Hyperbaric Oxygen Therapy for Diabetic Foot Ulcers Based on Wagner Grading: A Systematic Review and Meta-analysis

Abstract

Background:

Diabetic foot ulcers (DFUs) are common complications of uncontrolled diabetes mellitus that can result in infection and amputation of the lower extremities. This study compared the benefits and risks of hyperbaric oxygen therapy with those of other DFU treatments, based on the Wagner grading system.

Methods:

Systematic searches for randomly controlled trials using hyperbaric oxygen therapy for DFUs were performed using PubMed, the Cochrane Library, and Embase. Data regarding demographics, wound healing, minor and major amputations, operative debridement, nonhealing wounds, and adverse effects were analyzed based on Wagner grades, using RevMan 5.4.1 and Microsoft Excel.

Results:

Hyperbaric oxygen therapy was significantly superior to other treatments for wound healing rates 8 or more weeks after the final treatment (RR = 2.39; 1.87–3.05; P < 0.00001) minor/distal amputations (RR = 0.58; 0.43–0.80; P < 0.007), and major/proximal amputations (RR = 0.31; 0.18–0.52; P < 0.00001) for the 14 studies analyzed. In addition, this therapy increased the rate of complete wound healing for Wagner grades II (RR = 21.11; 3.05–146.03; P = 0.002), III (RR = 19.58; 2.82–135.94, P = 0.003), and IV (RR = 17.53; 2.45–125.44; P = 0.004); decreased the minor/distal amputation rate for grade III (RR = 0.06; 0.01–0.29; P = 0.0004) and the major/proximal amputation rate on for grade IV (RR = 0.08; 0.03–0.25; P < 0.0001); and decreased the operative debridement rate for Wagner grade II (RR = 0.09; 0.01–0.60; P = 0.01).

Conclusions:

Moderate-quality evidence revealed that adjunctive hyperbaric oxygen therapy improved DFU wound healing for Wagner grades II, III, and IV; prevented minor and major amputations for grades III and IV, respectively; and prevented operative debridement in grade II wounds.

Takeaways

Question: Which Wagner grades benefit from hyperbaric oxygen therapy for diabetic foot ulcers?

Findings: A systematic review and meta-analysis of 14 randomized controlled trials revealed moderate-quality evidence that adjunctive hyperbaric oxygen therapy significantly improved wound healing rates 8 or more weeks after the last treatment for Wagner grades II–IV and reduced the need for minor and major amputations in more severe cases, with no notable adverse effects.

Meaning: Encourage hyperbaric oxygen therapy in daily practice for diabetic foot ulcer patients with certain indications.

INTRODUCTION

Diabetes mellitus that is poorly controlled commonly leads to diabetic ulcers in regions of the foot that experience repetitive pressure and trauma.¹ Inadequate glycemic control, underlying neuropathy, peripheral vascular disease, or inadequate foot care can induce these ulcers, which can result in foot osteomyelitis and lower limb amputations.^1,2

In response to the frequent hospitalization of patients with diabetic foot problems, the International Diabetes Federation launched the “putting feet first” campaign to increase awareness of the common occurrence of amputations in diabetic patients worldwide.³

Foot ulcers occur in 6.3% of diabetic patients globally. Belgium reports a prevalence of 16.6%, whereas Asia has 5.5% and Australia registers 1.5%.⁴

The International Working Group on the Diabetic Foot and the Wound Healing Society identify the standard treatments for diabetic ulcers as wound debridement, infection control with empirical antibiotics, offloading/pressure relief, moist wound dressings, dressing selection (hydrocolloid, foam, and alginate), blood sugar control, and nutritional support.^5–9

In 1960, hyperbaric oxygen therapy (HBOT) was shown to increase tissue oxygen levels without the support of hemoglobin.¹⁰ Compared with standard wound care, this therapy promoted faster healing and a reduced need for amputation with no serious adverse effects. Of the 20 diabetic foot ulcer (DFU) patients assessed, 18 experienced complete wound healing and only one required amputation.

