Transcutaneous CO 2 Therapy Improves Tissue Oxygenation,Ulcer Healing and Pain in Recurrent Hard-to-Heal Diabetic Foot Ulcers: A Randomized Controlled Trial

Abstract

This trial evaluates whether transcutaneous gaseous carbon dioxide therapy (CO₂ therapy) combined with standard of care (SOC) improves healing of hard-to-heal recurrent diabetic foot ulcers (DFUs) compared with SOC alone. A total of 30 participants with recurrent hard-to-heal DFUs were included and randomized 1:1 to an intervention (CO₂ therapy plus SOC) or a control group (SOC only). The intervention consisted of 20 daily 50-minute CO₂ therapy sessions over 4 weeks. After 4 weeks, 46.7% of ulcers in the intervention group were completely healed, versus 0% in the control group (p = 0.005). The intervention group showed a median area reduction of 85.7% (IQR 54.9–100) compared to 17.2% (IQR 0.88–33.5) in controls (p < 0.001), with significantly better Falanga wound-bed scores (p < 0.001). The intervention group demonstrated a significant relative increase in StO₂ (29.6 ± 19.6%), whereas the control group exhibited a decrease (-2.2 ± 6.9%), resulting in a between-group difference of 31.8% (95% CI 20.5%–43.0%; p < 0.001). Pain intensity significantly decreased in the intervention group only (p = 0.002). This trial suggests that CO₂ therapy in combination with SOC is superior to SOC alone in healing recurrent, hard-to-heal DFUs.

Keywords

chronic wound diabetic foot ulcer wound healing

Introduction

Diabetic foot ulcers (DFUs) cause a substantial burden on public health.^1,2 If unmanaged, DFUs represent a life threat for patients with DFU, with a five-year mortality rate up to 30%.³ Even upon complete ulcer healing, the risk of ulcer recurrence is high - approximately 40% within one year and 60% within five years.⁴

A particularly challenging subset is hard-to-heal DFUs, typically defined by a failure to achieve a 50% area reduction after four weeks of standard of care (SOC). The specific clinical problem of these DFUs is primarily characterized by a persistent, chronic inflammatory state that prevents the normal progression of wound repair.⁵ This state is attributable to several intrinsic factors, such as biochemical imbalance (between proinflammatory cytokines and their inhibitors, as well as proteases and their antiproteases), failure of matrix synthesis (due to continuous influx of inflammatory cells and sustained production of inflammatory mediators), cellular barriers due to ischemia, infection etc., but also extrinsic factors such as delayed clinical intervention and lack of universal adoption of the guidelines published by the International Working Group on the Diabetic Foot (IWGDF).⁵ Next, patient-specific factors such as non-adherence to treatment, mental health comorbidities, and socioeconomic barriers further complicate the wound healing.⁵ Finally, Meloni et al.⁶ report in their large cohort study that prevalence of purely neuropathic DFUs in patients with diabetes mellitus in Europe is decreasing, becoming less common than ischemic/neuro-ischemic ulcers. This epidemiological shift underscores the importance of addressing tissue hypoxia as one of the key barriers to healing.

Consequently, there is a critical clinical need for a safe, cost-effective, and non-invasive adjunctive therapy capable of clinically addressing the local tissue hypoxia. In this context, prior work⁷ in the field of DFUs indicates that repeated CO₂ therapy can improve cutaneous microvascular function. Collectively, these findings support CO₂ therapy as a promising adjunct treatment method for DFUs with the potential to mitigate local hypoxia and improve microcirculatory function, which are essential for sustained ulcer healing. Building on this rationale, our trial extends the evidence base by adding serial hyperspectral imaging (HSI) of superficial tissue oxygenation (StO₂) as an objective physiologic response marker, demonstrating a progressive increase in oxygenation across the prespecified weekly trial visits during the 4-week period. The importance of implementing newer treatment technologies in healing hard-to heal DFUs is also recognized by IWGDF.⁸ Although the need to improve ulcer healing in patients with hard-to-heal DFUs is widely recognized, the overall quality of evidence from studies evaluating new treatments is, unfortunately, poor⁸.

Transcutaneous application of gaseous carbon dioxide (CO₂) (hereinafter CO₂ therapy) is a promising adjuvant treatment method for non-invasive management of DFUs due to its accelerated ulcer-healing rate^9,10, improved microcirculatory function⁷, and reduced ulcer recurrence rates¹⁰, with no observable systemic or local adverse effects^7,9–11.

