Localised insulin administration for wound healing in non-diabetic adults: A systematic review and meta-analysis of randomised controlled trials

Insulin has the potential to restore damaged skin and due to its affordability and global availability, it is an agent of interest when it comes to pioneering new remedies to accelerate wound healing. The aim of this study was to explore the efficacy and safety of localised insulin administration on wound healing in non-diabetic adults. Studies were systematically searched, using the electronic databases Embase, Ovid MEDLINE and PubMed, screened, and extracted by two independent reviewers. A total of seven randomised controlled trials that met the inclusion criteria were analysed. Risk of bias was assessed using the Revised Cochrane Risk-of-Bias Tool for Randomised Trials and a meta-analysis was performed. The primary outcome, which explored rate of wound healing (mm 2 /day), concluded that there was an overall significant mean improvement in the insulin treated group (IV = 11.84; 95% CI: 0.64 – 23.04; p = 0.04; I 2 = 97%) compared to the control group. Secondary outcomes concluded that there is no statistical difference between the healing time (days) of the wound (IV = (cid:1) 5.40; 95% CI: (cid:1) 11.28 to 0.48; p = 0.07; I 2 = 89%); there is a significant reduction in wound area in the insulin group; no adverse events were noted with the administration of localised insulin; quality of life improves drastically as the wound heals, irrespective of insulin. We conclude that although the study showed an improved wound healing rate, other parameters were not statistically significant. Therefore, larger prospective studies are warranted to fully explore the effects of insulin on different wounds, where an appropriate insulin regime can be developed for clinical practice.

individuals in the United Kingdom alone, where more than £5.3 billion is spent annually on treatment and management of associated complications 2 ; this highlight how wounds can impose a significant burden not only to individuals, but to the healthcare system as a whole.
Wound healing is a complex biological process that involves clot formation, inflammation, granulation tissue development and remodelling. 3 This can be influenced by various agents such as insulin-like growth factor 1 (IGF-1) and vascular endothelial growth factor (VEGF). 4 A variety of therapeutic methods are available to accelerate wound healing including skin grafts, hydrocolloid dressings, hyperbaric oxygen therapy, skin substitutes and negative pressure wound healing. 5,6 However, these can have associated complications and may not be feasible for certain patients due to wound-type, preference or expense. In the early stages of healing, the crucial process of reepithelialisation takes place-this involves proliferation, migration and differentiation of keratinocytes from the wound margins. 3,7 Research suggests that insulin has the potential to enhance many of these processes, as well as increase blood flow and promote granulation tissue regeneration that would contribute to wound healing. [8][9][10] Insulin is a well-known peptide hormone and growth factor that has the potential to restore damaged skin. 10 It has the ability to reduce inflammation by changing the expression of pro-and anti-inflammatory cytokines in the body. 11 This includes the activation of antiinflammatory cytokines, including interleukin (IL)-10, IL-4 and VEGF, which inhibits cell apoptosis whilst promoting cell proliferation. 11,12 It also suppresses the protein transcription factor nuclear factor kappa beta (NFkβ) P50/P65, which decreases the expression of proinflammatory markers IL-6, IL-12 and tumour necrosis factor alpha (TNF-α). 11 This in turn accelerates regeneration and healing. Research show that insulin can also impact glucose metabolism, protein biosynthesis and lipid biosynthesis, which can promote wound healing. 11 Due to its affordability and global availability, 13 insulin is an agent of interest when it comes to pioneering new remedies to accelerate wound healing. Systemic insulin treatment has presented to be effective but has drawbacks of inducing hypoglycaemia and hypokalaemia. 14 However, the limited research on localised insulin treatment has shown to overcome this and is a promising future therapeutic for the treatment of wounds. 15 The aim of this systematic review and meta-analysis was to explore the efficacy and safety of localised insulin application on acute and chronic wounds, and its ability to promote wound healing.
Although some studies have previously addressed this, including a review by Sridharan and Sivaramakrishnan, 15 these usually included diabetic patients. This study specifically focused on non-diabetic adults, since there is a lack of evidence exploring this population despite prevalence of different kinds of wounds in this group.

