/hydrogels, silver-impregnated dressings, sucrose octasulfate- impregnated dressings, micro-organism-binding dressings, etc. Unfortunately, these widely used local advanced dressings have few evidence on their effectiveness and safety and their recommendation would require a formal evaluation on the basis of published randomized control trials.

This systematic review and meta-analysis of randomized control trials intends to assess the effects of local advanced dressings commonly used in the treatment of DFU. The aim of the prespecified analyses reported in this paper was to answer to a clinical question [9] (PICO):

"In patients with neuropathic (non-ischemic and non-infected) foot ulcers, is the use of an active dressing preferable to standard of care, to reduce adverse outcomes and increase the chance of ulcer healing” (PICO #9 [9])?

The present meta-analysis was, in fact, performed in the process of developing the Italian guidelines for the treatment of Diabetic Foot Syndrome (DFS) [9]. These guidelines, which have been promoted by the Italian Society of Diabetology (Società Italiana di Diabetologia, SID) and the Italian Association of Clinical Diabetologists (Associazione Medici Diabetologi, AMD), are being developed for inclusion in the Italian National Guideline System (INGS), designed as a standard reference for clinical practice in Italy, using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) method [10].

Methods

We conducted this systematic review and meta-analysis in conformity with PRISMA checklist (Table S1) [11] and following a protocol previously published [9].

Search strategy and selection criteria

This meta-analysis is a part of a wider meta-analysis of studies on DFS conducted for the development of the Italian guidelines on the treatment of diabetic foot syndrome [9]. The present analysis includes all RCTs including diabetic patients or reporting subgroup analyses on diabetic patients with foot ulcers comparing advanced wound dressing with placebo/no therapy/standard of care, with a duration of at least 12 weeks. Animal studies were excluded, whereas no language or date restriction was imposed. A Medline and Embase search were performed up to April 1st, 2024 using the following search string: diabetic foot ulcer AND local AND treatment. Detailed information on the search strategy is reported in Table S2. Further studies were manually searched for references from retrieved papers.

Selection criteria

To be eligible, a study should enroll patients with diabetes and foot ulcers assessing the effects of advanced wound dressing compared either with placebo or standard of care.

The adjunctive treatments to be assessed were divided into two groups based on their main action and/or characteristics:

  • Advanced wound dressings

    1. (1)

      Alginate dressings: highly absorbent calcium alginate or calcium sodium alginate which can be combined with collagen. The alginate forms a gel at the wound surface level, which can increase absorbency.

    2. (2)

      Foam dressings: polyurethane foam dressings designed to absorb ulcer exudate and maintain moist wound surface.

    3. (3)

      Hemoglobin spray: topically application of hemoglobin to the wound bed to promote oxygen diffusion and increase granulation tissue formation.

    4. (4)

      Hydrocolloids: are occlusive dressings usually composed of a hydrocolloid matrix. This matrix forms a gel to provide a moist environment. Fibrous hydrocolloids have been developed for being more absorbent than standard hydrocolloid dressings.

    5. (5)

      Hydrogels: dressing made of insoluble polymers and designed to absorb wound exudate or rehydrate a wound. Hydrogels are supplied in either flat sheets or as beads.

    6. (6)

      Hyaluronic acid: is a naturally occurring, non-immunogenic, biodegradable polysaccharide acting with fibrin to support the influx of fibroblasts and endothelial cells into the wound site and the subsequent formation of granulation tissue. Interventions assessed: HA-impregnated inert pads, HA gel, or cream; pad or matrix composed entirely of HA.

    7. (7)

      Protease-modulating matrix dressings: capable of regulating the activity of proteolytic enzymes in chronic wounds.

    8. (8)

      Sucrose octasulfate-impregnated dressings (Technology Lipido-Colloid-Nano-OligoSaccharide Factor): the potassium salt contained in the sucrose octasulfate inhibits the excess of matrix metalloproteinase and interacts with growth factors promoting the restoration of their biological functions and contributing to new tissue formation.

  • Dressings with antimicrobial properties

    1. (9)

      Silver-impregnated dressings: different types (foam, hydrocolloids, etc.) of dressing containing silver ions usually used for treating infected ulcers.

    2. (10)

      Honey-impregnated dressings: medical-grade honey dressings proposed as an antimicrobial/anti-inflammatory local treatment which can be used both for acute and chronic ulcers.

    3. (11)

      Microorganism-binding dressings: are capable of binding microorganisms in a two water-repellent surfaces and subsequently removed at dressing change.

To be included in the present meta-analysis, all these dressing, assessed to be included in the Italian Guidelines for the treatment of DFS, should be EMA-approved.

