A parallel open‐label trial to evaluate microbial cellulose wound dressing in the treatment of diabetic foot ulcers

Abstract

The purpose of this study was to compare the rate of wound healing in diabetic foot ulcers (DFU) using either a microbial cellulose (MC) wound dressing or Xeroform™ Petrolatum gauze. In a parallel, open‐label trial in which the primary outcome was the rate of wound healing and the time to wound closure, 15 ulcers in type II diabetic patients received an MC dressing. Wounds in 19 control patients with type II diabetes were treated with a Xeroform gauze dressing. All wounds were non infected, Wagner stage II or III and received standard care including debridement, non weight bearing limb support and weekly wound evaluation. The mean time to heal in the MC (±SE) treated group was 32 days ± 2·5 and for controls it was 48 days ± 4·7 (P < 0·01). The rate of weekly wound closure (mean ± SE) was 1·7 times faster in the MC‐treated group (cellulose treated, −5·04% per week ± 0·38 versus control, −2·93% per week ± 0·19), (P < 0·001). Among covariants tested by univariate regression, only the original wound area correlated with the time to wound closure (P < 0·001). In conclusion, with the provision of current standards of care, the application of an MC dressing to a diabetic ulcer may enhance the rate of wound healing and shorten the time course of epithelisation.

INTRODUCTION

Treatment of diabetic foot ulcers (DFU) demands a multidisciplinary approach that includes metabolic control, debridement and off‐loading, control of infection and vascular insufficiency, and a modern and integrated approach to wound care (1). In the absence of effective management of the diabetic patient, a relentless increase in the complications of diabetes will result in a prevalence of DFU of 8–10%, a population‐based incidence of 1·0–4·0% and a lifetime incidence of perhaps 25% (2). In addition, a DFU precedes 85% of non traumatic lower extremity amputations with a significant increase in mortality at 1–5 years (3). Although trauma and dystrophic skin changes that result from the neuropathy of diabetes are the harbingers of perhaps 76% of DFU, this is frequently complicated by tissue hypoxia from vascular disease and infection. Effective treatment mandates appropriate off‐loading, revascularisation and control of infection when appropriate.

At the tissue level endothelial, fibroblastic and inflammatory cell lineage appear to be adversely affected by the diabetic condition and may result in abnormal and prolonged wound healing (4). Current wound management requires debridement of necrotic and inflamed tissue, control of infection and an appropriate wound dressing maintaining a moist environment (5, 6). The moist wound environment is important for fibroblast and keratinocyte proliferation and migration, collagen synthesis and an improved cosmetic effect. The ideal wound dressing should maintain a moist wound environment while absorbing excess exudate, provide a bacterial barrier, reduce pain, provide mechanical stability and be non toxic and biocompatible (7). It should be easy to use and is inexpensive. Recently, we have reported the application of a dry microbial cellulose (MC) membrane, Dermafill™ (AMD Ritmed, Tonawanda, NY, USA), in the management of chronic non healing ulcers of the lower extremity and in the treatment of skin tears in a population of frail elderly living in a skilled nursing facility (8, 9). Attributes of the MC membrane include closure of non healing wounds, pain reduction, requirement of a single application and ease of use without the need for secondary dressings. In this report, we present data on the use of MC dressings in a group of patients with DFU and offer a physicochemical explanation for the efficacy of MC as a wound dressing. We compare the results with a separate control group of DFU treated with a standard wound dressing.

METHODS

Fifteen DFU in 11 diabetic patients were studied in a parallel, open‐label trial using an MC dressing. Controls consisted of 19 diabetic patients with DFU treated with a standard dressing of Xeroform™ Petrolatum gauze. Patients were recruited from a single centre. Inclusion criteria included a non infected diabetic ulcer of the lower extremity and an ankle/brachial index (ABI) of 0·4 or greater. We showed the Wagner criteria for wound assessment (10). Patients with osteomyelitis, heart failure, chemotherapy, dialysis for ESRD or a Wagner grade IV or V wound were excluded. Following the Ethic's Committee approval of the study conducted in compliance with the 1975 Declaration of Helsinki, written informed consent was obtained from all patients.

At the initial visit we assessed the clinical status of the patient and the duration and status of each DFU. Target wounds were photographed and the greatest width and length were measured in centimetres. Periwound skin, the presence or absence of undermining, tunnelling and wound bed characteristics were assessed. Baseline glycated haemoglobin was recorded and oscillometry was performed to assess vascular status. Presence of neuropathy was assessed using the Semmes–Weinstein 5·07/10‐g monofilament. Patients were managed in an outpatient wound clinic.

