A Borate‐Based Bioactive Glass Advances Wound Healing in Non‐Healing Wagner Grade 1 Diabetic Foot Ulcers: A Randomised Controlled Clinical Trial

ABSTRACT

A novel advanced synthetic bioactive glass matrix was studied in patients with non‐healing diabetic foot ulcers (DFUs). Bioactive glasses can be constructed to be biocompatible, with water‐soluble materials in multiple geometries including fibre scaffolds that mimic the 3D architecture of a fibrin clot. In this trial, chronic, Wagner Grade 1 DFUs were randomised to receive borate‐based bioactive glass Fibre Matrix (BBGFM) plus standard of care (SOC) therapy for 12 weeks or SOC alone. The primary study endpoint was the proportion of subjects that obtained complete wound closure at 12 weeks. Secondary endpoints included time to achieve complete wound closure at 12 weeks. In the modified intent‐to‐treat (mITT) analysis, 48% (32/67) treated with BBGFM plus SOC healed at 12 weeks compared to 24% (16/66) with SOC alone (p = 0.007). In the per protocol (PP) population, 73% (32/44) of subjects treated with BBGFM plus SOC healed versus 42% (16/38) in the SOC group (p = 0.007). Based on the success of this trial, BBGFM demonstrates faster healing of DFUs compared to SOC and should be considered in the treatment armamentarium for Wagner Grade 1 DFUs. Future trials should investigate the use of BBGFM for healing deeper chronic DFUs, other wound aetiologies, or complex surgical wounds.

Summary.

• Mirragen Advanced Wound Matrix (ETS Wound Care; Rolla, Missouri) is a novel, borate‐based, bioactive glass fibre matrix (BBGFM) that has been shown to effectively treat DFUs in previous clinical studies.

• In the modified intent‐to‐treat (mITT) population, 48% (32/67) of subjects treated with BBGFM plus SOC healed at 12 weeks compared to 24% (16/66) in the SOC group (p = 0.007) and the mean time to heal within 12 weeks for the BBGFM plus SOC group was 9.1 weeks (95% CI: 8.1–10.0) versus 10.4 weeks in the SOC group (95% CI: 9.6‐11.1) (adjusted p = 0.042).

• In the per protocol population, 73% (32/44) of subjects treated with BBGFM plus SOC healed at 12 weeks compared to 42% (16/38) in the SOC group (p = 0.007) and the mean time to heal within 12 weeks for the BBGFM plus SOC was 8.2 weeks (95% CI: 7.0–9.4) compared to 9.7 weeks in the SOC group (95% CI: 8.6–10.7) (adjusted p = 0.084 [not statistically significant]).

1. Introduction

Diabetes‐related lower extremity complications (DRLECs), including foot ulceration, infection and amputation, represent some of the greatest challenges faced across the entire healthcare spectrum. The lifetime risk of developing a diabetic foot ulcer (DFU) may exceed 34% [1], with an estimated 18.6 million people around the world developing a DFU every year [2]. As of 2016, DRLECs affected 131 million people, or 1.8% of the global population [3]. DFUs impart a tremendous burden on healthcare systems, patients and their caregivers. The estimated cost for treating one DFU ranges from $11 700 to $16 883 [4]. The 5‐year mortality rate for people with a DFU is 30%, and 70% for those with a major amputation, exceeding many forms of cancer [5].

Greater than half of DFUs become infected, with nearly 20% of these infected wounds requiring amputation [2, 6]. Up to 30% of DFUs progress to chronic wounds despite appropriate standard of care (SOC) therapy [7, 8]. Thus, it is imperative to develop highly efficacious, safe, and cost‐effective wound healing modalities to combat the growing challenges posed by DFUs. This includes products like tissue grafts—also referred to as skin substitutes, cellular and/or tissue‐based products (CTPs), and cellular, acellular and matrix‐like products (CAMPs)—that can potentially facilitate granulation formation, re‐epithelialisation and/or promote other wound healing processes [9].

Bioactive glasses are biocompatible water‐soluble materials that dissolve into their base constituents when immersed in body fluids. Historically, the focus of these materials has been on the development of a scaffold for bone tissue engineering [10, 11, 12]. However, more recently, there has been a growing interest in the wound healing potential of bioactive glass. Specifically, through processes involved with improving angiogenesis and microvessel density [13, 14], increasing metabolic activity and cell recruitment/proliferation within the scaffold microstructure [15] and serving as antimicrobial agents [16, 17].

Boron‐containing compounds have been used extensively in medicine due to several mechanisms of action. These include the inhibition of phosphodiesterase‐4 (PDE4) and reduction of pro‐inflammatory cytokine production [18], broad‐spectrum antifungal activity [19], treatment of various cancers such as non‐Hodgkin's lymphoma and multiple myeloma [20, 21], antimicrobial activity against Gram‐negative bacteria [22] and antiprotozoal activity [23].

In particular, borate has been shown to accelerate wound healing [24], playing a key role in the actions of several enzymes in fibroblasts, including elastase, trypsin‐like enzymes, collagenase and alkaline phosphatase [25]. Further studies have shown that it also regulates mRNA expression of many extracellular‐matrix proteins, including collagen type 1, osteopontin, bone sialoprotein, osteocalcin, alkaline phosphatase and bone morphogenetic proteins [26, 27]. Borate‐based bioactive glass fibres have been shown to stimulate VEGF secretion [28] and promote angiogenesis [29].