The primary objective of this trial was to determine if combining CO₂ therapy with SOC improves healing rates in patients with hard-to-heal, recurrent DFUs compared to those receiving SOC alone.

Materials and methods

This prospective, open-label, parallel-group, superiority, single-center randomized controlled trial (RCT) with allocation ratio of 1:1 was carried out (including recruitment) at the Department of Vascular Surgery, General hospital Novo mesto (Novo Mesto, Slovenia), between September 2024 and March 2025. The trial was conducted in accordance with the Declaration of Helsinki and approved by the National Medical Ethics Committee of the Republic of Slovenia (Approval ID: 0120- 316/2024-2711-3, date of approval: August 20, 2024). There were no modifications to the study design, pre-specified outcomes, or planned statistical analyses after trial commencement. The trial protocol, statistical analysis plan, de-identified participant data and statistical code are accessible via the corresponding author following a formal proposal. Patients and the public were not involved in the design, conduct or reporting of this trial. The trial reporting adheres to the CONSORT 2025 guidelines for reporting RCTs. Written informed consent was obtained from all participants. The sample size chosen was determined by feasibility considerations.

The primary hypothesis of this trial was: CO₂ therapy, combined with SOC, is superior to SOC only in terms of ulcer healing in patients with recurrent, hard-to-heal DFUs (often accompanied by structural foot deformity and prior minor amputation).

The primary endpoint was the proportion of healed ulcers after 4 weeks; complete ulcer healing was defined as full epithelialization of the ulcer surface without exudate, and no requirement for dressings, confirmed at two consecutive visits. The secondary endpoints were (1) relative change of StO₂ from baseline after 4 weeks $% Δ St O_{2 i} =$ $(\frac{S t O_{2} {(w e e k 4)}_{i} - S t O_{2} {(b a s e l i n e)}_{i}}{S t O_{2} {(b a s e l i n e)}_{i}}) x 100$ and (2) change in pain intensity after 4 weeks. For the first secondary endpoint, StO₂ was measured using HSI at baseline and after 4 weeks; in addition, StO₂ was measured after 1, 2 and 3 weeks at synchronized weekly trial visits. For the second secondary endpoint, pain intensity was assessed using the 11-point (from 0 to 10) numerical pain rating scale (NPRS) at baseline and after 4 weeks. Additional supportive endpoints were (1) percentage ulcer area reduction from baseline after 4 weeks, (2) the proportion of ulcers achieving ≥50% area reduction after 4 weeks, and (3) change in wound-bed appearance and exudate assessed using the Falanga wound bed score after 4 weeks (assessed in unhealed ulcers only).

The inclusion criteria were (1) age 18 years or older, (2) confirmed diagnosis of recurrent hard-to-heal DFU, (3) the ability to understand the trial procedure and provide written informed consent, and (4) no comorbidities constituting contraindications to CO₂ therapy (for the list of contraindications refer to the study by Ban Frangež et al.¹¹). Exclusion criterion was clinically infected ulcer at baseline. A hard-to-heal ulcer was defined as an ulcer failing to show a 50% reduction in its area after 4 weeks of SOC. Note that formal power calculation was waived in favor of a feasibility-based recruitment strategy.

Patient baseline characteristics included sex/gender, age, body mass index (BMI), smoking status, ankle-brachial index (ABI), glycated hemoglobin (HbA1c), functional status, psychological symptoms, mobility status, and amputation history. ABI was measured using AngE™ Pro 8 ABI + (SOT Medical Systems, Maria Rain, Austria). Ulcer baseline characteristics included ulcer type (etiology), ulcer classification, ulcer area, and ulcer duration.

The values of characteristics were obtained from structured interviews, physical examinations, device-based measurements, blood testing, and medical records. HbA1c was analyzed in a certified hospital laboratory using standardized methods. An ulcer area was measured using the imitoMeasure app (imito AG, Zürich, Switzerland). Ulcer type was defined as neuropathic (in the presence of distal symmetrical polyneuropathy), ischemic (if ABI<0.90), or neuro-ischemic (if criteria for both, neuropathic and ischemic ulcer type were met). Ulcer classification used the SINBAD score and the Falanga wound bed score. Measurements of StO₂ were carried out using the TIVITA® Tissue HSI system (Diaspective Vision GmbH, Am Salzhaff, Germany), with the center of the ulcer defined as the primary measurement site; the acclimatization period consisted of 15 minutes of supine rest in a room with ambient temperature of 22 °C. HSI was carried at weekly trial visits (after weeks 1, 2, 3 and 4); in the intervention group, measurements were carried out 30 min after completion of the 5^th (week 1), 10^th (week 2), 15^th (week 3) and 20^th (week 4) CO₂ therapy, while in the control group the measurements were obtained at the same study time points, synchronized with the intervention schedule. The measurement site for StO₂ was defined before the baseline measurement, and all subsequent measurements were performed at the same site. To ensure reproducibility across measurement sessions, a reference calibration was performed, and a fixed focal distance of 50 cm was maintained. Consistency was further ensured by automatic alignment of the regions of interest with a chosen anatomical landmark, i.e. the center of each participant’s ulcer.

