Hyperbaric Oxygen Therapy as an Adjunctive Treatment for Diabetic Foot Wounds: A Comprehensive Review With Case Studies

Abstract: Complications associated with diabetes are often expensive to treat, and commonly include foot ulceration. While most diabetic foot ulcers heal with standard treatment, when standard treatment measures fail, adjunctive therapies must be considered. We review the theory and evidence for the use of systemic Hyperbaric Oxygen Therapy (HBOT) as an adjunctive treatment for chronic lower extremity diabetic ulceration. Two clinical cases of patients treated with HBOT for refractory diabetic foot ulceration at Georgetown University Hospital are presented. A growing body of evidence suggests that oxygen plays a major role in whether wounds heal normally or remain nonhealing chronic wounds. An oxygen gradient, from high oxygen levels in the edges of the wound to hypoxic conditions in the wound center, seems to optimally stimulate wound healing. Wounds that are instead surrounded by hypoxic tissue lack this oxygen gradient and seem less likely to heal. In wounds with adequate perfusion, HBOT may overcome periwound hypoxia to create an oxygen gradient and stimulate healing of otherwise nonhealing wounds. The clinical trials assessing the effectiveness of HBOT in diabetic wound healing have been inconclusive. However, considering the evidence supporting oxygen’s role in wound healing and the potential for HBOT to decrease medical costs related to the care of chronic diabetic ulcers, there is a need for more extensive clinical trials to evaluate HBOT efficacy.
Address correspondence to: John S. Steinberg, DPM Georgetown University Hospital Georgetown University School of Medicine 3800 Reservoir Rd. NW, Bles First Floor Washington, DC 20007 Email: jss5@gunet.georgetown.edu
     Diabetes has become a global epidemic. As the prevalence of diabetes has increased, so has the burden on the healthcare system to provide treatment for the complications associated with the disease. These complications are often expensive to treat and may include foot ulceration, secondary infection, and limb amputation.^1,2      Every year, 1.9% of persons with diabetes develop foot ulcers.³ These wounds are often refractory to standard treatment, ultimately requiring amputation in 15%–20% of diabetic patients within 5 years of ulcer development.⁴ In the United States alone, the cost of diabetic ulcer treatment accounts for more than half of the $4.6–$13.7 billion spent each year to treat diabetic peripheral neuropathy.⁵ Such costs intensify the need for improved, cost-effective methods in the prevention and therapy of diabetic ulcers.      In diabetic patients, peripheral neuropathy seems to be the most important risk factor in the development of foot ulceration.⁶ It has been shown that up to 80% of diabetic foot wounds are associated with diabetic peripheral neuropathy.⁷ Loss of innervation to sweat glands due to peripheral neuropathy may cause the skin surface to become dry and thus more prone to cracking, infection, and subsequent tissue damage and ulceration. The effects of peripheral neuropathy on protective sensation and on foot biomechanics have also been reported to drive diabetic ulcer formation.⁸ The loss of protective sensation due to nociceptive neuron loss can allow wounds to go unnoticed (ie, unrecognized trauma), increasing the likelihood of infection and development of a chronic, nonhealing wound. Altered foot biomechanics due to peripheral neuropathy may cause inappropriate weight distribution, which can exert high pressure on focal weight bearing areas of the foot. Ulcers may develop over time as patients apply constant micro-trauma to the skin. Tissue ischemia, uncontrolled hyperglycemia, infection, poor nutrition, and improper shoe gear also contribute to the chronic, nonhealing nature of diabetic ulcers. Over time these ulcers may involve the deeper layers of soft tissues including muscles, tendons, and ligaments, which may progress to involve pedal osseous structures if untreated. Treatment of osteomyelitis may involve long-term antibiotic therapy, surgical debridement, and possible amputation of all or part of the affected limb.⁹ More than 85% of diabetic amputations are precipitated by a foot ulcer deteriorating to deep infection or gangrene,¹⁰ underscoring the importance of early and aggressive intervention in diabetic wound healing.      Standard treatment of diabetic foot ulceration may involve the methods listed in Figure 1.^9,11,12 However, of these methods, only off-loading has been proven effective in clinical trials.⁹ Most ulcers heal with a combination of debridement, offloading, and moist healing environment, but when these fail, adjunctive therapies are considered. Commonly used wound healing adjuncts include negative pressure wound therapy¹³ and growth factors that can be applied topically to the wound.¹⁴ Granulocyte colony stimulating factor (GCSF)¹⁵ and hyperbaric oxygen¹⁶ can be administered as systemic treatment adjuncts. Of these therapies, evidence-based guidelines only support the systemic administration of GCSF for infected ulcers and hyperbaric oxygen therapy (HBOT) for chronic, nonhealing ulcers.⁹ This review is limited to the theory and evidence found for the use of HBOT as an adjunctive treatment for chronic lower extremity diabetic ulceration.