Furthermore, the degradation rate of bioactive glass can be altered by changing its composition [13, 30]. Borate‐based bioactive glasses have been formulated to degrade in wounds over a period of days or weeks. Mirragen Advanced Wound Matrix (ETS Wound Care; Rolla, Missouri) is a novel, borate‐based bioactive glass fibre matrix (BBGFM) with a target composition of 53B₂O₃–20CaO–12K₂O–6Na₂O–5MgO–4P₂O₅ wt% (Figures 1, 2). Bioactive glass material can be manufactured on a large scale with a relatively low cost of source materials, which is of great benefit for potential cost savings. The rate of degradation depends in part on the wound exudate. Therefore, the bioactive glass degrades naturally as wound healing occurs. Another key contributing factor to the degradation rate is the geometric shapes within the scaffold. BBGFM features fibres (a few microns in diameter) that typically dissolve in 1–2 weeks, as well as microspheres (several hundred microns in diameter) that persist for 3 weeks or longer.

FIGURE 1.

Gross illustration of BBGFM (Mirragen).

FIGURE 2.

SEM image of BBGFM at 800× magnification, featuring the individual fibres and the microspheres that comprise the scaffold structure.

Recent clinical studies have demonstrated the efficacy of BBGFM in managing chronic wounds [12, 31]. In an initial randomised controlled clinical trial (RCT) conducted by Armstrong et al., chronic DFUs treated with BBGFM demonstrated a statistically significant increase in healing rate at 12 weeks (BBGFM 70% vs. SOC 25%; p = 0.004) and a doubling of the rate of wound area reduction (BBGFM 79% vs. SOC 37%; p = 0.027). The purpose of this study is to build upon those initial findings by evaluating the efficacy and clinical utility of BBGFM in a larger, multi‐centre trial. In doing so, it aims to validate BBGFM as an efficacious solution for the treatment of chronic, Wagner Grade 1 DFUs, positioning this novel synthetic CAMP as a viable alternative to traditional biologic skin substitutes.

There is a significant need for new, synthetic, non‐biologic, and cost‐effective options that can enhance wound healing in these complex clinical scenarios. Moreover, without the concerns often related to biologic products including processing steps, ordering and shipping logistics, product size limitations, storage requirements and now cost implications for providers and patients [32, 33].

2. Materials and Methods

In this multi‐centre, parallel, two‐group, single‐blind randomised RCT, 148 patients with a chronic, full‐thickness, non‐infected, non‐ischemic DFU (University of Texas 1A/Wagner 1) were treated with either SOC therapy or bioactive glass fibre matrix (BBGFM) plus SOC for a period of 12 weeks. Both groups received SOC, which included weekly sharp debridement (removal of all nonviable tissue in the wound bed) as indicated, appropriate offloading with a boot (Foot Defender, Defender Operations, Miami, FL) or total contact casting (TCC) if the subject's foot was too large for a boot, infection management as needed (systemic antibiotics used only in conjunction with debridement), and application of the primary wound therapy followed by a three‐layer, padded dressing. Patients in the active treatment group received weekly application of BBGFM. Patients in the comparator group received a collagen alginate dressing (Fibracol; 3 M, Minneapolis, MN). As BBGFM is completely bioabsorbable, only loose sections from prior applications were removed as needed at subsequent visits. The study protocol, developed and utilised by the senior authors, follows the standard protocols utilised for DFU RCTs over the last two decades and was reviewed and approved by Advarra IRB (No. Pro00065638). Informed consent was signed by all participants. The study, registered at ClinicalTrials.gov (NCT06403605), was conducted under all applicable state and federal regulations in the United States, adhered to Good Clinical Practice guidelines, and was in compliance with the Declaration of Helsinki.

Patients with a chronic, full‐thickness, non‐infected, non‐ischemic, University of Texas 1A/Wagner Grade 1 DFU, located on the foot with at least 50% of the ulcer below the malleolus, a minimum surface area of 1 cm² and a maximum surface area of 20 cm², and present for a minimum of 4 weeks, but no longer than 52 weeks, were eligible to be included in the trial. Patients with osteomyelitis, or wounds that probed to the joint capsule or bone, glycosylated haemoglobin (HbA1c) greater than or equal to 12% taken at or within 3 months of the initial screening visit, and serum creatinine ≥ 3.0 mg/dL within 6 months of randomisation were excluded. The complete list of inclusion and exclusion criteria is listed in Table 1.

TABLE 1.

Inclusion and Exclusion Criteria.

