Topical Oxygen Therapy Shifts Microbiome Dynamics in Chronic Diabetic Foot Ulcers

Introduction. Bacterial biofilm in wounds prevents healing by acting as a physical barrier to wound closure and hyperactivating local inflammatory processes, thus making its removal a high priority. The authors previously have shown that adding topical oxygen to standard wound care increased healing of Texas Grade II and III diabetic foot ulcers (DFUs), which they hypothesized was a result of alterations of the wound microbiome/biofilm. Objective. This study aims to determine the mechanism of action of topical oxygen in DFUs by examining the diversity of bacterial genera present in DFUs treated with topical oxygen. Materials and Methods. Six patients with chronic DFUs had their wounds swabbed weekly over an 8-week period of continuous topical oxygen treatment, and microbiome diversity was assessed by metagenomic 16S rDNA sequencing using a next-generation sequencing platform. Results. The wound microbiome shifted toward a diverse flora dominated by aerobes and facultative anaerobes with oxygen therapy in 5 healed wounds. In contrast, anaerobic flora persisted in a single nonhealing ulcer in the present study cohort. Conclusions. Although the sample size was small, this study suggests topical oxygen therapy may have the ability to encourage the growth of aerobic members of the wound microbiome and be an effective alternative to antibiotics in this area. 

Human skin is normally colonized by a consortia of bacteria, fungi, and viruses (together referred to as the skin microbiome), which occupies this niche to prevent colonization of the skin by more pathogenic bacteria.¹ When a break in the skin occurs, members of the skin microbiome may colonize the wound as part of a developing wound biofilm. Although the clinical diagnosis of wound biofilm remains controversial, it is now recognized that biofilms are present in 60% to 100% of chronic wounds,^2-4 where they form a physical barrier to wound closure and maintain a state of chronic inflammation that can damage surrounding tissue.⁴ Biofilm removal is therefore a therapeutic priority, but treatment options are limited due to the fact that biofilms are intrinsically tolerant to antibiotics.⁵ The diabetic foot ulcer (DFU) microbiome is enriched in anaerobic bacteria when compared with other wounds⁶ and harbors a greater number of opportunistic pathogens when compared with contralateral intact skin.⁷ The abundance of anaerobes may be due to host factors (eg, poor tissue perfusion) but also may be inherent to the biofilm mode of growth. 

Recognizing deep, 3-dimensional wounds are often anaerobic, the authors recently introduced a topical oxygen delivery device into their wound care practice with excellent clinical results.⁸ The present study evaluated whether topicaloxygentherapy (TOT) causes shifts in the wound microbiome from anaerobic species toward an aerobic community, thereby favoring healing. Given that the wound microbiome represents the building blocks of biofilm formation, understanding its composition and response to treatment is the first step toward improving treatment of biofilms in chronic wounds.

This study was approved by the Research Ethics Board at St. Michael’s Hospital in Toronto, Canada. Following informed consent, 6 patients with diabetes and DFUs present for more than 4 weeks in duration were recruited in accordance with the inclusion and exclusion criteria listed in Table 1. 

Patients were assessed weekly and received best practice standard care tailored to their wound (ie, iodine-based dressings, sharp debridement, and offloading as needed) as well as continuous topicaloxygen via the NATROX Oxygen Delivery System (ODS; Inotec AMD, Cambridge, UK). This small, battery-operated oxygen generation device delivers a continuous stream of humidified oxygen to the wound bed via a small tube and sterile adhesive pad worn under the dressing. Patients wore the ODS device 24 hours per day for the full duration of the 8-week study so that their DFUs were continuously exposed to higher than ambient concentrations of oxygen. Ulcers were photographed and swabbed using Levine’s technique at each weekly visit. 

Genomic DNA was extracted from wound swabs as described by Stearns et al,⁹ and the variable region 3 of the 16S rRNA gene was amplified.¹⁰ Positive amplicons were normalized using the SequalPrep Normalization Plate Kit (Thermo Fisher Scientific, Waltham, WA) and sequenced on a MiSeq System (Illumina, San Diego, CA). Reads were filtered and trimmed using Cutadapt¹¹ and processed using DADA2.¹² Taxonomy was assigned using the SILVA database version 1.3.2 (Max Planck Institute for Marine Microbiology and Jacobs University, Bremen, Germany).