Vascular Disease and Wound Healing

     Patients with diabetes are at high risk for vascular disease related ulcerations. Diabetes predisposes these patients to ulceration through decreased oxygen delivery to the skin and wound bed due to both macro- and microcirculatory disorders. There is a clear association between vascular disease and hyperglycemic control with measured endpoints of cardiovascular disease, cerebrovascular accidents, retinopathy, nephropathy and nonfatal microvascular disease.¹⁷ The DECODE study has shown a nearly two-fold increase in cardiovascular disease in type 2 diabetic individuals,¹⁸ highlighting the importance of tight glucose control in association with both cardiovascular disease and wound healing.      Microvascular complications arise due to thickening of capillary basement membranes secondary to nonenzymatic glycosylation in the setting of hyperglycemia. This in turn can lead to nitric oxide pathway dysfunction and impaired intracellular signaling with a reduction in effective and efficient tissue perfusion, impairing oxygen delivery to a wound.      Macrovascular disease also frequently develops in patients with diabetes. Curiously, macrovascular disease affects the lower extremity to a greater extent than the upper extremity, although the reason for this remains unclear. Atherosclerosis of large vessels may be a result of hyperglycemia, increased free radical production and oxidized lipids. In addition to atherosclerotic plaque production, increased vascular smooth muscle proliferation has been found in patients with diabetes and vascular disease.      Ischemic limbs can be divided into five subgroups based on the ankle brachial index (ABI). Marston et al¹⁹ reviewed 169 limbs having an element of peripheral arterial disease (PAD). They found a correlation with increased risk of limb loss with increasing severity of PAD.¹⁹ This underscores the importance of adequate blood supply for the delivery of oxygen and nutrients in diabetic wound healing. Once it has been determined that adequate macrovascular flow is present, transcutaneous oxygen measurements can be taken to evaluate the effectiveness of hyperbaric oxygen therapy.      Role of oxygen in the healing process. The physiological pathways involved in wound healing are complicated and remain incompletely elucidated. As basic science strives to better understand the complexities involved in healing, a growing body of research has emerged to suggest that oxygen plays a major role in whether wounds heal normally or remain nonhealing, chronic wounds.      Wound hypoxia is directly caused by ischemia due to the disruption of the local microcirculation. The inflammatory and immune responses to injury exacerbate this hypoxia by increasing the oxygen demand in the wound bed. Fibrin, which is part of the immediate clotting response, stimulates platelets to release cytokines that recruit macrophages and fibroblasts to the site of injury. Leukocytes also migrate to the site of injury, and their increased activity incurs further local oxygen debt.¹⁶ This initial wound hypoxia persists in chronic wounds.²⁰      Hypoxia is required to initiate some elements of the healing process, while other elements are dependent on the presence of oxygen including the oxidative killing of bacteria by macrophages²¹ and the post-translational modification and deposition of collagen by fibroblasts.²² Immediately after injury, hypoxia and high lactate levels in the wound stimulate neovascularization (angiogenesis, endothelial cell growth),^23,24 adherence of leukocytes, and the formation of collagen and bone. However, many of these processes, although initially stimulated by hypoxia, actually require oxygen to be sustained. Thus, when wounds fail to heal as in chronic, nonhealing ulcers, the persisting hypoxia may actually inhibit or impair neovascularization, collagen production, cellular energy (ATP) production, and the oxidative killing of bacteria by macrophages.²³      Oxygen tension in wounds may be less than 3 mmHg; however, oxygen tensions in the periphery of the wound, or periwound area, typically hover close to 20 mmHg.²⁵ Interestingly, this oxygen gradient, from high oxygen levels in the edges of the wound to hypoxic conditions in the wound center, seems to optimally stimulate wound healing.²⁶ Wounds that are instead surrounded by hypoxic tissue lack this oxygen gradient and seem less likely to heal.^25,27 If successful healing depends on an adequate supply of oxygen in the periwound area, can the periwound hypoxia observed in chronic, nonhealing ulcers be reversed to heal these refractory wounds? It is on this premise of reversing periwound hypoxia, and thus stimulating healing of otherwise nonhealing wounds, that HBOT’s proponents argue its efficacy in healing lower extremity, chronic diabetic ulcers.      In a study of 20 patients with nonhealing ulcers, TcPO2 measurements were found to be less than 10 mmHg in ambient air at the periphery of the wound. When these patients were administered 100% oxygen at 2.5 atmospheres, wounds with a TcPO2 less than 50 mmHg showed consistent failure to heal while patients with a TcPO2 greater than 100 mmHg had improved healing compared to controls.²⁸ Other studies evaluating healing and hyperbaric oxygen therapy have demonstrated similar results.²⁹      On the other hand, the repeated cycles of sudden hyperoxygenation of hypoxic tissue involved in HBOT may, in theory, induce an ischemia-reperfusion injury. The exposure of hypoxic tissue to supraphysiologic levels of oxygen has been investigated in relation to the production of reactive oxygen species and possible ischemia-reperfusion injuries. Gürdöl et al³⁰ looked at the production of malondialdehyde (MDA) and advanced oxidation protein products (AOPP) before, during, and after HBOT. These are both markers of oxidative stress and oxidant-mediated injury. The 30 patients they examined showed an increase of MDA after the first therapy session, but remained unchanged after 15 sessions of HBOT. AOPP levels decreased after the end of therapy, suggesting that increased oxygenation of tissues in the diabetic subject due to HBOT may activate a protective effect through endogenous factors to guard against ischemia-reperfusion injuries.