| Literature search
The search strategies recommended by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting guidelines 16 were used for this review. Studies were searched systematically from inception to 10 July 2022 by two reviewers (Zunira Areeba Bhuiyan and Zubair Ahmed) using the Embase, Ovid MEDLINE and PubMed electronic databases. A filter was applied to the PubMed database to narrow search results to include only clinical trials and randomised controlled trials (RCTs). Key terms used for this search included: 'insulin', 'adult', 'human' AND 'wound healing OR ulcer healing' where Boolean operators were utilised (full search strategy available in Supporting Information Table S1). Referencing lists of reviewed articles were further screened for additional relevant studies.

| Eligibility criteria
The inclusion and exclusion criteria for this review is provided in Supporting Information Table S2. No limitations were set with regard to publication date, where all studies published before the date of search (10 July 2022) were eligible for inclusion. However, non-English studies were excluded. Studies were limited to RCTs only, as these are classed as the highest level of primary research in the hierarchy of evidence. 17 Any conflicts on eligibility criteria were resolved through discussion with the senior author (Zubair Ahmed).

| Data extraction
Studies retrieved by the literature search were initially screened using their title and abstract by two independent reviewers (Zunira Areeba Bhuiyan and Oluwasemilore Adebayo). Subsequently, relevant studies underwent full-text analysis to determine eligibility, where any disagreements were resolved through discussion with the senior author (Zubair Ahmed). Data extraction was then performed, where the following information were extracted onto a spreadsheet: study characteristics (first author, year, title, country of study, inclusion and exclusion criteria, intervention and control protocols, outcome measures and results), population characteristics (sample size, age, gender, wound location [upper or lower limb]) and outcome measures. Where information was unavailable, the tables were left blank as no assumptions were made. The primary outcome extracted for this review was the rate of healing of the wound (mm 2 /day). Secondary outcomes extracted include healing time (days), reduction in wound area (cm 2 ), safety evaluation and adverse effects of insulin and quality of life (QoL). Trials with multiple treatment arms were included and analysed in our systematic review. However, data from the best treatment effects in these trials were used to compare studies for the meta-analysis.

| Risk of bias
The risk of bias of the included studies were assessed using the Revised Cochrane Risk-of-Bias Tool for Randomised Trials (RoB 2) 18 by two independent reviewers (Zunira Areeba Bhuiyan and Oluwasemilore Adebayo). Five bias domains were considered including: bias arising from the randomisation process, bias due to deviations from intended interventions, bias due to missing outcome data, bias in measurement of the outcome and bias in selection of the reported results. Risk-of-bias judgements were scored as either 'low', 'some concerns' or 'high' for each domain. The overall risk of bias was also determined where individual scores of the domains were considered. Any disagreements were resolved through discussion with the senior author (Zubair Ahmed).

| Study selection
The literature search initially yielded 518 studies with an additional two studies identified through other sources (i.e., reference lists).
Overall, 347 studies were extracted for screening after removal of duplicate studies. Then, 329 records were excluded following the initial title and abstract screening having applied the proposed inclusion and exclusion criteria. A further 11 studies were excluded after a fulltext analysis, with reasons provided on the PRISMA flow chart in Figure 1. Subsequently, a total of seven studies were included in the qualitative and quantitative synthesis of data in this review.
A RCT by Martínez-Jiménez et al., 19 which explored the use of insulin on wound healing in non-diabetic patients, was excluded from being selected in this systematic review. This is because the control and insulin intervention were applied in different areas of the same wound of individual patients. Therefore, the effects of the control and insulin intervention may be compromised, and interpretation of the data may be biased.