Two independent reviewers (A.S. and M.A.) screened all titles and abstracts of the identified studies for inclusion. Discrepancies were resolved by a third, independent reviewer (M. M.).

Data extraction and collection

Variables of interest were ulcer healing, major and minor amputation, all-cause mortality, time-to-healing, frequency of dress changes, quality of life, and back-to-walk, as previously reported [9].

Data extraction was performed independently by two of the authors (B.R. and A.S.), and conflicts were resolved by a third investigator (M.M.).

Titles and abstracts were screened independently by the authors, and potentially relevant articles were retrieved in full text. For all published trials, results reported in published papers and supplements were used as the primary source of information. When the required information on protocol or outcomes was not available in the main or secondary publications, an attempt at retrieval was performed by consulting the clinicaltrials.gov website. The identification of relevant abstracts and the selection of studies were performed independently by all the authors. Data extraction and conflict resolution were performed by two investigators (A.S. and B.R.). The risk of bias in RCTs was assessed using the Cochrane recommended tool [12], which includes seven specific domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other bias. The results of these domains were graded as ‘low’ risk of bias, ‘high’ risk of bias, or ‘uncertain’ risk of bias.

Endpoints

The primary endpoint was ulcer healing at the endpoint. Secondary endpoints were major and minor amputation, all-cause mortality, time-to-healing, back-to-walk, pain, quality of life, and frequencies of dressing changes at the endpoint.

Statistical analyses

Heterogeneity was assessed by I2 test, whereas Funnel plots were used to detect publication bias for principal endpoints with at least 10 trials.

If data from more than one study on a given outcome were available, a meta-analysis using a random-effects model as the primary analysis was performed. Mantel-Haenzel Odds ratios and 95% confidence intervals (MH-OR, 95% CIs) were either calculated or extracted directly from the publications. Weighted mean differences (WMD) and 95% CIs were calculated for continuous variables.

For all the explored endpoints, active dressings were analyzed together and individually.

For the primary endpoint only, several pre-specified sub-group analysis was performed comparing trials with shorter (< 6 months) and longer duration (≥ 6 months) and those including and not including infected ulcers.

All analyses were performed using Review Manager (RevMan), Version 5.3 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014) for RCT and Comprehensive Meta Analysis v. 2.

The Grading of Recommendations, Assessment, Development and Evaluations (GRADE) methodology10 was used to assess the quality of the body of retrieved evidence, using the GRADEpro GDT software (GRADEpro Guideline Development Tool. McMaster University, 201,526. Available from gradepro.org).

Results

Retrieved trials

The study flow summary is reported in Fig. 1S of Supporting Information. The search of Medline and Embase databases allowed the identification of 257 items; after excluding studies by reading the title (n = 211), a further 31 studies were excluded after reviewing full-text (Table 3S of Supporting Information). Out of 15 studies [7, 13,14,15,16,17,18,19,20,21,22,23,24,25,26] fulfilling all inclusion criteria, only two [7, 15] were placebo-controlled trials; the remaining 13 trials were open-label in comparison with standard of care (see Table 1). All trials were short-term studies (follow-up < 24 weeks), with the exclusion of.

Fig. 1
figure 1

Effects of active dressings on ulcer healing in comparison with standard therapy/placebo (control)

Full size image
Table 1 Main characteristics of included trials
Full size table

three trials: two versus hydrocolloids [21, 22] and one versus hydrogel [20].

No trials for alginates, foams, and microorganism-binding dressings have been retrieved preventing any reliable assessments.

The principal characteristics of the included studies, comparing active dressings (n = 540) with placebo or SoC (n = 468), are reported in Table 4S of Supporting Information. The mean age, proportion of women, and baseline HbA1c were 61 years, 24, and 7.6% (58 mmol/mol), respectively.

The quality of studies was heterogeneous and relatively low due to open-label design and possible relevant selection bias. Only one study7 was a high-quality trial with no relevant bias (Fig. 2S of Supporting Information).

Fig. 2
figure 2

Effects of active dressings on the ulcer time-to-healing in comparison with standard therapy/placebo (control)

Full size image

Ulcer healing and time to ulcer healing

Only two studies did not report information on proportion of healed patients [18, 19]. Advanced dressing, considered together, were associated with a significantly higher ulcer healing rate (MH-OR: 1.82 [1.32, 2.51], p = 0.0003 I2 = 33%; Fig. 1) in comparison with SoC/Placebo. Funnel plot suggested a possible publication bias which downloads the strength of this association (Fig. 3S). Differences across individual dressing did not reach a fully statistical significance (test for subgroup differences: P = 0.050, I2 = 51.1%); however, some dressing reported only a trend toward reduction of healing rate (i.e., hemoglobin spray) or even neutral effects (i.e. hydrocolloids/hydrofibers, protease-modulating matrix, and silver dressings) in comparison with SoC.

