Dose-Response Relationship of Therapeutic Oxygen: More Is Not Necessarily Better and May Be Inferior to No Supplemental Oxygen – Part 1: Proof and Nature

Abstract

In the early 1990s, the authors had consistently positive experience in the treatment of typical sports injuries and cosmetic surgical wounds with hyperbaric oxygen. These treatments generally consisted of oxygen at 2.0 atmospheres absolute (ATA) for 60 minutes. We were thus surprised to find that recognized experts in the field of hyperbaric medicine did not believe this modality to be effective for such normal wounds. Consequently, we asked Eric Kindwall, M.D., a professional acquaintance and published proponent of this prevalent view, what the bases for his beliefs on this matter were. Starting with references provided by Dr. Kindwall, we began an extensive literature review to try to reconcile our practical successes with the prevailing hyperbaric medical dogma. During our analysis, we found that the outcomes of oxygen therapy have a hormetic dose-response relationship. This finding explains the mistaken conclusions drawn concerning the treatment of uncompromised or normal wounds with hyperbaric oxygen as more oxygen will not provide benefits without limit. At some dose point, even below an inspired partial pressure (P_iO₂) of 3.0 atm, benefits will begin to decline and, in due course, more oxygen will produce negative impact in comparison with no hyperoxic supplementation.

1. Introduction

This review article concerning oxygen dose response is presented in two parts. In Part 1, we provide evidence for a dose-related response to oxygen therapy and its hormetic nature. In Part 2, we discuss the implications and consequences of dose-response with respect to the use of oxygen as a therapeutic agent. These uses include not only those applications for which it is already employed but also new indications and improved therapeutic strategies for both normobaric and hyperbaric administration.

1.1 Oxygen chemistry, physiology, and pathology

Oxygen is the third most abundant element in the universe. It is nonmetallic and highly reactive forming compounds with almost all other elements. Physiologically, in aerobic organisms, oxygen is essential for producing adenosine triphosphate (ATP), which provides chemical energy for many processes within cells including biosynthetic reactions, motility, and cell division. Further, short-lived free radicals formed from oxygen, reactive oxygen species (ROS), play an important role in many processes including the killing of pathogens by the host’s immune system and cell signaling. As a result of its many critical physiological roles, the provision of oxygen to the cells of aerobic organisms in accordance with metabolic and functional demands is a necessity for life.

Because of their reactivity, ROS such as superoxide (O₂·), hydroxyl radicals (OH·), singlet oxygen (O₂·), nitric oxide (NO·), and hydrogen peroxide (H₂O₂) have the potential to produce not only beneficial effects but also adverse ones including damage to macromolecules and disruption of cellular metabolism. A number of chemical compounds such as superoxide dismutase, glutathione reductase, catalase, thioredoxins, peroxidases, and antioxidant nutrients (e.g., alpha tocopherol, ascorbic acid, and carotenoids), generally classified as antioxidant defenses, minimize free-radical-induced damage in the body and contribute to the repair of such damage when it does occur. Despite this, chronic adverse effects of ROS are believed to play a role in aging, some disease processes, and cell death. Adverse effects from acute overexposure to oxygen (i.e., oxygen toxicity) can impact all cells, but have been found to be particularly prominent in the central nervous system, lungs, and eyes of mammalian life forms.

1.2 Clinical uses of oxygen

In medicine, oxygen is commonly administered to:

• Compensate for and/or help correct:

  • Primary delivery insufficiencies (e.g., emphysema, pneumonia, pulmonary embolism, congestive heart failure, carbon monoxide poisoning, and exceptional blood loss anemia);

  • Local delivery insufficiencies (e.g., crush injury, compartment syndrome, and acute arterial insufficiency; non-healing wounds; refractory osteomyelitis; delayed radiation injury; compromised skin flaps and grafts; thermal burns; and frostbite).

• Enhance:

  • Immune system function in bacterial killing (e.g., gas gangrene, necrotizing soft tissue infections, and intracranial abscess);

  • Inert gas elimination and bubble resolution in gas embolic disorders (i.e., arterial and venous gas embolism, and decompression sickness).

1.3 Delivery of hyperoxic gases to patients

In some cases, oxygen is effective when delivered to patients at normal barometric pressure, primarily by mask or nasal cannula. In other cases, to obtain the necessary inspired oxygen pressure, oxygen must be delivered to patients at increased barometric pressure using a whole-body chamber. This is known as hyperbaric oxygen therapy (HBO₂). In only two of the conditions listed above for medical oxygen administration is the chamber pressure associated with HBO₂ known with certainty to play a direct role in the therapy. This is in the treatment of gas embolism and decompression sickness, both of which are gas embolic/bubble disorders. In these cases, increased hydrostatic pressure reduces the size of any gas phase in closed systems in the body in inverse relationship to the increase in absolute hydrostatic pressure, thus helping to eliminate circulatory blockage and tissue deformation caused by the bubbles. Reducing gas volumes also concentrates the components of the gas phases, thus, hastening their return to solution and the ultimate dissolution of the bubbles.

The explicit details of HBO₂ administration depend on the type of chamber used. In multiplace (i.e., multiple-occupant) chambers, oronasal masks or flow-through hoods are usually used (Figure 1). In monoplace (i.e., single-occupant) chambers, the chamber is usually flushed with oxygen so that no special gas delivery apparatus is required (Figure 2). The patient simply breathes the chamber atmosphere. Less commonly in monoplace chambers, the patient breathes oxygen by an oronasal mask or flow-through hood, while the chamber is compressed and flushed with air.

Figure 1.

Clinical multiplace hyperbaric chamber (courtesy of Fink Engineering, Warana, Queensland, Australia).

Figure 2.

Clinical monoplace hyperbaric oxygen chamber (courtesy of Environmental Tectonics Corporation, Southampton, PA, U.S.A.).

2. Regulation of clinical uses of hyperoxic gases

In medical applications, oxygen is often considered to be a drug and in many countries is regulated by national healthcare authorities (e.g., the Food and Drug Administration in the United States, Health Canada in Canada, Medicines and Healthcare Products Regulatory Agency (MHRA) in the UK, and the Therapeutic Goods Administration in Australia), and requires the prescription of a licensed physician for its use. Despite this and the fact that dose-response relationships are of fundamental importance to the safe and effective use of pharmaceutical agents, the authors are aware of no broad-scope, in-depth discussion of this aspect of oxygen therapy, either hyperbaric or normobaric, and only a relatively small number of passing references or references severely limited in scope [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]. This would seem to go hand-in-glove with the fact that few studies have been conducted to investigate therapeutic outcomes systematically as a function of dose (i.e., inspired oxygen partial pressure, treatment duration, treatment frequency, total number of treatments, and changes in any of these parameters over time).

3. Objectives of this review

Since oxygen dose-response relationships are a function of a myriad of both beneficial and detrimental effects of oxygen, the ultimate outcome of any oxygen treatment will be the net result of all these factors acting at the site of interest over the time course of treatment and recovery from the treatment. As such factors are only incompletely understood, even at this time (i.e., March of 2024), establishing any particular dose-response relationship could involve a great many relatively time-consuming hyperoxic treatments. Consequently, comprehensive studies of oxygen dose effects could be difficult to design, and costly and time-consuming to carry out.

The above notwithstanding, we have found that there is information in the scientific and clinical literature sufficient to provide important definitive and practical insights into the basic nature of the oxygen dose-response relationship and some of its clinical implications. This understanding evolved as the result of an extensive literature review we conducted to examine the opinion prevalent in the hyperbaric medical community that HBO₂ is not an effective treatment modality for normal wounds.