Medical personnel continuously monitored participants for local reactions (irritation/edema) during CO₂ therapies and observed for signs of distress. Systemic effects—including dizziness, headache, or palpitations—were patient-reported. Physical examination of ulcer site and patient reports were collected at each visit to identify any adverse events occurring between sessions.

After obtaining all baseline characteristics, participants were randomized to intervention and control groups in a 1:1 allocation ratio using a computer-generated randomization sequence (based on permuted block randomization method with a fixed block size of 4) in Microsoft Excel (Microsoft, Redmon, Washington, USA) prepared by an independent researcher not involved in treatment or outcome assessment. Allocation was concealed using sequentially numbered, opaque, sealed envelopes, opened only after completion of the baseline assessments. Electronic laboratory notebook platform was not used.

Intervention and control procedure

Participants in the intervention group received SOC alongside 20 sessions of CO₂ therapy. These 50-minute sessions were administered once daily on weekdays (Monday–Friday) over a 4-week period. CO₂ therapies were performed using the PVR system® medical device (Derma Art d.o.o. & PVR med d.o.o., Brežice, Slovenia). The participant's lower body was sealed in a single-use, biocompatible, medical-grade wrap. The wrap was filled with medical-grade CO₂ to a 99.9% concentration for a 50-minute session. After 50 minutes, CO₂ was pumped out and the wrap removed. Participants in the control group received only SOC without the administration of a placebo intervention. SOC included sharp debridement as clinically indicated at trial visits, local ulcer care with standardized dressings and scheduled changes three times per week (an amorphous hydrogel covered with a polyurethane foam secondary dressing), offloading using a half-shoe in the case of forefoot/hindfoot ulcers or a combination of an offloading shoe with a donut cushion/pad for midfoot ulcers, and infection screening at each trial visit (any infection developing during the trial was managed according to the local hospital’s protocols). SOC was provided consistently in both groups. To enhance adherence and minimize loss to follow-up, participants in both groups received reminder calls/SMS messages and, if necessary, transportation to and from the hospital. StO₂ measurements were repeated after weeks 1, 2, 3, and 4. After week 4, ulcer healing was assessed, along with reassessment of ulcer area, ulcer classification (only in unhealed ulcers), StO₂, and pain intensity.

Statistical analysis

Categorical variables are presented as counts and proportions. Continuous variables are summarized as mean ± standard deviation (SD) when normally distributed and as median (interquartile range (IQR)) when non-normally distributed. Normality was assessed using the Shapiro-Wilk test. Homogeneity of variances for between-group comparisons was evaluated using Levene’s test; when the equal-variance assumption was violated, the unequal-variance (Welch) t-test was applied.

Between-group comparisons of continuous variables were performed using the independent-samples t-test (or Welch t-test, as appropriate) for normally distributed data and the Mann-Whitney U test for non-normally distributed data. Within-group paired comparisons were performed using the paired-samples t-test or the Wilcoxon signed-rank test, as appropriate. For within-group repeated measurements across more than two time points, the Friedman test (ANOVA by ranks) was used. Categorical variables were compared using the chi-square test; Fisher’s exact test (two-sided) was used when expected cell counts were small (Fisher–Freeman–Halton exact test for r × c tables), with mid-p values reported where applicable.

All tests were two-sided, with statistical significance defined as p < 0.05. Analyses were conducted using IBM SPSS Statistics, version 21.0 (IBM Corp., Armonk, NY, USA).

Since complete baseline data were a prerequisite for enrolment, no data imputation method was required.

Results

A total of 33 patients with DFUs were enrolled in our RCT and screened. Three of these were not included in the trial after they declined to participate, having confirmed they would not be able to do so under the RCT's protocol. As a result, 30 patients were randomized in a 1:1 allocation ratio to the intervention (n = 15) and control (n = 15) groups, yielding group sizes of 15 each. The flow of participants through the RCT was documented according to protocol (Figure 1). All participants completed the trial as planned, with no losses to follow-up and no exclusions from the intention-to-treat analysis. Both study groups followed the protocol exactly as described.