Inclusion criteria

Exclusion criteria

Subjects must be at least 18 years of age or older. Subjects must have a diagnosis of Type 1 or 2 diabetes mellitus. At randomization subjects must have a target diabetic foot ulcer with a minimum surface area of 1.0 cm2 and a maximum surface area of 20.0 cm2 measured post debridement with a photographic planimetry app. The target ulcer must have been present for a minimum of 4 weeks and a maximum of 52 weeks of standard of care prior to the initial screening visit. The target ulcer must be located on the foot with at least 50% of the ulcer below the malleolus. The target ulcer must be full thickness on the foot or ankle that does not probe to bone. Adequate circulation to the affected foot as documented by any of the following methods performed within 3 months of the first screening visit: ○ TCOM ≥ 30 mmHg ○ ABI between 0.7 and 1.3 ○ PVR: Biphasic ○ TBI > 0.6 ○ As an alternative arterial Doppler ultrasound can be performed evaluating for biphasic dorsalis pedis and posterior tibial vessels at the level of the ankle If the subject has two or more ulcers, they must be separated by at least 2 cm. The largest ulcer satisfying the inclusion and exclusion criteria will be designated as the target ulcer. Target ulcers located on the plantar aspect of the foot must be offloaded for at least 14 days prior to randomisation. The subject must consent to using the prescribed off‐loading method for the duration of the study. The subject must agree to attend the weekly study visits required by the protocol. The subject must be willing and able to participate in the informed consent process.

A subject known to have a life expectancy of < 6 months is excluded. If the target ulcer is infected or if there is cellulitis in the surrounding skin, the subject is excluded. Presence of osteomyelitis or exposed bone, probes to bone or joint capsule on investigator's exam or radiographic evidence. A potential subject cannot have an infection in the target ulcer or in a remote location that requires systemic antibiotic therapy. A subject receiving immunosuppressants (including systemic corticosteroids at doses greater than 10 mg of Prednisone per day or equivalent) or cytotoxic chemotherapy is excluded. The topical application of steroids to the ulcer surface within 1 month of initial screening is not permitted. A subject with a previous partial amputation on the affected foot is excluded if the resulting deformity impedes proper offloading of the target ulcer. If a subject has a glycated haemoglobin (HbA1c) greater than or equal to 12% taken at or within 3 months of the initial screening visit he/she is excluded. If a subject has a serum creatinine ≥ 3.0 mg/dL within 6 months of randomization he/she is excluded. The subject is excluded if the surface area measurement of the target ulcer has reduced in size by more than 30% in the 2 weeks prior to the initial screening during the 2‐week screening phase: the 2 weeks from the initial screening visit (SV1) to the TV1/randomisation visit during which time the subject received SOC. A subject with an acute Charcot foot, or an inactive Charcot foot, that impedes proper offloading of the target ulcer is excluded. Women who are pregnant or considering becoming pregnant within the next 6 months are excluded. A potential subject with end stage renal disease requiring dialysis is excluded. A subject who participated in a clinical trial involving treatment with an investigational product within the previous 30 days is excluded. A subject who, in the opinion of the Investigator, has a medical or psychological condition that may interfere with study assessments is excluded. A subject treated with hyperbaric oxygen therapy or a cellular and/or tissue product (CTP) in the 30 days prior to the initial screening visit is excluded.

During the screening phase, all subjects underwent a 2‐week run‐in period. During this time, a series of assessments were performed to determine eligibility, including physical examination, medical history and demographics, concomitant medications, index wound assessment (wound measurement using the eKare inSight System [Fairfax, VA] that is FDA 21 CFR Part 11 compliant as well as digital photographs), probe to bone test (X‐rays and bone biopsy performed if osteomyelitis suspected), HbA1c and serum creatinine blood draws, pregnancy testing in females of childbearing age, Semmes Weinstein 10‐point monofilament exam for peripheral neuropathy, pain assessment and evaluation of circulation utilising any one of the following: transcutaneous oximetry measurement (TCOM ≥ 30 mmHg); ankle brachial index (ABI between 0.7 and 1.3); Pulse Volume Recording (PVR: Biphasic); Toe Brachial Index (TBI > 0.6); arterial Doppler ultrasound evaluating for at minimum biphasic dorsalis pedis and posterior tibial vessels at the level of the ankle.

After obtaining consent, the target ulcer was selected by the Investigator. Each subject could only have one DFU selected as the index ulcer, and if more than one ulcer was present, the largest DFU meeting eligibility criteria was selected. The index ulcer was evaluated for infection, cleaned and debrided as necessary. Subjects treated with SOC for the 2‐week run‐in period were eligible for randomisation once all inclusion and exclusion criteria were met. Subjects were excluded if the surface area measurement of the target ulcer reduced in size by more than 30% during the 2‐week run‐in period.

Following randomization, the treatment phase occurred for up to 12 weeks. Subjects meeting all eligibility requirements were randomised to one of two groups: [1] BBGFM plus SOC or [2] SOC with collagen alginate. Subjects were evaluated on a weekly basis, including weekly assessment of ulcer healing and measurements of ulcer size using digital imaging to measure wound area and provide photos of wounds, including photos of healed wounds. Subjects were seen weekly (±3 days) until the ulcer was healed (and confirmed healed 2 weeks later) or they exited the trial. At each weekly treatment visit (TV) following randomisation, the following assessments and activities were performed: medical history and physical examination changes from the previous visit; changes in concomitant medications and prohibited therapies; vital signs; pain assessment; blood glucose; offloading; Semmes Weinstein 10‐point monofilament exam; recording of adverse effects and adverse events (AE); index ulcer assessments, including signs of clinical infection; ulcer cleaning, debridement of the ulcer if appropriate; digital imaging of the index ulcer; recording of index ulcer measurements and application of appropriate primary dressing. If the index ulcer was noted to be completely epithelialized, the subject was scheduled for a healing confirmation visit 2 weeks later. While primary wound closure and confirmation of wound closure was identified clinically by the on‐site investigator, it was further confirmed by an independent adjudication committee made up of wound care experts. By 6 weeks, if the index ulcer demonstrated less than 50% healing, the subject proceeded to the end of study (EOS) visit and was recorded as a treatment failure for study analysis to allow for the most appropriate and compassionate care for the non‐healing wound. Importantly, for primary endpoint analysis, the treatment failure in this instance was included in the final study results.