This opinion runs contrary to our considerable practical experience that HBO₂ can be an effective adjunct for enhancing the healing of so-called normal wounds in sports and cosmetic surgery. Our view is supported by not only our own experience in the form of case reports but also by many controlled human and animal studies conducted by numerous investigators over almost 50 years as of the early 2020s. The review and analysis presented here show how the evidence for and against the effectiveness of hyperoxic therapy for normal wounds has been rationalized into a unifying concept of oxygen dose-response. Some important medical implications of this relationship are then discussed.

3.1 Normal wounds and hyperoxic therapy

Margolis et al. [13] provide a definition of normal wounds, a “wound that heals normally,” which effectively separates this category from others such as delayed effects of radiation, refractory osteomyelitis, and some lower-extremity diabetic ulcers for which special intervention like HBO₂ may be required to produce healing. Based on this definition, any uncompromised injury, regardless of its origin, whether combat, workplace, domestic, automotive, surgical, sport, recreational, or any other that does not require special intervention for ultimate resolution, would be classified as a normal wound.

Even for normal wounds, however, it has been established that oxygen levels resulting from breathing air at one atmosphere absolute (1 ATA) do not fully meet the increased metabolic demand of inflamed and healing tissue. U.S. military studies have shown that the optimal oxygen pressures for fibroblast proliferation and collagen production [14] and for macrophage migration [15] are approximately twice those found in normally oxygenated tissue. These determinations were made with in vitro cell cultures having continuous exposure to the given oxygen concentrations. Another study, conducted in vivo with rabbits, confirmed that in bone, the optimal oxygen tissue pressure for collagen synthesis is impossible to achieve when breathing air at one atmosphere [16].

Abbot et al. [17] have found that increased oxygen demand resulting from the inflammatory response, together with the effects of associated edema and fibrin in the interstitial fluid, prevent the maintenance of normal tissue oxygen levels despite the hyperemia caused by inflammation. Sheffield [18], through extensive studies of tissue oxygen pressures, has determined that oxygen levels in “normal wounds of any magnitude” are not only suboptimal for healing but are also hypoxic, and the rate at which such wounds heal is oxygen dependent. Hunt et al. [19] measured wound oxygen levels and determined that even for soft-tissue wounds in the early stages of healing under normal circumstances, more oxygen than delivered by the breathing of 100% oxygen at normobaric pressure would be used effectively if it were available at the wound site. Yablon and Cruess [20] and Niinikoski and Hunt [16] have concluded that oxygen is a fundamental and limiting factor in the healing of fractured bones.

Thus, to us, the view that oxygen administration in appropriate dose does not enhance the healing of normal wounds seemed illogical on a first-principle basis. This has, however, been the commonly held opinion, effectively dogma, in the hyperbaric medical community as indicated by clear, emphatic, and unchallenged statements published over almost four decades.

• “In short, hyperbaric oxygen seems to have little role in healing clean, normal, incised wounds. Its role in hypoxic wound healing, however, appears to be established” [21].

• “Hyperbaric oxygen will not heal normal wounds more rapidly but may, under certain circumstances, induce problem wounds to heal more like normal ones” [22].

• “In conclusion, HBO₂ can significantly improve the healing of ischaemic incisional wounds if the treatment takes place during the first few days of wound healing. In contrast, HBO₂ therapy has no effect on the healing of normal wounds” [23].

• “HBO₂ may not speed healing in normal wounds, but it proves particularly useful when an injury occurs in an area of local hypoxia or decreased blood flow” [13].

• “Current evidence suggests that popular alternative therapies such as massage, cryotherapy, and hyperbaric oxygen exposure as currently practiced on humans have little effect on recovery from minor muscle damage such as induced by exercise. While further research is still needed, hyperbaric oxygen exposure shows clear promise for potentially being a successful adjunct treatment for enhancing muscle repair and recovery from more severe crush or contusion injury in humans” [24].

Despite this prevalent view, interest in HBO₂ to treat normal wounds was stimulated when some professional sports teams began this practice in the early-to-mid 1990s. The first reported studies of HBO₂ as an adjunctive therapy for sports injuries were conducted by professional football (i.e., soccer) teams in Scotland beginning in the late 1980s. These studies, uncontrolled except for the teams’ physios’ extensive past experience with similar cases, showed that HBO₂ when added to their standard treatment regimens, significantly accelerated recovery from typical injuries without increased risk of relapse [25].¹ Following this research and an accumulation of related anecdotal experience, a number of professional sports teams in the United Kingdom and North America began the routine use of hyperbaric oxygen therapy to treat injured athletes. This in turn fostered research intended to explore, confirm, or perhaps in some cases, to refute [26, 27, 28] the efficacy of this application.

At present, controlled animal and human studies with uncompromised trauma to ligaments, tendons, muscles, and bone support the fact that hyperoxic therapy can significantly enhance the healing of normal wounds. A study of cutaneous wound healing using an animal model also attests to this fact, as do several reports on the practical application of HBO₂ to cosmetic surgery and research related to its use in uncompromised reconstructive surgery.

3.2 Enhanced and diminished healing

In the sections below, enhanced healing and diminished healing are critical aspects in discussing the outcomes of hyperoxic therapy regimens. By enhanced healing, we are referring to an outcome that is either significantly better in an absolute sense and/or has occurred significantly faster, temporally, than the outcome of some alternate therapy regimen or no therapy, whatsoever. By diminished healing, we are referring to an outcome that is either significantly inferior in an absolute sense and/or has occurred significantly slower, temporally, than the outcome of some alternate therapy regimen or no therapy, whatsoever.

While the superiority of a better absolute outcome (e.g., greater wound bursting strength, greater ligament stiffness, less scaring, greater survival of transferred autologous fat, and/or transplanted hair follicles) is obvious, the importance of faster healing of a normal wound may not be so apparent. After all, by definition, at some point in time, the outcomes of different therapy regimens for normal wounds should be identical [29]. There are, however, circumstances where faster healing can be of considerable importance.

As an example, if a surgeon suffers a severe knee injury while skiing and this prevents him from working until his injury has effectively healed, an actual situation we were approached about by an American surgeon seeking treatment with hyperbaric oxygen, a significantly faster recovery could reduce the work and income he would lose because of his accident. Likewise, in U.S. workers’ compensation cases, if the cost of hyperoxic therapy to return an injured employee to work more quickly were more than offset by the reduction in the temporary total disability payments to the employee, then the insurance company would have a net reduction in its costs. At the same time, the employee would more quickly return to earning his or her full income. Thus, if the economics were born out in practice, this would be a win-win situation for both the injured employee and the worker’s compensation insurer or equivalent in countries with different (e.g., national) healthcare management programs. Multiplied many times over, the amount that might be saved by insurance companies or national healthcare programs in such cases could be quite substantial.

An obvious situation for the benefit of faster recovery is in athletics. More rapid recovery from injury with the attendant reduction in deconditioning over the healing period translates into less lost competitions. In the case of professional sports where some athletes can make amounts as much as US$1M or even more per event, enabling that individual to compete through use of adjunctive HBO₂ when he would not have been able to do so otherwise could, in some cases, have profound financial impact and greatly influence the outcome of a season for the individual and/or his team. In many such cases, the cost of administering hyperbaric oxygen therapy would be moot. Consequently, the speed of recovery as well as the extent of recovery can be an important aspect of an outcome. Research that demonstrates the enhanced healing of normal wounds by hyperbaric oxygen therapy is summarized below.

3.3 Supporting evidence for treatment of normal wounds with HBO₂

Connective Tissue Healing: A prospective, randomized, double-blind study of the effects of HBO₂ on the healing of second-degree (Grade 2) medial collateral ligament sprains of the human knee demonstrated statistical differences in measured parameters and the way these values changed over time during the first 3 weeks of recovery [30]. In particular, range of motion, maximum flexion, reduction of edema determined by MRI, and functional running were significantly better in the HBO₂-treated subjects than in the control subjects.