Figure 1.

Flow diagram of the progress through the phases of our RCT of the intervention and study groups.

Table 1 presents baseline person and ulcer characteristics for participants in both groups.

Table 1.

Baseline person and ulcer characteristics of the control and intervention groups

Variable/characteristic	Control group (n = 15)	Intervention group (n = 15)	p
Sex
Male	12 (80%)	13 (86.7%)	0.624^a
Female	3 (20%)	2 (13.3%)	0.624^a
Age, mean ± SD [year]	61.5 ± 6.9	61.3 ± 10.4	0.967^c
HbA1c, mean ± SD [%]	6.75 ± 0.96	6.77 ± 0.76	0.967^c
BMI:
Underweight (<18.5 kg/m²)	2 (13.3%)	0 (0.0%)	0.534^b
Normal (18.5-24.9 kg/m²)	4 (26.7%)	5 (33.3%)
Overweight (25.0-29.9 kg/m²)	5 (33.3%)	4 (26.7%)
Obesity (≥ 30 kg/m²)	4 (26.7%)	6 (40.0%)
Smoking status:
Yes	8 (53.3%)	7 (46.7%)	0.715^a
No	7 (46.7%)	8 (53.3%)	0.715^a
ABI:
< 0.90	4 (26.7%)	4 (26.7%)	0.957^b
0.90–0.99	4 (26.7%)	5 (33.3%)
1.00–1.30	5 (33.3%)	5 (33.3%)
> 1.3	2 (13.3%)	1 (6.7%)
Functional status:
Independent	10 (66.7%)	10 (66.7%)	0.819^a
Needs assistance	3 (20%)	2 (13.3%)
Fully dependent on caregivers	2 (13.3%)	3 (20%)
Psychological symptoms:
Depression symptoms	1 (6.7%)	3 (20.0%)	0.473^b
Anxiety/fearful symptoms	10 (66.7%)	9 (60.0%)	0.705^a
Mobility status:
Walks normally	0 (0.0%)	0 (0.0%)	0.477^b
Altered gait pattern	10 (66.6%)	11 (73.3%)
Difficulty walking	4 (26.7%)	2 (13.3%)
Unable to walk	1 (6.7%)	2 (13.3%)
Amputation history:
Yes	4 (26.7%)	8 (53.3%)	0.136^a
No	11 (73.3%)	7 (46.7%)	0.136^a
Revascularization history:
Yes	0 (0.0%)	0 (0.0%)	1.000^a
No	15 (100.0%)	15 (100.0%)
Ulcer type:
Neuropathic	7 (46.7%)	5 (33.3%)	0.693^a
Ischemic	4 (26.7%)	4 (26.7%)
Neuro-ischemic	4 (26.7%)	6 (40.0%)
Ulcer location:
Forefoot	8 (53.3%)	7 (46.7%)	0.715^a
Midfoot/hindfoot	7 (46.7%)	8 (53.3%)	0.715^a
Ulcer duration, median (IQR) [month]	6 (4-8)	8 (6-14)	0.056^d
Ulcer area, median (IQR) [cm²]	5.86 (1.52-11.20)	6.97 (1.52-15.09)	0.744^d
StO₂, mean ± SD [%]	51.3 ± 9.9	44.8 ± 9.8	0.082^c
SINBAD total score:
0	0(0.0%)	0 (0.0%)	0.906^b
1	1 (6,7%)	0 (0.0%)
2	2 (13.3%)	3 (20.0%)
3	3 (20.0%)	2 (13.3%)
4	6 (40.0%)	6 (40.0%)
5	3 (20.0%)	4 (26.7%)
6	0 (0.0%)	0 (0.0%)
Falanga wound bed score (appearance):
A	0 (0.0%)	0 (0.0%)	0.059^b
B	7 (46.7%)	1 (6.7%)
C	5 (33.3%)	9 (60.0%)
D	3 (20.0%)	5 (33.3%)
Falanga wound bed score (exudate):
1 (none/minimal amount)	2 (13.3%)	0 (0.0%)	0.545^b
2 (moderate amount)	7 (46.7%)	9 (60.0%)
3 (very exudative wound)	6 (40.0%)	6 (40.0%)
NPRS:
No pain	0 (0.0%)	0 (0.0%)	0.386^b
Mild (1–3)	4 (26.7%)	1 (6.7%)
Moderate (4–6)	7 (46.7%)	7 (46.7%)
Severe (7–9)	4 (26.7%)	6 (40.0%)
Very severe (10)	0 (0.0%)	1 (6.7%)