HBOT uses an increased inhaled oxygen pressure to enhance oxygen transfer from lung alveoli to capillaries, thereby boosting dissolved oxygen levels in the blood and promoting tissue oxygenation, toxin removal, neovascularization, fibroblast growth, enhanced bacterial defense, and inflammation reduction. This increased oxygen content favors tissue recovery, wound healing, and scar remodeling, allowing HBOT to effectively treat osteomyelitis, radiation-induced bone damage, and diabetic foot ulcers with vasculitis and infections.¹¹

Studies treating DFUs with HBOT have produced diverse results, from significant improvements in wound healing and amputation rates^12–14 to similar benefits to those of standard care.^15,16 This study compared the benefits and risks of HBOT with those of other DFU treatments.

METHODS

Data Sources and Searches

A systematic review of studies published before February 23, 2023 was performed based on the preferred reporting items for systematic reviews and meta-analyses (PRISMA) checklist using three online databases (PubMed, Cochrane Library, and Embase). The keywords “hyperbaric oxygen therapy,” “high pressure oxygen,” “oxygen,” “diabetic foot ulcer” “wound healing,” “amputation,” and “standard wound treatment” were used, along with filters for randomized controlled trials and the English language.

Study Selection

Our inclusion criteria were (1) randomized controlled trials comparing HBOT with other treatment methods for DFUs; (2) participants with diabetes mellitus type 1 or type 2 who presented with ulcers in the lower extremity for at least 30 days and were unresponsive to standard treatment, with or without vascular insufficiency; (3) HBOT treatments administered at pressures of 2.0–3.0 ATA for a minimum of 45 minutes once or twice daily in a mono- or multi-place chamber; and (4) control groups receiving the “standard treatment” used by the health center for wound healing or treatments other than HBOT.

Our exclusion criteria were patients who underwent amputations on the ulcer extremity before the trial and those with chronic ulcers that were not diabetic-related (pressure, arterial, venous ulcers, etc.)

Data Extraction

The study characteristics, risk-of-bias parameters, pretreatment conditions, wound healing rates, amputation rates, and Wagner grade outcomes were extracted and reviewed. In addition, the author(s); study site; publication year; methods (population studied, treatment duration, follow-up duration, control and treatment group criteria); and outcomes (outcomes measured, participant numbers for each group, mean, SD or number of events for each outcome, and time at which each outcome was assessed) were collected.

Statistical Analysis

The patient outcomes after standard wound care with adjunct HBOT (intervention group), standard wound care alone, or treatments other than HBOT (control group) were compared using Review Manager (version 5.4.1.) at a statistical significance level of P less than 0.05. We presented pooled results as the mean difference for continuous data, used odds ratios or risk differences for categorical data, assigned I² less than or equal to 50% to indicate no significant statistical heterogeneity among the studies, and used a fixed-effects model for the pooled analysis. Unpooled data were summarized descriptively.

Risk of Bias Assessment in the Included Studies

The Cochrane Handbook (version 5.1.0) randomized controlled trial risk assessment tool was used to categorize the study quality into three levels based on selective, implementation, measurement, follow-up, and reporting biases as follows: low risk (four or more low-risk criteria), moderate risk (two or three low-risk criteria), and high risk (one or zero low-risk factors).

RESULTS

The database searches yielded 785 published studies. Duplicates were eliminated and the titles and abstracts of the remaining studies were reviewed. The full texts of 61 articles were obtained, of which 37 did not use randomized controlled trials, four involved unsuitable outcomes, three did not provide the full text, and two had unsuitable treatment or control groups. After excluding these studies, 15 trials were included for analysis. Figure 1 shows the PRISMA flowchart.

Fig. 1.

Preferred reporting items for systematic reviews and meta-analysis (PRISMA) flowchart.

Characteristics of Included Populations

The 15 studies were conducted in Malaysia, Taiwan, Egypt, Turkey, China, Sweden, India, France, the Netherlands, the United Kingdom, Canada, and Italy. (See table, Supplemental Digital Content 1, which displays baseline characteristics of the study, http://links.lww.com/PRSGO/D115.) (See appendix, Supplemental Digital Content 2, which displays the additional characteristics of the studies. http://links.lww.com/PRSGO/D116.)