In another human study of HBO₂ treatments for Grade 2 medial collateral ligament injuries of the knees of rugby players [31], visual analog scales done on the same day immediately before and after HBO₂ showed significant improvement when walking and jogging (p < 0.001 in both cases). Further, the mean time to return to competition was significantly reduced by 12.5 days in comparison with untreated (with HBO₂) control subjects (p < 0.05).

In a third human study of HBO₂ treatments for connective tissue involving Grade 2 ankle sprains, thirty-six patients were divided evenly into three groups (A, B, and C) and received seven treatments over 5 days breathing 100% oxygen in a multiplace chamber, each treatment with a duration of 90 minutes [32]. Two treatments per day were received on Days 1 and 2, and one treatment per day was received on Days 3, 4, and 5. Group A received treatments at 1.0 ATA; Group B received treatments at 2.0 ATA; and Group C received treatments at 2.5 ATA. At the end of the course of treatments, edema in the injured ankle of Group C was reduced by 97%, in Group B by 80%, and in Group A by only 30%. At the same point, Groups B and C had no pain in their injured ankle and Group A had reduced pain [32].

These three human studies have been complemented by a number of animal studies in which significant differences between HBO₂-treated and control animals in the healing of ligaments and tendons, assessed morphologically, histologically, biochemically, and mechanically were consistent with more rapid healing in the HBO₂-treated human groups.

• HBO₂ significantly enhanced recovery in the stiffness of experimentally lacerated medial collateral ligaments in rat knees and produced a strong tendency for ligament strength to recover more quickly, as well [33, 34].

• HBO₂ resulted in the more rapid increase of pro-a1 (I) mRNA assessed biochemically and the more rapid formation of collagen fibers assessed histologically following experimental laceration of the patellar tendons of rat knees [35].

• The healing of experimentally lacerated ligaments in the hind limbs of rats with HBO₂ treatments at 2.0 ATA for 60 minutes, 2.0 ATA for 30 minutes, and 1.5 ATA for 30 minutes were assessed morphologically, histologically, and biochemically and found to be enhanced when compared to controls breathing air at 1.0 ATA. Of the three HBO₂ conditions, 2.0 ATA for 60 minutes produced the best outcome [36].

• HBO₂ accelerated the healing of knee ligaments of rats from which a 2-mm segment was surgically removed [37]. Comparisons with controls included the formation of scar tissue assessed by macroscopic inspection, Type I procollagen gene expression assessed histochemically, and tensile strength and stiffness assessed mechanically.

• HBO₂ enhanced the outcome of anterior cruciate ligament reconstruction in a rabbit model [38]. Blindly assessed parameters including histological (i.e., fibrocartilage formation, new bone formation, tendon graft bonding to adjacent tissue, and the number of blood vessels in the area of tendon-bone junction); biomechanical (i.e., tensile strength of tendon and failure modes including graft pullout from the bone tunnel, graft rupture, and avulsion fracture of the tendon-bone junction); electron microscopic (i.e., nature and organization of regenerated collagen fibers).

Soft Tissue Healing: A study using a well-defined, blunt trauma model of skeletal muscle injury (i.e., contusions) in rats showed that HBO₂ inhibits the muscle wasting normally associated with such injuries [26]. A complementary study of muscle contusions in rats showed that HBO₂ accelerates the recovery of muscle function (i.e., twitch and tetanic force) in response to electrical stimulation [39].

Another study using calf muscle contusion in rats as a skeletal muscle injury model [40] found that among other effects, HBO₂ in the acute phase suppressed elevation of circulating macrophages and reduced both lower limb volume and the wet weight of the contused muscles. It also increased the amount of regenerated muscle fibers and, at 7 days, promoted isometric muscle strength. Thus, HBO₂ reduced early inflammation and accelerated myogenesis in muscle contusion injuries [40].

A fourth study investigating the effects of HBO₂ on reproducible muscle tears in rats found significantly increased hydroxyproline levels in the urine of treated rats compared to untreated control rats at 1- and 2-weeks post-injury [41]. It was concluded that HBO₂ enhances collagen metabolism which may be helpful for muscle healing.

A fifth study using a rabbit model to investigate the effects of HBO₂ on the healing of muscle stretch injuries showed highly significant functional (i.e., ankle isometric torque) and marked morphological differences (i.e., inflammation and damage around the muscle-tendon junction) between the HBO₂-treated and untreated control groups, with the HBO₂-treated group recovering more quickly [42].

Three other studies investigated the effects of HBO₂ on the functional and morphological properties of rat muscles regenerating from myotoxic injury. Two found significantly increased force-producing capacity and larger regenerating muscle fibers in the extensor digitorum longus [43] and the soleus [44], respectively, in HBO₂-treated animals compared to untreated control animals. In the soleus muscles, the difference between test and control muscles was significant at 14 days but not at 25 days [44].

The third study investigated recovery with HBO₂ from injury to the tibialis anterior muscles of rats induced by injecting cardiotoxin (CTX) into them [45]. The HBO₂ treatments consisted of 120 minutes of oxygen at 2.5 ATA once a day, 5 days per week for 2 weeks. In addition to untreated controls, other post-injury management utilized air at 2.5 ATA and oxygen at 1.0 ATA with the other treatment parameters similar to those for the HBO₂-treated subjects. Assessments comparing the HBO₂-treated rats to untreated controls, normobaric oxygen-treated animals, and hyperbaric-air-treated animals included histological analyses and measurement of the maximum force-producing capacity of the regenerating muscle fibers. The results showed that HBO₂ treatments accelerated satellite cell proliferation and myofiber maturation in rat muscle that was injured by a CTX injection, particularly at the end of the 2-week course of treatments. This suggested to Horie and associates [45] that HBO₂ treatments accelerate the healing and functional recovery of skeletal muscle after injury.

Finally, two studies evaluated the efficacy of HBO₂ for treating human muscle injuries. The first involved 20 patients who had sustained sports-related muscle injuries and were admitted to the hospital of Tokyo Medical and Dental School in Tokyo, Japan for therapy [46]. There, they received daily sessions of HBO₂ for 60 minutes at 2.8 ATA for 1 to 7 days. These treatments significantly improved before- and after-course-of-therapy results of visual analog scales (VAS) for resting pain and motion pain, the patient’s subjective evaluation of edema, muscle stiffness measured with a muscle tonometer, and leg volume measured with a water-filled volumetric gauge for patients with gastrocnemius muscle injury [46].

The second was a blinded study of recovery from exercise-related (i.e., baseball) first-degree strains to skeletal muscles of the extremities [47]. This was conducted with HBO₂-treated human subjects compared to sham-treated control subjects. The HBO₂ treated subjects received 10 sessions of HBO₂ administered at 2.5 ATA twice per week for 5 weeks. Their hyperbaric oxygen treatments were given for 70 minutes at maximum pressure with two 5-min air breaks included in those periods. The control subjects breathed air at 1.3 ATA for 70 minutes with the same frequency. Significant reduction in levels of serum creatine phosphokinase, glutamic oxaloacetate transaminase, myoglobin, and other muscle enzymes were found in the HBO₂-treated subjects at the end of the course of therapy and persisted for the next 2 weeks over which they were monitored. No such changes were found in the control subjects. At the end of therapy, the Brief Pain Inventories for the HBO₂-treated subjects demonstrated significant reductions in the pain intensity subscale not found in the control subjects. It was concluded that the HBO₂ treatments accelerated recovery of humans from routine, sports-related muscle injuries.

The enhanced healing of normal wounds has also been reported outside the sphere of sports-type musculoskeletal injuries. The healing of a standardized wound to the ears of homozygous hairless mice was used to investigate the effect of HBO₂ on the reepithelialization and microvascular perfusion in normal and ischemic skin tissue [48]. It was found that HBO₂ at the dose employed significantly accelerated healing in normal tissue as well as ischemic tissue, though the relative improvement in healing was greater for ischemic wounds. As final examples of enhanced healing in normal soft-tissue and connective-tissue wounds, HBO₂ has been effectively utilized to improve outcomes in uncompromised reconstructive surgery [49, 50] and in laser face resurfacing by cosmetic surgeons [51, 52, 53].