Legend: Categorical data are presented as counts and proportions and continuous data as mean ± SD or median (IQR), as appropriate. p-values refer to between-group comparisons. ^a - χ² test, ^b - Fisher’s exact test (two-sided mid-p; Fisher–Freeman–Halton for r × c tables). ^c – independent samples t-test (Welch correction when appropriate), ^d - Mann–Whitney U test. Abbreviations: ABI: ankle–brachial index; BMI: body mass index; DFU: diabetic foot ulcer; IQR: interquartile range; NPRS: numerical pain rating scale; SD: standard deviation; StO₂: superficial tissue oxygenation.

Age, HbA1c, and StO₂ were normally distributed in both groups, whereas the distribution of all other variables was non-normal. At baseline, there were no statistically significant differences between the groups in any of the variables characterizing persons and ulcers. In addition, when it comes to the SINBAD score, there were no statistically significant differences between the scores for each category between the two groups. All previous amputations were classified as minor, occurring distal to the ankle (e.g., toe, ray, transmetatarsal, Lisfranc, or Chopart amputations).

Table 2 shows characteristics of the intervention and control groups after 4 weeks.

Table 2.

Ulcer characteristics and pain scores of the control and intervention groups after 4 weeks

Variable/characteristic	Control group (n = 15)	Intervention group (n = 15)	p
Number of healed ulcers	0 (0.00%)	7 (46.7%)	0.005 ^a
Ulcer area, median (IQR) [cm²]	2.69 (1.00–12.18)	1.1 (0.00–4.97)	0.031^b
Percentage ulcer area reduction [%]	17.2 (0.9–33.5)	85.7 (54.9–100)	<0.001^b
Number of ulcers achieving ≥50% area reduction	3 (20.0%)	14 (93.3%)	<0.001^a
Falanga wound bed score (appearance)*:
A	3 (20%)	8 (100.0%)	<0.001^a
B	7 (46.7%)	0 (0.0%)
C	4 (26.7%)	0 (0.0%)
D	1 (6.6%)	0 (0.0%)
Falanga wound bed score (exudate)*:
1 (none/minimal amount)	2 (13.3%)	8 (100.0%)	<0.001^a
2 (moderate amount)	7 (46.7%)	0 (0.0%)
3 (very exudative wound)	6 (40.0%)	0 (0.0%)
NPRS:
No pain	0 (0.0%)	6 (40.0%)	0.002^a
Mild (1–3)	6 (40.0%)	8 (53.3%)
Moderate (4–6)	8 (53.3%)	1 (6.7%)
Severe (7–9)	1 (6.7%)	0 (0.0%)
Very severe (10)	0 (0.0%)	0 (0.0%)

Legend: Categorical data are presented as counts and proportions and continuous data as median (IQR). Area of healed ulcers was set to 0 cm². p-values refer to between-group comparisons. * - Falanga wound bed score (both appearance and exudate) was assessed in unhealed ulcers only (control n = 15; intervention n = 8). ^a - Fisher’s exact test (two-sided mid-p, where appropriate). ^b - Mann–Whitney U test. Abbreviations: IQR: interquartile range; NPRS: numerical pain rating scale.

In addition to these results, in 7 unhealed ulcers in the intervention group after 4 weeks, the ulcer area was reduced by more than 50%, i.e., 55.3% (52.3% - 76.9%). In the control group, only 3 such ulcers were observed. In addition, the ulcer area of one ulcer in the control group even increased. There was a statistically significant difference in the number of ulcers, with their area reduced by 50% or more, between the intervention and control groups (p < 0.001; Fisher’s exact test (two-sided; mid-p where applicable). Falanga wound bed scores were evaluated only in ulcers that remained open at week 4; therefore, between-group comparisons reflect differences among persistent wounds Lastly, the Wilcoxon signed rank test revealed a statistically significant difference in ulcer area before and after 4 weeks in control (p = 0.006) and treatment group (p = 0.012). To sum these results, CO₂ therapy in combination with SOC is superior to SOC alone in terms of (1) proportion of healed ulcers after 4 weeks (primary endpoint), (2) percentage ulcer area reduction, (3) proportion of ulcers with ulcer area reduction of ≥50% and (4) formation of granulation tissue/reduction of exudate in ulcers after 4 weeks (all supportive endpoints), and (5) statistically significant change in pain intensity (second secondary endpoint).