The primary study endpoint was the proportion of subjects that obtained complete closure over a 12‐week treatment period. Secondary endpoints included a comparison between groups of: time to achieve complete wound closure of the target ulcer by the end of 12 weeks; percentage wound area reduction from TV1 to TV13 as measured weekly with digital photographic planimetry and physical examination; number and type of treatment emergent adverse events; change in peripheral neuropathy of the target foot as assessed using the standard 10‐point Semmes‐Weinstein monofilament exam; changes in pain of the target ulcer assessed using the numeric pain rating (NPRS) scale; and change in quality of life using the WOUND‐Q assessment.

3. Statistics

Descriptive statistical methods were used to summarise the data from the study, with hypothesis testing performed for primary and secondary endpoints. Summary statistics include the number of subjects (n), mean, median, standard deviation, minimum, and maximum for continuous data and frequencies and percentages for categorical data. All statistical testing was two‐sided and performed using a significance (alpha) level of 0.05. All statistical analyses were conducted using IBM SPSS Statistics (v30).

The populations defined for the analysis included modified intent‐to‐treat (mITT), per protocol (PP), and safety populations. The primary analysis was conducted on the mITT and PP populations. The mITT population consisted of all randomised subjects, excluding: (a) subjects who were randomised but later found to be ineligible for the study because they failed to meet all inclusion and exclusion criteria; (b) subjects randomised but who actually received different treatment than that allocated by randomisation. The PP population consisted of randomised subjects with analysis conducted according to treatment received. The following subjects were excluded: (a) subjects randomised but later found to be ineligible for the study because they failed to meet all inclusion and exclusion criteria; (b) subjects randomised but who actually received different treatment than that allocated by randomisation; (c) subjects who did not complete the study or were lost to follow‐up; and (d) subjects with major protocol violations. The safety population included all subjects who were randomised and received at least one treatment, including subjects who were randomised in error despite not meeting inclusion/exclusion criteria for enrolment in the study.

The original sample size calculations used a general sequential design based on the primary endpoint with provision for an interim analysis when approximately 100 subjects had completed the trial (approximately 42% of subjects). The calculations assumed a sample size of 120 in each treatment group and at the final look achieved 100% power to detect a difference of 0.35 in terms of proportions of wounds healed by 12 weeks between group 1 (Mirragen Advanced Wound Matrix) (proportion 0.70) and the control group (proportion of 0.35) at the 0.049 significance level (alpha) using a two‐sided Z‐Test (Pooled).

A very brief interim analysis was conducted when 76 subjects had completed the trial (very similar to the PP population), which showed an 80% statistical power for the primary endpoint if the sample size was projected to 100 subjects. Based on these results, the trial was stopped early.

Missing complete wound healing endpoint data by 12 weeks was imputed by the analyst for the mITT population by scoring the following situations as not healed: subject died, subject was withdrawn from the study or lost to follow‐up before the index wound healed, or subject experienced an amputation on the index foot that obliterated the index ulcer. Other endpoint data was not imputed with the exception of percent change in the surface area of the index ulcer (PAR) when the wound was healed (set to 100% at the week of initial healing and all subsequent time points).

Summary statistics for complete wound healing at 12 weeks were calculated for each treatment group. Primary endpoint analysis was conducted using Fisher's exact test. A generalised linear model (logit link function) was created to adjust for differences between groups based on available patient and wound‐related variables that could influence wound healing significantly if the Fisher exact tests were statistically significant. The independent binary variable was whether the wound healed or not by 12 weeks. Generalised linear models with additional variables were built using stepwise addition of variables starting with treatment group; model parsimony was checked using stepwise deletion of all available variables. Besides treatment, no other variables were statistically significant, so the final model only included treatment as the independent variable.

Secondary endpoint analysis of time to heal within 12 weeks was conducted using the Kaplan–Meier approach and the log rank test with treatment as the factor. The PAR was calculated using the following formula: ((A ₁–A ₂)/A ₁) × 100 where A ₁ is the baseline area (at randomisation), and A ₂ is the area at the specified time point. Analysis was conducted using linear mixed modelling (LMM; repeated measures) with treatment, study visit, and treatment‐by‐visit as fixed factors, area at randomisation as a covariate, and percentage change as the response. For WOUND Q, the mean change in scores between randomisation and EOS visits was calculated for each treatment group and analysed for each of the four domains (wound concerns, wound drainage, wound smell and effect of social life) using the Mann–Whitney test. Type 1 error control was achieved through the use of the step‐down Holm procedure.