Fracture/Bone Healing: With respect to bone repair/fracture healing, a number of studies have shown that hyperbaric oxygen enhances the various processes involved. In addition, anecdotal results of the benefits of HBO₂ for fracture healing in professional athletes have been noted [54], though literature surveys to support such application of HBO₂ have not produced formally reported findings to the standards of evidence-based medicine [54, 55, 56]. Despite this, a number of research publications have reported statistically significant enhanced bone repair in comparison with controls for fracture/bone damage properly classified as normal.

• Coulson et al. [57] produced closed fractures in one femur of rats and found that daily HBO₂ increased Ca⁺⁺ deposition in the fracture callus and increased the breaking strength of the healing bone relative to that of the other, uninjured femur.

• Yablon and Cruess [20] studied the healing of experimental fractures of the femurs of rats. While no difference in healing between the HBO₂-treated and control animals was shown by standard X-rays after 9 days, significant differences in the time course of events were evident after 22 and 40 days with the HBO₂-treated animals healing more quickly.

• Natiella et al. [58] found that the healing of experimentally produced cryogenic injuries to the femurs of dogs was accelerated by HBO₂ though the 2 months this therapy was administered but did not produce a different long-term outcome.

• Wilcox and Kolodny [59] conducted a study on the effects of HBO₂ on the healing of mandibular and maxillary osteotomies in humans. Controls were retrospectively paired with untreated cases, and outcomes were assessed by two experienced surgeons. It was found that HBO₂ hastened the decrease in fracture mobility, thus reducing the need for post-immobilization retainers and training elastics. Wilcox and Kolodny [59] concluded that “hyperbaric oxygen augments clinical bone healing subsequent to osteotomy procedures.”

• Karapetian et al. [60] reported that the healing of non-inflamed mandibular fractures in rats was enhanced by HBO₂ at 2.0 ATA.

• Barth et al. [61] found that holes drilled in the proximal and distal metaphyses of rat femurs filled in 4 weeks through enchondral ossification (i.e., differentiation to cartilage than bone) without HBO₂ and in 3 weeks through primary ossification (i.e., direct differentiation to bone) with a course of 20 HBO₂ treatments administered once a day 5 days per week.

• Johnsson et al. [50] found that a 21-day course of HBO₂, once a day at 2.8 ATA for 90 minutes, significantly increased the interface strength of a bone-titanium implant, both following irradiation and with no irradiation. The benefit was relatively larger in the case of irradiation, however.

• Ehler et al. [29] studied the effects of HBO₂ on autogenous cancellous bone graft healing in the canine femur. The breaking strength of the bones was assessed over time, and it was found that HBO₂ improved the outcomes early but produced no different result over the long term.

• Kitakoji et al. [62] studied the effects of HBO₂ on a lengthened callus model in rabbits. It was found that bone mineral density was significantly greater in the HBO₂-treated animals than in the control animals after 1, 2, 3, and 4 weeks, with these differences tending to normalize after 5 and 6 weeks.

• Kawada et al. [63] studied the effect of HBO₂ on the healing of an open fracture model in mice. At 2 weeks post-fracture, the callus was significantly larger in the HBO₂-treated group than the control group. At four- and six-weeks post-fracture, radiographic findings showed accelerated healing in the HBO₂-treated group. At 6 weeks, bone stiffness and maximum load were significantly higher in the HBO₂-treated group. These results suggested that HBO₂ enhances bone anabolism and accelerates fracture healing.

In view of such factors as the recognized essential role of oxygen in all aspects of bone healing [16, 20], the substantial evidence for oxygen’s enhancement of the healing of uncompromised trauma such as that cited in the approximately 30 studies above, and the considerable positive practical experience in the application of HBO₂ to actual sports injuries and cosmetic surgery cases² (e.g., [47, 51, 52, 53, 64, 65, 66, 67, 68, 69, 70]), understanding the basis for what we believed to be the hyperbaric medical community’s mistaken dogma that hyperbaric oxygen therapy has no efficacy for normal wounds was of considerable interest to the authors. Consequently, an investigation was begun with several references suggested by a professional acquaintance, Dr. Eric Kindwall³ who was a recognized expert in the field of hyperbaric medicine and a published proponent of the prevailing view on the treatment of normal wounds Kindwall et al. [22]. These citations were provided by Dr. Kindwall when he was asked about the foundation of his opinion on this specific issue. The analysis of these references led the authors to review other reports and the eventual formation and validation of a hypothesis concerning the general nature of oxygen dose-response described below.

3.4 Case against effectively treating normal wounds with HBO₂

In addition to the three references suggested by Dr. Kindwall (i.e., [1, 71, 72]), two similar references (i.e., [2, 73]) were discussed by Hunt et al. [21] in a chapter on the effects of oxygen in wound healing in an early text on hyperbaric oxygen therapy [74]. Because of the findings of the latter two research reports, particularly those of Penttinen et al. [73], Hunt and his associates concluded that hyperbaric oxygen therapy has no role in the healing of normal wounds [21]. This would seem to be the origin of the dogma in question that has persisted from 1977 to the present time (i.e., about 47 years as of early 2024).

As noted, the references listed immediately above would appear to have formed the foundation of the view that hyperbaric oxygen therapy is not of benefit for the treatment of normal wounds. Through critical analysis of these reports and their comparison to similar research conducted with different partial pressures of oxygen, we determined that oxygen has a biphasic dose-response relationship which impacts some outcomes of oxygen therapy differently than envisioned by its practitioners. So that readers may more easily understand the conclusion we reached in this endeavor, the references below are presented in greater detail than those discussed in the prior section of this report.

Lundgren and Sandberg [1]: A study to determine the effects of HBO₂ on the tensile strength of the healing skin of rats was conducted by Lundgren and Sandberg [1]. Incisions long enough to facilitate measuring the force required to rupture the healing wound three times were made on one side of a rat’s back. The animal was then allowed to heal for 5 days, either with or without receiving HBO₂. After measuring the rupture force three times on that side, a similar procedure was performed on the other side of each rat’s back with healing taking place in the manner opposite to that of the first case (i.e., with HBO₂ if HBO₂ were not administered initially; without HBO₂ if HBO₂ were administered initially). This permitted each rat to be its own control in the study. Five groups (i.e., A–E) of 7–8 animals, each, were included in the study.

• Group A received oxygen at 2.0 ATA for three, 120-minute sessions each day for 5 days. The tensile strength of the wound was 43% less with HBO₂ than without it, but three animals died from what was likely oxygen toxicity. Thus, this difference was not statistically significant due to the small group size of the treated animals, though in view of the results of Group B, the difference would certainly have been significant had there been more HBO₂-treated subjects alive at the end of the oxygen treatments.

• Group B received oxygen at 2.0 ATA for two, 120-minute sessions each day for 5 days. The tensile strength of the wound was 27% less with HBO₂ than without it. This difference was statistically significant (p < 0.01).

• Group C received oxygen at 1.5 ATA for two, 120-minute sessions each day for 5 days. The tensile strength of the wound was 6% less with HBO₂ than without it. This difference was not statistically significant.

• Group D received oxygen at 2.0 ATA for two, 120-minute sessions each day from Day 2 through Day 5 (i.e., a total of 4 days). The tensile strength of the wound was 2% more with HBO₂ than without it. This result was not statistically significant.

• Group E had the primary artery to the wound area ligated, reducing local blood flow, and received oxygen at 2.0 ATA for two, 120-minute sessions each day for 5 days as with Group B. The tensile strength of the wound was 8% less with HBO₂ than without it. This result was not statistically significant.

Clearly, these results indicate that HBO₂ in the higher doses administered in this study impedes healing in comparison with no HBO₂. This would superficially support the contention that HBO₂ is not efficacious for normal wounds.

