Diabetic foot ulcers are among the most disabling complications of diabetes. Between 19% and 34% of people with diabetes develop a foot ulcer during their lifetime, and these lesions precede approximately 85% of non-traumatic lower-limb amputations.(1) The lifetime risk of a foot ulcer is therefore comparable to that of other severe microvascular complications, and the wound, once established, behaves as a chronic condition in its own right, with recurrence and reulceration common after the first episode.(1) Incident ulcers carry a substantial subsequent risk of amputation, and population analyses show that amputations secondary to diabetes and peripheral artery disease have not declined in several settings despite improvements in revascularization and wound care.(2,3) The consequences extend beyond limb loss, because amputation is followed by high mortality, reduced mobility and heavy use of health services, which makes wound closure a central goal of diabetic foot management.

The rationale for hyperbaric oxygen therapy rests on the correction of local tissue hypoxia. Ulcer tissue in diabetes shows impaired oxygen delivery, and oxygen tension below the threshold required for fibroblast proliferation, collagen deposition, angiogenesis and leukocyte bacterial killing.(4) Hyperbaric oxygen increases the amount of dissolved oxygen in plasma, reduces edema, stimulates neovascularization and supports granulation tissue formation, which provides a biologically coherent basis for using it as an adjunct to standard wound care.(4) The physiological argument is nevertheless specific: hyperbaric exposure raises arterial and tissue oxygen tension, corrects the hypoxia that stalls fibroblast activity and collagen cross-linking, and augments the oxidative burst that neutrophils use to clear infection, effects that have been documented in wound models and in human tissue studies.(4) The intervention is nevertheless expensive, requires specialized chambers and repeated attendance, and carries a small risk of barotrauma, which means that its use should be justified by evidence of a meaningful clinical benefit rather than by pathophysiological plausibility alone.

The available evidence is conflicting. A Cochrane review concluded that hyperbaric oxygen improved short-term healing of diabetic foot ulcers but found insufficient evidence for its effect on major amputation, and the confidence intervals were wide.(5) Subsequent meta-analyses reached opposite conclusions about healing and amputation, in part because they pooled small trials of variable quality and different ulcer populations.(6-8) Clinical guidelines list hyperbaric oxygen among adjunctive interventions but attach low certainty to the recommendation for routine use, and the most recent guideline updates continue to place it below standard care elements such as offloading, revascularization and infection control.(9) Since 2016, several randomized trials and observational cohorts have been published that were not captured consistently by earlier syntheses, including a sham-controlled trial and a multicentre trial in ischemic ulcers, and a synthesis restricted to the current treatment era is lacking.

The practical stakes reinforce the need for a current synthesis. Hyperbaric chambers are a scarce resource, treatment courses require twenty to forty sessions, and the cost per course is substantial, so a decision to treat commits patients and services to a resource-intensive pathway whose benefit has not been established across the full spectrum of ulcer types. At the same time, standard wound care has evolved, with structured offloading, infection control and revascularization now defined by updated guidelines, which means that the comparator in older trials may differ from current practice and that the incremental value of hyperbaric oxygen must be judged against this modern baseline.(9) These considerations make an updated synthesis valuable for clinicians who must decide, case by case, whether the expected gain justifies the cost and the commitment of repeated attendance.

The objective of this review was to synthesize the evidence on the effects of hyperbaric oxygen therapy on ulcer healing and major lower-limb amputation in adults with diabetic foot ulcers compared with standard wound care alone, and to characterize the adverse events, mortality and secondary outcomes reported across studies.

This review was developed in accordance with the PRISMA 2020 statement.( 10) Studies were eligible when they met the following criteria. The population comprised adults with a diabetic foot ulcer, including neuropathic, ischemic and neuroischemic ulcers, with no restriction by ulcer grade. The intervention was systemic hyperbaric oxygen therapy delivered in a chamber as an adjunct to standard wound care, and the comparison was standard wound care alone or an equivalent control without hyperbaric oxygen. The primary outcomes were complete ulcer healing and amputation, including major amputation above the ankle and minor amputation below the ankle.

Complete healing was defined as full epithelialization of the target ulcer, and major amputation as an amputation above the ankle; when a study used a different definition, the study definition was retained and noted. Randomized trials, prospective comparative studies, retrospective cohorts and case-control studies were eligible. Ulcer severity was graded in the included studies with the Wagner system or the University of Texas system, and the definition of standard care followed the current classification and infection guidelines.(11,12) Case reports, case series without a comparator, reviews, animal studies, studies of local or topical oxygen delivery, and studies restricted to outcomes other than healing or amputation were excluded.

Duplicate records across the four sources were removed before screening by matching digital object identifiers and, when identifiers were absent, by matching titles and years. The remaining records were screened in two stages. In the title and abstract stage, records clearly outside the scope were excluded, and all uncertain records were carried forward to full-text assessment.