All wounds were treated using current standards of care including debridement, off‐loading with total contact casting or other non weight bearing devices, or dry compression bandages using Coban™ and Webril™. When indicated, wounds were debrided using sharp surgical debridement followed by cleansing with saline. Dressings were tailored to the size of the wound. After applying the MC membrane to the wound, wrinkles were removed with saline soaked gauze. In the majority of cases only a single application of the MC dressing was required for complete healing. Off‐loading devices or dry compression dressings were applied weekly to all wounds. (MC membranes were supplied by AMD Ritmed, Tonawanda, NY, USA).

Wound area was determined by measuring the greatest length times the greatest width. At weekly evaluation of the wound the condition of the MC membrane, characteristics and size of the wound and compliance with off‐loading were assessed. In controls, Xeroform gauze dressings were changed weekly. Adverse events were recorded. Photographs and measurements of all wounds were taken at each visit. Wound healing was defined as complete closure of the wound surface (11). The time measured in days at which the target wound was fully epithelised was considered the date of the last clinic visit.

STATISTICAL ANALYSIS

The primary outcome events were the rate of relative reduction in the target wound area and the time to complete wound closure. A baseline measurement taken at the first visit was used for the calculation of the percentage reduction in wound area as follows: ({Area_(Day0)–Area_{(Day x)}/Area_(Day0)}× 100). Rate of closure of each wound was computed from the linear regression of percent original area versus the time of treatment. The rate is given as percent reduction in wound area per week. Significant difference in the mean time to complete wound closure between MC membranes and Xeroform gauze‐treated groups was determined by Student's t‐test. Differences in baseline patient characteristics and wound parameters were analysed using one‐way analysis of variance (ANOVA). Comparison in the rate of wound closure and time to heal in different treatment groups was analysed by univariate regression including analysis of covariant effects on outcomes. Two‐tailed significance was set at P < 0·05 (SPSS Software, version 10, Chicago, IL, USA).

RESULTS

In 11 patients (81% men), 15 DFU were managed with an MC dressing. Data were compared with 19 control patients (53% men), with a DFU managed with a standard Xeroform gauze dressing changed weekly. Both groups consisted entirely of type II diabetes and all wounds were either Wagner stage II or III. The topography of DFU was predominantly the plantar surface, dorsum of foot, toes and ankles. All DFU were non infected and showed evidence of granulation tissue. None were considered to be chronic. Following initial evaluation, DFU were covered with either an MC dressing or Xeroform gauze dressing, and off‐loading devices or a dry compression was applied.

Baseline characteristics of the study population are shown in Table 1. The mean age of the MC‐treated group was 55 years and the controls 69 years. The mean (±SE) healing time for the MC dressing‐treated group was 32 ± 2·5 days and for controls 48 ± 4·7 days (P < 0·01). There were no reported adverse events in either group. Patients treated with MC membranes reported less discomfort throughout the trial, in part because of far fewer dressing changes.

Table 1.

Baseline characteristics of the study population

Variables	MC Membrane	Xeroform gauze	P‐value
Number of patients	11	19
Number of DFU	15	19
Age (mean ± SD)	55 ±15	69 ± 12	0·04
Sex (% males)	87%	53%	0·04
Initial wound area (mean ± SE)	3·0 (cm2) ± 0·69	5·0 (cm2) ± 1·28	NS
Duration of wound prior to treatment (weeks)	6 ± 0·9	15 ± 1·5	NS
HbA1c (%)	7·20	6·60	NS
Left ABI (mean ± SD)	0·68 ± 0·22	0·69 ± 0·22	NS
Right ABI (mean ± SD)	0·57 ± 0·39	0·69 ± 0·24	NS
Number with neuropathy (%)	80	50	NS
Number with PAD	40%	74%	0·05
Number with hypertension (%)	53	74	NS

MC, microbial cellulose; DFU, diabetic foot ulcer; HbA1c, haemoglobinA1c; ABI, ankle/brachial index; PAD, peripheral arterial disease.