It is also clear, however, that as the dose of inspired oxygen was progressively reduced, the experimental outcome (i.e., wound tensile strength) progressively improved. Further, a comparison of Groups B and E, which had the same HBO₂ treatment schedule, shows that a lower dose of oxygen local to the wound site, caused by ligation of the primary artery, improved the outcome.

In discussing their results, Lundgren and Sandberg [1] observed that the outcomes appeared dose-dependent (Figure 3). In view of the outcome for Group E and since oxygen delivery to the wound sites in Groups A, B, C, and D would be directly related to the oxygen doses of the hyperbaric oxygen treatments, it appears that the outcomes of this study were inversely related to the oxygen dose at the wound site. Thus, the lower the dose of oxygen used of those included in this study local to the wound, the better the outcome.

Figure 3.

Data for Lundgren and Sandberg’s [1] five groups showing a decline in outcomes as the dose of oxygen was increased. What we shall call the relative index of oxygen dose was calculated as the PO₂ of treatment (Tx) * duration of Tx * frequency of Tx * days of Tx for Groups A, B, C, and D. Group E, with the ligated primary artery to the wound region impacting negatively on healing, it was arbitrarily assumed that the dose index was reduced by 1/2 after calculation as described for the other groups, above.

As a final point, some of the doses of oxygen used in the Lundgren and Sandberg study [1], particularly for Group A, were greater than is typical for clinical cases involving the treatment of tissues with impaired oxygen delivery. The fact that three of the seven HBO₂-treated animals in this group died after 3 days of treatment from what could readily be construed to be oxygen poisoning strongly suggests that this dose of oxygen was excessively high [1].

Kulonen and Niinikoski [71]: Research was conducted by Kulonen et al. [2] on the impact of hyperoxic therapy at normobaric pressure on normal wound healing. They found significantly positive impacts of normobaric hyperoxygenation in direct relation to the PO₂ (i.e., 35 and 70% oxygen in comparison with 21% oxygen Kulonen et al. [2]). Kulonen and Niinikoski [71] then wished to extend these findings to hyperbaric oxygen conditions. In this latter study, the tensile strengths of healing skin wounds and granulomas were determined for groups of HBO₂-treated rats and control animals that did not receive HBO₂. Hyperbaric oxygen was administered at 2.0 ATA for two, 120-minute sessions each day with a 120-minute interval between them. These treatments started in the morning immediately after wounding and extended for 3 days or 9 days. Assessments were made the day after treatment ended (i.e., Days 4 and 10 for the two groups, respectively). The tensile strength of the healing skin wounds and the granulomas were significantly less in the HBO₂-treated rats than in the control animals (p < 0.0005) at both 4 and 10 days. All animals gained weight normally over the course of the study and there were no signs or symptoms of oxygen poisoning (i.e., convulsions) in the treated groups. Considering their positive outcomes with hyperoxic breathing gas at 1 atm, Kulonen and Niinikoski [71] concluded that the optimal concentration of oxygen for healing had been exceeded in this hyperbaric oxygen study.

Wray and Rodgers [72]: A study to determine the effects of HBO₂ on the breaking load of tibial fractures in rats was conducted by Wray and Rodgers [72]. Fractures were produced with a rotary saw, and HBO₂ at 2.0 ATA was administered 6 hours a day for from 20 to 26 days starting 1 day after a fracture. A control group received no HBO₂.

Results showed that despite increased callus activity on histological examination, the breaking loads for healing fractures with HBO₂ were lower than those in the untreated control group. This result was not statistically significant but was strongly suggestive (0.05 < p < 0.07). The authors concluded that their hyperbaric oxygen regimen produced local tissue oxygen levels that were too high and resulted in adverse (i.e., toxic) effects on the cells, perhaps altering the lysosomal membrane [72].

Other notable points included that the HBO₂-treated rats gained less weight than the control rats over the study period, though this difference was not statistically significant; and nine HBO₂-treated rats died from “pulmonary complications” when one of the 6-hour treatments was inadvertently allowed to overrun by about 14 hours.

Niinikoski [3]: In this report, Niinikoski comprehensively covers material addressed in a number of prior publications [2, 71, 75, 76, 77]. The latter of these Kulonen and Niinikoski (i.e., [71]) was one of the reports cited by Kindwall and has been discussed above. As the pertinent information from Niinikoski [3] concerning the impact of HBO₂ on normal wounds has, therefore, already been presented in detail, it will not be repeated here.

Penttinen et al. [73]: The effects of HBO₂ on fracture healing in rats were assessed. Both tibias of the subject rats were broken in the middle of the diaphysis by manually bending them to an angle of 90^o under anesthesia. HBO₂ was begun immediately and administered at 2.5 ATA for 120 minutes twice daily until sacrifice [73].

In comparison with untreated control animals 11–21 days post-fracture, the healing fractures of the HBO₂-treated animals had significantly greater collagen formation and accumulation of minerals. The tensile strength of the calluses of the HBO₂-treated fractures, however, were not different from those of untreated control fractures, nor were the RNA/DNA ratios of callus cells as reported in a related publication different than those of untreated control fractures [78]. These results suggested to Hunt and associates [21] that hyperbaric oxygen by itself does not accelerate bone healing in normal wounds.

3.5 Further consideration of oxygen dose response

In contrast to the results of Lundgren and Sandberg [1] and Kulonen and Niinikoski [71] in which the bursting force for healing skin wounds was generally less for HBO₂-treated rats than control animals, Meltzer and Myers [79] reported greater bursting strength for healing incisional skin wounds in HBO₂-treated rats than in untreated controls at 3 days and 7 days post-wounding. In Meltzer and Myers’ study, however, HBO₂ was administered at 2.5 ATA for 120 minutes only three times, once a day at 0-, 24-, and 48-hours following creation of the experimental incisions.

In comparison, Lundgren and Sandberg [1], at the extreme (i.e., Group A), administered HBO₂ at 2.0 ATA for 120 minutes three times daily for 5 days. In the next most extreme case of oxygen exposure (i.e., Lundgren and Sandberg Group B), HBO₂ was administered at 2.0 ATA for 120 minutes twice daily for 5 days. As noted above, the wound bursting strength of this latter group was significantly lower than for wounds healing in the untreated (i.e., without HBO₂) control animals and the wound bursting strength for the former (i.e., A) group was only insignificantly lower than the untreated controls because of the subject fatalities almost certainly due to oxygen toxicity.

Likewise, Kulonen and Niinikoski [71] administered greater doses of oxygen than Meltzer and Meyers [79]. Their exposures were comparable to Lundgren and Sandberg’s Group B, 2.0 ATA for 120 minutes twice daily for 3 or 9 days. As with Lundgren and Sandberg’s Group B outcome, the bursting strength of the healing wounds in Kulonen and Niinikoski’s study was significantly lower for the HBO₂-treated rats than for the untreated controls.

Parallel results have been reported for HBO₂ administered for bone repair. Wray and Rodgers [72] found decreased breaking force (not quite statistically significant) in HBO₂-treated animals in comparison to their control animals for the healing fractures. Penttinen et al. [78] found unchanged tensile strength in the healing fractures of the femurs of rats for HBO₂-treated animals in comparison to untreated controls. Wray and Rodgers sectioned one tibia of each subject rat with a rotary saw and administered HBO₂ at 2.0 ATA for 6 hours a day for from 20 to 26 days to the test group. The rebreaking loads for the healing fractures of the test and control groups were then compared to the breaking load for the animals’ uninjured tibias. The HBO₂-treated animals were found to have lower relative rebreaking loads than the untreated controls. Penttinen et al. [78] administered HBO₂ to their subject rats at 2.5 ATA for 120 minutes twice daily from the fracture creation until sacrifice and found no difference in the rebreaking load compared to the untreated controls.