The objective of this review was to determine whether hyperbaric oxygen therapy improves healing and reduces major amputation in diabetic foot ulcers compared with standard care. The synthesis indicates that healing and amputation do not answer the question in the same direction. Small randomized trials and several observational cohorts reported large improvements in healing, whereas the two largest trials with the most rigorous control groups found no significant benefit for either healing or amputation, and the observational evidence for amputation is dominated by single-arm cohorts in which the crude rates reflect case mix as much as treatment effect.

The certainty of the evidence is low for healing and very low for amputation. For healing, the randomized evidence is inconsistent and dominated by small unblinded trials, and the observational evidence is at high risk of bias. For amputation, the two adequately controlled trials are imprecise, and the observational estimates come from single-arm cohorts. Judged against the criteria of risk of bias, inconsistency, indirectness, imprecision and publication bias, the body of evidence does not support a high-confidence conclusion in either direction, and the plausible benefit is therefore best described as uncertain rather than established or refuted. Publication bias is plausible in this literature, because the smallest trials with the largest effects dominate the favorable end of the evidence and the observational series that report null or modest outcomes are less likely to be submitted or published, but the presence of two large neutral trials mitigates the concern that the overall direction is driven entirely by selective reporting.

Several features of the review support its interpretation. The search was executed across four complementary databases with Scopus as the first source, the queries were recorded verbatim with their dates and yields, and the flow of records is documented arithmetically in the PRISMA diagram rather than approximated. The eligibility criteria were applied to a large candidate set, the risk of bias was assessed with instruments matched to the design of each study, and every effect estimate in the narrative is traceable to a named report. The restriction to the 2016 to 2025 window focuses the review on the era in which the two sham or multicentre controlled trials were published, which is the period most relevant to current practice.

The healing findings are consistent with the landmark trial by Löndahl, which reported more complete healing at one year but no difference in amputation, and with the Cochrane synthesis, which found a benefit for healing but low-certainty evidence for amputation.(5,32) More recent meta-analyses echo the same divergence. Sharma reported improved healing with hyperbaric oxygen in controlled trials, and Zhang reported a reduction in major amputation when smaller and predominantly observational evidence was pooled, whereas the overviews emphasize that the estimates depend heavily on the inclusion of uncontrolled series.(6,7,33) Reviews restricted to ischemic ulcers reached cautious conclusions, with Brouwer finding no consistent benefit in the presence of arterial insufficiency and Lalieu noting that the nonischemic population is where most positive signals arise.(34,35) The stratified analysis by Wagner grade reported the largest gains in advanced grades only when treatment adherence was high, which matches the dose-response pattern observed here in the Ennis and Lalieu cohorts.(23,26,36) The most recent systematic reviews of clinical outcomes and of oxygen therapy for chronic wounds reach the same conclusion, namely that the benefit is plausible but conditional on patient selection and reporting quality.(37,38)

For amputation, the two sham or adequately controlled trials deserve the greatest weight. Fedorko used a sham exposure and found no reduction in amputation indication, and the DAMO2CLES trial found point estimates favoring hyperbaric oxygen that did not reach significance, with a signal of benefit only in the subgroup that completed the protocol.(15,16) This contrast between per-protocol and intention-to-treat results suggests that session completion and patient selection, rather than the intervention itself, may explain part of the apparent benefit in uncontrolled series. The high five-year amputation estimate in the Danish registry cohort is a reminder that patients referred for hyperbaric oxygen are a selected, severely affected population, which inflates crude amputation rates in single-arm studies.(30) A Cochrane overview and the meta-analyses that included only randomized evidence likewise found no convincing reduction in amputation, which reinforces the conclusion that the amputation benefit reported in small trials is fragile.(5,6) The design difference between the two large trials and the smaller ones is the central explanation for their divergent results. Fedorko used a sham arm in which participants breathed air at reduced pressure, so that neither patients nor assessors could distinguish the intervention from the control, and the trial therefore isolates the specific effect of hyperbaric oxygen rather than the effect of being treated in a chamber. The DAMO2CLES trial, although not sham-controlled, was multicentre and prespecified limb salvage and wound healing at twelve months, and it enrolled a uniformly ischemic population. The smaller trials, by contrast, were single-centre, unblinded and used surrogate or short-term wound measures, and their very large effect sizes are therefore best interpreted as the upper bound of the plausible effect rather than as the expected effect in routine practice.