Table 3 presents StO₂ values measured after weeks 1, 2, 3, and 4, together with the between-group differences in StO₂ (ΔΔStO₂). The ΔΔStO₂ values favored the intervention group at all post-baseline trial visits (p < 0.001).

Table 3.

Superficial tissue oxygenation StO2 measured with hyperspectral imaging: weekly values and between-group differences

Time point	StO₂, mean ± SD (Control group (n = 15))	StO₂, mean ± SD (Intervention group (n = 15))	ΔΔStO2, mean (95% CI) [%]	p
Week 1	50.1 ± 9.2	48.0 ± 9.1	4.3 (2.0 to 6.7)	<0.001^a
Week 2	50.1 ± 9.3	50.6 ± 8.5	6.9 (3.9 to 9.9)	<0.001^a
Week 3	50.7 ± 10.1	52.9 ± 8.5	8.7 (5.7 to 11.8)	<0.001^a
Week 4	50.1 ± 10.2	57.0 ± 9.9	13.3 (9.3 to 17.4)	<0.001^a

ΔΔStO₂ denotes the between-group difference (intervention minus control) in mean change from baseline. p-values refer to between-group comparisons. ^a - independent-samples t-test with Welch correction. Abbreviations: CI: confidence interval; SD: standard deviation; StO₂: superficial tissue oxygenation.

The %ΔStO₂ values were calculated for both groups at weeks 1, 2, 3, and 4 (Figure 2). In the intervention group, %ΔStO₂ increased progressively over time, whereas the change in the control group remained stable, resulting in a widening between-group separation throughout the trial. After 4 weeks, the relative change in StO₂ from baseline was 29.6 ± 19.6% in the intervention group compared to -2.2 ± 6.9% in the control group (between-group difference was 31.8% (95% CI 20.5%–43.0%) (p < 0.001).

Figure 2.

Relative change of superficial tissue oxygenation from baseline (%ΔStO₂) over time. Individual measurements are shown as scatter points. The horizontal axis labels refer to synchronized weekly trial visits conducted after 1, 2, 3, and 4 weeks.

Finally, no adverse events were observed in either group, and no participants reported negative sensations or discomfort during the trial.

Discussion

Results of our RCT demonstrate that CO₂ therapy, when added to SOC, can meaningfully improve outcomes in a high-risk subgroup of DFUs with recurrent, hard-to-heal ulcers. After 4 weeks, 46.7% (7 out of 15) of ulcers in the intervention group were completely healed compared with none (0 out of 15) in the control group (p = 0.005). Consistently, the intervention group achieved a markedly greater percentage ulcer area reduction (85.7% vs. 17.2%, p < 0.001) and a substantially higher proportion of ulcers reaching the clinically relevant ≥50% area-reduction threshold (93.3% vs. 20.0%, p < 0.001). Together, these results suggest that CO₂ therapy may help overcome the stagnant healing trajectory characteristic of hard-to-heal DFUs. Importantly, baseline superficial tissue oxygenation tended to be lower in the intervention group, suggesting that the observed differences in healing were unlikely to be attributable to more favorable initial perfusion status. Compared to other trials studying the effect of CO₂ on DFU healing^9,10, the achieved proportion of healed ulcers in the intervention group is lower. This is most likely attributable to a deliberate inclusion of a particularly challenging subgroup, i.e., patients with recurrent, hard-to-heal DFUs, which makes the observed healing rate clinically meaningful, particularly given the fact that none of the ulcers in the control group healed after 4 weeks.

Beyond the observed proportion of healed ulcers in the intervention group, the proportion of achieving percentage area reduction of ≥50% and marked improvements in wound-bed appearance and exudate control in the intervention group suggest that CO₂ therapy favorably modifies the local ulcer environment, promoting granulation and progression toward healing even when full epithelialization is not yet achieved.