4. Results

In this trial, 148 subjects were randomised at 14 sites between September 2022 and January 2025. Ninety‐one subjects were considered screen failures (37.1%). There were 77 subjects assigned to the BBGFM plus SOC group and 71 subjects assigned to the SOC group. The mITT and PP populations comprised 67 and 44 subjects, respectively, in the BBGFM plus SOC group. While in the SOC cohort, the respective populations were 66 and 38 subjects. A total of 53 subjects were withdrawn from the trial. The majority were withdrawn early due to PAR less than 50% at week 6 (67.9%). In the BBGFM plus SOC group, four subjects were removed due to AEs, two withdrew consent, one subject was lost to follow up, and one subject died. In the SOC group, two subjects were removed due to AEs, three withdrew consent, and four subjects for other reasons. A CONSORT diagram summarising subject disposition is shown in Figure 3.

FIGURE 3.

CONSORT diagram.

For the mITT population, key subject‐related variables were well balanced between treatment groups with the exception of subject age at which first DFU appeared (significantly lower age for SOC compared to BBGFM plus SOC group) and prior DFU count, which was significantly higher in the SOC group compared to the BBGFM plus SOC group (Table 2). In the PP population, key subject‐related variables were well balanced between treatment groups with the exception of HbA1c, which was higher in the BBGFM plus SOC group compared to the SOC group (Table 3). A comparison by treatment group for key wound‐related variables showed that variables were well balanced between treatment groups in the mITT and PP populations (Tables 4, 5).

TABLE 2.

Comparison by treatment group for key subject‐related variables (mITT population).

Variables	BBGFM	SOC	p
Age (years)	60.1 (16.81)	59.1 (14.26)	0.72
Race			0.60
White	57 (85)	60 (91)
Black/African American	7 (10)	4 (6)
Hispanic	2 (3)	2 (3)
Asian	1 (2)	0 (0)
Ethnicity			0.42
Hispanic/Latino	19 (28)	14 (21)
Sex			0.85
Male	47 (70)	48 (73)
Female	20 (30)	18 (27)
BMI	33.9 (8.45)	35.0 (8.98)	0.46
Diabetes duration (years)	17.1 (13.59)	15.9 (9.58)	0.85
	Median: 13.5	Median: 15
	IQR: 20	IQR: 10
Subject age when first DFU appeared (years)	58.2 (11.56)	53.6 (11.60)	0.025
Prior DFU count	3.1 (4.27)	4.4 (4.96)	0.032
	Median: 2	Median: 3
	IQR: 2	IQR: 5
History of DFU recurrence	32 (48)	31 (47)	1.0
Major amputation	2 (3)	5 (8)	0.27
Minor amputations			0.70
0	41 (61)	37 (56)
1	11 (16)	14 (21)
2	5 (8)	7 (11)
≥ 3	10 (15)	8 (12)
Creatinine (mg/dL)	1.09 (0.42)	1.24 (0.54)	0.07
HbA1c
At randomization	7.7 (1.69)	7.4 (1.65)	0.42
At EOS	7.6 (1.73)	7.2 (1.53)	0.07

Note: Figures are counts (percentages) for categorical variables and means (SD) for continuous variables, with medians and IQRs also for non‐normal variables.

Abbreviations: DFU, diabetic foot ulcer; EOS, end of study.

TABLE 3.

Comparison by treatment group for key subject‐related variables (PP population).

Variables	BBGFM	SOC	p
Age (years)	60.2 (16.34)	57.7 (15.60)	0.49
Race			0.51
White	37 (84)	34 (90)
Black/African American	5 (12)	2 (5)
Hispanic	1 (2)	2 (5)
Asian	1 (2)	0 (0)
Ethnicity			1.0
Hispanic/Latino	10 (23)	9 (24)
Sex			1.0
Male	33 (75)	29 (76)
Female	11 (25)	9 (24)
BMI	34.3 (9.06)	36.4 (9.99)	0.31
Diabetes duration (years)	15.7 (9.75)	16.4 (14.18)	0.80
	Median: 14.5	Median: 12.5
	IQR: 10	IQR: 9.30
Subject age when first DFU appeared (years)	58.9 (11.86)	54.4 (11.09)	0.083
Prior DFU count	2.5 (2.95)	3.6 (3.77)	0.27
	Median: 2	Median: 2
	IQR: 3	IQR: 5
History of DFU recurrence	18 (41)	16 (42)	1.0
Major amputation	1 (2)	4 (11)	0.18
Minor amputations			0.22
0	31 (70)	21 (55)
1	6 (14)	10 (26)
2	3 (7)	3 (8)
≥ 3	4 (9)	4 (11)
Creatinine (mg/dL)	1.09 (0.46)	1.19 (0.46)	0.35
HbA1c
At randomization	7.7 (1.61)	7.4 (1.57)	0.17
At EOS	7.7 (1.79)	7.1 (1.60)	0.049

Note: Figures are counts (percentages) for categorical variables and means (SD) for continuous variables, with medians and IQRs also for non‐normal variables.

Abbreviations: DFU, diabetic foot ulcer; EOS, end of study.

TABLE 4.

Wound‐related variables (mITT population).