In contrast, Coulson et al. [57] administered HBO₂ at 3.0 ATA for 120 minutes to rats once daily from the time a 3/32-inch defect was created in the rats’ left femurs with a dental drill until sacrifice at 2 weeks. With this reduced therapy regimen, the strain necessary to break the healing femur was found to be 93% of that of the uninjured right femur, on average, for the HBO₂-treated rats, and only 82% of the uninjured right femur, on average, for the untreated control rats.

A study supplementing the bone repair findings summarized above was conducted by Barth et al. [61]. As noted previously, they found that HBO₂ at 2.0 ATA for 120 minutes once a day for 20 days significantly accelerated bone repair in rats and resulted in notably greater blood vessel in-growth into the new bone than that occurring in untreated control animals. In contrast, HBO₂ administered at 2.0 ATA for 120 minutes twice a day for 10 days significantly retarded bone repair and blood vessel in-growth into the holes compared to untreated controls. In the case of the rats treated twice a day with HBO₂, the holes were filled by enchondral ossification in 5 weeks, a full week (i.e., 25%) longer than the untreated control animals and two full weeks (i.e., 67%) longer than the animals treated only once per day with HBO₂. Further, at 5 weeks, there were still islets of cartilage in the forming plugs of the animals treated twice per day.

While the studies summarized immediately above do show that HBO₂, under certain circumstances, can produce healing of normal wounds not better than or even inferior to that without adjunctive HBO₂, there is a consistent pattern through all of them. When healing in either soft tissue, connective tissue, or bone was adversely affected by hyperoxic therapy, the doses of oxygen were high, either in absolute terms or relative to other experimental protocols which produced the opposite (i.e., beneficial) outcomes. In other words, higher doses of oxygen hindered healing for uncompromised wounds! In support of the Barth and associates’ findings, Karapetian et al. [60] reported that while the healing of non-inflamed mandibular fractures in rats was enhanced by HBO₂ at 2.0 ATA, it was hindered by HBO₂ administered at 2.5 ATA.

This insight suggested to the authors that the oxygen dose-response relationship has a descending limb at high doses. Thus, we developed the biphasic dose-response relationship shown in Figure 4 which at the high dose end eventually goes below the level of response to normal conditions (i.e., no supplemental oxygen administration, whatsoever).

Figure 4.

Oxygen dose-response relationship envisioned by the authors circa 2000 to illustrate the understanding they had evolved through extensive research and analysis of the scientific and medical literature on hyperoxic treatments of a variety of wounds to soft tissues, connective tissues, and bones [80, 81].

4. Is hyperoxic therapy dose dependent?

4.1 Common dose-response relationship

When dealing with therapeutic drugs, the most common dose-response relationship considered seems likely to be the one that is for a substance that has only beneficial effects or that is used at dose levels where only beneficial effects occur. After reaching a threshold for effects to begin to take place, the response increases until a maximum level is reached. Then, the response goes no higher. This type of pharmaceutical agent has the typical sigmoidal response when the dose is plotted on a logarithmic scale (Figure 5).

Figure 5.

A simulated dose-response curve for the stimulation of a tissue by an agonist with a log₁₀ (EC50) of 0.70. This curve was generated in GraphPad Prism with a randomized relative error of 5% and with a fitted curve following the Hill Equation. This shows a typical, sigmoidal dose-response relationship with low doses insufficient to generate a response while higher doses ultimately reach a maximum response and go no higher (JM Goodman 17 July 2019. Licensed by Wikipedia under the Creative Commons Attribution-Share Alike 4.0 international (https://crelativecommons.org/licenses/by-sa/4.0/deed.En) license).

In view of absolutely no mention of a different type of dose-response relationship, this must certainly be the one most healthcare professionals and researchers are familiar with and/or think of when considering higher doses of oxygen. Clearly, without statements to the contrary, when considering hyperbaric oxygen doses in the treatment of normal (i.e., uncompromised) wounds, Hunt et al. [21] could not have imagined there to be a biphasic oxygen dose-response relationship with a down-side at higher doses (Figure 4). Nor could Margolis et al. [13]; Kindwall et al. [22]; Quirinia and Viidik [23]; nor Tiidus [24] have considered such a dose-response relationship when they published their statements about HBO₂ and normal wounds. It seems that they all had lost sight of the fact that unlike most pharmaceuticals where amount (equivalent to the partial pressure of a gas (oxygen) and frequency of administration) is the critical aspects of dose, hyperoxic treatments are also dependent upon the duration of exposure. Thus, any given P_iO₂, and most especially higher values within the 2.0 to 3.0 atm envelope can span values across the entire dose-response range with longer and longer treatments Clark et al. [82]. In addition to Lundgren and Sandberg’s [1] observation that their results of utilizing hyperbaric oxygen therapy to treat cutaneous wounds in rats appeared to demonstrate dose dependence, though it was not stated that this relationship was the inverse of how dose-response is routinely thought of (Figure 3). Barr and Perrins [4] made a similar general observation for human patients:

“The response to HBO₂ is clearly dose-dependent. Some ulcers (i.e., non-healing wounds) respond to 1.5 ATA while others require up to 3.0 ATA. Some fail to respond to 2 hours a day, while others heal with less than 1 hour a day. This anomaly probably results from differing vascularity in the affected limbs and how late HBO is applied in the destructive phase.”

In view of the conflicting research outcomes that have led to the belief that hyperbaric oxygen therapy is not effective for normal wounds, if hyperoxic therapy is dose-dependent, then this relationship must be along the lines of that shown in Figure 4 which was constructed by the authors in the early 2000s [80, 81]. This relationship has a progressively increasing benefit over the lower range of doses, then at higher doses, has a progressively decreasing or negative impact, ultimately falling below the no-hyperoxic-therapy-at-all response, as per the Lundgren and Sandberg [1] findings (Figure 3).

In considering this hypothesis from a superficial standpoint, it seemed to fit our practical experience in treating sports injuries and cosmetic surgery cases. At the outset, the chamber most frequently used for these treatments was the HYOX HTU (Figure 6). This chamber has an upper pressure limit of 2.0 ATA, and the treatments conducted in it were usually only an hour in duration once a day, though some treatments for fractures were 90-minutes long, and occasionally, serious injuries occurring during sports competition were treated twice a day, but only for a day or two. Thus, the doses used to treat uncompromised sports injuries and cosmetic surgery wounds were, in our experience, typically lower than those administered in routine clinical HBO₂.

Figure 6.

HYOX HTU is the 2.0 ATA monoplace chamber used in many treatments of athletes and cosmetic surgery cases. The chamber is compressed with air and oxygen is breathed by mask or hood. The gas exhaled by the patient is dumped overboard.

With this view in mind, we set out to determine if there were sufficient data in the scientific and clinical literature to conclusively establish that the oxygen dose-response relationship is of the general nature as shown in Figure 4. Fortuitously, the literature from the 1960s into the 2000s is rife with research relevant to this inquiry, including the data already presented, which concerns a variety of tissues, wound types, and methods of assessment.