Systemic Administration of HBOT

     Hyperbaric oxygen therapy is a treatment modality in which the patient breathes 100% oxygen at pressures greater than 1 atmosphere, generally between 2.0 and 3.0 atmospheres absolute (ATA). HBOT typically consists of 15–30 treatments of 60–120 minutes each, with treatment frequency either once or twice daily. Treatments may take place in a single-capacity, monoplace chamber (Figure 2A³¹) pressurized with 100% oxygen, or in a larger, multiplace chamber (Figure 2B³²) that can accommodate several patients at once. In the latter case, the space can be quite complicated, with several patients breathing 100% oxygen delivered via endotracheal tube, mask, or head tent.³³      In the United States, where HBOT is typically used with pressures of 2.0–2.4 ATA, arterial oxygen tension (PaO2) of about 1200 mmHg can be attained. At such elevated PaO2, hemoglobin is fully saturated (SaO2 = 100%) and there is a significant increase in the amount of oxygen dissolved in the plasma. This increase in oxygen dissolved in plasma accounts for a substantial increase in the distance that oxygen can diffuse from the end-arteriole into tissues, from 60 microns normally to 250 microns with HBOT.³⁴      Due to this increased diffusion of oxygen into tissues, HBOT may be able to overcome periwound hypoxia, but only in situations where perfusion is sufficient, as in the case of hypoxia due to small vessel disease or edema. Hopf et al³⁵ evaluated angiogenesis and hyperbaric oxygen in a murine model. They evaluated vessel ingrowth to a Matrigel plug. Results showed hyperoxia increased angiogenesis significantly and in a dose-dependent manner. Increased oxygen tension was also seen with levels of 439 mmHg ± 232 mmHg at 2.5 ATA in these subjects.³⁵ However, in situations where perfusion is insufficient, as in the case of ischemia due to the disease of large blood vessels, the increase in the diffusion of oxygen into tissues due to HBOT is not enough to overcome periwound hypoxia.³⁶ Therefore, HBOT is not a panacea for all chronic, nonhealing wounds, but rather only appropriate for select problem wounds.      When used as adjunctive treatment for selected problem wounds, HBOT in humans has been associated with several beneficial effects (Figure 3).³⁶ Hyper-oxygenation of periwound tissue due to HBOT in selected wounds decreases edema via the vasoconstrictive effects of oxygen on vasculature and enhances the oxidative killing of bacteria by leukocytes and macrophages.^33,38,39 By improving periwound tissue oxygenation, previously hypoxic cells relying on glycolysis to produce ATP anaerobically can now take advantage of aerobic respiration to produce ATP much more efficiently. HBOT potentiates the effects of antibiotics, especially the aminoglycosides,⁴⁰ and the improved oxygen gradient from wound periphery to hypoxic wound center promotes neoangiogenesis.⁴¹ There is also evidence that epithelial migration as well as collagen production and deposition by fibroblasts are enhanced by HBOT.²⁰ Finally, studies in rabbits suggest that HBOT may increase the uptake of platelet-derived growth factor beta peptide (PDGF-BB), thereby improving the formation of granulation tissue and the likelihood of successful healing.^42,43      In some HBOT studies^44–47 wounds that had not healed completely by the end of the study period raised questions concerning HBOT efficacy. The effectiveness of HBOT does not lie solely in the direct healing of lower extremity wounds, but rather in the changes it brings to the periwound area. In addition to the production of growth factors, angiogenesis is stimulated through the production and release of VEGF, which can lead to activation and mobilization of local stem cells. HBOT has been shown to reduce inflammation and apoptosis in the wound allowing for a more hospitable environment for healing to occur.      Adverse effects due to HBOT, although rare, have also been reported. Damage to the ears (2%–4% of patients), sinuses (< 2% of patients), and lungs (estimated pneumothorax incidence < 1 in 1 million HBOT treatments) attributed to barotrauma, temporary exacerbation of myopia (< 10% of patients), generalized seizures believed to be induced by hyperoxia (0.03% of patients), and claustrophobia have been reported.⁴⁸ Accordingly, although these effects are rare, HBOT cannot be considered a completely benign treatment option.