Variables	BBGFM	SOC	p
Wound area (cm2) a	3.1 (2.81)	3.3 (2.57)	0.16
	Median: 2.1; IQR: 2	Median: 2.9; IQR: 2.2
Wound age (weeks) a	20.9 (13.26)	20.9 (14.82)	0.86
	Median: 16	Median: 15
	IQR: 21	IQR: 19.6
Vertical location			0.67
Plantar	52 (78)	54 (82)
Dorsal	15 (22)	12 (18)
Position			0.61
Medial	35 (52)	37 (57)
Lateral	32 (48)	28 (43)
Anatomical location			0.85
Toe	13 (19)	10 (15)
Forefoot	29 (43)	24 (36)
Midfoot	12 (18)	17 (26)
Hindfoot	2 (4)	2 (3)
Heel	9 (13)	10 (15)
Ankle	2 (3)	3 (5)
Number of debridements	9.0 (3.54)	9.2 (3.34)	0.71

Note: Figures are counts (percentages) for categorical variables and means (SD) for continuous variables (median [IQR]) additionally for non‐normal variables).

^a At randomization.

TABLE 5.

Wound‐related variables (PP population).

Variables	BBGFM	SOC	p
Wound area (cm2) a	2.7 (2.37)	3.1 (2.43)	0.14
	Median: 1.9	Median: 2.9
	IQR: 1.92	IQR: 2.2
Wound age (weeks) a	19.4 (12.99)	21.3 (14.55)	0.48
	Median: 15.1	Median: 16.1
	IQR: 22.5	IQR: 17.8
Vertical location			1.0
Plantar	34 (77)	29 (76)
Dorsal	10 (23)	9 (24)
Position			0.66
Medial	26 (59)	20 (54)
Lateral	18 (41)	17 (46)
Anatomical location			0.58
Toe	13 (30)	6 (16)
Forefoot	18 (41)	15 (39)
Midfoot	5 (11)	7 (18)
Hindfoot	1 (2)	1 (3)
Heel	6 (14)	6 (16)
Ankle	1 (2)	3 (8)
Number of debridements	9.4 (4.09)	9.7 (3.85)	0.71

Note: Figures are counts (percentages) for categorical variables and means (SD) for continuous variables (median [IQR]) additionally for non‐normal variables).

^a At randomization.

For the mITT population, 48% (32/67) of subjects treated with BBGFM plus SOC healed at 12 weeks compared to 24% (16/66) in the SOC group (p = 0.007) (Figure 4). Posterior power calculations showed a power of 83%, which is an adequate value. The mean time to heal within 12 weeks for the BBGFM plus SOC group was 9.1 weeks (95% CI: 8.1–10.0) versus 10.4 weeks in the SOC group (95% CI: 9.6–11.1) (adjusted p = 0.042). The Kaplan–Meier plot comparing healing time between the two groups is shown in Figure 5.

FIGURE 4.

Complete wound healing at 12 weeks (mITT and PP populations). Bar graph illustrating the proportion of subjects achieving complete wound closure by 12 weeks in both the modified intent‐to‐treat (mITT) and per‐protocol (PP) populations. In the mITT population, 48% (32/67) of subjects receiving borate‐based bioactive glass fibre matrix (BBGFM) plus standard of care (SOC) achieved complete healing compared to 24% (16/66) in the SOC‐alone group ( p = 0.007). In the PP population, 73% (32/44) of BBGFM plus SOC‐treated subjects healed versus 42% (16/38) in the SOC group ( p = 0.007).

FIGURE 5.

Kaplan–Meier plot of time to heal within 12 weeks (mITT population). Kaplan–Meier survival curves comparing time to complete wound closure within 12 weeks for subjects in the modified intent‐to‐treat (mITT) population. Subjects treated with BBGFM plus SOC healed faster than those receiving SOC alone. Mean healing times were 63.4 days (95% CI: 56.6–70.1) versus 72.5 days (95% CI: 67.3–77.7), respectively (adjusted p = 0.042).

For the PP population, 73% (32/44) of subjects in the BBGFM plus SOC group healed at 12 weeks versus 42% (16/38) receiving SOC alone (p = 0.007). Posterior power calculations showed an adequate power of 81%. The mean time to heal within 12 weeks for the BBGFM plus SOC group was 8.2 weeks (95% CI: 7.0–9.4) compared to 9.7 weeks in the SOC group (95% CI: 8.6–10.7) (adjusted p = 0.084 [not statistically significant]). The Kaplan–Meier plot comparing healing time between the two groups is shown in Figure 6.

FIGURE 6.

Kaplan–Meier plot of time to heal within 12 weeks (PP population). Kaplan–Meier survival curves comparing time to complete wound closure within 12 weeks for subjects in the per‐protocol (PP) population. BBGFM plus SOC significantly reduced time to healing compared to SOC alone. Mean healing times were 57.5 days (95% CI: 49.3–65.7) versus 67.8 days (95% CI: 60.5–75.2), respectively (adjusted p = 0.084; not statistically significant).

For the other secondary endpoints, including PAR, changes in peripheral neuropathy of the target foot as assessed using the standard 10‐point Semmes‐Weinstein monofilament exam, pain in the target ulcer assessed using the NPRS scale, and quality of life using WOUND‐Q assessment, the differences between the treatment groups were not statistically significant.