Nine trials reported information on time to healing [14, 17, 18, 20,21,22,23, 26, 27] showing a significantly shorter time to healing in favor of active dressings (WMD: − 24.38 [− 42.90, − 5.86] days, p = 0.010, I2:100%).

A pre-specified analysis including only trails with a treatment duration of at least 6 months (n = 3 trials [20,21,22]) reported no significant between-group differences in healing rate (Fig. 4S, Panel A) and a nonsignificant trend toward reduction of time-to-healing (Fig. 4S, Panel B).

A further pre-specified analysis excluding trials enrolling infected ulcers [13,14,15, 17, 18, 24], revealed similar favorable effects of advanced wound dressings on both ulcer healing (Fig. 5S, Panel A) and time-to-healing (Fig. 5S, Panel B). A post-hoc analysis comparing AWD with different comparators has been performed in order to verify the possible role of different type of SoC in healing processes. No significant between-group differences were observed, suggesting no effects of different type of SoC in modulating AWD effectiveness (Fig. 6S of Supplementary Materials).

Amputations

Out of seven studies reporting information on major amputations, four reported at least one case with a MH-OR for active dressings of 0.72 [0.17, 2.97], p = 0.58, I2: 0%; Fig. 7S).

Only five trials [17, 18, 21, 22, 25] reported information on minor amputations. Three studies [17, 21, 22] reported at least one minor amputation among either patients allocated to active dressings or SoC/placebo, showing no between-group differences (MH-OR: 1.26 [0.50, 3.15], p = 0.62; I2: 0%; Fig. 8S) (Fig. 2).

All-cause mortality

Only four studies reported at least one event (n = 21 deaths) [7, 18, 25, 26] showing no effects of active dressing on all-cause mortality (MH-OR for active dressings vs SoC/placebo: 0.48 [0.19, 1.26], p = 0.14; I2 = 0%; Fig. 9S).

Other outcomes: back-to-walk, frequency of dressing changes, and pain

No studies reported information on back-to-walk and pain, which were prespecified endpoints. Frequency of dressing changes was reported for all the included studies, the majority of whom left the dressing change to the discretion of the investigators or followed a pre-specified protocol. Only two studies have assessed this parameter [7, 21], the first with sucrose octasulfate and the latter with a silver dressing, with no between group differences observed (WMD: − 0.16 [− 0.49, 0.17] times/week, p = 0.34, I2: 0%; Fig. 10S).

Quality of life

Only two trials reported information on quality of life [7, 21], using two different tools and therefore preventing any formal meta-analysis. Active dressings in both studies was not associated with better quality of life scores in comparison with SoC/placebo.

Quality of retrieved evidence

Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology [10] was used to assess the quality of the body of retrieved evidence. For the primary endpoint, the quality of evidence was rated as “moderate” (Table S4 of Supporting Information) due to the relatively low quality of the included studies (selection and performance bias) and a possible publication bias.

Discussion

Diabetic foot syndrome represents a clinical and economical challenge for any healthcare system. Patients with DFU are at higher risk for local and systemic complications, including infections/sepsis, amputations, need for revascularization, and all-cause mortality. As the prevalence of diabetes and diabetic foot ulcers continue to rise globally [28], the development of innovative and effective therapies, including local dressings, is an essential tool to increase the chance of ulcer healing and reducing the risk of MALE.

The panel responsible for the development of the Italian Guidelines for the treatment of Diabetic Foot Syndrome identified as relevant, a clinical question regarding the use of advanced wound dressings, choosing ulcer healing, time-to-healing, all-cause mortality, major and minor amputation, frequency of dressing changes, pain, and quality of life [9].

In the present systematic review, which was performed to support the development of these guidelines, the primary endpoint was ulcer healing, which were scored as the most important outcome by the guideline panel [9].

The novelty of the present paper consists in collecting all RCT exploring the efficacy of advanced wound dressings on top of standard of care for the treatment of DFU. In fact, with the exception of a retract network meta-analysis comparing nine different dressings, no previous meta-analyses assessed the effectiveness of common local dressings in patients with DFU.

We collected RCT performed on 11 commonly used medical dressings, retrieving few trials for each individual dressing, with the exception of alginates, foams, and microorganism-binding dressings, for whom no trials have been never performed in patients with DFU. Differently from other meta-analyses [29,30,31] which included both randomized and nonrandomized clinical studies, we decided to restrict our analyses only to RCT to reduce the risk of prescription and selection bias and reducing heterogeneity. We have also chosen to exclude bio-engineered dressing, such as growth factors, mesenchymal stem cells, skin grafts/substitutes (considered all together by other meta-analysys [32]) which will be the topic of a further PICO [9], and that are used for recalcitrant ulcers and not for routinely local treatments.