Clinical Evidence

     The Undersea and Hyperbaric Medical Society (UHMS) currently recommends HBOT for the adjunctive treatment of selected problem wounds, including sufficiently perfused lower extremity diabetic ulcers (Figure 4).³³ In 2002, the Centers for Medicare and Medicaid Services (CMS) started approving reimbursement for the use of HBOT as an adjunctive treatment for lower extremity chronic wounds in diabetic patients.⁴⁹ Blue Cross and Aetna now also reimburse the adjunctive use of HBOT in the treatment of lower extremity chronic diabetic ulcers. Although recognized and reimbursed by insurance companies, is there sufficient clinical evidence to justify the widespread adoption of HBOT as an adjunctive treatment of lower extremity, chronic diabetic ulcers?      Clinical studies examining the use of HBOT in the adjunctive treatment of diabetic foot ulcers have been systematically reviewed on several occasions,^50–52 with the review performed by the Cochrane Collaboration⁵² being the most rigorous. The common denominator among these reviews is the conclusion that there is little evidence to suggest HBOT expedites ulcer healing, but HBOT might significantly decrease the rate of major amputation. These results, however, should be interpreted with caution because the studies from which they are derived are lower on the evidence-based medicine grading scale.      In its systematic review, the Cochrane Collaboration found 26 studies examining HBOT in the context of chronic wound healing. Of these studies, only five were considered sufficiently rigorous enough to deserve further review, four of which^44–47 looked specifically at the effect of HBOT in healing diabetic foot ulcers. The methodological quality of these four studies were graded and assigned a Jadad score.⁵³ Each of three core criterion (randomization, double-blinding, and a description of how withdrawals are handled) earns 1 point. Two more points are assigned to studies using a reliable randomization method and a sham treatment arm. The studies conducted by Doctor et al⁴⁴ and Faglia et al⁴⁵ both received a Jadad score of 2, the study by Lin et al⁴⁶ earned 4 points, and the 2003 study by Abidia et al⁴⁷ earned the maximum possible 5 points (Table 2).      Table 2 highlights the methodology, while Table 3 compares the outcomes measured in each of the four studies. Two of the four studies lacked a sham treatment arm.^44,45 Three of the four studies lacked an intention to treat (ITT) population,^44–46 an epidemiological metric designed to avoid concluding a completely ineffective treatment to be beneficial. Two^44,45 of the three studies lacking the ITT population reported an improvement in major amputation rate with HBOT, while the only study to include an ITT population⁴⁷ reported the same rate of major amputation with and without HBOT (Table 3). Three of the four studies^45–47 may have suffered from bias due to management decisions made in a nonblinded fashion. For example, the decision of whether or not to amputate may have been made with the knowledge of whether or not HBOT treatment had been administered.⁵²      The lack of standardized patient inclusion criteria is also evident in the last column of Table 2. Patients were either included if the lesion was chronic⁴⁴, if the size of the lesion satisfied a specific area range criteria (1 cm–10 cm) for > 6 weeks⁴⁷, or if the severity of the lesion was classified within a specific range of Wagner severity (2, 3, 4 in Faglia et al⁴⁵; 0, 1, 2 in Lin et al⁴⁶). Clearly the ulcers treated in each of the four studies could have been quite different, and accordingly some studies would be more or less likely to achieve healing than others. Therefore, even when two or more studies report the same outcome measure (eg, major amputation rate), the results of any comparisons between studies should be carefully examined.      In addition to differences in ulcers treated, the follow-up period and the outcome measures reported also varied from study to study (Table 3). With respect to variability in follow-up period, data was either reported immediately after HBOT treatment,⁴⁶ upon hospital discharge,^44,45 or at intervals up to and including one year following treatment.⁴⁷ The outcome measures also varied significantly among the four studies. While the 1992 study by Doctor et al⁴⁴ only reported the proportion of patients requiring major and minor amputation, the 1996 study by Faglia et al⁴⁵ reported the proportion of patients requiring major amputation as well as the transcutaneous oxygen tension (TcPO2) in the affected foot after HBOT treatment. In the highest quality and most thorough study, conducted by Abidia et al⁴⁷ in 2003, the proportion of ulcers healed at the end of HBOT treatment (6 weeks), at 6 months, and at 1 year was reported in addition to the proportion of patients requiring major and minor amputation. On the other end of the spectrum, the 2001 study by Lin et al⁴⁶ did not report clinical endpoints of any kind, only transcutaneous oxygen tension (TcPO2) in the affected foot after HBOT treatment.      A significant improvement in ulcer healing was not demonstrated in any of the four studies. A significant decrease in amputation rate was observed in two of the studies^44,45; however, this beneficial effect was not observed in the only study with both a sham treatment arm and an ITT population⁴⁷ (Table 4). In a study by Kessler et al⁵⁶, 28 patients were examined and followed to evaluate healing of foot ulcers with nonischemic chronic diabetic foot ulcers. After completion of hyperbaric oxygen therapy the ulcer size decreased significantly (41.8 ± 25.5% versus 21.7 ± 16.9% in controls). They found oxygen tension to increase 20 fold over baseline measurement under treatment conditions. The difference in ulcer size was lost after 30 days post treatment. Kessler et al⁵⁶ concluded HBOT doubles the mean healing rate of diabetic ulcerations after hyperbaric treatment.      Recent work with reconstructed peripheral nerves has shown increased healing potential when treated with hyperbaric oxygen. The study evaluated transected and repaired sciatic nerves of rats and the effect of HBOT. Researchers found a 27.8% increase in the number of axons as well as increased capillary formation in the treatment group.⁵⁷ These findings led to improved nerve regeneration, as evaluated through motor latency, EMG, and histopathologic analysis compared to controls. This new information may lead to new uses of HBOT in the treatment of peripheral neuropathy. Previous studies have validated the regenerative potential of nerves in hyperbaric treatment.^58,59