For the subjects who were withdrawn at 6 weeks for failing to reach 50% PAR, the median values and PAR range for each treatment group were: BBGFM: median: −35.8%; range: 311.1; min/max: −277.8 to 33.3. SOC: median: −16.2%; range: 166.8; min/max: −123.1 to 43.8.

An AE refers to any undesired medical occurrence in a patient receiving treatment, whether or not it is causally related to the intervention. A serious adverse event (SAE) is one that results in significant outcomes such as hospitalisation, persistent disability, life‐threatening conditions, or death. Within the safety population, which included all randomised patients, there were a total of 70 reported AEs: 47 in the BBGFM plus SOC group and 23 in the SOC group. The AE rates over 12 weeks on a per subject basis were 33% for the BBGFM group and 24% for the SOC group, which were consistent with rates experienced in comorbid DFU patient populations.

Although there were more than twice as many AEs in the BBGFM group compared to the SOC group, the difference was mostly accounted for by AE clusters in three BBGFM‐treated subjects. A comparison of by severity showed similar percentages for various categories (Table 6). In the BBGFM group, there were 10 SAEs originating from six subjects. One subject died at home from underlying comorbidities. With the exception of three SAEs involving cystitis, acute kidney injury and acute toxic encephalopathy from possible drug overdose, the other SAEs involved infection or sepsis with or without osteomyelitis. In the SOC group, there were four SAEs in three subjects: cellulitis in the right leg, anasarca, bilateral pleural infusions and acute respiratory failure.

TABLE 6.

Categorisation of AEs by severity. Figures in parentheses represent percentages for each treatment group.

AE category	BBGFM	SOC a
Severity
Mild	16 (34)	9 (41)
Moderate	25 (53)	11 (50)
Severe	4 (9)	0 ()
Life‐threatening	1 (2)	2 (9)
Fatal	1 (2)	0 ()
Total	47 (100)	22 (100)

^a The severity status of one AE is unknown.

There were no unexpected safety‐related occurrences noted during the study. Of the AEs attributed, or potentially attributed, to the intervention, three within the BBGFM plus SOC group were noted as ‘possibly related’, one related to the secondary dressing, one related to cellulitis of the right foot and one related to infection of the index ulcer. The SOC group had one AE that was ‘definitely related’ to the intervention product and involved an index ulcer infection.

Representative case examples of enrolled patients are shown in Figures 7, 8, 9.

FIGURE 7.

Case example 1 is a 75‐year‐old male with a chronic, plantar midfoot DFU. The baseline index wound area was 1.1 cm². At screening, creatinine: 1.07 mg/dL and HbA1c: 6.7%. Ulcer age at screening was 20 weeks. The wound healed at 6 weeks.

FIGURE 8.

Case example 2 is a 65‐year‐old male with a chronic, plantar forefoot DFU. The baseline index wound area was 1.5 cm². At screening, creatinine: 0.88 mg/dL and HbA1c: 6.0%. The ulcer age at screening was 4.5 weeks. The wound healed at 6 weeks.

FIGURE 9.

Case example 3 is a 58‐year‐old male with a chronic, plantar‐lateral DFU and prior forefoot amputation. Healed. The baseline index wound area was 6.2 cm². At screening, creatinine: 2.84 mg/dL and HbA1c: 8.5%. The ulcer age at screening was 29 weeks. The wound healed at 10 weeks.

5. Discussion

In the current trial, there was a higher proportion of wounds healed at 12 weeks in subjects receiving BBGFM plus SOC (48%) compared to SOC alone (24%) in the mITT population. In addition, the mean healing time within 12 weeks for the BBGFM plus SOC group was 9.1 weeks versus 10.4 weeks in the SOC group and was statistically significant (p = 0.042). The proportion of wounds completely healed within 12 weeks is consistent with the preceding study conducted by Armstrong and colleagues, in which there was a 2.8‐fold increase in healing rate at 12 weeks among DFUs treated with BBGFM [12]. The significantly increased healing rates and rapid wound area reduction demonstrate that BBGFM can promote wound healing in multiple phases, particularly granulation tissue formation and re‐epithelialisation in the proliferative phase. BBGFM is conformable to all wound shapes and remains in place in deep and tunnelling wounds. It can be stored at room temperature for up to 5 years.

The BBGFM scaffold closely resembles the architecture of a fibrin clot in terms of microstructure and porosity. This scaffold allows for infiltration and proliferation of native cells and maintains sufficient space for native collagen deposition and blood vessel formation. As the matrix dissolves into its base constituents such as boron and calcium, a further environmental effect in the wound bed occurs, stimulating critical processes like angiogenesis, which may be facilitated via upregulation of VEGF due to the addition of bioactive glass fibres [13, 34]. Borate‐based glass fibres, in particular, have been shown in preclinical models to stimulate greater VEGF secretion from human fibroblast cells [28]. Modulation of the inflammatory response by borate‐based glass fibres increases wound healing with less scarring [13, 35]. In fact, a recent case series demonstrated accelerated wound healing and decreased scar formation in complex wounds treated with BBGFM [31].