As a critical test, we determined if the hyperbaric oxygen doses administered in the treatment of normal wounds which produced negative outcomes were significantly greater in comparison with those hyperbaric oxygen therapy doses which produced positive outcomes with normal wounds. To do this, we utilized both the human and animal data described previously in this report computing a parameter we have called the mean relative index of oxygen dose per day (i.e., Tx PO₂ * Tx duration * #Tx/day) as the values to compare.⁴ These data include a variety of soft tissue wounds (e.g., muscle and skin), connective tissue wounds (e.g., ligaments and tendons), and bone breaks/injuries. In addition, the research consisted of many different approaches to assessing healing including morphological, histological, mechanical, biochemical, radiological, and functional. Of the data related to positive outcomes for the treatment of normal wounds with hyperbaric oxygen therapy, thirty-six reports with forty-one different outcomes have been used in the analysis presented here [20, 26, 29, 30, 31, 32, 33, 34, 37, 38, 39, 40, 41, 42, 43, 44, 46, 48, 49, 50, 51, 52, 53, 57, 58, 59, 60, 61, 62, 63, 71, 79, 83, 84, 85, 86]. Regarding the negative outcomes, 10 reports with thirteen different outcomes have been utilized [1, 60, 61, 71, 72, 73, 78, 85, 87, 88, 89]. The statistical parameters calculated and included in this comparison are given in Table 1. The oxygen doses producing negative outcomes when treating normal wounds were very significantly higher than those oxygen doses which produced positive outcomes (i.e., P < 0.00005). Thus, we conclude that the outcomes of oxygen treatments are definitely dose-dependent with significant relative overdose producing inferior outcomes to those achieved with lower doses.

Parameter	Negative outcomes	Positive outcomes
Number of samples	13	41
Mean dose/day	805.4	222.0
Standard deviation	233.5	107.2
One-tailed t-value	8.722
Degrees of freedom	52

Table 1.

Statistical comparison of mean relative index of oxygen dose per day for hyperbaric oxygen treatments that produced negative outcomes with normal wounds to hyperbaric oxygen treatments that produced positive outcomes with normal wounds [90]. The number of degrees of freedom was computed with this same data [90]. A table of one-tailed t-test values was used to determine the statistical significance of the amount by which the mean index of oxygen dose per day for the negative outcomes exceeded the mean index of oxygen dose per day for the positive outcomes (Table of Critical Values) [91].

4.2 Oxygen dose-response relationship

In 2011, while researching general references related to dose-response on the Internet, the authors encountered the term, “hormesis” [92]. In toxicology, hormesis is a dose-response phenomenon typically characterized by low-dose stimulation and high-dose inhibition [93, 94, 95]. This relationship produces a biphasic dose-response curve [94] as shown in Figure 7. This matched almost exactly the dose-response relationship the authors had independently derived for oxygen therapy, Figure 4, which underlays the hormetic curve below. Upon further investigation, we found that oxygen was already on lists of hormetic agents [97], though not for the same reason as our independent development of an hormetic dose-response relationship.

Figure 7.

Example of hormetic dose-response relationship from Wikipedia article circa April 2021 “Hormesis”, which is released under the Creative Commons Attribution-Share-Alike License 3.0 [96]. Our independent derivation of the dose-response relationship for oxygen (i.e., Figure 4) is shown in the background [80].

An important aspect of hormesis is that agents which exhibit it are likely to produce preconditioning to stressful situations with exposures in the lower-dose range. For this reason, hormesis is also referred to by a number of different terms such as preconditioning, conditioning, pretreatment, cross-tolerance, adaptive homeostasis, and rapid stress hardening (mostly low temperature rapid cold hardening) [94].

Oxygen, however, is outside typical patterns for hormesis as we are routinely exposed to a finite level (i.e., air that has a P_iO₂ just lower than 0.21 atm, 0.20954 atm). Thus, regarding preconditioning, oxygen has the potential to have hormetic relationships in both the higher dose (i.e., hyperoxic) and lower dose (i.e., hypoxic) directions, and this has indeed been found to be the case.

4.3 Hypoxic preconditioning

For over 45 years, low oxygen has been one of the leading stressors known to confer protective preconditioning/hormetic effects [98]. These effects frequently relate to animals, both invertebrates and vertebrates, withstanding prolonged periods of anoxia or hypoxia. Such benefits of lower than normal oxygen are based on a hypothesis of “preparation for oxidative stress” (i.e., POS) [99], where, upon experiencing a decrease in oxygen levels, mitochondria prepare for the ensuing oxygen reperfusion by elevating cellular defenses, particularly oxidative stress defenses [100, 101, 102]. The need for antioxidant defenses with reperfusion of an ischemic wound results from the buildup of substances in the wound during the hypoxic or anoxic phase which, when the wound is reperfused with oxygen laden blood leads to the formation of reactive oxygen species with the potential to cause serious secondary injury [103].

An example of preconditioning at lower levels of oxygen than in room air entails a protective effect on the central nervous system. This was demonstrated through vascular remodeling which took place via two separate pathways in mice after 7 days of continuous hypoxic preconditioning with an oxygen concentration between 21 and 8% [104]. When breathing 13% oxygen, endothelial hyperplasia was triggered and expanded the capillary network. When breathing 12% oxygen, endothelial hypertrophy was triggered leading to expansion of large vessels and arteriogenesis. Such hypoxic conditioning had biphasic dose-response relationships with the maximum response when the mice were breathing about 10% oxygen [104].

4.4 Hyperoxic preconditioning

Exposures to hyperoxic conditions have also been demonstrated to protect against reduced delivery of oxygen and glucose to various tissues.

• Huang et al. [105] investigated the effects of HBO₂ PC (i.e., HBO₂ preconditioning) against reduced delivery of oxygen and glucose (OGD: oxygen glucose deprivation) as stress to the spinal cord. Heat shock proteins (HSP) were found in in vitro spinal neuron culture after only one exposure to HBO_2. When OGD or H₂O₂ stresses with and without HSP inhibitors were compared, the peak effect was found to be at 12 hours post-preconditioning, but active immediately afterwards. It was determined that HBO₂-PC protects by upregulating HSP 32 expression.

• Duan et al. [106] investigated hyperoxic preconditioning for stroke. Several regimens of hyperbaric oxygen preconditioning including once every other day for five sessions (A), once every other day for three sessions (B), once every day for five sessions (C), and once every day for three sessions (D) caused infarct size to be significantly reduced following middle cerebral artery occlusion. The greatest difference was for the A group. Preconditioning patterns of once every day for two sessions (E), once every 12 hours for four sessions (F), once every 12 hours for two sessions (G), and a single session (H) produced results statistically similar to untreated controls. The benefits of the hyperoxic preconditioning appeared to be related to stimulation of angiogenesis.

• Tähepõld et al. [107] reported on heart protection against simulated stroke in rats and mice from prior preconditioning with hyperoxic gas (i.e., up to 95% oxygen) at normal barometric pressure (i.e., 1.0 atmospheres absolute). They found that the PPO₂ and duration of the preconditioning impacted on effectiveness and there were also species variations.

• Touleimat et al. [108] found hyperbaric oxygen therapy administered at 2.4 ATA for 90 minutes 5 days in a week produced a small but significant change in lung diffusing capacity. When Dr. Conoscenti was queried about this following the report, he noted that after the subsequent week of hyperbaric oxygen therapy for the group included in the study, the lung function changes that were found after 1 week had normalized. Thus, the initial hyperbaric oxygen exposures apparently produced accommodation to the reported changes [109]⁵

Other supporting facts for the hormetic nature of the oxygen dose-response can be found among conditions that have been treated with clinical HBO₂. These mirror the results of research and practical experience with normal wounds. With very high doses of oxygen, HBO₂ has been found to be of no benefit or even counterproductive and sometimes lethal with animal subjects due to oxygen poisoning. With lower doses, positive outcomes and/or practical benefits have been achieved. Some examples of this come from the literature on the survival of ischemic skin flaps:

• Negative outcomes: Kernahan et al. [87]; Caffee and Gallagher [88].

• Positive outcomes: Champion et al. [85]; Tan et al. [83]; Perrins [110]; Jurell and Kaijser [111].

All in all, since beginning to evolve our model of oxygen dose-response in 1996, some 26-plus years ago, the authors have yet to encounter research or practical application that did not fit the concepts of the model we have established and described here.