A core barrier to DFU healing is local tissue hypoxia driven by impaired microcirculation. The significant improvement in StO₂ serves as an objective physiological marker of the observed clinical response. Mechanistically, the observed improvement in tissue oxygenation is likely multifactorial and may involve complementary physiological pathways, including (1) enhanced oxygen unloading from hemoglobin associated with increased CO₂ (Bohr effect; rightward shift of the oxyhemoglobin dissociation curve)¹², (2) CO₂-associated vasodilatory responses at the microcirculatory level, and (3) changes in endothelial (including NO-mediated) and neurogenic regulation of cutaneous blood flow⁷. These findings are consistent with prior clinical reports of improved wound outcomes following transcutaneous CO₂ therapy^9,10, supporting the clinical relevance of CO₂-evoked improvements in local microcirculation. To properly interpret StO₂ results, it is important to note there is a well-established physiological relationship exists between StO₂ and transcutaneous oxygen pressure tcpO₂, as both parameters are bound by the hemoglobin-oxygen dissociation curve, even though both parameters measure different physiological parameters. In undisturbed, resting chronic wounds, StO₂ and tcpO₂ correlate well.¹³ However, this correlation disappears when physiological stress is introduced (e.g., acute surgical trauma¹⁴). This divergence occurs because StO₂ reacts dynamically to regional fluid and hemoglobin volume shifts, whereas tcpO₂ captures forced capillary oxygen pressure under induced hyperemia.

CO₂ therapy was also associated with a clinically meaningful reduction in pain. At week 4, 40.0% of participants in the intervention group reported being pain-free, compared with 0% in the control group (p = 0.002), while moderate-to-severe pain remained common in the control group (60.0%) but was uncommon in the intervention group (6.7%). Given the well-established impact of pain on quality of life in DFU patients¹⁵, this suggests an additional patient-centered benefit beyond ulcer healing.

Finally, no adverse effects were observed in the intervention group, supporting the feasibility and local safety profile of CO₂ therapy, consistent with previous studies^9–11.

This RCT is subject to several limitations. First, the small sample size (n = 30) reduces the statistical power (due to type II error) and limits the generalizability of the findings. Second, the feasibility-based recruitment without formal power calculation potentially may have precluded the detection of subtle, but clinically significant differences; thus, non-significant results should be interpreted with caution. Third, the open-label design (necessitated by the nature of the intervention) may have introduced performance or attention bias, which must be considered when interpreting the magnitude of the observed effects. Finally, due to the lack of a follow-up period, it is not possible to draw definitive conclusions regarding long-term sustained ulcer healing or ulcer recurrence rates. Nevertheless, this RCT provides exploratory efficacy data and mechanistic insight and lays the groundwork for future adequately powered multicentre studies with extended follow-up.

Conclusions

CO₂ therapy in combination with SOC is superior to SOC alone in healing recurrent, hard-to-heal DFUs, including persistent ulcers at prior minor-amputation sites. It is associated with significant improvements in tissue oxygenation, reduced pain and consequent improvement of quality of life. CO₂ therapy represents a promising non-invasive adjuvant to standard care for hard-to-heal DFUs.

Footnotes

Acknowledgments

The authors have no acknowledgments to declare.

ORCID iDs

José Luis Lázaro-Martínez

Nikolaos Papanas

Igor Frangež

Author Contributions

M.P.: Investigation, Formal Analysis, Writing – original draft; B.A.: Data Curation, Formal Analysis, Investigation, Visualization, Writing – original draft; J.L.L.-M.: Writing – Formal Analysis, Review & Editing; M.B.: Writing – Formal Analysis, Review & Editing; N.P.: Writing – Formal Analysis, Review & Editing; I.F.: Conceptualization, Methodology, Project administration, Supervision, Writing – original draft

Ethical Considerations

The trial was conducted in accordance with the Declaration of Helsinki and approved by the National Medical Ethics Committee of the Republic of Slovenia (Approval ID: 0120-316/2024-2711-3, date of approval: August 20, 2024).

Consent to participate

Written informed consent was obtained from all individual participants included in the study.

Consent for Publication

Not applicable.

Funding Statement

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interest

Miloš Potkonjak no conflicts of interest to declare. Bernardka Ančimer no conflicts of interest to declare. José Luis Lázaro-Martínez no conflicts of interest to declare. Marianne Brodmann no conflicts of interest to declare. Nikolaos Papanas has been an advisory board member of Astra-Zeneca, Bayer, Boehringer Ingelheim, Menarini, MSD, Novo Nordisk, Pfizer, Takeda and TrigoCare International; has participated in sponsored studies by Astra-Zeneca, Eli-Lilly, GSK, MSD, Novo Nordisk, Novartis and Sanofi-Aventis; has received honoraria as a speaker for Astra-Zeneca, Bayer, Boehringer Ingelheim, Eli-Lilly, Elpen, Menarini, MSD, Mylan, Novo Nordisk, Pfizer, Sanofi-Aventis and Vianex; and has attended conferences sponsored by TrigoCare International, Astra-Zeneca, Boehringer Ingelheim, Eli-Lilly, Galenica, Menarini, Novo Nordisk, Pfizer and Sanofi-Aventis. Igor Frangež no conflicts of interest to declare.