Case Reports

     Patient 1 is a 56-year-old man with diabetes, vascular disease, and end stage renal disease. He underwent a transmetatarsal amputation for chronic osteomyelitis in the forefoot with subsequent dehiscence of the operative site. He underwent a revisional amputation at the Lisfranc joint, infra-popliteal angioplasty, and local wound care to the operative site with cadexomer-iodine.      Before HBOT, the distal incision again broke down, and measured 6.1 cm x 0.9 cm x 0.4 cm (Figure 5). The wound was cultured and no organisms were isolated. The patient then underwent 30 treatment sessions of HBOT. Four weeks after the final HBO treatment, the distal wound showed two small ulcerations with healthy margins and a fibrous base (Figure 6). In this case, hyperbaric oxygen therapy may have potentiated wound healing and allowed salvage of the patient’s leg to avoid additional lengthy and invasive surgical procedures.      Patient 2 is an 82-year-old woman with a medical history complicated by peripheral vascular disease and hypertension. She initially underwent a partial hallux amputation for a chronic nonhealing wound at the distal aspect of the toe. The patient had an angiogram and subsequent angioplasty of the popliteal artery for stenosis prior to the amputation. The distal soft tissues became ischemic and necrotic (Figure 7A) and the amputation was revised to a partial first ray amputation with primary closure (Figure 7B).      The first ray incision began to turn dusky postoperatively (Figure 7C) and a small dehiscence occurred measuring 1.2 cm x 0.6 cm. A dry eschar formed over the nonhealing wound. The deep soft tissues of the wound base were cultured and deemed to be free of infection. The patient received 30 sessions of HBOT. The wound began to show signs of healing with new vascular ingrowth and tissue formation. A small hyperkeratotic area remained that eventually healed with local wound care of cadexomer-iodine.

Conclusion

     The lack of oxygen in the periphery of a chronic wound may prevent healing. In select wounds with adequate perfusion, this hypoxia may be reversed by the systemic administration of HBOT. Clinical trials assessing the effectiveness of HBOT in diabetic wound healing have been inconclusive. Additionally, there are possible cost savings achieved by reducing diabetic amputations and hospitalizations.⁶⁰      Double blind, randomized, placebo controlled, prospective studies enrolling sufficient numbers of patients are needed to further elucidate the role of HBOT in healing recalcitrant diabetic ulcers. Large vessel perfusion should be optimized prior to study inclusion. Pre- and post-HBOT treatment measures should include wound size, depth, classification, and periwound PcO2. Overall time to healing, and the long-term minor and major amputation rate should be reported.      HBOT is a promising therapy for healing diabetic ulcers that have not responded to standard treatment and deserves further study.