4.5 Oxygen toxicity

That oxygen is a biological toxin is well established [82, 112, 113, 114, 115, 116, 117, 118, 119]. In sufficiently high doses, it is believed that oxygen is detrimental to all living cells [120]. Such effects result from the formation of reactive oxygen species (ROS) which damage many molecules in cells including DNA, RNA, and amino acids in proteins [118, 119, 121]. ROS also inactivates enzymes by oxidizing co-factors [122]. Polyunsaturated fats in membrane phospholipids are particularly sensitive to oxidation by ROS causing increased membrane rigidity, decreased activity of membrane-bound enzymes, altered activity of membrane receptors, and altered membrane permeability [119]. These effects are on the plasma and mitochondrial membranes of cells as well as their endomembrane system [122]. In view of such broad molecular and physiological impact of relatively high doses of oxygen, it is a logical conclusion that such actions should ultimately produce a net negative impact on the healing of damaged/wounded tissue which is overexposed to oxygen. Several studies on the impact of reactive oxygen species on tissue healing confirm this. A study by Ŝenel et al. [123] showed that ROS can adversely affect healing in ischemic wounds. A study by Stewart, Moore, Bennett, et al. [124] showed that ROS can adversely affect the healing of wounds treated with hyperbaric oxygenation.

• Ŝenel et al. [123] conducted studies to determine the effect of oxygen radicals on wound healing in ischemic tissue. For this, they assessed the bursting force and, as an indicator of collagen metabolism, hydroxyproline content of the healing cross-member of an “H” incision on the back of rats as previously described by Quirina and Viidik [23]. These assessments were done at 7- and 14-days post-surgery in each of four groups: Untreated control group; group treated with superoxide dismutase (SOD), an oxygen radical scavenger; group treated with allopurinol, a xanthine oxidase inhibitor; control group with a normal, non-ischemic incisional skin wound (i.e., not an “H” incision) and no post-surgical treatments Ŝenel et al. [123]. In addition, a histopathological evaluation was conducted to assess the amount and organization of collagen in the healing cross-members of the “H” incisions on the backs of the rats in each of the three groups that received these.

  In this study, Ŝenel and Associates [123] found that allopurinol administration increased the bursting force of the “H” cross-member in comparison with the control group by 52% at 7 days (p < 0.002) and by 109% at 14 days (p < 0.001). The SOD administration, on the other hand, increased the bursting force of the “H” cross-member by 69% in comparison to the control group at both 7 and 14 days (p < 0.003 and p < 0.002, respectively). With respect to the other assessed parameters, the allopurinol-treated group was found to have a 75% greater hydroxyproline content than the control group on Day 7 (p < 0.03); the SOD-treated group was found to have 113% greater hydroxyproline content than the control group on Day 7 (p < 0.001). From the histopathological evaluation, the amount and organization of collagen was more prominent in the allopurinol- and SOD-treated groups than in the ischemic controls. Based on these outcomes, Ŝenel and Associates [123] concluded that oxygen free radicals play an important role in the failure of ischemic wound healing, and that antioxidants partially improve the healing of ischemic skin wounds.

• Stewart et al. [124] conducted studies of random-pattern skin flap healing in rats to determine whether treatment with hyperbaric oxygen (i.e., 2.5 ATA for 90 minutes once per day) and administration of free-radical scavengers (i.e., either superoxide dismutase plus catalase, or α-tocopheral acetate) or a free-radical inhibitor (i.e., allopurinol) would result in increased survival over hyperbaric oxygen, alone. Flap survival was determined by fluorescein injection, and tissue damage was determined by the degree of lipid peroxidation.

  At 7 days post-flap creation, hyperbaric oxygen plus both free-radical scavengers significantly improved flap survival (p < 0.05) while hyperbaric oxygen alone, allopurinol alone, and allopurinol in combination with hyperbaric oxygen did not produce a significantly different outcome than no treatments, whatsoever. At the same point post-surgery, lipid peroxidation associated with the hyperbaric oxygen was prevented by the administration of both free-radical scavengers, but not by the free-radical inhibitor.

  With respect to the effectiveness of the two different scavengers (i.e., superoxide dismutase (SOD) plus catalase, or α-tocopheral acetate), the α-tocopheral acetate was found to be slightly superior. In conclusion, the investigators determined that moderate doses of free-radical scavengers or antioxidants in combination with conservative hyperbaric oxygen can improve the survival of random pattern skin flaps [124] in comparison with no hyperoxic therapy, or hyperoxic therapy administered without such antioxidant defenses.

In addition, the 1990 study by Barth et al. [61] introduced previously in this report, demonstrates that the holes drilled in proximal and distal metaphyses of rat femurs filled in 3 weeks through primary ossification when HBO₂ was administered once a day, in 4 weeks through enchondral ossification without HBO₂, and in 5 weeks through enchondral ossification when HBO₂ was administered twice a day. In prior research conducted by Bassett and Herrmann [125], it was determined that in vitro connective tissue cells, cut free from a clot of such cells and exposed to modestly elevated levels of oxygen (i.e., 35% with 5% CO₂ and balance N₂) would compact and differentiate directly to osseous tissue (primary ossification). Connective tissue cells cut free from the larger mass and exposed to reduced levels of oxygen (i.e., 5% with 5% CO₂ and balance N₂) would compact and differentiate to cartilage (enchondral ossification). This suggests that while one treatment per day of HBO₂ produced the expected result with hyperoxic treatments in the Barth study, rapid filling of the hole through primary ossification, no HBO₂ treatments and two treatments per day both behaved as if exposed to hypoxic levels of oxygen and filled the holes through enchondral ossification. Thus, in the case of two HBO₂ treatments per day in the Barth study, the benefits of the first treatment were more than offset by the second so an outcome less than that of no HBO₂ treatments was achieved. Consequently, the logical explanation for the dose-response relationship shown in Figure 4 and supported by the studies discussed above is that as the treatment dose increases, the impact of the toxic effects of oxygen progressively inhibits, interferes with, and eventually more than cancels out the beneficial effects of hyperoxic therapy for healing the wounded tissue.

In view of the above, the outcome of hyperoxic treatment can be seen as the net result of the various beneficial and adverse effects at the specific wound site, and here, we are referring to “wound” in the broadest sense, so this could be an ulcerating wound such as a Wagner Grade 5 diabetic foot or the mildest of delayed onset muscle soreness (DOMS). This provides for a continuum of results with hyperoxic treatments from non-existent to optimally beneficial to neutral to negative depending on the local balance between all factors involved.

Conflict of interest

The authors are the Founding Partners of Microbaric® Oxygen Systems, LLC, now known as NEMO Therapeutics, LLC, which was established to provide an organization through which research in applications of normobaric oxygen therapy, particularly autism spectrum disorders, could be investigated and, as possible and appropriate, made available as a mainstream medical therapy for autism and other neurological conditions for which there is currently no effective treatment. This organization has been funded in full by the Founding Partners. No outside funding or income has been received to support the research and writing of business-related activities or this review of oxygen dose response.

Acknowledgments

By Russell E. Peterson, Ph.D: I would like to thank my mother, Pauline W. Peterson, whose interest in medicine and the medical sciences led to my career in diving and hyperbaric medicine and physiology. I would also like to acknowledge my graduate school mentor, Christian J. Lambertsen, M.D., whose guidance to believe in one’s own observations rather than the beliefs of others was a critical element in the evolution of this publication.

By Michael W. Allen: I would like to thank my parents, Phyllis P. and William G. Allen, who blessed me with an inquisitive mind and a need for constant learning. I would also like to thank Philip B. James M.D., Ph.D, my early mentor and lifelong friend who fostered my interest in pressure physiology and the potential offered by the use of oxygen in medicine.

Funding statement

The entire literature review, development of concepts, and preparation of this report were funded by the authors. No outside funding was sought or received from any source.

Data availability

All publications cited as journal or book references in this manuscript are available from the National Center for Biotechnology Information, U.S. National Library of Medicine: 8600 Rockville Pike, Bethesda MD, 20894. www.ncbi.nlm.nih.gov/pubmed and other scientific and university databases.