The biological plausibility of a healing effect is strongest in neuropathic ulcers, where the tissue is hypoxic but still perfused, and weakest in ischemic ulcers, where the inflow problem cannot be corrected by raising dissolved oxygen alone.(4) This distinction fits the pattern in the present review, because the positive signals came from nonischemic or neuropathic populations and the largest null results came from the ischemic DAMO2CLES population, and it also explains why offloading and revascularization remain the first-line interventions in the guideline hierarchy.(9,12) The dose-response relationship between the number of completed sessions and healing or amputation observed in the Ennis and Lalieu cohorts is consistent with this mechanism, although it may equally reflect selection of patients who were well enough to attend repeated sessions.(23,26) The same logic explains the emphasis on vascular assessment before referral, because a hyperbaric course in a limb with uncorrected inflow is unlikely to restore tissue oxygen to the levels required for repair, and the clinical guidelines therefore place revascularization ahead of adjunctive oxygen.(12)

Mortality was reported in only three studies, and in each of them it tracked the severity of the underlying disease rather than the intervention. The Kaplan cohort recorded higher mortality in the amputated group, the Danish cohort recorded a five-year mortality above half, and the small prospective study reported lower mortality in the hyperbaric group without adjustment for baseline risk.(22,30,31) These figures cannot support a survival claim for hyperbaric oxygen, but they confirm that the populations enrolled in this literature are at high baseline risk, which magnifies the influence of selection bias in the observational comparisons.

The safety profile appears acceptable but is incompletely reported. Barotrauma was the most common adverse event, and seizure risk in patients with a seizure history has been examined separately, with hyperbaric oxygen appearing tolerable in non-emergent settings.(24,39) The economic analysis based on the DAMO2CLES trial found no significant difference in cost or quality-adjusted life-years, which weakens the case for routine use outside selected patients.(40) Cost-effectiveness syntheses and shared decision-making reviews converge on the same operational conclusion, namely that hyperbaric oxygen should be offered selectively and with explicit patient information about uncertain benefit.(41,42)

Several sources of heterogeneity explain the conflicting results. The trials differed in ulcer grade and ischemia, with positive signals concentrated in neuropathic and nonischemic ulcers and null results in ischemic populations. Hyperbaric protocols varied in pressure, session duration and number, and adherence was low in the largest trials, with more than a third of patients failing to complete the protocol in the DAMO2CLES trial.( 16) Control groups differed as well, because sham exposure was used in only one trial, and the smaller trials that reported the largest effects were unblinded. Outcome definitions also differed, because complete healing was defined variously as full epithelialization or as an 80% reduction in surface area, and major amputation was defined inconsistently across studies. These differences prevented a pooled estimate and mean that the apparently favorable observational rates cannot be interpreted as treatment effects. The heterogeneity is therefore more than a technical obstacle to pooling, because it is itself an informative finding that points to the subgroups and the adherence conditions under which hyperbaric oxygen is most likely to add clinical value.

The findings must also be read against the components of standard care. Offloading, revascularization and infection control have their own evidence base and remain the first priorities in the guideline hierarchy, and hyperbaric oxygen is an adjunct rather than a substitute for any of them.(9,12) The trials included here applied hyperbaric oxygen on top of comprehensive wound care, and the largest of them found no additional benefit, which is consistent with the view that the marginal effect of hyperbaric oxygen is small once the basic elements of care are optimized. The positive signals in the smaller trials may reflect differences in baseline care intensity as much as the biological effect of oxygen, because none of those trials used a sham control.

For podiatric practice, the review has a direct operational meaning. Diabetic foot ulcers are managed in multidisciplinary foot clinics in which podiatrists coordinate offloading, debridement and footwear, and the decision to refer for hyperbaric oxygen usually arises when a wound fails to close despite these measures. The evidence reviewed here indicates that such a referral is reasonable when the ulcer is neuropathic or nonischemic and when the patient can attend a full course of sessions, but that the expected gain is in wound closure rather than in limb salvage, and that this expectation should be communicated before treatment begins.

For practice, the findings support offering hyperbaric oxygen as an adjunct in selected patients with neuropathic or nonischemic ulcers that have not responded to standard wound care, while counseling that the effect on major amputation is unproven and that the treatment requires substantial attendance and carries a small risk of barotrauma. For research, adequately powered sham-controlled trials that stratify by ischemia, report protocol completion and measure major amputation as a prespecified endpoint are the clearest route to resolving the question that the current literature leaves open.

Future trials would also benefit from standardized outcome definitions. The present literature used at least three definitions of complete healing and two definitions of major amputation, and this variation prevented any pooled estimate and weakens comparisons between trials. Adopting the outcome definitions proposed in the current guideline framework, reporting major and minor amputation separately, and prespecifying session adherence as a stratification factor would make the next generation of trials directly combinable.(9,11) Until such trials exist, the clinical decision will continue to rest on individualized judgment informed by the low-certainty evidence summarized here.

Hyperbaric oxygen therapy may improve complete ulcer healing in selected patients with diabetic foot ulcers, but the most rigorous controlled trials do not demonstrate a consistent reduction in major amputation. The evidence for amputation is further limited by the predominance of single-arm observational cohorts, and adverse events appear acceptable although they are inconsistently reported. The certainty of the evidence for both healing and amputation remains low, and the decision to use hyperbaric oxygen should therefore be individualized and accompanied by explicit discussion of the uncertain benefit.