The change in wound area over time is shown for both treatment groups in Figure 1. These data show a difference in the rate of wound closure between the group treated with MC membranes and Xeroform gauze dressings. There is a decrease in the rate of wound closure in the Xeroform‐treated group which is apparent at about 21 days. This decay in rate of closure is not seen in the MC‐treated group. Themean rate of healing was assessed from the change in initial wound area using linear regression. The rate (mean ± SE) of weekly wound closure was 1·7 times faster in the MC‐treated group (−5·04% per week ±0·38) versus controls (−2·93% per week ±0·19), (P < 0·001). When other covariants (Table 1) were evaluated by univariate regression only original wound area was found to be significantly correlated with both the rate of wound healing and the time to complete wound closure (both P < 0·001).

Figure 1.

Rate of wound closure by treatment group. Wound area was evaluated weekly for each patient. The change in wound area is shown during treatment as the percentage of the original wound area determined at each clinic visit over the treatment period. The data show the mean (±SE) value at each week for wounds treated with either microbial cellulose (MC) dressings or Xeroform Petrolatum gauze.

DISCUSSION

Cellulose is a linear polymer in which β‐1, 4 glycosidic moieties are joined to form cellobiose repeat units. Regardless of the origin of the cellulose, plant or bacterial, the two most important characteristics that affect function include the degree of polymerisation which can span a tenfold difference, and the supramolecular organisation of the chains which can vary from amorphous to crystalline in structure (12, 13, 14). Hydroxyl groups form intra and intermolecular hydrogen bonds which are responsible for its chemical stability, rigidity and tensile strength.

MC is obtained through biosynthesis of the bacteria Glucanacetobacter xylinum, a gram‐negative organism able to convert glucose and other organic substrates into cellulose within a few days (15). By comparison to plant cellulose, MC has greater crystallinity, greater water content and a tensile strength that is almost twice that of plant origin (13). It appears that these attributes allow dry (or wet) membranes of MC to function as efficient dressings in a variety of wounds.

At present, we can only speculate how this material could affect wound closure. The pore size created by the microfibrils and ribbons of the membrane are able to trap platelets and initiate coagulation bringing about both haemostasis and adherence to the wound surface which may also impart pain reduction. The same structure could act as a regenerative tissue scaffold to effect fibroblast, endothelial and keratinocyte function enhancing granulation tissue formation and epithelisation. The greater absorptive capacity of the membranes and its porous nature allow for the dissipation of exudate while maintaining a moist wound interface. Finally, the tensile strength of the material allows it to remain in place for longer periods with little need for dressing change. Alvarez et al. has offered a similar speculation that MC forms a semi‐permeable occlusive barrier that enhances macrophage activity, fibrinolysis and angiogenesis (16).

Recently, we showed that the application of an MC membrane resulted in healing of wounds that had been stalled in excess of 300 days (8). With an MC dressing, 75% reduction in wound area occurred in 81 days, a significant difference from baseline (P < 0·001). In addition to the factors noted above, the dressing might also be affecting the protease imbalance and compromised growth factors in the chronic non healing wound (4).

In the present study of non infected DFU with granulation tissue and standard care defined as debridement, off‐loading or compression, use of a cellulose dressing results in the epithelisation of all wounds and, by comparison to previous reports, appears to enhance the rate of wound closure. The small sample size limits the conclusions that can be drawn. However, in agreement with Margolis et al. we also noted that initial wound size was a predictor of wound healing (17). In agreement with Zimney et al. we found that the fastest healing times occurred in weeks 1–7 (Figure 1) (18). A graph by these authors showing the rate of healing in patients with a neuropathic aetiology [reference 12, Figure 1, p. 330] is remarkably similar to our data for larger wounds and suggests a bimodal distribution. The initial inertia in healing of larger wounds is rapidly replaced by an accelerated phase that leads to complete wound closure. We would suggest that the MC membrane might help overcome this early inertia in the rate of healing.

As our earlier data suggest and other authors have noted, the longer duration of DFU and the presence of limb ischaemia both have a negative impact on healing rates (8, 19). In agreement with Margolis et al., we did not find using univariate regression that patient age or glycated haemoglobin values were predictive of outcome (17).

In summary, MC membranes possess physical and chemical attributes that are both unique and different from plant cellulose and which have proven to be effective in the management of acute and chronic wounds. The wounds presented here would not be considered difficult to manage or chronic because they all began a rapid rate of closure within the first month of treatment and none stalled in reaching complete epithelisation (20). DFU can be added to the list of wounds effectively managed with an MC membrane.