The average age of the participants was 43.8–72 years for the HBOT group and 45–70.6 years for the control group. The proportion of men varied between 33.3% and 80.8% across all studies. The patient characteristics were as follows: diabetes mellitus diagnosis for 9.8–24.8 years, DFUs for 4.98–57.2 weeks, HbA1c values of 7.8%–9.9%, TcpO₂ measurements of 21.3–5.5 mm Hg, and ulcer sizes of 0.8–29.9 cm². In the studies that reported Wagner grade, grades I to IV DFUs were present, with most patients presenting grades II (39.8%) or III (37.7%).

The sample sizes consisted of eight to 60 participants in each arm and the HBOT groups received one to two daily sessions and four to 40 treatments, delivered using either single or multiple hyperbaric chambers. Some studies involved 2- to 10-week sessions or continued until the wound was completely closed. Follow-ups varied from none to 1 year after treatment.

All studies provided standard wound care for both the control and treatment groups, and some included full standard wound care (offloading, infection control, debridement, dressing changes, blood sugar control, and nutritional support).^15,17–21 Five studies did not provide a complete explanation of standard wound treatment.^12,22–25 Three did not include details of the standard wound treatment provided,^26–28 and one used only normal saline and debridement.²⁹

The HBOT settings were generally consistent, with 100% oxygen and ATA pressures of 2.4–2.5, although some studies^23,24 used a pressure of 3 ATA. All studies paired HBOT with standard wound care.^20,24 No uniform pattern in HBOT settings was noted across countries, as each healthcare center followed distinct protocols.

Most studies compared HBOT and standard wound treatment with standard treatment only, whereas one study²⁹ divided the samples into three different treatment option groups: antiseptic dressings, HBOT, and platelet-derived growth factor gel dressings. Another report²⁸ compared HBOT with extracorporeal shockwave therapy. Two studies reported sham treatments for the control groups.^15,21

Risk of Bias

Figure 2 shows the estimated risk of bias assessment of the 15 randomized controlled trials. The overall quality of the studies is represented in Figure 3.

Fig. 2.

Methodological study quality using the Cochrane risk-of-bias tool for randomized controlled trials.

Fig. 3.

Risks of bias presented as percentages.

Nine studies properly randomized the samples using a computer generator,^17,18,25,27 number generator,²³ or randomization table or block.^19,25,26 Three studies did not mention randomization, one did not explain their randomization method, and two mentioned a sealed envelope but did not indicate whether shuffling occurred^20,22; therefore, selection bias was deemed unclear for those studies. Five studies clearly established allocation concealment methods using an opaque sealed envelope with a code or a computerized method.^{17,20,21,25,27} Two studies used sealed envelopes that were not identified as opaque or sequenced.^15,22 The remainder of the studies did not identify a concealment method. Three studies used a sham treatment for the control, whereby the participants were completely blinded.^12,15,20 One study was not blinded,²⁸ and other studies were considered high risks for performance bias because of the lack of a sham treatment. The assessors were blinded in eight studies,^{15,17,19–21,25–27} while one study²⁸ was not blinded, and the remaining studies provided no information regarding blinding.^{12,18,22–24} The largest proportion of withdrawn patients occurred in the study by Chaudhary et al,²⁹ with five patients withdrawn from the HBOT group (25%) and six removed from the control group (30%). Eight studies reported the loss of patients for follow-up due to conditions such as claustrophobia, hospitalization, amputation, death, vascular intervention, and barotrauma otitis.^{15,17,20,21,25,27–29} Six studies mentioned the reasons for the patient losses and all participants completed the follow-up in three studies, which were considered low risk for attrition bias. Two studies did not include the reasons for participant withdrawals and were considered high risks for attrition bias.^17,29 Two of the studies had registered clinical trial numbers (NTR3944 and NCR01219127),^27,28 whereas the studies that did not indicate protocols were categorized as having unclear selective reporting biases.

Results of the Meta-analysis

Primary Outcomes

Wound healing was evaluated in seven trials,^{15,17–20,26,28} with a total of 326 participants divided into HBOT (165 participants) and control groups (161 participants). Two studies were excluded due to a lack of events.^15,19 The remaining studies showed nonsignificant decreases of total ulcer healing in the HBOT groups compared with the control groups at the less than 8-week follow-up (RR = 1.08; 95% CI, 0.69–1.69; P = 0.75; I² = 75%). The trial with the greatest influence (83.3%) and heterogeneity²⁸ was excluded. The test for overall effect significantly favored the HBOT group (RR = 4.16; 95% CI, 1.62–10.66; P = 0.003; I² = 0%; Fig. 4).