| Study characteristics
The seven RCTs included in this review were published between August 2009 and November 2021. [20][21][22][23][24][25][26] These were undertaken in either of the following countries: China, Egypt, India or Iran. Four of the seven studies had a two-armed RCT design, where the therapeutic group explored insulin as a treatment for wound healing and the control group had a standard saline placebo. 20

| Patient characteristics
Various information on patient demographics were provided in all of the included studies, which are detailed in Table 2 Five studies assessed the primary outcome of exploring the rate of healing in wounds due to insulin intervention. 20,21,23,25,26 However, only four were included in the quantitative analysis as shown in Table 3. 20,21,23,25 This was due to the fifth study using a different parameter to explore the rate of healing-percentage at Days 7 and 14 instead of mm 2 /day. 26 Despite omission, the study showed a statistically significant difference favouring the insulin groups, where p < 0.001 at both Days 7 and 14. Table 3 displays the mean rate of healing for each study with the standard deviation and p-value.

| Results of individual studies-Primary outcome
Three studies reported wound healing rate in mm 2 /day, 20,23,25 whilst one study reported the rate in cm 2 /week. 21 The mean rate of wound healing ranged from 46.1-53.5 mm 2 /day, with one study reporting healing rate of <1 mm 2 /day. 21 To assess the strength of evidence present, a meta-analysis with these four studies ( Figure 2) showed an overall significant weighted mean improvement of rate of healing by 11.84 mm 2 /day (95% CI: 0.64-23.04; p = 0.04; I 2 = 97%) favouring the insulin treatment group.

| Results of individual studies-Secondary outcomes
Five of the included studies explored the time taken for wounds to heal, measured in days. These data have been collated in Table 4, where three studies concluded a significant difference found in the insulin group compared to the control group. 21 as these expressed the biggest difference in results.
A meta-analysis exploring healing time was carried out using these five studies, as expressed in the forest plot in Figure 3.

| Risk of bias in studies
The risk of bias of the seven included RCTs are summarised in    23 This observation further supports this theory, but additional research is warranted to further explore this concept and correlation.
Other secondary findings of this review concluded that insulin administration is a potentially safe procedure with no adverse events or side-effects reported. However, the absence of reported risks does not necessarily imply that insulin administration is safe and further high-quality studies are required to reach definitive conclusions.
Nonetheless, low levels of insulin administration does minimise the risk of wound infections and improves healing. 24 This finding is also reported in another study where, compared to normal saline, topical insulin increased bacterial clearance rate in wounds, thereby reducing the chance of wound infection. 30 Studies show that sustained skin wounds are likely to become contaminated with bacteria that can encourage unfavourable complications; thus, it is important to prioritise methods to reduce this. 7,13 Insulin is therefore a potential intervention to explore, as it can directly combat infections and reduce complication rates.
Another outcome studied in this systematic review is QoL, which drastically improved with wound healing irrespective of intervention (insulin or control). 21,25 It can be argued that as the rate of healing in the insulin groups are relatively faster, it may have an accelerated impact on QoL and hence can be deemed as the more favourable treatment.

| Role of insulin
Insulin is a peptide hormone that has the ability to alter inflammation by promoting anti-inflammatory cytokines (such as IL-10, IL-4 and VEGF) as well as decreasing expression of pro-inflammatory cytokines (such as IL-6, IL-1β, IL-12 and TNF-α) via suppression of NFkβ P50/P65. 11 Research shows that NFkβ P50/P65 inactivation by insulin can also induce glucose uptake by cells and store it as glycogen. 11 This reduces the risk of a hyperglycaemic environment, which is proinflammatory and correlates to vascular damage and reduced tissue oxygenation. 11,31 This in turn reduces inflammation and promotes wound healing.
Insulin is also shown to stimulate lipogenesis (fatty acid synthesis), which can further supress TNF-α-mediated inflammation. 11,30 Additionally, protein synthesis is stimulated by insulin that promotes cell growth and differentiation, whilst proteolysis is inhibited via inactivation of the Fork head box protein O1 (FOXO) pathway. 11 This contributes to the acceleration of wound healing.