Through a variety of mechanisms, BBGFM disrupts the integrity of the cell membrane of certain gram‐positive bacteria, such as MRSA, and inhibits the growth of gram‐negative bacteria, including Escherichia coli, Shigella sonnei, Vibrio natriegens, Staphylococcus epidermidis and Serratia marcescens [36]. Boron‐based bioactive glass fibrous matrix has been shown to disrupt biofilm in vitro, a critical impediment to healing wounds [15, 16]. The antibacterial activity of boron can be further attributed to its reaction with water that increases free radicals, such as hydroxyl radicals, and damages the bacterial cell membrane [37, 38, 39, 40]. A recent study demonstrated a sustained antimicrobial effect of BBGFM and significant reduction in pathogens common in DFUs [41]. Most CTPs do not contain antimicrobials, although some synthetically add antimicrobials to their products. By contrast, the antimicrobial properties possessed by BBGFM are intrinsic, potentially mitigating production costs, allergic reactions and bacterial resistance [31].

It is well recognised that many chronic wounds have some level of bacterial contamination and/or biofilm formation. While this can certainly impact the wound healing process, it does not create the clinical signs and symptoms associated with infection and has more recently been classified as chronic inhibitory bacterial load (CIBL) by Armstrong and colleagues [42, 43]. Based on published case series and scientific studies in the literature, there is data to suggest that BBGFM has a positive impact against wound relevant pathogens such that the presence of CIBL is not prohibitive of wound healing progress with treatments [12, 31, 43]. This effect by BBGFM on the wound healing environment and on a broad spectrum of potential bacterial and fungal species could explain the clinical results in previously published series and the relatively low risk of clinical infections with treatment [12, 31, 43]. Moreover, in this trial, uninfected Wagner Grade 1 DFUs were included because the aim was to compare BBGFM to SOC. If either group had a statistically different number of infected ulcers, the results could be skewed.

The strengths of the current study include a robust trial design, appropriate screening procedures, multiple investigational sites, a standardised approach to SOC, ITT analysis and appropriate adjustments for multiple statistical testing. Regarding weaknesses, blinding of the investigator was not possible due to lack of a sham product, and subjects not demonstrating sufficient healing by 6 weeks were withdrawn to provide compassionate care. Importantly, for subjects withdrawn due to lack of healing by 6 weeks, regardless of cohort, they were designated as a treatment failure and included in the analysis. This protocol is consistent with previously validated protocols published in RCTs studying other CTPs and skin‐substitutes [10, 44, 45]. However, since this cohort comprised 36 subjects (24% of the mITT population), it is unknown how many of these wounds might have healed within the 12‐week timeframe had they not been withdrawn. Additionally, multiple study sites may have led to heterogeneity across practices in some of the secondary endpoint analyses such as the monofilament test. Based on the success of this study, future trials should investigate the use of BBGFM for healing chronic DFUs with greater depth and complexity, including wounds that probe to tendon, capsule, or bone, and other chronic wound aetiologies like VLUs and complex surgical wounds.

This study represents a larger randomised controlled trial demonstrating the significant clinical impact of BBGFM in DFU treatment. The results of this 148‐patient trial further support the successful clinical results published in a previous 40‐patient RCT, as well as additional clinical case studies and case series in complex, atypical wounds and refractory wounds [12, 31].

6. Conclusion

Every 1.2 s around the world, a DFU occurs, and every 20 s a limb is amputated [46]. The increasing prevalence, cost and complexity of the diabetic foot, and challenges associated with healing DFUs demand safe and effective wound healing modalities. In this study, BBGFM demonstrated significantly improved wound healing at 12 weeks in Wagner Grade 1 DFUs compared to SOC, consistent with a previous pilot study and should be considered in the treatment arsenal for these types of chronic ulcers.

Conflicts of Interest

David Armstrong, DPM, MD, PhD received research funds from PERI to design and administrate the study and also assist with the writing and review of the manuscript. Dennis Orgill, MD, PhD received research funds to serve as a plastic surgeon to review study photos and assist with the writing and review of the manuscript. Robert Galiano, MD received research funds to serve as a plastic surgeon to review study photos and assist with the writing and review of the manuscript. John Lantis, MD received research funds to assist in review of the final data analysis and writing and review of the manuscript. Paul Glat, MD received research funds to serve as a plastic surgeon to review study photos and assist with the design of the protocol. Marcus Gitterle, MD received funds for speaking at a medical education forum at a national health care conference for ETS Wound Care. Marissa Carter, PhD received research funds to provide assistance with study design as well as statistical analysis plan, and provide the statistical analysis for this trial and assist with writing of the result section of the manuscript. Nathan Young DPM has nothing to disclose. Charles M. Zelen, DPM is the medical director of the PERI and his company received research funds to administrate the clinical trial and write the paper for publication. There are no other conflicts of interest with any of the authors in relationship to this study, or with regard to ETS Wound Care IRB conflicts of interest statements are on file with PERI.

Supporting information

Data S1: Supporting information.

Armstrong D. G., Orgill D. P., Galiano R. D., et al., “A Borate‐Based Bioactive Glass Advances Wound Healing in Non‐Healing Wagner Grade 1 Diabetic Foot Ulcers: A Randomised Controlled Clinical Trial,” International Wound Journal 22, no. 10 (2025): e70763, 10.1111/iwj.70763.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.