Fig. 4.

Rate of complete wound healing at less than 8 weeks after final treatment.

Eight trials involving a total of 517 participants, 259 in the HBOT group and 258 in the control group, reported outcomes at 8 weeks or more after the final treatment.^{15,18,20,21,23,25,27,29} The trial that had the greatest impact²⁷ contributed 48.3% of the weight. The HBOT group showed a significant increase in total ulcer healing compared with the control group (RR = 2.39; 95% CI, 1.87–3.05; P < 0.00001, I² = 82%; Fig. 5).

Fig. 5.

Rate of wound healing at 8 weeks or more after final treatment.

Nine studies, with a total of 577 participants, 290 in the HBOT group and 287 in the control group, reported minor or distal amputation rates.^{12,15,18,20,21,23–26} One study²⁶ showed no events in either group and was excluded from the analysis. A significant decrease in minor/distal amputations was observed in the HBOT group compared with the control group (RR = 0.58; 95% CI, 0.43–0.80; P < 0007, I² = 71%). The study with the greatest contribution (34.9%) to the results²³ affected the heterogeneity of the pooled effect and was excluded. This final result was not significant (RR = 0.81; 95% CI, 0.58–1.12; P = 0.20; I² = 28%; Fig. 6).

Fig. 6.

Rate of minor/distal amputations.

Nine studies involving 540 participants, 275 in the HBOT group and 265 in the control group, reported major or proximal amputation rates.^{12,15,18,20,23–27} Two of these^18,26 were excluded from the analysis, as there were no events in either group. One trial²³ contributed to the majority (32.1%) of the weight. A significant decrease was observed in major/proximal amputations in the HBOT group compared with the control group (RR = 0.31; 95% CI, 0.18–0.52; P < 0.00001; I² = 36%; Fig. 7).

Fig. 7.

Rate of major/proximal amputations.

Secondary Outcomes

The overall effect for all subgroups of two trials,^23,25 with a total of 154 participants implied a significant increase in complete wound healing based on Wagner grade (Fig. 8).

Fig. 8.

Rate of complete wound healing based on Wagner grade.

The same two trials^23,25 revealed significant decreases in minor/distal amputations based on Wagner grade (Fig. 9) overall for all subgroups.

Fig. 9.

Rate of minor/distal amputations based on Wagner grade.

Three trials^12,23,25 with a total of 222 participants reported significant decreases in major/proximal amputations based on Wagner grade (Fig. 10); however, the results for grades II and III were not sufficient for analysis.

Fig. 10.

Rate of major/proximal amputations based on Wagner grade.

Two trials^23,25 with a total of 154 participants showed an overall significant decrease in operative debridements, based on Wagner grade for all subgroups (Fig. 11).

Fig. 11.

Rate of operative debridements based on Wagner grade.

The same two trials^23,25 revealed an overall effect for all subgroups of a nonsignificant increase in nonhealing wound rate based on Wagner grade (Fig. 12).

Fig. 12.

Rate of nonhealing wounds based on Wagner grade.

No significant difference in the graft or flap closure requirement between the HBOT and control groups was found in the two applicable studies^23,25 (Fig. 13).

Fig. 13.

Rate of graft/flap closures required.

The results for the four studies that presented rates of adverse events^15,20,26,27 showed slight increases in adverse side effects in the HBOT groups compared with the control groups, but the differences were not statistically significant (Fig. 14).

Fig. 14.

Rate of adverse effects.

Publication Bias

An asymmetrical pattern was observed on the funnel plot for the selected outcomes of complete wound healing 8 weeks or more after the final treatment, which indicates a publication bias (Fig. 15). In addition, Begg’s test revealed a significant publication bias (P = 0.03).

Fig. 15.

Funnel plot of complete wound healing at 8 weeks or more after final treatment.