Outcomes of DFUs
Of the 6 DFUs, 5 healed over 8 weeks of TOT (Figure 1). The DFU that did not heal in the cohort was 1 of 2 ulcers that was significantly larger than the others at the outset (Figure 2). Clinically, the nonhealed ulcer (patient 2; ~2 cm³) was on the plantar surface of the first metatarsal head and had been wrapped in a compression dressing with acetic acid washes for suspected Pseudomonas aeruginosa for the 8-week study period (Figure 1B). The largest healed ulcer (patient 3; ~2.4 cm³) in the study cohort was on the first digit (hallux tip) (Figure 1B), which was offloaded using a DARCO shoe (DARCO International, Inc, Huntington, WV) and intermittently cleansed with povidone-iodine when deemed clinically necessary. Despite being larger than patient 2’s DFU at the outset of treatment, patient 3 healed completely by week 7 of TOT. The clinical course of all 6 wounds in the current study is presented in Figure 1. 

Microbiome results 
Sequencing identified typical DFU pathogens (ie, Staphylococcus, Streptococcus, Pseudomonas) as well as some less commonly encountered genera (ie, Anaerococcus, Helcococcus, Mycoplasma, Haemophilus) in all ulcers in the present study (Figure 3). There was no obvious association between the presence of a specific pathogen and the duration/severity of an ulcer. In keeping with the results of Loesche et al,¹³ it was found that the microbiome of the 5 healing DFUs in the present study cohort diversified and transitioned over time (Figure 3, Figure 4). In contrast, the microbiome of the 1 nonhealing ulcer in the present cohort (patient 2) was the least diverse and did not change over time despite standard care plus TOT (Figure 3, Figure 4). The most abundant 1% of organisms identified according to oxygen utilization were classified (Table 2), and their abundance over the course of treatment was plotted (Figure 2, Figure 5). Averaged results from all 6 ulcers demonstrated a transition from a flora dominated by anaerobes at the outset of monitoring to a flora rich in aerobic species after 8 weeks of continuous (24 hours/day) topical oxygen. In contrast, the wound microbiome of the nonhealed DFU remained anaerobic, with the exception of a timepoint in which the patient missed their scheduled acetic acid wash (Figure 2). Although it is recognized this is a single data point, it is speculated acetic acid washes may have sterilized the commensal aerobic flora that were able to proliferate in other patients’ healed ulcers.

This study, to the best of the authors’ knowledge, is the first to document changes to a chronic wound microbiome with TOT. However, the sample size was limited to 6 patients due to the longitudinal nature of the study. Because of the small sample size, the authors were unable to determine keystone species common to healing versus nonhealing wounds. The authors are currently conducting a similar, larger trial to map changes in the microbiome on a finer scale and calculate the statistical significance of the findings. 

Topicaloxygen therapy can enrich aerobes to drive the microbiome toward a flora that more closely resembles that of intact skin. These changes may contribute to healing through suppression of dysregulated inflammatory cascades induced by chronic wound colonization. Understanding the mechanism-of-action of topical oxygen may help in determining which patients will benefit most from its use and guide treatment decisions by distinguishing responders from non-responders earlier. 

Note: The authors wish to thank Laura Rossi and J.C. Szamosi for their expert technical assistance.

Authors: Paul Hunter, BSc¹; Elisa Greco, MD²; Karen Cross, MD, PhD^1,3; and Julie Perry, PhD^1,4

Affiliations: ¹Division of Plastic Surgery, St. Michael’s Hospital, Toronto, Canada; ²Division of Vascular Surgery, St. Michael’s Hospital; ³Associate Scientist, Keenan Research Centre for Biomedical Science, Toronto, Canada; and ⁴Faculty of Dentistry, University of Toronto, Toronto, Canada

Correspondence: Julie Perry, PhD, Senior Research Associate, St. Michael’s Hospital, 30 Bond Street, Toronto, Ontario M5B 1W8; julie.perry@utoronto.ca

Disclosure: This work was partially funded by Inotec AMD (Cambridge, UK). The authors disclose no other conflicts of interest.