The results of our paper showed that taken overall, advanced dressings are more effective than standard of care (i.e., wet-to-moist dressing, saline-moistened or sterile gauze, etc.) in increasing the rate of ulcer healing and reducing the time-to-healing- However, despite the scarce number of trials prevented any reliable conclusion for any of the explored dressing, some local treatment seemed to show better profiles, such as sucrose-octasulfate, honey dressing, hyaluronic acid, and hydrogel. Hydrocolloid/hydrofibers, hemoglobin spray, protease-modulating matrix, and silver dressings have been less studied and/or did not report any significant advantages over the SoC in term of ulcer healing and/or time-to-healing.

Sucrose-octasulfate has been recently assessed in a relatively large-scale, high-quality trial [7], investigating the effect of sucrose octasulfate dressing, also known as technology lipido-colloid with nano-oligosaccharide factor The rationale of this dressing consists in the inhibition of matrix metalloproteinase (MMPs), which are over produced at the wound bed level, leading to abnormal tissue breakdown and prolonged healing time. The Explorer study [7] showed significant beneficial effects in the treatment of noninfected, neuroischemic, hard-to-heal DFUs on several outcomes, such as healing rate and time-to-healing. Despite, similar biological actions, protease-modulating matrix dressings did not report any advantages over the SoC. However, this dressing has been evaluated in one single, low-quality, short-term trial, preventing therefore any reliable conclusion.

Hydrogels have several remarkable properties, including moisturization, permeability, and the ability to promote autolytic debridement, which could explain the favorable results obtained in the present meta-analysis. Hydrogels, in fact, reproduce the natural environment of the extracellular matrix, providing an ideal setting for cell proliferation [33].

Honey dressings has emerged as a promising therapeutic strategy for treating DFU due to their antibacterial properties and physical–chemical characteristics. In fact, honey exhibits antibacterial activity by releasing hydrogen peroxide and lowering pH wound levels and promotes granulation tissue formation due to high sugar and phenolic contents [29]. Despite similar antibacterial effects, silver dressings did not show any significant beneficial effects on ulcer healing and time-to-healing. These neutral results are somewhat surprising, since in other different settings, such as that of leg ulcers, silver dressings showed favorable effects in term of ulcer healing and time-to-healing [34]. Once again, such neutral results can be due to the paucity of included trials, rather than an ineffective action of these dressings. Similar considerations can be made for other local treatments, such as hydrocolloids/hydrofibers and hemoglobin spray reporting conflicting [21, 22] or neutral [24] results.

Several limitations should be acknowledged and considered in interpreting the obtained results. The quality of a meta-analysis always depends on the quality of included parent studies. The present meta-analysis, unfortunately, included several low-quality, short-term and not adequately sized trials possibly introducing biases and affecting the overall results. The majority of the included studies were open-label and randomization and allocation procedures not always well described. The heterogeneity of inclusion and exclusion criteria, study procedures, and definition of endpoints could have further contributed to introduce another bias (inconsistency bias). However, despite these possible sources of heterogeneity, for the primary outcome, we did not detect any relevant heterogeneity, which was low-moderate. This allows us to consider the obtained results reliable. Another limitation is represented by the adopted selective and restrictive inclusion and exclusion criteria (i.e., only randomized studies, with a minimum duration of at least 12 weeks, and including only diabetic foot ulcers) which allowed us to include a relatively scarce number of trials. This could have reduced the overall heterogeneity, but, at the same time, limited the statistical power for some of the included outcomes, possibly underestimating the effects of some active local dressings on DFU prognosis. Moreover, the included trials enrolled both plantar and nonplantar DFU, which could have different time-to-healing and benefit differently from pressure relief systems. A further limitation is represented by the relatively short duration of the included studies. Only three studies fulfilled the inclusion criteria reported in our previous methodological paper (i.e., studies with a duration of at least 6 moths [9]). The results obtained are therefore coming from post-hoc analyses. These limitations suggest caution in interpreting the overall results of the present meta-analysis and highlight the urgent need for higher quality studies with longer follow-up, enrolling larger cohorts of patients, and taking into account also economic aspects.

In conclusion, advanced wound dressings, particularly sucrose-octasulfate, hydrogels, hyaluronic acid, and honey dressings, can actively promote wound healing and shortening time-to-healing in patients with non-infected, predominantly neuropathic DFU together with SoC (including a multidisciplinary assessment [5], plantar pressure relief [35], and local debridement [36]).