DISCUSSION

Poorly controlled diabetes can cause foot ulcers that lead to osteomyelitis and amputation, resulting in frequent hospitalizations.^1,3 In 1960, HBOT was shown to increase tissue oxygen, resulting in faster healing and reduced amputation rates in patients with DFUs.¹⁰ Subsequent studies have reported significant improvements while others have shown few benefits of HBOT compared with standard care.^12–16

The 15 randomized controlled trials reviewed showed moderate overall quality due to a high level of performance bias, as most trials did not implement a sham treatment, and an unclear reporting bias since most studies did not involve a protocol. Previous meta-analyses found a lack of participant blinding in the trials.^16,30

DFU patients who were adjunctively treated with HBOT had significantly higher likelihoods of complete wound healing at 8 weeks or more after the final HBOT session in 14 of the randomized controlled trials. In addition, HBOT significantly reduced the risk of minor/distal and major/proximal amputations and slightly (not significantly) increased wound healing at less than 8 weeks after the final treatment, which resulted in a significant four-fold increase after one study²⁸ was excluded from the pooled analysis.

One study³¹ revealed that HBOT treatment (n = 793) increased the overall amputation (6.7% compared with 2.1%) and major amputation (3.3% compared with 1.3%) rates and lowered the wound healing (43.2% compared with 49.6%) percentage after 16 weeks of follow-up compared with treatment without HBOT.³² Other meta-analyses were consistent with our study (HBOT increased wound healing^33–36 and reduced major amputations^{33,34,36–38}), despite some studies reporting no differences in minor amputation rates.^33,35 In 2015, a Cochrane review¹⁶ found that HBOT for chronic wounds significantly improved ulcer healing over the short term (at 6 weeks) but not over the long term (after 1 year). However, the clinical heterogeneity and methodological shortcomings of the clinical trials weakened these findings.³⁹

Our analysis identified three trials that had reported outcomes based on the Wagner grading system.^12,23,25 Despite the limited study and participant numbers, a significant increase was found in complete wound healing for Wagner grades II, III, and IV in the HBOT group, while significant decreases were observed in minor/distal amputations for grade III, major/proximal amputations for grade IV, and the risk of operative debridement for grade II patients. The rates of nonhealing wounds, graft or flap closure requirements, and adverse side effects between the groups were not significantly different. The most severe side effects of HBOT were oxygen-induced seizures and tympanic membrane perforations, both of which were remedied without further complications.²⁷ The inability to equalize ear pressure was reported in three trials.^15,20,27

An observational study³⁰ reported that healing of severe DFU (Wagner III–IV) wounds increased from 54% to 60% and up to 75% for patients who completed the prescribed number of hyperbaric treatments. Furthermore, a retrospective observational study with 130 participants found no differences between patients with grades II and V ulcers; however, the treatment was more effective in patients with grades III (87.5%) and IV (84.6%) ulcers.³²

Fat grafting and platelet-rich plasma (PRP) are potential wound-healing approaches.^40–42 Fat grafts containing adipose-derived stem cells promote neovascularization, whereas PRP enhances fat graft survival and adipose-derived stem cell differentiation, aiding in angiogenesis and tissue remodeling. However, the ideal PRP concentration for adipose-derived stem cell growth is unclear. Limited clinical evidence exists on the use of these techniques in combination; therefore, further research is needed to establish the efficacy of this method.^40,41 One study compared control, fat grafting, and fat/PRP grafting groups over 12 weeks and identified no significant differences in wound size changes, complete healing rates, or healing duration.⁴³

This study involved several constraints: the analysis was restricted to a small number of participants who were only classified into Wagner grades II, III, and IV, which may have affected the precision of the findings; the quality of the trials was moderate because of a high-performance bias; and notable heterogeneity and bias existed in one of the parameters, complete wound healing at 8 weeks or more. In addition, this study involved heterogeneous conditions of wounds with standard treatments that may not have been clearly defined and a variable number of HBOT sessions. Therefore, our results should be used with caution.

CONCLUSIONS

HBOT, in combination with standard wound treatment, may result in significant wound healing at 8 weeks or more after the final HBOT session and a reduction in the necessity for minor and major amputations in DFU patients. Wagner grades II, III, and IV ulcers had higher likelihoods of wound healing, and minor and major amputations were less likely for grade III and IV ulcers, respectively. In addition, operative debridements were required less often for Wagner grade II patients. However, the limitations of our study should be considered before interpreting the results. Prospective randomized studies are essential to resolve this important issue.

DISCLOSURE

The authors have no financial interest to declare in relation to the content of this article.