| Previous studies
It is well established that insulin has beneficial effects on wound healing, as well as its role in regulating glucose, lipids and proteins. 32 Most of these previous studies, however, have focussed on wounds in experimental animal studies and, more recently, diabetic patients. A limited number of primary research has been undertaken targeting the non-diabetic population, where this systematic review is the first compilation of RCTs exploring this group.
In 2017, a similar systematic review and meta-analysis was undertaken by Sridharan and Sivaramakrishnan, where the efficacy of topical insulin in wound healing was explored (inclusive of diabetic and non-diabetic patients). 15 Overall, findings concluded that no significant differences were observed in the healing rate between the study groups ( p = 0.96). A sub-group analysis for the non-diabetic population was also undertaken (three studies were included), 23  However, both used 10 units (0.1 mL) of insulin as their intervention, which is considerably higher than the noted optimum doses mentioned. This is due to the differences in insulin administration in these studies (insulin spray and insulin-soaked sterile gauze, respectively, vs. insulin injection in the prior studies). Hence, as previously stated, it is a priority to explore the different routes of administration, so optimum doses and their efficacy can be assessed with each procedure. It may also be worthwhile to investigate the use of different agents alongside insulin, such as zinc as reviewed in the study by Attia et al., 25 where it may be regarded as more effective than insulin therapy alone.

| Clinical implications
IGF-1 is a hormone that has a similar molecular structure to insulin where it can bind to the insulin receptor. 23,34 This potentially allows it to be an agent of interest when exploring wound healing therapeutics. However, studies show that IGF-1 activates the insulin receptor at approximately 0.1 times the potency of insulin, hence is not as effective. 35,36 Nevertheless, insulin has its own benefits, where it is deemed as a cost-effective drug, which is globally available in large-scale production. 13 Cost of other growth factors currently used in wound healing (e.g., epidermal growth factor, transforming growth factor beta and platelet derived growth factor, etc.) can range from 1500 to 10,000 USD per mg, while insulin is a much cheaper alternative. 13 Furthermore, research on insulin has been undertaken for decades, where its mechanism, safety and long-term consequences are relatively well understood.
It is important to consider the characteristics of different types and severity of wounds and appreciate that the differences in their pathophysiology and anatomy may affect wound healing time and rate. For example, superficial burns are predicted to heal faster than deep burns. This review, however, did not acknowledge these differences when analysing the data for the meta-analysis. This was primarily due to the limited number of studies available; further research is warranted to explore insulin therapy for specific wound types, so an insulin regime can be aptly tailored clinically for each wound.
Overall, the clinical usage of insulin for wound healing appears to be a promising concept. It has the potential to be an appropriate alternative or addition to many of the therapies available in practice today.
However, despite supporting literature, it has been challenging to standardise insulin therapy for practical application. 19 A better understanding of its mechanism of action in the context of wound healing, appropriate dosage and optimal route of delivery is required. Largescale, prospective RCTs inclusive of different types of wounds are warranted to explore this, where future policies can be made to allow insulin therapy to be clinically available for wound healing.

| Risk of bias
The leading bias observed in the RCTs using the RoB 2 tool were the bias arising from the randomisation process in five of the included studies. This was primarily due to a lack of: (i) randomisation in the allocation sequence, and/or (ii) concealment of allocation sequence until participants were enrolled and assigned to interventions. To overcome the former, details of the randomisation process are warranted, where allocation methods such as repeated coin-tossing, throwing dice, shuffling cards or using computer-generated random numbers would suffice as a successful randomisation. For the latter, a form of remote or centrally administered method to allocate interventions to participants is needed, where the process of allocation is controlled by an external unit or organisation that is independent of the enrolment personnel. This could be done by sequentially numbered drug containers (with insulin) prepared by an independent pharmacy, or using sequentially numbered, opaquely sealed envelopes to conceal allocation.

| Limitations
Limitations of this review are mostly due to the restricted number of

| CONCLUSION
In conclusion, our systematic review and meta-analysis shows that localised insulin administration in non-diabetic adults was beneficial for wound healing, with an improvement in the rate of healing, reduction in wound area and QoL. The use of insulin was also observed to have no adverse events or side-effects. Furthermore, high-quality RCTs are warranted to fully explore the effects of localised insulin on different types and severity of wounds, also considering factors such as age, gender, cause, location and classification of wounds. This would further support the development of an appropriate insulin regime, considering optimum dosage and appropriate route of administration, in the clinical setting.