Data Availability

The datasets generated and analyzed during the study are available from the corresponding author on request.

The trial was registered at ClinicalTrials.gov (ID: NCT07418073; ) on February 10, 2026.

Statements and Declarations

References

Boulton

Vileikyte

Ragnarson-Tennvall

, et al. The global burden of diabetic foot disease. Lancet 2005;366(9498):1719–1724; doi: 10.1016/S0140-6736(05)67698-2.

Armstrong

Swerdlow

Armstrong

, et al. Five year mortality and direct costs of care for people with diabetic foot complications are comparable to cancer. J Foot Ankle Res 2020;13(1); doi: 10.1186/S13047-020-00383-2.

Armstrong

Tan

Boulton

AJM

, et al. Diabetic Foot Ulcers: A Review. JAMA 2023;330(1):62–75; doi: 10.1001/JAMA.2023.10578.

Armstrong

Boulton

AJM

Bus

SA.

Diabetic Foot Ulcers and Their Recurrence. New England Journal of Medicine 2017;376(24); doi: 10.1056/nejmra1615439.

Nube

Frank

White

, et al. Hard-to-heal diabetes-related foot ulcers: current challenges and future prospects. Chronic Wound Care Management and Research 2016;Volume 3:133–146; doi: 10.2147/CWCMR.S84990.

Meloni

Izzo

Giurato

, et al. Prevalence, Clinical Aspects and Outcomes in a Large Cohort of Persons with Diabetic Foot Disease: Comparison between Neuropathic and Ischemic Ulcers. J Clin Med 2020;9(6):1–11; doi: 10.3390/JCM9061780.

Finžgar

Frangež

Cankar

, et al. Transcutaneous application of the gaseous CO2 for improvement of the microvascular function in patients with diabetic foot ulcers. Microvasc Res 2021;133; doi: 10.1016/j.mvr.2020.104100.

Chen

Vilorio

Dhatariya

, et al. Guidelines on interventions to enhance healing of foot ulcers in people with diabetes (IWGDF 2023 update). Diabetes Metab Res Rev 2024;40(3); doi: 10.1002/DMRR.3644.

Macura

Ban Frangez

Cankar

, et al. The effect of transcutaneous application of gaseous CO2 on diabetic chronic wound healing—A double-blind randomized clinical trial. Int Wound J 2020;17(6):1607–1614; doi: 10.1111/IWJ.13436.

10.

Bulum

Poljičanin

Badanjak

, et al. Effect of Transcutaneous Application of Carbon Dioxide on Wound Healing, Wound Recurrence Rate and Diabetic Polyneuropathy in Patients with Neuropathic, Ischemic and Neuroischemic Diabetes-Related Foot Ulcers. 2025;15(4):618; doi: 10.3390/LIFE15040618.

11.

Ban Frangež

Rodi

Miklavčič

, et al. The Effect of Transcutaneous Application of Gaseous CO2 on Diabetic Symmetrical Peripheral Neuropathy—A Double-Blind Randomized Clinical Trial. Applied Sciences 2021, Vol 11, Page 4911 2021;11(11):4911; doi: 10.3390/APP11114911.

12.

Jensen

FB.

Red blood cell pH, the Bohr effect, and other oxygenation-linked phenomena in blood O2 and CO2 transport. Acta Physiol Scand 2004;182(3):215–227; doi: 10.1111/J.1365-201X.2004.01361.X.

13.

Jafari-Saraf

Wilson

Gordon

IL.

Hyperspectral Image Measurements of Skin Hemoglobin Compared With Transcutaneous PO2 Measurements. Ann Vasc Surg 2012;26(4):537–548; doi: 10.1016/J.AVSG.2011.12.002.

14.

Kluckner

Hitzl

Enzmann

, et al. Real-Time Perfusion Assessment with Hyperspectral Imaging After Revascularization in Peripheral Artery Disease. J Clin Med 2026;15(4):1667; doi: 10.3390/JCM15041667.

15.

Ribu

Rustøen

Birkeland

, et al. The Prevalence and Occurrence of Diabetic Foot Ulcer Pain and Its Impact on Health-Related Quality of Life. Journal of Pain 2006;7(4):290–299; doi: 10.1016/j.jpain.2005.12.002.